REESE LIBRARY
OF THE
UNIVERSITY OF CALIFORNIA.
Class
i~-''
T\
THE LATE PROFESSOR JOHANN BAUSCHINGER
(See Appendix A.
Frontispiece.
THE
MATERIALS OF CONSTRUCTION.
A TREATISE FOR ENGINEERS
ON THE
STRENGTH OF ENGINEERING MATERIALS.
J. B. JOHNSON, O.E.,
Professor of Civil Engineering in Washington University, St. Louis, Mo. ; Member
of the Institution of Civil Engineers ; Member of the American Society of
Civil Engineers ; Member of the American Society of Mechanical
Engineers ; Corresponding Member of the American Insti-
tute of Architects ; Member of the International
Association for the Standardizing of
Methods of Testing Materials;
etc., etc. ,
FIRST EDITION.
FIRST THOUSAND.
NEW YORK:
JOHN WILEY & SONS.
LONDON: CHAPMAN & HALL, LIMITED.
1897.
<\0
Copyright, 1897,
BY
J. B. JOHNSON.
ROBERT DRUMMOND, ELECTROTYPER AND PRINTER, NEW YORK.
PREFACE.
THE rational designing of any kind of construction involves a knowledge
of
The external forces to be resisted, transformed, or transmitted;
The internal stresses resulting therefrom;
The mechanical properties of the materials to be employed to accomplish
the objects sought.
Of these three coordinate departments of knowledge the first two are
founded on the sciences of mathematics -and applied mechanics. The last
one, however, does not rest on any deductive science, as this information
can only be gained by patient, expensive, and competent research. For
this reason the third essential named above has not kept pace with the other
two kinds of engineering science; but, on the other hand, it furnishes very
much greater rewards to the skilled investigator.
During the past twenty-five years the number of such investigators has
increased from a scattering few to hundreds and even thousands, and these
are now found in all enlightened nations. The results of their original
studies and experiments are pouring in upon us from all countries, in many
languages; and no practising engineer can hope to even scan, much less to
appropriate and assimilate, more than a very small part of this vast wealth of
experimental knowledge.* In the following work the author presents to his
readers a condensed and concise summary of such portions of the knowledge
now available on this subject as he has found suitable for such a work. He
is fully aware of its incompleteness and of its more or less fragmentary
character. Yet with all its faults he believes it contains sufficient reliable
information, not commonly accessible elsewhere, to justify its .publication in
this form.
* The author has one list of original contributions to the subject of the Strength of
Materials filling 140 quarto pages.
Ill
iv PEEFACE.
When the work is used as a text-book in schools of engineering the
instructor would do well to assign only such portions of Part I as are required
to supplement the student's course in applied mechanics; to have his students
read Part II if they do not get this information in other ways; to dwell
longer and with more care on Part III; and to call attention only to such
portions of Part IV as pertains to the particular course the students are
taking. In this way the hook may be made intelligent and familiar to the
student, and so become to him a great and lasting aid in designing, testing,
and inspecting, without requiring more time than can be devoted to the
subject. This course should precede or accompany an experimental course
in the testing laboratory, with which all American schools of engineering
are now equipped.
An unusual use has been made of stress-diagrams and other forms of
graphical representation of facts and laws, no pains or expense having been
spared in this direction. So far as possible tables have been omitted and the
original tabular data have been incorporated in diagrams. A law of relation-
ship cannot be perceived from data arranged in a tabular form. When
plotted to significant arguments the law not only becomes evident at a
glance, but when once impressed on the mind through the sense of sight it
cannot well be forgotten. To obtain this lasting benefit, however, ttr
diagram must be intelligently read and understood. The reader is urged,
therefore, to give great care to the study of all the diagrams which accom-
pany the text on any subject, for, as a rule, the facts, laws, and conclusions
to be drawn from them are not fully expressed in the text. The diagrams
must be considered as a part of the text, and they should be read with even
greater care than is bestowed on the word- embodied ideas.
Throughout the book, with few exceptions, both in the diagrams and in
the text, the English units of weight and measure (pound and inch) have
been employed. The author is of the opinion that until the metric system
has been definitely adopted it is best to use the old unit >, and that a double
system of units is confusing. The revising of books to put them in harmony
with the decimal system will be but a very small part of the total expense
entailed by the formal adoption of that system by our Government. As a
very large part of the data given in the diagrams comes from Continental
sources, all of which were expressed in the metric system, a great amount of
labor was required to bring this material into the English system of units.
Even the results obtained from English sources were generally expressed in
long tons per square inch, so that this also required reduction to bring it to
pounds per square inch.
Some of the author's usages may be regarded as unwarranted innova-
tions. Especially may this be the case in the matter of the new elastic limit,
which he proposes for general adoption, and which is discussed in Arts. 13,
261, 262, and 263. The author bespeaks for these articles a careful consid-
eration and also a study of the many stress-diagrams scattered through the
book, before his views are condemned. The fact is, something must be done
PREFACE. V
in this matter, as now no one knows what is meant by " elastic limit " with-
out an explanation which explanation is not usually given.
The relatively large space given to the subject of timber is not more than
its importance as a structural material, and the general absence of scientific
information on the subject would seem to demand. Probably the reason
little has been given on this subject hitherto, in such works as this, is because
little has been known. Until the Forestry Division of the U. S. Agricul-
tural Department began the systematic study of timber and timber-trees,
some five or six years ago, very little accurate or scientific information was
obtainable as to the mechanical and other properties of American timber.
The author's intimate connection with these investigations is a further reason
why he should here present an adequate account of the work done to date.*
It has been no part of the author's aim to give working rules for using
materials in structures of various kinds, or to propose original specifications
to be used in the purchase of materials. He has tried to impart a knowl-
edge of the properties of materials; on what these depend; the ordinary
causes of variation and defects, and how these should be discovered; thus
making the reader competent to draw his own specifications and to make his
own rules.
The latest forms of investigation of metals and building-stones by means
of the microscope are briefly treated (the former in Appendix B) ; and a
chapter has been given on the magnetic properties of iron and steel, and the
methods to be employed in determining these. This chapter need be read
by electrical students only.
The author has acknowledged his sources of information in the text, and
especially in the legends accompanying the cuts. In addition to these he
desires to make a special acknowledgment here of his obligation to Pro-
fessors Bauschinger, Tetmajer, and Martens; to the French Commission
Report; and to Mr. Henry M. Howe, Prof. Thomas Turner, Mr. G-. R.
Redgrave, Prof. J. Q. Arnold, Mr. Thos. Andrews, Mr. H. H. Campbell, and
to Dr. B. E. Fernow. His thanks are also due to Dr. Wm. Trelease for
assistance in obtaining illustrations of American trees in Chap. XIII; to
Mr. II. A. Wheeler, E.M., for the chapter on the manufacture of paving-
brick; to Mr. W. A. Layman, M.S., for the chapter on the magnetic prop-
erties of iron and steel; and to Prof. H. Aug. Hunicke, E.M., for revising
the manuscript of Chapters VII to XI inclusive.
There are to-day a few exceptionally fertile sources of exact information
on subjects pertaining to the materials of construction, prominent among
which may be named :
1. The annual publications of the Results of Tests made at the U. S.
Arsenal, Watertown, Mass., beginning in 1882.
* The author has had entire charge of the mechanical tests, some 40,000 of which have
"been made in his laboratory at St. Louis.
Vl PREFACE.
2. Bauschinger's Communications from the Laboratory of the Technical
School at Munich, Germany.
3. Tetmajer's Communications from the Laboratory of the University at
Zurich, Switzerland.
4. Martens' Communications from the Laboratory of the University of
Berlin, Germany.
5. The Report of the French Commission (of 115 members) on the
Standardization of Tests of the Materials of Construction, in four quarto
volumes, 1895.
6. The Monthly Journal, Baumateridlenkunde, published in Zurich, as
the organ of the International Society for the Standardization of the Tests
of Materials of Construction.
The entire engineering profession is so indebted to the late Prof. Johann
Bauschinger for the work he has done in developing the scientific testing of
materials that the author of this work has chosen to express his feeling of
gratitude to him by using his portrait as a frontispiece and giving a brief
account of his life in Appendix A.
That this work may contribute somewhat towards more rational, safe,
and economic practices in the designing of all kinds of construction has been
the purpose and is now the hope of
THE AUTHOR.
ST. Louis, Mo., Jan. 1897.
TABLE OF CONTENTS.
PART I.
SYNOPSIS OF THE PRINCIPLES OF MECHANICS UNDER-
LYING THE LAWS OF THE STRENGTH OF
MATERIALS.
CHAPTER I.
GENERAL NATURE OF DEFORMATION AND STRESS.
PAGE
Elastic and Plastic Bodies Stress and Deformation Proportionality of Stress and
Deformation inside the Elastic Limits Kinds of Deformation and Stress
Longitudinal and Lateral Deformation under Direct Stress Angular Defor-
mation under Direct Stress Relation between Shearing and Direct Stresses
Shearing Modulus of Elasticity .* 1
CHAPTER II.
MATERIALS UNDER TENSILE STRESS.
General Phenomena accompanying Tensile Tests Significant Results of Tensile
Tests True Elastic Limit Apparent Elastic Limit Ultimate Strength
Percentage of Elongation Reduction of Area of Cross-section 10
CHAPTER III.
MATERIALS UNDER COMPRESSIVE STRESS.
Two Classes of Engineering Materials Crushing Strength of Plastic or Viscous
Materials Crushing Strength of Brittle Materials Relation . of Crushing
Strength tc Shearing Strength Crushing Strength of Prisms' Relative
Strength of Prisms and Cubes Loading on a Portion of Cross-section only-
General Laws of Crushing Strength Strength of Columns Weakening
Effects of Eccentric Loading 24
vii
Till TABLE OF CONTENTS.
CHAPTER IV.
MATERIALS UNDER SHEARING STRESS.
PAGE
Two Manifestations of Shearing Stress Moment of Torsion Shearing Deforma-
tions 38
CHAPTER V.
MATERIALS UNDER CROSS-BENDING STRESS.
Historical Sketch Fundamental Equations of Equilibrium Moment of Resistance
and Stress on Extreme Fibre Resistance of Beams of Various Forms of
Cross-section Resistance of Beams beyond their Elastic Limits Distribution
of Stress and Position of Neutral Axis at Rupture Moduli of Rupture in
Cross-breaking Distribution of Shearing Stress in a Beam Wooden Beams
in Shearing and Cross-bending Deflection of Beams General Formulae
Various Cases analyzed Table of Moments, Stresses, and Deflections
Deflection from Shearing Forces Determination of Young's Modulus of
Elasticity Rational Designing of Flitched Beams Steel and Concrete in
Combination Flat Plates computed approximately ". . . 42
CHAPTER VI.
THE RESILIENCE OF MATERIALS.
Resilience defined Varieties of A Measure of Shock-resistance Impact Stresses
Resilience Areas in Stress-diagram's Resilience in Direct Stress In Cross-
bending In Torsion Comparative Table 75
PART II.
MANUFACTURE AND GENERAL PROPERTIES OF THE
MATERIALS OF CONSTRUCTION.
CHAPTER VII.
CAST IRON.
General Classification of Iron and Steel Physical Properties of Cast Iron Carbon
in Silicon in Remarkable Effects of Silicon Sulphur, Phosphorus, and
Manganese in Grading Pig Iron Foundry Practice The Cupola Kernel t-
ing Moulds Moulding Sand Size and Shape Shrinkage Mechanical
Properties Hardness Strength in Compression, Tension, and Cross-bending
Malleable Cast Iron Method of Manufacture Mechanical Properties 87
TABLE OF CONTENTS. ix
CHAPTER VIII.
WROUGHT IRON.
Methods of Manufacture The Puddling- Process Oxidation in Puddling Muck
Bars Reheating and Rolling Repeated Reheatings Imperfections in Fin-
ished Product Mechanical Properties Crystalline Fracture Welding
Effect of Reduction in the Rolls on the Strength ; 117
CHAPTER IX.
STEEL.
Methods of Manufacture Crucible Process Bessemer Process Open-hearth Proc-
essBasic and Acid Processes Comparison of Bessemer and Open-hearth
Processes Molecular Structure of Wrought Iron and Steel Structure as
affected by Heat Treatment Mechanical Qualities of Steel Commercial
Classification Quality as determined by Chemical Composition Influence
of Carbon on Iron Three States of Carbon in Iron Change in the Carbon
at a Low Yellow Heat Hardening and Tempering Effects of Carbon on
the Mechanical Qualities On Tensile Strength On Ductility On Compres-
sive Strength Effects of Silicon Of Manganese Manganese Steel Of
Sulphur Red -shortness Sulphide of Iron Dangerous Of Phosphorus On
Ductility On Strength Hardening Tempering Annealing Corrosion. . , 133
CHAPTER X.
THE MINOR OR AUXILIARY METALS OF CONSTRUCTION AND
THEIR ALLOYS.
Copper Zinc Tin Aluminum Nature of Metallic Alloys Copper-zinc-tin
Alloys The Brasses The Bronzes Alloyed Aluminum Aluminum in
Steel Fusible Alloys 172
CHAPTER XL
LIME, CEMENT, MORTAR, AND CONCRETE.
Quick, or Fat, Lime Hardening of Lime-mortar Hydraulic Lime Natural
Cement Portland Cement Historical Account of Ingredients of The
Clay Silica and its Compounds Alumina Sulphur Compounds Chemical
Reactions in the Furnace Chemical and Physical Changes in Setting and
Hardening Slag-cements Sources of Raw Materials for Portland Cement
Processes used in Pulverizing and Mixing Processes used in Burning-
Grinding the Clinker. 181
CHAPTER XII.
THE MANUFACTURE OF VITRIFIED PAVING-BRICK.
Definition of Clays employed Physical Properties of Clays Preparation of the
Clays Moulding Drying and Burning Annealing Sorting 196
X TABLE OF CONTENTS.
CHAPTER XIII.
TIMBER.
PAGE
Structure and Appearance Classes of Trees Sapwood and Heartwood Annual
Rings Spring and Summer Wood Anatomical Structure of Broad leaved
Trees Minute Structure Grains of Wood Color and Odor Resonance
Weight a Function of Structure and Moisture Variation of Weight in Single
Trunk and in Species Moisture Distribution Drying Timber Shrinkage
explained Effects of Shrinkage Amount of Shrinkage Mechanical Proper-
ties Stiffness Strength as a Beam In Tension and Compression In
Shearing Influence of Weight and Moisture on Strength Hardness
Cleavability Flexibility Toughness Practical Conclusions Chemical
Properties and Technological Products Wood as a Fuel Charcoal Prod-
ucts of Wood-distillation Durability and Decay All Decay produced by a
Fungus-growth Prevention of Decay Structure as a Key to Identification
of Species A Structural Key to Species Characteristic Structural Features
Use of the Key Descriptive List of the More Important Trees in the U. S.,
with Illustrations of Leaf and Fruit. . , .205
PAKT III.
TESTING-MACHINES AND METHODS OF TESTING
MATERIALS OF CONSTRUCTION.
CHAPTER XIV.
MECHANICAL TESTS IN GENERAL.
General Observations Tests Classified Testing-machines Effect of Rate of Load-
ing Significant Limits of Deformation All Absolute Elastic Limits
unsatisfactory The " Apparent Elastic Limit" 302
CHAPTER XV.
TENSION TESTS.
Significance of Tension Tests Selection of Test Specimens Preparation of Speci-
mens Standard Dimension of Tension-test Specimens Tetmajer's Analysis
of the Elongation Time Function of Tension Tests Tension-test Machines
Gripping Devices Special Machines The Emery Testing-machine fully
Described Exteusometers Autographic Diagram Apparatus Gauging Im-
plements 312
TABLE OF CONTENTS. XI
CHAPTER XVI.
COMPRESSION TESTS.
PAGtt
Objects of Compression-test Specimens Bedding the Specimen in the Machine
Compressometers Column Tests Strength of Column the same as its
Apparent Elastic Limit Considered Results Tetmajer's Results Formulae
for Strength of Columns Spring Testing-machines 353
CHAPTER XVII.
CROSS-BENDING TESTS.
Object of Essential Conditions of Deflection measured in Testing Cast Iron-
Modulus of Rupture Modulus of Elasticity Impact-testing Machines 369
CHAPTER XVIII.
IMPACT AND HARDNESS TESTS.
Object of Impact Tests Essential Conditions of Energy of the Blow Hardness
Denned Test of Permanency of Form The Rodman Punch Standardized
Test for Permanency of Substance Turner's Apparatus 375
CHAPTER XIX.
SHEARING AND TORSION TESTS.
Essential Conditions of Shearing Tests Occurrence of Shearing Stress in Practice
Shearing-test Appliances Torsion Tests Torsion-testing Machines 385
CHAPTER XX.
COLD BENDING AND DRIFTING TESTS.
Significance of Cold-bending Tests Methods of making them Results of Cold-
bending and Tension Tests compared Effects of Punching and Drilling
developed by Cold-bending Tests Combined Specified Requirements in
Tension and Cold-bending Results of Tension, Cold-bending, and Impact
Tests compared Drifting Tests standardized 394
CHAPTER XXI.
THE TESTING OF CEMENT.
Standard Scientific Tests of Cement Test of Fineness Significance of Fineness-
Thoroughness of Burning tested by Specific-gravity Test Apparatus for
Rate of Setting Automatic Apparatus for Registering Vicat's Needle
xii TABLE OF CONTENTS.
PAGE
Tests for Soundness The Boiling Test Tests of Strength Fixed Relation
between. Tensile and Coin press! ve Strength Standard Consistency of Neat-
cement Briquettes Ell'ects of Varying Percentages of Water Normal or
Standard Sand Effect of using Different Sands Consistency of Standard
Mortar, 1 C : 3 S Formation of the Briquettes Form of the Briquette A New
Form proposed Distribution of Stress over tlie Minimum Section The
Clips Cement testing Machines Eccentricity of Briquette in Clips Cross-
breaking Tests Standard Tests of Adhesion Normal Variations of Volume
of Cement mortars in Air and in "Water Recommendations of the French
Commission for testing Permanency of Volume Test of Permeability of
Cement-mortar Test for Decomposing Action of Sea-water 407
CHAPTER XXII.
TESTS OF THE STRENGTH OF STONE AND BRICK.
Crushing Tests of Stone Tests for Paving-brick The Cross-breaking Test The
Crushing Test The Rattler Test Standardized Standard Tests of Com-
mittee of the National Association of Brick Manufacturers 456
CHAPTER XXIII.
TESTS OF THE STRENGTH OF TIMBER.
Important Deductions from the U. S. Timber Tests Description of the U. S.
Timber Tests Mechanical Tests Cross-bending Test Crushing-endwise
Tests Crushing across Grain Shearing Tension 462
PART. IV.
THE MECHANICAL PROPERTIES OF THE MATERIALS OF
CONSTRUCTION AS REVEALED BY ACTUAL TESTS.
CHAPTER XXIV.
THE STRENGTH OF CAST IRON.
Tensile Strength Composition and Strength of High-grade Cast Iron Compres-
sive Strength Cross-breaking Strength Modulus of Elasticity Kirkaldy's
Results Shrinkage Stresses Strength Increased by Impacts Pipes and
Columns 469
CHAPTER XXV.
THE STRENGTH OF WROUGHT IRON.
Strength with the Grain Strength' across the Grain Time Function Compressive
Strength Shearing Strength Effect of Stressing beyond the Elastic Limit
Strength of Chains '. 482
TABLE OF CONTENTS. xiii
CHAPTER XXVI.
THE STRENGTH OF STEEL.
PAGE
Tensile and Compressive Strength Effect of Varying Percentages of Carbon
Effect of Thickness Effect of Finishing at n Low Red Heat Effects of An-
nealing on Low-carbon Steel Tests of Sleel by Punching Quenching and
.Annealing Billet Tests Characteristic of Final Rolled Forms Elongation
and Reduction Compressive Strength same as the Elastic Limit Elastic
Limit in Compression for Various Kinds of Contact Areas of Contact be-
tween Wheels and Rails Moduli of Elasticity in Tension and Compression
Annealing Effects after Overstressiug Effects of Varying Lengths of
Reduced Section Nickel Steel Effects of Forging and Rolling Steel-
welded Tubes I Beams and Plate Girders Effects of Stressing beyond the
Elastic Limit Shearing Strength Fiictional Resistance of Riveted Joints
Friction per Square Inch of Rivet Section Bearing Resistance of Plates
Tensile Strength of Grooved Plates Injurious Effects of Punching and
Shearing Influence of Form of Thread on Strength of Screw-bolts Steel
Specifications 490
CHAPTER XXVII.
THE FATIGUE OF METALS.
Fatigue Defined Micro-flaws in Steel Wohler's Fatigue Tests Limits of Stress
for an Indefinite Number of Repetitious A New Universal Formula for
Dimensioning 537
CHAPTER XXVIII.
STRENGTH OF THE COPPER-ZINC-TIN ALLOYS.
Strength of Copper Annealing Copper Wires and Plates Strength of Brass
Strength of Bronze Special Bronzes 548
CHAPTER XXIX.
THE EFFECTS OF TEMPERATURE ON THE MECHANICAL
PROPERTIES OF METALS.
Effects on the Strength of Iron and Steel The Change in the Elastic Limit
Change in the Modulus of Elasticity Effect on Resistance to Impact-
Effects on Copper and Bronze 557
CHAPTER XXX.
RESULTS OF TESTS ON CEMENTS, CEMENT-MORTARS,
AND CONCRETES.
Strength of Natural Cements Strength of Portland Cements Modulus of Elas-
ticity of Cement-mortars Strength of Sand cement Mortars Variation of
xiv TABLE OF CONTENTS.
PAGE
Strength with Increasing Proportions of Sand Variation of Strength of
Mortars with Varying Size of Sand-grains Relative Economy of Coarse and
Fine Sand-grains Experiments with Sands of Artificial Grauulometric Com-
positionPorosity of Mortars as Affected by Size of Sand-grains Effect of
Long Storage on the Strength of Cement Effect of Regaugiug after Set
begins Effect of Carbonic-acid Gas on the Hardening of Cement-mortars
Adhesive Strength of Cement-mortars Compressive Strength and Elasticity
of Cement and Concrete Strength and Economy of Cement-mortars and
Concretes Filtration through Concrete Effects of Freezing on Cement-mor-
tars and Concretes Anti-freezing Mixtures Concrete Mixtures Concrete
Structures in Sea-water Fire-resisting Qualities of Concretes Properties of
Cinder-concrete Mixtures Cinder-concrete with Expanded Metal Base 568
CHAPTER XXXI.
RESULTS OF TESTS ON STONE AND BRICK.
The Building-stonesWeathering of Building-stones Freezing Tests The Sul-
phate-of-soda Test Chemical Tests Microscopic Tests The Absorption
Test The Specific-gravity Test Compressive Strength Table of Physical
Qualities of American Building-stones Elastic Properties and Crushing
Strength, with Stress-diagrams Bauschinger's Results Resistance to Abra-
sion Bauschinger's Abrasion Tests and Results Strength and Elastic Prop-
erties of Brick and Brick Piers, with Stress-diagrams Results of Tests of
Paving-brick Results of Tests on Building-brick 630
CHAPTER XXXII.
EXPERIMENTAL VALUES OF THE STRENGTH OF TIMBER.
The Mechanical Tests of the U. S. Timber Investigations List of Species Tested
Ultimate Ends of the Investigation The Moisture Factor Tables of Results
on Thirty-two Species Special Investigations Relation between Strength
and Weight The Factor of Safety Table of Safe Loads on Beams-
Strength of Wooden Columns How to distinguish between Short-leaf and
Long- leaf Pine Lumber Geographical Distribution of Southern Pines
Holding Force of Nails 664
CHAPTER XXXIII.
STRENGTH OF IRON AND STEEL WIRE AND WIRE ROPE.
The Strength of Wire Strength of Steel-wire Rope Methods of Testing the
Strength of Wire Ropes Shop Tests of Wire The Albert-lay Rope 691
TABLE OF CONTENTS. XV
CHAPTER XXXIV.
THE MAGNETIC TESTING OF IRON AND STEEL.
PAGE
Magnetic Properties defined Hysteresis Measurement of Permeability Induc-
tive Methods Traction Methods Measurement of Hysteresis Results of
Tests Development due to Testing Conditions affecting Magnetic Quality
Importance of Magnetic Testing Useful Data on Conductivity 702
APPENDICES.
A. BIOGRAPHICAL SKETCH OP PROF. JOHANN BAUSCHINQER 723
B. STUDY OF IRON AND STEEL BY MICROGRAPHIC ANALYSIS 725
C. COMPARATIVE ANALYSIS OF THE RESOLUTIONS OF THE CONVENTIONS, OF
THE FRENCH COMMISSION, AND THE AMERICAN SOCIETY OF MECHANICAL
ENGINEERS 737
D. STANDARD SPECIFICATIONS FOR IRON AND STEEL.. 756
THE
MATERIALS OF CONSTRUCTION.
PART I.
SYNOPSIS OF THE PRINCIPLES OF MECHANICS
UNDERLYING THE LAWS OF THE
STRENGTH OF MATERIALS*
CHAPTER I.
GENERAL NATURE OF DEFORMATION AND STRESS.
1. Elastic and Plastic Bodies. An elastic body is one which, when
deformed under the application of an external force, will recover its origi-
nal dimensions when the deforming force has been removed. A plastic
body is one which will not recover its original dimensions after deforma-
tion. A body which will fully recover its original dimensions after defor-
mation is said to be perfectly elastic. When it will only partially recover
its original dimensions after deformation it is said to be partially elastic,
and to the extent of its failure to recover its original form it may be said
to be plastic.
All solid bodies are nearly or quite perfectly elastic up to a certain limit
of deformation, beyond which they become partly elastic and partly plastic.
This limit within which the body is nearly or quite perfectly elastic is
called the elastic limit. When deformed beyond this limit the body will
recover a portion of such deformation, the rest remaining as a permanent
change or set. Beyond the elastic limit, therefore, a body may be said to
be partly elastic and partly plastic. Practically all the materials used in
engineering design may be said to be perfectly elastic within certain limits;
and as these elastic limits are well beyond the limit of maximum loading in
* This Part is intended to be supplementary to the matter contained in text-books on
applied mechanics, rather than to replace such courses.
2 THE MATERIALS OF CONSTRUCTION.
practice, it is customary to regard all engineering materials as perfectly
elastic for all practical purposes.
2. Stress and Deformation. The deformation which a solid body suffers
on the application of an external force has commonly been called strain,
but in this work it will be designated deformation simply.* The deforma-
tion which is fully recoverable on the removal of the external force may be
called the elastic deformation. That which remains as a permanent set
after the external force has been removed may be called the plastic defor-
mation. Within t.he elastic limit the deformation is wholly elastic.
The relative deformation is the proportionate distortion, or the linear
change divided by the original length or dimension in the direction of the
deforming force. Thus if a bar 10 inches long be stretched 0.01 inch,
then this 0.01 inch is the deformation, and the relative deformation is 0.01
divided by 10, or 0.001 main. Thus the actual deformation is a concrete
quantity, and is measured in units of length, while the relative deformation
is an abstract number, and may be defined as the ratio of the distortion to
the original length. This relative or proportionate deformation may also
be defined as the deformation per unit of length.
Stress may be defined as the resistance a solid body offers to the defor-
mation produced in it by the action of an external force, and it may also be
defined as the resistance to this external force directly. Under the law
that action and reaction are equal, the stress must be quantitatively equal
to the external force, and it may be regarded as resisting this external
force, or these two may be regarded as being in equilibrium. Since the
application of an external force to a solid body, however, is always accom-
panied by a deformation of that body, and since this deformation disap-
pears on the removal of the external force, the internal stress in the body
may be said to be developed as a resistance to this deformation, and in this
sense the deformation may be regarded as the immediate cause of the stress,
the ultimate cause being the external force.
3. Proportionality of Stress and Deformation Inside the Elastic Limit.
Within the limits of perfect elasticity of solid bodies the deformation is
directly proportional to the external force producing that change of form;
and since the internal stress is of necessity equal to the external force, we
arrive at this important proposition : Inside the elastic limit the stress is
directly proportional to the deformation ivhich accompanies it.\ This
* The word strain is used in common language in several other senses, so that its use
in this specific scientific sense, though warranted, is of doubtful propriety. The author
has so used it, however, in his previous works.
f This law was first announced by Robert Hooke in 1676, in the form of an anagram,
as " The true theory of elasticity or springiness ceiiinosssttuv." Two years later the
key to this anagram was given in the Latin phrase "Ut tensio sic vis," a free rendering
of which would be, "As the extension so is the strength." This law of proportion-
ality, therefore, between the stress and the deformation within the elastic limit is fre-
quently referred to as Hooke's Law.
GENERAL NATURE OF DEFORMATION AND STRESS. 3
proposition may be stated in another way by saying that inside the elastic
limit the stress per unit area divided by the proportional deformation is a
constant for any particular solid body. Since this constant is the ratio of
the deforming force to the accompanying deformation of any particular
solid body, it is evidently an important function, and it has therefore been
given a name. The name of this ratio is the modulus of elasticity. We
have, therefore,
Modulus of elasticity = E = -= -. -, . . . . (1)
deformation
wherein by stress is meant stress in pounds per square inch, and by defor-
mation is meant a proportionate change, or the deformation per unit of
length. Thus if an external pull, P, be applied to a bar whose cross-section
p
is A y then the unit stress is ; and if the length of the bar be I, and its
actual extension under the application of this external force be a, then the
deformation, or proportionate distortion, would be y, whence we should have
t
a Aa a"*
7
wherein p is the stress in pounds per square inch. Thus if the external
force of 60,000 pounds be applied to a bar whose length is 10 inches and
whose cross-section is 2 square inches, and if the extension under this load
be 0.01 inch, we should have
That is to say, such a material would have a modulus of elasticity of
30,000,000; and this is about the average value of the modulus of elasticity
of steel.
Example. If steel rails be welded together at a temperature of 80 F., what
will be the total tensile stress in an 80-pound rail at a temperature of 20 below
zero, and what will be the compressive stress in this rail at a temperature of 140 F.,
the coefficient of expansion being assumed as 0.0000065 per degree F.?
In solving such a problem as this, since the length of the rail cannot
change for a change of temperature, the contraction which would occur if free
to move is overcome by the application of a sufficient external force coming
from the surrounding bodies to prevent this contraction. In other words,
an external force is developed just sufficient to stretch the body as much as
it would contract under a fall of temperature, and similarly an external
force is exerted to compress the body as much as it would expand under a
4 THE MATERIALS OF CONSTRUCTION.
rise of temperature. We have then only to determine the amount of the
contraction or expansion from temperature and call this the deformation
produced by the application of an external force, and then by the aid of the
modulus of elasticity find the amount of this external force, and divide it by
the area of the cross-section of the rail, thus obtaining the internal stress in
pounds per square inch. The cross-section of the rail is indicated by its
weight. The weight of rails is always given in pounds per yard, and it so
happens that a bar of iron or steel one inch square and 36 inches long
weighs just ten pounds. This unit is called an inch-yard. Therefore an
80-pound rail has just eight inches of cross-section. With the above
information the student is prepared to solve the problem. It is evident that
the length need not be considered; or any length may be chosen, as, for
instance, one inch, since only the proportionate change of length need be
considered in either case. The answers to the problem are 156,000 pounds
total stress in tension at the lower temperature and 93,600 pounds total stress
in compression at the upper temperature, the stress per square inch being
19,500 pounds in tension and 11,700 pounds in compression, respectively.
Since the elastic limit in both tension and compression of this grade of steel
is about 45,000 pounds per square inch, it is evident that these stresses are
well within these elastic limits, and hence no injury to the rail would ensue
from the prevention of expansion and contraction in this manner.
4. Different Kinds of Deformation and Stress. Under the application of
suitable external forces there are commonly recognized five kinds of defor-
mation, namely: Extension, Compression, Angular, Bending, and Twisting;
and corresponding with these are five kinds of stress, namely: Tensile,
Compressive, Shearing, Bending, and Tortional. The last two kinds of
stress are really combinations of the other three. Thus, bending stress may
be resolved into tension and compression, with or without shearing, and a
tortional stress is a particular kind of shearing stress. For any particular
kind of material there is a definite relation between these several deforma-
tions and their corresponding stresses. The numerical values of the ratios
of these corresponding deformations and stresses are the moduli of elasticity
in the several cases. It so happens, however, that the modulus of elasticity.
or the ratio between the stress and the deformation in tension, is usually the
same as it is in compression. Both tension and compression are called direct
stresses, and hence we may in general speak of the modulus of elasticity in
direct stress, and the modulus of elasticity in shearing, in cross-bending, and
in torsion. Since cross-bending distortion gives rise mostly to distortion
in extension and compression, and their corresponding stresses, the modulus
of elasticity in cross-bending may also be said to be the same as that in
direct stress.
The modulus of elasticity, therefore, which is used in tension, compres-
sion, and cross -bending, is one and the same, and is sometimes spoken of as
Young's modulus. That is to say, it is the ratio between direct stress in
GENERAL NATURE OF DEFORMATION AND STRESS. 5
pounds per square inch and the corresponding proportionate linear defor-
mation.
5. Longitudinal and Lateral Deformation under Direct Stress. The lon-
gitudinal deformation of a solid body in the direction of the deforming force
is A/, where I is the original length in this direction and A is the proportion-
ate deformation. Hence we may write, for Young's modulus,
_ stress per unit area _ p
deformation per unit length A* ^ '
It is a fact of observation that when a metal body is elongated by an exter-
nal force from ltol-\-\l (inside the elastic limit), it contracts laterally about
one fourth of its proportionate elongation. Hence if the original diameter
were d, its diameter after stretching would be d -d. This ratio of lateral
4)
to longitudinal deformation., under longitudinal external forces, is called
Poisson's ratio. It is usually taken as J- for all metals, but for india-rubber
it is J. The true values of this ratio, for some of the more common mate-
rials, are : *
Glass .... 0.2451
Steel .... 0.2686
Copper . . . 0.3270
Brass 0.3275
Delta-metal .... 0.3399
Lead 0.4282
6. Change of Volume under Direct Stress. If the length of the body is
incfeased by \l, and its lateral dimensions are decreased by jAfZ, the new
volume for a rectangular bar having lateral dimensions of b and d would be
But the original volume was Ibd, hence the change of volume is Ibd-
2
and the relative change is Ibd divided by the original volume = -, or the
* &
volume has been increased by one half as great a percentage as the length
was increased.
If we should now apply an equal direct tension in the direction of b, we
would increase this dimension by A#, and the volume by -J,bd 9 and similarly
for a tensile force in the direction of d. Hence for a direct tensile force in
* Taken from Wertheim and given in the Report of the French Commission des
MetJtodes d'Essai des Materiaux de Construction, 1895, vol. in. p. 6.
f Since A is very small as compared to unity. The product of (I -f- m)(l -f- ri)(l -f- p),
etc., where m, n, and p are very small fractions, is I -f- (m + n -f p), since the products
of the auxiliary terms can be neglected.
G
THE MATERIALS OF CONSTRUCTION.
all three planes the volume would be increased by |A times its original vol-
ume, and each dimension by -J/\ times its original measure.
For a compressive force in all directions the volume would be diminished
to (1 | A,) times its original volume, and each lineal dimension to (1 JA)
times its original measure..
The volumetric change of a solid body for an equal stress applied in all
directions is therefore f of the change of the dimension in the direction of
an equal simple longitudinal stress. Thus the longitudinal proportionate
deformation for a direct stress of p pounds per square inch is A, or
-But since the relative volumetric change for stress in all directions is
f A, we have as the ratio between volumetric stress and deformation under
an equal stress in all directions, as a fluid pressure for instance,
whence
= E.
That is to say, the volumetric modulus of elasticity of a solid body for an
equal stress in all directions is f of Young's modulus, which applies only to
direct stress in one plane and its accompanying deformation.*
7. Angular Deformation under Direct Stress. We will here consider one
case only of angular deformation under
direct stress, and that is for equal di-
rect stresses of opposite signs on planes
at right angles to each other, as shown
in Fig. 1. If the original length of
each side of this cube be /, then the
dimension in the direction of l^ will be
increased as much by the action of the
vertical compression V as it will be
diminished by the action of the hori-
zontal tension H, since V = H in
pounds per square inch. Also the
cube will be shortened in the direction / 3 by an amount XI due to the force
* This statement applies only to bodies in which Poisson's ratio is . Since this
ratio is very nearly for india-rubber, it follows that the cross-section is reduced as
much as the length is increased, under a tensile stress m one plane, and hence the
volume remains unchanged. Similarly, for a compressive stress in all directions the
volume is unchanged (almost); so that while Young's modulus of elasticity for this
material is very small, the volumetric modulus is very great: and if Poisson's ratio were
quite \, the volumetric modulus would be infinite, or it would tje quite incompressible.
It is probably the most incompressible of any known substance.
GENERAL NATURE OF DEFORMATION AND STRESS. 7
V, and by J- this amount due to the lateral force H. Also the dimension in
the direction , will be elongated by A from the action of the horizontal
force H, and by this amount from the force F. Hence the final dimen-
sions in these directions will be 1(1 |A) and 1(1 -j- H) respectively.
If in the front face of this cube the lines ABGD be drawn, joining the
middle points of the edges before deformation, this figure is a square. After
deformation, if we make the point at A common to the two figures, we
have the points B, C, and D moved to B', C', and D' respectively. This
produces an angular movement of one of these lines equal to the angle
BAB' , which we will call 6. This is now one half the deviation of the
angles B'AD', B'C'D', AB'C', and AD' Q' from right angles.
But since BG GB f = J(J A/), we may assume that B' falls on BC,
since is very small.
Also,
BB' ,, . .BG
tan = = (from similar triangles) r-=, = * . ' = f A.
-?>(/
Or, since 6 is small, we may say,
6 = f A, where is given as arc in terms of the radius as unity.
But 8 = J the deviation of the angles AB'C', B'C'D', etc., from right
angles. Hence we have
20 = angular change = 2(f A) ...... (5)
equals twice the linear deformation.
That is to say, two direct stresses at right angles to each other and of
opposite signs produce in the plane of the stresses an angulnr deformation
equal to twice the proportionate linear deformation. This result will be
used in Art. 9 in obtaining the ratio of the modulus of elasticity in shearing
to that in direct stress.
8. Relation between Shearing and Direct Stresses. In Fig. 2 let the
square ABGD represent a very small portion
of a longitudinal section of a body, taken in
the plane of the forces. Assume also that
there are shearing forces acting on the body,
which have developed at this point in this
plane a shearing stress on the vertical sides /
equal to s, pounds per square inch, these
forming a couple and producing a turning
moment. Evidently this particle can only
be held from turning in this plane by the
development of an exactly equal shearing
stress (or resistance) on the horizontal faces,
which will give an opposing couple and
moment of resistance equal to the turning moment of the origina. shearing
s,
8 THE MATERIALS OF CONSTRUCTION.
forces. If the lengths of these sides be equal, we shall then have s a = s 1 ir
pounds per square inch. Hence we may say:
A shearing stress in one direction at any point in a body develops an
equal opposing shearing stress at right angles to it in the plane of the
resultant external forces.
But the two sets of shearing forces indicated in the figure will tend to
deform the body by elongating it in the direction BD, and shortening it in
the direction AC. The internal resistance to such a deformation develops
in the body a direct tensile stress or resistance along the line A C and a
compressive stress along the line BD.
If S l = $ 2 represent the total shearing stresses on the vertical and hori-
zontal sides of this particle, respectively (s, and *, being equal intensities of
stress, or stress in pounds per square inch), then we may resolve these along
the diagonals and obtain
total tensile stress on AC = Vs* + S 9 * = Vs'33* + 8
= sAC= T
= total compressive stress on BD S? -j- S* = (?,
or these two direct stresses also are equal.
But the stress per square inch is the total stress divided by the area
over which it acts; hence AVC have for the intensities of the tensile and
compressive stresses
Hence we have the larger conclusion that
A shearing stress in one direction at any point in a body develops an
equal opposing shearing stress at right angles to it in the plane of the exter-
nal forces, and these opposing shearing stresses produce two opposing direct
stresses 'acting at 45 with the shearing stresses and at right angles to each
other, these tensile and compressive stresses having the same intensities, in
pounds per square inch, as the original shearing stress.
9. The Shearing Modulus of Elasticity. The modulus of elasticity in
shearing may be defined as the ratio of the shearing stress in pounds per
square inch to the accompanying angular deformation. By angular
deformation is here meant the angular change, as derived in Art. 7, where
20 is a pure ratio, being the ratio of arc to radius. From the last article
we know that a shearing stress gives rise to direct stresses at right angles to
each other, of opposite signs, but of equal intensities; and when such stresses
act, we learned in Art. 8 that the proportionate angular change was twice
the proportionate linear change when equal direct stresses were acting at
right angles to each other. But when both of these stresses were acting
we found the linear change to be f A, or f that due only to the deforming
GENERAL NATURE OF DEFORMATION AND STRESS. 9
force in that direction; and, as found in equation (5), 20 2(fl), we have
26 = f A.
But E = Young's modulus of elasticity j-,
o
and E 8 shearing modulus of elasticity ~^
_ shearing stress per square inch
angular deformation
But we have shown, when s = p, 20 = f A,; hence we have
That is to say, E s = \E, or ^e shearing modulus of elasticity = f of the
linear or Young's modulus.*
* This conclusion is based on a value of Poisson's ratio of . The general relation
between E s and E is E s = h = - r~\^ where m is the reciprocal of Poisson's ratio.
Z'J ,i 771 -f- 1
Thus if this ratio be i, which it is approximately for brass and copper, then m = 3 and
E s = %E, while for india-rubber, where m = 2, we have E s = ^E. Prof. Bauschiuger's
tests on round bars of steel give E s = 13,600,000, while for square bars of the same
s material he found E s = 11,500,000, thus showing a failure of the theory to harmonize
results on these two forms of cross-section even inside the elastic limit. See Rep.
French Commission, vol. m. p. 208, for Bauschinger's results.
CHAPTER II.
MATERIALS UNDER TENSILE STRESS.
10. General Phenomena accompanying Tensile Tests. When a body of
uniform cross-section is subjected to the action of an external force which
tends to pull it asunder, it is elongated in the direction of this force by a
proportionate amount equal to the average force per square inch divided
by its modulus of elasticity; thus
A = the proportionate elongation = ^,
where p is the external force, or internal stress, in pounds per square inch,
and E is the modulus of elasticity (Young's modulus).
At the same time its lateral dimensions are reduced by one fourth as
great a percentage as that which represents the proportionate elongation, as
described in Art. 5. This rate of elongation in the direction of the force,
and contraction in its transverse dimensions, continues in strict proportion
to the amount of the external force, until the elastic limit is reached, when
both the longitudinal elongation and the transverse contraction begin to
increase at a more rapid rate, until finally, with the more ductile metals,
the condition of perfect plasticity or viscosity is reached, and the body
elongates under a constant force, while the lateral dimensions reduce more
and more, until rupture finally occurs.
If the external force or load, in pounds per square inch, be represented
by vertical ordinates, and the corresponding elongations be represented by
horizontal abscissae, then the action of the specimen under test may be
indicated by what is known as a stress-diagram, the vertical coordinates
representing stress, and the horizontal coordinates the corresponding defor-
mations. In Fig. 3 such stress-diagrams are shown for timber, ca3t iron,
wrought iron, and steel. These lie on the upper side of the horizontal
axis. If the same materials were to be subjected to compressive external
forces, corresponding stress-diagrams might be drawn in opposite direc-
tions, that is to say, downward and to the left, as indicated in Fig. 3, below
the horizontal axis.
In a complete stress-diagram of a ductile metal there are four signifi-
cant points which need to be noted. These are: the true elastic limit,
the apparent elastic limit, the ultimate strength, and the breaking-point*
10
MATERIALS UNDER TENSILE STRESS.
11
These four significant points in a tension stress-diagram are indicated by
the letters A, B, C, and D in Fig. 4, where the same diagram is drawn to
widely different horizontal scales.
Thus the point A is the true elastic limit, or the ratio of the stress to
the deformation is a constant from the origin to this point. This requires
that the stress-diagram should be a perfectly straight line from to A.
Beyond the elastic limit, or above A, the deformation sometimes increases
somewhat more rapidly than it did belew A, and the locus then becomes
somewhat curved from A to B. At B a very marked change occurs in the
FIG. 3. Typical Stress-diagrams of Timber, Cast Iron, Wrought Iron., and Steel in
Tension and Compression, drawn to the same scales.
specimen in the case of wrought iron and structural steel. If the test be
continued slowly at this point, it will be found with the more ductile
metals that the specimen elongates a considerable amount under a nearly
constant load, as shown in the diagram, from B to B' . This point is
called the " apparent elastic limit," or the " yield-point " or the " breaking-
down point." In ordinary commercial testing of. iron and steel this point
is always called the " elastic limit;" and the true elastic limit, or the point
A, is not found. This results from the rapid and somewhat crude methods
12
THE MATERIALS OF CONSTRUCTION.
used in making commercial tests, and the author of this work has some-
times called this " apparent elastic limit " the " commercial elastic limit,"
since it is the so-called " elastic limit " found in practically all the tests
made by American inspection bureaus and rolling-mills. Since this yield -
point has been so long regarded as the "elastic limit," whereas the point
A is the true elastic limit, persons who wish to be accurate and at the same
SCALE FOR MAGNIFIED DEFORMATION
005 0.1 0.15 0.2 0.25
JO* 15* 20#
DEFORMATION
FIG. 4.
25%
time to be understood find difficulty in conveying their meaning.* The
terms "yield-point" and "breaking-down point" are not in common use,
while the term " elastic limit " is commonly misused. In the present state
of knowledge on the subject, therefore, the terms " true elastic limit " and
" apparent elastic limit " probably would best describe the points A and B
respectively.! It has been the practice of the author, in making tests to
be used commercially, to call the point B the " elastic limit," without any
explanation or exception, when he desired his results to be comparable
with those made elsewhere for commercial purposes.
Just what happens to the specimen at the point B is well shown on
Plate I, I which is a reproduction of a photograph of specimens of polished
* Fortunately, in the case of soft, or structural, steel these true points are practically
identical, so that in this material no such distinction of terms as is here proposed are
necessary. See Figs. 5, 6, 7, and 8.
f The French Commission use this term " apparent elastic limit " for the point B.
j The author has not seen elsewhere as clear indications of the action of such ma-
terials M the ."yield-point." The tests shown on Plate I were made by him and photo-
graphed in March, 1892. The bars were polished to a mirror surface before testing.
These photographs were exhibited at the Engineering headquarters at the World's Fair,
hicago, 1893, and while they were much observed and studied, it did not appear that
^ny one had ever seen such clear " breaking-down " indications before. The significant
fact is that these effects come instantly, as to any particular marking, and that they sue-
PLATE I.
PHOTOGRAPHS OF A POLISHED STEEL BAR, 1 IN. x 2 IN., AFTER BENDING AND AFTER
PULLING, SHOWING THE " BREAKING DOWN " OF THE METAL.
The tensile test was interrupted before the breaking-down action had extended
entirely throughout the length of the bar. (Tested and photographed by the author,
1892.)
MATERIALS UNDER TENSILE STRESS.
steel subjected respectively to a uniform bending moment and to a tensile
stress. This photographic reproduction shows how the tension specimen
fails or " breaks down " its molecular arrangement in detail by shearing on
inclined sections, beginning at the end of the specimen where it was held
a/si
0.S0M- /.00
FIG. 5. Autographic Stress- diagrams of Mild Steel, taken simultaneously with the
Gray Exteusometer Apparatus. Time, %^ minutes.
by the grips. The breaking down proceeded from the ends towards the
centre. In this case the test was stopped before it had reached the middle
portion. This central portion, therefore, is in its original or normal con-
dition, while the remaining portions have been broken down in an irregular
ceed each other regularly along the bar, like the formation of ice-crystals on freezing
water. The markings on the tension bar, or on the tension side of a beam, are depres-
sions, while on a compression bar, or on the" compression side of a beam, they are
swellings.
14
THE MATERIALS OF CONSTRUCTION.
weblike pattern. If the test had been continued, this action would have
gone on from the ends towards the centre, until the entire specimen had
yielded in this manner; and when this breaking-down action had developed
over the entire length o| the specimen, the point B' in the diagram would
have been reached. This breaking-down action, therefore, all occurs over
40000
340001
20,000
v/rr/
'MT/CW
at as
0.3
FIG. 6. Typical Stress-diagram of Mild Steel, plotted to two scales. (From
records of Tests of Metals, Wat. Ars., 1886.)
the entire length of the specimen between the points B and B', and the
reason why B stands above B' seems to be that it requires a greater force to
start this breaking-down action than is necessary to continue it and extend
it throughout the length of the specimen after it has once been started.
See Figs. 5, 6, 7, and 8. In Figs. 7 and 8 the true elastic limit is well
above the yielding resistance of the metal, or the point A is above B.
MATERIALS UNDER TENSILE STRESS.
15
PlG. 7. Tensiou Tests of Wrought-iron Shafts 1 in. in diam., used for endurance testSc
Average ultimate strength = 50,400 Ibs. per sq. in.; average elongation 27$ on a
length of 11 diameters. ( Wat. Ars. Rep., 1890.)
16 THE MATERIALS OF CONSTRUCTION.
After this breaking-down action has extended over the entire length of
the specimen, a further increase in the load will continue to stretch the
specimen nearly uniformly throughout its length, with a uniform reduction
in cross-section, until at last thf elongation and reduction continue under a
constant load. That is to say, the stress-diagram becomes horizontal at the
point C, Fig. 4. This marks the load under which the material is perfectly
plastic or viscous, or for which the distortion continues with no increase of
load.
MATERIALS UNDER TENSILE STRESS. 17
After passing the point C the specimen begins to show the marked
reduction of cross-section at a particular point, which will ultimately be
the plane of rupture. This action is indicated in Fig. 10. As soon as this
"necking-down" begins, the elongation continues under a diminishing
load, as shown by the dropping of the locus in the stress-diagram, and the
remaining portion of the elongation of the specimen nearly all occurs in
this immediate vicinity. The area of cross-section becomes less and less,
until at rupture it is perhaps less than half the original area, as shown
in Fig. 10.
11. The Significant Results of a Tensile Test. There are five signifi-
cant results of a tensile test, namely:
The Modulus of Elasticity;
The Elastic Limit;
The Ultimate Strength;
The Percentage of Elongation;
The Reduction of Area of Cross-section.
The Modulus of Elasticity is found by dividing any stress per square
inch below the elastic limit by the corresponding proportionate deforma-
tion. Since the stress-diagram is a straight line from the origin to the
elastic-limit point, any point on this portion of the locus may be selected
for the determination of the modulus of elasticity. For instance, if the
point which represents an elongation of 0.1 of one per cent be chosen, the
deformation being 0.001 (see Fig. 6), the modulus of elasticity is found
at once by multiplying the corresponding stress in pounds per square inch
by 1000.* In other words, the modulus of elasticity is the tangent of the
angle which that portion of the stress-diagram below the elastic limit forms
with the horizontal axis when the two coordinates are properly evaluated
by the vertical and horizontal scales respectively.
It is a very remarkable fact that the modulus of elasticity of all grades
of wrought iron and rolled steel, from the softest up to the highest grade of
spring steel, is nearly constant, and has a value from 29,000,000 to
30,000,000, being perhaps always within the limits of 27,000,000 and
31,000,000 pounds per square inch. The ultimate strength of these metals
varies from about 45,000 to several hundred thousand pounds per square
inch for the strongest steel wire; but through this range of variation of
strength the ratio of the stress to the corresponding deformation re-
mains nearly constant. The modulus of elasticity is, therefore, a very
valuable quality of such materials and one which is made great use of in
* If the diagram is not straight to this point (has its elastic limit below this point),
then draw a tangent to the diagram at the origin, aud note where it ciits the ordinate
marking a deformation of 0.001, aud this stress multiplied by 1000 is the modulus of
elasticity. This modulus can be read off in this manner from any of the stress dia-
grams for tension and compression found in this work.
18 THE MATERIALS OF CONSTRUCTION.
engineering design. It may be called the modulus of stiffness, since it is a
direct measure of the rigidity of a body, or an inverse measure of its flexi-
bility.*
12. The True Elastic Limit is, in general, from 50 to 70 per cent of the ulti-
mate strength of the material, while the apparent elastic limit is from GO to
70 per cent of the ultimate strength of the material. The apparent elastic
limit, or the breaking-down point, is also the ultimate strength for practical
purposes, since almost all materials lose their value in structural designs
after they have been deformed beyond this limit.
The true elastic limit may be defined either as the deformation where
permanent set begins, or as a point beyond which a given increment of load
produces a greater increment of deformation, which is the point where the
ratio of the stress to the deformation ceases to be a constant and begins to
diminish. This is also the upper extremity of the straight portion of the
stress-diagram. If a material like wrought iron or structural steel be loaded
beyond its true elastic limit, and even beyond its yield-point, and the load
removed, the material has been permanently elongated; but if it again
be subjected to a load, it will be found to be perfectly elastic up to the limit
of its previous loading. In other words, its elastic limit has been raised to
the value of its previous loading. In this way the elastic limit can be raised
practically up to the ultimate strength. When the term " elastic limit " is
used in a scientific sense without modification, the true or primitive elastic
limit (point A, Fig. 4) is always to be understood; but when used in a com-
mercial sense, the apparent elastic limit or yield-point (point B) is to be
taken.
As stated previously, the elastic limit is usually found in commercial
testing by noting the action of the weighing-beam in dropping under an
increasing stretch, this being in fact the breaking-down point. To determine
the true elastic limit it is necessary to use very delicate measuring appli-
ances, which will enable the observer to discover when the ratio of stress
to deformation has begun to change. Even when using such devices the
readings must be plotted to a large scale to detect the deviation from a
straight line.
13. " The Apparent Elastic Limit" is defined by the French Commission
as " the load per square millimeter of the original section, where the defor-
mation begins to increase sensibly with no increase in the external force
applied (corresponding to the dropping of the beam in testing-machines)."!
Since in most kinds of materials there is no such point other than the ulti-
* A modulus of flexibility would be the reciprocal of the modulus of elasticity, or .
IL
but Prof. A. B. W. Kennedy has taken for such a modulus of "specific extension" the
stretch in thousandths of an inch on a length of 10 inches under a stress of 1000 Ibs. per
sq. in. Its reciprocal multiplied by 10,000,000 is the modulus of elasticity,
f Report of the French Commission, vol. i. p. 207.
MATERIALS UNDER TENSILE STRESS.
19
mate strength, and since in these materials an elastic limit corresponding to
sensible deformations is required for practical purposes, the author proposes
to extend the meaning of this term so as to make it applicable to all elastic
materials, and at the same time to make it serve as the " elastic limit " to
be universally used in all kinds of practical tests. For this purpose he
employs the following definition :
The apparent elastic limit is the 2wint on the stress-diagram of any
material, in any kind of test, at which the rate of deformation is fifty per
cent greater than it is at the origin*
This point is found either by comparing increments of deformation with
given increments of load, or better by plotting the stress-diagram and draw-
ing a tangent to it which has an inclination to the vertical 50 per cent greater
than has the tangent to the diagram at the origin, as shown in Fig. 9. To
/V
0..
l/M.
T /
'M
O N
&'5
ft
A/-C
\
F IG . 9. Stress- diagram of Hard-drawn Steel Wire. ( Wat. Ars. Hep., 1890.)
do this lay off the tangent to the curve at the origin (or inside the true
elastic limit, where it is a straight line), and then fix a point on any hori-
zontal line of the paper, 50 per cent farther from the vertical axis than the
point where this tangent cuts it. Lay a parallel ruler on this point and the
origin, and move it till it becomes tangent to the curve and draw the tan-
* This definition should not be made to apply, however, to materials not perfectly
elastic within any limits. Thus certain stones and concretes have stress-diagrams which
are reversed curves, their rates of deformation being greater at first than after they are
heavily loaded, and any load produces a permanent set, as shown in Chapter XXXI.
Here the modulus of elasticity is different for every increment of load, and no kind of
"elastic limit" can be attributed to them. That is to say, they are not perfectly
elastic for any load however small.
20 THE MATERIALS OF CONSTRUCTION.
gent line. Then fix the point of tangency by the eye, and call this the
apparent elastic limit.*'
This fixes a point which in all cases corresponds to an extremely small
permanent deformation. In Fig. 9 the permanent deformation at this
limit, for hard-drawn steel wire, is about 0.0003 of the length, or T fg- of one
per cent, while the limit so fixed is some 22,000 Ibs. per sq. in. above the
true elastic limit. Although this test was made at the U. S. Arsenal at
AVatertown, Mass., and on the Emery testing-machine, with extreme accu-
racy, as shown by the accordance of the results when plotted to the large
scale in Fig. 9, yet the " elastic limit " as set down in the published record
(which " elastic limit " is supposed to be the " true elastic limit ") lies some
8000 Ibs. per sq. in. higher than the " apparent elastic limit " fixed by the rule
here laid down! This same state of affairs is shown in numberless cases in
the recorded results of tests made at this the most fruitful and accurate labo-
ratory in the world. f While, therefore, objection would be quickly raised
to the criterion herein proposed for fixing an " apparent elastic limit-" in so
arbitrary a manner, and apparently so far beyond the " true elastic limit," yet
no one would be inclined to question the records of the U. S. Watertown
Arsenal tests, in the fixing of a "true elastic limit," even though this should
in nearly all cases lie beyond this conventional "apparent limit "! After a
great deal of thought and research given to this subject, the author believes
no better criterion can be found for fixing a practical " elastic limit" which
will be one and the same limit for a given material in the hands of all ex-
perimenters, and on all machines. For all materials which have a definite
" yield-point " this " apparent elastic limit," determined as here described,
will agree with it exactly; but for such materials it would never be deter-
mined in this manner, since it is then so much more readily found by the
" drop of the beam," or even by a pair of dividers set to given marks on the
specimen. For all materials which have no point of "yielding under a
fixed load" at this stage of the test, this criterion would always accomplish
the following results :
1. It would always fix one and the same well-defined point.
2. This point would always correspond to so small a permanent defor-
mation as to be, for all practical purposes, the true elastic limit.
3. It is equally applicable to all materials which have an elastic field.
4. It is equally applicable to all kinds of tests, whether on specimens or
on finished members or structures, where deformations of any kind can be
correctly measured.
While the 50 per cent increase in the rate of deformation is purely arbi-
*The author has done this in his U. S. timber tests since 1891, calling this point in
his cross-bending stress-diagrams "the relative elastic limit."
f See other instances in records selected therefrom for this work in Chapters XXV
and XXVI.
MATERIALS UNDER TENSILE STRESS. 21
trary, it is not large enough to fix a point having an appreciable permanent
set, but it is large enough to fix a well defined point on the stress-diagram.
A very extended experience in its application, therefore, serves but to con-
firm the author in its continued use, and in the recommendation of its gen-
eral adoption which is here put forth for the first time. *
14. The Ultimate Strength of a specimen subjected to tensile stress is.
measured by the maximum load carried, and is indicated on the stress-
diagram by the true maximum point in that curve. It is found by dividing^
the maximum breaking load by the original area of cross-section. In case
of the more plastic metals, the area of the broken section is usually about
one half the original area, so that the ultimate strength of the actual section
at rupture when found by dividing the breaking load by the final area of
this section would be about twice the ultimate strength as computed on the
original section. That is to say, the drawing down and pulling out of the
metal has nearly doubled its strength per square inch. The term " ultimate
strength" however, ahvays refers to the original section, and is found bij
dividing the maximum load by the original section.
15. The Percentage of Elongation is found by dividing the increase of
length after rupture has occurred, by the original length. By original
length is meant a certain portion of the specimen which has been reduced
to a uniform cross-section before testing. A standard length for tensile-'
test specimens in America and in England is eight inches, while in
Germany and France it is twenty centimeters, these standard lengths being
practically identical. The elongation of a test specimen of the plastic
metals may be divided into two portions: (a) that part of the elongation
which is uniformly distributed over the section; (I)) that part of the
elongation which occurs in the vicinity of the section which finally breaks.
Thus in Fig. 10 are shown four sets of test specimens of mild steel, there
being three specimens in each set. All the specimens of one set were
originally of the length indicated by the untested specimen which stands
on the left side of each group. The specimen next adjoining it on the
right has been stretched to the limit of the elongation indicated in (a)
above, or until there is an indication of a local reduction of area. The
right-hand specimen in each group shows the local elongation and reduc-
tion, but the specimen has been removed from the testing-machine before
rupture occurred. The middle specimen of each group has been tested to-
the ultimate strength of the material, since, when the specimen begins to-
reduce locally, the ultimate strength has been passed, and the strain-diagram
begins to fall, or it is developed under a diminishing load.
By the amount, therefore, that the right-hand specimen in each of these
groups is longer than the middle specimen of the group, by so much has
the length been increased by the local drawing out on the section where
failure will finally occur. The first elongation, therefore, is that portion
which is uniformly distributed over the specimen, and the second is that
* See also Arts. 261, 262, and 263, pages 306-311.
THE MATERIALS OF CONSTRUCTION.
MATERIALS UNDER TENSILE STRESS. 23
which is concentrated in the vicinity of the final failure. Both of these
elongations are, however, measured and included in the total elongation,
from which the percentage of elongation is determined. The total elonga-
tion is obtained after rupture has occurred, by placing the two ends together
and measuring the distance between the primitive gauge-marks. In the
case of specimens having shoulders at their ends the gauge-marks should
be at least one-half inch inside of the shoulder, since the metal adjacent to
the shoulder does not elongate fully, because of the strengthening effect
of the enlarged cross-sections at the ends.
It will at once be apparent from a study of these specimens that the
(1} elongation, or that which is locally developed in the vicinity of final
rupture, is nearly the same in all these specimens; whereas the (a) elonga-
tion, or that which is uniformly distributed over the specimen, is always
directly proportional to the length. The total elongation, therefore, will
not be proportional to the length. In other words, the percentage of total
elongation will be greater for the short specimen than for the long ones.
This shows the necessity of using standard lengths of these specimens when
the percentage of elongation is to be found.
The percentage of elongation is the result which indicates the ductility
of the material, this being one of the most important qualities of the
metals used in structural designing.
16. The Reduction of Area of Cross-section is found by determining the
area of the broken cross-section, subtracting this from the original area of
cross-section, and dividing the difference by the original area. This is not,
so important an indication or result as the others described above, but it is
customary to determine it, and to add it to the record. For the ductile
metals this reduction of area may be as much as from fifty to sixty per cent,
of the original cross-section.
CHAPTER III.
MATERIALS UNDER COMPRESSIVE STRESS.
17. Two Classes of Engineering Materials. Engineering materials may
be divided into two general classes, according to their manner of failure in
compression.
Plastic or viscous materials are those which will flow without showing
any other indication of failure under a sufficient compressive load.
Brittle or comminuible materials are those which will crush to a pow-
der, or crumble to pieces, or fail by shearing on definite angles under a
compressive load.
In the former class are such materials as wrought iron, soft and medium
steel, the alloys, lead, copper, zinc, and the like. Of the latter class are cast
iron, hard or tempered steel, brick, stone, cement, etc. The laws of failure
of these two classes are very different, and they will, therefore, have to be
discussed separately.
18. Crushing Strength of Plastic or Viscous Materials. There is no
such thing as an "ultimate strength "in compression of a plastic body.
There is, however, a definite "apparent elastic limit," the same as in ten-
sion. Beyond this limit the material simply spreads, and increases the area
of its cross-section indefinitely under an increasing load, as shown in Plate
II. The elastic limit in compression of such a material is the greatest load
from which the specimen will fully recover, or it is the greatest load within
which the stress and deformation bear a constant ratio to each other. This
elastic limit in compression for wrought iron and steel is, fortunately,
about the same in pounds per square inch as the elastic limit in tension.
It is not customary, therefore, to test such materials in compression, but
to assume that they have the same elastic limit in compression which they
are found to have in tension.
19. The Law Governing the Strength in Compression of a Brittle or Com-
minuible Material. Experiments show that all such materials when sub-
jected to a compressive load fail by shearing on certain definite angles.
The resistance to movement along these angles is made up of two parts:
first, the strength of the material to resist shearing; and second, the
frictional resistance to motion along this plane. The sum of these two
resistances must equal the shearing component of the load imposed when
24
PLATE II.
Wrought Iron.
Steel.
Wrought Iron.
Steel.
Steel.
Wrought Iron.
Wrought Iron. Steel.
RELATIVE MALLEABILITY OP WROUGHT IRON AND SOFT STEEL.
All the specimens were originally of the shape of the one remaining umleformed.
The wrought iron specimens uniformly show large cracks. (From von Tetmajcr's Com-
munications, vol. iv, PI. V. N
MATERIALS UNDER COMPRESSIVE STRESS.
resolved along the shearing plane. To find what this angle should be, we
may equate the two resistances here described with the shear-
ing force, and find the angle of rupture, the determining
condition being that this angle shall be that which offers the
least total resistance to failure under a crushing load. This
angle may be found in the following manner:
Let s = shearing strength of the material per square inch;
A = area of prism = 1 square inch;
6 = angle of rupture;
p = crushing load per square inch.
The tendency to slide on the plane of rupture is p sin 0.
The resistance to sliding is s sec -\- fp cos 0, where /is
the coefficient of friction = tan 0, where = angle of re-
pose. Hence, at failure,
p sin = s sec -\-fp cos 0. . . .
FIG. 11.
(1)
It is evident that the angle of rupture will be such as to cause failure
under the least load; hence if be taken as the independent variable, we
shall have at rupture
j- = - s (cos 2 - sin 2 -f 2f sin cos 0) = 0,
,,-r
cos'" sin 2
' ~ 2 sin cos
cos
sin 20
= - cot 20. , . . (2)
Whence, since/ = tan 0, we have
tan = - cot 20 = tan (90 - 20) = tan (20 - 90),
or
= 20-90 and = - ^- = 45 + ^. . . . (3)
That is to say, the angle of rupture is 45 plus one lialf the angle of repose.
If the friction had been omitted, we should have had
p sin = s sec 0; whence -L = s(cos a sin 2 0) = 0;
l-2sin 2 = 0; 2 sin 1 = 1, or = 45. . . . (4)
It has been customary to neglect the friction, and to state that the
plants of rupture make this angle of 45 with the horizontal;* but the
actual plane of rupture, when the specimen has sufficient height, is about
* Coulomb is responsible for this -theory, while Navier has given the true analysis.
Most writers, including Rankine, have followed Coulomb, however.
26 THE MATERIALS OF CONSTRUCTION.
55 with the horizontal, or 35 from the direction of the applied load. See
Figs. 12 and 13, showing tests on sandstone made by Prof. Bauschinger.)
Mr. Charles Bouton has shown* that the theoretical angle of rupture is
FIG. 12. Bausehinger's Compression Tests oil Sandstone.
borne out in practice with many kinds of materials. (See Fig. 14 for
photographic views of crushed specimens of cast-iron cylinders of various-
heights, showing angle of rupture.)
The following table gives the results of Mr. Bouton's determinations of
the theoretical and the actual values of this angle:
* In u thesis for the degree M.S. at Washington University, 1891, entitled Theory
and Experiments on the Laws of Crushing Strength of Short Prisms. Mr. Bouton also de-
rived the formulae in this article and afterwards found that Navier had anticipated him.
MATERIALS UNDER COMPRESSIVE STRESS.
27
Observed
Observed
Theoretical
Number
Angle
Angle
Angle
Material.
of
Experi-
of Rupture.
of Repose.
of Rupture.
th
Differences.
ments.
*
45- + |
41 F " cast irou
24
54. 8 0.2
20. 6
55. 3
- 0.5
" C. W." cast iron
24
55 .0 .2
16 .9
53 .4
+ 1 .6
Limestone
4
62.2
33 .4
61 .7
-f .5
Asphiilt paving mixture
"Milwaukee brick
3
4
59.7
58.2
27 .3
27 .0
58 .6
58 .5
+ 1 .1
0.3
FIG. 13. Bauschinger's Compression Tests on Sandstone.
28 THE MATERIALS OF CONSTRUCTION.
The u F." cast iron was good foundry iron, having a tensile strength of
22,000 pounds per square inch and a modulus of elasticity of 14,500,000;
the " C. W." iron was car-wheel iron, having a tensile strength of 20,000
FIG. 14. Bouton's Compression Tests on Cast Iron.
pounds per square inch and a modulus of elasticity of 6,500,000, or less
than one half of the former.
20. Relation of Crushing Strength to Shearing Strength. To show the
relation of the crushing strength to the shearing strength, we have, from
equation (1) in the previous article,
s = ;?(sin 6 cos 6
also, from equation (2),
' 0);
~ cos 2^ cos 2 sin 7 9
j __ nQr v 7 __
sin 26 2 sin 6 cos 6
Substituting this value of/, we find
_ p cos _
(5)
MATERIALS UNDER COMPRESSIVE STRESS. 29
or
p = 2s tan 0, .......... /g\
where p = compressive strength in pounds per square inch, arid
s = shearing strength in pounds per square inch.
This relation was also shown by Mr. Bouton to be well borne out in
practice. The great trouble to prove such a relation is to find s experi-
mentally on brittle materials without introducing bending stresses (See
Art. 37.)
21. Relation of Crushing Strength to Relative Dimensions of Specimen. _
This is a very important matter. Hitherto nearly all crushing-test speci-
mens of brittle materials have had a cubical form. So long as the theoreti-
cal angle of rupture was thought to be 45 this was proper; but since this
theoretical angle approaches 60, it is evident that the height of the speci-
men should be at least one and one half times the least lateral dimension,
in order to allow of failure on a normal angle. Prof. Bauschinger has
studied this question very exhaustively, and the following conclusions are
drawn from studies of his results of tests on a very uniform quality of fine
Swiss sandstone, all possible refinements as to appliances having been
introduced : *
He recommends the formula
for all shapes of cross-section and for all relative heights, where
p = crushing strength per unit of area;
A = area of cross-section ;
u = perimeter of cross-section;
h = height of specimen;
a and & = constants.
For rectangular cross-sections the following formula serves very well :
4/~j
p =*+**, . ( 8 )
where k and k' are constants.
The application of this formula is shown in Fig. 15, in which the tests were
on three sets of sandstone prisms of the dimensions 2-j- in. by 5 in., 3f in.
by 5 in., and 5 in. by 5 in. in cross-section, respectively, the heights of each
set varying from one-half to five times the least lateral dimension. It is
* Mittheilungen aw dem Mechanisch TechniscJien Laboratorium der K. TecJinischen
Hoclischule in MuncJien, von J. Bauschinger, vol. vi, 1876.
30
THE MATERIALS OF CONSTRUCTION.
V
X
4000
1
1
wr
and =-, and the curve as shown
i Fig. 17 is the result.*
Thus, from this mean curve, we have the equation
strength of prism
Q
strength of cube h '
r here b l = least lateral dimension, and li = height of prism.
This equation shows that the strength of a stone prism whose height is
* This law holds between the limits h = 0.4& and h = 56, these being the limits of
ie observations.
32
THE MATERIALS OF CONSTRUCTION.
one and one half times its least lateral dimension has a strength equal
92$ of the strength of a cube of the same material.
This height of =- = - was found to be necessary to allow the materi
i
to shear on the theoretical angle of 45 -j- ~. Hence when the cubic
form is used for test specimens in crushing, the results are
if the proper height of specimen had been chosen.
greater th;
ay
0.4 d<3 /.2 {6 2.4 2.8 3.2 d 40 44 48 S.2
FIG. 17. Relation between the Crushing Strength of Prisms and Cubes.
Also, if a brick, for instance, be tested flatwise, in which positi
- 1 = 2, we find from this curve it will give a result 22$ greater than tl
for a cube, and 33$ greater than that for a specimen in which r ' = -
ft O
other words, the results from tests on cubes are 9$ too large, and on brie
flatwise they are 33$ too large.
It will also be noted that, so far as these tests go, the unit strength
the material is no function of the size of the specimen, but only afunch
of its form.
23. Effects of Loading a Portion of the Cross-section.* (a) Chamfei
Edges. If the edges of a cube or prism be chamfered off as shown
Fig. 18, and the load applied uniformly over the reduced area, the law
* All the tests discussed in this article are taken from Prof. Bauschmger's publisl
reports, but the'author of this work has discussed them witn the results as given.
MATERIALS UNDER COMPRESSIVE STRESS.
the variation of strength with varying areas of compressed surface is
shown by the curves on this figure.
Thus, as the area of pressed surface approaches that of the full cross-
section, the load carried per unit of pressed surface decreases, as shown by
the curved locus at the top, while the average load on the full cross-section
^000
/q000
n
000
0000
J000
2000
FIG. 18. Crushing Streugth of Cubes with Chamfered Edges. (Bausclringer.)
increases uniformly, as shown by the straight locus of Fig. 18, the two loci
meeting at 9500 pounds, the strength per square inch of a full cube.
These results show clearly that the bearing surface should be that of
the full cross-section of the specimen if normal results are to be obtained.
The contrary has sometimes been asserted that the strength of the
specimen was not increased appreciably by the material outside the bearing
surface. In other words, crushing -test specimens should be true prisms in
form, without chamfered edges or rounded corners.
Since the locus of unit strength for bearing surface, Fig. 18, comes
nearly into a horizontal direction as the pressed surface approaches the full
area of cross-section, it follows that when the pressed surface is nearly
equal to that of the full cross-section of the specimen the error introduced
34
THE MATERIALS OF CONSTRUCTION.
by considering only the pressed surface is very small. For instance, if the
area of the compressed surface is 0.8 that of the full cross-section (dimen-
sions of cross-section 0.9 those of the full section), the error introduced by
considering the pressed surface only would be by this curve -^VW 3.2^.
(b) Square Bearing, Symmetrically Placed. AVhen the pressed surface
is square and placed symmetrically on a larger cube, the relation of the
resistance per unit of pressed surface to the strength of the cube is shown
on Fig. 19. Here the curves are given for the small bearing on one side
I,
w
$ <
S000
6000
S000
#000
FIG. 19. Effect of Loading a Portion only of the Surface of a Cube. (Bauscliinger.)
and also on opposite sides, and the crushing resistance computed and
plotted per unit of bearing surface and also per unit of cross-section of the
cube. Evidently the loci must all meet at a point where the bearing area
equals the total area on each side, and this point will be the strength of a
cube of this material, which was 9500 pounds per square inch, the same as
shown in Fig. 18, the material being the same.
MATERIALS UNDER COMPRESSIVE STRESS.
35
(c) Bearing Surface Rectangular and Extending Entirely Across the
Cube. In this case the resistance per square inch is a function of the dis-
tance of the pressed surface from the edge of the cube. This law is shown
in Fig. 20. The material being the same as before, the strength of a cube
26000
22000
/sooo
/OOOO
&OOO
2000
s
Iff LL'L/l-
//zz
/0 20 30 40 JO ffl 7O 00 W
FIG. 20. Effect of Loading a Zone on the Surface of a Cube. (Bauschinger.)
would be 9500 pounds per square inch. This corresponds to a distance
from the edge of the cube equal to 8$ of the half-width. As the bearing-
surface had a width equal to 10$ of the half-width of the specimen, it fol-
lows that the outer line of the pressed surface came within 3$ of the half-
width, or \\% of the total width from the edge of the specimen when the
normal strength of the material was developed.*
24. General Laws of Crushing Strength. The laws of crushing strength
shown in Figs. 15 to 20 apply specifically to a particular quality of sand-
stone. In Fig. 16 it is shown that cast iron follows a different law. In all
probability each kind of material, or at least materials which have different
angles of rupture (that is to say, different coefficients of friction), will show
different curves for the several relations indicated in these plates. In the
absence of any more definite information, however, on this subject, it is
thought the curves shown upon these plates will serve to indicate in a
general way the laws of the variation of crushing strength with the varying
conditions here indicated.
* See figures 12 and 13 for methods of failure for cases (a), (5), and (c).
36
THE MATERIALS OF CONSTRUCTION.
By referring to Fig. 14 [it will be observed that the cylinders all
swelled more or less in the middle before rupture occurred. This is doubt-
less due to the restraining action of the friction against lateral motion on
the end bearing surfaces. It is difficult to take this source of strength
fully into account in a theoretical analysis of resistance to crushing.
25. Strength of Columns. When a compression member is so long as
to fail in compression by lateral deflection, its failure is a function of the
elastic-limit strength and of the stiffness (modulus of elasticity) of the
material, rather than of the ultimate strength of the material in compres-
sion. The discussion of this case properly comes in works on mechanics
and on framed structures. The author has fully expressed his views on this
question in his work on Modern Framed Structures, and to some extent in
Chapter XVI of this work, and hence he will not occupy space with it here.
26. Weakening Effects of Eccentric Loading. Few persons are aware
of the great increase of stress on the near side of a member subjected to a
n n n ~~~ direct stress (either tension or compres-
sion) caused by an eccentricity of the
load-line with reference to the gravity-
axis of the member. This eccentricity
may result from an eccentric imposi-
tion of the load itself; or from the
member being bent; or from the ad-
dition of material on one side of the
member, such addition usually prov-
ing a source of weakness instead of
strength. These three cases are shown
in Fig. 21. In each case we have
Total load = P;
" area = A ;
Eccentricity = a\
Width =h;
Moment of inertia of section = /;
Radius of gyration of section = r;
Distance of extreme fibre from the
gravity-axis = y l = with symmetrical
A
T
a
FIG. 21.
sections;
Total stress on nearest outer fibre =/.
h
Hence we have for symmetrical cross-sections, where y, = ,
A
* The stress due to the bending moment Pa is found from the equation m = or
#i
/= , where m = Pa, and / = Ar*.
MATERIALS UNDER COMPRESSIVE STRESS. 37
For solid rectangular cross-sections we have r 2 = = ; hence for
A. LA
such sections
The proportionate increase in the stress, therefore, over that which
would obtain for a concentric load is given by the fraction y . In other
words, when a = ^h the stress on the outer fibre on the near side is doubled,
compared with that for a central loading.
To discover the weakening effect of additional material added to one
side of a member, assume a central loading on a straight symmetrical
member having an initial width = h, (a = o). If additional material, to a
thickness of x, be now added on one side of this member, the new total width
x
becomes h -\- x, and the eccentricity is a = . Assuming the member to be
solid rectangular in cross-section, with original dimensions of b and h, the
new dimensions of section are b and h-\-x; the former area was A = bh, and
P P
the latter A' = b(h -{- x). Before the addition we should have/= = .
A Oil
After the addition we should have, from (14), for the stress on the
near side,
0.)
Hence the increase of the stress due to an unsymmetrical addition of
material is
(16)
This is zero for x = and for x = 2h, and it is a maximum for x = -,
when it becomes
Hence we may say that the addition of material on one side of a mem-
ber subjected to a direct stress symmetrically placed weakens it until the
added material has exceeded twice the original thickness of the member, the
maximum weakening occurring when the added material is one half the
original thickness, when the enlarged member is only three fourths as
strong as the original member.*
* Attention was called to this fact by Mr. Carl G. Earth in Trans. Engrs. Club of
Phila., Oct. 1891, p. 307.
CHAPTER IV.
MATERIALS UNDER SHEARING STRESS.
27. Two Manifestations of Shearing Stress. When all the opposing
external forces which act on a body lie in one plane,* but not in one and
the 'same line, the resisting stresses are those of simple shear and cross-
bending, without torsional stress.
When the opposing external forces do not lie in one plane the resisting
stresses are those of torsional shear, with or without cross-bending and
simple shear.
In any case these three kinds of stress are determined separately, as
follows :
(a) For Parallel External Forces in One Plane. The moment oi
resistance of the bending (direct) stresses at any transverse section is equal
to the algebraic sum of the moments of tho external forces on either side oi
that section taken about the neutral axis in that section.
The simple shearing stress on any section is equal to the algebraic sum
of the transverse components of the external forces on either side of that
section.
(b) For Parallel External Forces Not in One Plane. First replace all
the forces by equal parallel forces acting in the plane of the axis of the body;
and by couples equal in value in each case to the force multiplied by
its displacement. Then the moments of resistance and the simple shearing
stresses will be the same as in the last case, and in addition there will be*
the moment of torsion.
The torsional moment at any transverse section is equal to the algebraic
sum of the moments of the couples of the displaced forces, acting on either
side of the transverse section in question,
(c) For Non-parallel Forces Acting in Any Manner. Resolve all forces
into horizontal and vertical components at their points of application, and
then solve for bending moments, shears, and torsions at any section in these
two planes. ;'
The bending moment at this section will then be the square root of the
sum of the squares of the bending moments at right angles to each other.
* When a force is distributed over an area it is here supposed to act at the centre-
of gravity of these force-elements.
MATERIALS UNDER SHEARING STRESS.
39
The total shear will also be the square root of the sum of the squares of
the primary shears at right angles to each other.
The total moment of torsion will be the algebraic sum of the two
moments of torsion found from the two sets of forces.
28. The Moment of Torsion gives rise to a shearing stress over the entire
cross-section, which is zero at the centre of gravity of the section, and
which increases in intensity directly as the radial distance from the gravity
axis.
For various forms of sections, the following intensities of shearing stress
are found, by the principles of mechanics, for the corresponding forms of
cross-section.
The general equation for resistance to torsion is
.a)
Figure.
Dimensions.
i Radius = r.
Outer radius = r (.
Inner " = t\ (
Side = b.
3uter dimension = b
[uner " = b 1
Side = a.
Radius of circumscribed
circle = r
Longer axis = 2a I
Shorter axis = 26 )
longer semi-axes=a&ct 1
Shorter " " =6&6 X )
Area.
&a
6 2 - 6^
3a 2 _
2,-a |/2
irab
rr(ab a 1 6 1 )
J*
nr*_
~2~
+ 2^2)
4 -
2r
31/2
3 1/2
1.083o
1.876r
~(os + & a )
40 THE MATERIALS OF CONSTRUCTION.
where M total torsional moment;
s = shearing stress on extreme fibre;
J* = polar moment of inertia of cross-section about the gravity axis
r distance from neutral axis to the extreme fibre having th
shearing stress s.
Whence we have, for the forms figured, the relations given in th
table.
29. Shearing Deformations. As shown in Arts. (7) and (8), a shearin
action of external forces results in angular deformation of the body. L
the case of simple shear, or where the forces lie in one plane, the angula
deformation from shear is very small, the bending being mostly due to th
longitudinal deformations resulting in the direct tensile and compressiv
resisting stresses on the two sides of th
neutral plane respectively. When th
forces do not lie in one plane, or wher
there is a moment of torsion, the angula
deformation gives rise to a twist of th
body about the neutral longitudinal axis
Thus in Fig. 23 assume the solid cylindei
anchored at o, to have a length / and
radius r. Let the torsional moment be Pa = M t . Then the shearin.
stress on the extreme fibre is, by equation (1),
_ M t r _ 2Pa
where J"is the polar moment of inertia = twice the rectangular moment o
inertia in this case.
In Art. 9 the shearing modulus of elasticity was defined as
__ shearing stress per sq. in.
angular deformation
If we take the stress and angular deformation of the outer fibre i
Fig. 23, we have :
2 Pa
Shearing stress per sq. in. = s - r .
Tangent of the deformation angle -= = deformation angle,
since this angle is small.
Hence we have
_ _ si
8 ~ ~
* The student's attention is called to the fact that the polar moment of inertia
equal to the sum of the true rectangular moments of inertia about the principal ax
through the centre of gravity of the section.
MATERIALS UNDER SHEARING STRESS. 41
or
In general, for any cross-section we have
M t l si M t l si
s = ~ = =
where y^ distance from the neutral axis to the extreme fibre in which the
stress is s.
In Art. 9 it was shown that the shearing modulus of elasticity f of
Young's modulus, or E s = \E. Hence in terms of Young's modulus of
elasticity, which is that ordinarily given, we have
5 MJ,^5sl_
2 ' EJ ZEyS
where 6 = angular movement in terms of the radius;
M t = torsional moment on the bar;
I length of bar between sections representing a relative angular
movement of 0;
s = shearing stress on outer fibre;
E s shearing modulus of elasticity of the material;
E = the ordinary modulus of elasticity;
J = polar moment of inertia = I x -\- I y , where these are the rectan-
gular moments of inertia about the principal axis through the
centre of gravity of the section ;
y^ = distance from neutral axis of outer fibre in which the shearing
stress is s.
CHAPTER V.
MATERIALS UNDER CROSS-BENDING STRESS.
30. Historical Sketch.* For two hundred and fifty years the true theory of the
strength of a beam has been a much-mooted question amongst physicists, engineers,
and mathematicians.
Galileo was the first of whom we have any record who undertook to discuss the
problem. In his famous Dialogues (Leiden, 1638, from which Fig. 24 is taken) he
FIG. 24.
propounds a theory based on an assumed absolute rigidity of the material, and con-
cluded that the fibres of the beam were subjected to a uniform tension which acted I
about the base of the beam as a fulcrum. On this theory the moment of resistance
* This historical review of the development of the true theory of the beam is derived!
mostly from Saint-Venant's Historique Abrege des Recherches sur la Resistance et sur-
I ' Elasticite des Corps Bolides, prefixed to his Navier's "Lecons," Third Edition, Paris,,
1864, and from Totlhunter's History of the Theory of Elasticity, Cambridge, Eng.,
1886. It is here reprinted from the author's joint work on Modern Framed Structures.
42
MATERIALS UNDER CROSS-BENDING STRESS. 43
of a solid rectangular beam would be , where / is the ultimate strength of the
material in tension.
Robert Hooke first published his famous law of the relation between deformation
and stress in 1678, discovered by him he says 18 years previously, and kept secret for
the purpose of procuring patents on some applications of the principle to springs' for
watches, clocks, etc. Two years previously he had ventured to publish the law in
an anagram at the end of another book, in this form, " ciiiinosssttuv" which
being interpreted reads, "Ut tensio sic vis," or, ' 4 as the extension so is the resist-
ance." Hooke makes this law apply to all "springy" bodies, amongst which he
names nearly all ordinary solids. This is still known as Hooke s Law.
Mariotte showed by experiment in 1680 that the fibres on one side of the beam
were extended and on the other side compressed, and assumed that the neutral
surface passes through the centre of gravity of the section.
Varignon, in 1702, undertakes to harmonize the theories of Galileo and Mari-
otte, by admitting the extension of the fibres, but puts the neutral plane at the bot-
tom, as Galileo did, and assumes the tensile stress as uniformly varying from there
to the other side. This would make the strength of a solid rectangular beam - -,
3
which agrees almost exactly w r ith the facts for cast iron at rupture when f is the
tensile strength.
James Bernouilli made an important advance by applying Mariotte's law to
obtain deflections of beams (1694 and 1705), and argued that the position of the
neutral axis is a matter of indifference, which was a great error. He denied the
truth of Hooke's law, which we know is not applicable to all substances, nor to the
point of rupture with any substance. He first constructed stress diagrams, but
his work in the field of hydraulics was of even greater importance than in the study
of solids.
A. Parent, a French academician, seems to have been the first to perceive (1713)
the mechanical necessity of equilibrium between the tensile and compressive stresses,
which condition, together with that of a uniform variation of stress, fixes the posi-
tion of the neutral axis at the centre of gravity of the section. This important dis-
covery seems, however, to have passed unnoticed.
Coulomb reannounced this relation in a memoir to the French Academy in 1773,
or sixty years after its first publication by Parent. Saint- Venant credits Coulomb
with never having seen Parent's work, as no writer of that century, has mentioned
it. But even after this second publication of so important a necessary truth, such
workers as Girard, Barlow, and Tredgold all misconceived the mathematical necessi-
ties in the problem, and resorted to various makeshifts to explain the strength of
beams. -
Navier finally, in 1824, put the matter on a solid mathematical basis, although
he also at first went entirely astray. He stated in his first edition that the moment
of resistance varied as the cube of the depth of the beam, and in his second edition
this error was corrected, but the moment of the stresses on one side of the neutral
axis was said to be equal to the moment of the stresses on the other side, about that
axis, an equality which does not exist except on symmetrical sections. Navier also
fully developed the theory of the deflection of beams as we now use it.
Saint-Venant, a student of Navier's, has finally (1857) in his notes on Navier's
Lemons given a complete analysis of both the elastic and the ultimate strength of a
beam, with suitable equations which will give theoretical results agreeing with the
actual tests, when the empirical constants are properly evaluated. This great engi-
neer, physicist, and teacher has done more than any other one to bring theory and
practice into harmony and to put both on a thoroughly scientific basis, so far as the
strength and elasticity of engineering materials is concerned.*
In spite of these various true expositions of this subject the source of
strength in a beam continues still to be very imperfectly understood by
* He died January 6, 1886.
THE MATERIALS OF CONSTRUCTION.
many engineers, and even by current writers on applied mechanics, and
gross errors in this direction are still common. It is in consideration of
this state of the science that the problem is treated so fully here.
31. Fundamental Equations of Equilibrium. When a solid body is in
equilibrium under the action of non-concurrent external forces, the follow-
ing propositions hold true for the body as a whole:
I. The sum of the vertical components of /he external forces is equal
to zero.
II. The sum of the horizontal components of the external forces taken in
any plane is equal to zero.
III. The sum of the moments of the external forces taken about any
point is equal to zero.
When a solid body is subjected to the action of non-concurrent forces
acting in one plane the body may be regarded as a beam, since the effect of
the external forces is to bend the body and develop in it what are commonly
called cross-bending stresses. If a section
be passed through the body perpendicular
to the plane of the forces, and the portion
of the body on one side of this section be
removed, the other portion may beheld in
equilibrium with the external forces act-
ing upon it, by means of the stresses exist-
ing in the body on this cross-section, these
stresses now being regarded as external
forces, as indicated in Fig. 25. Since the
remaining portion of the body now under
consideration is in equilibrium under the
action of external forces and of internal
stresses, which for the time may be re-
garded as external forces, the three propositions given above will apply.
Or, stating these propositions now so as to equate the real external forces
with the internal stresses developed at the section, they would read as
follows :
If a transverse section be passed through a beam
I. The sum of the vertical components of the stresses acting at the sec-
tion is equal to the sum of the vertical components of the external forces
acting upon the body on either side of that section.
II. The sum of the horizontal components of the stresses acting on the
section is equal to the sum of the horizontal components of the external
forces acting upon the body on either side of that section.
III. The sum of the moments of the stresses acting on that section i?
equal to the sum of the moments of the external forces acting on the body on
either side of that section.
It follows from the above that if all the external forces acting upon
FIG. 25.
MATERIALS UNDER CROSS- B ENDING STRESS. 45
a beam are parallel vertical forces, the end reactions or supports being
regarded as external forces the same as any primary weights or loads, and
if no horizontal forces act upon the beam, then we should have for any ver-
tical section
I. The shearing stress is equal to the algebraic sum of the external forces
acting on either side of the section.
II. The algebraic sum of the horizontal stresses acting on the section is
equal to zero.
III. The algebraic sum of the moments of the stresses acting on that
section, which is commonly Called the moment of resistance, is equal to the
sum of the moments of the external forces about any point in that section.
The effect of the action of cross-bending forces upon a beam is to bend
or deflect it, thus shortening the lengths of the fibres or elements on the
concave side of the beam, and lengthening them on the convex side. So
long as this action does not exceed the elastic limits of the material, the re-
sisting stresses are directly proportional to the deformations. Hence there
is always found a compress! ve stress on the concave side and a tensile stress
on the convex side of a beam, and therefore there will be a plane near the
centre of the beam the elements of which are neither lengthened nor
shortened, and on which there will be no longitudinal stress. This is
called the neutral plane or " neutral axis" of the beam.
Furthermore, a geometrical effect of the bending of a beam is to pro-
dace deformations which are zero at the neutral plane and which increase
uniformly outward to the extreme convex and concave sides, and hence the
longitudinal resisting stresses developed by these deformations also increase
uniformly outward. AVithin the elastic limits, therefore, the direct stresses
increase uniformly from the neutral plane to the extreme fibre*.
Since from Proposition II, as stated above, the summation of the hori-
zontal stresses on the cross-section is zero, in simple cross-bending, where
the external forces have no horizontal components, it follows that the total
summation of the tensile stresses on the convex side of the neutral plane must
always exactly equal the total summation of the compressive stresses on the
concave side. Also by Proposition III the sum of the moments of all these
stresses taken about any point in this plane must equal the sum of the
moments of the external forces acting on either side of the section taken
about the same point. If this centre of moments be taken in the neutral
plane itself it will at once be evident that the moment of the tensile forces
on one side has the same sign as the moment of the compressive forces on
the other side, and that they are, therefore, to be added together numer-
ically in order to equal the algebraic sum of the moments of the external
forces acting on either side of the section. While, therefore, the sum of the
moments of the tensile stresses may be numerically equal to the sum of the
moments of the compressive stresses (which is the case for symmetrical
cross-sections), yet since they are to be added together numerically, in order
46
THE MATERIALS OF CONSTRUCTION.
to equal or hold in equilibrium the moments of the external forces on one
side of the section, there is evidently no mathematical necessity why the
moments of the compressive stresses should equal the moments of the tensile
stresses; and in unsymmetrical sections, and even in symmetrical sections
beyond the elastic limit, these moments are not equal to each other.
Since the stresses on any cross-section of a beam subjected to the action
of bending forces increase uniformly from the neutral plane to the extreme
sides, it is evident that it is only the stress found to exist in the extreme
fibres or elements of the beam, which needs to be determined. That is to
say, if the maximum stresses are kept within the working limits, it is imma-
terial what the particular stresses are on other portions of the cross-section.
It is common, therefore, to find the relation between the total moment of
resistance of a beam (which of necessity is always numerically equal to the
bending moment of the external forces), and the stresses on the extreme
fibres or elements of the cross-section of the beam. This general relation
between the bending moment and the stresses on the extreme fibres is made
the subject of the following article.
32. Relation between the Moment of Resistance and the Stress on the
Extreme Fibre. In Fig. 26 let the load P be applied at (7, and this will
!
LH ,
(S) \y
c /K i -
j
k
;
\
p
FIG. 26.
produce a bending moment on AB of Pd. On this plane the moment of
the longitudinal stresses makes up the moment of resistance which holds in
equilibrium (and hence is always numerically equal to) the bending moment
of the external forces. That is to say, M= Pd = M , the moment of resist-
ance. We shall here assume the cross-section to be irregular and unsym-
metrical, as shown in the figure. The direct stress varies uniformly across
the section in all cases. The following notation will be used :
M = bending moment of the external forces.
M = moment of resistance of the direct stresses = M.
p = intensity of the direct stress at the distance ?/ from the neutral
plane ay, where a = intensity of direct stress at a unit's
distance.
f = intensity of the direct stress at the extreme side of the beam.
y 1 = distance of extreme fibre on one side from the neutral axis.
v ' = " " " " " the other side from the neutral axis.
MATERIALS UNDER CROSS-BENDING STRESS. 47
/ = y*dxdy moment of inertia of the cross-section about the centre
of gravity axis,
y = distance from axis of reference to the centre of gravity of the
cross-section.
Intensity of stress on any fibre = p = ay\ (1)
Total stress on fibre having an area of dxdy = pdxdy = ay dxdy ', . . (2)
Moment of stress on fibre dxdy = pydxdy = ay* dxdy; (3)
/+!
y*dxdy al. . (4)
2/1'
But SLS p = ay, so/= ay t and/' = ayf\ or
Therfeore
fl* f'l
M - M n = al = - - = J r (5)
I y* y'
This is the general equation between the moment of resistance and the
stress on either extreme fibre. When the section is symmetrical, y ^ = yf\
hence f = f' 9 and only one side need be considered.
When the cross-section is solid and rectangular, equation (5) becomes
(6)
The above demonstration assumed that the neutral axis or plane of the
beam passed through the centre of gravity of the cross-section, since /was
referred to this gravity axis. This remains to be proved.
From equation (2) we have, the stress on any element is ay dxdy, where
y is measured from the neutral axis. But for simple cross-bending the
algebraic sum of these direct stresses over the whole section is zero; hence
we have
But
/+Vl /+! /+01
aydxdy a I ydxdy a I ydA = ..... (?)
-VT! "-Hi! * / -Vi'
fydA=yA,\ ............. (8)
*Both equation (5), M = , for any section, and equation (6), M Q \fbli* t for solid
y\
rectangular section, should be thoroughly memorized by the student, as they are of con-
stant application.
f The symbol y denotes the distance from the axis of y to the centre of gravity of the
f ydxdy f ydxdy
cross-section, and it equals .
Jdxdy A
48
THE MATERIALS OF CONSTRUCTION.
since the sum of the statical moments of the elementary areas about any
axis is equal to the moment of the total area into the distance to its centre
of gravity. Therefore we have, for reference to the neutral axis,
/H-2/1
i I yd A = 0,
*'.'
or yA 0.
(9)
But yA can only equal zero when reference is made to the gravity axis-
Therefore these two axes must coincide. In other words, the neutral plane
always traverses the centre of gravity axis of the beam, so long as the
stresses remain inside the elastic Hunts of the material in both tension
and compression, and also provided the modulus of elasticity is the same
for both kinds of stress.
33. Moments of Resistance (Strength) of Beams of Various Forms of
Cross-section. The moment of resistance of a beam of any form of cross-
Form of
Cross- section.
Distance of Centre
of Gravity,
or Neutral Axis,
from the
Most Distant Fibre.
Moment of Inertia
about the
Centre of Gravity
of the
Section.
= 7
Moment of Resistance
in
Terms of the Stress
in the
Most Distant Fibre.
--
*6 >
h
If
6/1,3
12
>
O:
d
~2
vd*
64
/ d>
le"
1
24^ ; '
i* " o "*
T7V
h
12
7
V>
6 ^ //IJ
">'^?T"
1
bh* - (b - t'yh - 2f)3
57^3 _ ^ - t'yh - 2t) 3 f
'"2Ei.
12
6/1
/IE
lt'h* + t(b -t')(h- JO
b/j,3 _ (b - t'yh - t) 3
fl
t'h + <(6 - V)
3
Vi
^* .
6 + 26' fc
^rBft-f*' (6+26')^
//^r3(36 + 6')(6 + 60
/^ \ *
6 + 6' * 3
k L 12 18(6 + 6') J
eL 2(6 + 26') ( 6 + 26 U
$'
MATERIALS UNDER CROSS-BENDING STRESS. 49
section was found to be, by equation (5), M a = , where /= intensity of
stress on the extreme fibre which lies at a distance from the neutral plane
equal to y lt and /is the rectangular moment of inertia of the cross-section
about the neutral or gravity axis. In the table on p. 48 are given the
values of ?/,, /, and M for various forms of sections which are commonly
used as beams. For tabular and graphical methods of finding the moments
of inertia of irregular forms, see Modern Framed Structures, pages 127-130.
The values given in the above table are true for all values of /inside the
elastic limit. When this limit is exceeded the stress no longer varies uni-
formly across the section, but the stresses near the neutral axis are larger
than the above theory allows, and hence, for a given actual stress on the
extreme fibres beyond the elastic limit (as the breaking-stress, for instance),
the moment of resistance is much more than would-be obtained by using
the breaking value of /(in tension Of compression) and substituting this in
the above formulae. It must be understood, therefore, that in no case are
these formulae, true at rupture, but only inside the elastic limits of the ma-.
terial. It is for this reason that the values of /as found from cross-bending
tests carried to failure, and as computed from the above formula?, differ so
largely from the breaking values of the material in direct tension or com-
pression.* Thus, cast iron, which has a tensile strength of 20,000 pounds
per square inch and which breaks on the tension side in cross-breaking, has
a value of/, when computed by the above formulae from a breaking-load, of
from 30,000 to 40,000 pounds per square inch, depending somewhat on the
shape of the cross-section of the specimen. The more the material is con-
centrated near the neutral plane the more the value of /differs from the
tensile strength. This value of/ computed from the breaking moment, is
called the modulus of rupture in cross-breaking. It is from 1.5 to 2 times
the tensile strength of the metal.
In timber beams the reverse is the case; that is to say, the crushing
resistance being less than the tensile resistance, the modulus of rupture in
cross-breaking is greater than the former and less than the latter, and it is
in fact nearly a mean of the two.
34. Strength (Moment of Resistance) of Beams beyond their Elastic Limits.
After the stress on the extreme fibres on one or both sides of the beam
has passed the elastic limit, the distribution of stress over the section is no
longer uniformly varying as was assumed in deriving the formulae of the
last article, and the law of this variation will now be examined.
In all cases the variation of stress across the transverse section of a,
beam subjected to simple cross-bending, with or without shearing stress,
folloivs the law of the variation of the stress ordinates to a stress-diagram
* See a full discussion of this subject iii the author's work on Modern Framed
Structures, Chapter VIII.
50
THE MATERIALS OF CONSTRUCTION.
in which the extreme ordinate represents the stress on the extreme fibre of
the beam.
Thus in Fig. 28, suppose the beam to be cast iron, and to be bent until
the stress on the extreme fibre on the tension side has become f t . Passing
FIG. 28.
now to the tension portion of the stress-diagram for this material,* we see-
that this stress,/,, is found far beyond the elastic limit of the metal in ten-
sion. Let us now recur to the fact that the deformation of the longitudinal
fibres of the beam increases uniformly outward from the neutral axis, even
beyond the elastic limit, since the section remains sensibly plane, and hence
the uniform increase of the deformation is a geometrical necessity. In
view of this fact it becomes evident that the law of increase of stress from
the neutral axis outwards, or the law of the increments of stress correspond-
ing to equal increments of deformation, is exactly that represented by the
stress-diagram, since here we have the increments of stress shown for equal
increments of deformation. Hence it follows that if f t is the stress on the
extreme fibre of the bent beam on the tension side the stresses on all
other fibres on the tension side are truly indicated by the lengths of the-
corresponding ordinates on that side of the neutral axis, when the position
of the stress ordinate/ in the stress-diagram is taken as the position of the
extreme tension side of the beam, and the origin in that diagram is taken
as lying on the neutral axis of the beam. Evidently the same argument
would apply to the compression side.
35. Distribution of Stress and Position of the Neutral Axis at Ruptuie.
In a brittle material like stone or cast iron, failure occurs on the tension
side; while in the case of wood, failure usually occurs first on the compres-
sion side of the beam. The diagrams shown in Fig. 28 may fairly be taken
as representing the facts in the case of cast iron, and those in Fig. 29 in the
case of timber. Since timber is much stronger in tension than in compres-
sion, it fails first on the compression side. Furthermore, after the fibres
have buckled, or broken down, in compression, they are able to support only
about three fourths as much of a load as before, so that the compression
stress-diagram has the peculiar form shown in the accompanying figure.
At failure, therefore, the tensile and compressive stresses are distributed
over the section in a manner entirely different from that which obtains
* See Chap. XXIII for complete stress-dingrams for cast iron of various qualities.
MATERIALS UNDER CROSS-BENDING STRESS.
51
within the elastic limit. The statement made in Art. 25, however, regard-
ing the equality between the sums of the tensile and compressive stresses
must still hold, as this is a mathematical or mechanical necessity; and as
this total stress is graphically represented by the area of the stress-diagram
FIG. 29.
shown on the sections of the beams in Figs. 28 and 29, it follows that these
stress areas on the two sides of the neutral axis must be equal.
Thus in the case of timber, for instance, the neutral plane at first lies in
the centre of gravity of the cross-section, but after the material has begun to
crush on the compression side, the neutral plane rapidly moves towards the
tension side of the beam and often, at final rupture, it lies very near this side,
the tension stress area being a triangle of very long base (stress on extreme
fibre) and very short altitude (distance to neutral plane). It is evident
that, although the beam has long since failed in compression, if it be con-
tinuously deflected, failure must ultimately occur also in tension. When
the material is weaker in tension than in compression such double failure
cannot occur, since the tension failure parts the body, and the rupture is
complete. Evidently no general law can be given for distribution of the
stress across the section after the elastic limit has been passed, other than
to say it is that of the corresponding stress-diagrams of that material in
direct tension and compression respectively.
36. Moduli of Rupture in Cross-breaking. Prom the facts related in the
preceding article it is evident that the formulae of Articles 32 and 33 cannot
apply at rupture, and that if the breaking-load be used for computing the
so-called ultimate strength of the material in pounds per square inch (the
"modulus of rupture in cross-breaking/ 7 and the quantity /"in those
formulas when P is the breaking-load, or when M is the ultimate bending
moment), the result obtained as the value of /is a purely fictitious quantity,
and that it does not really represent any actual tensile or compressive stress
on the extreme fibres at all. It may, however, be called the " modulus of
rupture in cross-breaking" in pounds per square inch, and used to indicate
the strength of the material when loaded as a beam ; but it must- not be con-
fused with, or assumed to have any fixed relation to, either the tensile or the
compressive strength of the material. As a matter of fact it always lies
somewhere between these two latter values, but it does not have any uni-
THE MATERIALS OF CONSTRUCTION.
versal relation to them. It is always dependent largely on the form of the
cross-section of the beam, as to the concentration of material near the neu-
tral axis or near the extreme sides. Thus the elastic-limit strength of a
rolled I beam can be very closely approximated by using for /in equation
(5) the tensile or compressive elastic-limit strength of the material in either
tension or compression, while the elastic-limit strength of a solid round
bar could not be determined very closely by so doing. Also the ultimate
strength of a cast-iron beam of an I-shaped cross-section could be deter-
mined approximately by using the tensile strength of the material for the
value of / on the tension side of the beam in eq, (5), but the ultimate
strength of a round or square cast-iron bar would be nearly twice as much
as would be shown by the use of eq. (5) if the tensile modulus of rupture
were taken.
37. The Distribution of Shearing Stress in a Beam. (a) The Relation
between Shear and Bending Moment at any Section. In Fig. 30 assume
any two adjacent sections dx apart. Let the total
shearing force acting here be S. Call the bending
moment at the first section M, and that at the other
M'. Assume the beam to be cut at the section
where the moment is M 9 and the left portion re-
moved and replaced by the direct tensile and com-
pression stresses, and also by the total shear, S.
Then it is evident the moment at the adjacent
section is
.... (10)
M M
M M
But
FIG.
M' -M= dM,
hence we have
M'-M=dM=Sdx,
or o
dM
dx'
(11)
That is to say, the total shear on any transverse section of a beam is
equal to the first differential coefficient of the bending moment.
It follows from this that
(1) Where the bending moment is constant the shear is zero.
(2) Where the shear is zero the bending moment is at a maximum
or a minimum.
(b) The Distribution of the Shearing Stress across any Transverse Sec-
tion. In Fig. 31 take two transverse sections,
dx apart, as before, on which the moments are M
and M' respectively. By eq. (5), Art. 26, we
have for the stresses on the outer fibres at these
two sections
and
MATERIALS UNDER CROSS-BENDING STRESS. 53
Also for any horizontal section, as cc 1 ', the fibre stresses will be
P =f?L=%> and ,' = *
The breadth of the beam must be regarded as variable to obtain a gen-
eral solution, and it will be denoted by b, a variable quantity.
Now the total shearing stress on the section cc' , whose area is b'dx,i& the
difference between the total direct stress on a'c' and on ac. But the stress-
on ac is
A/17 ir ^
X//i / bMyd
pbdy / ' T
Jy> 1
Similarly, the total direct stress on a'c' is
The difference is
f'b(M'- M)ydy _ C^ldMydy.
Jy I "Jt 7
v'
But dM '= Sdx by the previous article, hence we have at last
C yi bSdxydy
total stress on plane cc' I ~~f~~ ....
t/ y*
But the area of this section on cc' is bdx. Hence the intensity of the
stress on this plane is /
c/ y*
I Vl bSdxydy
b'ldx"
Now S, I, and b' are constant for any given beam, transverse section, and
: plane of shear cc' '; hence these quantities can be removed from under the
integral sign, and we have
Intensity of shearing stress at any point in a beam, distant y from tlie
neutral axis, is
/*/!
Now / bydy is the statical moment of that portion of the cross-section
of the beam outside the line cc' on which the shearing stress is obtained,
taken about the neutral axis; hence we may say:
The intensity of the shearing stress at any point in the cross-section of
a beam is equal to the total shearing stress on that cross-section, multiplied
by the statical moment of the area of that portion of the cross-section out-
side the longitudinal plane of shear in question, about its axis in the neu-
54 THE MATERIALS OF CONSTRUCTION.
tral plane, divided by the product of the moment of inertia of the entire
cross-section into the breadth of the section at that point.
This applies to solid sections of beams of all possible shapes.
For a solid rectangular section b is constant and / bydy = ~(y* y z ).
Jy
Hence for this case, where h = 2y lt and b b' (a constant), we have
=[(!)' -/]- 1
Hence the shear at the extreme sides, where y -, is zero, and at the
fy
neutral axis
That is to say, the maximum intensity of the shearing stress on a solid
rectangular section is f of the mean intensity.
It is evident, also, that eq. (14) is the equation of a parabola with it
vertex on the neutral axis, whic
Beam uniformly loaded . , , -, . ,, A ,
is also the axis of the curve.
'eutral
any particular longitudinal plane
also, the shearing intensity varie
from end to end of the beam, a
the total shearing stresses on th
cross-sections vary, as shown i
Fig. 32. By applying equation (14) to various forms of cross-section it ca
be shown that
1. The maximum shearing intensity in a beam of a solid rectangula
section = |- the mean shear.
2. For a solid circular section it is f the mean shear.
3. For I beams and plate girders it is practically equal to the total shea
divided by the area of the web portion alone.*
38. To Dimension the Cross-section of a Beam. (a) For Direct Stress O't
the Outer Fibre. If the beam be of a solid rectangular form of cross-section
use eq. (G). If of any other form, use eq. (5) and evaluate by the accom
J \
panying table, if the form be one there given. If not, it will be necessar
to compute the moment of inertia. If the form be irregular or unsym
metrical, it may be best to obtain the moment of inertia graphically.! Ii
the case of unsymrnetrical sections the neutral axis lies at different distance
from the outer fibres on the tension and compression sides, and it may b<
necessary to compute both of them. Since f\ = jr- 1 , it is evident thes--
* See Modern Framed Structures, Art. 130.
\ Ibid., Art. 127.
MATERIALS UNDER CROSS BENDING STRESS. 55
stresses per square inch are to each other directly as the distances of their
fibres from the neutral axis. Thus in the case
of a cast-iron beam the cross-section is made
larger on the tension side, as in Fig. 33. Here
the outer fibres on the compression side are
much farther away from the neutral axis than
the outer fibres on the tension side, and hence the maximum stress in com-
pression is much greater than it is in tension, which is as it should be.
For a solid rectangular section we have
M=M. = -, or bk" = ...... (16)
Take M maximum bending moment on the beam, in inch-pounds,
and f = working value of the stress on the outer fibre. This gives the
value of bit*, and 1} and h can now be chosen at pleasure, conditioned on
Wf being equal to the right-hand side of the equation.
(b) For Shearing Stress along the Neutral Axis. Since timber is very
weak in shearing, as compared to its strength in tension and compression,
timber beams and joists of ordinary lengths will usually fail by shearing,
and hence they should be dimensioned to safely withstand this shearing
action. Lanza shows* that the shearing strength of spruce and white- and
yellow-pine beams is about -fa of the transverse modulus at rupture, but he
recommends a much smaller factor of safety for shearing than for transverse
rupture. If the factor of safety for shearing be two thirds that for trans-
verse strength, we would have the working stress in shearing -fa that in
cross-breaking. In order to show what length of wooden beams would
require dimensioning for shearing, using this relation of working stresses,
we have
/=20 ?0 ......... (17)
For a beam loaded at the centre
e W , ,, Wl
S = and M .
ii 4
For a beam uniformly loaded
W Wl
S = and M = .
/w O
Also from (16), for cross-breaking,
Wi* --
and from (15)
Applied Mechanics, 4th ed., p. 696.
66 THE MATERIALS OF CONSTRUCTION.
For a beam loaded at the centre
For a beam uniformly loaded
QM 3 Wl
bh -"T ~-*T ........ (20)
From (17), (18), and (19) we find, for beams loaded at the centre, they
are equally strong in shearing and in cross-breaking when
<>
and for beams uniformly loaded they are equally strong in these two ways
when
1 - f
For shorter lengths wooden beams are weaker in shearing than in cross-
breaking. Hence we have the following
PROPOSITIONS.
Wooden Beams in Shearing and Cross-breaking.
I. For a centre load the beam should be dimensioned for a shearing
stress wlien the ratio of length to height is less than one half the ratio of the
cross-breaking to tlie shearing working stress.
II. For a uniformly distributed load the beam, should be dimensioned
for a shearing stress when the ratio of length to height is less than the ratio
of the cross-breaking to the shearing working stress.
Thus, for white- and yellow-pine and spruce beams we may take
^ = 20.* (23)
Whence
All pine and spruce beams should be dimensioned for shearing failure:
I
For concentrated centre load ivhen T ^ 10.
( 24)
For uniformly distributed load 'when -^ <20.
In dimensioning for cross-breaking use equations (19) and (20), and
for shearing use equation (18), for both concentrated and distributed loads.
The following working values of /and q may be used.
* Here q is not the true shearing resistance of sound timber, but the shearing resist-
ance of large beams along their neutral axis, where they are usually season -checked.
MATERIALS UNDER CROSS-BENDING STRESS.
57
Species.
Working Values of
Cross-breaking
Modulus in Pounds
per Square Inch.
(f)
Working Values of
Shearing Modulus in
Pounds per Square
Inch.
(:
64
THE MATERIALS OF CONSTRUCTION.
03 P
I
CO Ir^
CO
s
IB
Sit
8 =
II
ftJ ^ flfe
W w ^ lo
V II A 11
+ " S
cS
cs a ^
O S "O
'-CO ^
I 1 ^
V II A
g
I
01
S
I I
V II A
a
MATERIALS UNDER CROSS-BENDING STRESS.
65
*B
II
liS-*
Si
II II
3 ^
* !>?
0,100
I
THE MATERIALS OF CONSTRUCTION.
46. Deflections from Shearing Forces. For short beams it is necessary
to take into account the shearing forces also. Since the modulus of
elasticity in shearing is the ratio of the shearing stress to the angular
distortion (transverse distortion per unit of length, since the angle is equal
to its tangent), we may say that for a distance along the beam of dx where-
A'*
the shearing stress per unit area is s = r -, the differential deflection from.
j
shear is
(36)
To integrate this we have to express 8 and A as functions of x. The
cross-section, A, will be assumed as constant, and for a concentrated load
8 is also constant and equal to the supporting force on that side of the-
load. For a uniformly distributed load 8 is equal to the algebraic sum of
the forces on one side of the plane of shear, which here must be taken
normal to the deflection. Thus for any section distant x from the end of
p
the beam we have for a concentrated load at the centre 8 = , a con-
/v
stant, while for a beam uniformly loaded (supported at the ends in both
cases) S = ~-px=p(^-- xj.
' Using these values in eq. (36), we have for the
Deflection of a learn from shear when supported at its ends and
loaded at the centre,
Ys = ZETA X > r at C6ntre As = 4^1 ..... < 37)
Deflection of a beam from shear when supported at its ends and
uniformly loaded t
or
y s = -J-1 \ or at centre A s = - * . . . . (38)
Hence the total deflections at the centre for these two cases are
A = P ( l * , M
Us-Si ^ 4^^y
and
5r
r z 2 \
LE/+^Z> ( 40 >
* Assuming also that the shearing stress is uniformly distributed over the cross-
section.
MATERIALS UNDER CROSS-BENDING STRESS. 67
for beams supported at the ends and loaded with a single concentrated load
P at the centre, and with a uniformly distributed load of p per unit of
length, respectively.
For the metals take E $ = \E, while for wood take E, = \E. The
fibrous character of wood may explain the apparent anomaly.
For solid rectangular wooden beams, therefore, we have
For load at centre, from (39), making E 8 = \E, and r-= - -,
A. J. L JL
and
For beam uniformly loaded, from (40),
These equations show that when a rectangular wooden beam loaded at
the centre has a length less than seven times the height, the deflection
from shear is more than ten per cent of the total deflection, while for such
a beam uniformly loaded the deflection from shear exceeds ten per cent of
the total when the length is less than about six times the height.
47. Determination of Young's Modulus of Elasticity from Bending-
Tests. Since E enters in all the expressions for deflection of beams, it is
evident that it may be found from a bending test where all the dimensions,
loads, and deflections are observed. Thus for a beam of uniform, solid,
rectangular cross-section, supported at the ends and loaded at the centre,
we should have, from eq. (33), for a long beam where deflection from
shearing forces could be neglected,
Since in testing a beam the stress on the extreme fibre is also desired, the
i last form of this equation may be useful in case/ is also to be computed.
However, this value of/ must be inside the elastic limit in order to use it
in computing E.
It is best to measure a series of coincident loads and deflections, and
plot them as in Pig. 49, then draw a tangent to the curve at the origin
and use this in finding E. Thus the tangent line OA is used for comput-
i ing E, and the coordinates of any point on this line may be taken. It is
convenient to take a point representing a deflection of unity. On this
curve this corresponds to a load of 6250. The dimensions of the beam
were I 140 in., I = 4.02 in., h = 8,04 in., and the material was long-leaf
8 THE MATERIALS OF CONSTRUCTION.
yellow pine (Pinus palustris}. Using the second form of eq. (43)
we have
. Q ni s = 2,070,000 pounds per square inch.
The maximum load was 13,500 pounds, from which we find, by eq. (6)
tlie computed maximum stress on the outer fibre to be
f TT^J = p- jj-j = 10,000 pounds per square inch.
15000
10000
' 1
I
5000
1 Deflection 2 in inches 3
FIG. 49.
The elastic-limit load might be taken as 9000 pounds, whence the fibre-
stress at this limit would be
3 X 9000 X 140
- 2 X 4.02 X 8.04- = 73 ^ ^ ^ are inoh '
48. The Rational Designing of Flitched Beams.* A flitched beam is one
composed of two sticks of timber enclosing between them a wrought-iror
or steel plate of the full length of the sticks, these three members beinf
rigidly bolted together, preferably along the neutral plane, in such a waj
that they will act as one solid member when deflecting under a load. Ir
order that these two materials may come to their working stresses simul
taneously, the iron or steel plate should always be of less depth than thai
of the timber.
To find the relative depths of steel (or wrought iron) plate and tht
timber sides in order that they shall come to their working stresses at th<
* This problem is introduced here, not because it is very common or important ii
itself, but because it is a good type of composite systems and illustrates the methoc
of analyzing such systems.
MATERIALS UNDER CROSS-BENDING STRESS. 69
same time, we must utilize the principle that when two or more members
jointly carry a single load, they share this load in direct proportion to their
relative rigidities. The rigidity of a beam is the inverse of its flexibility,
and the flexibility is measured by the deflection under a given load. Hence
the rigidity will be measured by the reciprocal of the deflection. The
equation representing the deflection of a solid rectangular beam, in terms
of the stress on the outer fibre, is, from eq. (33), since
I PI
M = -fbh 2 - - for a load P at the centre.
o 4
pr 1
~
But since the rigidity is measured "by the reciprocal of the deflection,
we have as a measure of the rigidity of a rectangular beam, in terms of the
stress on the outer fibre,
5 = rigidity = -^ = ^ (45)
_/ /
We may now write the proportion:
Deflection of the t deflection of the t the rigidity of e the rigidity of
wooden beam steel plate the plate the beam,
or
But
Hence we have for a flitched beam, in which A w A s ,
A w R s E s h s f w
A -
A s : R -. R
, or 4- = *
(46>
Rs
A s R:
W EJi s f w
(47>
(48)
where R w = rigidity of the timber sides;
R a = same for the steel plate;
A w = deflection of the timber sides;
A s = same of the steel plate;
E w modulus of elasticity of timber = from 1,000,000 in white pine
to 1,800,000 in long-leaf yellow pine;
Es = modulus of elasticity of wrought iron and steel = 28,000,000;
P = total load on flitched beam ;
P w = load carried by the timber sides;
P 8 = same for the steel plate;
/, = working fibre-stress for timber = from 1000 in white pine to-
1800 in long-leaf yellow pine;
/ 8 = same for steel = 12,000 to 18,000 pounds per square inch;
70 THE MATERIALS OF CONSTRUCTION.
li w = depth of the timbers in inches;
h s = same for the steel plate;
b w = total thickness of both timbers in inches;
b s = same for the steel plate.
From eq. (48) we may derive many important relations:
(a) To find the relative depths of steel plate and wooden learns to give
simultaneous working stresses in each. Eq. (48) may be written
Example : Let E s = 28,000,000, E w = 1,400,000, f s = 16,000, / w = 1600;
then
h _ 28,000,OOOJ<_1600_ __ 9
1T S ~ 1,400,000 X 16,000 ~" ~*
That is to say, the wooden sides must be twice as deep as the steel plate,
regardless of their respective thicknesses, in order to give a working stress-
in the \vooden sides of one tenth that in the steel plate.
(b) To find the relative stresses on the outer fibres when the plate is of'
the full depth of the timber sides. We now put eq. (48) in the form
/KAX
f s E ...... (50)
Using the same values of fias above, and making h w = h s , we have
.U _ E w _ 1
fs El ~ 20'
Hence when the steel or iron plate has the full depth of the wooden sides,
the stress in the outer fibres of the timber is only one twentieth that in the
plate. This does pretty well for a white-pine and steel combination.
But in the case of white pine we should not take E w higher than
1,000,000. Hence we have for white pine and steel of equal depths
fw_E L _ 1,000,000 1
fa E 8 28,000,000 ~~ 28'
That is, the maximum stress in the timber is only ^ that in the steel plate.
For an elastic limit of steel of 40,000 pounds per square inch we may have
a working fibre-stress of 20,000 pounds per square inch. This would give a
fibre-stress of 700 pounds per square inch in the timber sides, which is-
hardly a sufficiently high working stress for white pine. All these conclu-
sions are quite independent of the relative thicknesses of plate and sides.
To find what part of the total load P is carried by the timber sides and
by the steel plate respectively, we may let P w and P s represent these loads,
so that P w -f P s = P. Also the total load P divides itself between the
parts in proportion to their respective rigidities, these rigidities being now
MATERIALS UNDER CROSS-BENDING STRESS. 71
taken as the reciprocals of the deflections when expressed in terms of the
equal loads W instead of fibre-stresses. From eq. (44) we have
R s = -T = pi S and R w = -p =
" s -L s i ^W
whence we have
P 7? F J
ft = t = fi ^
(53)
But for solid rectangular sections /= -fabh*', hence we have
P 9
P w E w b w li w
Jut P w = P P s , which substituted in (51) and reduced gives
P > = JOIO ( 54 >
Similarly,
(55)
Thus if the depths and thicknesses of the plate and of the timber sides be
Inown, the parts of the total load which they will carry can be found from
Iquation (54) or (55), or their relative values may be found at once from
Iquation (53).
EXAMPLE: Dimension aflitclied learn 24 feet long to carry a distributed
\oadof2000 Ibs. per foot.
Assume a depth of timber sides of 16 inches, and let the plate be the
all depth of the timbers. If we use "long-leaf" pine, we may take
?, = 1,400,000, while E s = 28,000,000 for the steel plate. Eq. (50) now
/ -
fives us = . That is, the maximum fibre-stress in the timber sides is
j s ~u
ne twentieth that in the steel plate. We will also assume the plate to be
inch thick. If it is stressed to 20,000 Ibs. per square inch, the load it
lone will carry is found from eq. (G). Thus
M s = ~ = M =' f ^~> or A = 12,000 Ibs., nearly.
|v
; ^is leaves 36,000 pounds to be carried by the timber sides.
But when the stress in the plate is 20,000 pounds, that in the timber sides
5 but 1000 pounds. Hence we must now find the combined breadth of the
wo sides to carry 36,000 pounds with this fibre-stress. Here again we have
M o =f^^, or b w = l^- t = 30 inches, nearly.
O *J w tl w
72 THE MATERIALS OF CONSTRUCTION.
As this thickness is out of the question, we might double the thickness
of the steel plate, making it 1 inch, when it will carry 24,000 pounds, leaving
24,000 pounds for the timber sides. This would reduce them to 20 inches
in width, or two sticks, 10 in. by 16 in. each.
If it were practical to obtain timbers 18 inches deep, they would serve
the purpose much better. (The student might redirnension the beam on
this assumption.)
It is evident from the above that there is no economy in combining iron
and wood in this manner. An iron or steel I beam or a plate girder should
always be used in such a case when this is practicable. The problem has-
been inserted here as a valuable exercise.
49. Steel and Concrete in Combination. It is now common to employ
steel wires or bars to strengthen the tension sides of concrete beams. To
analyze this case it is necessary to know the modulus of elasticity of the;
particular concrete employed, and at the age when its working strength is-*
first required. This property of concrete has seldom been observed (see;
Chapter XXX), but for good Portland-cement concrete it may be taken at:
1,000,000. For cinder concrete, such as is used in fire-proof flooring int
buildings, it is very much less, possibly not over 100,000.
Kef erring again to the general proposition that in composite structures the'
load divides itself between the systems in direct proportion to their relative-
rigidities, we conclude that for like areas, similarly placed, the rigidities are'
to each other as their moduli of elasticity. Since the modulus of elasticity
of steel is 28,000,000 and of the concrete, say, 1,000,000, it follows that one
square inch in section of steel resists for equal deformations as much as 28-
square inches of concrete similarly placed. To find the resistance of the-
combined material, therefore, substitute an amount of concrete for the steel 1
wire or bar equal to twenty-eight times-
its cross-section, adding this in the hori-
zontal plane of the steel bar, and then
treat this new form of section, as shown
in Fig. 50, as an actual beam of concrete.
By finding its moment of inertia, the
strength of the beam, when the concrete*
fails by cracking on the tension side, mayi
FIG. 50. Steel and Concrete in , fl
Combination. be found from the equation M 9 == ^-,
where/ is the ultimate tensile strength of the concrete ; /is the moment of 1
inertia of the transformed section; y l is the distance from the neutral axis>
of this section to the tension side of the beam ; and M is the moment of
resistance of the actual beam when the concrete cracks.*
* For a discussion of the case where the concrete cracks and the elastic limit of the
steel bar is reached, as well as for the case where the concrete cracks on the tension side
and then fails in compression because of the strength of the steel bar, including also the
case here treated, see an article by the author in Engr. News, Jan. 3, 1895, p. 10.
MATERIALS UNDER CROSS-BENDING STRESS. 72a
If b = breadth of concrete beam,
h = height "
a = area of steel bar,
77T
A = substituted equivalent area of concrete = a ~ 9
E s = modulus of elasticity of steel,
E c = " " " " the concrete used,
e distance from centre of beam to centre of bar,
d " " " " " " new neutral axis,
?/, = " " neutral axis to tension side of beam,
y t = " " " compression side of beam,
f t = stress on outer portion of the concrete on the tension side,
/= " " " " " " " " " compression side,
/ = moment of inertia of the transformed cross-section,
C* -I/"
/ = stress on outer portion of beam if no steel bar were used = ~ t ,
Oil'
m . -jrf, used for convenience (but is in fact the ratio of the lon-
a ti s
gitudinal rigidity of the beam to that of the steel bar),
then we have
Tensile stress on concrete at
bottom
Compressive stress on con- | _ /. _ My 9 __
crete at top j ~~ * c ~ / ~" ^
1 + TO+ 12_
If the steel bar be a flat plate and this be placed at the bottom of the
beam, but buried in the concrete, then e and we have
Tensile stress on concrete at bottom = f t ' = fA - -\ (A')
' \m-\- 4:)
and similarly
Compressive stress on concrete at top =// =/( -r). (B')
If the steel rod, or plate, be removed still farther from the body of the
concrete, by placing it in the lower side of a projecting rib of concrete, then e
becomes greater than -. Equations (A) and (B) will still apply to this case,
6
merely using the true values of e and h, not counting the projecting rib as
any part of the concrete beam. Thus if a concrete floor 4 inches thick be
supported by ribs every two feet, in the bottoms of which are steel rods 1
726 THE MATERIALS OF CONSTRUCTION.
inch square, so placed as to be 10 inches below the centre of the concrete
floor, then from equations (A) and (B) we have
f> = -f" = - 007/ and f ' = ' = -
f GM M
where / = m- = OQT-
If any particular ratio of compressive to tensile strength of the concrete
is to be developed, we may impose the condition^ = &; whence for the steel
/*
placed at the bottom side of ttye beam we have, from equations (A') and
whence, iox =
- - \j 9 if __ _
f t m a E s Jc 1
- V)E C
Thus if ^ = 5, we have, for ~ = , a = j|.
That is to say, if the steel plate covered the entire base of the beam, it
would have to be -fa as thick as the concrete and steel combined to satisfy
this condition, it being assumed in this and all former cases that the con-
crete does not crack on the tension side. Evidently it is impracticable to
develop the full compressive strength of the concrete by this construc-
tion, on condition that the concrete is to remain unbroken on the tension
si'de.
To find the total stress in the steel bar, we assume it to stretch the same
as the parts of the concrete beam adjacent to it; hence for any given position,
distant e from the centre of the beam, we have
Total stress on steel bar = ^(e d)-^ a = - - =j (D)
/J l + + ljf
If e --, this becomes, for the bar at the bottom,
Z
Total stress on steel lar at bottom ~ 4 ("ir
ill ~\~ t\ fv
For this case the tensile stress in the steel rod, in pounds per square
7~r Tji
inch is -W 1 /,, or it is -^ times as much as that in the concrete adjoining it.
H/ c lit c
This stress in the steel bar can never be more than from 2000 to 5000
pounds per square inch in rock or gravel concrete, but in cinder concrete it
would be very much more. To utilize the strength of the steel, therefore,
in rock concrete, it is necessary either to allow the concrete beam to crack
on the tension side or to remove the steel bars to the lower portions of
projecting ribs.
MATERIALS UNDER CROSS-BE2VD1NG STRESS. 73
50. Approximate Determination of the Strength of Flat Plates under
formal Forces.* (a) Flat Circular Plate Supported at the Circumference
and Uniformly Loaded. Assume a diametral strip 1 in. in width to be loaded
over its full width at the ends, but the loaded surface to reduce to a zero
width at the centre, this load to be p Ibs. per square inch. The total load
on the strip will then be pr, and each end support will be - . The bending
moment at the centre will be
pr pr pr*
M = . r . f r - - (56)
But for a solid rectangular section we have
M = ^fblf, or, for #, = 1;
whence
h = r\ l - (58)
where h thickness of plate in inches;
r = radius of plate " " ;
f = stress in extreme fibre in pounds per square inch;
p = pressure on plate in " " " "
From a very elaborate analysis, Prof. Grashof finds for this case
(b) Square Flat Plate Supported at the Periphery and Uniformly Loaded.
Since the corners are more distant from the centre and therefore carry a
less proportion of the load, we may assume that the opposite sides act inde-
pendently, so far as the bending moment at the centre is concerned. On
this assumption the plate may be regarded as supported at two sides only
and loaded with one-half the actual load, whence we have
M Q = ^pW=^fbh\ ........ (59)
or
1/4,
* = 11 =0.6111, ...... (60)
y / /
where I = length of one side of the square plate.
(c) Same Cases when the Plates are Fixed in Position at their Periph-
eries. Since the maximum bending moment on a beam fixed at the ends
and uniformly loaded is only J that of a beam supported at the ends and
* These proximate solutions are offered as illustrative of simple approximate methods
which may often be applied to very complicated problems of this class.
74 THE MATERIALS OF CONSTRUCTION.
similarly loaded, we may assume the same relations would hold here, thus
giving for a circular plate, rigidly held,
For a square plate, rigidly held,
or
(d) For Elliptical and Rectangular Plates. Here the plate fails by
cracking along its greater axis; and since the deflection of a beam for a
given load is as the cube of the length, it is evident that the ends carry but
a small part of the total load. Where the longer axis is more than twice the
shorter one, we may neglect these end bearings entirely when we have the
case of a flat plate supported at two opposite sides, which then becomes a
simple beam: and this is the proper assumption to make in such a case.
Making this assumption, and calling b the smaller dimension of the opening,
we have
Prof. Bach gives for this case
where is somewhere between f and 1.
When the longer axis is about 1J times the shorter, as is common with
manhole- covers, assume that f of the total load is carried at the sides, thus
giving, from (64),
f- or A -
-*' 4 K" - 4 r /"
CHAPTER VI.
THE RESILIENCE OF MATERIALS.
51. Resilience Defined. Resilience is literally the springing back of a
deformed body after the deforming force has been removed. As used in
mechanics, however, it is the work done by the body in this springing back,
which is the same as the work done on the body in deforming it, so long as
this is inside the elastic limits. Beyond the elastic limit the work of
deformation always exceeds the work given back by the body. The body
then does not fully recover its initial position, shape, or dimensions.
Sometimes the work of deformation, whether inside or beyond the elastic
limit, is spoken of as the resilience, but this is improper. The resilience
proper is the amount of work, or energy, in foot-pounds, ivhich can be stored
in an elastic bod?/, up to a given stress per square inch, and which can be
given out again by the body as useful work, if desired* That portion of the
energy spent in deforming the body but not given back as resilient work is
spent in permanently deforming the body, by causing the particles to move
or slide over each other, thus developing heat. The elastic deformation
of a body does not develop heat. Since work is measured by a force
acting over a distance, the work of deformation may be measured by
the product of the deforming force into the distance through which it
acts. But the deforming force is zero at first and increases uniformly as
the deformation increases (inside the elastic limit); hence the total work
done in deforming a body is the average value of the force into the total
deformation. Since the force increases uniformly with the deformation, its
average value is always one half its final value (inside the elastic limit), so
that the ivork of deformation, or the energy stored in the body, is one half
the product of the final force (or resistance}, into the deformation. Inside
the elastic limit the stress-diagram (for all kinds of stress) is a straight line,
and here also the resilience, or work given back, is equal to the work of
deformation. Hence the elastic resilience is equal to the triangular area of
the stress-diagram, included between this curve, the axis of deformation,
and an ordinate parallel to the axis of loads, to the extremity of the locus
developed. As similar areas are to each other as the squares of their like
*This is the sense in which Young first used the term in 1807, but he did not so
clearly define it since he assumed bodies to be perfectly elastic to rupture.
75
76 THE MATERIALS OF CONSTRUCTION.
parts, or dimensions, it is evident that the elastic resilience, or energy,
stored in a body is as the square of the unit stress under the finals load, this
unit stress being equal or proportional to the load ordinate in the stress-
diagram. It will be shown hereafter that this is true for all kinds of
resilience, both inside and beyond the elastic limit.
52. Three Varieties of Resilience. There are three kinds of resilience
commonly recognized, namely, of tension or compression, of bending,
and of torsion. These, of course, correspond to the three corresponding
kinds of stresses and deformations. It will be shown below that the
elastic resilience of a body, in foot-pounds or inch-pounds, is always equal
to the product of three factors, namely :
(a) A numerical coefficient, which is different for each of the three kinds
of resilience, and for different forms of cross-section, and for different
methods of applying the external forces.
f*
(b) The factor ~, or the square of the maximum stress divided by the
modulus of elasticity.
(c) The volume of the body.
That is to say, for any particular kind of stress and form of cross-
section the elastic resilience varies directly as the square of the stress-
intensity and as the volume of the body, and inversely as its modulus of
elasticity, or
f
R 8 = lc== . volume (1)
That is to say, the resilience, or energy, which can be absorbed, or stored,
in a body of a given material and form, up to a given fibre-stress, is no
function of the relative dimensions of the body, but only of its volume.
In other words, one cubic inch of steel will absorb and give out the same
amount of work, or energy, as the same volume of fine wire, if the load is
applied in the same manner, or if the stress is of the same kind, so long-as
the form of cross-section remains the same.*
53. Resilience a Measure of the Ability of a Body to Resist a Shock or
Blow. The magnitude or effect of a blow, or of a falling body, is measured
by the energy stored in the moving body at the instant of impact. In the
case of a body which has fallen freely in space under the action of gravity,
its energy is Wh, where W is the weight of the body (force of gravity), and
h is the distance through which the body has fallen freely (distance through
which the force of gravity has acted). In any case, the energy of the body
Wv*
is - , where v is the velocity in feet per second, and g is the acceleration.
y
* When the stress is direct (tension or compression), the form of cross-section is
immaterial. In bending and torsion, however, the form of section is important.
THE RESILIENCE OF MATERIALS.
77
of the force of gravity, or 32 feet per second. If a moving body, as a
falling weight, is stopped by striking a fixed solid body, which is here
assumed to be a test specimen, the energy of the moving body is spent in
one or all of the following ways :
(a) In deforming the moving body itself, either within or beyond its
elastic limit.
(b) In a local deformation of both bodies at the surface of contact, within
or beyond the elastic limit.
(c) In moving the fixed body as a whole, with an accelerated velocity,
the resistance consisting of the inertia of the body.
(d) In moving the fixed body against its external supports and resistances.
(e) Finally, in deforming the fixed body as a whole against the resisting
stresses developed thereby.
If the moving body be very hard and rigid; if the surfaces of contact are
comparatively unyielding; if the specimen have a small mass as compared to
the moving body, and if it be very rigidly supported upon or against a very
great mass or weight which is relatively unyielding; and, finally, if the spec-
imen which is to receive and absorb the energy of the blow is quite yield-
ing or flexible, and in short if there is nearly absolute rigidity in all parts of
the apparatus except in the body struck, and if this yields only as a whole
and not at the. point of contact or at its supports, then, and only then,
can nearly all the energy of the moving or falling body be absorbed by the
deflection or stretch or compression or twisting of the specimen. It is prac-
ticable, by making the energy of the falling body consist mostly of weight,
and only to a small degree of velocity, that is, by having a heavy weight drop
through a short distance, to absorb up-
wards of 90$ of it in the specimen. It
goes without saying that it is impossible
to get it all stored in the specimen
under any circumstances; and if great
care is not exercised in arranging the
test, but a very small percentage may
be given over to the specimen, the
rest being dissipated in the other ways
named above.
In the stress-diagram shown in Fig.
51 let the vertical ordlnate represent
total resistance in pounds and the hor-
izontal ordinate represent deformation
of the body, as a whole, measured at the
point of contact, in inches; whether this
deformation be a bending, extension,
compression, or twist is not now material,
formed to
d l it is resisting this action with a force of p
When this body has been de-
when deformed
78 THE MATERIALS OF CONSTRUCTION.
to t/ 2 it is resisting with a force of p^ , etc. When the deformation passes
the elastic limit the resistance does not increase as rapidly as the deformation,
and hence the diagram is no longer a straight line, but becomes curved.
A deformation of d 3 now develops a resistance of p z , and d 4 of p 4 , etc.
Now since the work of resistance is the sum of the products of the in-
stantaneous resistances into the corresponding deformations, it is properly
represented by the area of the stress diagram up to the maximum deforma-
tion and resistance. That is to say, the work done on the body to deflect or
deform it to d l is indicated by the area of oq l d l , the work required to de-
form it to d 9 , and also the energy stored in the body when deformed to this
point, is indicated by the area oq 9 d^ , etc. So long as the point q falls on the
stress diagram inside the elastic limit, this amount of energy stored will all
be given back again by the body. But when this point q falls beyond the elas-
tic limit point of the diagram, the body is no longer able to fully recover its
original form, but it remains permanently deformed. The amount of- this
permanent set can always be found by dropping lines q t q 9 'q 4 q t ' 9 etc., from the
extremity of the diagram which marks the maximum load imposed, parallel
to the straight portion of the curve. These lines are the return paths which i
the specimen follows on the removal of the deforming forces or loads. They !
are always parallel to the elastic path of the body, or to that part of the
curve below the elastic limit. This is true for all kinds of stresses and dia-
grams, whether tension, compression, bending, or torsion; and whether the
vertical ordinate represents total loads or resistances, or loads per square
inch, or intensities of stress on extreme fibres.
In case the specimen has had to absorb an amount of energy, or work, ,
represented by the area oq^d^, therefore, it will give back only so much as is
represented by the area q 3 'q 3 d 3 . The remainder, oq a q a ', represents work
which has been spent in permanently deforming the specimen, and which it
can never give back, this having been transformed into heat by friction.
Under our definition of resilience, therefore, we should have to say that the
resilience of the specimen, for the resistance p a , is q 3 'q 3 d 3 , and not the full
area 0^3. This latter represents the work done in deforming the specimen,,
but it cannot properly be called resilience. Similarly, when the body is dis-
torted to d t with a developed resistance of p^ the resilience now is q^^^,*
and oq 4 q t ' has been lost in the permanent deformation of the specimen, or
in heat.
The student will readily perceive that the areas of the triangles whose
bases are od l} od^, q 3 'd 3 , and q/d t , respectively, are to each other as the squares
of these bases, or as the squares of their altitudes, p l , p^ , p s , and p t , respec-
tively, since they are all similar, their sides being parallel. If their altitudes
represented stresses per square inch, which they might, then we could say
the resilience of this specimen varied as the square of the stress developed
in it, as stated in Art. 52, and as will be further shown analytically, whether
this maximum stress be inside or beyond the elastic limit.
THE RESILIENCE OF MATERIALS.
79
Thus far in studying Fig. 51 we have spoken of the "work of deforma-
tion " without stating whether this work was developed bv a load slowly ap-
plied, or by one quickly applied, as
by a falling weight. In fact it does
not matter how this work is done, a
given number of foot-pounds of en-
ergy producing exactly the same ef-
fect, and developing the same stress
diagram, provided we assume that
all the energy of the quickly applied
load goes into the specimen, to pro-
duce this particular deformation.
This conclusion is also based on
another assumption, which is, that the
relation betiveen the deformation and g^
its corresponding resistance devel-
oped in the body is the same for a def-
ormation produced suddenly as for ^
one produced by a slower application
of external force. This equality of
relationship has never been shown, as
between static and impact applica-
is Drobable FlG ' 52 '~ Showin that a reater . impact
stress is required to produce a given de-
formation.
344.)
(Fr. Com. Rep., vol. n. p.
tions of the load; but it
ithat this relation is very nearly inde-
pendent of time, inside the elastic
limit, and with brittle bodies up to rupture, since it is in this case a molec-
iular resistance to relative deformation, and not a resistance to t flow or rel-
iative displacement. In the case of plastic or ductile bodies, however, it has
jbeen shown that beyond the elastic limit the stress diagrams developed by
impact and by static loads are very different, the former being in the case
of soft iron wire some 30$ greater in area. This means that for such mate-
rials the actual energy absorbed by the specimen under impact is some 30$
more than it is under a static load. See Fig. 52.
Assuming now that all the energy of a blow is spent in deforming the
specimen in the manner represented on the static stress-diagram, we come
to this very important conclusion: the energy of the blow, in foot-pounds, is
equal to the area of the stress-diagram developed by that blow, properly eval-
uated to the scales of the drawing, when this diagram is drawn to co-ordi-
nates representing deformation and resistance thereto. Thus if a weight
falls on a body, as a beam, and if we may assume that very little of the
energy spends itself otherwise than in bending the specimen, if .the speci-
men deflects by the amount d^ , for instance, then we" assume that the corre-
sponding resistance at the instant of maximum deflection is p^, and that the
energy of the blow was somewhat greater than the area oq //; or by know-
80 THE MATERIALS OF CONSTRUCTION.
ing the energy of the falling body ( Wli), and observing the deflection pro-
duced, we could determine the amount of energy absorbed by the specimen
if we only had a stress-diagram of this specimen under impact, as shoivn in
Fig. 52. Few such diagrams have ever been obtained. It has been custom-
ary to use for this purpose static test-diagrams, carried beyond the elastic
limit, and perhaps to failure. This, of course, destroys the specimen for
impact tests; but by having two specimens, presumably just alike, a static
test may be made on one of them, from which a static stress-diagram can be
drawn, and then the impact test on the other specimen can be interpreted by
this static diagram. Having done this, we should still fail to find the area
of stress-diagram developed by a single blow fully equal to the energy of the
blow, because of the dissipation of a portion of this energy in other ways,
and also because the impact stress-diagram lies above, or outside of, the
static diagram.
It is common to test materials by means of falling weights, and often
the height of drop is regularly increased until failure occurs. Let us follow
the course of such a test, referring again to Fig. 51. Thus we will suppose
all the energy of the blow goes into the specimen (or it would serve as well
to suppose a certain fixed percentage is absorbed by the specimen), and that
the first blow deformed the specimen to d l , the second to d^ , the third to 6? 8i
and the fourth to d t . Now what were the energies of these blows, if all
went into the specimen each time ? Evidently the energy of the first blow
was that indicated by the area oq^l^ of the second by oq^d^ of the third
by oq^d^ and of the fourth by q^q^q^d^. Thus we see that all the first area;
is included in the second, all of both first and second in the third, and i
large part of the third (q s 'q z d s ) in the fourth. These areas are therefore
not mutually exclusive, so that the sum of the energies of all the blow*
(2W7i) is not equal to the total area of the stress-diagram developed b}
them, oq 4 d t . Neither is the energy of the last blow equal to this area, and
in fact there is no relation between the total area of the stress-diagrams
oq 4 d 4 , and the energy of one or all of the blows given. If we now add to
this statement the evident fact that we can never know in practice whai
portion of the energy of any blow is spent in deforming the specimen (ano
often we cannot tell what this proportion is within over 50$, and sometime) 1
it has been assumed that it all went into the specimen when there could no-
have been more than five per cent of it so spent !), it becomes patent thai
NO ABSOLUTE CONCLUSION WHATEVER CAN BE BASED ON IMPACT TESTS
Some relative conclusions may be drawn by subjecting two or more liki
specimens to exactly identical treatment and finding which withstands thlj
greater number of blows. Even then the apparent relative strength depend'
largely on wliat particular magnitude of blow be selected for making tests
Also a very small difference (apparently) in the character of the foundawH
on which the specimen rests may make a very great difference in the per
centage of the total energy which goes into the specimen. Hence suc!>
THE RESILIENCE OF MATERIALS. 81
omparative tests should always be made on the same foundation, and all
he elements of the test exactly duplicated.*
In order to obtain the absolute characteristics of any material, a com-
dete stress-diagram should be obtained by static tests. The area of such a
iagram up to its elastic limit indicates the total energy of the single shock
r blow it could withstand up to that limit, or without taking a permanent
et (provided this energy all went into the specimen), and the area of this-
.iagram, up to rupture, indicates the total energy of the single blow it
ould absorb without actually breaking.
It must also be observed that in addition to the impact stress being
reater for given deformations, beyond the elastic limit with ductile mate-
ials, the total elongation at rupture is also much greater when produced by
sudden blow than when produced in a static test, and hehce the area of
tie stress diagram thus developed may be very much larger than the static-
jest diagram on the same material.
For an absolute measure of a given material to withstand a shock or
low, therefore, it is necessary to give it a static test in some kind of a test-
bg-machine, whether this test be in tension, or in compression, or in cross-
lending, or in torsion. Then the area of the stress-diagram up to the elastic
imit, divided by the volume of the specimen under test, is a. measure of the
bility of the material, per unit of volume, to absorb and give out energy ',
r to resist repeated shocks without injury ; and the total area of the stress-
liagram is its measure to resist a single blow without rupture.\
It is necessary in this connection to guard, the student against several
iaisconceptions.
(a) By a " slow " or " static " test is meant such a gradual imposition of
he load as will give to the moving parts an inappreciable velocity, or
nomentum, or vis viva. Evidently any ordinary test in a testing-machine
ulfils this condition.
(b) By an impact test, or a shock, or a blow, is meant a genuine striking
>r impact, in which the force of the blow is nearly all due to the speed or
velocity of the moving body or falling weight, and only slightly due to its
static weight alone.
(c) Aside from the two methods which alone have been under dis-
cussion in this article, there is another method of loading, called a " sudden
imposition of load." Thus in the case of placing a load on a beam, if
the load be brought into contact with the beam, but its weight sus-
tained by external means, as by a cord, and then this external support be
*lt is not uncommon to find impact tests described by giving only the weight of
hammer and height of fall, with no description of the character of the supports. It
has also been customary to rest stamp-mills on spring-timbers to lessen the force of the
blow !
f Except that for ductile materials, in which the impact stress-diagram is greater
than the static stress-diagram, as shown in Fig. 52.
82 THE MATERIALS OF CONSTRUCTION.
suddenly (instantaneously) removed, as by quickly cutting the cord, thei
although the load is already touching the beam (and hence there is no re;
impact), yet the beam is at first offering no resistance, as it has as y<
suffered no deformation. Furthermore, as the beam deflects the resistam
increases, but does not come to be equal to the load until it has attained ii
normal deflection. In the meantime there has been an unbalanced force (
gravity acting, of a constantly diminishing amount, equal at first to th
entire load, but now reduced to zero when the resistance has come to t
equal to the load, at the normal deflection. But at this instant both th
load and the beam are in motion, the hitherto unbalanced force having pn
duced an accelerated velocity, and this velocity of the weight and beai
gives to them an energy, or vis viva, which must now spend itself in ove:
coming an excess of resistance over and above the imposed load, and th
whole mass will not stop until the deflection (as well as the resistance) he
come to be equal to twice that corresponding to the static load imposet
Hence we say the effect of a suddenly imposed load is to produce twice th
deflection and stress of the same load statically placed. It must be eviden-
however, that this case has nothing in common with either the ordinal)
" static" tests of structural materials in testing-machines, or with impac
tests. It is introduced here to prevent a confusion of mind in these mattei
often found to exist with persons whose conceptions of such problems i
mechanics are not clear.
54. Resilience Areas in Stress-diagrams. It was shown in the previoo
article, in discussing Fig. 51, that the shaded triangular areas represents
the resilience of the specimen for the several loads imposed. It will now fc
shown that these areas may be represented as one figure with continuousl'
added increments.
Referring again to Fig. 51, if the permanent set, oq s , be laid off on p^
from p 3 giving q 9 ", oq t on p 4 q t from p 4 giving rm cross-section and of a definite length is subjected to the action of
xternal forces, producing direct tension or compression, the deformation
rod need in the body, from eq. (2), Chapter I, is a = '-^. If A = the
lit
ross-section of the body, then the total external force applied is P =. pA.
he total external work is then
Pa _pA pi _ 1 p*
T '~2'^~2W Al (3>
But since this is equal to the internal work of resistance, and since
I volume of the specimen, we have
1 ff
R d resilience in direct stress .-^ . volume; ... (3)
& -&
r per unit of volume,
*=*f <*>
Since p and E are given in pounds per square inch, the volume must
so be in cubic inches.
If p is made equal to the elastic limit of the material, the corresponding
alue of r d is the primitive elastic resilience in inch-pounds per cubic inch.
eyond the elastic limit, the elastic resilience is indicated by the triangles
Va^s* y/Vtd*! e tc., in Fig. 51, corresponding in each case to the new or
irtificially-raised elastic limits p a , p t , etc. These subsequent elastic resili-
ence values may be called the artificially-raised elastic resilience.
As defined in the previous article, all resilience is elastic resilience, but
:he term "elastic" is retained here in order to insure that it is not confused
vvith the term "total resilience," which is sometimes misused and made to
nean the total area of the stress-diagram, which the author of this work
vvill not admit is resilience in any sense.
56. Resilience in Cross-bending. The deflections of beams loaded and
supported in different ways, in terms of the stress on the extreme fibre, are
* When the stress passes a maximum and both these curves begin to descend, the
included area here becomes negative.
f True under the author's definition of resilience, but not true when this term is
made to mean the work or energy absorbed instead of the energy given back.
THE MATERIALS OF CONSTRUCTION.
given in column four of the table on pages 62-65. For any case the
/7 2
deflection may be represented by the term k '~r 9 where k is a numerical
coefficient which varies for the different cases, but the values of which are
given in that table. Since the resilience of a beam when developed by
falling weights, or other impact loads, would produce deflections corre-
sponding to concentrated loads, only concentrated-load deflections as given
in the table need be here considered. Thus, for a beam supported at the^
ends and loaded at the centre, the deflection, in terms of the stress on the 1
outer fibre, is A = - ~r. But the load which will produce the stress /"on
Jjjil
the outer fibre is (see column five of table) P = -. The external work;
done on the beam in deflecting it, which must equal the internal work of !
resistance, or the resilience, if / is inside the elastic limit, is
- r i
Resilience = = -.-.1
For a solid rectangular cross-section, / =
Substituting this in eq. (5), we have
Resilience of a solid rectangular beam loaded at the centre and
supported at the ends =
1 /" 2 1 /' 2
. ^ . bhl = . --JT . volume.
lo Hi lo HJ
(6)
If the bending moment had been uniform throughout its length, as i&
the case with a spiral or helical spring when under a bending stress, the
movement of one end of the spring should be determined, and this multi-
plied by one half the final force applied at this point. But since the
internal work of resistance is always equal to the external work of deforma-i
tion, we may measure up the internal work and call this the resiliences
The case of a beam (or a spring) under a uniform bending moment is a
favorable case for this purpose. Thus -assume a spiral or helical spring
made of a steel bar having a rectangular cross-section whose original-
dimensions were I, b, and h. When coiled into a spring (the dimensions oi
the coil being immaterial for our purpose)
and a couple producing bending moment
applied to it, thus developing in the spring
throughout its entire length a moment oi>
resistance which we will suppose is such as
to give rise to the elastic-limit stress / on
the outer fibres throughout the entire length
of the coiled bar, we are to measure up the
total internal work of resistance, or th(
energy thus stored in the spring. Since the fibre-stress is uniformly vary-
FIG. 53.
THE RESILIENCE OF MATERIALS. 85
ag across the section of the bar, and is /on the outer fibres on each side,
3 is evident that
il 2f
The stress on any fibre = p = ay = f ~-y 9 . . . (7)
y i
'here y = distance of fibre from neutral axis, and
y^ = distance of outer fibre from neutral axis = ;
f = stress per square inch on outer fibre.
But
The stretch of any fibre = a = -~r = -by* .... (8)
here p = stress per square inch;
/ length of bar of which spring is composed;
E = modulus of elasticity.
Therefore
2^2 ^
The ivork of resistance of any fibre = - . -=-, . y*. . . (9)
J^J fir
The work of resistance of any zone of fibres bdy in area of cross-section,
istant y from the neutral axis, would be - . ^ . y*dy, and
r + Y
he total work of resistance = resilience = R = I L. _ .
J h E li
y*dy
_/' . N. *! - L.f- m
i~^ y 13-6 ^
~ 2 ~ 2
1 /"'
= -p. volume of spring ... .......... . . (10)
Comparing this with eq. (5) we see that fifty per cent more energy can
e absorbed by a beam or spring when subjected to a uniform bending
tioment than when the moment increases uniformly from the ends to the
entre, or from one end to the other,
57. Resilience in Torsion. Referring to Fig. 23 we see that the external
pork is , where is the distortion angle and a length of arm of the
ouple, whose forces are P. But from eq. (14), Chapter III, when the
aoment Pa is on the specimen the stress on the outer fibre is
= -& or
Uso, from eq. (15),
2Pal
86
THE MATERIALS OF CONSTRUCTION.
Pa6
-rr,we
Substituting here the value of Pa above, we have
%nr*E ~ 2 rJ
Combining this with the value of Pa again to get the value of
have
~P 'n f) \ f 2
Work of torsion on solid cylinder = -.%-. 7tr*l
2 8 E
= IT E' volume -
58. Comparative Resilience of Bodies under Different Kind of Stress.
For bodies of uniform cross-section we have the following table of values of
resilience in inch-pounds per cubic inch, and their relative capacities to ab-
sorb and give out energy, taking the capacity in direct stress as unity.
COMPARATIVE RESILIENCE OF BODIES.
Figure.
Kinds of Stress.
Resilience in inch-
pounds per cubic
inch. = r.
Relative Capacity
for Absorbing and
Giving out Energy.
22
'///M L
T W///////,
Direct tension or compression.
H'W
1
\f
Cross-bending with bending
moment uniformly increas-
ing longitudinally.
1 /*
US' E
i
9
'x
I
HSi ! i U^
^VJ 1 iS^
1
jjjjii^ \
Cross -bending with bending
moment uniform longitudi-
nally.
i r
6 ' E
1
3
f
C
(
v Ei'iiljii! !ii!!/t
izfomnA)
py
i
1 <
Torsion.
5 / 2
5
_ _
8 ' E
EXAMPLES ON PART I. 86a
EXAMPLES ON PART I.
1. A section of a steel bar 1 in. in diameter and 8 in. long elongates 0.01 in. for
an increase in tensile stress of 30,000 Ibs. What is the modulus of elasticity ?
2. What reduction in temperature would bring a wrought-iron bar, immovably
fixed at its ends, to its elastic limit of 26,000 Ibs. tensile stress per square inch?
Take E = 28,000,000 and the coefficient of expansion = 0.0000065 per degree F.
Am. 142. 8
3. Find the proportionate change in volume of a brass cube which is subjected to
a compress! ve stress in one direction of 10,000 Ibs. per square inch. Take E
15,000,000. What is its change in volume for a fluid pressure of this amount in all
directions? Ans. 0.00023- ; 0.00069.
4. What is the shearing modulus of elasticity for steel if # = 29,000,000 and
Poisson's ratio = 0.27 ? Ans. E 8 = 0.39^.
5. Find the modulus of elasticity of steel from Fig. 6, making allowance for the
locus cutting the vertical axis at 1000 pounds above the origin. Use the deformation
of 0.001 and its corresponding stress-increment in pounds per square inch. From
the same diagram find the elastic limit, the ultimate strength, and the percentage
of elongation.
6. The following is a record of a test on cast iron :
Loads per square inch
in pounds 1000 5000 10000 15000 20000 25000 30000 31040
Proportionate deforma-
tions .00022.00055 .00097 .00150 .00220 .00368 broke
Plot this record and determine from it: (1) The modulus of elasticity; (2) The ap-
parent elastic limit ; (3) The total percentage of elongation (by extending the plotted
curve till the breaking load is reached); (4) The work required to break the speci-
men in foot-pounds per cubic inch of metal (obtained by finding the area of the
diagram and evaluating it to the scales of the drawing, see Art. 53).
7. Assume a brick to be 8 in. long, 4 in. wide, and 2 in. thick. From equation
(12), p. 31, find the relative crushing strength of the brick per unit area when
i tested flatwise, edgewise, and endwise, taking the strength of a cubical specimen of
! the same material as unity. Ans. 1.22; 0.89; and 0.83.
8. A stone cube two inches on a side has its edges chamfered or rounded so that
the bearing surfaces are but 1.8 in. square. What is its total crushing strength as
compared to the strength of a full cube ? What is its strength per square inch of
bearing-surface as compared to the strength per square inch of a full cube ? (See
Fig. 18.) Ans. 85 per cent; 103 per cent.
9. By how much is a centrally loaded column 12 in. square weakened by adding
four inches of the same material to one side of the column without shifting the load ?
Ans. The maximum stress in the column is increased by 31 per cent.
10. A steel rod 1/4 in. in diameter and 30 in. long is used as a torsional spring for
closing a door. What will be the increase in the moment of torsion from giving
the rod an additional twist through 90, the shearing modulus of elasticity being
taken as 12,000,000 ? What will be the maximum increased shearing stress in the
rod due to this angular movement ? What will be the increase in the force required
to hold the door in this position, the door-knob being 30 inches from the hinges?
Ans. 244 inch- pounds; 78,500 Ibs. per square inch ; 8 pounds.
11. A wooden beam 8 in. by 16 in. in cross-section and 20 ft. long carries a
uniform load of 1000 Ibs. per running foot. Find the maximum direct stress on the
outer fibres and the maximum shearing stress in the beam.
A j 1170 Ibs. per square inch direct stress;
Ans. | 11? k4 tl tl M shearing u
12. For the same beam and load as in Ex. 11, find the deflection of the beam,
taking E = 1,200,000. If the deflection were observed to be 3/4 in., what would be
the modulus of elasticity? Ans. 1.1 in deflection; 1,760,000 modulus.
13. A flitched beam is composed of two sticks 4 in. by 16 in. by 16 ft. long-, and
a steel plate 3/4 in. by 16 in. of the same length, and carries a load of 2000 pounds
866 THE MATERIALS OF CONSTRUCTION.
per running foot. Find the portion of the load carried by each part, the maximum
fibre-stresses resulting, and the deflection at the centre, taking E = 80,000,000 for
the steel and 1.500,000 for the timber
A S Steel: 1305 Ibs. per foot; 15,660 Ibs. ; 0.25 inch.
\ Timber: 695 " " " 782 " " "
14. How many foot-pounds of energy per pound of steel can be stored in a steel
helical or spiral spring coiled about an axle, by winding it up until the stress in
the outer fibre is 80,000 Ibs. per square inch, E being taken equal to 30,000,000 ?
Am. 17.8.
15. How much would such a spring weigh which could absorb the energy of a
street-car weighing 20,000 Ibs., and moving at the rate of six miles per hour on a
down grade just sufficient to overcome the frictional resistances ? Would the size of
the cross-section of such a spring affect its necessary weight ?
Could such a spring be designed so as to reach this fibre-stress when the car had
stopped, and also so as to be exerting the maximum torsional moment on the car-
axle without causing the wheels to slip ? Is such a device practicable ? *
Am. 2315 Ibs.
16. To what extent can energy be stored in metallic springs of any sort ? Could
they ever be used for the storing of motive power ? (This has often been attempted.)
17. A pendulum, mounted on knife-edges, weighs 50 Ibs., and its centre of gravity
is 8 feet from the pivot-supports. It is moved to an angle of 30 from the vertical,
and is allowed to swing and strike the centre of a cast-iron bar 1 in. square, resting
on absolutely rigid supports (or assumed to be such) 2 feet apart. The pendulum in
falling breaks the bar and moves a horizontal distance of 24 in. beyond its true ver-
tical position before it comes to a stop. What is the shock- resisting capacity of the
iron in inch- pounds per cubic inch of metal in cross-breaking under a concentrated
load? Ans.ZOA.
18. Assuming the stress-diagram of a static test of such a bar in cross-bending to
be a triangle, what would its final deflection be if the breaking load were such as to
correspond to a modulus of rupture of 40,000 Ibs. per square inch ?
Ans. 0.88 inch.
* This is a favorite device with " car-starter" inventors. See an article by the Author in Jour. Assc,
Eng. Socs., vol. iv. p. 393.
PART II.
MANUFACTURE AND GENERAL PROPERTIES OF THE
MATERIALS OF CONSTRUCTION.
CHAPTER VII.
CAST IRON.
GENERAL CLASSIFICATION OF IKON AND STEEL.
59. Importance of the Subject. While the use of iron in a small way,
for offensive and defensive weapons of war and for utensils, is doubtless
older than authentic history,* it is only since its manufacture has become
possible on a grand scale, by the aid of steam-power, that it has become a
common material of engineering and architectural construction. It has
now nearly replaced the use of timber in engineering works, and it is rapidly
replacing the use of wood, stone, and brick in architeccural designing. So
dependent now are all kinds of construction on the use of iron, that the
condition of the iron-manufacturing industry is universally regarded as a
true index of the general state of trade and commerce the world over. Since
iron, therefore, in its various states, is more used in engineering construc-
tion than all other kinds of materials combined, a corresponding amount of
space is given to a study of it in this work.f
* There is now in the British Museum () a sickle-blade found by Belzoni under the
base of a sphinx near Thebes; (&) a blade found by Col. Vyse embedded in the mortar of
one of the Pyramids ; (c) a portion of a cross-cut saw exhumed by Layajd at Nimroud.
These may be of meteoric origin. The reason more specimens of iron and steel are not
found may be due, however, to their rapid oxidation when exposed to air and moisture.
The stone and bronze implements have resisted this action, and hence many have assumed
that in the "stone " and " bronze " ages no iron was in use.
f A chronological review of the greatest discoveries and inventions in iron manufac-
ture is here given :
4000 B c. to | Wrought iron by the direct process from the ore in small quantities
about 1500 A. D. ) by means of charcoal, and this made into cement- or blister-steel.
About 1500 A.D. Cast iron made in Germany with charcoal.
1620-1735. Cast-iron made by Dud Dudley in England with coke, but the prac-
87
88 TEE MATERIALS OF CONSTRUCTION.
60. Classifications of Iron and Steel. Iron and steel may be classified
according to its qualities, structure, and composition, or according to its
methods of manufacture. Apparently the former is the more significant
basis of classification, but in English- and French-speaking countries the
latter basis has come to be universally adopted. We will, however, here first
classify these products according to their more significant qualities (the
method used in Germany).
IRON AND STEEL CLASSIFIED ACCORDING TO QUALITIES.
Malleable.
Cast, when molten, into a malleable mass or ingot.
Ingot Iron cannot be hardened by sudden cooling.
Ingot Steel can be hardened by sudden cooling.
Aggregated from pasty particles without subsequent fusion
(puddling process).
tice lapsed till revived in 1735 by Abraham Darby. Blast from leather
bellows driven by water-power.
1740. Cement-steel melted in crucibles by Huntsman near Sheffield, England.
1760. The steam-engine of Watt applied to produce the blast for making pig-
iron, and to drive rolls and hammers in working the wrought iron
and steel.
1783-4. Grooved rolls of various forms, driven by the steam-engine, and
wrought iron made from pig iron by "dry-puddling," both by
Cort, England. White iron used in the "dry" process. These
inventions lie at the base of the supremacy of Great Britain in the
iron trades.
1829. Hot blast, used in blast-furnaces in Scotland by Neilson, thus greatly
cheapening the cost of production.
1830. The "wet-puddling" process of making wrought-iron, or "pig-boil-
ing," introduced by J. Hall, England.
1840. Use of manganese in making crucible cast steel, introduced at Sheffield
by /. M. Heath, which reduced the cost of steel by 50 per cent.
1856. The Bessemer process of making steel, patented by Sir Henry Bessemer
in England (son of a French refugee, born 1813), this " being of far
more importance to the world than all the gold of California and
Australia."
1861. Invention of the regenerative gas furnace by Sir W. Siemens in Eng-
land (born in Hanover, 1823), and educated at the Magdeburg Poly-
technicum and at Gottingen).
1863. Application of the Siemens furnace to the open-hearth process of mak-
ing steel by P. and E. Martin in France, thus originating the Sie-
mens-Martin process of steel-making now employed for nearly all the
soft and mild steel used in structural work, and for steel castings.
1878. The invention of the basic process of making steel, by which the phos-
phorus of the ore is eliminated, by 8. G. Thomas and P. C. Gilchrist
(cousins) in England. (First public demonstration April 4, 1879.)
By this process the range of ores which can be used for steel-making-
is enormously increased, especially in Europe, while it is used often.
in America.
CAST IRON. 89
Weld Iron cannot be hardened by sudden cooling.
Weld Steel can be hardened by sudden cooling.
Semi-malleable.
Steel Castings malleable metal cast into final forms.
Malleable Cast Iron non-malleable metal (cast iron) cast into
final forms and then brought to a semi-malleable condition.
Non-malleable.
Cast Iron.
Hard Cast Steel.
The significant criterion here employed to distinguish between iron and
steel consists in the hardening effects of sudden cooling from a bright-red
heat. This is not a very satisfactory criterion, however, since all such
metal is hardened somewhat by sudden cooling.
What are commonly known as wrought iron and steel, however, are
made by radically different processes one being the formation of the
product in a melted, or liquid, state and then casting it into a mould,
forming what is called an ingot; the other consisting in forming the
product in a pasty or spongy state in a bath of melted or liquid foreign
matter, from which it is lifted and immediately forged or rolled. When
formed in the melted state it is purified from all foreign matter except
such as enters into its own composition, while when formed in the pasty
or spongy state, in a bath of melted foreign matter, a considerable
proportion of this foreign matter or slag is, of necessity, lifted out with
the pasty aggregation, called a " puddle-ball," and some of this slag
remains distributed through the iron even after it is rolled, thus giving
it a kind of fibre or grain. While, therefore, the mechanical qualities of
the puddled product may be almost identical with those of the cast product,
there is always a sufficient difference in their structure, resulting from the
radical differences in their methods of manufacture, to clearly distinguish
them by simply examining the fracture, and to warrant a classification on
this basis also: and this is the customary basis of classification in this
country.* We have, therefore,
IRON AND STEEL CLASSIFIED ACCORDING TO METHOD OF MANUFAC-
TURE.
Malleable.
Wrought Iron rolled or forged from a puddle-ball; it contains
slag and other impurities, and cannot be hardened by sudden
cooling.
Steel rolled or forged from a cast ingot and free from slag and
similar matter.
* It has also been recommended by a sub-committee of the recent French Commis-
sion.
90 THE MATERIALS OF CONSTRUCTION.
Soft Steel will weld (with care), and cannot be hardened by
sudden cooling (Ingot Iron). Same uses as Wrought Iron.
Medium Steel will weld imperfectly except by electricity), and
will not harden by sudden cooling. Used in Structural Work.
Hard Steel will not weld, and will harden by sudden cooling.
Tool-steel, Spring-steel, etc.
Semi-malleable.
Steel Castings Malleable metal cast into final forms.
Malleable Cast Iron non-malleable metal cast into final forms
and then brought to a semi-malleable condition.
Non-malleable.
Cast Iron.
Hard Cast Steel.
Neither of these classifications must be construed too rigidly, but they
fairly define the common usage, so far as the employment of these materials
in engineering design is concerned.
THE PHYSICAL PROPERTIES OF CAST IRON.
61. General View. While cast iron has been known and commonly em-
ployed since the Middle Ages, it has not been critically and scientifically
studied till within a very few years. In the last quarter of a century, the
attention of metallurgists engaged in iron-manufacturing industries has
been almost wholly confined to the manufacture of steel. The great advances
which have been made in this direction have caused cast iron to be very
largely replaced by steel in structural designing, and in other directions,
and since 1885 steel has also been cast in final forms, the same as cast iron,
so that the use of cast iron has been very much diminished, relatively to the
total iron and steel output. For many purposes, however, cast iron will
probably never be replaced by any other material, especially since great
improvements have been made in this direction, as a result of scientific
study and experiments devoted in recent years to the manufacture of cast
iron.
Much of the matter here given on this subject has been quoted directly
from the Metallurgy of Iron by Thomas Turner, Associate of the Royal
School of Mines, England. This work was published in 1895 and contains
the latest results of scientific research on the subject there treated.*
62. General Properties." Cast iron consists of metallic iron, together
with at least 1.5 per cent of carbon. It also contains silicon, sulphur, phos-
phorus, manganese, and other elements in greater or less proportion, but
these may be regarded as impurities, though their presence is often useful
or even necessary for the purposes for which cast iron is applied. The pro-
* When not otherwise credited the quoted paragraphs are from this work. (Chas,
Griffin & Co., London, and Lippincott, Philadelphia.)
CAST IRON. 91
portion of elements other than iron is usually about 7 per cent of the total
weight, though this varies considerably and is sometimes very much more.
Cast iron is fusible at a temperature of about 1200 C. (2200 F.); when cold
it is hard and brittle, some varieties being much more so than others;
it is not malleable or ductile, like wrought iron or mild steel, nor can it be
hardened and tempered like ordinary carbon steel, f^he iron-founder distin-
guishes between pig iron, or the form in which the metal is obtained from
the blast-furnace, and cast iron, or the form it assumes after it has been
again melted; but no such difference is recognized by the chemist, and pjer
iron is merely a variety of cast iron which is produced in a particular formTN
63. Carbon in Cast Iron. " Cast iron, when fused, consists of a saturatecTf
or nearly saturated, solution of carbon in iron. The amount of carbon
which molten iron can thus dissolve is about 3J per cent of its own weight,
though the solubility is largely influenced by the presence of other elements.
With much chromium the maximum solubility of about 12 per cent of car-
bon is reached; with much manganese up to 7 per cent of carbon may be
dissolved; while with about 20 per cent of silicon the minimum solubility
of carbon is obtained, and only about 1 per cent of carbon then dissolves.
Apart from special alloys, such as those mentioned, it is very unusual to
meet with less than 2 per cent or more than 4.5 per cent of carbon in cast
iron.
" So long as iron containing some 3 per cent of carbon remains in the
fluid condition the composition is uniform throughout, and the carbon has
no tendency to separate from the metal, except with very gray iron; in this
case a layer of graphite, which often occurs in beautiful plates and is known
as kish, may be formed. But when the molten cast iron is cooled to a tem-
perature at which it begins to solidify, it may either retain tke carbon and
solidify in a relatively homogeneous form, called white iron; or it may, in
solidifying, precipitate the greater part of the carbon in the form of small
scales of graphite, which, being entangled by, and uniformly distributed
through, the iron, impart to it a somewhat spongy nature, and produce the
dark color and soft character met with in gray iron. When about half of
the carbon is precipitated as graphite, and the rest retained in combination,
the result is the production of dark gray portions in a matrix of white, and
the iron is then said to be mottled.
" The condition which the carbon assumes on the solidification of the
mass is dependent partly on the rate of cooling, and still more on the
nature and quantity of the associated elements. In connection with the
influence of cooling, cast iron obeys the laws which govern other solu-
tions, for it is well known that slow cooling assists the production of
crystals, and leads to the formation of crystals of larger size, -while with
rapid cooling both solvent and the substance dissolved may solidify together.
In a similar manner slow cooling tends to produce graphitic carbon, and
the slower the cooling the larger are the flakes of graphite which sepa-
92 THE MATERIALS OF CONSTRUCTION.
rate. Some kinds of white iron may thus be rendered * gray by slow cool-
ing, while some kinds of gray iron may be made perfectly white by rapid
cooling or 'chilling/ It is, however, only with intermediate irons that
the rate of cooling produces a marked effect, for irons which are either
very white or very gray cannot be changed in this manner. The influence
exerted on the condition of the carbon by the other elements present in
cast iron is of the greatest importance; thus manganese and chromium,
which increase the solubility of carbon in iron, lead to a greater percentage
of total carbon in the fluid metal, and when the iron solidifies this carbon is
retained in solution, so that irons rich in manganese and chromium are white
and no amount of slow cooling will alter this character. On the other hand, v
silicon and aluminum diminish the solubility of carbon in iron; if much of
either of these elements be present in the fluid metal, it is capable of dis-
solving less carbon, and retains it with less energy when it solidifies; as a
result the carbon is precipitated as graphite, and gray iron is produced.
Just as irons which contain much manganese or chromium are permanently
white, so metal rich in silicon or aluminum is permanently gray.
" The proportion of total carbon in iron to be employed for a given pur-
pose is often of secondary importance; it is governed by furnace conditions,
and by the proportion of other elements. A moderate alteration in total car-
bon, or in the graphite, will frequently have little effect on the physical
properties of the product, while a small change in the combined carbon
will profoundly alter the strength and hardness of the casting. Probably
no other constituent in cast iron is of importance equal to that of combined
carbon, and the influence of the other elements is largely due to the effect they
produce in increasing or diminishing the combined carbon. The following
percentages of combined carbon will usually be found suitable for the pur-
poses specified:
* Combined Carbon in
parts of one per cent.
Extra soft siliceous gray iron 0.08
Soft cast iron 0.15
Maximum tensile strength 0.47
Maximum transverse strength 0.70
Maximum crushing strength over 1.00
These figures are, however subject to some variation according to the
size of the casting and the proportion of other elements. The hardness of
the metal increases regularly with the increase of combined carbon."
64. Silicon in Cast Iron. "All cast iron contains silicon, in quantities
* Chemical ingredients of iron and steel are always given in hundredths of one per
cent. Thus " twenty carbon " and " eight phosphorus " signifies 0.20 and 0.08 of one per
cent of each, respectively. Even the common workmen use these terms, though they
may not always understand them. The word point is often added, as " twenty-point
carbon." J. B. J.
CAST IRON. 93
varying in ordinary cases from under 0.5 to over 4 per cent; while 'silicon
pig ' is made in the blast-furnace with from 10 to 18 per cent of silicon.
No factor is of greater importance in determining the suitability of a sample
of cast iron for any purpose in the foundry than its content of silicon, as
this element is so constantly present, and its proportion is so. variable, while
the influence it exerts on the condition of the carbon present, and conse-
quently on the hardness and fluidity of the rnetal, is so marked. It was
formerly very generally held that silicon was injurious in all proportions,
and the less there was present in iron for foundry purposes the better. It
is true that Sefstrom had observed, long ago, ' that the carbon in gray iron
in which much silicon exists, say from 2 per cent to 3 per cent, is wholly,
or nearly so, in the graphitic state/ * A similar observation was made by
Snelus in 1870, and was still more plainly stated by Ledebur in 1879. It
was also known in the United States that certain irons from Ohio which
were rich in silicon could be used as ' softeners ' in foundry practice, and
certain Scotch irons -were in favor for similar purposes, though the reason
of this was not understood. It may, however, be claimed that no general ap-
plication of these facts, or accurate knowledge of the principles underlying '
them, existed before the researches of the author, on the ' Influence of Sil-
icon on the Properties of Cast Iron/ published in 1885. f For the purpose
of these experiments cast iron as free as possible from silicon was specially
prepared by heating wrought iron with charcoal to a high temperature in
closed crucibles. This was then remelted with a silicon pig containing
about 10 per cent of silicon in proportions necessary to yield any desired
composition. The trials were made with sufficient material to allow of
proper mechanical tests being performed, and a graduated series of mixtures
was prepared. The tension, compression, and ductility tests were performed
by Professor A. B. W. Kennedy with the testing-machine at University Col-
lege, London, while the hardness determinations were performed by the
author with a weighted diamond point (see Chapter XVIII) as described
in his paper on the ' Hardness of Metals/ J The chemical analyses were
checked by J. P. Walton, at that time chemist to the Glasgow Iron Com-
pany, Wishaw."
"The original pure cast iron was white, hard, and brittle; on adding
silicon this became gray, soft, and strong; but with a large excess of silicon
it once more became weak and hard. The results of the mechanical and
chemical tests are shown graphically in Fig. 55, and it will be observed that
the proportions of silicon corresponding to the various properties were as
follows:
* Percy, p. 131.
f Journ. Chem. Soc., 1885, pp. 577, 902.
\ Birm. Phil. Soc., Dec. 1886.
94
THE MATERIALS OF CONSTRUCTION.
Maximum hardness under 0.80 per cent*
" crushing strength about 0.80 "
" modulus of elasticity " 1.00 "
" combined crushing and tensile strength; trans-
verse strength about 1.40 "
" tensile strength , " 1.80 "
" softness and working qualities, " 2.50 "
/?/SS/0M .
" 7FA/3/0M o
' , &M$l$ffffXMM0 x
* ,, Wff.ffffMST. a
,, //JffffA/f? m
44000
35000
02 40 8/0
FIG. 55. Showing the Influence of Silicon on the Strength and Hardness of Cast Iron,
(From Turner's Iron.)
" It must be borne in mind that these values are only true for the author's
experiments. Experience has since proved that these are approximately cor-
rect in other cases, and that the order is as above given; but in practice the*
size of the casting and the proportion of other elements will have an impor-
tant influence."*
The influence of silicon on the shrinkage of cast iron, in various sizes tip
to 4 inches square, is well shown in Fig. 56. These results have been well
* See also Arts. 76 and 79.
CAST IRON.
95
established by Mr. W. J- Keep of Detroit. His results of transverse tests of
strength and deflection, for varying proportions of silicon, are given iniigs..
57 and 58.
4/40
a/00
M70
0.0S0
\\
&
m
.M .2$ \.30 .4$ .W &
1 2"*
W//
.30 /.W
FIG. 56. Showing the Influence of Silicon on the Shrinkage of Cast-iron Specimens
of Various Areas of Cross-section. (Keep.)
" A small addition of silicon eliminates blowholes and produces sound
castings. As soon as the metal is sound, with the least graphite, tJie great-
est crushing strength is obtained; this condition also gives th'e maximum
density, Further addition of silicon leads to the formation of graphite, di-
minishes the brittleness, and gives the greatest transverse and tensile
strength. When the graphite increases beyond this point, the metal is di-
vided by the interspersed graphitic material, and the strength and hardness
decrease. The deflection also increases with the increase of graphite, but
when the maximum separation of graphite has taken place any further addi-
tion of silicon causes stiffness or brittleness, and so diminishes the deflection.
White iron shrinks during solidifying more than gray iron, while highly
96
THE MATERIALS OF CONSTRUCTION.
siliceous iron shrinks still more than white. Hence on adding silicon to
white iron the shrinkage is diminished, but an excess of silicon, on the other
hand, leads to increased shrinkage. Shrinkage appears to closely follow the
hardness of cast iron, hard irons almost invariably shrinking most; and as
carbon into the graphitic form the maximum softness was obtained. With
further additions of silicon the metal became harder owing to the hardening
effect of silicon itself, and for this reason an excess of silicon, beyond about
3 per cent, is injurious to the working qualities of the metal."
* W. J. Keep, Silicon in Cast Iron, p. 22. \ Kohn, Iron Manufacture, p. 57.
\ The French Commission recommend this tesc for hardness, using impacts of known
value in place of static pressure. See Chap. XVIII. J. B. J.
CAST IRON.
107
76. Hardness and Strength of Cast Iron. " When cast iron has to be
turned or otherwise worked the hardness is of considerable importance,
while in some cases smoothness of surface and general perfection of the cast-
ing are of the utmost moment. Hard cast iron is brittle, deficient alike in
crushing, transverse, and tensile strength, and seldom gives smooth clean
castings. With metal which is a little less hard the maximum crushing
strength is obtained; while on rendering it a little softer, or, as the workman
would call it, ' moderately hard,' the maximum transverse strength is
observed. With slightly softer cast iron the highest tensile tests are
obtained, while still softer metal works with the utmost facility, though it is
deficient in strength. It will be seen, therefore, that when the general
connection between hardness and strength has been fully grasped, the iron-
founder requires only the information how to harden or soften his. metal at
will, by the use of silicon or other agents, to be able to produce castings in
which crushing, transverse, or tensile strength shall predominate as desired,
or in which softness and fine surfaces shall be the most characteristic feature.
" There is a somewhat prevalent idea among founders that if considerable
strength is required a hard iron must be employed. Doubtless this is to some
extent true in connection with crushing and transverse tests, but is certainly
not correct with tensile strength. In all specimens of exceptionally high
tensile strength examined by the author the metal was a soft good working
iron, specially suited for engineers' purposes. In the accompanying table
is a summary of the author's results on the tenacity and hardness . of cast
iron, as affected by alterations in the proportion of silicon.* The working
qualities of the specimens are also given, and it will be seen that the hard-
ness as determined by the sclerometer agrees very closely with the observa-
tions of the workman. It will be noticed, however, that hardness and
tensile strength do not vary together, but on the contrary high tensile
strength is met with in the softer irons."
TABLE I. INFLUENCE OF SILICON ON THE
OF CAST IRON.
HAEDNESS AND TENACITY
Tensile
Hardness
No.
Silicon
per cent.
Strength
in Lbs. per
by
Sclerom-
Working Qualities, as determined by the Workman.
Sq. In.
eter.
1
0.19
22,700
72
Very bard indeed.
2
0.45
27,600
52
Very hard, though not so hard as No. 1.
3
0.96
28,500
42
Hard, though softer than No. 2.
4
1.96
35,200
22
Good, sound, ordinary, soft-cutting iron, of ex-
cellent quality.
5
2.51
32,800
22
Rather harder than No. 4.
6
2,96
27,400
22
Like No. 4.
7
3.92
25,300
27
Like No. 6, but rather harder.
8
4.75
22,700
32
Rather harder than No. 7, though not unusually
hard.
9
7.37
12,000
42
Still harder, cutting very like No. 10.
10
9.80
10,600
57
Hard-cutting iron, though still softer than No. 1.
* Journ. Cliem. Soc., 1885^
108
THE MATERIALS OF CONSTRUCTION.
77. Crushing Strength. " Cast iron possesses an exceptionally high
crushing strength, and for the majority of purposes the founder relies upon
this, and does not perform special tests. Usually the tensile strength is not
above one sixth of the crushing strength ; hence if power to resist a tensile
force is assured, the crushing strength is usually sufficient for ordinary work,
In performing compressive tests it is necessary to have perfectly parallel sur-
faces, and to bed the specimen as true as possible, otherwise the results will
be low."
As shown in Articles 19 and 22, the height of a crushing-test specimen
of such a material as cast iron should be not less than about twice its
least lateral dimension. Test specimens of cast iron are usually cylinders
which have been turned up in the lathe, and hence the length of such a
specimen should be not less than twice the diameter." The following table
of values contains the results of a great many tests of cast iron in compression
on short cylinders, the dimensions being usually one inch in diameter and
from two and one-half to three inches high.
TABLE II. CRUSHING STRENGTH OF CAST IRON.
Experimenters.
Pounds per Square Inch.
Authorities.
Max.
Min.
Mean.
Hod"kinson
146,000
121,000
140,000
215,000
207,000
82,000
55,000
44,500
92,000
77,000
107,000
86,000
91,000
Pairbairn, Iron, 1869, p. 218.
Pole, Iron Construction, p. 84.
Report, 1858, p. 2.
B. A. Report, 1853, p. 87.
J. Chem. Soc., 1885, p. 907.
Hodgkiiison (1849)
Woolwich (1858)
Fairbairn
Turner
" The average crushing strength of British cast iron is thus about 90,000
Ibs. per square inch; exceptionally, results so low as 45,000 Ibs. have been
observed, while, on the other hand, a strength of upwards of 200,000 Ibs.
has been produced in some instances. In the above experiments no special
pains were taken to produce an iron possessing a high crushing strength; on
the contrary, only such irons were taken as were met with in commerce.
In the light of modern researches, iron could doubtless be produced with a
crushing strength of 225,000 Ibs. to the square inch, while a strength of
150,000 Ibs. could, if necessary, be regularly assured. A series of sketches
illustrating the fractures of test-pieces with different proportions of silicon,
when subject to a compressive force, are given in the author's paper on
' Silicon in Cast Iron.' * The samples were prepared by the author, and the
mechanical tests performed by Professor Kennedy at University College.
From these experiments it is probable that the maximum crushing strength
would be obtained with about 0.75 per cent of silicon and 2 percent of
combined carbon."
* Journ. Chem. Soc., 1885, p. 909.
<3-
CAST IRON.
109
78. Transverse Strength. " As before stated, the maximum transverse
strength is obtained with metal a little softer than that which possesses the
highest crushing strength. Transverse strength depends, at least in part,
on the power to resist both a crushing and a tensile force; hence transverse
strength is intermediate between crushing and tensile so far as the charactei
of the iron is concerned. This combination of properties imparts to the
metal characters which are most valuable in certain cases. For transverse
tests many shapes and sizes of test-bar have been adopted, and, for scientific
purposes, the results so obtained are converted by calculation into breaking-
stress on the extreme fibres in pounds per square inch, which is called
the modulus of rupture. See Art. 33. For a load in the centre of a
3 PI
rectangular bar we have/ = - c-y^, where
/v bli
f modulus of rapture in cross-breaking;
P = breaking load, at centre ;
I length of bar in inches ;
1) = breadth of bar in inches;
li height of bar in inches.
^
TABLE III. MODULUS OF RUPTURE OF CAST IRON.
Modulus of Rup-
ture in Pounds
Experimenters.
per Square Inch.
f = llL.
Authorities.
Robert Stephenson, 1847. . .
( Max.
/ Min .
58,000
37,000
Pole, Iron for Construction, p.
88.
( Max.
47,500
Hodgkinson and Fairbairn.
j Min
29,500
Box, Strength of Materials, p.
186.
( Mean
37,000
*
( Max.
42,500
Woolwich 1858
j Miu .
9,700
Report, p. 2.
( Mean
20,700
Fairbairn 1853 ....
. Max.
50 000
B. A. Report 1853, p. 87.
Turner 1885
. . . Max.
03,500
Inst. Journ., 1886, p. 1.
" It will be noticed that the modulus of rupture varies from the excep-
tionally low value of 9700 Ibs. to 63,500 Ibs. The average for common iron
is about 30,000, while 45,000 is required for better-class castings. Foi
specially good work some South Staffordshire founders can produce u
strength of 60,000 with tolerable regularity. In performing transverse tests.
care should be taken to avoid even the slightest twist on the specimen, and
the weights used should be added very gradually, otherwise low and irregular
results are obtained. The size of bar used has also an influence on the
strength (as shown in Figs. 57 and 58), the smaller sectional areas giving
much higher values. It should be remembered that the strength of a test-
bar does not accurately represent the strength to be expected in the casting,
if the size of the latter, and the circumstances of pouring, do not pretty
closely agree with those of the test-bar itself."
110
THE MATERIALS OF CONSTRUCTION.
79. Tensile Strength. " In many of the less important foundries tensile
tests are omitted, but in the better works such tests are generally performed,
and appear to be growing in favor. It was shown by the American Ordnance
Experiments (1856) that the tenacity of cast iron usually serves as a guide
to its mechanical value, and practical experience quite confirms this view.
Tensile test-pieces are of various forms: they are sometimes used with the
skin on, at others the surface is carefully turned ; sometimes small pieces are
cast separately, while other founders cast the pieces on to the object which
is being made. At Rosebank Foundry, Edinburgh, the practice is to cast a
test-piece-on to the top and bottom of each important article; these pieces
are afterwards broken off, and carefully turned down to a suitable size before
breaking. Such a method is calculated to give a result very nearly approach-
ing what may be expected in the casting itself; for not only is the test-piece
of the same composition as the casting, but it is also cast under as nearly as
possible the same conditions as to temperature, pressure of metal, and rate
of cooling, all of which have a considerable effect on the strength of the
product.
" The follbwing table condensed from a paper by the author will serve
to illustrate the results obtained by different observers : *
TABLE IV. TENSILE STRENGTH OF CAST IRON.
Experimenters.
Pounds per Square Inch.
Authorities
Max.
Min.
Mean.
Minard and Desnormes, 1815 ....
Hodgkinson and Fairbairn, 1837.
" , 1849.
Woolwich 1858
20,300
22,000
23,500
34,300
35,200
40,800
12,200
13,400
9,200
10,600
16,000
16,800
15,300
23,300
Tredgold, 4th Ed., p. 230.
B. A. Report, 1837, p. 339.
Pole, Iron Construction, p. 79.
Report, 1858, p. 2.
J. Chem. Soc. , 1885, p. 580.
Turner, 1885
Rosebank 1886
" It will be seen that the highest tensile strength of British iron above
recorded (40,800 Ibs.) was obtained in the experiments at Rosebank Foundry
in 1886. The average tensile strength obtained by earlier experimenters
was about 16,000 Ibs., while in 1858 the mean was raised to 23,000 Ibs.
This increase represents a real improvement in the metal tested, and was due
to a selection of the more suitable irons as a result of increased knowledge.
Foundry practice has since improved, and some engineers now stipulate that.
a bar one inch in section shall be capable of bearing a weight of 22,400 Ibs.
for twenty-four hours without fracture, and this apparently severe test has.
been complied with. Contracts are now satisfactorily executed in which a
minimum strength of 27,000 Ibs. per square inch is required, and to produce
this nothing but Cleveland iron is employed. The author has also succeeded
in regularly producing an iron of excellent working qualities, with a tensile
*/. S. C. I., vol. v. p. 289.
CAST IRON.
Ill
strength of 29,000 Ibs. per square inch, from a mixture costing under two
pounds (ten dollars) per ton, and* consisting of cast-iron scrap and siliceous,
iron. This is a striking instance of the value of combined chemical arid
mechanical knowledge to the iron-founder.
" In foreign cast iron some tensile strengths .have been recorded which
have not yet been equalled in Britain, though probably these results are to
be regarded as quite exceptional. Thus Professor Ledebur records a tensile
strength of 42,700 Ibs. per square inch with German iron,* while the Ameri-
can Commission on Metal for Cannon, in 1856, obtained a maximum of
46,000 Ibs., and at the Wassiac furnaces, New York, 47,500 Ibs. have been
obtained. f Much difference of opinion has been expressed as to the value
of tensile tests for cast iron, as the metal is now never used in tension.
Professor Ledebur, who is probably the best authority on this subject in
Germany, states that tensile tests should always be made ; and the author's
experience leads to the conclusion that where a complete system of tests,
such as that of "VV. J. Keep, cannot be adopted, no other test affords so good
an indication of the value of the metal, as cast iron with high tensile strength
is almost invariably soft, sound, and fluid. In the following table seven
analyses by the author of samples of cast iron of unusually high tensile
strength are given, together with the results obtained at Woolwich in 1856,
and at Wassiac. Full details of the preparation of these samples are given
in the original paper." J
, TABLE V. COMPOSITION OF CAST IRON HAVING A HIGH TENSILE
STRENGTH.
0, ^
$
d
ifc
B
Rosebank Irons, 1886.
Dumbarton
Irons.
|
o>
l & *
ll
1
1
Tensile Strength
Pounds per sq. in.
>....
35,000
40,700
38,200
37,200
36,700
37,000
34,000
41,200
37,500
Per Cent
Per Cent
Per Cent
Per Cent
Per Cent
Per Cent
Per Cent
Per Cent Per Cent
Per Cent
Graphitic Carbon
2 59
1 62
2.90
2.60
2 31
Combined Carbon
0.56
0.36
0.58
0.52
0.40
0.32
0.30
0.78
0.475
Silicon
1 42
1 96
1 29
1 50
1 13
1 33
1.34
1 63
1 31
1 434
Phosphorus
39
28
56
047
0.41
70
1.09
1.10
29
587
Sulphur
06
03
06
07
06
05
14
12
08
074
Manganese . . .
0.58
0.60
1.00
1.00
1.33
0.65
1.38
1.29
1.51
1.037
" The average composition shown in the above table may be regarded as
typical for good cast iron when the maximum strength is desired, together
with soundness and good working qualities. By increasing the- silicon the
* Inst. Journ., 1891. vol. n. p. 252.
f Inst. C. E., vol. LXXIV. p. 373.
\J. 8. C. I., vol. vii. p. 200
112 THE MATERIALS OF CONSTRUCTION.
metal becomes more soft and fluid, while by diminishing the silicon the
transverse and crushing strength, together with the tendency to chill, are
increased."
MALLEABLE CAST IRON.
80. The Product Denned. White cast iron, or that which has its carbon
all in the combined form, can be made " malleable," or somewhat ductile,
and nearly doubled in strength, by a process of annealing, by which the carbon
separates from the iron without forming a mesh or matrix in which the
remaining iron crystals are imbedded and surrounded, as is the case in
ordinary gray cast iron. It has hitherto commonly been assumed that this
change in the iron was effected by means of a decarbonizing agent (oxide of
manganese, or iron oxide scale), which caused the carbon to leave the iron
and enter into combination with the surrounding materials, thus making the
iron malleable. But Mr. II. B. Stanford has shown,* confirming the results
of Led e bur
1. That only about 10 or 20 per cent of the total carbon is lost in the
process; and,
2. That the same results are effected when the castings are packed in'
clean river sand at least so far as the interior portion of the cross-section is
concerned. It further appears from his investigations that the carbon of
this interior portion is simply changed from the combined to the graphitic
form (Ledebnr's " temper-carbon ") at a bright cherry-red, the temperature
not being high enough to allow the excluded carbon to unite into a more or
less continuous mesh, but that it is kept in very minute, separated aggre-
gates, thus leaving the decarbonized iron crystals in immediate contact, as
they are in ingot metal (steel). The advantage-of this process lies in getting
the final forms run from a perfectly fluid iron, at a comparatively low tem-
perature thus obtaining smooth, full, and solid castings, which can then be
" decarbonized " (chemically speaking), without allowing the excluded car-
.Jwn to form in a graphitic matrix. The alternative is to melt and cast
direct the decarbonized iron (steel), thus making what are known as steel
castings. This requires a very much higher melting temperature, and thus
the metal is less fluid, and it is apt to contain gases, or to generate them in
the mould from the excessively high temperature at which it must be poured.
The effect is that steel castings are apt to be rough and unsightly on the
exterior and more or less porous on the interior.
When cast iron cools slowly from the melted state the carbon, which is
wholly in solution or in chemical combination, passes largely into the
graphitic form, (as shown in Fig. 72), this graphite forming a complete
matrix, and producing the dark, leaden appearance of gray cast iron. In all
large or thick castings the cooling is of necessity slow (except when pur-
posely chilled at the surface), and hence such forms are not radically changed
* Trans. Am. Soc. C. E., vol. XXXTV (1895), "Notes on Manufacture and Prop-
erties of Malleable Cast Iron," by H. B. Stanford, Assoc. M. Am. Soc C. E.
CAST IRON.
in their molecular composition by the annealing which constitutes the essen-
tial feature of the malleable process. Only small castings, therefore, are
suited to this process, unless a very hard white cast iron is used, which does
not change in cooling to gray iron when cast in thicker or heavier masses.
It is essential to the successful working of the process that the original cast-
ings shall have the carbon wholly, or nearly so, in the combined form, which
then changes to an ununited graphitic form when kept some five days at a
bright cherry-red heat. The only essential characteristics of the packing
material are, according to Mr. Stanford, such as prevent it from fusing or
adhering to the castings, or from caking together in hard lumps, and that
it should not be too expensive. The packing serves only to exclude the air
and to hold the cast forms to their normal shapes, when heated and softened,
except as to the decarbonizing action of iron-scale packing on the superficial
portion of the forms so treated, as described below.
81. Method of Manufacture. 7 The original castings, which have hitherto
been made only of charcoal pig iron, may as well be made from coke pig
iron,* provided the sulphur be kept low, the essential requirement being
that the castings shall show all white iron (all carbon in the chemically
combined form). These are then packed carefully in cast-iron annealing-
pots, about 18 inches by 24 inches in cross-section, and four feet high, made
in three sections for convenience of packing, the sections fitting together
with bell and spigot ends. The castings are so placed in the pots that such
settlement as occurs in the oven will not deform them. They are surrounded
by the packing material, which is usually a decarbonizing agent for their
outside portions, but which serves mainly only as a suitable bedding or pack-
ing for large cross-sections. When large forms are to be treated, they are
placed and covered in the oven, without the use of annealing-pots. The iron
pots waste away rapidly by oxidation, being able to serve only about five
heats, of five days each.
The annealing-oven may be any suitable oven in which the temperature
may be kept nearly constant, and uniform over -its entire bed. This requires
that a portion of the combustion shall be completed in the oven itself.
About five days are required to fully effect the change in the condition of
the carbon, after which the furnace is allowed to cool down, the first 24
hours with closed doors. The cost of the treatment is from one-fourth to
one-half cent a pound
The effect of sulphur in the cast iron is to greatly delay the change in
the carbon state, thus largely preventing it, for the ordinary periods of treat-
ment or requiring a greatly extended period of annealing to fully accomplish
it. Thus iron .containing 0.04 per cent sulphur will anneal in three and one-
half 'days, while iron in the same sizes containing 0.20 per cent sulphur re-
quires about nine days. Hence if coke iron is used the sulphur ingredient
must be looked to
* On the authority of Mr. Stanford. See paper quoted above.
114 THE MATERIALS OF CONSTRUCTION.
Clean, heavy-forge iron-scale seems to be the best material to use for
packing the castings in the annealing-pots, and these, being composed of
iron oxide, having a strong affinity for carbon, do extract a large part of the
carbon from the exterior y 1 ^ inch of the surface of all castings so treated,
leaving a bright-colored envelope (on the fracture), containing very little
graphitic carbon. This skin is very much stronger than the interior, as
shown by Fig. 59, from which it appears that the removal of this portion
reduces the strength per square inch by 25 per cent. This argues that the
outer skin is more than twice as strong as the interior.* It is necessary to
establish this fact by further experiment before accepting it as a general
truth.
82. Mechanical Properties of Malleable Iron. From what has been said
it is evident that " malleable iron " is an extremely various product, depend-
ing on the materials used in the cast, and also on the treatment. The
following table is taken from Mr. Stanford's paper, which shows what may
be accomplished by this process. The test specimens were j-f inch in.
diameter, and were tested without dressing down either on the gripped .endsf
or on the reduced portion. The small reduction in total carbon (which all
occurred in the outer portions) and the change from combined to graphitic
carbon are here shown conclusively. The average tensile strength of 49,800
Ibs. per square inch is probably about twice that of the original castings,
while the average elongation of 6.6 per cent indicates a very considerable
ductility. This elongation of T V inch to the inch, together with an assumed
corresponding compression, makes it apparent that small sections would
submit to a considerable amount of bending distortion and other kinds oi
abuse before breaking. In other words, the iron is now twice as strong to
resist a static load, and probably many hundred times as strong to resist the
force of sJiocks or blows, as was the original cast iron.
The elastic limit in compression is very low, but the compressive defor-
mation may be very great.
The relative strength and ductility acquired under varying periods of
annealing, from 3 to 9 days, is shown in Fig. 59, where the averages of all the
results given by Mr. Stanford are plotted, after being reduced to a common
standard of reference. The few tests on turned-down specimens are also
plotted, but these are too few to give either very accordant or very trust-
worthy results.
In Fig. 60 are shown the results of tension tests on malleable cast iron of
^ inch and T \ inch in thickness, and also of ^-inch plates which had been
welded together. These last show a greater strength than the unwelded
bars.
* Due allowance being made for the relative portions of each.
f These parts should have been dressed iu order to prevent any want of symmetry
in the applied forces, or forms like that shown for cast iron in Chapter XV could have
been used. Some of the discrepancies in these results are due doubtless to the rough,
surfaces in the grips.
CAST IRON.
115
TABLE VI. MALLEABLE CAST IRON".
CHEMICAL COMPOSITION (UNANNEALED AND ANNEALED) AND PHYSICAL PROPERTIES.
' TESTS OF ANNEALED SPECIMENS.
fl
o
o
6
Unannealed or
Annealed.
1
|I
~5
Mn.
Si.
P.
S.
Loss of
Carbon.
||5
' 2
=*
~ oo u
c
5
W
'2
1
llf
!1
!!
"Pi
ll
&
H
O
o
O
5 CG Cu
l
K'Z
go<
K =
391
j Before annealing
I After
3.02
2.95
2.96
0.15
0.06
2.80
OA8
0.19
0.69
0.69
0.138
0.140
0.066
0.064
[-0.07
55,100
5.5
5.2
3
1C8
395
j Before
\ After
3.09
2.98
2.99
0.53
0.10
2.45
0.190.60
0.190.62
0.133
0.130
0.060
0.061
[0.11
43,800
6.1
6.3
3
108
400
J Before
"1 After
3.09
2.56
2.96
0.31
0.13
2.25
0.180.75
0.19'0.74
0.142
0.142
0.065
0.061
[-0.53
44,900
7.7
10.3
3
108
AnR i Before
406 1 After
3.07
9.91
2.77
0.08
0.30
2.83
0.190.65
0.200.64
0.163
0.164
0.064
0.064
j^O.16
47,000
7.2
6.2
3
108
,00 ! J Before
*"* | 1 After
3.26
2.95
2.65
0.36
0.61
2.59
0.17'0.69
0.180.61
0.156
0.151
0.071
0.068
j-0.31
45,300
9.1
8.2
3
108
.00 j Before
433 1 After
2.85
2.77
2.72
0.72
0.13
2.05
0.180.74
0.180.72
0.161
0.162
0.073
0.071
|o.08
64,500
1.3
2.8
3
108
.,00 j Before
436 1 After
2.88
2.58
2.73
0.51
0.15
2.07
0.18:0.90
0.18'0.87
0.196
0.192
0.069
0.069
[-0.30
38,900
4.3
6.2
3
108
44t - j Before
' 1 After
2.97
2.66
2.75
0.31
0.22
2.35
0.180.T7
0.19,0.76
0.148
0.145
0.073
0.075
[-0.31
69,100
2.6
4.0
2
108
451
I Before
/ After
3. OS
2.15
2.85
0.08
0.23
2.07
0.18,0.96
0.190.96
0.123
0.129
0.033
0.039
[0.93
45,400
7 2
8.0
2
108
458
j Before
1 After
3. OS
2.03
2.82
0.28
0.26
1.75
0.180.74
0.180.71
0.151
0.150
0.036
0.037
tl 05
56,700
8.4
8 2
2
108
459
j Before
1 After
3.03
2.06
2.81
0.27
0.22
1.79
0.19:0.70
0.200.70
0.195
0.192
0.038
0.037
j-0.97
44,200
2.1
4.2
3
108
j Before
3.09
3.00
0.09
0.200.77
0.127
0.022
' n oq
n f
7
"Iftft
469
I After
2.86
0.32
2.54
0.19
0.75
0.129
0.023
J-0.23
51,olHJ
i . i
< .u
lUo
474
J Before
| After
3.08
2.67
2.92
0.09
0.16
2.58
0.230.70
0.220.70
0.158
0.156
0.023
0.023
[-0.41
46,600
9.8
8.5
3
108
495
j Before
1 After
2.84
2.58
2.62
0.06
0.22
2 52
0.310.63
0.320.66
0.182
0.178
o.oi4j I 26
0.018 f 026
48,100
8.6
8.7
3
108
510
j Before
1 After
3.26
3.12
3.23
0.53
0.030.400.64
2.590.400.67
0.136 0.040
0.1351 039
[-0.14
46,000
5.9
5.3
3
108
Av.
i Before annealing
) After
3.04 2.85
2 66 0.31
0.190.21
2.350.21
0.73
0.72
0.154
0.153
0.050
0.050
|o.38~
49,810
6.23
6.61
42
108
NOTE. The above test -bars were all cylindrical in section and { inch in diameter
3 4 6 6 7 S
FIG. 59. Tensile Strength and Per Cent of Elongation of Cylindrical Test-specimens of
Malleable Cast Iron \\ in. in Diameter. Figures show Number of Tests averaged for
each Point Plotted. (Stanford, Tr. Am. Soc. C. E., vol. xxxiv, 1895.)
116
THE MATERIALS OF CONSTRUCTION.
In Plate IV the plain bar represents the original form of malleable iron
from which all the other forms on the plate were worked. In one case it is
FIG. 60. Tension Tests of Malleable Cast Iron. Each Curve the Mean of Two or Three
Tests. (Berlin Testing Laboratory, 1886.)
folded over and the ends are welded together, while in another it has been
forged like wrought iron. All the other forms were bent cold.
PLATE IV.
EXAMPLES OP COLD-BENDING, FORGING, AND WELDING OF MALLEABLE CAST-IKON
SPECIMENS, ALL BEING ORIGINALLY LIKE THE UNDEFORMED ONE IN THE
CENTRE OF THE PLATE.
(Berlin Testing Laboratory Communications, vol. iv, PI. III.)
CHAPTER VIII.
WROUGHT IRON.
83. Definition. Wrought iron may be defined as nearly pure iron inter-
mingled with more or less slag. As will appear from a study of the methods
of production to be described, the iron is formed in a bath of melted slag
(somewhat as butter is formed in churning), and when it aggregates into a
pasty mass and is removed from the furnace, to be squeezed and rolled, some
of this slag remains intimately associated with the iron. This gives to
wrought iron a fibrous appearance (see Fig. 322) not found in any other
metal.* In the most carefully made iron this fibrous appearance is uniform
throughout the entire cross-section. It is not uncommon, however, especially
in the cheaper grades of wrought iron, to find it largely and coarsely crystal-
line. Wrought iron melts only at a very high temperature, and assumes a
perfectly plastic state through a considerable range of temperature below
this melting heat, in which condition it is easily and perfectly welded. It
is more or less ductile when cold, and will not harden when heated and
quenched in water.
The oldest methods of production of wrought iron were alj direct pro-
cesses, obtaining the malleable product at one operation, or directly from
the ore. While some modern processes also proceed on this plan, practically
all the wrought iron of to-day is made from pig iron and various kinds of
scrap by a puddling process.
METHODS OF MANUFACTURE.
84. The Puddling Process Briefly Stated, f " The ordinary puddling
furnace, is a single-bedded reverberatory of simple construction, formed
externally of cast-iron plates, tied together with wrought-iron rods, and
provided with suitable openings in front for the fire-hole and the working-
door, and lined internally with refractory fire-brick. The crown of the
furnace is also of fire-brick, and is open to the air. The bottom of the
furnace is composed of- three cast-iron plates, which rest upon an iron frame.
The grate of the furnace has wrought-iron fire-bars, and is large in propor-
* To develop this fibrous appearance nick a bar of wrought iron on one side and
bend it double, and if possible split it down like a stick of timber. See Fig. 322.
t The quoted paragraphs on wrought iron have mostly been taken from Turner's
Metallurgy of Iron, Lippincott & Co. , 1895.
117
118
THE MATERIALS OF CONS2EUCTION.
tion to the bed or crucible part on account of the very high temperature
required, particularly towards the end of the process. Each puddling
furnace is provided with a separate flue, which is either connected to a
simple rectangular stack, provided with an iron damper, or which passes into
a boiler-flue so as to economize the waste heat of the furnace. A sectional
view of such a furnace is shown in Fig. 61.
FIG. 61. Plan and Section of a Simple Reverberutory Furnace.
" Two men are employed at each furnace, and are called the ' puddler*
and the ' under-hand ' respectively. The work is very laborious, while it
entails no little skill if good results are to be obtained. Usually six heats
are worked in a turn of twelve hours, but exceptionally seven heats are
obtained.
" The furnace is first charged with a sufficiency of fluxing cinder or
* hammer slag ' (oxide of iron) which has been squeezed out under the
hammer from previous balls, and there is then introduced about 500 Ibs. of
good gray forge-iron. The door is closed and the charge is then heated to
melt the iron, and the most favorable results are obtained when the iron
and the cinder, charged as above described, become pasty, and melt down
together. When the iron has thoroughly melted down and has become
fluid, it is carefully watched until it has * cleared,' and until a number of
small blue jets of flame issue from the surface of the liquid. The damper is
now ' put down,' or closed, so as to fill the furnace witfo a reducing (non-
oxidizing) atmosphere, and to lower the temperature somewhat. In a short
time the jets of blue flame almost cease, and the mixture of iron and cinder
rises in the furnace to a height of some 8 or 10 inches, and during this stage
constant stirring or ' rabbling ' is necessary to prevent the iron settling on
WROUGHT IRON. 119
the bottom of the furnace, and to assist the decarburization by bringing the
(carburized) iron and cinder (iron oxide) into uniform and intimate contact.
The whole mass should now be in motion, and bubbles of gas should rise and
burn with a blue flame, tinged more or less with yellow, at the surface.
When the ' boil ' is thus in full progress, or ' well on,' the damper may be
raised somewhat, and the iron will soon be observed to ' come to nature ' or
to separate from the cinder. The first sign of this is the appearance of
small bright spots on the surface of the cinder, which alternately appear and
disappear. The cinder now gradually sinks, and leaves the iron as an irreg-
ular mass, not unlike the small globules or grains of butter produced by the
churn; and as in good butter-making so in good puddling, the grains should
be small and uniform throughout the mass. The temperature should now
be raised to the highest point so that the iron may be at a welding heat; the
puddler after first lifting the metal and turning it over, by inserting a bar
underneath in order to prevent the bottom becoming colder than the top,
and breaking it up, proceeds to collect it into balls, which are taken to the
hammer."
-9^5. Oxidation in Puddling. " The following remarks on the oxidation
of cast iron under different conditions, will explain the differences between
the old and newer processes of puddling.
" It is usual to speak of atmospheric air as oxidizing and removing the
impurities present in cast iron, but if a globule of cast iron be melted in the
air, and then exposed to a blast of air or oxygen, it will be observed that the
impurities are not the only substances that are oxidized. It is true that
under very special conditions either the carbon or the silicon may be
separately oxidized. But on performing the experiment above indicated it
will be found that the iron itself is oxidized in about the same, relative pro-
portion as the other elements, and the result is that practically a layer of
impure magnetic oxide of iron is formed outside the globule, while the por-
tion of metal that is left is of nearly the same composition as the original
iron. If the cinder be allowed to run away as rapidly as it is formed, ulti-
mately the whole of the iron would be converted into magnetic oxide, and
the last particles of cast iron so removed would have nearly the same com-
position as the original metal. In this case oxidation has taken place, but
no purification has resulted.
" If, now, the same experiment be tried, but the fluid oxide be allowed
to remain, and to cover the fused metal, the oxidation of the iron will pro-
ceed very little further; a reducing action will then be commenced whereby
the silicon, carbon, and other easily oxidizable elements will be removed,
but at the same time a corresponding weight of iron will be returned to the
globule from the surrounding slag.
" If, thirdly, a globule of cast iron be covered with magnetic oxide of
iron to protect it from the air and to supply the necessary cinder, and it be
then strongly heated, it will be found that the globule has not lost in weight,
but has become distinctly heavier during the process. It is scarcely neces-
120 TEE MATERIALS OF CONSTRUCTION.
sary to say that the waste which takes place daring reheating or remelting,
corresponds to the first condition above given. The oxide runs away as it is
formed, and this is an example of waste of iron pure and simple. The only
redeeming feature is that sometimes the oxide produced may be of value for
other purposes. The early open-hearth processes for producing wrought
iron in fineries, and the original method of puddling, resemble the second
case, for part of the iron is wasted to produce the cinder needed to remove
the impurities from the remainder of the metal. The larger the proportion
of these impurities the greater will be the loss of iron necessary to make the
required cinder, and for this reason a comparatively pure iron is needed, in
order to obtain the least waste, while at best the waste is comparatively
great. A deficiency of fluid cinder in the early stages of ordinary puddling,
or pig ' boiling,' has an exactly similar effect, and leads to waste for the
same reasons,
" In the modern method of working, on the other hand, the object is to
imitate the conditions of the third case previously supposed. Oxide of iron
can be bought much more cheaply than it can be made from pig iron, and
besides the oxidation of pig iron requires the expenditure of time and fuel.
Oxide of iron is, therefore, supplied in its cheapest and most readily available
form, and as much of this oxide as possible is reduced and converted into
wrought iron. To do this it is necessary that the iron and fluid oxide
should be brought into actual and frequent contact, and so perfect fluidity
and constant rabbling are needed. There is, of course, a practical limit to
the amount of carbon which can be present, due to the fact that cast iron
cannot take up more than a certain amount, say 4 per cent, of this element.
There is also a practical limit in the case of both silicon and phosphorus;
the first being regulated by the increased consumption of time and fettling
with excess of silicon, and the second being determined by the inferior
quality of iron produced, with large proportions of phosphorus. But within
these practicable limits it is advantageous to reduce as much of the oxides
of iron supplied as possible."
86. Details of the Puddling Process. " The working heat of puddled
iron may be conveniently divided into four stages, which will be separately
described, namely:
" (1) Melting-down stage, lasting about half an hour, by the end of
which most of the silicon and manganese and a considerable proportion of
phosphorus have been removed.
" (2) Quiet fusion or ' clearing ' stage, lasting about ten minutes,
during which the rest of the silicon and manganese and a further quantity
of phosphorus are removed.
" (3) The boil, which lasts nearly half an hour, during which the greater
part of the carbon is eliminated, together with a further quantity of phos-
phorus.
" (4) Balling-up stage, which occupies some twenty minutes, and by
WROUGHT IRON. 121
which time the purification, except as regards the removal of slag, has prac-
tically ceased.
" 1. The furnace having been suitably prepared, and hot from a previous
heat, the pig iron is charged as before described ; the door is then closed, and
the working opening in the bottom of the door covered with an iron plate
and rendered as far as possible air-tight by means of a little fine cinder
thrown with the shovel. The fire is also made up, and heating proceeds for
some twenty minutes, by which time the top of the pig iron is red-hot, and
the flux begins to soften. The pigs are now turned so as to heat them more
uniformly and the door is again closed ; in a few minutes the iron begins to
melt, and if carefully watched may be seen to trickle down into the cinder
in drops. The workman now introduces an iron rod, stirs up the mass, and
brings up any pieces of iron which have not completely melted, and which
might otherwise remain covered and take longer to melt. When the whole
is thoroughly fluid and well mixed the melting-down stage is finished.
" 2. One of the workmen, generally the under-hand, now introduces a
bar which is bent at the end at right angles, and so acts as a scraper or
stirrer, and the whole charge is well stirred and exposed to the action of the
fettling and cinder, and also to some extent to the oxidizing influence of the
air. The temperature is maintained as high as possible during this stage.
" The iron is thus thoroughly ' cleared ' or purified from silicon, the
point at which clearing is completed being judged by the appearance of the
charge, and upon the skill of the workman at this stage much of the subse-
quent success depends.
"3. When the metal has cleared, and is in a state of tranquil fusion, the
next point is to bring on the ' boil.' The puddler, therefore, diminishes the
draught, or ' puts his damper down,' so as to fill the furnace with a smoky
flame and lower the temperature. In some cases also the door is opened and
water thrown in at this stage, as this promotes rapid cooling and supplies
oxygen at the same time. The (carbonized) metal being thus somewhat
thickened, and being vigorously stirred "during the whole time, becomes
intimately mixed with the (iron oxide) cinder; the carbon is thus oxidized,
producing carbon monoxide, which burns in blue flames as the bubbles of gas
rise and^urst. These flames are sometimes called ' sulphur ' or ' puddler's
candles,' on account of their pale blue color. The charge thus swells up and
rises some six inches in the furnace (like boiling molasses), and as the heat
increases and the damper is opened somewhat, a quantity of red-hot slag
flows over the fire-plate (at the door) into a cast-iron slag-wagon placed
ready to receive it. The violence of the action now gradually diminishes,
the iron ' comes to nature, ' and the charge settles in the furnace ; the less
fusible wrought iron is in the form of a porous cake, and the residue of slag
collects chiefly underneath.
" 4. In the fourth, and last, stage the puddler has to manipulate the
iron into convenient forms for subsequent treatment. For this purpose the
122 THE MATERIALS OF CONSTRUCTION.
cake of metal is broken up by inserting a bar underneath, and is worked at
a welding heat into one uniform mass or ball. This is now divided into
about six balls, of approximately equal size, each of which weighs about 80
Ibs., and these are in turn withdrawn from the furnace and taken to the.
hammer, where the slag is to a great extent expelled, and a bloom of iron is
obtained. This is rolled, without reheating, into 'puddled bar,' which is
the name given to the crude wrought iron produced as above described."
87. Production of Muck Bars. " The balls of crude wrought iron, hav-
ing been produced in the puddling-f urnace as before described, have now to
be compressed to expel the slag and render the material more uniform in
character; they are afterwards rolled into bars, which receive the name of
* muck bars.' For compressing the iron various forms of hammers or
squeezers are used, while for the production of bars, grooved rolls, as intro-
duced by Cort in 1783, are generally employed, though in a few exceptional
cases, where water-power is available, bars are still produced by the hammer
or ' battery,' as in ancient times.
" Various forms of squeezers have been introduced from time to time,
chiefly with the object of preventing the jar or shock due to the action of the
hammer, though such appliances have not met with very general application.
The more usual forms may be conveniently divided into two classes :
" (1) Those in which compression is produced by means of a lever, as in
the ' alligator ' or ' crocodile ' squeezers, which are so called by the workmen
from the resemblance between the motion of this class of squeezer and that
of the mouths of the animals above mentioned.
" (2) Those in which a revolving cam is employed.
" Though squeezers appear at first sight to have many advantages over
hammers, particularly on account of their even and quiet action, they do not
seem to have grown in general favor in recent years, it being stated that the
iron worked in squeezers is less uniform in character, and that the slag is
not so completely expelled by squeezers as with hammers.
" Steam hammers are used for shingling puddled balls in almost all
modern works, and are now always double-acting. The hammer-block in
this instance weighs about ten tons, and is heavier than is generally employed
in forges, though lighter than is usual for manipulating large masse^f steel.
Forge-hammers seldom exceed three tons in weight, while steam-hammers
for forgings of the largest size weigh 50 tons and upwards.
" The iron, having been thus compressed and consolidated by some form
of hammer or squeezer, and a considerable portion of the slag expelled, is
now taken while still hot to the puddle-rolls, where it is converted into bars.
The bars are allowed to cool, and are afterwards cut np with shears into
suitable lengths; these are then made into bundles, or ' piles,' of the required
weight and size. When a specially smooth surface is required, as in the
production of sheet iron, it is usual to make the top and bottom of each pile
of ' scrap bars ' ; these are made by reheating the crop ends of finished bars
or other good wrought-iron scrap, and are therefore more uniform in char-
WROUGHT IRON. 123
acter, and possess a smoother and cleaner surface than ordinary puddled
iron. ' '
88. Reheating the Muck Bars. " The puddled iron having been pre-
pared as before described, is now taken from the forge to the other part of
the works, which is known as the ' mill.' This is usually covered with a
tolerably lofty roof, but is open at the sides; it contains reverberatory furnaces
for heating the piles of puddled iron, and also rolls of various sizes, with the
necessary engine and connections required for producing the various ' sec-
tions ' of finished iron. A steam-hammer is also provided if forgings are
produced, but otherwise this is not required.
" The temperature employed in mill-furnaces is a white heat, and suffi-
ciently high to cause the metal to weld together when it is passed through
the rolls, to which it is taken from the furnace."
89. Rolls. " The rolls used in iron- works are classified according to
their shape and the method adopted in their production. They are gener-
ally made from a strong close-grained cast iron, usually that obtained from a
blast-furnace, in which cold blast is employed. Occasionally steel rolls are
used, and these appear to be somewhat growing in favor in recent years.
" Rolls may be classified according to their shape into
" (1) Flat or plain rolls, which are used for rolling sheets or plates.
" (2) Grooved rolls, which are required for the production of bars, rods,
angle and channel iron.
" According to their method of production rolls are classified as
" (1) Grain rolls, which are produced in moulds of green or dry sand,
and in which the surface of the roll after turning down shows the ordinary
grain of the cast iron from which it is made. These are used for all rough-
ing purposes and for sections, and in other cases if the metal is finished hot.
" (2) Chilled rolls, which are produced in cast-iron moulds or chills.
TKey, therefore, have a hard white surface of chilled iron, which varies in
thickness from about i to f of an inch, according to the size of the casting
and the class of work for which it is intended. Rolls of this kind are more
costly, and are employed for the production of sheets, plates, or strip, or
in other cases where specially fine surfaces are required."
Small rolls tend more^to elongate than to spread the materials, while
large rolls tend both to elongate and to spread the metal.
90. Effect of Repeated Reheating of Iron. "As it is well recognized
that puddled iron is much improved in quality by being cut up, piled,
reheated, and rolled or hammered, and that the iron is further improved
by repeating the operation, it might be assumed that by continuing this pro-
cess the properties of the metal might be again and again further improved.
In practice, however, this is not found to be the case, and it is only in
special cases that it is advantageous to reheat puddled iron more than twice.
It has been shown by experiments, in which puddled bar was reheated and
rolled as many as twelve times, that after about six workings the metal began
to seriously deteriorate, and even in the earlier workings, after the third no
124 THE MATERIALS OF CONSTRUCTION.
corresponding advantage was obtained for the fuel and labor expended and
the waste incurred. The results obtained were as follows ( Useful Metals,
p. 318):
Tensile Strength in
Pounds per Square Inch.
Original puddled bar 43,900
2d working 52,860
3d " 59,580
4th " 59,580
5th " 57,340
6th " 61,820
7th " 59,580
8th 57,340
9th " 57,340
10th " 54,100
llth 51,970
12th " . . . . 43,900
"If it be assumed that the result in the fifth heating was accidentally
low, it will be seen that all the other tests follow in a regular succession, the
maximum tensile strength being obtained with the sixth working. Probably
with iron of different composition or character the maximum would be
reached at a different point, but in all cases the gradual original improve-
ment and subsequent deterioration would be observed. When the metal
passes into the bands of the smith it is found that if it has been worked
during its previous preparation so as to bring it to its best condition, it has a
tendency to ' go back ' in forging; while, on the other hand, if the iron has
not been unduly 'worked, it improves when properly smithed. For this
reason also it is not advantageous to often reheat and work iron during the
process of manufacture, and ' best,' ' best best,' or ' treble best ' irons are
obtained, not by frequent heatings, as is sometimes stated, but by the careful
selection of all the materials employed, and by systematic and frequent tests
of the iron during the various stages of manufacture."
91. Sections of Finished Iron. " The shape into which finished iron is
rolled varies according to the purposes f<;^ which it is designed, the chief
divisions being plates, sheets, strips, bars, angle-irons, and rails, the last being
relatively of much less importance than formerly. Among the more usual
shapes or ' sections ' may be mentioned the following : bars, including round,
half-round, square, flat, round-edged fiats, oval, octagon,, together with
levelled and bulb iron, and rods; tee (or T-shaped) iron, tee^with round top
or edges; angle- (or L-shaped) iron, angle-iron with unequal sides or round
back; channel-iron, I iron, Z iron; rails, including single-headed, double-
headed, and flange; and horseshoe-iron, which is rolled sftigle-grooved,
double-grooved, or concave. Numerous other forms are also Required from
time to time for various purposes; so that the number of rolls which have to
WROUGHT IRON. 125
be kept in stock at a large works with a general trade is very great, not
unfrequently amounting to hundreds. As each pair of rolls is generally only
capable of finishing one section of iron, the cost of the supply and main-
tenance of rolls forms a considerable item of the expenditure of an iron-
works."
92. Imperfections in Finished Iron. " The three chief varieties of
imperfection in the appearance of finished iron are rough edges, spilly
places, and blisters.
" (a) Rougli edges, when not due to imperfections in the rolls or to care-
less working, are a sign of red-shortness, and are particularly noticeable in flat
bars or strips. Red -shortness may be due to an excess of carbon, or to the
presence of sulphur, particularly if copper is also present. Usually, how-
ever, if iron has been properly puddled, practically the whole of the sulphur
is eliminated, and the red-short condition is due to the ' dryness ' of the iron.
Iron is said to be dry when.it is deficient in fusible or welding cinder, which
may be readily squeezed out from between the particles when the iron is
worked, and so enable clean surfaces to be brought together to form a good
weld. A thick dry cinder, on the other hand, leads to red-shortness, and a
piece of brick or other foreign matter which crushes up in the rolls to form
a, dry powder acts in the same manner.
" (^) Spilly places are spongy or irregularly spotted parts which are not
unfrequently noticed in sheets, and which are occasionally met with in all
kinds of wrought iron. They are generally due to imperfect puddling,
whereby one part of the iron, when ' coming to nature,' has been oxidized
more than another. If the heat has been thoroughly well worked and the
iron uniformly mixed, spilly places are seldom observed.
" (c) Blisters are not unfrequently met with in sheets, and 'lead to con-
|siderable loss and inconvenience. They are much less common in steel sheets
than in iron, and some experiments conducted in 1803 led the author to
ittribute the formation of blisters to a reaction between carbon and oxide of
jiron in wrought iron of inferior quality. This view is in accordance with the
experiments of A. Friedmann, who collected and analyzed the gas contained
in a number of blisters. This gas was found to contain over 70 per cent of
carbon monoxide, the remainder being chiefly carbon dioxide, with some
nitrogen and hydrogen. Inside th ; blisters a quantity of scaly matter is
found, which" Friedmann states to consist of about two thirds silica and
nearly one third iron aluminate (FeAlOJ, together with small quantities of
other oxides."*
|
MECHANICAL PROPERTIES OF WROUGHT IRON.
L93. Crystalline Fracture. As explained in Art. 84, the fibrous appear-
ce of wrought iron when nicked and bent with a splitting action, similar
o that of a piece of timber treated in like manner, is due to the presence of
* Inst. Journ., 188o, vol IT. p. 645,
126 THE MATERIALS OF CONSTRUCTION.
the foreign matter which formed the slag or bath from which the puddled
ball was taken. Ordinarily when wrought iron is broken in tension in a test-
ing-machine the fracture appears to be wholly fibrous, somewhat like that
of soft steel, but with a darker and more ragged appearance. If a wrought-
iron bar be nicked and broken by bending, it will usually show a fibrous
appearance, whereas steel so treated will always show a crystalline fracture.
Occasionally, however, a part of all of the fracture of a test specimen of
wrought iron, whether broken in tension or by nicking and cross-bending,
will have a coarsely crystalline fracture. It is very common, also, to find
such a fracture when wrought iron breaks in service, as in the case of car and
wagon axles, steam-engine cranks and pins, etc. In such cases as these it
has been common to ascribe the failure to the crystallized condition of the
iron and to assume that the iron had changed to this condition in service.
This is called the theory of the cold crystallization of wrought iron. Those
who believe in it usually ascribe the change to a vibratory action. Whether
or not wrought iron ever does crystallize in service in this or in any other
manner has been a disputed question for the last half-century. It has,
however, remained a theory the truth of which has never been established
by actual experiment, and it is now one which seems to have no scientific
adherents. It has, however, become so thoroughly fixed in the minds of the
less educated users of iron and steel that it is met with on every hand, and;
this action is stated to be a fact with the most positive assurance by nearly
all mechanics and is commonly believed by the public generally. The views
of the author of this work on this subject may be summarized as follows:
I. The normal molecular arrangement of wrought iron is crystalline, but
the thorough admixture of the inert slag in a well-worked product prevents j
these crystals from forming in visible sizes. The ordinary fibrous fracture, j
therefore, exhibits rather a lateral view of these finely crystallized threads,
thus causing this to present a fibrous appearance.
II. When any portion of the puddled ball as removed from the furnact
is not intermixed with foreign matter, as may be the case from overheating j
and melting of some portion of the puddled ball, or from the inclusion in tin
mass of some unreduced melted cast iron, these portions being really of th
nature of ingot metal or steel, rather than of wrought iron, then thesi
masses of iron, free from foreign matter, when somewhat cooled and rolleo
into a bar, will form in that bar a part of the cross-section which will be abl<
to crystallize on a slow cooling in large-sized crystals, so as to be clearly visibl
to the naked eye. When such melted portions are due to overheating o
the puddled ball the iron is said to be burnt, but a too-rapid hurrying of th
boiling process under a low heat will also enable some of the unreduced cas
iron to be removed from the furnace in this way with a similar result.
III. With the ordinary and more inferior grades of wrought iron now o
the American market, it is very common to find large portions of the c
section of test-bars showing a crystalline appearance, even for tension-
specimens of standard form. Much more, therefore, are such irons likely
WROUGHT IRON. 127
have this appearance when nicked and broken across, or when nicked and
pulled in tension.
IV. All wrought iron when broken with extreme suddenness will show a
crystalline fracture. This is because time is not given for the drawing out
of the section, rupture occurring directly across the fibres, so that the frac-
ture shows only the end view of the same.
V. When a bar is nicked with a sharp chisel, or grooved in a lathe with
a sharp-pointed tool, and broken across, rupture begins at one side without
any elongation of the fibres, and extends from fibre to fibre across the section
in such a way as to produce a result similar to that caused by an instan-
taneous rupture cited in IV. In this way wrought iron will often show a
crystalline or granular 'fracture, when under the ordinary tensile test it
would be wholly fibrous. All steel or ingot metal will always show a
crystalline fracture when treated in this manner, although for all the soft
and medium grades of steel the fracture is always fibrous or silky when
broken in tension, with the usual accompanying elongation and contraction.
VI. Much of the so-called wrought iron on the market to-day consists
simply of rolled fagots of " scrap-iron," a large portion of which is really
scrap-steel. As these are heated only to a welding heat, and then rolled into
merchant bar, there is no real mixing of the metals, and the several com-
ponents form so many separate portions of the cross-section of the final rolled
forms. The crystallized steely areas found in the fractures of most wrought
irons of the common grades to-day can be largely traced to this source.
Wrought-iron railway-axles and other large forms are usually made up in
this way.
VII. When wrought iron breaks in service, therefore, and shows a
coarsely crystalline fracture, it does not prove that crystallization has
occurred in service. It proves only that this iron had such a structure
originally. If, however, the rupture occurs in practice in a suddenly con-
tracted area, as in a screw-thread or in a sharp angle, or if it has been pro-
duced with extreme suddenness, as in case of an explosion or shock of any
kind, if the appearance of the fracture is finely crystalline or granular, this
appearance may be wholly due to the method of failure. This is shown by
the fact that if a specimen be cut from the adjoining metal and tested in
tension with the standard form of specimen, it might show a wholly fibrous
fracture. In such cases, therefore, the crystalline appearance of the fracture
is due to the particular conditions as to shape of specimen and suddenness of
rupture and not to any molecular change which has taken place in the iron.*
* From the Report of U. S. Watertown Arsenal Tests of Metals for 1890, in which are
recorded many tests of specimens cut from the journals of old railway-axles, the follow-
ing note is taken :
"Axles have been examined which have had long-continued service, the journals of
which showed incipient cracks, indicating that rupture had begun and that further use
must result in complete rupture. It is a remarkable fact that the tests of the metal of
these journals near these cracks showed no loss of strength or ductility. No indications
128 THE MATERIALS OF CONSTRUCTION.
94. The Welding of Wrought Iron. It is a peculiar property of wrought
iron that it remains in a plastic condition throughout a considerable range of
temperature. If two pieces of wrought iron could be reduced to this plastic
state (a white heat) with perfectly clean surfaces, and pressed together
firmly and allowed to cool, the union would be so perfect as to be practically
as strong as any other portion of the material. The great difficulty in suc-
cessful welding lies in the fact that when the iron is heated in the presence
of oxygen the surface is oxidized, and this oxide of iron, being quite fusible
at this temperature, forms a complete coating of slag over the entire surface.
When two such surfaces are brought together, therefore, each being entirely
covered with melted iron oxide, which is practically a foreign substance, the
union effected is necessarily imperfect. The degree of imperfection depends
on the amount of this melted slag which succeeds in remaining in the joint.
In order to remove this liquid slag as much as possible, the two surfaces
should be convex to each other when they are brought together. That is to
say, they should first come in contact along the central portion of the weld
area, so that the hammering or the pressure by which the weld is effected
will, as perfectly as possible, squeeze out this liquid slag from the joint, thus
allowing the plastic iron surfaces to come into immediate and actual union.
If any portion of this melted oxide remains in the joint, it entirely prevents
a union of the surfaces over such area as it occupies, and to that extent
weakens the joint. As it is impracticable to heat the surfaces to be welded
or even to join them in a vacuum, or away from the oxygen of the air, it is
impossible to avoid entirely the presence of the melted oxide of iron in weld-
ing operations. With intelligence and care, however, in the performance of
the work, nearly all this oxide can be removed in the act of welding, and a
practically perfect union effected. To assist in removing this melted oxide,
borax is commonly used. This being a perfect solvent of the oxide, the
whole is changed to a thin liquid, which is the more perfectly squeezed out
of the joint in the welding process. In this way steel may be welded which
would not unite without it. When the parts to be joined are heated in an
ordinary forge the blast of air causes an excessive oxidation of the surfaces,
and thus gives rise to large quantities of the melted slag. By maintaining,
a thick fire most of the oxygen has been consumed to CO or C0 2 before*
of a tendency to crystallize were discovered, and inasmuch as the metal has gone
through all the phases of deterioration up to the limit of actual rupture without showing
a crystalline tendency, it is thought this demonstrates and proves that this material isi
incapable of cold crystallization when exposed to the conditions of service."
In one instance one of the old cracks which had developed at the inner shoulder of
the journal reached to a depth of 0.02 inch into the side of the test specimen, and yet I
the specimen broke two inches from this'sectiou. After rupture the end of the specimen
(1| in. diam.) containing this crack was bent cold 33 degrees with *' this crack at the j
middle of the bend on the tension side, which opened the crack in width and ah
developed numerous cracks in this vicinity," but without rupture. All the tests show*
fibrous fractures.
WROUGHT IRON. 129
reaching the iron, and hence less oxide is formed. If the parts were heated
in a reverberatory furnace or in a " muffler," raised to a sufficient tempera-
ture and kept out of the way of air-currents, a much less amount of this slag
would be formed, and the welding would be more readily performed. One
of the advantages of electric welding lies in the fact that no air-current is
employed, and by having the parts in contact during the time they are being
heated, the air is largely excluded from the welding surfaces, and hence
little or no oxide is formed there to prevent a perfect union. It is largely
for this reason that electric welding may be more perfect than hand welding.
In view of the inherent difficulties described above, it might well be
anticipated that welded joints are necessarily very unreliable even when done
with more than ordinary care. Many tests of the strength of welded joints
have shown that this strength may be anywhere from 30 to 100 per cent of
the strength of the parts which have been joined, and in the hands of careless
or incompetent workmen the strength of a welded joint may be almost zero.
With the most careful work, however, that is found to be practicable in the
best forging practice, the average strength of hand-welded joints has been
found by Kirkaldy * to be in the case of round iron tie-bars from 1 J to 3-J
inches in diameter, but 60 per cent of the average strength of the bars. In
the case of flat plates from 2J to 6 inches in width and from inch to 1
inch in thickness, the average strength of the welds was 71 per cent of the
trength of the plates. In the case of chain-link welds from to 2-J- inches
n diameter, the average of 216 tests showed an average strength of the welded
oints of 83 per cent of the strength of the iron rods. In the case of a
welded chain, where the strength of the chain is only that of its weakest link,
t would not be safe to rely on a strength of joint greater than 50 per cent
)f the strength of the iron from which the chain has been made.*
In 1885 Professor Bauschinger undertook an elaborate series of experi-
ments to determine the relative welding qualities of soft steel and wrought
ron, also the relative efficiencies of forging by hand and under a steam-
lammer. The results of his experiments have been condensed in the
ollowing tables. These results show a strength of welded soft-steel bars
jqual to 89.2 per cent of the strength of the original material, while the
fficiency of the welds of the wrought-iron bars was 95.6 per cent. The
elative value of hand and power forging is indicated in the second table,
where it is shown that the hand forging gave an efficiency of 84 per cent,
vhile the steam forging gave an efficiency of 97.2 per cent, on the soft-steel
)ars, while on wrought iron these were 87.9 per cent and 91.0 per cent
respectively.
These tests were made under the most favorable conditions, and they
probably represent the highest attainable efficiency in welding on both kinds
>f materials. These results should, therefore, not be taken as representing
Average results in practice, but rather as an ideal which may possibly be
* Kirkaldy's System of Mechanical Testing, London, 1891, report KK.
130
THE MATERIALS OF CONSTRUCTION.
reached with the greatest care. It will be noticed that the fourth soft-steel
specimen gave an efficiency of 99.6 per cent, the break occurring entirely
outside of the weld ; while the sixth set of specimens of soft steel gave an
efficiency of but 57.3 per cent, the break occurring in the weld. Another
specimen of soft steel would not weld at all.
TABLE vir. BAUSCHINGER'S TESTS OF THE STRENGTH OF WELDS
WITH LOW-CARBON STEEL (iNGOT IRON) AND WROUGHT IRON.
(Each line of " welded " results contains the mean of two tests.)
SOFT STEEL OR INGOT IRON
s
Cross-section
Ea^
P3 8
Dimen
of Test-bar.
v-
d
"c3
hM U
,660
65,400
29.8
65
1.26x0.55
0.60x0.54
0.32
Welded
Hand forging
34,840
60,640
92.7
15.4
68
Broke outside of wel(
1.18x1.18
d = 0. 70
0.39
Orig.
Two heats
46,640
69,960
22.8
42
1.18x1.18
0.39
Welded
Hand forging
36,970
40,100
57.3
Broke in weld
d = 1.10
d = 0.70
0.39
Orig.
Two heats
33,840
61.570
11 9
15
d = 1.06
d = 0.61
0.33
Welded
Hand forging
35,550
45,790
74.4
0.9
6
Broke in weld
d = 0.79
d - 0.44
0.15
Orig.
Two heats
44,080
66,830
23.2
67
d = 0.79
d = 0.44
0.15
Welded
Hand forging
39,100
58,230
87.1
8.7
17
Broke in weld
Average = 89.2
WROUGHT IRON.
3.27X0.71
3.27x0.71
2.39x0.71
2.36x0.59
1.70
1.39
Or;?.
Welu. '
Three heats
Steam hammer
32,700
24,170
52,600
50,050
95.1
26.1
13.4
42
23
Broke in weld
2.56x1 00
2.56X1.06
1.64X0.72
1.65x0.69
1.18
1.14
Orig.
Welded
One heat
Steam hammer
23,750
22,750
50,050
50,050
100.0
28.5
20.9
45
34
Broke in weld
1.65X0.47
1.65x0.47
0.84x0.47
0.86x0.43
0.40
37
Orig.
Welded
Two heats
T jand forging
27,730
22,750
51,190
48,910
95.6
11.4
8.1
18
14
Broke in weld
1.84x0.63
1.34x0.63
0.58x0.64
0.68x0.58
0.37
0.40
Orig.
Welded
i "K> heats
i-/.-?d forging
28,440
27,020
56,880
55,450
97.6
24.3
15.3
42
Broke in weld
1.02x1.02
1.02x1.02
d - 1.02
d= 1.02
d = 0.59
d - 0.59
>tf
0.27
Orig.
Welded
Orig.
Welded
Two heats
Hand forging
29,860
28,440
55,030
56310
102.3
31.5
15.6
18.3
9.2
39
17
36
16
Broke near weld
d = 0.59
d = 59
0.27
0.27
Two heats
Hand forging
29,860
27,020
61,000
50,620
83
Broke in weld
Average = 96.6
WROUGHT IRON.
131
TABLE viii. BAUSCHINGER'S TESTS OF THE RELATIVE VALUE OF
HAND AND STEAM FORGING.
(Welding Low-carbon Steel and Wrought Iron.)
SOFT STEEL OR INGOT IRON.
Dimen-
sions of
Original
Cross-
section
in Inches.
Cross-section
of Test- bar.
Condi-
tion of
Bar.
Method of
Welding.
43
a
Cu
i
42.660
47,200
47,490
Tensile Strength.
co . Original
Welded .
o^ 1 Percentage of Elon-
co-?co gation for 10 Inches.
Percentage of Re-
duction of Area.
Remarks.
Dimen-
sions.
Area.
1.26x0.47
1.26x0.47
1.26x0.47
0.81x0.41
0.82X0.36
0.80x0.35
d = 0.43
d = 0.43
d = 0.43
0.33
0.30
0.28
Orig.
Welde'd
Welded
Hand forging
Steam hammer
68,110
68,250
71,100
9.2
to
66
66
62.5
Broke outside of weld
Broke outside of weld
0.75X0.75
0.75X0.75
0.75X0.75
0.14 Orig.
14 Welded
0.14 Welded
Hand forging
Steam hammer
45,640 66,260
44,22061,710
39.810J 65,551)
93.1
98.9
21.7
9.1
13.0
68.
9
36
Broke in weld
Broke in weld
d = 1.02
d = 1^02
d = 0.59
d = 0.58
d = 0.58
0.2?
0.27
0.27
Orig.
Welded
Welded
Hand forging
Steam hammer
42,230
24,880
33,130
62,700
27,300
54,600
43.5
87.1
22.9
0.1
6.9
63
1
18
Broke in weld
Broke in weld
d = 0.9l
d = 0.91
d = 0.91
d = 0.51
d = 0.51
d = 0.51
0.21
0.21
0.21
Orig.
Welded
Welded
Hand forging
Steam hammer
49,200
44,930
44,930
09,530
68,960
68,400
99.2
98.4
23.0
15.8
14.8
62
28.5
20
Broke in weld
Broke outside of weld
Average hand forging = 84.0
steam =97.2
WROUGHT IIION.
d = 1.02
d= 1.02
d= 1.02
d = 0.58
d = 0.58
d = 0.58
0.27
27
0.27
Orig.
Welded
Welded
Hand forging
Steam hammer
24,880
24,880
27,020
55,460
48,770
50,480
87.9
91.0
23.1
9.9
12.7
41
10.5
35
Broke in weld
Broke outside of weld
95. The Effect of Reduction in the Rolls on the Strength of Wrought
Iron.* Other things being equal, the strength of wrought iron will vary
directly with the amount of reduction in the rolls from the size of the pile to
that of the finished specimen. If it is desired to obtain an equal strength
TABLE IX. EFFECT OF VARYING REDUCTION IN THE ROLLS ON THE
STRENGTH OF WROUGHT IRON.
Diameter.
Length.
Elastic Limit.
\b*r
Ultimate Strength.
Per Cent of
Elongation.
Ratio of Elastic
Limit to Ulti-
mate Strength.
1
4.75
38,000
57,533
17.1
66.1
n
6/25
34,600
54,85
21.65
62.9
it
7.50
8.50
32,600
34,800
58, 3f
52,6' >
23.5
26.0
60.8
65.7
H
0.75
33,100
52,400
25.3
63.3
li
8.175
34,325
54,175
22.0
63.4
If
9.50
33,175
52,000
23.15
63.6
n
9.75
31,875
50,325
23 2
63.8
2
10.50
31,800
49,725
22.3
63.9
* See also similar results on steel in Chap. XXV.
132
THE MATERIALS OF CONSTRUCTION.
for different finished sizes, it is necessary to make these several sizes from
the piles whose areas of cross-section bear a constant ratio to those of the
finished sections. The following table gives average results of four series of
tests on wrought iron on sizes from one inch to two inches in diameter.
As showing that a uniform reduction in the rolls may be made to produce
iron of equal strength for these same sizes, the following table of results is
given, the iron having been rolled and the tests of strength made expressly
to establish this fact.*
TABLE X. DIMENSIONS AND AREAS OF PILES, AEEAS OF BARS IN
PERCENTAGE OF AREAS OF PILES, TENSILE STRENGTH, ELASTIC
LIMIT, ETC., OF NINE BARS.
Size of Bar.
Dimensions of
Piles.
Area of Piles.
Area of Bars in
Per Cent of
Area of Piles,
Tensile Strength.
Elastic Limit.
Inches.
Inches.
Sq. In.
Per Cent.
Pounds.
Pounds.
2
8 X 10
80
3.92
50,763
33,258
If
8 X 10
80
3.45
53,361
35,032
If
8X9
72
3.34
53,154
35,323
If
8x8
64
3.24
53,329
33,520
tt
6X9
54
3.27
52,819
34,840
H
6X7
42
3.53
52,733
34,606
li
6x6
36
3.41
53,248
33,520
H
6X5
30
3.31
54,648
34,695
1
5X5
25
3.14
53,915
36,287
* These two tables of results are compiled from data given in the report of the U. S..
Board on Testing Iron and Steel, vol. i, 1881.
CHAPTER IX.
STEEL.
METHODS OF MANUFACTURE.
96. The Crucible Process is the oldest and simplest of those used at the
present time, and is still -used for the finer grades of tool-steel. A pure
grade of wrought iron is first rolled into flat bars and cut into convenient
lengths. These are then heated for from three to six days in " cementing
furnaces," where they are tightly enclosed in boxes separated by layers of
fine charcoal. This recarburizes the wrought iron at the rate of about -J-
inch in depth every twenty-four hours, and makes cement or blister steel.*
This was the steel of commerce until 1740, when it was first remelted in
crucibles (by Daniel Huntsman, in England), thus making what is still
known as crucible steel. These crucibles are now heated in a Siemens re-
generative gas-furnace, simliar to that described in Art. 98. Cheaper grades
of crucible steel are made by remelting in crucibles Bessemer scrap. The
cheaper Bessemer and open-hearfJt processes have now limited the use of
the crucible process to the manufacture of high-grade tool and spring steel
only. In 1896 the total annual capacity of crucible-steel furnaces in the
United States was about 100,000 gross tons.
97. The Bessemer Process. This consists of a decarburization' of crude
pig iron by means of finely divided air-currents blown through the iron when
in a melted state. The oxygen in the air burns out the silicon and carbon
from the melted cast iron, and this combustion so raises the temperature of
the melted mass that it remains a mobile fluid even after these foreign in-
gredients have been almost wholly removed. This requires a very high
temperature indeed, and one which could not be obtained in the ordinary
puddling-furnace. The purified iron is then "recarburized" by adding
melted spiegeleisen which contains from 10 to 20 per cent of manganese,
and also some carbon and silicon. This manganese unites with the large
amount of oxide of iron present, which was formed by the blast, and which
would cause the product to be red-short and to crumble in working, and
*Also called shear, double shear, or German steel.
133
134
THE MATERIALS OF CONSTRUCTION
at the same time the proportion of carbon is brought up to any desired
amount. The whole mass is then poured off into ladles, and thence into
cast-iron moulds. These masses of cast steel are now called ingots. This
process was invented by Sir Henry Bessemer of England, and perfected by
G. F. Goranssou of Sweden,* in 1858.
In this process the crude melted iron is tapped directly from the cupola
furnace, and in Sweden directly from the blast-furnace into the converter,
which is a large steel vessel, mounted on trunnions, lined with refractory
materials, with a removable bottom provided with many small openings
or tuyeres. This vessel is turned down into a horizontal position to
receive its charge. The blast is then started and the vessel raised
to a vertical position, the air-pressure being sufficient to keep the
melted iron from entering the air openings in the base. In Sweden,
where a very pure iron is used, the blast is stopped when the appearance of
the shower of sparks issuing from the mouth of the converter indicates the
TIME /N MINUTES
2 4- 6
F IGt 62. Chemical Reductions of au Open-hearth Converter. (Howe, Jour. IT. &
St. Inst., vol. n. p. 102.)
desired percentage of carbon, when the metal is at once poured off into the
moulds. As this criterion is a very uncertain one, it is customary in this
country to continue the blast till practically all the carbon has been con-
sumed, this stage being clearly indicated by the changed appearance of the
flames. The addition now of manganese and carbon, in any desired pro-
portions, is readily made. It is important to remember that by this process
no sulpliur or phosphorus is removed, and hence only pig irons compara-
tively free from these elements can be used for the Bessemer process, such
iron being known as Bessemer pig. The iron must contain from 1| to 2
per cent of silicon in order that by its combustion it may sufficiently heat
the charge to keep it fluid when the carbon is consumed. If there is as
much as 2J per cent of silicon in the pig-iron, from 10 to 15 per cent of
* See paper by Prof. Rich. Ackerman of Stockholm, in Trans. Am. Soc. Min.
Engrs., vol. xxn. p. 266.
STEEL.
135
cold steel scrap can also be worked into the charge without chilling it.
The rate of burning the silicon, carbon, and manganese is shown in Fig. 62.
The combustion of the silicon brings to the mass about nine times as much
heat as the combustion of the same amount of carbon. One reason for this
is that the products of the combustion of silicon form a slag which remains
FIG. 63. Plan and Sectional View of a Bessemer-steel Plant.
in the converter, while the product of the combustion of carbon is a gas
which passes off and carries much heat with it.
In Figs. 63 to 67 are shown the characteristic features of a standard
American Bessemer-steel plant. On the right of Fig 63, in plan, are
hown four cupola-furnaces, with a blower, for melting the pig iron. The
sectional view shows these to be placed at a high elevation, so that the
melted iron, received in the ladles K, which stand on platform-scales for
weighing the charge, can be poured into the spouts M N 9 and run directly
into the mouth of the converter, which is then turned into the position
shown in Fig. 65. The blast is now started through the base of the con-
verter, and it is raised to a vertical position, Fig. 66, and the blast kept on
136
THE MATERIALS OF CONSTRUCTION.
FIG, 64.-Views of the American Form of Bessemer Converter, showing the Movable
Bottom.
65. Receiving the Charge.
STEEL.
137
till first the silicon and then the carbon has been burned out. The con-
verter is then again revolved to a hori-
zontal position, and the blast stopped.
The proper amount of melted spiegel-
eisen which is kept melted in the two
reverberatory furnaces RR is then run
into the converter, whereupon it is at
once poured into the ladle, which is
operated by a crane which swings it in
FIG. 66. The Bessemer Converter in
Action.
FIG. 67. The Valvular Ladle.
the path of a circle over the several ingot-moulds, the metal falling
through a valve at the bottom, as shown in Fig. G7. The motions of the
converter and also of the crane, as well as the blast, are all controlled from
one platform by levers operating hydraulic machinery.
The Swedish practice of taking the iron directly from the blast-fur-
nace is growing in this country, but it is here first run into a large vessel
containing from 100 to 150 tons of melted iron, called a mixer, in order to
obtain a more uniform product. This mixer also serves as a reservoir for
equalizing the inequalities of supply from the blast-furnace and demand
from the converter. From this mixer it is drawn into ladles on cars, and
run to an elevated platform and poured into the converter. This is called
the direct process. It may be employed when the blast-furnaces are re-
moved from the Bessemer plant as far as one or two miles.
Recently a means of removing phosphorus has been found in the addi-
tion of calcined lime to the charge in the converter.* The phosphorus
unites with the lime and so passes into the slag. In this case the lining of
the converter must also be " basic " to keep the slag from uniting with it
and so rapidly consuming it, so the lining is then made of a calcined mag-
nesian limestone (dolomite) and tar, made into brick, or rammed into place.
This is called the basic Bessemer process, but its use had been abandoned
in America because of some unfortunate failures when first introduced.
*By S. G. Thomas and P. C. Gilehrist, England, 1878.
138
THE MATERIALS OF CONSTRUCTION.
The method has now (1896) been revived at Troy, N. Y., with marked
success.
The Bessemer is the cheapest known process of making steel. This
process alone has revolutionized many lines of industry, and has led to the
replacing of wrought iron by steel in all the more important uses of these
materials. The Bessemer process is now used exclusively for making steel
rails for steam and electric roads, and for all the cheaper grades of steel
plates and structural forms. For the better grades of structural material it
is being replaced by
98. The Open-hearth Process. In this process pig iron, cast iron, and
wrought-iron and steel scrap are converted into steel under the direct
action of an oxidizing flame in a regenerative gas-furnace. It was patented
in 1845 by Heath, but was not found to be successful until Siemens had
developed his regenerative gas-furnace about 1862. Since about 1870
Am.BiukNot.Co.N.I.
FIG. 68. Transverse Section of a Typical Open-hearth Regenerative Gas-furnace.
these furnaces have multiplied rapidly, and in 1896 the total capacity of
these furnaces in the United States was 2,400,000 gross tons, as against a
total capacity of 9,400,000 gross tons by the Bessemer process.
The more common type of furnace used for this purpose is shown in
Figs. 68 and 69. The fuel used is what is known as producer-gas. This
is a mixture of carbonic oxide and hydrocarbons, diluted with about 60
cent of nitrogen. It is formed in gas producers in which coal is burned h
air-tight ovens with an insufficient supply of air, this supply being fed ii
STEEL.
139
under pressure and in known volumes. This producer-gas is brought to
the hearth area of the open-hearth furnace through a passageway entirely
filled with red-hot fire-brick stacked to form an open checker-work, as
shown at E and F ir Figs. 68 and 69.
FIG. 69. Longitudinal Section of a Typical Open-hearth Regenerative Gas-furnace.
As this hot gas enters the furnace area it is mixed with streams of hot
air which has also been drawn in over red-hot brick surfaces, and the com-
mingling of these red-hot gases, in proper proportions to produce complete
combustion, develops the most intense heat possible to obtain by the com-
bustion of gases. In the reverberatory furnace., where the flame of an
ordinary coal fire is emplo}^ed, the maximum temperature attainable is
about 3,500 degrees F., but in the Siemens regenerative gas-furnace a tem-
perature of 4500 degrees F. may be maintained. The regenerative princi-
ple consists in the utilization of the heat of the escaping gases in reheating
the fire-brick placed in the air and gas passageways. To do this it is of
course necessary to alternate the incoming and the escaping gases in two
sets of passages, this being done by simply moving certain valves every
twenty or thirty minutes.
In Fig. 68 Jris the furnace hearth; EE are air-chambers and FF gas-
chambers, the checker brickwork being shown in one only of each, but it
really fills all four of these passageways. The red-hot gas enters the
furnace through the lower ports H, Fig. 68, and BE, Fig. 69; while the air
enters just above these through an annular space /, Fig. 68, and C, Fig.
69. The furnace itself, therefore, is like a great argand gas-burner in its
method of receiving and burning the gas. The depressed roof of the fur-
nace throws the heat strongly upon the materials placed on the hearth,
140 THE MATERIALS OF CONSTRUCTION.
while the gases themselves are forced to play upon the melting metal. The
flame has an excess of oxygen so that it is an oxidizing flame, and would
rapidly waste unmelted wrought-iron or steel scrap placed in it. It is cus-
tomary, therefore, to place first on the hearth pig or cast iron on which the
oxidizing flame acts by consuming first the silicon and then the carbon, at
the same time oxidizing some iron which by melting forms a slag which
floats on the bath of melted metal. After such a bath has been prepared
the wrought-iron and steel scrap can be thrown in, since these will now be
covered by the bath and so protected from the oxidizing flame. The
facility with which such scrap can be remelted and made over into new
ingots by this method is one of its chief elements of value. If one could
always choose his ingredients at pleasure he could so proportion them that
little or no decarburizatiou would be necessary, a simple melting together
giving the requisite proportions.
The final removal of any excess of carbon, after the products have melted,
is effected by means of the melted oxide of iron, which floats on the surface
at first, but which afterwards becomes thoroughly mixed with the mass by
the boiling action of the escaping gases when the temperature becomes high.
Some of the oxygen of this iron oxide combines with the carbon of the
melted iron and comes to the surface as carbonic oxide, where it is burned to
dioxide and passes out with the other consumed gases. This also restores
a corresponding portion of the iron of the oxide slag to the metal bath and
so adds to the product.
When a large amount of pig or cast iron is to be reduced, it is common
to charge a suitable air ant of oxide of iron ore to supply the requisite
amount of oxygen to decarbonize the cast iron, and so to hasten the process
and also to avoid the necessity of creating so much artificial oxide of iron by
the oxidizing flame. The remaining portion of the oxide slag not destroyed
by giving up its oxygen to the carbon in the bath is neutralized and chemi-
cally destroyed by adding a charge of spiegeleisen containing 20 or 30 per
cent of manganese, or an artificial f erromanganese containing some 80 per
cent of manganese, just before pouring. The manganese unites with the
oxygen of the slag, and restores the iron to the bath the same as is done in
the Bessemer process. There, however, it was usually desired to add carbon
also, and hence this process has come to be known as recarburization, or the
adding of a recarburizer. In the open-hearth process the carbon is not all
burned out, as it is in the Bessemer process, but by taking samples out with
a sn^l dipper or ladle, and casting these, cooling them in water, and break-
ing them, the operator can tell when he has the carbon ingredient brought
to the desired amount. He seldom wants to add carbon, therefore, in the
open-hearth process, but must add the manganese to destroy the oxide slag"
which, if poured off with the iron, would make it red-short, or unmali^uble
in the rolls. The manganese charge should more properly be called a
" deoxidizer," but, by analogy from the Bessemer process, the same term of
" recarburizer" is applied here to :' !j manganese charge.
STEEL. 141
The pervasive action of this manganese charge in the open-hearth furnace
is very remarkable. The manganese has so strong an affinity for the oxygen
in the iron oxide that it seems quickly to seek it out throughout all parts of
the bath, and even when added to the metal in the ladle, after teeming,
it seems to be equally effective. In the case of the Bessemer process, how-
ever, it is necessary not only to destroy the iron oxide, but to add a carbon
ingredient to the metal. To secure a uniform distribution of this carbon
element through the mass seems to require a thorough artificial mixing.
The only action of this kind which is secured in the Bessemer process is had
in the pouring off into the ladle and the drawing from the pouring-nozzle in
its bottom into the ingot-moulds. This does not insure a uniform distribu-
tion of the carbon, and hence it is not very uncommon to find great differ-
enced in the mechanical qualities of different portions of the same sheet of
Bessemer steel. In the open-hearth process little or no carbon is added in
the " recarburizer," so that the mixture retains the homogeneity it neces-
sarily secures from the violent boiling action of the bath. This greater
homogeneity and reliability, when judged by sample tests, has led to a
general preference for open-hearth steel by engineers in all kinds of struc-
tural designing.
Here, as in the Bessemer process, there is no elimination of the phos-
Iphorus and of the sulphur. This has greatly limited the range of materials
which could be fed to this furnace, and it has led here, as in the case of the
[Bessemer process, to the use of a charge of calcined lime to unite with the excess
f phosphorus* and hold it in the slag, which is then' drawn off. But, as with
(the Bessemer furnace, this lime would unite with ihe sand lining of the
(furnace to form a flux which would quickly destroy- this lining altogether.
prevent this, those furnaces in which lime is to be added to 4 the charge
re themselves lined with calcined dolomite limestone, and these are called
asic-lined furnaces, and this process has thus come to be known as the basic
pen-hearth process. It must be understood, however, that neither here nor
1 the Bessemer process docs the lining play any part in the process itself.
'he process, in each case, depends on the materials charged and not 011 the
urnace lining. The lining is simply made such as will not be attacked by
slag formed, and is always intended to be neutral, or inert. In 1896
bout one half of the open-hearth steel made in the United States was by
basic process.
To distinguish the ordinary open=hearth process, in which a sand or
ilica lining is used and no lime fed in the charge, that is to say, in will 'i
o attempt is made to remove any qf the phosphorus in the ingredients used,
rom this " basic " process, the former is now called the acid open-hearth
rocess. It was formerly known as the Siemens-Martin process, from its use
the ^iemens furnace and from the fact that the Messrs. Martin of France
* When steel is very low in carbon some phosphorus, as 0.03 or 0.04 per cent, seems
esiruble to add strength to the metal. -ru.f
142 THE MATERIALS OF CONSTRUCTION.
first employed tlie open-hearth in this way, but without the Siemens regen-
erative gas-furnace.
99. Comparison of the Basic and Acid Open-hearth Processes. From
what has been given in the previous article it is evident that poor steel and
steel high in phosphorus may be made by either process, due to ignorance,
carelessness, or inexperience. By the use of the basic process ingredients
high in phosphorus may be employed, and thus the available materials are
very much increased and hence cheaper grades can be employed. The
process itself is somewhat more expensive than the acid process. When the
acid process is used, unless there is a rigid inspection and control over the
product, there is a danger that the cheaper (phosphorus) ingredients may be
used, and so lead to a brittle product, whereas if the basic process be speci-
fied and employed the cheaper ingredients are anticipated, and the removal
of the phosphorus provided for. The maker can be trusted to reduce the
sulphur in order to make the product malleable when hot, so as to roll
smoothly, as defects here are patent to any one. Engineers now generally
specify the open-hearth process without prescribing the kind of lining, but-
naming a maximum proportion of phosphorus. This upper limit of phos-
phorus is now commonly taken at from 0.06 to 0.08 per cent, but Mr. II. H,t
Campbell,* who is the highest authority from the standpoint of the manu
facturer, says this upper limit should now be made 0.04 per cent. If tin
phosphorus limit is not specified, or if specified but not determined by actua
tests, then it would be safer to specify the basic open-hearth method, f
100. Comparison of Bessemer and Open-hearth Steel. Comparing th
products only (not the processes) of these two general methods of makin;
steel on a large scale, we may say :
1. While for like chemical analyses like mechanical properties may b
anticipated from these two methods, yet all the unexplained accidental failure
of steel have occurred on Bessemer steel. Engineers have become suspicion
of it. Open-hearth steel is therefore more reliable than Bessemer steel.
2. Test specimens cut from different parts of the same Bessemer ste<*
plates have shown extraordinary differences in their mechanical propertie
This has never been found in open-hearth plates. They are therefore mo\
homogeneous than Bessemer plates.
3. Bessemer steel products found on the general market are apt to
extremely irregular in their composition, though rolled into like forms ar
eold to serve the same purposes. Open-hearth products purchased in the op<
market and designed to serve the same purposes are more\uniform in qualit
4. The open-hearth steel may be tested before tapping off, and its coi
position adjusted at pleasure, and this is usually done. Bessemer st
* Superintendent Pennsylvania Steel Co., Steelton, Pa.
f In January, 1896, there were in operation in the United States opeu-hearth-st
plants having an annual capacity of 2,430,000 gross tons, 700,000 of which capacity 1
been added in the preceding two years, more than one half of them using the " basi
process. Thirteen of these new furnaces are to be used in making steel castings.
STEEL.
143
TABLE XI. TESTS SHOWING THE HOMOGENEITY OF OPEN-HEART META*L"\^
Heat 10,699. Acid Open-hearth.
Heat 10,910. Acid Open-hearth.
Test-bars, f" rolled rounds.
Test-bars, 1" rolled rounds.
Elastic Limit,
Lbs. per Sq. In.
Ultimate
Strength,
Lbs. per
Sq. In.
Elongation
in 8 In.,
Per Cent.
Reduction
of Area,
Per Cent.
Elastic Limit,
Lbs. per Sq. In.
Ultimate
Strength,
Lbs. per
Sq. In.
Elongation
in Bin.,
Per Cent.
Reduction
of Area,
Per Cent.
35,900
53,510
28.75
64.14
31,140
52,760
32.75
60.60
36,450
54,790
31.75
64.58
31,790
52,750
32.75
61.20
36,000
56,150 28.75
62.71
31,540
53,000
31.50
56.50
SO, 225
55,690 31.25
64.48
31,250
52,000
32.25
63.30
36,090
55,830
31.00
64.71
31,250
52,320
34 00
64.10
36,315
55,830
32.00
65.18
31,080
52,320
32.50
57.10
36.740
56,370
31.50
64.84
31,160
52,830
32.75
61.80
36,350
55,090
29.50
62.87
31,250
53,160
32.75
58.10
36,450
57,510
31.25
64.25
31,040
52,160
32.75
61.80
36,125
56,900
30:75
64.26
32,050
53,840
32.50
59.60
37,580
56,600
33.50
64.16
31,660
53,580
32.50
61.10
36,900
57,510 30.50
65.16
31,700
52,480
32.25
53.10
37,220
57,420
31.25
63.28
32,550
52,580
34.00
63.10
37,130
57,280
31.75
64.16
32,570
52,960
32.75
65.40
36,000
57,050
31.25
65 . 75
33,330
53,050
33.00
60.40
35,860
57,190
31.25
64.23
33,580
53,860
33.00
60.40
36 615
57,440
31.25
64.74
36,450
57,670
31.75
66.46
37,165
57,580
32 . 75
63.68
36,640
57,350
31.25
63 . 18
56,538
Av. 36,510
31.15
64.34
Av. 31,809
52,853
32.75 60.48
i
1
Heat 11,018. Acid Open-hearth.
Heat 1,820. Basic Open-hearth.
Test-bars, f" rolled rounds.
Test-bars, %'' rolled rpuuds.
Elastic Limit,
Lbs. per Sq. In.
Ultimate
Strength,
Lbs. per
Sq. lu.
Elongation Reduction
in 8 In., ( of Area,
Per Cent. Per Cent.
Elastic Limit,
Lbs. per Sq. In.
1 Ultimate
Strength,
Lbs. per
Sq. In
Elongation
in 8 In.,
Per Cent.
Reduction
of Area,
Per Cent.
36,700
58,400
28.00 66.20
33,065 ! 48.340
34.50
71.87
37,150
57,840
30.75
65.60
31,530 47,380
35.00 72.05
37,280
56,880
29.50
58.60
33,650
48,450
35.00
72.05
36,060
56,940
28.50
53.30
31,600
48,230
37.00
74.14
36,420
56,700
30.75
63.10
33,340
49, 1 75
36.25
70.09
36,060
57,180
30.25
65.50
32,760
48.560
33.75
79.25
35,780
56,800
31.00
65.40
33,260
47,730
35.00
74.49
36,700
57,440
30.00
63.80
32,130
48,785
34.00
71.80
35.780
56,800
31.00
65.40
32,935
48,640
34.25
71 ? 92
36,700
57,440
30.00
63.80
33,270
49,440
34.00
71.48
35,700
56,900
32.50 , 67.10
32,900
47,835
34.00
72.72
37,020
57,180
31.25 59.70
31,920
48,050
33.75
71.42
37,400
57,320
30.50 ! 68.10
32,185
48,360
36.25
74.28
37,260
56,780
30.25
67.20
33,880 48,400
33.75
73.64
37,480
57,420
31.00
66.30
....
34. 15
Av. 36,634
57,201
30.25
63.94
Av. 32,745
48,384
72.49
144 THE MATERIALS OF CONSTRUCTION.
usually goes as blown, without correction. The open-hearth product is
therefore under better control.
5. The remarkable homogeneity of open-hearth steel is indicated by the
preceding series of tests (Table XI) on specimens cut from different por-
tions of plates rolled from four different heats.*
101. Molecular Structure of Wrought Iron and Steel. One of the most
important facts for the engineer to fix in his mind is this: All grades of iron
and steel, originally formed in a molten state, will always thereafter, when
cooled after either a melting, forging, or rolling, take a crystalline form.
That is to say, all cast, or ingot, metal is always crystalline whether cast,
hammered, or rolled to its final forms. f This includes all the grades of
" steel " as given in the second classification, p. 89, as well as the cast iron
and cast steel.
It may also be said that wrought iron also shows a crystalline structure
whenever a portion of the iron in the puddle-ball was in a liquid condition
when removed from the furnace. This liquid iron may result either from
the entangling of unreduced but melted cast iron in the glutinous mass, or
from too high a temperature of the furnace, resulting in the melting down
(burning) of the reduced metal, or wrought iron proper, which when
".brought to nature" and at the proper temperature should be a spongy,
pasty mass, sufficiently firm to be handled with the puddling -bars. This is
only a special case of ingot metal, since so much of the iron in the puddle-
ball as comes out in a liquid form is, within itself, free from the slag which
covers the more pasty iron mass, within and without, as a slime might
adhere to the entire internal and external surface of a sponge which has been
lifted from it and squeezed. Any given portion of this puddle-ball, when
finally rolled out into plates and bars, will become a small filament or fibre
of the cross-section, but greatly extended in the direction of the rolling.
Thus, if a small pocket of liquid iron was entangled in the ball, this would
become a small crystalline thread throughout the bar. A larger mass of
melted metal would make a longer crystalline portion, of the cross-section.
When wrought iron is properly made, that is, when it is entirely reduced
or " brought to nature," and when the furnace is not so hot as to melt the
pasty mass, wrought iron will be found to be practically free from crystalline
formations, and to be wholly fibrous. The fibrous structure is due to the
continuous mixture of the slag with the iron, which, after repeated piling andi
rolling, leaves the slag so distributed through the mass in thin filaments as,
to prevent any visible crystalline arrangement of the molecules, although!
each such filament is really a series of distorted (usually elongated) crystalline
forms.
* From H. H. Campbell's paper on " The Open-hearth Process" before the World's
Engineering Congress. Trans. Am. Inst. Min. Engrs., vol. xxn. p. 352.
f That is to say, when shaped while hot. When shaped cold, as by cold rolling or
-wire-drawing, the crystalline form may be partly or even wholly destroyed.
STEEL. 145
102. Fracture Showing Structure. In order to obtain a normal fracture
of any malleable metal, or one which x shows the true character of the molec-
ular arrangement, uninfluenced by the distorting effects of the forces used
to produce the fracture, it is common to nick the specimen with a chisel and
bend it.* This, however, subjects one side of the uncut section to a com-
pression, and the other to a tension; and even if a fracture is effected, the
entire surface has not been treated alike. It is better, therefore, to turn a
sharp groove into the side of the specimen all around (QY nick it with a
chisel), and then to pull the specimen in two in a testing-machine. This
will always reveal the true structure of the metal, without the distorting
effects accompanying the cold drawing out (elongation) of the specimen which
is purposely sought in the ordinary tensile test. To insure against any
elongation whatever, the tool used must be perfectly sharp at the point, so
that the bottom of the groove is a true angle, and not a curve with a finite
radius. When good structural steel is tested in tension it elongates at the
ruptured section fully 100 per cent; that is, it stretches here to more than
twice its original length, and this cold drawing out of the metal wholly
destroys its original molecular arrangement, so that the fracture always look
" fibrous " or " silky." This universal appearance of soft and medium steel
when pulled in two leads many persons to suppose that this material has not
normally a crystalline arrangement. When, therefore, they find this same
material broken in use, as on a screw-thread, or at a shoulder, or at a sudden
reduction of section, and they discover it having a wholly crystalline struc-
ture, they conclude that this is abnormal, and that the material has
"crystallized in service." The simple test described above, of pulling a
grooved specimen, will prove that all grades of steel, even to the softest ingot
iron, has normally a crystalline structure. If the nicking tet be cited to
prove this fact, it is often claimed that the nicking produced such a jarring
of the metal as to cause it to instantly rearrange its molecules, while cold and
rigid, into the crystalline form! Surely the grooving in a lathe is not open
to even this shadowy ground of suspicion. -
If wrought iron be grooved and pulled as here described, it will be found
to be apparently wholly fibrous (if of a superior quality, the crystallized
filaments being so small), or containing occasional small crystalline patches,
(if of an ordinary quality), or sometimes nearly wholly and coarsely crystal-
line (if of a very inferior quality). It is therefore said to be normally of a
fibrous or non-crystalline structure. Now when wrought iron breaks in
service and reveals a coarsely crystalline structure, it simply indicates, in
the opinion of the author, the original poor quality of the material, and does
not prove that the material had crystallized in service as is generally sup-
* Metcalf affirms that a skilful workman can grade steel, by the fracture, for cus-
tomers, so closely that "year after year not one piece will vary in carbon more than
0.05 per cent above or below the mean for that temper." Steel, p. 6. This can only be
true of material from the same establishment produced under like conditions.
146 THE MATERIALS OF CONSTRUCTION.
posed. This subject has been discussed at considerable length in Article
93.
103. Structure of Steel as Affected by Heat Treatment. While steel or
ingot iron is entirely free from slag and similar foreign ingredients, it must
not be regarded as a simple or single mineral, or substance, but rather as a
substance,- like granite, made up of a number * of separate minerals or
" metarals " (a term suggested by Howe), each crystallizing out by itself,
or being left as a matrix after the more controlling minerals have crystallized
out. The particular final arrangement depends on which of these various
proximate combinations control in the crystallizing stage of cooling, and also
on the heat treatment it receives. Concerning the effect of the heat treat-
ment on the appearance of the fracture, the following statements are based
on the discussion of this subject by Howe ( 240-250). (See also Appen-
dix A.)
1st. There is a critical temperature, at a " low yellow " heat (lower for
high-carbon and higher for low-carbon steel), above which the material
forms rapidly into coarse crystals.
2d. If cooled either slowly or rapidly from above this temperature, it is
coarsely crystalline, the coarseness of the crystals depending on the time
allowance for their formation when at this higher temperature.
3d. If worked (forged or rolled) in cooling from this higher temperature,
the crystallization is that characterizing its temperature when leaving the
hammer or rolls.
4th. If raised from a temperature below a low red, just to this critical
temperature, whatever its previous condition, or if worked down to this
critical temperature in 'cooling from a higher, and cooled rapidly as by
quenching in water or oil, it is so finely crystalline as to appear amorphous,
or porcelanic, to the naked eye. If cooled slowly from this critical tempera-
ture, it is finely crystalline to the naked eye.
5th. Since there is no tendency to crystallize below a low red heat, it is
sufficient to cool rapidly from the critical temperature (low yellow) down to
a low red, or to continue to work the metal to this temperature, after which
the cooling may be slow, thus preserving the porcelanic fracture and obtain-
ing a greater toughness.
6th. As an illustration of the effects of these different treatments, we
have f
For slow cooling after forging, size of grain 0.1414 in. diam.
Reheated to the low yellow and cooled slowly, size of grain 0.0048 "
Eeheated to low yellow, quenched in water to low red, and
then slowly cooled, size of grain 0.0004 "
* Seven such having already been distinguished; see Howe's Metallurgy of Steel, 237.
Osmond has recently added at least two new ones, and furthermore diamond has now
been isolated from certain steels. Stahl u, Eisen, vol. xvi, No. 15, 1896.
f Quoted by Howe, 250, from Chernoff .
STEEL. 147
The size of this last is entirely too small to be discoverable by the naked
eye, and hence it would appear amorphous or porcelanic.
7th. The bright surfaces observed on a steel fracture may be either
cleavage planes across individual crystals, or their exterior sides, depending
on which surface offers the least resistance to rupture. In either case the
size of these individual surfaces is a true index of the coarseness or fineness
of the crystalline structure.
8th. The appearance of a steel fracture is thus a good indication of the
condition of the metal when it left the rolls, or of its subsequent treatment.
The student should himself verify these statements at the forge.
This subject will be discussed again when treating of hardening, temper-
ing, and annealing. See Arts. 130 to 134.
104. The Mechanical dualities of Steel. When the term steel is made
to include all grades of ingot metal as well as converted wrought iron, its
qualities are so various as to necessitate a series of trade names, such as
flange-steel and shell-steel, for boiler-plates; tank-steel, for plates of uncer-
tain quality and of cheap manufacture, often used where a better grade is
needed; structural steel, both mild and medium, used for all kinds of
structural shapes, as angles, I beams, channels, T's, etc. ; rail-steel, used
for railway rails, both steam, and electric; machinery-steel, especially adapted
to forging and welding; tool-steel, spring-steel, saw-steel, etc.
The special qualities required of these various grades of steel are approxi-
mately as follows:
Fire-box Steel, Flange-steel, and Rivet-steel used for locomotive fire-
boxes, boiler-heads, rivets, and other purposes where it is subject to great
deformations in service, or where it must be shaped, or dished, in manufac-
ture in such a way as is only practicable with a very soft, or pliable, semi-plastic
material. This grade of steel, therefore, must be ductile rather than strong,
and extremely tough and capable of resisting great abuse, either cold or hot,
without losing its strength or toughness. This steel has a tensile strength
of from 50,000 to 00,000 Ibs. per square inch; an elastic limit* of from
30,000 to 40,000 Ibrf. per square inch; an elongation of from "25 to 35 per
cent in eight inches; a reduction of area of from 50 to 65 per cent at the
fractured section. It will also bend cold through 180, and close down per-
fectly flat, as in Fig. 71, either under the hammer or in a press, up to a thick-
ness of f inch to 1 inch without showing any sign of fracture. One could
literally tie it in knots while cold, as in Fig. 70 without sign of rupture.
Made now mostly by the open-hearth process.
Shell-steel is used for boiler-shells, and for structural purposes, where a
greater tensile strength may be obtained at the expense of some ductility.
This steel has a tensile strength of from 55,000 to 05,000 Ibs. per square
inch; an elastic limit of from 33,000 to 44,000 Ibs. per square inch; an
elongation of from 25 to 30 per cent in eight inches; a reduction of urea of
* Here the commercial elastic limit is meant, or the break-down point.
148
THE MATERIALS OF CONSTRUCTION.
from 50 to 60 per cent at the fractured section. Made by the open-hearth
and the Bessemer processes.
PIG. 70. Knot of Rivet-steel, f in. in diameter, pulled to Incipient Fracture by the
Author.
FIG. 71. Flange-steel Plates, | in. thick.
Tank-steel has no particular limits of quality. It is a term which means
the cheapest grade of steel plate on the market; is sold with no guarantee;
its qualities usually unknown, or at least unrevealed ; is likely to be extremely
STEEL. 149
various in quality, even in different parts of the same plate; and should be
used only for indifferent purposes. Made by the Bessemer process. *
Structural Steel is used for bridges, roofs, steel skeletons of buildings,
etc., and should be of a superior quality. Several grades are recognized,
although, these names are used loosely, and have no precise meaning. Thus
soft and mild structural steel may be regarded as the same in quality as the
flange and shell steel respectively described above. Medium structural steel
might be considered as having a tensile strength of from 00,000 to 70,000
Ibs. per square inch; an elastic limit of from 35,000 to 45,000 Ibs. per square
inch; an elongation of from 20 to 25 per cent in eight inches; a reduction
of area of from 50 to 60 per cent at the fractured section. Hard structural
steel having a tensile strength of from G5,000 to 75,000 Ibs. per square inch
is now used but little, as it suffers too much from shearing, punching, and
assembling to make it as reliable as is desired for such material. Made by
"both the open-hearth and the Bessemer processes.
Rail-steel must be very hard, with a high elastic limit to resist abrasion
and wear, while it must also have great strength and resilience, or resistance
to shock. It is a hard steel, having a tensile strength of from 70,000 to
80,000 Ibs. per square inch; an elastic limit of from 40,000 to 50,000 Ibs.
per square inch; an elongation of from 15 to 20 per cent in eight inches; a
reduction of area of from 40 to 50 per cent at the fractured section. Made
by the Bessemer process.
Ordinary Tool-steel, Spring-steel, etc., are harder grades, capable of being
hardened and tempered, in which the tensile strength and ductility are of
less importance than its hardening qualities. The tensile strength here may
be from 90,000 to 160,000 Ibs. per square inch and the elongation very smalt,
depending on the particular temper given to the specimen. Made by the
Bessemer, open-hearth, and crucible processes.
The Finer Grades of Tool- and Spring-steel, especially such as are to be
used for edge-tools, are still made of crucible-steel. Metcalf ^'"es the follow-
ing tempers for their respective uses :
" 0.50 to 0.70 C for hot work and for battering-tools, and for tools of
dull edge.
"0.70 to 0.80 C for battering-tools, cold-sets, and some forms of reamers
and taps.
" 0.80 to 0.90 C for cold-sets, hand-chisels, drills taps, reamers, and
dies
"0.90 to 1.00 C for chisels, drills, dies axes, knives, and many similar
purposes.
"1.00 to 1.10 C for axes, hatchets, knives, large lathe-tools, and many
kinds of dies and drills if care be used in tempering them.
"1.10 to 1.50 for lathe-tools, graving-tools, scribers, scrapers, small
drills, and many similar purposes.
" The best all-around tool-steel is found between 0.90 and 1.10 C. This
can be adapted safely and successfully to more uses than any other temper."
* See Plate III for the sad effects of using such material iu an important structure.
150 THE MATERIALS OF CONSTRUCTION.
This is at and just above the point of complete saturation of combined
carbon.
105. Qualities of Steel as Affected by its Chemical Composition, Mr.
H. M. Howe, the highest authority on this subject, says (1): "I conceive
steel to consist (A) of a matrix of iron which is sometimes (as in ingot iron
and annealed steel), comparatively, or even quite pure, and sometimes (as in
hardened steel, manganese-steel, etc.), chemically combined with a portion,
or even the whole of the other elements which are present, probably in
indefinite ratios, its mechanical properties being greatly affected by them;,
and (B) of a number of independent entities which we may style ' minerals,' *
chemical compounds of the elements present, including iron, which crystal-
lize within the matrix, and by their mechanical properties, shape, size, and
mode of distribution also profoundly affect the mechanical properties of the
composite mass, though probably less profoundly than do changes of corre-
sponding magnitude in the composition of the matrix."
And again ( 237), " From the microscopic study of polished sections
iron (and steel) appears to be constituted, like granite and similar compound
crystalline rocks, of grains of several distinct crystalline minerals, of which
seven common ones have already been recognized, through peculiarities of
crystalline form and habit, color, lustre, hardness, and behavior towards
solvents. Their nature, size, shape, and orientation, and through these the
structure and physical properties of the metal as a whole, seem to depend
chiefly
1. On the ultimate chemical composition of the mass;
2. On the mechanical treatment which it has undergone;
3. On the conditions under which it has been heated and cooled, i.e.,,
its " heat-treatment," which may induce the ultimate components of the
mass to regroup themselves in new combinations, thus causing one set of
minerals to give place to another."
When the iron or steel is in a state of fusion the ingredients are in mutual
solution, and they do not separate until the fluid mass congeals or hardens,
when one or another of these mineral ingredients crystallizes out first, and
thus gives its own characteristics greater prominence than the other minerals
which form the matrix, or which form in crystals later, and subject to the
limitations as to form and size imposed by the previously formed crystals.
Just as the character of a granite rock, therefore, is to be judged from
the character of -its mineral constituents, as proximate chemical compounds,
and very imperfectly from ever so exact a determination of its ultimate ele<
ments, so we must learn to rely with less assurance on the ultimate chemica
analysis of iron and steel, and more on the proximate chemical compounds
formed therefrom. Unfortunately these latter are very difficult of deter
mination, or even of identification, and hence we know very little abou
them. It is for this reason that we are as yet unable to infer with any grea
* But for which Mr. Howe suggests the term ' ' metarals. "
STEEL. 151y
assurance the mechanical properties from the chemical analysis. Such con-
clusions as may be drawn from chemical composition are partially summarized
in the following articles.
INFLUENCE OF CARBON ON IRON.
106. Combination of Carbon with Iron. The effects of carbon on iron
are more pronounced and useful than those of any other known chemical
element. Iron absorbs carbon readily, becoming saturated with about 4.6
per cent of it, unless aided by manganese, when it may absorb as much as
7 per cent.
Cast Iron may be regarded as supersaturated with carbon, or as having
some 4 per cent of this element in some form.
Wrought Iron is nearly free from carbon in any form, having perhaps not
over 0.10 per cent.
Steel (ingot metal) may have anywhere from 0.05 to 1.50 per cent
of carbon, the upper limit usually being about one per cent. For extreme
hardness, steel may be made with as much as two or even three per cent
carbon, while 0.9 per cent C gives maximum working qualities for tool-
and spring-steels. (This is the point of perfect saturation of combined
carbon.)
Three States of Carbon in Iron. Carbon is found in iron in three rad-
ically different states:
1. Mechanically mixed, in the form of a graphite, this being thrown out,
or excluded, when cast iron crystallizes from a melted state.
2. Chemically combined in unknown proportions, this forming a very
hard and strong compound, and the carbon so combined being here called
hardening carbon.*'
3. Chemically: combined, as a carbide of iron (Fe s C), up to the satura-
tion-point of 0.9 C. It is intensely hard according to Sqrby, though it
does not appear to contribute to the hardness of steel to the same degree as
the hardening carbon, f
We shall therefore speak of the uncombiued carbon as graphite (often
called graphitic carbon), and the chemically combined carbon as hardening
carbon and cement carbon.
When the metal is fused all the carbon may be regarded as chemically
combined in the form of hardening carbon. When there is a great deal of
this, as in cast iron, a large proportion of it is thrown out as graphite in the
early stages of cooling, if sufficient, time be allowed for this action to complete
* Prof. J. O. Arnold (Sheffield) gives the formula Fe 24 C for this component. Trans.
Inst. Civ. Eng., vol. cxxin, 1896. Many authorities agree with Osmond in attributing
the hardness of quenched steel to an allotropic form of iron. Both sides are well
presented in the discussion of Arnold's paper, here cited. See Appendix A.
f Besides these diamond has recently been isolated, and Ledebur -adds "temper-
carbon."
152
THE MATERIALS OF CONSTRUCTION.
itself. When there is not over 0.9 per cent of total carbon, none of it will
appear as graphite in the cold product.
A change from hardening carbon to cement carbon occurs (time permit-
ting) at a low yellow heat, and no further change occurs below a low red
heat.
These changes and also the subsequent condition of the metal are indi-
cated graphically, in a general way, in Figs. 72 and 72a. Thus in Fig. 72
the total carbon in cast iron being about 4 per cent, as soon as it begins to
FIG. 72. Change of Carboii in Cast Iron.
FIG. 72a. Change of Carbon iii Steel.
congeal or crystallize, it begins to expel carbon in the form of graphite, and
this action is supposed to be completed when the metal has cooled to TV, at
which time the product has probably become wholly crystalline, having
perhaps less than one per cent of carbon left in the combined form, as
hardening carbon. It is new very granular in its nature, having little or no
cohesion, and this intermediate granular form will always prevent the rolling
of steel direct from the melted state. At the temperature W (low yellow) a
peculiar change occurs in the combined carbon, a large part of it passing
from the hardening to the cement form, if sufficient time be given at this
temperature for this to occur. This change in the carbon state is accom-
panied by a remarkable development of sensible heat, causing the color to
brighten up again, and this phenomenon is known as recalescence. This
marks the truly plastic state at which it should be worked. As shown in
Fig. 72a, there is no appreciable amount of graphitic carbon in steel, it all
being in chemical combination, but changing from the hardening to the
cement form, in a falling, and back again for a rising, temperature past
the critical low yellow heat.
The presence of the large amount of graphitic carbon in cast iron causes
it to fuse at a much lower temperature than steel, because of the recombin-
ing of this carbon, chemically, with the iron at this high heat. The fusing
temperature of steel is higher as the proportion of carbon is less.
\
STEEL. 153
107. Physical Effects in Steel of the Change in the Combined Carbon at
a Low Yellow Heat. Hardening and Tempering. This change in the
combined carbon of steel from hardening to cement and back again is
accompanied by a corresponding change in the crystalline arrangement, in
the appearance of the fracture, and in all its mechanical properties. Thus
if the region W V, Fig. 73, be passed quickly, as when the specimen is
quenched in water from a temperature above TF, there is very little change
in passing this critical temperature, and hence the carbon remains mostly in
the hardening state. This gives a very hard and brittle product (when the
percentage of carbon is high, or from 0.75 to 1.0 per cent), and in all cases
raises the elastic limit and the ultimate strength,* but reduces the ductility.
The crystalline arrangement, also, is now that which was formed on the first
cooling, above IF, it being very coarsely crystalline.
If the region W V be passed slowly, more especially if the specimen be
held at this temperature for a considerable period and then cooled slowly, the
combined carbon changes mostly to the cement state, and a great softening
of the material results. The only way to retain the carbon in the hardening
state, when cold, being to cool it quickly from a temperature above W.\
When steel has been hardened by sudden cooling from above IF, it can be
tempered, or softened, by heating again, to some temperature below V and
cooling slowly. The higher this tempering heat is, below a red heat, followed
by slow cooling (as in the air), the softer will be the product when cold, as
the more of the hardening carbon will be changed to the cement state. If
the reheating be carried to V or above, and cooled slowly, the carbon will be
(almost) wholly in the cement state, the temper then having been entirely
drawn. The particular temper required, therefore, is obtained by first
quenching from TF or above, then reheating to the required temperature
below F, and cooling slowly. This leaves the required portion of the carbon
in the hardening state, and gives the product the desired compromise
qualities of strength, hardness, and ductility, combined with toughness.
In the matter of the fracture, also, either a slow or a rapid cooling from
a white heat, without forging or rolling, leaves a coarse crystalline fracture.
If worked down to a red heat it gives a fine crystalline fracture.
It is of the utmost importance that the heating for both hardening and
for tempering should be uniform throughout the entire body of the specimen.
Evidently a liquid bath of some kind furnishes the ideal condition for both
heating and cooling. Thus a melted lead bath, kept stirred, may be used
for heating, and a mercury, brine, water, or oil bath for the quenching, or
sudden cooling. All hardening should be done by quenching from a rising
temperature, to preserve fineness of grain. The reheating of the hardened
* Quenching in water from a high temperature may impair the ultimate strength of
low carbon-steel. Quenching in oil seems always to increase the ultimate strength.
f If not uniformly heated when quenched, it is apt to break or crack from internal
stress. The different densities resulting from quenching from different temperatures,
may furnish a key to this action.
154
THE MATERIALS OF CONSTRUCTION.
steel for the purpose of tempering it, may be done by holding it over a fire,
or in contact with a heated mass of iron, or in boiling water, or hot steam,
or in some other way.
When clean iron or steel is heated in the open air, the oxide which forms
on the surface takes in succession the following well defined colors, namely :
light straw, straw, light brown, darkey brown, pigeon-wing (a purplish
brown), light blue, dark blue, and black. In tempering, the final " temper "
depends on which of these graduated colors has been reached, and followed
by slow cooling. Thus if only the first color indication, light straw, be
reached, and then the bar slowly cooled, evidently very little softening of
the hardened steel has resulted, and the product is left very hard, or it is
said to have a " very high temper "; whereas if the highest temperature had
been reached, at which the oxide had deepened to black, and then the bar
cooled slowly, it would be found to be quite soft, or the hardness would have
been entirely removed. The word " temper " then may have the following
meanings, according as it is used by the steel-maker or by the steel-user:
Steel-maker's Meaning.
Steel-user's Meaning.
Temper drawn at
Percentage of Carbon.
Temperature.
Name_of Color.
Verv hisrii
1 50 carbon
About 400 F
' Liffht straw "
Hiffh
1.00 to 1.20 C
450 F
* Straw "
Medium
.70 to .80 C
500 F.
' Brown "to
Mild
40 to 60 C
550 F.
"pigeon-wiug"
' Liglit blue "
Low
20 to .30 C
600 F
' Dark blue "
Soft, or dead soft
Under .20 C
' 650 F.
Black"
EFFECTS OF CARBON IN ITS VARIOUS STATES ON THE MECHANICAL PROP-
ERTIES OF IRON AND STEEL.
108. Not Fully Explained by Chemical Analyses. As shown in Art. 105.
the mechanical properties of iron and steel are not fully indicated by anj
ultimate chemical analysis of the material, bat are dependent on the pap
ticular combinations the elements may have formed. In Art. 106 it wa&
further shown that carbon is found in three distinct forms in iron and steel,
and that the physical qualities depend largely on these particular forms o
carbon. It is to be expected, therefore, that the mechanical qualities, o
the qualities shown by the material when resisting the action of externa,
forces, would also be found to be greatly dependent on these particular form
of carbon, combined and uncombined, or even on the total combined carbon
since this has been shown to exist in two very different states. The effort
therefore, of students of this subject to harmonize the results of mechanica
tests with the corresponding ultimate chemical analyses of the materials wa
foredoomed to failure. And since the proximate chemical analysis is as ye
impossible, we are wholly unable to predict mechanical properties froc
STEEL. ] 55
chemical analysis alone. When this is supplemented, however, with a full
knowledge of the heat treatment, as described in the preceding article, some
approximate knowledge of the mechanical properties is obtained. (See also
Appendix B.
109. The Hardening of Steel. A coarsely crystallized steel may be
reheated to a temperature between Fand IF, and cooled either slowly ^.or_
rapidly, and the fracture becomes finely crystalline or even porcelanic. (Jffli^t
what does occur in the hardening of high carbon steel is, and has long Aceri//
a matter of contention among our most distinguished metallurgical chemist^ ,
Osmond and his school contend for an allotropic form of iron (called " /?
iron," to distinguish it from the annealed form, which he called " a iron "),
not to be explained by a definite chemical compound, but containing carbon
in solution, while Prof. Arnold makes a very strong plea for a chemical
compound, Fe 24 C, which he calls a " sub-carbide " (Fe 3 C being the carbide),
this being he thinks the real composition of steel in a melted state, having
as much as 0.89 per cent C, which he calls the point of saturation.* When
this compound is cooled suddenly, this unstable sub-carbide hardens into a
solid without any change in its chemical composition; but when cooled
slowly, it passes at 400 C. into the carbide form, with pure iron (Fe 2l C =
Fe 3 C -f- 21 Fe) and with the evolution of heat. See Plates in Appendix B.
Prof. Arnold gives, as a general summary of his views, the following :f
I. The constituents of steel may be: (a) Crystals of pure iron which remain
bright on etching, (b) Crystals of slightly impure iron which become pale brown
on etching, probably owing to the presence of a small quantity of an intermediate
carbide of hypothetical formula FeioC. (c) Normal carbide of iron, FesC, which
exists in three distinct modifications, each one conferring upon the iron in which
it is found particular mechanical properties. (1) Emulsified carbide present in an
excessively fine state of division in tempered steels. (2) Diffused carbide of iron
occurring in normal steels in the forms of small ill-defined striae and granules.
(3) Crystallized carbide of iron occurring as well-defined laminae in annealed and
in some normal steels, (d) Subcarbide of iron, a compound of great hardness
existing in hardened and tempered steels and possessing the formuja Fe 2 4C. This
substance is decomposed by the most dilute acids, and at 400 C. it is decomposed
into Fe 3 C and free iron with evolution of heat. One of the most remarkable prop-
erties of this compound is its capacity for permanent magnetism, (e) Graphite
or " l temper-carbon.":}:
The existence of Fe 2 4C is proved by the fact that iron containing 0.89 per cent
carbon presents several correlative critical points when examined by different
methods of observation : (1) Well-marked saturation-points in the micro-structure
of normal annealed and hardened steels. (2) A sharp maximum in a curve the
coordinates of which are heat evolved or absorbed in recalescence and 'carbon per-
centage. (3) A point in the compression curve of hardened steels at which molec-
ular flow absolutely ceases. (4) A sharp maximum in a curve the coordinates of
which are carbon percentage and permanent magnetism in hardened steels.
II. The influence of annealing is (1) To increase the size of crystals and to
increase the intercrystalline cohesion when originally feeble or impaired. (2) To
convert elongated masses of iron containing diffused Fe 3 C into compact rounder
* With as much as 1 per cent, manganese he claims the point of saturation with carbon
is reached with 0.65 C. See Appendix B.
\ Trans, hist. Civ. Engrs., vol. cxxiu, 1896, p. 160. See also Appendix B.
\. Ledebur distinguishes between graphite and temper-carbon.
THE MATERIALS OF CONSTRUCTION.
bodies, containing laminae of crystallized Fe 3 C, between which the iron becomes
more or less dovetailed throughout the mass.
III. The approximate theoretical constituents of hardened and normal steels will
be in accordance with the figures given in Table IX. (These percentages, however,
can never be quite correct, because in practice hardened steels below the saturation-
point (0.89$ C) always contain a little Fe 3 C, and normal steels below the saturation-
point a small quantity of the intermediate carbide, FeioC(?) ). It is obvious that in
tempered steels an almost unlimited variety of constitutions and consequently of
mechanical properties is possible.
TABLE IX. APPROXIMATE THEORETICAL COMPOSITION OF HARDENED AND
NORMAL IRON AND CARBON STEELS REQUIRED BY THE SUBCARBIDE
THEORY HEREIN ENUNCIATED.
Hardened Steels.
Normal Steels.
Carbon
Fe.
Fe 24 C.
Fe 3 C.
Fe.
Fe,C.
Per cent.
Per cent.
Per cent.
Per cent.
Per cent
Per cent.
0.10
89
11
99
1
0.20
78
22
97
3
0.30
67
33
95
. 5
0.40
56
44
94
6
0.50
45
55
93
7
0.60
34
66
91
9
0.70
22
78
90
10
0.80
11
89
88
12
0.90
100
87
13
1.00
99
1
85
15
1.10-
97
3
84
16
1.20
95
5
82
18
1.30
93
7
81
19
1.40
91
9
79
21
1.50
89
11
77
23
IV. The snbcarbidc theory falls into line with the observations of e very-day
experience. For instance, the fact has long been known that pure carbon steel, con-
taining about 0.85 per cent of carbon, is the most suitable for steel which must carry
a cutting edge and yet be tough enough to withstand a sudden shock. Such steel is
therefore employed for cold sets.* It is also well known that a steel containing 1.3
per cent of carbon would be useless for such a purpose, as it would crack and
"snip." The reason is clear; such material is full of lines of weakness along the
junctions of the subcarbide granules with the surplus normal carbide membranes.
On the other hand, it is known that 'a steel harder than one carrying 0.9 per cent of
carbon is necessary for turning-tools. In such a case no shock has to be encountered,
so that the surplus Fe 3 C augments the hardness of the subcarbide with its own
intense hardness, and moreover adds 10 per. cent of a substance incapable of
"letting down" with the heat of friction. It is also clear that a steel with carbon
much below 0.9 per cent cannot carry a cutting edge, because of the presence of
particles of soft free iron amongst the mass of the hard subcarbide. See App. B.
110. Effect on Tensile Strength. While a diagram showing the relation
of tensile strength to percentage of carbon in steel gives a cloud of results
spread over a wide belt (see Howe's Metallurgy of Steel, p. 14), yet a
simple formula which is the algebraic expression of a line which traverses
this field well below its centre of gravity may be of some use. While many
* As required iu saws.
STEEL.
157
such formulas have been proposed, that of Salom * seems best to fit the
total assemblage of results and is easily remembered. It is
T 45,000 + 100,000(7, (1)
where T tensile strength of rolled steel in pounds per square inch
(up to C = 1.0 per cent);
C = percentage of carbon
The recorded tests show many results as much as 20,000 pounds per
square inch above this locus, and some 10,000 pounds below it. It may be
regarded, therefore, as traversing the lower edge of the middle third of the
cloud of recorded observations. As the maximum strength of steel is
reached with C = about 1.0 per cent, the above formula must not be used
above T = 145,000 and C 1.0. Higher values of tensile strength, as with
drawn steel wire, are due to the physical treatment and not to the chemical
composition. The elastic limit in both tension and compression may be
taken as GO per cent of the tensile strength.
As a result of a careful study of over four hundred tests accompanied by
their corresponding chemical analyses, made in the regular course of business,
Mr. William E. Webster f offers the following table, showing the variation
of strength of soft steel with varying percentages of carbon and phos-
phorus, assuming the manganese and sulphur are each zero. When either
or both of these latter are present, the values may be corrected by adding
the values given in the two auxiliary tables for the corresponding percent-
ages of these ingredients.
TABLE XII. ESTIMATED ULTIMATE STRENGTH OF STEEL FOR VARYING
PERCENTAGES OF CARBON AND PHOSPHORUS. \
On the assumption that neither manganese nor sulphur is present, the tabular values to be
increased for these ingredients by the amounts given in the two following auxiliary tables.
Carbon in Parts
of 1 per cent.
.06
.08
.10
.13
.14
.16
.18
.26
.22
.24
Percentage of
Phos. .000
39,550
41,150
42,750
44,350
45.950
47.550
49.150
50,750
52,350
53,950
" .01
40,350
41.950
43,750
45,550
47.350
49,050
50,650
52,250
53,850
55,450
u .02
41.150
42.750
44,750
46,750
48,7f>0
50,550
52,150
53,750
55.350
5(5.950
" .03
41.950
43,550
45,750
47,950
50.150
52.050
53,850
55,250
56.850
58.450
" .04
43,750
44.350
46.750
49.150 ! 51,550
53.550
55,150
56,750
58,350
59 950
" .05
43.550
45,150
47,750
50.3:,0
52.950
55,050
56,650
58,250
59,850
61,450
" .06
44,350
45,950
48,750
51.550
54,350
56 550
58,150
59.750
61,350
62,950
" .07
45.150
46,750
49,750
52.750
55,750
58.050
59,6:,0
61,250
62,850
64,450
" .08
45.950
47,550
50,750
53.950
57.150
59.550
61,150
62.750
64,350
65 950
" .09
46,750
48.350
51.750
55.150
58,550
61,050
62,650
64,250
65,850
67,450
" .10
47,550
49,150
52,750
56,350
59,950
62.550
64,150
65,750
67,350
68,950
.001 Phos. =
SO Ibs.
80 Ibs.
100 Ibs.
120 Ibs. 140 Ibs.
150 Ibs.
150 Ibs.
150 Ibs.
150 Ibs.
150 Ibs
*See Trans. Am. Inst. Mm. Engrs., xiv. p. 127, and also Howe's work aa
quoted above.
f Trans. Am. InsL Mm. Engrs., vol. xxm. p. 114.
J Campbell gives, iu his work on Structural Steel, p. 306.
Ultimate strength of acid steel = 33,000 + 1485 (7+ 1260 P,
" basic " =40,000 f 1085 C+ 1200 P.
\vher000 Ibs. per square inch,
with a very low elastic limit, of some 8000 Ibs. When rolled, or drawn into
wire, its strength may be raised to 50,000 or 60,000 Ibs. per square inch,
depending on the amount of work done upon it. It is then "hard-rolled"
or " hard-drawn," and it has very little ductility. Its elastic limit is then
very nearly equal to its ultimate strength. By heating it a bright cherry-
red and cooling it either slowly or quickly, it becomes softened again, or
annealed.
137. Zinc, which is commonly called "spelter" when cast, is a hard,
brittle, white metal, with a highly crystalline fracture. It becomes malleable
and ductile at about 200 to 300 F., but is brittle again at higher temper-
tures. Its specific gravity is 6.9 cast and 7.1 rolled. It melts at 800 F.,
and volatilizes at about 1900 F. It rapidly oxidizes in air at a red heat,
and at a bright red heat, at which copper melts, zinc distils. It is mostly
used as an alloy in brass, German silver, etc., and as a coating to iron and
steel sheets and wire, which process is called galvanizing. It is a common
* The Lake Superior coppers are among the purest in the world.
172
N THE MINOR METALS AND THEIR ALLOTS. 173
electropositive element in electric batteries. Its common impurities are
iron, lead, and arsenic.
138. Tin. Tin is a white, lustrous, and extremely malleable metal, as is
evidenced by its form in tin-foil. Its specific gravity is 7.3; it melts at 450
F., but does not readily volatilize. Commercial tin contains various portions
of many elements such as lead, iron, copper, arsenic, antimony, bismuth,
tungsten, and sometimes manganese and zinc. It is used for coating iron
plates, and to alloy with copper and zinc: Its low melting-point causes it
to be used for safety-plugs in boilers, as its melting-point corresponds to a
steam-pressure of about 400 Ibs, per square inch above atmospheric.
139. Aluminum is a white, soft, malleable metal of extreme lightness, its
specific gravity being only 2.56 when cast and 2.75 when rolled. It melts
at about 1150 F., but does not volatilize at ordinary melting temperatures.
It is especially free from oxidation and corrosion in air, as neither oxygen,
carbonic acid, carbonic oxide, sulphuric or nitric acid, sea-water, nor sulphu-
retted hydrogen has much effect on it. It is, however, readily dissolved
by hydrochloric acid and by caustic alkalies. Its strength pure, when cast,
is only about 18,000 Ibs. per square inch, with low elastic limits in tension and
compression. When rolled or drawn into wire its strength is raised to from
25,000 to 50,000 Ibs. per square inch with elastic limits of about one half
the ultimate strength. It is seldom used in a pure state because of its
softness, but makes with copper, iron, zinc, and tin remarkably strong and
malleable alloys, which will be discussed as aluminum alloys.
Aluminum may be rolled either hot or cold. It is annealed by bringing
it to a low red heat and cooling slowly. In casting aluminum care must be
taken to provide for the great shrinkage. It is best to cast in hot iron
moulds and to cool from the bottom artificially, keeping melted metal
supplied at the gate to supply the shrinkage. Casting under pressure also
gives good results.
It is difficult to obtain aluminum in a perfectly pure* state, and very
slight amounts of impurities largely affect its properties. The common
impurities are iron and silicon. It is now (189G) supplied regularly by the
Pittsburg Reduction Company, under .a guarantee of 98 % pure at 50 to 55
cents a pound, and will be furnished 99$ and 99.6$ pure at special rates.
THE ALLOYS.
140. Nature of Metallic Alloys. Any permanent mixture of two or more
metals is termed an alloy.* Neither the appearance nor the mechanical
properties of an alloy can be predicated upon those of the constituent metals,
and the surprising character of the results produced by various mixtures has
led to an enormous number ^of specially named products, each possessing
certain desirable qualities, the ingredients usually being, for a time at least,
* When mercury is one of the constituent metals the product is termed 'an amalgam.
174 THE MATERIALS OF CONSTRUCTION.
trade secrets. Between 1875 and 1880 the U. S. Test Board made so
thorough an examination of all possible mixtures of the more usual ingredi-
ents found in alloys (copper, zinc, and tin), that the proprietary or trade
names formerly used exclusively for these products are now giving place to
stated percentages of the constituent metals.
In a general way, mixtures composed almost exclusively of copper and
zinc are termed brass, while those composed mostly of copper and tin are
called bronze, while compositions of all three of these elements are called
composition metal, or perhaps also bronze. All these terms are used loosely,
however.
An alloy, though ever so uniformly mixed when in a melted state, is
usually a conglomerate mixture, after cooling, analogous with granite.
Some pure chemical unions are formed, and certain substances may crystallize
out, leaving the more fusible solution or mechanical mixture to form the
matrix for the entire mass when cold. In most cases there is a decided
tendency for the metals to separate before cooling, especially when they are
of different specific gravities, this separating action being called liquation.
To prevent this the mixture is stirred vigorously just before pouring, which
is done at as low a temperature as possible. The quicker the metal cools in
the mould, also, the better, so that it is common to cast alloys in iron moulds
in order to chill the metal, or to cool it suddenly. To obtain constant
mechanical qualities in any given alloy seems to be almost a practical
impossibility. To secure even approximately uniform results requires more
care and expert superintendence than the manipulation of any single metal.
The greater the number of the constituent metals, also, the greater are the
difficulties encountered. Manufacturers should be slow, therefore, to con-
tract for alloys having definite mechanical qualities of a high order, if they
have not had a considerable experience in meeting with similar demands.*
Almost as much seems to depend on the manipulation as on the metals and
proportions employed; but this subject is too large to be entered upon here.f
141. The Copper, Zinc, Tin Alloys. In a general way, some alloy of two
or moflre of these metals, copper being always one, is used for all purposes
where strength, hardness, or malleability is desired in a non-corrosive metal.
In other words, zinc and tin, one or both, are added to copper to harden and
strengthen it. Formerly an alloy was used also for large cast guns (then
called gun-metal), but these are now made of hollow-forged steel. As shown
by Fig. 76, the valuable alloys are those in which copper forms the control-
ling element. This diagram is based on that principle in geometry which
makes the sum of the normals from any point on the interior of an equi-
lateral triangle equal to the altitude of the triangle. If the three altitudes
be each taken as a scale of equal parts on which are indicated proportions (per-
centages) of copper, zinc, and tin respectively, these ranging from zero to
*The author has known of many failures of contractors in this field,
t See Mixed Metals, by Prof. Hiorns, 1890, Macmillan & Co.
THE MINOR METALS AND THEIR ALLOYS.
175
100, then to the same scale the sum of the three normals from any point in
the triangle will be 100, and hence these three normals may be nsed to indi-
cate the percentages of the three metals which unite to form that alloy which
is represented by that point in the triangle.* An alloy of any two of these
finds its place along one side of the triangle, of which the three apices make
the 100-per-cent ends of the three metal scales. A little study of Fig. 76
will make this clear.
The contour-lines on this figure were drawn by the author after plotting
on this triangle the tensile strengths of cast bronze of 'known composition
77/V Z//VC
FIG. 76. Showing the Tensile Strength of the Cast Copper-Tin-Zinc Alloys of all
Possible Mixtures. (Plotted by the author from the results of tensile tests reported
by the U. S. Test Board in 1881.)
from all reliable sources. Dr. Thurston's chart, after which this is modelled,
was made from torsion tests by using a constant factor to reduce to equiva-
lent tensile strength. The author finds the tension tests themselves* do not
agree very well with the values given on that chart, and hence he has drawn
* This method of representing these triple alloys was first used by Dr. R. H. Thurston,
Trans. Am. Soc. C. E., 1881.
176 THE MATERIALS OF CONSTRUCTION.
a chart from the tension tests themselves. It must be understood, however,
that, as stated in the previous article, so much depends on the purity of the
ingredients, and on the manipulation of the process of melting and casting,
that this chart, or any similar record, must be taken as showing what may be
obtained rather than as what will be obtained from the use of these particular
mixtures.
THE BRASSES.
142. The Brasses Copper and Zinc. The most valuable brass alloys con-
tain from 65 to 85 per cent of copper and 35 to 15 per cent of zinc (3 to 5
of copper to 1 of zinc). These mixtures are all strong and ductile, not too
hard to be readily worked in the lathe (a little tin, say 2 per cent, helps it
for this purpose), are readily rolled into plates and drawn into wire, and
under various names are the brasses of commerce. The French standard
mixture for sheet brass is 67 Cu to 33 Zn (2 Cu to 1 Zn), and care is taken
to use only the purest metal for this purpose, as a very slight amount of iron
or silicon (or lead in case of wire-drawing) greatly lessens its ductility.
Rolled or hammered brass is annealed by heating to a cherry-red and cooling
either slowly or rapidly.
Muntz-metal and Sterro-metal are used for ship-coverings in place of
copper. The former contains 3.8 per cent of zinc, the large proportion of
zinc producing a corroded surface which prevents the attachment of
barnacles. The latter contains, in addition, 1.5 to 2 percent of iron, which
greatly strengthens it. It is also used for hydraulic cylinders carrying
very great pressures.
Brass Castings should contain some tin when used for bearings, as this
increases the hardness. Two or three per cent is sufficient. One or two
per cent of lead increases its adaptation to turning, filing, and polishing,
while from i to 6 per cent aluminum adds greatly to its strength and duc-
tility.
In all cases, when melting copper, brass, or bronze, great care must be
exercised to keep the air from the metal, in order to prevent oxidation.
This is done by covering the metal, in the crucible, with a thick layer of
powdered charcoal. The copper is first melted alone, in a deoxidized flame,
and then the scrap brass and zinc (previously melted, these fusing at a much
lower temperature) are added and the whole stirred vigorously to effect a
thorough mixing. Sometimes this mixing is done after the crucible is
removed from the furnace. If it is done in the furnace, the dampers should
be nearly closed to prevent an excessive heat, which would vaporize the zinc.
A new brass-melting furnace is shown in Fig. 77,* in which crucibles are
not used. The metal is charged at the upper door upon a sloping hearth
from which it falls, when melted, upon the hearth proper, from which it is
drawn from the tap-hole as shown. The flame from the adjacent fire can
* Designed and built by J. W. Bennett & Co., Pittsburg. From Engr. News, Oct. 1,
1896.
THE MINOR METALS AND THEIR ALLOTS.
177
be turned into either of these chambers by the opening of suitable dampers,
or into both at once, or stopped off entirely by the damper at the top of the
flue. A furnace Jiaving 600 to 800 Ibs. bosh capacity (5000 to 6000 Ibs.
per day) occupies a space of only 30 to 40 sq. ft. By means of the top
damper the character of the flame may be so controlled as to prevent exces-
sive oxidation.
If iron moulds are used, they should be heated and the interior surfaces
FIG. 77. A New Brass- melting Furnace.
coated with a mixture of resin (3 pts.) and lard-oil (1 pt.) to prevent adhe-
sion. In pouring, the metal must be very carefully skimmed. The pattern
should be made to allow a shrinkage of J in. per foot. For common cast-
ings green sand is used, but for fine work the moulds are dried.
143. Delta-metal, which is an improvement on sterro-metal, is a pro-
prietary composition, or brass, placed on the market since 1883 by a Mr.
Alexander Dick (England), who used the Greek form of the initial letter of
his own name to designate his product. His process consists in incorporating
a fixed amount of iron by making first a saturated solution of iron (about 5
per cent) in molten zinc. To prevent all oxidation a little phosphorus is
added to the melted copper. The proportions are varied for different pur-
poses, having from 50 to 65 per cent copper, 50 to 30 per cent zinc, 0.1 to
5 per cent iron, and sometimes 0.1 to 1 per cent tin. This metal is as strong
and ductile as mild steel, having a tensile strength, when rolled and annealed,
of from 60,000 to 80,000 Ibs. per square inch, with elongations in eight
inches of from 40 to 14 per cent, respectively, at these limits.* When cast
* Tests made at Lloyd's Proving-house, as given by Hiovns.
178 THE MATERIALS OF CONSTRUCTION.
in sand its tensile strength is 45,000 Ibs., with an elongation of 10 per cent.
It also resists corrosion perfectly.
144. Tobin Bronze is very similar to sterro-metal and delta-metal, the
iron ingredient being somewhat less. Its composition is approximately 60
per cent copper, 38 per cent zinc, 1 to 2 per cent tin, with small portions
(0.1 to 0.3 per cent) of iron and lead. Its remarkable properties are due to
its rolling and annealing. As placed on the market,* its tensile strength is
from GO, 000 to 80,000 Ibs. per square inch, with an elastic limit of GO per
cent of its ultimate strength, and an elongation of from 25 to 15 per cent in
eight inches at these limits respectively. It may be regarded as having all
the mechanical qualities of structural steel, with the advantage of being non-
corrosive. It can be procured in sheets from ^ inch to 1-g- inches thick,
and in round rods from J inch to 5 inches in diameter. It is readily forged
at a cherry-red heat either by hand or by machinery. It also works well in
the lathe. It seems, therefore, to be a practically perfect non-corrosive,
engineering metal.
THE BRONZES.
145. The Bronzes Copper and Tin. Since tin is added to copper solely
to harden it (it strengthens it very little), the copper-tin bronzes may be re-
garded as a kind of hardened copper. The ancients used this combination
for their cutting-tools, and it is used largely at the present time to produce
a very hard, non-corrosive metal, useful for many engineering purposes.
If more than 25 per cent tin is used, the alloy, though hard, becomes very
weak and brittle. The most common mixture is that of gun-metal, which
consists of 90 per cent copper and 10 per cent tin. If more than 5 per cent
tin is used, the metal loses most of its malleability when cold. With about
20 per cent tin the metal is very hard and sonorous, making it suitable for
bells, gongs, and wind-instruments. The copper- tin alloys are annealed by
sudden cooling, as by quenching in water from a red heat, while by slow
cooling they are hardened. They differ in this respect from nearly all other
metals.
By using 33 per cent tin (2 copper to 1 tin), a beautiful, hard, perfectly
white alloy is produced, called speculum-metal, suitable for polishing for
mirrors, f
146. Phosphor-bronze is a plain copper-tin alloy made by using a little
phosphorus as a deoxidizer. It is also claimed that the phosphorus causes
the tin to form a crystallized compound with the copper. It is mainly, how-
ever, as a cleanser of the melted metal from the oxide of copper that it is
valuable. When used properly it forms a slag, and is skimmed off, and is
not found in the finished product. The phosphorus is added in the form
of phosphor-copper or phosphor-tin, these containing phosphides of copper
* By the Ansonia Brass and Copper Co., New York.
f Lord Ross's great telescopic reflector was made of this alloy.
THE MINOR METALS AND THEIR ALLOTS. 179
or of tin. For a malleable product, to be rolled or drawn into wire, the tin
should not exceed 4 or 5 per cent, and the phosphorus should not exceed
yL of one per cent. For hard castings of great strength, as for pinions,
valves, bearings, or bushings, use 7 to per cent of tin and \ to 1 per cent
of phosphorus. A greater amount of phosphorus, up to 4 per cent, increases
the hardness and brittleness. More than 4 per cent phosphorus will make
the product useless.
147. Silicon-bronze is now used extensively in Europe for electric con-
ductors, as it has 70 per cent of the conductivity of copper, while phosphor-
bronze has but 30 per cent, and steel 10,5 per cent. By using silico-bronze
wires the poles may be put much farther apart than when copper wires are
used.* While the proportion of silicon remaining in the alloy is very small,
it has an excellent cleansing action, like phosphorus, without danger of devel-
oping brittleness. The "dose" of silicon, to be added to the melted copper
or bronze, is prepared by the inventor, Weiller, as follows: " Take potassium
silico-fluoride 450 parts by weight, powdered glass GOO parts, common salt
250 parts, carbonate of soda 75 parts, carbonate of lime GO parts, and dried
chloride of calcium 500 parts. Heat these in a covered plumbago crucible
a little below the temperature where they begin to act on each other, when
the whole is added to the melted copper or bronze, and vigorously stirred."
The resulting slag is skimmed off.
148. Aluminum Bronze has now come to be regarded as one of the most
valuable made. It is composed of from 5 to 12 per cent aluminum with
from 95 to 88 per cent of copper. These alloys have remarkable ductility,
combined with great strength. Thus the 5 to 7| per cent aluminum bronzes
have, when rolled or forged, an ultimate tensile strength of from 70,000 to
80,000 Ibs. per square inch, an elastic limit of over 40,000 Ibs., and an elong-
ation in 8 inches of over 30 per cent. With 10 per cent of aluminum, the
rolled bars have an ultimate tensile strength of 100,000 Ibs. per square inch,
an elastic limit of 60,000 Ibs., and an elongation of 10 per bent in 8 inches.
If further rolled, it hardens and strengthens to 130,000 Ibs. tensile strength
with 5 per cent elongation. The 5 to 7 per cent bronzes can be hammered,
rolled, and forged at a red heat, and are very similar in every way to mild steel.
They are almost absolutely non-corrosive. It hardens by cold working, but
may be annealed by heating to a red heat and quenching in water. It has
a modulus of elasticity of about 18,000,000, which is higher than that of the
alloys of copper, zinc, and tin.
On account of the excessive shrinkage of this alloy in hardening, it is
necessary to provide a large sinking head in casting, and to so locate this as
to supply to the cast form the necessary fluid metal to give a sound casting
as it shrinks away in cooling.
* The Austrian Railway Company puts its poles from 328 to 720 feet apart in open
country when using silico-brouze wires.
180
THE MATERIALS OF CONSTRUCTION.
149. Alloyed (or Hardened) Aluminum. Just as a small percentage of
aluminum added to copper greatly hardens and strengthens it, without de-
stroying its ductility, so a small percentage of copper added to aluminum
works a similar change in this soft metal. If more than 15 or 20 per cent
o'f either be added to 85 to 80 per cent of the other, however, the resulting
mixtures become hard, weak, and brittle, and entirely worthless as commer-
cial products.
Both tin and zinc, up to 15 per cent, are used to harden and strengthen
aluminum, while an alloy of 15 per cent zinc, 3 per cent tin, and S2 per
cent aluminum is especially recommended. There are a number of secret
mixtures of hardened aluminum, some of which are used for casting bicycle-
frames.
150. Aluminum in Steel. If about one pound of aluminum per ton of
steel be added to the heat just before drawing or teeming, it prevents the
formation and escape of gases, and gives solid ingots or castings. For steel
castings two to three pounds per ton is now commonly added to the melted
steel by throwing small pieces into the ladle as the steel is drawn from the
furnace. In both methods it permeates the entire mass without artificial
stirring, as manganese does, and seems to have very much the same effect.
Its effect on cast iron are the same as those of silicon, but as it is much more
expensive it is not used in this way. In steel, however, its use is common,
as nothing seems to take its place. Besides preventing blow-holes it adds
to the ductility of the product.
151. Alloys which Fuse below the Boiling-point. The following remark-
able alloys, all of which fuse at very low temperatures, may be used as safety-
plugs in automatic fire-spraying pipe-systems in mills and for similar pur-
poses.
TABLE XVII. FUSIBLE ALLOYS.
Name.
Percentage of Ingredients.
Fusing
Tempera-
ture.
Bismuth.
Lead.
Tin.
'Cadmium.
Newton's . ....
50
50
50
50
50
31
28
25
24
27
19
22
25
14
13
12
10
95 C.
100 C.
93 C.
66-71 C.
60 C.
Rose's
Darcet's
Wood's
L/ipoirtz's . . .
CHAPTER "XL'
LIME, CEMENT, MORTAR, AND CONCRETE.
LIME AND NATURAL CEMENT.
152. Quick, or Fat, Lime. If carbonate of lime (CaC0 3 ), as found in
ordinary limestone, or marble, or chalk, be heated to a temperature of
about 800 F., when it becomes a cherry-rod, the carbon dioxide (C0 a ) is
driven off, and the oxide of calcium (CaO) remains, and is called quick-
lime. In a pure carbonate of lime 4-1 parts by weight of carbon dioxide
(carbonic acid) are combined with 56 parts by weight of oxide of calcium,
or quicklime. Since the rock will contain some moisture, the amount of
quicklime obtained from burning limestone will never be more than one
half the weight of the stone charged. The calcium oxide, or quicklime,
cannot be decomposed by heat, but it has a very strong affinity for water.
When water is added to it, it rapidly rises in temperature, swells, and falls
into an impalpable powder, and increases its volume to about three times
its initial volume before the water was added. This process is called slack-
ing, and the product is then called hydrated, or fat, lime (calcic hydrate), or
slacked lime, or lime paste or putty when further diluted with water. The
quicklime, or calcium oxide, will slack by absorbing moisture from the
atmosphere, unless kept in closed vessels. It is therefore not kept in stock
for any great length of time, as it becomes bulky and difficult to handle
when slacked. It can be kept indefinite^ without deterioration in the
form of lime paste, or putty, if kept wet so as to exclude the air. Quick-
lime is not found, as such, in nature, since it has a tendency to recombine
with carbonic acid from the atmosphere, and form carbonate of lime.
Rocks composed of nearly pure carbonate of lime are found in all parts of
the world, and they have been used in this way for the manufacture of
-quicklime for mortar from the most ancient times. In slacking, 18 parts
by weight of water unite with 56 parts by weight of quicklime, making 74
parts of calcic hydrate, Ca(OH) 2 . The heat generated in slacking greatly
facilitates the process, and some limes will slack in boiling water whicn
cannot be slacked by the use of cold water. Limes of this latter class are
called " poor," in distinction from those which slack readily, which are
commonly termed "fat."
153. Hardening of Lime-mortar. When quicklime has been slacked
and mixed with sand it forms what is commonly called lime-mortar, which
181
182 THE MATERIALS OF CONSTRUCTION.
is used for laying brick and stone masonry, for plastering houses, and
the like, where the mortar-joints will be exposed to the action of the air
only. Because of the great shrinkage of lime-paste in drying, it cannot be
used neat, but must always be mixed with several times its volume of sand.
"When exposed to atmospheric action, the hydrated lime, Ca(OH) 2 , slowly
unites with carbonic acid (C0 2 ), which is always present in the atmosphere,
thus changing a portion of the hydrated lime back to its original form of
carbonate of lirne, leaving another portion in the hydrated form. Since the
carbonic acid can have access to the lime only by the circulation of air
through it, it follows that this chemical change occurs mostly at the outer
and exposed surfaces of lime-mortar joints, and does not take effect at a
distance from the surface to any appreciable extent, except through the
lapse of long periods of time. In all cases, therefore, where it is neces-
sary for the mortar to harden in a comparatively short time, lime-mortar
must not be used.
154. Hydraulic Lime. When a limestone contains from 10 to 20 per
cent of clayey matter, new combinations of lime and the silica in the clay
are formed in the furnace, if the temperature is sufficiently high, which
causes the product to slack less readily, and with a much less increase of
volume, than in the case of quicklime. Hydraulic lime is partially slacked
on drawing from the kiln by adding from 15 to 20 per cent of its weight of
water, and it is then thrown into large heaps. The steam thus formed
causes it to slack in the course of a week, after which it is screened and
packed for market. It cannot be kept in the form of paste, as fat lime
always is, as it would harden, like cement. If this same rock be calcined at
a high heat and reduced to a clinker but not fused, and then ground with-
out slacking, it forms the natural cement described in the next article. It
is changed from the one product to the other by the chemical reactions
which occur at the higher temperature in the kiln. Mortar made with this
lime will harden somewhat under water, by a process of partial crystalliza-
tion, and hence it is called hydraulic lime. Limestones having a composi-
tion suitable to make hydraulic lime are very common in England and
Europe, but are not common in America; hence what is there known as
hydraulic lime is not known in America as an article of commerce.
155. Natural Cement. Carbonate and magnesian limestone rocks con-
taining from 20 to 40 per cent of clay, when calcined to a clinker, just
short of fusion, and finely ground, give a product which sets or hardens
quickly on the addition of about 25 per cent of its weight of water, without
any increase of volume, and forms a permanent artificial stone which in-
creases in strength and hardness for many years. This product is known
as natural cement* because it is produced wholly from a natural rock. It
* In England and on the Continent this kind ef cement is commonly called Roman
cement, from a supposed similarity to the cement the Romans used on their hydraulic
LIME, CEMENT, MORTAR, AND CONCRETE. 183
has become customary to give to natural cements local geographical names,
indicating the place of their manufacture. This is more especially appro-
priate since the natural cements made in a given locality will have the
same general characteristics, because they are all made from the same sedi-
mentary rock. These cements are very largely used in America, some of the
principal varieties being the " Rosendale " cement, made near the Hudson
River in Ulster County, N. Y., the " Utica " cement, made at Utica, 111.,
the "Louisville" cement, made mostly on the Indiana side of the Ohio
River in the vicinity of Louisville, Ky., and the " Milwaukee" cement,
made at Milwaukee, Wis. Such cements are made at various other places
in the United States and Canada, and are known by their corresponding
local geographical names. These cements are now very cheap, and often-
times are found to vary greatly in quality. While the better grades of
natural cement are quite sufficient in strength for nearly all kinds of
engineering works, the want of uniformity in their hardening properties
is a serious objection to their use.
Some of the American natural cements are very quick setting, which is
a further objection to them, since it is difficult to use the mortar or concrete
made from them before it begins to set, or harden.
The old Roman cement used by the Romans in their hydraulic masonry
constructions was made by mixing volcanic ashes with lime in proper pro-
portions.
PORTLAND CEMENT.
156. Historical. An artificial mixture of lime and clay in proper propor-
tions, calcined to a clinker at a temperature of incipient fusion, and finely
ground, is called Portland cement. It received this name in 1824 in Eng-
land, where it was first made, from its similarity in appearance when
hardened to the noted oolitic limestone from the " Isle of Portland " * long
used in England for building purposes. It was patented in that year by
Mr. Joseph Aspdin, a Leeds brickmaker, as an "artificial stone." He
mixed pulverized limestone, taken from the public macadamized roads,
with clay, by adding water enough to reduce it to a liquid form. This was
then dried and burned " in a furnace similar' to a lime-kiln till the car-
bonic acid is entirely expelled." The necessity^! burning to a clinker
was not given in the specification, and was pro&ably not known at that
time, neither was the proper proportion of clay mentioned. His success was
therefore something of an accident, as was doubtless the discovery of the
engineering works. There are few suitable rocks in Europe for making this cement. It
is extremely irregular in composition, and not to be compared with the very uniform
beds found in inexhaustible quantities in the United States. If such natural-cement
rocks as we have, had been common in England and on the Continent, it is almost certain
that the artificial Portland cement would never have been discovered.
* This is really a peninsula on the south coast of England, in Dorset, near Weymoutb,
noted for its building-stone. The Westminster cathedral is built of this stone.
184 THE MATERIALS OF CONSTRUCTION.
hydraulic property of the mixture itself. Aspdin began manufacturing his
cement at Wakefield * in 1825.
Previous to his time a kind of natural cement had become common
under the general name of "Roman cement." This was made by calcining
nodules (geodes) of a clayey limestone found along the seacoast, " at a heat
nearly sufficient to vitrify them," and grinding the product. (Patented by
James Parker in England in 1796.)
The discovery that the hydraulic property of certain limes was due to
the clay ingredient is due to Smeaton (about 1756), who had some knowl-
edge of chemistry. f The occasion of these investigations was the building
of the first Eddystone lighthouse. This, therefore, marks the beginning
of all intelligent study of the subject of hydraulic cements. J
Although Aspdin began manufacturing Portland cement in the north
of England in 1825 (and continued to 1853), and it was introduced exten-
sively on the Continent, it was not known in London till made by J. M.
Maude and Son (with Aspdin's son) in 1843 under Aspdin's patents in what
is now a part of London, and by J. B. White and Sons, in Kent, in 1845.
In tests made in 1843 for the new Houses of Parliament this Portland
-cement was shown to be superior to the Eoman cement then in common
use, but engineers and architects were slow to grant the fact. Public com-
petitive tests between the above-named firms were conducted in 1848
which further proved the superiority of the Portland cement, | and after
the Exhibition in 1851, at which many tests were made, its use soon became
general in England.
Many failures marked the first thirty years of the Portland-cement
manufacture, from an entire neglect of the chemical analysis of the ingredi-
ents. Reliance was placed solely on the empirical knowledge of workmen
ignorant of chemical science, and much sophistry and deception were used
to cover up their failures. It is now known that good Portland cement
* A small city in Yorkshire near Leeds.
f A report of his investigations and conclusions was not published till 1791, in Book
IV of his Narrative of the Building, etc., of the Eddystone Lighthouse.
^ For a very good account of the early history of > this subject see Redgrave's Cal-
careous Cements, London, 1895.
This was a Roman cement factory, but Mr. I. C. Johnson, their manager, after
long search and experimentation, the Aspdin processes being secret and purposely mys-
tified, discovered the secret of burning to a clinker. He also at last discovered the
proper proportions. From an account by Mr. Johnson himself in The Building News
(London), 1880.
|| These tests consisted in building out brick beams from solid walls, and in crushing-
tests of large cement prisms. As late as 1845-6 Sir Robert Peel announced in Parlia-
ment his intention of taxing the use of the clay nodules of which the Roman cement was
then made, to prevent their complete exhaustion, and to retain sufficient of them for
government works. Aspdin thereupon addressed him a personal note describing his
artificial cement, and the proposed measure was dropped.
LIME, CEMENT, MORTAR, AND CONCRETE. 185
can be made anywhere by properly combining, burning, and grinding a
mixture of carbonate of lime and a suitable clay, the only elements of
commercial success being economy and scientific direction.
Since a good Portland cement, with or without sand, gravel, and broken
stone, makes an artificial compound equal to almost any natural stone in
hardness, strength, and durability, and since it can be moulded to any form
and is much cheaper than quarried and cut stone, it is constantly finding
wider and wider fields of application. This material has already worked a
revolution in engineering construction nearly equal in significance to that
following upon the general use of the Bessemer and open-hearth processes
of making steel. The character of Portland cement also has constantly
improved, until now it has reached practical perfection. Within the past
ten years the improvement has been very marked, as a result of the universal
system of testing now in vogue, and of the general employment of compe-
tent scientific supervision of the works, made necessary by these tests on
the part of the user. Portland cement is now made on a gigantic scale in
Germany, Belgium, France, and England, and its manufacture is rapidly
increasing in the United States.
157. The Ingredients of Portland Cement. All mixtures, natural or
artificial, of carbonate of lime (CaC0 3 ) and clay in the proportions of from
72 to 77 per cent of the" former to 20 to 25 per cent of the latter will, when
calcined at the proper temperature, produce a Portland cement of fair
quality. After calcining, and driving the carbonic acid (C0 2 ) from the
carbonatelSf lime (CaC0 3 ), the proportions of lime (CaO) and clay (silicate
of alumina (A1 3 8 , 2Si0 2 , 2H 2 0) are about 60 to 65 per cent of lime and
fr0in 25 to 30 per cent of clay, with some 5 per cent of other ingredients,
such as sulphate of lime, magnesia, iron oxides, etc. "A variation in the
lime ingredient of one per cent above the true amount will give a cement
liable to crack on long exposure to water, and a deficiency of one per cent
of lime will reduce the strength of the cement and also make the mixture
liable to fuse in the kiln."* The most competent chemical supervision and
continual analyses of the ingredients are therefore necessary to secure the
best results.
158. Chemical Characteristics of the Ingredients. The carbonate of
lime should be nearly free from all other substances except clay (silica
and alumina). While magnesia and iron in small amounts are not injuri-
ous, they are probably inert, and the sulphur compounds are a positive
injury, above a two or three per cent limit.
159. The Clay. " The best clays for the cement-manufacturer are" those
having a greasy, unctuous, feeling, quite smooth to the touch. As a rule,
clays which stain the fingers should be avoided, as being either too much
* Prof. Spencer B. Newberry in the Engineering Magazine, Juue, 1894. Mr. New-
berry is chemist and manager of the Sandusky, O., Portland Cement Works. This
Statement is probably a little too strong.
186 THE MATERIALS OF CONSTRUCTION.
impregnated with iron compounds, or containing a large proportion of
organic or other impurities. This does not hold good in the case of the
carboniferous shales, some of which are rich in matters which assist in the
calcination of the cement. Shales which contain much alum, selenite, or
iron pyrites, and many of the shales having a high percentage of carbonate
of lime, need great care in manipulation, as they are apt to fluctuate widely
in composition and to lead to mistakes in the proportions of the ingredients.
Some clays contain a high percentage of sandy particles, or of nearly pure
silica not in combination with lime, iron, or alumina, and these clays, though
useful to the brick-maker, are ill adapted for cement-making. They are
generally characterized by. a harsh gritty touch when tested between the
finger and thumb, and it is possible to wash out a considerable percentage
of sandy particles."*
160. Silica and its Compounds. " It will be necessary, in order to
understand the chemistry of cements, to treat in some detail of silica and
its compounds. Silica, the oxide of the element silicon, is found very
widely distributed in nature, sometimes pure, but more often in combina-
tion with other substances, as it has a great tendency to forrr complex
salts, known as silicates. It plays the part of an acid, and combines with
lime, alumina, iron, and the alkalies in a vast number of different propor-
tions. It is found that 28 parts by ^weight of silicon and 32 parts by weight
of oxygen are present in silicic anhydride or silica, having the chemical
formula Si0 2 . Clay, a hydrous silicate of alumina, may oe taken as a type
of the. silica compounds, while quartz, flint, and chalcedony consist of
almost pure silica. Porcelain clay, which contains about 47 per cent of
silica, 39.2 per cent of alumina (A1 2 3 ), and 13.7 per cent of water, and
corresponds to the chemical formula Al 2 3 2Si0 3 + 2II 2 0, or clay proper,
with a molecular weight of 258.4, may represent th,e silicates. There are,
however, an enormous number of clays in which silica and alumina are
present in very varying proportions, and which contain in addition iron,
alkaline matters, lime, etc. For certain of these clays it becomes almost
impossible to propound any reliable chemical formula to express their
composition; and alumina, while it may combine in certain definite pro-
portions with the silica as a base, is also capable of acting as an acid, and of
combining with lime and the alkalies, especially at high temperatures, to
form certain more or less unstable and little known compounds termed
aluminates." *
161. " Alumina is the oxide of the metal aluminum which has the atomic
weight of 27.2, and two parts of aluminum combine with three parts of oxy-
gen, equal 48, to form its only known oxide, termed alumina, amounting in
all to 102.4. It will not be necessary to study in detail the combinations of
silica and alumina with iron and the alkalies soda and potash though
these compounds play a very important part in cement action." *
* Redgrave.
LIME, CEMENT, MORTAR, AND CONCRETE.
187
162. Sulphur and its Compounds.
Portland cement may be associated
with a small percentage of gypsum
or sulphate of lime (calcic sulphate
CaS0 4 , 2H 2 0), the water of which is
driven off in the calcining process,
reducing this compound to what is
commonly known as plaster of paris
(GuS0 4 ). Sulphur may also be intro-
duced in the fuel used for burning,
or from the clay which sometimes
contains iron pyrites. The sulphuric
acid relieved from these compounds
ma unite with the free lime in
The carbonate of lime used for making
5
F.G. 78. Effect of Plaster of Paris on
Time of Setting of Cement. (Wheeler,
Rep. Chf. Engrs. 1895, p. 2938.)
/.a 2.0 3.0
FIG. 79. Showing the Effect of Plaster of
Paris 011 the Strength of Portland-cement
Mortar, 1 C. : 3 S. (Tetmajer, vol. vn. p
39.)
furnace and form an additional portion of calcic sulphate. The effect of
this calcic sulphate or plaster of paris in quantities not exceeding two or
three per cent is to greatly delay the time of setting (Fig. 78), but to
increase slightly the final strength of the cement (Fig. 79). When present
in quantities exceeding four or five per cent both these effects are lost, and
it is also considered injurious in other ways, since it is comparatively sol-
uble in water, and when present in the kiln in considerable quantity it
leads to the formation of calcic sulphide, which decomposes the iron com-
pounds in the cement, thus leading to disintegration. The German standard
rules allow a proportion of calcic sulphate not to exceed two per cent, but an
effort has recently been made to have this limit raised to three per cent.
163. The Chemical Reactions Produced in Calcining. It has been com-
monly agreed by chemists that the following combinations are effected by cal-
cining an intimate mixture of lime and clay to the point of incipient fusion :
188
THE MATERIALS OF CONSTRUCTION.
Silicate of lime (SiO Q , 3CaO)
j Silii
|Lirr
Aluminate of lime (Al a O a , 3CaO) j !
Proportions by
Weight.
Silica 23
Lime 43
Alumina 17
jme 28
Silica 15
Double silicate of lime and alumina (Si0 3 (Al 2 3 -fCaO) 3 ) J Alumina 51
Lime 28
Magnesia probably remains inert, and does not combine with alumina
and silica. It is harmless if not forming over five or six per cent of the
whole.* Oxide of iron is also a useless ingredient.
M. Le Chatelier has been able to identify the aluminate of lime by a
microscope, with polarized light, in both cement clinker and in an artificial
synthetic compound. He also thought he determined two other substances
in this manner, they being a silicate of lime (2CaOSi0 2 ) found crystallized
4
out
COMP/JESS/ON
/ N
DAYS
6000
/40 28O
FIG. 80. Showing the Inferior Character of the Furnace-dust compared with the
Ground Clinker, when used in Mortar, 1C.: 3S. (Tetmajer, vol. vn. p. 12.)
in a matrix of an alumino-ferrite of lime, with a formula 2(AlFe) 2 3 3CaO.
These chemical reactions in the furnace are, however, not yet known with
certainty.
Since there is no further mixing of the lime and clay ingredients in the
furnace in the calcining action, it is absolutely necessary, in order to secure
* The German Cement Manufacturers' Association has allowed five per 'cent since
1893.
LIME, CEMENT, MORTAR, AND CONCRETE. 189
perfect results, to have the lime and the clay perfectly and uniformly
mixed before going into the furnace; that is to say, each particle of lime
should have adjacent to it its particle of clay with which to unite when the
proper temperature has been attained. Since it is, of course, impossible to
intermix these materials to this degree of perfection, there must of necessity
result from the burning more or less inert or un combined clay and lime
without cementing qualities, which inert matter forms a large part of the
furnace-dust. (See Fig. 80.) If the ingredients were actually fused or melted
into a liquid mass, and the chemical action were to take place after the
ingredients were in the liquid form, a much more perfect union of the ele-
ments would of course be effected. In the formation of the clinker which
is ground into Portland cement, however, the ingredients are not fused,
since fusion would be fatal, and hence the elements of the mixture are
incapable of uniting except they be in immediate juxtaposition. The
further improvement of Portland cement evidently lies in the direction of
more perfect and more uniform mixture of the raw materials in a finely
divided state before they are burned. From experimental tests which
have been made in this direction, it would seem that the strength of
Portland cement might be made at least twice what it is now, by more
perfectly satisfying this requirement.
164. The Chemical and Physical Changes involved in Setting and
Hardening. By the setting of cement is meant its initial change from a soft
or plastic mortar to a friable solid. This change is usually effected with
great suddenness, after it begins, as shown by the curves in Fig. 333, and
it has been shown to be always accompanied by the evolution of heat.
After the cement has become thoroughly set it still is very weak, and is
readily pulverized in the fingers. If left undisturbed, however, it increases
in hardness and strength, sometimes for several months, but generally for
many years. There is no relation between the time elapsing after wetting
before setting takes place, and the period of time required to attain to
nearly its ultimate strength. The setting of cement is thought to be effected
by the crystallizing out of the silicate and the aluminate of lime, which
are soluble in water in their anhydrous form. After dissolving in the
water they pass to the hydrated state in which they are insoluble, and hence
are precipitated in a crystalline form, with a development of heat. This
process is greatly hastened at higher temperatures.
The hardening of cement is due to a continued crystallization of salts
from solution, and to further chemical and physical changes which develop
slowly, but which continue for long periods of time. M. Fremy regards the
aluminate of lime as the chief source of the hardening property, and he also
thinks the silica and the alumina of the clay are separated by calcining and
take on allotropic forms, ready to unite into new compounds with the quick-
lime when water is added. There are so many kinds of combinations of
various substances which will serve to produce the final characteristics of
190
THE MATERIALS OF CONSTRUCTION.
hardened Portland cement, that there must be many different chemical
compounds which, after calcining, will harden on the addition of water.
The problem is so complicated
^.. ^ ,, ^ . --^Jl-- .^. . that ^ nas as T?e ^ Defied a complete
chemical analysis.
165. Slag-cements. Many va-
rieties of iron blast-furnace slags
will make an excellent cement
when ground with hydrated or
slacked lime, without further cal-
cining. The slag is " granulated "
by running it from the blast-fur-
nace into water, where it forms into
a brittle, porous, pumice-like mass
resembling caked sand, and in this
condition it is called " slag-sand/'
It is now easily crushed into
powder, but retains the water in
its meshes so that it is very diffi-
cult to dry it. Sometimes this
" slag-sand " is calcined at a low
heat simply to dry it.
This method of suddenly cool-
ing the melted vitreous slag in
water has no effect upon it further
than to simply reduce it to the
porous, friable condition, since
when allowed to cool in the ordi-
nary way, into a solid mass, and
then crushed to powder and lime
added, it has no hydraulic proper-
ties (see Fig. 81). It seems prob-
able, therefore, that the sudden
cooling leaves the chemical com-
pounds in a mere unstable con-
dition, so that when powdered
and intimately mixed with hy-
drated lime, and water added,
they are ready to enter into new
chemical combinations with the lime. To three parts by weight of the
dry "slag-sand" is added one part of hydrated lime (CaII 2 2 ), and these
are thoroughly ground together and intermixed by suitable mechanical
appliances. This cement does not deteriorate appreciably by lapse of
time.
a-
\
LIME, CEMENT, MORTAR, AND CONCRETE. 191
The burdening properties of slag-cement, with its small proportion of
lime, and its being an artificial mixture of free lime and slag without subse-
quent calcining, have greatly disturbed the previously accepted theories of
the chemical transformations in the furnace and after melting, in the case
of Portland cements, and it seems that now the whole matter will have to
await a further progress of chemical knowledge in this direction.
Slag-cements make a more unctuous mortar and are much liked by
architects for laying brick Avails and piers, and for making floors, sidewalks,
etc. It is slow setting and does not stain the masonry in outside walls. It
is often mixed with additional amounts of lime-putty to further delay the
time of setting, or to cheapen the mortar, or to make it work smoother
under the trowel, or perhaps for all these reasons combined.
166. Sources of the Raw Materials Used in Making Portland Cement.*
" Portland cement is made from carbonate of lime and clay. These materials
may be naturally mixed, as in the case of argillaceous limestones, or entirely
separate. In all cases, however, it is necessary to bring the material to cor-
rect composition by artificial additions and thorough mixing. In England
chalk is the form of carbonate of lime employed. In Germany the chief
material is marl (mergel), by which is understood a more or less hard lime-
stone rock containing clay. In some German factories a pure soft marl
(weisenkalk), or fresh-water chalk, is used, consisting chiefly of carbonate
of lime and similar to the marl deposits of this country.
" In the United States the materials used are very similar to those of
Germany. Most of our clay limestones are highly magnesian, and there-
fore unsuitable for Portland cement, though they are used on an immense
scale for natural-rock cements. At certain localities, however, as in Lehigh
County, Pa., at Phillipsburg, N. J., and in the far West, limestones con-
taining sufficient clay and nearly free from magnesia are abundantly found,
and in the above localities and from* this material most of our Portland
cement is made. In the Lehigh County region, the chief seat of the
American Portland-cement industry, the different strata of rock are care-
fully selected and mixed in such proportions as to give a material of the
right composition.
" In central New York and at a few points in Ohio and Indiana large
deposits of pure white marl are found. This is generally called ' shell-
marl/ and was supposed to result from the disintegration of fresh-water
shells. In the opinion of the writer, however, these marl-beds are generally
pulverulent deposits from calcareous springs, and are not formed from
shells to any great extent. At the localities above mentioned this material,
artificially mixed with clay, is largely used for the manufacture of Portland
cement. Owing to the soft, fine-grained character of the marl, the mixing
can be much more cheaply done than in the case of limestone, though this
* This and the following article are taken from the paper on Portland Cement by
Prof. Spencer B. Newberry, in the U. S. Geol. Surv. Report for 1894, Part IV, p. 581.
192
THE MATERIALS OF CONSTRUCTION.
advantage is largely compensated for by the necessity of drying out the 40
to 50 per cent of water which the marl generally contains. It must be
remembered also that in the argillaceous limestones the ingredients are
already uniformly mixed in nearly the proper proportions, while with the
pure lime and clay this mixing must be wholly effected by artificial means.
The leaving of any free lime in the final product, from imperfect mixing,
has often led to the disintegration of the mortar by sea-water, and by fresh
water containing carbonic acid in solution.
"As already stated, most American Portland cement is made from argil-
laceous limestone, as shown by the following table.
NUMBER OF AMERICAN" CEMENT FACTORIES USING LIMESTONE COMPARED
WITH THE USERS OF MARL (1894).
Factories Using
Number.
Quantity.
Lim6stoii6
17
Barrels.
611 829
Marl
7
186 928
Total
24
798 757
"The first group includes G factories in the Lehigh County region in
Pennsylvania, producing over 400,000 barrels; 1 at Phillipsburg, N. J. ; and
10 at other points. The second group, using marl, includes 4 factories in
New York, 2 in Ohio, and 1 in Indiana."
167. Processes Used in Pulverizing and Mixing the Raw Materials.
There are, in general, three processes employed in preparing this intimate
mixture of the raw materials, which may be designated " The Wet Process/*
" The Semi-wet Process," and " The Dry Process."
1. The Wet Process was originally employed in England and in France,
and was used for the admixture of crushed chalk and clay. These were
pulverized and mixed in " wash-mills " with such an excess of water as to
form a thin liquid. This was stirred by such an arrangement as that shown
in Fig. 82, the escape being at the top over a lip or weir. The coarsest
particles settled in this wash-mill, and such granulated matter as escaped
in the liquid was intercepted on its way to the " backs," which were open
tanks some four feet deep, with earth or gravel bottoms. The mixture was
now allowed to settle for some days, when the clear water was siphoned off
and the " slurry" left to dry in the open air until it could be handled with
a shovel. It was then wheeled upon drying-floors and dried by artificial
heat into irregular clods or masses, when it was sent to the furnace. Even
in summer this process required many weeks' time for its completion, and
in the first settlement the chalk and clay ingredients would sometimes have
a different specific gravity, and hence they would not settle simultaneously.
This would give an uneven mixture, which would be so far fatal, since it
LIME, CEMENT, MORTAR, AND CONCRETE.
193
could not be corrected. This process is going out of use even for such
material as is suited for this method of treatment.
2. The Semi-wet Process consists in mixing the ingredients in the state
of a soft paste. This may be done either by grinding them together in this
condition or by means of "edge-runners." These consist of heavy cast-iron
cylinders of short length, mounted on a horizontal axle which is made to-
FIG. 8 - 3. Wash- mill used in the Wet Process of making Portland Cement.
swing about a vertical axis, thus causing the heavy cylinder to roll about on
a bed-plate. Sometimes the roller-axle maintains a fixed position and the
plate revolves on which it rests. This is an efficient pulverizer and mixer
when the ingredients are comparatively soft. It does not produce as
uniform a pulverization, however, as a grinding-mill. Both these processes
are used in America.*
3. The Dry Process is used in Germany, and in Pennsylvania and New
Jersey in this country, where the materials consist of argillaceous lime-
stone having nearly the proper composition for making Portland cement.
The rock is first crushed and then ground. The final mixture of limestone
and clay-shale is made before the material is ground, so that the process of
grinding effects a very thorough mixing. The ingredients must be reduced
to an impalpable powder in order to make possible that thorough mixing
necessary to enable each molecule of lime to associate itself with its mole-
cule of clay in the calcining process, so as to produce the true chomical
combinations in the clinker. If the limestone used has primarily nearly
the composition required, which is sometimes the case in America, then it
is evident that any want of perfection in the first grinding and mixing of
the raw materials is not so injurious, since the native mixture is not only
* At Bellefontaine, O., the "edge-runners" are used, and at Sandusky, O., the
grinding-mill, the material in both cases being a soft marl and clay.
194 THE MATERIALS OF CONSTRUCTION.
nearly correct as to proportions, but so far as it goes the admixture is prac-
tically perfect. When pure carbonate of lime is used (as in the case of soft
marl) with clay, there is no primary mixture of the lime and clay at all, and
hence the necessity of a much more elaborate artificial mixing process than
when these ingredients are found intimately associated in a natural rock
and to nearly the correct proportions. On the other hand, the soft marl
and clay (of Ohio, for instance) are much more easily worked than the hard
limestone and clay shales (of Pennsylvania). In order to enable the hard
materials to compete successfully with the soft, it is necessary that the
limestone should contain primarily nearly the proper proportion of clay.
After the dry grinding and mixing of the raw materials, the dust is wet
sufficiently and moulded into bricks (thus obtaining a further mixing), and
then dried and burned. The raw powder cannot be calcined in the ordinary
furnaces without first compacting it in aggregate forms to allow of a draft
of air through them. In the tubular rotating furnaces it is calcined as a
dry powder, and this is one of the great advantages of that process.
168. Processes Used in Burning Portland Cement. "There are three
distinct forms of kiln used in burning Portland cement in America. These
are (1) intermittent or dome kiln, (2) continuous kiln, of the Dietzsch or
Shofer type, (3) rotary furnace. In the old-fashioned intermittent kiln the
bricks of cement mixture are charged into the kiln with coke in alternate
layers, and the whole allowed to burn out and cool down before emptying.
The Dietzsch or Shofer continuous kiln is continuously charged with bricks
of cement mixture and soft coal, and the burned clinker periodically with-
drawn at the bottom. It presents the great advantage of cheaper fuel and
economy of labor, and burns the dry powdered material. The rotary fur-
nace consists of a rotary cylinder heated by a blast of air and gaseous fuel,
the material being continuously run in at one end, and issuing as burned
clinker at the other. This process was patented by Mr. Frederick Ransome,
in England, in 1885, and has been subsequently modified and improved by
others. Many difficulties have been met with in carrying out this plan, but
it is now successfully operated at a number of works in this country. It
would seem to be the most rational method of carrying on the burning of
cement, since it effects an enormous saving in time and labor, and allows
the temperature to be regulated far more exactly than is possible in the
older processes. Crude or fuel oil is used as a source of heat at all points in
America where this kiln is employed, though producer-gas -with or without
regenerative furnaces might be employed.*
" In the United States most of the Portland cement produced is burned in
the old-fashioned intermittent kilns. The Dietzsch kiln is used at Harper
and Middle Branch, Ohio. The Shofer kiln is to be used at new works now
* This process requires a much greater fuel expense than the kilns and seems to be
used only where the raw material will not adhere sufficiently by wetting to form
briquettes which can be burned in kilns. J. B. J.
LIME, CEMENT, MORTAR, AND CONCRETE.
195
beginning operations at Glens Falls, N. Y. The rotary furnace is in oper-
ation at Colton, Cal., Phillipsburg, N. J., Coplay, Pa., and Sandusky,
Ohio. The following table shows the number of barrels of cement made
during 1894 and 1895 in vertical kilns (continuous and intermittent) and
the rotary furnace.
AMOUNT OF PORTLAND CEMENT MADE IN KILNS OF VARIOUS KINDS.
Rotar} r furnace
1893.
Barrels.
149.000
1894.
Barrels.
242,176
Vertical kilns (continuous and intermittent) . . .
441 653
556 581
Total..
590,653
798.757
It thus appears that the output of rotary furnaces has increased much
more rapidly than that of vertical kilns. The recent rapid advance in the
price of crude oil is a great obstacle to the use of the rotary furnace. At-
tempts are being made to substitute producer-gas for crude oil in burning
cement. There is no reason why this should not be successfully done, and
the change will greatly reduce the cost of
burning cement at all points where the
rotary process is used."
169. Grinding the Clinker. " For
grinding the finished product the Griffin
steel mill is used at the larger factories.
Some of the older works still use buhr_
stones. The Griffin mill* consists of a
steel ring, against the inside surface of
which a heavy steel roll revolving on a ver-
tical shaft presses by centrifugal force*
Fig. 83. The mill is provided with screens
which allow powder of the requisite fine-
ness to pass through, while the coarser
particles drop back into the mill. This
mill is an American invention, and is
rapidly finding its way into the leading
cement-works of Germany."
FIG. 83. Perspective View of the
Griffin Mill.
* Made by the Bradley Pulverizer Co., Boston, Mass. It is used for all kinds of
pulverizing where buhrstones and stamp-mills have hitherto been employed. It works
in either wet or dry material, and is an extremely ingenious and successful grinding
machine. J. B. J.
CHAPTER XII.
THE MANUFACTURE OF VITRIFIED PAVING-BRICK.
By H. A. WHEELER, E.M.*
170. Definition. As there is a lack of harmony in the use of the term
vitrified brick, it is necessary to define what is meant by vitrified. There
is a popular idea that a vitrified brick must be glassy, in accordance with
the etymology of the word; whereas a truly glassy brick is impracticable to
make at least to a reasonably large percentage; and unless annealed with
very much more care than is now given to paving-brick, such a brick would
be too brittle for paving purposes, besides being badly misshapen. It is true
that samples of excellent paving- brick frequently exhibit to an eminent
degree a glassy or vitreous surface; but these vitreous faces are due to air-
checks (caused by the hot brick being struck by cold air), and if the brick
is broken along an unchecked or solid face it will not exhibit a glassy
surface: it will there present a very close, dense, homogeneous, stone-like
.fracture, and this fracture is what is recognized and accepted as character-
istic of a vitrified brick. There is a total absence of the individual parti-
cles of the clay in such a f lecture the presence of which characterizes build-
ing and fire-brick. Furthermore, such a vitrified brick has a hardness of 6.5
to 7, on Moh's scale of hardness, or is about as hard as quartz (the hardest
mineral in granite), and it ..readily scratches glass or the hardest steel. While
this typical vitrified fracture is easily recognized by the experienced eye,
there is no sharp line of demarcation between it and the glassy fracture
on the one side (when the brick is overturned), and a hard but unvitrified
brick on the other hand (when underlurned)\ for clay gradually passes
through a transition, when highly heated, from (1) an eminently porous,
strong, and rather hard condition just previous to tjie vitrifying-point; (2)
to a very much harder, tougher, slightly porous condition when vitrified;
and finally (3) to a very dense, glassy, non-porous condition when com-
pletely vitrified, in which latter condition it is very apt to be decidedly brit-
tle. These three stages of burning can usually be found in every kiln of
paving-brick, with all intermediate transitions from one extreme to the other,
* Formerly Assistant Geologist Missouri State Geological Survey,, in charge of inves-
tigations made on clays, and now (1896) manufacturing paving-brick in St. Louis, Mo.
196
THE MANUFACTURE OF VITRIFIED PAVING-BRICK 197
though from 60$ to 90$ usually come within the second or properly vitrified
stage.
171. Clays employed for Paving-brick. Three radically different classes
of clays are employed in the manufacture of paving-brick, viz.:
I. Surface Clays;
II. Inferior Fire-clays;
III. Shales.
Surface Clays.
By surface clays are meant those soft, unconsolidated clays at or near
the surface which have been deposited during or since the glacial period,
or that have resulted from the atmospheric decay of the underlying rocks.
This class of clays was more frequently used in the earlier development
of the paving-brick industry, but they have been almost completely
given up on account of the great difficulty in successfully vitrifying
a large percentage of the brick ; for, as a rule, they are apt to be so very
siliceous (or have from 60$ to 80$ silica), or else so very calcareous (or
have from 10$ to 25$ lime), that there is usually a very narrow range of
temperature at which they can be vitrified. Hence the brick are apt to be
either too soft (underburned), or else over burned and badly misshapen, so
that these clays have been generally abandoned for the safer-burning shales
and fire-clays.
Inferior Fire-clays.
The inferior or impure fire-clays, which are frequently known in the
trade as " bastard fire-clay " or " pipe-clay," have been quite largely used in
the past, and are still employed to some extent in the manufacture of
paving-brick. They are a class of fire-clays thaT con tain sufficient fluxing
impurities to enable them to be slightly vitrified, and the more impure
the fire-clay the more successfully it can be used for this purpose. When
the physical properties of the fire-clay are suitable, these impure fire-clays
make an excellent quality of fire-brick, though they always show rather
high absorption, or from 2$ to 5$ of water after soaking 24 hours in water.
They never show a glassy fracture, and are rarely misshapen or kiln-marked,
as in the other two classqs of clays; but they are very much more apt to be
soft, and therefore short-lived, from underburning, as it requires a very high
heat to vitrify them. When properly vitrified, .they make a very satisfac-
tory paving-brick, on account of their toughness, and some of the oldest
paving-brick in the country were made from this class of clays, notwith-
standing that they exhibit a very high absorption of water.
Shales.
The shales, or those hard, consolidated, laminated, rock-like clays that
are also popularly called "soapstone" and "soft slate," are now almost
exclusively used in the manufacture of paving-brick. They occur in very
198 THE MATERIALS OF CONSTRUCTION.
much larger and thicker bodies than either the surface clays or fire-clays,
often outcropping as low hills, and they can usually be cheaply worked by
steam-shovels in open pits. While they are usually non-plastic as they
occur in the bank, they can be easily ground to powder, when they readily
work up into a plastic mass with water. The shales are usually very high
in fluxing impurities, and this is the reason why they are so favorably
adapted for paving-brick, as this enables them to be readily vitrified. The
average composition of the shales that have proved eminently satisfactory
for this purpose is as follows:
Silica (Si0 2 ) 56 per cent.
Alumina (A1 2 3 ) - , 22 "
Ignition loss (chemically combined water). 7 "
Moisture (H 2 0) 2 "
Total non-fluxing constituents 87 per cent.
Sesquioxide of iron (Fe 2 3 ) 7 per cent.
Lime (CaO) 1
Magnesia (MgO) 1 "
Alkalies (K 2 ONa 2 0) 4
Total fluxing constituents 13 per cent.
Grand total 100 per cent.
While the best shales range quite closely around the preceding analysis,
quite a range in the fluxing constituents is permissible, as the chemical
analysis is always very secondary in the consideration of clays. All clays,
for any purpose whatever, depend primarily on their physical properties,
and if these are not favorable the chemical composition is of no importance.
A very elaborate discussion of the chemical composition and the influence
of the impurities of clays is given by the writer in Part I. of the "Report
on the Missouri Clays," to which the reader is referred for details which it
is impossible to discuss in this brief chapter.*
172. Physical Properties of Clays. The physical properties of clay, on
which depend its manufacture and uses, consist of the following factors :
I. Plasticity;
II. Shrinkage in drying and burning;
III. Speed in drying, burning, and cooling;
IV. Point of incipient, complete, and viscous vitrification;
V. Density before and after burning;
VI. Colors of burned ware;
VII. Strength of burned ware.
Plasticity.
Plasticity is the most important quality of any clay, as its ability to be
moulded depends upon this property. When mixed with the proper
* To be obtained from the State Geologist, Jefferson City, Mo.
THE MANUFACTURE OF VITRIFIED PAVING-BRICK. 199
amount of water it is called fat when it is very plastic, and the more
plastic the clay the stronger the brick will be. In making paving-brick,
excessive plasticity is found to increase the defect of laminations, which
is a great source of weakness if excessively developed. To counteract this
trouble the clay is either mixed with a less plastic one, or with sand, " grog,"
or other lean materials which reduce the plasticity.
Shrinkage.
The shrinkage is a very important factor in determining the size of
moulds and dies to produce a given-sized brick after burning. The drying
shrinkage is the reduction of volume which takes place when the soft mud
brick becomes dry from the elimination of the water used in moulding,
which amounts to 3 to 7 per cent. A second shrinkage occurs when the
dried brick is burned, which is greater the harder the brick is burned, until
thoroughly vitrified, when it ceases to shrink. The fire shrinkage varies
from 4 to 8 per cent, and the total shrinkage ranges from 7 to 15 per cent.
Speed of Drying, etc.
The speed of drying, burning, and cooling are extremely important
factors to the manufacturer in determining the size of his plant, besides
being of great importance in affecting the strength of the brick. Some
clays can be rapidly dried, burned, and cooled without 'having their strength
seriously impaired, while others are very much weakened, if not actually
cracked or ruptured, unless this is carried on very slowly. As a broad rule,
the more plastic a clay the more slowly it must be dried, burned, and cooled,
while the coarser and leaner clays can be treated much more rapidly with-
out detriment. This is a factor that is keenly appreciated by the manu-
facturer, but is rarely understood or appreciated by the engineer, yet it
affects the strength of the brick more than any one factor. It does not
follow that a clay that requires to be slowly dried must necessarily be
slowly heated and cooled, or vice versa; for there is an individuality about
clays that requires a separate determination of each of these factors as to
their amount and influence.
Vitrification.
The stages of (1) incipient, (2) complete, and (3) viscous vitrification
are extremely important, as paving-brick should be raised to at least the
stage of incipient vitrification to secure the requisite density, hardness, low
absorption, and toughness; while it must not be raised to the point of
viscous vitrification, as it then loses its shape. In the shales suitable for
paving-brick the first stage is reached at from 1500 to 1800 F., and the
second at 1800 to 2200 F., or at a very bright cherry-red.
Density.
The denser the clay the denser the brick made therefrom will be, and
the higher the density the more durable the brick. The specific gravity
200 THE MATERIALS OF CONSTRUCTION.
of shales usually range from 2.10 to 2.60, and the specific gravity of the
brick will be about the same. The impure fire-clay brick are generally some-
what lighter, or vary from 1.95 to 2.30, but the brick have a specific
gravity somewhat lower than that of the original fire-clay.
Color.
The color of paving-brick is of great local importance in estimating the
degree to which it has been burned, and the care witli which it has been
handled. If the shale is high in iron (which is usually the case) the result-
ant brick varies from red to very dark brown in color, while if the clay is
low in iron and high in lime it is light in color. Furthermore, the skill
of the burner is able to largely influence the color by the manipulation of
his fires, so that general rules for determining the quality of paving-brick
by color only are dangerous, though for specific cases and a given burner
they are of very great aid in quickly arriving at the quality of the brick.
THE MANUFACTURE OF PAVING-BRICK.
173. Preparing the Clays. The surface clays are usually obtained by
either the pick and shovel, plough and scraper or clay gatherer, or the
steam-shovel and cars, according to the size of the yard and local conditions.
The fire-clays, as they usually occur underground, are mined by the room-
and-pillar system, like coal, which is very much more expensive. The
shales are sometimes worked by the room-and-pillar system, where they
occur underground, but in most cases they are worked in open pits, by
blasting, or else worked direct from the bank into the cars by powerful
steam-shovels.
The clays are sometimes pulverized by toothed rolls," and occasionally by
centrifugal disintegrators, but in most cases a revolving dry-pan with a
perforated grate bottom is employed especially for shales and fire-clays.
The crushed clay is usually screened in either revolving trommels, or
fixed or shaking riddles, with 4 to 16 meshes to the linear inch. The
degree of fineness of the screen is a very important matter, as the finer the
clay the more plastic it is, and hence the stronger the brick. In some
cases, however, excessive fineness causes checking and cracking in drying
or burning, and aggravates the trouble from laminations, so that the fine-
ness of the screen should be determined for each specific clay. Sometimes
the clay is not screened any further than is accomplished by the screen-
plates of the dry-pan, which are usually ^ to ^ inch in width.
The screened clay is next mixed with water to a more or less plastic
mass in a pug-mill. The pug-mill consists of a trough containing a revolv-
ing shaft that is armed with blades set at an angle. It should revolve at
such a speed, or the blades should be set at such low pitch, or the length
should be sufficiently great, or the amount of clay to be pugged should be
so restricted, as to secure a thorough, uniform mixture of the clay and
water; but frequently the pug-mills are too short to accomplish this, or they
THE MANUFACTURE OF VITRIFIED PAVING-BRICK. 201
are overcrowded, or speeded too high, or the clay is run through too quickly
by the blades being given an excessive pitch, and consequently the clay
comes out with variable amounts of water. This causes checking and
cracking in the drying, and sometimes in the burning, with marked varia-
tions in the strength of the brick, besides causing the bar of clay to rag as
it leaves the brick machine. The more thoroughly a clay is pugged, the
more plastic it is rendered, and the more uniform and reliable will be the
quality of the brick, and this department could be remodelled to decided
advantage in most paving-brick plants.
174. Moulding. Three processes are employed for moulding paving
brick, to wit:
The Soft-mud Process;
The Stiff-mud Process;
The Semi-dry Process.
In tlie Soft-mud Process the clay is mixed with sufficient water to
make a very soft, extremely plastic mud, which is moulded by hand or
soft-mud machines into imperfectly formed brick; these are allowed to
partially dry, to a firm, stiff condition, and are then repressed into perfectly
formed brick. This process makes an excellent quality of brick, but it
necessitates a second handling of the brick during the drying stage. Out-
side of a few small yards, it has been quite generally given up in the paving-
brick trade, on account of the expense of the extra handling and breakage,
besides considerable risk of injuring the strength of the brick, if they arc
allowed to get too dry.
In the Stiff -mud Process the clay is pugged with sufficient water to
make a stiff, plastic mud, which is forced through a die by a continuous-
working auger or intermittent plunger, as a bar of clay, which is then cut
by wires into suitable lengths. This is the process that is almost univer-
sally employed in the manufacture of vitrified brick, as the mud is stiff
enough to be made into perfectly shaped brick, which can be loaded on to
cars without risk of being marked or injured in handling.
In extruding the bar of clay from the brick machine, two types of dies
are employed; in one the bar of clay is -approximately 3" X 4" in section,
which is cut into 9" lengths, and is known as the '* end-cut system"; while
in the other the die is approximately 4" X 9" in section, and the bar is cut
into 3" lengths, which is known as the "side-cut system." There is con-
siderable difference of opinion as to the relative merits of these two methods
of moulding, which too frequently is founded on very diversified facts. If
the clay is lean or sandy, the side-cut brick is apt to be of better quality
than the end-cut; while if the clay is very fine and eminently plastic, the
end-cut system gives fewer laminations and a superior quality to the side-
cut.
In the Semi-dry Process the clay is mixed with just enough water to
dampen it, so that it adheres slightly when firmly pressed. The brick are
202 THE MATERIALS OF CONSTRUCTION.
moulded by feeding the damp clay into a mould-box, in which it is sub-
jected to a very heavy pressure by a reciprocating plunger. This process
has been used to a very limited extent for paving-brick, as the brick are
not as tough as when made by the mud process, while it is much more
difficult to burn a large percentage of No. 1 grade. As there is consid-
erable difficulty in feeding the mould-boxes with damp clay, the mouids
are frequently only imperfectly filled, which prevents the brick from
receiving the heavy pressure necessary to bond it, besides causing imperfect
faces.
Repressing. Kecently there has been a heavy demand for repressing
the brick made by the stiff-mud process immediately after it leaves the
brick machine. In the repressing process the brick is exposed to a mod-
erate vertical pressure in a metal mould-box, while still in a plastic condi-
tion, which thoroughly fills out the edges and angles, and rounds them if
desired. This results in a brick of uniform size and perfect shape, so that
the appearance of the brick is greatly improved ; but as the pressure is
moderate, it is doubtful if the quality of the brick is enhanced by this
extra operation. Where there have been opportunities for testing the
relative merits of the same clay in repressed and unrepressed brick, the
facts indicate that the strength of the brick is endangered by breaking the
structure formed in the slow-acting brick-machine by subjecting it to such"
radically different forces as occur in a vertical-acting, quickly applied
repress. Thus some Purington unrepressed brick have been exposed for
two years on Lasalle Street, Chicago, to the heaviest kind of metropolitan
traffic, which it has very successfully withstood; while repressed brick
from the same plant has not stood so well on other Chicago streets with
much less traffic. This is a matter that needs further investigation and
more facts, and the above is the most important evidence known to the
writer that bears directly on this question. *
175. Drying and Burning. Drying. The moulded brick are hacked
on cars, in open checkerwork, direct from the brick-machine, which are run
in drying-tunnels, where they are exposed for 24 to 60 hours to light open
fires, or to a heated blast, or to the radiation of an extensive series of steam-
pipes, in order to expel the water used in moulding. Some clays can be
safely dried in 18 to 30 hours without checking or cracking, while others
have their strength seriously impaired unless the drying takes from 48 to
72 hours. Usually the finer and more plastic the clay the greater the time
required, while the coarser and leaner the cl^y the more rapidly it can be
dried.
Burning. Three classes of kilns are employed in burning: the up-
draught, the down-draught, and the continuous. The up-draught kiln,
which is the type usually employed in burning building-brick, has a series
of parallel fires at the bottom of the kiln, from which the heat rises through
the brick, and escapes at the top of the kiln. The brick that are directly
exposed to the fire at the bottom receive too mmh heat, while the brick at
* Recent (1896) rattler tests by Prof Orion, on bricks made from the same clay, on
different machines, and burned together in the same kiln, indicate clearly that repress-
ing an end-cut brick benefits it, while repressing a side-cut brick injures it.
THE MANUFACTURE OF VITRIFIED PAVING-BRICK. 203
the top of the kiln do not receive sufficient heat to vitrify them. There
is, consequently, a goodly percentage of overburned, misshapen brick at the
bottom of the kiln, and a heavy percentage of soft, unburned brick at the
top, the central portion being the only part that receives the proper degree
of heat. As the percentage of No. 1 brick, or those suitable for paving,
ranges from 35 to 65 per cent, according to the skill of the burner, the tip-
draught type of kiln is seldom employed for paving-brick. In the down-
draught type of kiln the heat rises to the top or crown of the kiln from a
series of outside fires, and then passes down through the brick to fines at
the bottom of the kiln, and then escapes to one or more stacks. The brick
are protected from excessive heat, and the heat is more thoroughly and
completely distributed through the brick than in the up-draught type. The
percentage of first-class brick is therefore much greater, as with intelligent
handling from CO to 90 per cent of No. 1 pavers can be obtained.
The down-draught kilns were formerly of the round or beehive type,
which hold from 25,000 to 75,000 brick; but in recent practice the long
rectangular design is preferred, which hold from 100,000 to 300,000 brick,
and most of the paving-brick are burned to-day in kilns of this design.
In -the continuous type of kiln the coal is fed directly in among the brick,
which are piled in a long tunnel, and the heat is drawn through them by a
high stack at the opposite end. This results in a great economy of fuel
over both the up-draught and down-draught types of kiln; but the shrinkage
and the difficulty of securing uniformity in burning is so great that they
only yield from 40 to 70 per cent of No. 1 pavers, and they are not generally
used.
The practice of glazing paving-brick with salt, similar to sewer-pipe, was
formerly employed to a considerable extent, as it gave the brick a dark color,
which was supposed to indicate hardness, besides rendering defects less con-
spicuous. As the glaze is superficial, it adds nothing to the durability of
the brick, while it greatly increases the difficulty of sorting by color, and
enables soft brick to be overlooked unless very thoroughly inspected. The
practice is to be strongly deprecated, and it is dying out.
176. Annealing. After the brick have been burned, the kiln should be
tightly closed to shut off the access of cold air, and the longer the time given
the brick to cool and anneal the tougher the brick will be. Bricks made
from the best clays can be ruined by cooling off the kiln too rapidly. If
this does not result in checking the brick it will at least make them brittle.
The conductivity of clay for heat is so feeble that unless the brick are
very slowly cooled internal stresses are produced very much as in rapidly
cooled steel or glass which interfere with the toughness of the brick.
This annealing or toughening by slow cooling is not appreciated by
engineers, though well understood by the brickmakers. They claim they
cannot afford to take the time in cooling off the kilns that they would like
to. where the price is the criterion that will determine the successful bidder,
while quality is made subordinate. The kilns are the most expensive portion
204 THE MATERIALS OF CONSTRUCTION.
in a paving-brick plant, and delay in emptying them by slow cooling adds
considerably to the expense of manufacture; so that unless the brick-
makers are paid accordingly, they cannot afford to anneal with the care
that is demanded for the best quality of brick. The usual practice in
brickyards is to " turn " or fill a kiln once a month, which allows from six
to nine days for cooling off. If the kiln capacity of a yard could be in-
creased 25 per cent so as to give the kilns twice the time to cool off, it
would result in a very much tougher, more uniform, and reliable brick;
but it would necessitate a price commensurate with this increased outlay
of capital, as there would be no increase in the quantity of brick produced.
Where the very best quality of paving-brick is required, a matter of $1.00
or $2.00 increase in the cost of the brick to insure thorough annealing would
prove to be very great economy, and a very judicious investment, in greatly
increasing the durability of the pavement.
177. Sorting. In emptying the kiln there are usually three grades of
brick made in the vitrified trade. In the down-draught type of kiln, one or
two top courses are liable to be air-checked and more or less brittle if the
kiln is either improperly designed or improperly handled in burning, while- r
the top layer is always covered with soot and ashes that mars and stains the
surface of the brick. As these brick get the highest heat they are usually,
the hardest, and while not generally tough enough for paving purposes, they
are very desirable brick for sewers, foundations, and sidewalks, especially as
they are free from kiln-marks, and are seldom misshapen. The first two or
three layers are therefore usually set aside and sold as sewer and sidewalk
brick.
The bottom portion of the kiln, or the lower two to ten courses, do not
usually receive sufficient beat to be properly vitrified, and are known as No.
2 or building brick, as they are well adapted for foundations or for backing-
stock brick.
The intermediate or central portion of the kiln are No. 1, or strictly
first-class, paving-brick, which are distinguished by the fracture, toughness,
hardness, and the color from the other two grades of brick. They should be
perfectly uniform on the fracture, homogeneous, very dense, very hard,
tough, and reasonably free from " kiln-marks," or indentations made by
overlying brick.
Kiln-marks are a splendid guide that the brick have received sufficient
heat to vitrify them, and the greater the depth of the kiln-mark the more
thoroughly the brick is usually vitrified; but if too deep they make a rough,
uneven pavement. There is usually a limit as to what is allowable for the
depth of the kiln-mark, which is a matter of opinion for the engineer, and
is placed at ^ to f inch. Except in fire-clays, it is seldom that a properly
vitrified brick is entirely free from slight indentations, unless from the
very top of the kiln; with the exception of this one place, a total absence of
such marks is apt to indicate underburning.
CHAPTER XIII.
TIMBER.*
CHARACTERISTICS AND PROPERTIES OF WOOD.
178. Structure and Appearance. The structure of wood affords the only
reliable means of distinguishing the different kinds. Color, weight, smell,
and other appearances, which are often direct or indirect results of struc-
ture, may be helpful in this distinction, but cannot be relied upon entirely.
In addition, structure underlies nearly all the technical properties of this
important product and furnishes an explanation why one piece differs as to
these properties from another.
St T ucture explains why oak is heavier, stronger, and tougher than pine;
why it is harder to saw and plane, and why it is so much more difficult to
season without injury. From its less porous structure alone, it is evident
that a piece of a young and thrifty oak is stronger than the porous wood of
an old or stunted tree ; or that Georgia or long-leaf pine excels white pine
'ii weight and strength. Keeping especially in mind the arrangement and
direction of the fibres of wood, it is clear at once why knots and " cross-
grains " interfere with the strength of timber.
It is due to structural peculiarities that " honeycombing " occurs in
rapid seasoning, that "checks "or cracks extend radially and follow pith-
rays, that tangent or "bastard" boards shrink and warp more than quar-
tered lumber. These same peculiarities enable cherry and oak to take a
better finish than basswood or coarse-grained pine.
Moreover, structure, aided by color, determines the beauty of wood.
All the pleasing figures, whether in a hard-pine ceiling, a desk of quar-
tered oak, or in the beautiful panels of "curly "or "bird's-eye" maple
decorating the saloon of a ship or a palace-car, are due to differences in the
structure of the wood. Knowing this, the appearance of any particular
* This chapter is taken from Bulletin 10 of the U. S. Forestry Division, Agricultural
Department' 1895 ; B. E. Feruow, Chief of the Division. The matter contained in this
bulletin U ijiiostly the result of original studies made by Mr. Filibert Roth, this work
beiug one c'/epartuient of the " U. S. Timber Investigations."
205
206 THE MATERIALS OF CONSTRUCTION.
section can be foretold, and almost unlimited choice and combination are
thereby suggested.
Thus a knowledge of structure not only enables us to distinguish the
different woods, judge as to their qualities, and explain the causes of their
beauty, but it also becomes an invaluable aid to the thoughtful worker,
guiding him to a more careful selection and a more perfect use of his
material.
179. Classes of Trees. The timber of the United States is famished
by three well-defined classes of trees: the needle-leaved, naked-seeded coni-
fers (pine, cedar, etc.); the dicotyledonous (with two seed-leaves), broad-
leaved trees (oak, poplar, etc.) ; and to an inferior extent by the monocoty-
ledonous (with one seed-leaf), palms, yuccas, and their allies, which last are
confined to the most southern parts of the country.
Broad-leaved trees are also known as deciduous trees, although, es-
pecially in warm countries, many of them are evergreen,* while the coni-
fers are commonly termed "evergreens/' although the larch, bald cypress,
and others shed their leaves every fall, and even the names " broad-leaved "
and "coniferous/' though perhaps the most satisfactory, are not at all
exact, for the conifer ginkgo has broad leaves and bears no cones.
In the lumber trade, the woods of broad-leaved trees are known as
*' hardwoods," though poplar is as soft as pine, and the coniferous woods
are " soft woods," notwithstanding that yew ranks high in hardness even
when compared to " hardwoods."
Both in the number of different kinds of trees or species and still
more in the importance of their product the conifers and broad-leaved trees
far excel the palms and their relatives.
In the manner of growth both conifers and broad-leaved trees behave
alike, adding each year a new layer of wood which covers the old wood in
all parts of the stem and limbs. Thus the trunk continues to grow in
thickness throughout the life of the tree by additions (annual rings) which
in temperate climates are, barring accidents, accurate records of the tree.
With the palms and their relatives the stem remains generally of the same
diameter, the tree of a hundred years being no thicker than it was at ten
years, the growth of these being only at the top. Even where a peripheral
increase takes place, as in the yuccas, the wood is not laid on in well-
defined layers; the structure remains irregular throughout.
Though alike in their manner of growth, and therefore similar in their
general make-up, conifers and broad -leaved trees differ markedly in the de-
tails of their structure and the character of their wood. The wood of all
conifers is very simple in its structure, the fibres composing tjie main part
of the wood being all alike and their arrangement regular! The wood of
broad-leaved trees is complex in structure; it is made up of several differ-
V
* In CeylOn even the cultivated cherry has become an evergreen^
TIMBER.
ent kinds of cells and fibres and lacks the regularity of arrangement so
noticeable in the conifers. This difference is so great that in a study of
wood structure it is best to consider the two kinds separately.
180. Sapwood and Heartwood. Examining a smooth cross-section or
end face of a well-grown log of Georgia pine or Norway pine, we distin-
guish an envelope of reddish, scaly bark, a small whitish pith at the centre,
and between these the wood in a great number of concentric rings.
A zone of wood next to the bark, 1 to 3 or more inches wide, and con-
taining thirty to fifty or more annual rings, is of lighter color; this is the'
sa^wood, the inner, darker part of the log beirig the heartwood. In the
former many cells are active and store up starch and otherwise assist in
the life-processes of the tree, although only the last or outer layer of cells
(the cambium layer) forms the growing part and the true life of the tree.
In the heartwood all cells are lifeless cases, and serve only the mechanical
function of keeping the tree from breaking under its own great weight, or
from being broken by the winds.
The darker color of the heartwood is due to infiltration of chemical sub-
stances into the cell-walls, but the cavities of the cells in pine are not filled
up, as is sometimes believed, nor do their walls grow thicker, nor is their
wall any more lignified than in the sapwood. Sapwood varies in width and
in the number of rings which it contains, even in different parts of the
same tree; the same year's growth which is sapwood in one part of a disk
may be heartwood in another. Sapwood is widest in the main part of
the stem and varies often within considerable limits, and without apparent
regularity. Generally it becomes narrower toward the top and in the
limbs, its width varying with the diameter, and being least, in a given disk,
on the side which has the shortest radius. Sapwood of old and stunted
pines is composed of more rings than that of young and thrifty specimens.
Thus in a pine two hundred and fifty years old, a layer of wood or annual
ring does not change from sapwood to heartwood until seventy or eighty
years after it is formed, while in a tree one hundred years old, or less, it
remains sapwood only from thirty to sixty years. The width of the sap-
wood varies considerably for different kinds of pines; it is small for long-
leaf and white pine, and great for loblolly and Norway pines. Occupying
the peripheral part of the trunk, the proportion which it forms of the
entire mass of the stem is always great. Thus even in old trees of long-leaf
pine the sapwood forms about 40 per cent of the merchantable log, while
in the loblolly and in all young trees the bulk of the wood is sapwood.
181. The Annual Rings. The concentric, annual, or yearly, rings
which appear on the end face of a log are cross-sections of so many thin
layers of wood. Each such layer forms an envelope around its inner
neighbor, and is in turn covered by the adjoining layer without, so that
the whole stem is built up of a series of thin hollow cylinders,- or rather
cones. A new layer of wood is formed each season, covering the entire
208 THE MATERIALS OF CONSTRUCTION
stem, as well as all the living branches. The thickness of this layer, or
the width of the yearly ring, varies greatly in different trees and also in
different parts of the same tree. In a normally grown, thrifty pine log
the rings are widest near the pith, growing more and more narrow
toward the bark. Thus the central twenty rings in a disk of an old
long-leaf pine may each be one-eighth to one-sixth inch (3 to 4 mm.) wide,
while the twenty rings next to the barK may average only one-thirtieth
inch (0.8 mm.). In our forest trees rings of one-half inch in width occur
only near the centre in disks of very thrifty trees of both conifers and hard
woods; one-twelfth inch represents good thrifty growth, and the minimum
width of about one'two-hundredths inch (0.12 mm.) is often seen in stunted
spruce and pine. The average width of rings in well-grown old white pine
will vary from one-twelfth to one-eighteenth inch, while in the slower grow-
ing long-leaf pine it may be one twenty-fifth to one fiftieth of an inch.
The same layer of wood is widest near the stump in very thrifty young
trees, especially if grown in the open part, but in old forest trees the same
year's growth is wider in the upper part of the tree, being narrowest near
the stump and often also near the very tip of the stem. Generally the
rings are widest near the centre, growing narrower towards the bark. In
logs from stunted trees the order is often reversed, the interior rings being
thin and the outer rings widest. Frequently, too, zones or bands of very
narrow rings, representing unfavorable periods of growth, disturb the
general regularity. Few trees, even among pines, furnish logs with truly-
circular cross-sections; usually they are oval, and at the stump commonly
quite irregular in figure. Moreover, even in very regular or circular disks
the pith is rarely in the centre, and frequently one radius is conspicuously
longer than its opposite, the width of some of the rings, if not all, being
greater on one side than on the other. This is nearly always so in the
limbs, the lower radius exceeding the upper.
In extreme cases, especially in the limbs, a ring is frequently conspicuous
on one side and almost or entirely lost to view on the other. Where the
rings are extremely narrow, the dark portion of the ring is often wanting,
the color being quite uniform and light. The greater regularity or irregu-
larity of the annual rings has much to do with the technical qualities of
the timber.
182. Spring and Summer Wood (Coniferous Trees}. Examining the
rings more closely, it is noticed tha. each ring is made up of an inner,
softer, light-colored, and an niter, or peripheral, firmer and darker-colored
portion. Being formed in the fore part of the season, the inner, light-
colored part is termed spring wood, the outer, darker portion being the
summer wood of the ring. Since the latter is very heavy and firm, it de-
termines to a large extent the weight and strength of the wood, and as its
darker color influences the shade of color of the entire piece of wood, this
color effect becomes a valuable aid in -.distinguishing heavy and strong
TIMBER. 209
from light and soft pine wood. In most hard pines, like the long-leaf,
the dark summer wood appears as a distinct band, so that the yearly
ring is composed of two sharply defined bands an inner, the spring wood,
and an outer, the summer wood. But in some cases, even in hard pines,
and normally in the wood of white pines, the spring wood passes gradually
into the darker summer wood, so that a sharply defined line occurs only
where the spring wood of one ring abuts against the summer wood of
the previous year's growth. It is this clearly defined line which en-
ables the eye to distinguish even the very narrow rings in old pines
and spruces. In some cases, especially in the trunks of Southern pines,
and normally on the lower side of pine limbs, there occur dark bands of
wood in the spring-wood portion of the ring, giving rise to false rings
which mislead in a superficial counting of rings. In the disks cut from
limbs these dark bands often occupy the greater part of the ring and appear
as "lunes" or sickle-shaped figures. The wood of these dark bands is
similar to that of the true summer wood the cells have thick walls, but
usually lack the compressed or flattened form.
Normally, the summer wood forms a greater proportion of the ring in
the part of the tree formed during the period of thriftiest growth. In an
old tree this proportion is very small in the first two to five rings about the-
pith, and also in the part next to the bark, the intermediate part showing
a greater proportion of summer wood. It is also greatest in a disk taken
from near the stump and decreases upward in the stem, thus fully account-
ing for the difference in weight and firmness of the wood of these different
parts. In the long-leaf pine the more substantial summer wood often forms
scarcely 10 per cent of the wood in the central five rings; 40 to 50 per cent
of the next one hundred rings; about 30 -per cent in the next fifty, and
only about 20 per cent in the fifty rings next to the bark. It averages 45
per cent of the wood of the stump and only 2-4 per cent of tl]at of the top.
Sawing the log into boards, the yearly rings are represented on the faces
of the middle board (radial sections) by narrow, parallel stripes (see Fig. 84),
an inner, lighter stripe, and its outer, darker neighbor, always corres-
ponding to one annual ring.
On the faces of the boards nearest the slab (tangential or "bastard"
boards) the several years' growth should also appear as parallel but much
broader stripes. This they do only if the log is short and very perfect.
Usually a variety of pleasing patterns is displayed on the boards, depend-
ing on the position of the saw-cut and on the regularity of growth of
the log. (See Fig. 84!)
Where the cut passes through a prominence (bump or crook) of the log,
irregular, concentric circlets and ovals are produced, and on almost all
tangent boards V-shaped forms occur.
183. Anatomical Structure of Coniferous Woods. Holding a well-
smoothed disk or cross-section one-eighth inch thick toward the light, it is
210
THE MATERIALS OF CONSTRUCTION.
readily seen that pine wood is a very porous structure. If viewed with a
strong magnifier, the little tubes, especially in the spring-wood of the rings,
are easily distinguished and their arrangement in regular straight radial
rows is apparent. Scattered through the summer-wood portion of the rings,
numerous irregular grayish dots (the resin-ducts) disturb the uniformity
and regularity of the structure. Magnified one hundred times, a piece of
spruce, which is similar to pine, presents a picture like that shown in Fig.
85. Only short pieces of the tubes or cells of which the wood is com-
posed are represented in the picture.
FIG. 85. Wood of Spruce.
FIG. 84. Board of Pine. C8, cross-section ;
US, radial section ; TS, tangential section ;
sw, summer wood ; spw, spring wood.*
1, natural size;
2, small part of one ring magnified 100
times. The vertical tubes are wood- fibres,
in this case all " tracheids." m, medullary
or pith ray; n, transverse tracheids or pith-
ray; a, b, and c, bordered pits of the tra-
cheids, more enlarged.
The total length of these fibres is one-twentieth to one-fifth inch, being
smallest near the pith, and is fifty to one hundred times as great as their
width (Fig. 86). They are tapered and closed at their ends, polygonal or
rounded and thin-walled, with a large cavity (lumen) or internal space in
the spring wood, while they are thick-walled and flattened radially, with the
internal space or lumen much reduced, in the summer wood. (See right-hand
portion of Fig. 85.) This flattening, together with the thicker walls of the
cells which reduces the lumen, produces the greater firmness and darker
* This figure is deceptive inasmuch as the more open or porous spring wood is repre-
sented by a plain white surface, as though it were solid, while the more solid summer
wood is represented by a shaded surface as though it were more porous. The reverse is
of course the case. J. B. J.
TIMBER.
211
mr.
color of the summer wood ; that is to say, there is more
material in the same volume. As shown in the figure,,
the tubes, cells, or " tracheids "are decorated on their
walls by circlet-like structures, called " bordered pits/'
sections of which are seen more magnified at a, b, and
c, Fig. 85. These pits are in the nature of pores, cov-
ered by very thin membranes, and serve as waterways
between the cells or tracheids.
The dark lines on the side of the smaller piece (l r
Fig. 85) appear when magnified (in 2, Fig. 85) as tiers
of eight to ten rows of cells, lying in vertical radial
planes, and are seen as bands on the radial face, and as
rows of pores on the tagential face. These kinds or
tiers of cell-rows are the " medullary rays " or "pith-
rays," and are common to all our lumber woods. In
the pines and other conifers they are quite small, but
they can readily be seen, even without a magnifier, if a
radial surface of split wood (not smooth) is examined.
The entire radial face will be seen almost covered,
with these tiny structures, which appear as fine but
conspicuous cross-lines. As shown in Fig. 85, the cells
of the medullary or pith rays are smaller and very
much shorter than the wood-fibres or tracheids, see
b, Fig. 90, and their long axis is at right angles to that
of the fibres. In pines and spruces the cells of the
upper and lower rows of each tier or pith -ray have
" bordered " pits like those of the wood-fibres or tra-
cheids proper, but the cells of the intermediate rows,
and of all rows in the rays of cedars, etc., have only
"simple" pits, i.e., pits devoid of the* saucer-like
"border" or rim.
In pine, many of the pith-rays are larger than the
FIG. 87. Block of Oak. C.S., cross-section; JR.8., radial
section; T.S., tangential section ;m. r. , medullary or pith ray;
a, height, b, width, mid e, length of a pith- ray.
212
THE MATERIALS OF CONSTRUCTION.
majority, each containing a whitish line, the horizontal resin-duct, which,
though much smaller, resembles the vertical ducts seen on the cross-section.
The larger vertical resin-ducts * are best observed on the removal of the bark
from a fresh piece of white pine, cut in winter, where they appear as con-
spicuous white linos, extending often for many inches up and down the stem.
Neither the horizontal nor the vertical resin ducts are vessels or cells,
but are openings between cells, i.e., intercellular spaces in which the resin
accumulates, freely oozing out when the ducts of a fresh piece of sap-
wood are cut. They are present only in our coniferous woods, and even
here they are restricted to pine, spruce, and larch, and are normally absent
in fir, cedar, cypress, and yew.
Altogether the structure of coniferous wood is very simple and regular,
the bulk being made up of smallfabres called tracheids, the disturbing ele-
ment, of pith-rays and resin-ductsfceing insignificant, and hence the great
unifo/mity and great technical val^of coniferous wood. -
184. Anatomical Structurff"vP"Sjtad-leajed Trees. On a cross-section
of oak, the same arrangement of
pith and bark, of sapwood and
heartwood, and the same disposi-
tion of the wood in well-defined
concentric or annual rings occurs,
but the rings are marked by lines,
or rows, of conspicuous pores or
openings which occupy the
greater part of the spring wood
of each ring (see Fig. 87, also
Fig. 89) and are, in fact, the
openings through the vessels cut
by the section. On the radial
section, or quarter-sawed board,
the several layers appear as so
many parallel stripes (see Fig.
88) ; on the tangential section or
" bastard " face, patterns similar
to those mentioned for pine wood
are observed. But while the pat-
terns in hard pine are marked by
the darker summer wood and are
FIG. 88. Board of Oak. CS, cross-section; ES, composed of plain, alternating
radial section; TS, tangential section ;i>, vessels stripes of darker and lighter
or pores, cut through; A, slighVcurve in log WQod> the fig ures j n oa k ( an( J
which appears in section as an islet. ofcher broad . leaved woods ) are
due chiefly to the vessels, those of the spring wood in oak being the most
* See rd t Fig. 118.
TIMBER.
213
conspicuous (see Fig. 88); so that in an oak table the darker, shaded parts
are the spring wood, the lighter parts the summer wood.
On closer examination of the smoothed cross-section of oak, the spring-
wood part of the ring is found to be formed, in great part, of pores: large,
round, or oval openings through long vessels. These are separated by a
FIG. 89. Cross-section of Oak magnified about 5 times.
grayish and quite porous tissue (see Fig. 89) which continues here and there
in the form of radial, often branched, patches (not the pith-rays) into and
through the summer wood to the spring wood of the next ring. The large
vessels of the spring wood, occupying 6 to 10 per cent of the volume of a
log in very good oak, and 25 per cent or more in inferior and narrow-ringed
lumber, are a very important feature, since
it is evident that the greater their share in
the volume, the lighter and weaker the
wood. They are smallest near the pith, and
grow wider outward; they are wider in the
stem than limb and seem to be of indefinite
length, forming open channels in some cases
probably as long as the tree itself.
Scattered through the radiating gray
patches of porous wood are vessels similar '
those of the spring wood, but decided^
smaller. These vessels are usually fewer ^and
larger near the spring wood, and smaller and
more numerous in the outer portions of the FIG. 90. Portion of the Firm Bodies
ring. Their number and sizecan be utilized of Fibres with Two Cells of a small
to distinguish the oaks classed as white oaks Pilh - ray ' mr ' Highly magnified.
from those classed as black and red oaks; they are fewer and larger in red
oaks, smaller but much more numerous in white oaks. The summer wood,
except for these radial grayish patches, is daik-colored and firm. This firm
portion, divided into bodies or strands by these patches of porous wood and
also by fine wavy concentric lines of short, thin-walled cells (see Fig. 89),
consists of thick-walled fibres (see Fig. 90) and is the chief element of
214
THE MATERIALS OF CONSTRUCTION.
strength in oak wood. In good white oak it forms one half and more of the
wood; it cuts like horn, and the cut surface is shiny and of a deep chocolate-
brown color. In very narrow-ringed wood and in inferior red oak it
is usually much reduced in quantity as well as quality.
The pith-rays of the oak, unlike those of coniferous woods, are at
least in part very large and conspicuous (see Fig. 87, their height in-
dicated by the letter a, and their width by the letter Z>). The large
medullary rays of oak are often twenty and more cells thick and sev-
eral hundred cell-rows in height, which amount commonly to one or
more inches. These large rays are con-
spicuous on all sections. They appear as
long, sharp, grayish lines on the cross-sec-
tion, as short, thick lines, tapering at each
end, on the tangential or "bastard " face,
and as broad, shiny bands, the "silver
grain " or " mirrors," on the radial section.
In addition to these coarse rays, there is
also a large number of small pith-rays,
which can be seen only when magnified.
On the whole, the pith-rays form a much
larger part of the wood than might be sup-
posed. In specimens of good white oak it
has been found that they formed about 1C
to 25 per cent of the wood.
185. Minute Structure. If a well-
smoothed, thin disk or cross-section of
oak (say one-sixteenth inch thick) is held
up to the light, it looks very much like a
sieve, the pores or vessels appearing as
clean-cut holes; the spring wood and gray
patches are seen to be quite porous, but
the firm bodies of fibres between them are
dense and opaque. Examined with the
magnifier it will be noticed that there is
no such regularity of arrangement in
straight rows as is conspicuous in the pine;
on the contrary, great irregularity prevails.
At the same time, while the pores are as
as large as pin-holes, the cells of the denser
wood, unlike those of pine wood, are too
small to be distinguished. Studied with
the microscope, each vessel is found to be a vertical row of a great
number of short, wide tubes, joined end to end (Fig. 91, 6-). The porous
spring wood and radial gray tracts are partly composed of smaller vessels,
FIG. 91. Isolated Fibres, and Cells.
a, four cells of wood-parenchyma;
b, two cells from a pith-ray; c, a
single joint or cell of a vessel, the
openings x leading into its upper
and lower neighbors; d, tracheid;
e, wood-fibre proper.
TIMBER.
215
but chiefly of tracheids like those of pine, and of shorter cells, the "wood-
parenchyma/' resembling the cells of the medullary rays. These latter, as
well as the fine concentric lines mentioned as occurring in the summer wood,
are composed entirely of short, tube-like parenchyma-cells with square or
oblique ends (Fig. 91, a and b). The wood-fibres proper, which form the
dark, firm bodies referred to, are very fine, thread-like cells one twenty-fifth
to one-tenth inch long, with a wall commonly so thick that scarcely any
empty internal space or lumen remains (Figs. 91, e, and 90).
If instead of oak a piece of poplar or basswood (Fig. 92) had been used
in this study, the structure would have been found to be quite different.
FIG. 92 Cross-section of Basswood (magnified), v, vessels ; mr, pith-rays.
The same kinds of cell-elements, vessels, etc., are present, but their com-
bination and arrangement is different, and thus from the great variety of
possible combinations results the great variety of structure and, in conse-
quence, of the qualities which distinguish the wood of broad-leaved trees.
The sharp distinction of sapwood and heartwood is wanting; the rings are
not so clearly defined, the vessels of the wood are small, very numerous,
and rather evenly scattered through the wood of the annual ring, so that
the distinction of the ring almost vanishes and the medullary or pith rays,
in poplar, can be seen, without being magnified, only on the radial section.
186. Different "Grains" of Wood. The terms "fine-grained/' "coarse-
grained," "straight-grained," and " cross-grained " are frequently applied
in woodworking. In common usage, wood is " coarse-grained " if its
annual rings are wide, " fine-grained " if they are narrow; in the finer wood
industries a " fine-grained " wood is capable of high polish, while a " coarse-
grained " wood is not, so that in this latter case the distinction depends
chiefly on hardness, and in the former on an accidental case of slow or
rapid growth.
Generally the direction of the wood-fibres is parallel to the axis of the
stem or limb in which they occur, the wood is straight-grained, but in
many cases the course of the fibres is spiral or twisted around the tree as
THE MATERIALS OF CONSTRUCTION
shown in Fig. 93, and sometimes (commonly in butts of gurn and cypress)
the fibres of several layers are oblique in one Direction, and those of the
FIG. 93. FIG. 94.
FIG. 93. Spiral Grain. Season-checks, after removal of bark, indicate the direction of
the fibres or grain.
FIG. 94. Alternating Spiral Grain in Cypress. Side and end view of same piece. When
the bark was at o the grain at this point was straight. From that time each year it
grew more oblique in one direction, reaching a climax at a, and then turned back in
the opposite direction. These alternations were repeated periodically, the bark
sharing in these changes.
next series of layers are oblique in the opposite direction, as shown. in
Fig. 94; the wood is cross- or twisted-grained. Wavy grain in a tan-
gential plane as seen on the radial section is illustrated in Fig. 940, which
represents an extreme case observed in beech.
This same form also occurs on the radial plane,
causing the tangential section to appear wavy or
in transverse folds. When wavy grain is fine,
i.e., the folds or ridges small but numerous, it
gives rise to the "curly" structure frequently
seen in maple. Ordinarily, neither wavy, spiral,
nor alternate grain is visible on the cross-section;
its existence often escapes the eye even on smooth,
longitudinal faces in sawed material, so that the
T ly saf ! Sui 1 e to their discovery lies in splitting
the wood m the two normal planes.
Generally the surface of the wood under the bark, and therefore also
that of any layer in the interior, is not uniform and smooth, but is chan-
nelled and pitted by numerous depressions which differ greatly in size and
form. Usually, any one depression or elevation is restricted to one or a
few annual layers (i.e., seen only in one or a few rings), and is then lost,
TIMBER.
217
being compensated (the surface at the particular spot evened up) by growth.
In some woods, however,' w ny depression or elevation once attained grows
from year to year and reaches a maximum size which is maintained for
many years, sometimes throughout life.
In maple, where this tendency to preserve any particular contour is
very great, the depressions and elevations are usually small (commonly less
than one-eighth inch, but very numerous. On
tangent boards of such wood the sections of these
pits and prominences appear as circlets and give
rise to the beautiful. " bird Veye " or "land-
scape " structure. Similar structures in the burls
of black ash, maple, etc., are frequently due to
the presence of dormant buds, which cause the
surface of all the layers through which they pass
to be covered by small conical elevations, whose
cross-sections on the sawed board appear as ir-
regular circlets or islets each with a dark speck,
the section of the pith or " trace " of the dor-
mant bud in the centre.
In the wood of many broad-leaved trees the
wood-fibres are much longer when full grown
than when they are first formed in the cambium
or growing zone. This causes the tips of each
fibre to crowd in between the fibres above and
below, and leads to an irregular interlacement of
these fibres, which adds to the toughness but
reduces the cleavability of the wood.
At the junction of limb and stem the fibres
on the upper and lower sides of the limb behave
differently. On the lower side they run from the
stem into the limb, forming an uninterrupted
strand or tissue and a perfect union. On the
upper side the fibres bend aside, are not con-
tinuous into the limb, and hence the connection
is imperfect (Fig. 95).
Owing to this arrangement of the fibres, the
cleft made in splitting never runs into the knot,
if started on the side above the limb, but is apt
to enter the knot if started below, a fact well
understood in woodcraft. When limbs die, de-
cay, and break off, the remaining stubs are sur-
rounded and may finally be covered by the growth
of the trunk, and thus give rise to the annoying "
55. Section of Wood
showing Position of the
Grain at Base of a Limb
which has been Dead Three
Years. P, pith of both
stem and limb ; 1-7, seven
yearly layers of wood; a,
b, knot or basal part of a
limb which lived four
years, then died and broke
off near the stem, leaving
the part to the left of a, b,
a " sound " knot, the part
to the right a " dead "
knot, which would soon be
entirely covered by the
growing stem.
dead " or " loose " knots.
187. Color and Odor. Color, like structure, lends beauty to the wood,
218 THE MATERIALS OF CONSTRUCTION.
aids in its identification, and is of great value in ^.he determination of its
quality. Considering only the heartwood, the black color of the per-
simmon, the dark brown of the walnut, the light brown of the white oaks,
the reddish brown of the red oaks, the yellowish white of the tulip and
poplar, the brownish red of the redwood and cedar, the yellow of the papaw
and sumac, are all reliable marks of distinction; and color together with
lustre and weight are only too often the only features depended upon in
practice. Newly formed wood, like that of the outer few rings, has but
little color. The sapwood generally is light, and the wood of trees which
form no heartwood changes but little, except when stained by forerunners
of disease.
The different tints of colors, whether the brown of oak, the orange-
brown of pine, the blackish tint of walnut, or the reddish cast of cedar, are
due to pigments, while the deeper shade of the summer-wood bands in pine
and cedar, or in oak or walnut, is due to the fact that, the wood being
denser, more of the colored wood substance occurs on a given space, i.e.,
there is more colored matter per square inch.
Wood is translucent, a thin disk of pine permitting light to pass
through quite freely. This translucency affects the lustre and brightness
of lumber. When wood is attacked by fungi it becomes more opaque, loses
its brightness, and in practice is designated "dead" in distinction from
"live "or bright timber. Exposure to air darkens all wood; direct sun-
light and occasional moistening hasten this change and cause it to pene-
trate deeper. Prolonged immersion has the same effect, pine wood becom-
ing a dark gray, while oak changes to a blackish brown.
Odor, like color, depends on chemical compounds, forming no part of
the wood substance itself. Exposure to weather reduces and often
changes the odor, but a piece of dry long-leaf pine, cedar, or camphor wood
exhales apparently as much odor as ever when a new surface is exposed..
Heartwood is more odoriferous than sapwood. Many kinds of wood are
distinguished by strong and peculiar odors. This is especially the case
with camphor, cedar, pine, oak, and mahogany, and the list would com-
prise every kind of wood in use were our sense of smell developed in keep-
ing with its importance. Decomposition is usually accompanied by pro-
nounced odors; decaying poplar emits a disagreeable odor, while red oak
often becomes fragrant, its smell resembling that of heliotrope.
188. Resonance. If a log or scantling is struck with the axe or hammer,
a sound is emitted which varies in pitch and character with the shape and
size of the stick, and also with the kind and condition of wood. Not only
can sound be produced by a direct blow, but a thin board may be set vibrat-
ing and be made to give a tone by merely producing a suitable tone in its
vicinity. The vibrations of the air, caused by the motion of the strings of
the piano, communicate themselves to the board, which vibrates in. the
same intervals as the string and reenforces the note. The note which a
TIMBER. 219
given piece of wood may emit varies in pitch directly with the elasticity,
and indirectly with the weight, of the wood. The ability of a properly
shaped sounding-board to respond freely to all the notes within the range
of an instrument, as well as to reflect the character of the notes thus
emitted (i.e., whether melodious or not), depends, first, on the structure of
the wood and next on the uniformity of the same throughout the board.
In the manufacture of musical instruments all wood containing defects,
knots, cross grain, resinous tracts, alternations of wide and narrow rings,
and all wood in which summer and spring wood are strongly contrasted in
structure and variable in their proportions, is rejected, and only radial
sections (quarter-sawed 'or split) of wood of uniform structure and growth
are used.
The irregularity in structure, due to the presence of relatively large
pores and pith-rays, excludes almost all our broad-leaved woods from such
use, while the number of eligible woods among conifers is limited by the ne-
cessity of combining sufficient strength with uniformity in structure, absence
of two pronounced bands of summer wood, and relative freedom from resin.
Spruce is the favored resonance wood; it is used for sounding-boards
both in pianos and violins, while for the resistant back and sides of the latter
the highly elastic hard maple is used. Preferably resonance wood is not
bent to assume the final form; the belly of the violin is shaped from a
thicker piece, so that every fibre is as nearly in its original unstrained con-
dition as possible, and therefore free to vibrate. All wood for musical
instruments is, of course, well seasoned, the final drying in kiln or warm
room being preceded by careful seasoning at ordinary temperatures often
for as many as seven years or more. The improvement of violins, not by
age but by long usage, is probably due not only to the adjustment of the
numerous component parts to each other, but also to a change in the wood
itself; years of vibrating enabling any given part to vibrate much more
readily. \
"* SPECIFIC GRAVITY, OR WEIGHT.
189. Weight Dependent on Structure and Moisture. A small cross-
section of wood, as in Fig. 96, dropped into water, sinks, showing that the
substance of which wood-fibre or wood is built up is heavier
than water. By immersing the wood successively in heavier
liquids, until we find a liquid in which it does not sink,
and comparing the weight of the same with water, we find
that wood-substance is about 1.6 times as heavy as water,
and that this is as true of poplar as of oak or pine.
c, , . , T-T -i Cross - section
Separating a single cell, as shown in Fig. 97, a, drying of ^ Grou D of
and then dropping it into water, it floats. The air-filled Wood fibres,
cell-cavity or interior reduces its weight, and, like a corked
empty bottle, it weighs less than the water. Soon, however, -water soaks
into the cell, when it fills up and sinks.
220 THE MATERIALS OF CONSTRUCTION.
Many such cells grown together, as in a block of wood, sink when all or
most of them are filled with water, but will float as long as the majority are
empty or only partly filled. This is why a green, sappy pine pole soon
sinks in " driving " (floating). Its cells are largely filled before it is thrown
in, and but little additional water suffices to make its weight greater than
that of the water.
In a white-pine log, composed chiefly of empty cells (heart wood), the
water requires a very long time to fill up the cells (five years would not
suffice to fill them all), and therefore the log may float for many months or
even years. When the wall of the wood-fibre is very thick (five-eighths or
more of the volume), as in Fig. 97, 1, the fibre sinks whether empty or filled.
This applies to most of the fibres of the dark summer-wood
bands in pines, and to the compact fibres of oak or hickory,
and many, especially tropical woods, have, such thick- walled
.cells and so little empty or air space that they never float.
Here, then, are the two main factors of weight in wood:
The amount of cell-wall, or wood-substance, constant for any
given piece, and the amount of water contained in the wood,
variable even in the standing tree, and only in part eliminated
in drying.
The weight of the green wood of any species varies chiefly
as the second factor, and is entirely misleading if the relative
weight of different kinds is sought. Thus some green sticks
of the otherwise lighter cypress and gum sink more readily
than fresh oak.
* The weight of sapwood, or the sappy peripheral part of
our common lumber woods, is always great whether cut in
winter or summer. It rarely falls much below 45 pounds and'
commonly exceeds 55 pounds to the cubic foot, even in our lighter wooded
species.
It follows that the green wood of a sapling is heavier than that of an
old tree, the fresh wood from a disk of the upper part of a tree often
heavier than that of the lower part, and the wood near the bark heavier
than that nearer the pith, and also that the advantage of drying the wood
before shipping is most important in sappy and light kinds.
When kiln-dried, the misleading moisture factor of weight is uniformly
reduced and a fair comparison made possible. For the sake of convenience
in comparison, the weight of wood is expressed either as the weight per cubic
foot or, what is still more convenient, as specific weight or density.
190. Variation in Weight in a Single Trunk. If an old long-leaf pine
is cut up as shown in Fig. 98, the wood of disk No. 1 is heavier than that
of disk No. 2, the latter heavier than that of disk No. 3, and the wood of
the top disk is found to be only about three fourths as heavy as that of
disk No. 1.
TIMBER. 221
Similarly, if disk No. 2 is cut up as in the figure, the specific weight of
the different pieces is:
a about 0.52
I about 0.64
c about 0.67
d, e,f about 0.65
Showing that in this disk, at least, the wood formed during the many years'
growth, represented in piece a, is much lighter than that of former years.
It also shows that the best wood is the middle part, with its large proportion
of dark summer-wood bands.
Cutting up all disks in the same way, it will be found that the piece a
of the first disk is heavier than piece a of the fifth, and that piece c of the
first disk excels the piece c of all the other disks. This shows that the wood
grown during the same number of years is lighter in the upper parts of the
stem; and if the disks are smoothed on their radial surfaces and set up one
on top of the other in their regular order for sake of comparison, this decrease
in weight will be seen to be accompanied by a
decrease in the amount of summer wood. The
color effect of the upper disks is conspicuously
lighter.
If our old pine had been cut one hundred
and fifty years ago, before the outer, lighter wood
was laid on, it is evident that the weight of the wood
of any one disk would have been found to increase
from the centre outward, and no subsequent decrease
could have been observed.
In a thrifty young pine, then, the wood is heav- FlG " 98 -~
ier from the centre outward, and lighter from below
upward; only the wood laid on in old age fails in weight below the
average. The number of brownish bauds of summer wood are a direct indi-
cation of these differences.
If an old oak is cut up in the same manner, the butt cut is also found
heaviest and the top lightest, but, unlike the disk of pine, the disk of oak has
its firmest wood at the centre, and each successive piece from the centre out-
ward is lighter than its inner neighbor.
Examining the pieces, this difference is not as readily explained by the
appearance of each piece as in the case of pine wood. Nevertheless, one con-
spicuous point appears at once: the pores, so very distinct in oak, are very
minute in the wood near the centre and thus the wood is far less porous.
Studying different trees it is found that, in the pines, wood with narrow
rings is just as heavy as, and often heavier than, the wood with wider rings,
but if the rings are unusually narrow in any part of the disk the wood has
a lighter color; that is, there is less summer wood and therefore less weight.
In oak, ash, or elm trees of thrifty growth, the wider rings (not less than
222
THE MATERIALS OF CONSTRUCTION.
one-twelfth inch) always form the heaviest wood, while any piece with very
narrow rings is light. On the other hand/the weight of a piece of hard
maple or birch is quite independent of the width of its rings, since the
structure here is uniform across the entire width of the annual ring.
The bases of limbs (knots) are usually heavy, very heavy in conifers, and
also the wood which surrounds them, but generally the wood of the limbs is
lighter than that of the stem, and the wood of the roots is the lightest.
191. Weight of Different Species. In general, it may "be said that none
of the native woods in common use in this country are, when dry, as heavy
as water, i.e., G2 pounds to the cubic foot. Few exceed 50 pounds, while
most of them fall below 40 pounds, and much of the pine and other conif-
erous wood weighs less than 30 pounds per cubic foot.
The weight of the wood is, in itself, an important quality. Weight assists
in distinguishing maple from poplar. Lightness, coupled with great strength
and stiffness, recommends wood for a thousand different uses. To a large
extent weight predicates the strength of the woody at least in the same species,
so that a heavy piece of oak will exceed in strength a light piece of the same
species, and in pine it appears probable that, weight for weight, the strength
of the wood of various pines is nearly equal, all being reduced to the same
*ee of dry ness.
WEIGHT OF KILN-DKIED WOOD OF DIFFERENT SPECIES.
Common Name of Species.
Approximate.
Specific
Gravity.
Weight of-
1 cubic
foot.
1000 feet
of lum-
ber.
(a) Very heavy woods:
* Hickory, oak, persimmon, osage orange, black locust,
hackberry blue beecb best of elm and ash
0.70-0.80
.60- .70
.50- .60
.40- .50
.30- .40
Pounds.
42-48
36-42
30-36
24-30
18-24
Pounds.
3700
3200
2700
2200
1800
(b) Heavy woods:
Ash, elm, cherry, birch, maple, beech, walnut, sour gum,
coffee-tree, honey-locust, best of Southern pine, and
tamarack
(c) "Woods of medium weight:
Southern pine, pitch-pine, tamarack, Douglas spruce,
Western hemlock, sweet gum, soft maple, sycamore,
sassafras, mulberry, light grades of birch and cherry. .
(d) Light woods:
Norway and bull pine, red cedar, cypress, hemlock, the
heavier spruce and fir, redwood, basswood, chestnut,
butternut, tulip, catalpa, buckeye, heavier grades of poplar
(e) Very light woods :
* For the scientific names of timbers see the list of Useful American Timbers at the
end of this chapter.
TIMBER. 223
Since ordinary lumber contains knots and also more water than is here
assumed, and also since its dimensions either exceed or fall short of perfect
measurement, the figures in the table are only approximate.
Thus 1000 feet, 13. M., of long-leaf pine weighs:
Pounds.
Rough and green 4500
Boards rough but seasoned 3500
Boards dressed and seasoned 3000
Flooring, matched dressed and seasoned 2500
Weather-boarding, bevelled and dressed. 1500
MOISTURE IN WOOD.
192. Moisture Distribution. Water may occur in wood in three con-
ditions: (1) it forms the greater part (over 90 per cent) of the proto-
plasmic contents of the living cells; (2) it saturates the walls of all
cells; and (3) it entirely or at least partly fills the cavities of the life-
less cells, fibres, and vessels. In the sapwood of pine it occurs in all
three forms; in the heartwood only in the second form, that is, it
merely saturates the walls. Of 100 pounds of water associated with
100 pounds of dry wood-substance in 200 pounds of fresh sapwood of
white pine, about 35 pounds are needed to saturate the cell walls,
less than 5 pounds are contained in living cells, and the remaining 60
pounds partly fill the cavities of the wood-fibres. This latter forms the
sap as ordinarily understood. It is water brought from the soil, containing
small quantities of mineral salts, and in certain species (maple, birch, etc.)
it also contains at certain times a small percentage of sugar and other
organic matter. These organic substances are the dissolved reserve food
stored during winter in the pith-rays, etc., of the wood and bark; generally
but a mere trace of them is to be found. From this it appears that the
solids contained in the sap, such as albumen, gum, sugar, etc., cannot
exercise the influence on the strength of the wood which is so commonl^
claimed for them.
The Avood next to the bark contains "the most water. In the species
which do not form heartwood the decrease toward the pith is gradual; but
where this is formed, the change from a more moist to a drier condition is
usually quite abrupt at the sapwood limit. In long-leaf pine the wood of
the outer 1 inch of a disk may contain 50 per cent of water, that of the
next, or second inch, only 35 per cent, and that of the heartwood only 20
per cent. In such a tree the amount of water in an entire cross-section
varies with the amount of sapwood, and is therefore greater for the upper
than the lower cuts, greater for limbs than stems, and greatest of all in the
roots.
Different trees, even of the same kind and from the same place, differ
as to the amount of water they contain. A thrifty tree contains more water
224 THE MATERIALS OF CONSTRUCTION.
than a stunted one, and a young tree more than an old one, while the wood
of all trees varies in its moisture relations with the season of the year.
Contrary to the general belief, a tree contains about as much water in
winter as in summer. The fact that the bark peels easily in the spring
depends on the presence of incomplete, soft tissue, the rapidly growing
cambium layer found between wood and bark during this season, and has
little to do with the total amount of water contained in the wood, of the
stem.
Even in the living tree a flow of sap from a cut occurs only in certain
kinds of trees and under special circumstances; from boards, timber, etc.,
the water does not flow out, as is sometimes believed, but must be evapo-
rated.*
193. Drying Timber. The rapidity with which water is evaporated,
that is, the rate of drying, depends on the size and shape of the piece and
on the structure of the wood. An inch board dries more than four times
as fast as a 4-inch plank and more than twenty times as fast as a 10-inch
timber. White pine dries faster than oak. A very moist piece of pine or
oak will, during one hour, lose more than four times as much water per
square inch from the cross-section, but only one half as much from the
tangential as from the radial section.
In a long timber, where the end or cross-sections form but a small part
of the drying surface, this difference is not so evident. Nevertheless, the
ends dry and shrink first, and being opposed in this shrinking by the more
moist adjoining parts they check, the cracks largely disappearing as season-
ing progresses.
High temperatures are very effective in evaporating the water from
wood no matter how humid the air. A fresh piece of sap wood may lose
weight in boiling water, and can be dried to quite an extent in hot steam.
Kept on a shelf in an ordinary dwelling, wood still retains 8 to 10 per
cent of its weight of water, and this percentage is always greater than the
percentage of moisture in the surrounding air. Nor is the amount of
water in dry wood constant; the weight of a panful of shavings varies;
with the time of day, being on a summer day greatest in the morning and
least in the afternoon.
Desiccating the air with chemicals will cause the wood to dry, but wood
thus dried at 80 R will still lose water in the kiln. Wood dried at 120
F. loses water still if dried at 200 F., and this again will lose more water
if the temperature is raised. Absolutely dry wood cannot be obtained;
chemical destruction sets in before all the water is driven off.
On removal from the kiln the wood at once takes up water from the air, j
*The seeming exceptions to this rule are mostly referable to two causes, namely.';
(a) Clefts or "shakes" will allow water contained in them, to flow out. (b) Froroj
sound wood, if very sappy, water is forced out whenever the wood is warmed, just as ;
water flows from green wood in the stove.
TIMBER. 225
even in the driest weather. At first the absorption is quite rapid; at the
end of a week a short piece of pine, 1 inches thick, has regained two
thirds of, and in a few months all, the moisture which it had when air-dry,
8 to 10 per cent, and also its former dimensions.
In thin boards all parts soon attain the same degree of dryness; in
heavy timbers the interior remains moister for many months, and even
years, than the exterior parts. Finally an equilibrium is reached, and then
only the outer parts change with the weather.
With kiln-dried wood all parts are equally dry, and when exposed the
moisture coming from the air must pass in through the outer parts, and
thus the order is reversed. Timber seasoned out of doors requires months,
or even years, before it is at its best; kiln-dry timber, if properly handled,
is prime at once.
Dry wood when soaked in water soon regains its original volume, and
in the heartwood portion it may even surpass it; that is to say, swell to a
larger dimension than it had when green. With the soaking it continues
to increase in weight, the cell-cavities filling with water, and if left many
months all pieces sink. Yet even after a year's immersion a piece of oak
2 by 2 inches and only 6 inches long still contains air, i.e., it has not
taken up all the water it can. By rafting, or prolonged immersion, wood
loses some of its weight, soluble materials being leached out, but it is not
impaired either as fuel or as building material. Immersion and, still more,
boiling and steaming reduce the hygroscopicity of wood and, therefore, also
the troublesome " working," or shrinking and swelling.
Exposure in dry air to a temperature of 300 F. for a short time re-
duces, but does not destroy, the hygroscopicity and with it the tendency to
shrink and swell. A piece of red oak which has been subjected to a tem-
perature of over 300 F. still swells in hot water and shrinks in the kiln.
In artificial drying, temperatures of from 158 to 180 F. are usually
employed. Pine, spruce, cypress, cedar, etc., are clried fresh from the saw,
allowing four days for 1-inch boards; hard woods, especially oak, ash,
maple, birch, sycamore, etc., are air-seasoned for three to six months, to
allow the first shrinkage to take place more gradually, and are then exposed
to the above temperatures in the kiln for about six to ten days for 1-inch
lumber. Freshly cut poplar and cotton wood are often dried directly in
kilns.
By employing lower temperatures, 100 to 120 F., green oak, ash, etc.,
can be seasoned in dry kilns without danger to the material.* Steaming
the lumber is commonly resorted to in order to prevent checking and
" casehardening," but not, as has frequently been asserted, to enable the
board to dry. Yard-dried lumber is not dry, and its moisture is too un-
* The dry kiln shown in Fig. 99 is operated at this temperature, and -live steam is
admitted once or twice a day to prevent checking. J. B. J.
226
THE MATERIALS OF CONSTRUCTION.
evenly distributed to insure good behavior after manufacture. Careful
piling of the lumber both in the yard and kiln, is essential to good drying.
nnnnnnnnnnci^
:
o o o o o
'
,
f Steam
Coils
Lumber Dryer for U.S. Test Timbers
Dimensions: Length 3O ft.
Heigth 16 ft.
Width 12. ft.
FIG. 99.
Piling boards on edge or standing them on end is believed to hasten drying.
This is true only because in either case the air can circulate more freely
.around them than when they are piled in the ordinary way. Boards on
TIMBER.
227
end dry unequally; the upper half dries much faster than the lower half
and horizontal piling is, therefore, preferable.
Since the proportion of sap- and heart- wood varies with size, age, species,
and individual, the following figures must be regarded as mere approxima-
tions :
POUNDS OF WATER LOST IX DRYING 100 POUNDS OF GREEN WOOD
IN THE KILN.
Common Names of Species.
Sap wood or
Outer Part.
Heartwood
or Interior.
45-65
16-25
(2) Cypress, extremely variable
50-65
18-60
(3) Poplar, cot ton wood basswood
60-65
40-60
(4) Oak, beech, ash, elm, maple, birch, hickory, chestnut, wal-
nut, and sycamore ...
40-50
30-40
The lighter kinds have the most water in the sapwood; thus sycamore has more than
hickory.
SHRINKAGE OF WOOD.
194. Shrinkage Explained. When a short piece of wood-fibre, such as
that shown in Fig. 100, A, is dried it shrinks, its wall grows thinner (as in-
dicated by dotted lines), its width, ab,
the thickness of the fibre, becomes
smaller, and the cavity or opening
larger, but, strange to say, the height
or length, lc, remains the same. In a
similar piece of fibre with a thinner
wall (Fig. 100, B) the effect is the
same, but the wall being only half as
thick the total change is only about
half as great.*
If sections or pieces of fibres are
dried and then placed on moist blotting-
paper, they will take up water and
swell to their original size, though the
water has been taken up only by their
walls and none has entered into their
lumina.
openings or
This indicates FIG. 100. Short Pieces of Wood-fibres,
one thick-, the other thin-walled; mag-
nified.
that the water in the cavity or lumen of
a fibre has nothing to do with its di-
mensions, and that if the cell-walls are saturated it makes no difference
* Though generally true, it must not be supposed that the fibres of all species, or
even the fibres of the same tree, shrink exactly in proportion to the thickness of their
walls.
228
THE MATERIALS OF CONSTRUCTION.
in the volume of a block of pine wood whether the cell-cavities are empty
as in the heartwood or three fourths filled as in the sapwood.
If an entire fibre, as shown in Fig. 101, is dried, the wall at its ends
a and b, like those of the sides, grows thinner, and thereby the length of
the entire cell grows shorter. Since this length is often a hundred or
more times as great as the diameter, the effect of this shrinkage is inap-
preciable; and if a long board shrinks lengthwise, it is largely due, as
we shall see, to quite another cause.
A thin cross-section of several fibres (see Fig. 102, A) like the piece
of a single fibre shrinks when dried, the wall of each fibre becomes
FIG. 101.
Isolated Cell.
FIG. 10'2. Warping of
Wood.
thinner, and thus each piece smaller, and the piece on the whole necessarily
shares this diminution of size, the distances ab and cd each becoming
shorter. Where the cells are very similar in size and in the thickness of
their walls, as in the case of piece A, Fig. 102, ab and cd become shorter
by about the same amount ; but if the piece is made up of fibres some
of which have thin and others thick walls, as piece B, Fig. 102, then, the
row of thick-walled cells shrinking much more than the row of thin-walled
cells, the piece becomes unevenly shrunk or warped as shown in Fig. 102,
C. Not only is the piece warped, but the force which led to this warping
continues to strain the interior parts of the piece in different directions.
Since in all our woods cells with thick walls and cells with thin walls
are more or less intermixed, and especially as the spring wood and summer
wood nearly always differ from each other in this respect, strains and
tendencies to warp are always active when wood dries out, because the
summer wood shrinks more than the spring wood, heavier wood in general
more than li^ht wood of the same kind.
TIMBER.
229
X
If the piece A, Fig. 102, after drying is placed edgewise on moist blot-
ting-paper, the cells on the under side, at cd, take up moisture from the
paper and swell before the upper cells at ab re-
ceive any moisture. This causes the under side ^
of the piece to become longer than the upper
side, and, as in the case of piece C, warping c
occurs. Soon, however, the moisture penetrates
to all the cells and the piece straightens out. A
thin board behaves exactly like this minute
piece, only the process is slower and more easily
observed. But while a thin board of pine curves
laterally, it remains quite straight lengthwise,
since in this direction both shrinkage and swell-
ing are small. A thin disk or cross-section
swells, and when moistened on one side warps as
readily in one direction as in another. If a green
board is exposed to the sun upon one side, warp-
ing is produced by removal of water and conse-
quent shrinkage of that side, and the course of
the process is simply reversed.
As already stated, wood loses water faster
from the end than from the longitudinal faces.
Hence the ends shrink at a different rate from
the interior parts.
195. Effects of Shrinkage. In a timber, the width AB (Fig. 103, X)
may have shortened (Fig. 21, Y), while a short distance from the end cd
the original width is still preserved. This should produce a bending of
the parts toward the centre of the piece as shown in exaggeration at Y, but
the rigidity of the several parts of the t.i-.iber prevents such bending and
the consequent strain leads to their separation as shown at Z, the end sur-
face of the timber being " checked."
As the timber dries out, the line cd becomes shorter, the parts 1 to 6
are allowed to approach again, and the checks close up and are no longer
visible.
The faster the drying at the surface, the greater is the difference in the
moisture of the different parts, and hence the greater the strains and con-
sequently also the amount of checking. This becomes very evident when
fresh wood is placed in the sun, and still more in a hot kiln. While most
of these smaller checks are thus only temporary, closing up again, some
large radial checks remain and even grow larger as drying progresses.
Their cause is a different one and will presently be explained.
The temporary checks not only occur at the ends, but are developed on
the sides also, only to a much smaller degree. They become especially an-
FIG.
103. Formation of
Checks.
230
THE MATERIALS -OF CONSTRUCTION.
noying on the surface of thick planks of hard woods, and also on peeled
logs when exposed to the sun.
So far we have considered the wood
as if made up only of parallel fibres all
placed longitudinally in the log. This,
however, is not the case. A large part of
the wood is formed by the medullary or
pith rays. In pine over 15,000 of these
occur on a square inch of a tangential sec-
tion, and even in oak the very large rays,
which are readily visible to the eye, repre-
sent scarcely a hundredth part of the num-
ber which the microscope reveals.
As seen in Fig. 104, the cells of these
rays have their length at right angles to
the direction of the wood-fibres.
If a large pith-ray of white oak is
whittled out and allowed to dry, it is
... / ' .
found to shrmk S^^J m the direction
from c to d (Fig. 104), while, as we have
stated, the fibres to which the ray is
firmly grown in the wood do not shrink in the same direction. There-
fore, in the wood, as the cells of the pith-ray dry, they pull on the longitu-
dinal fibres and try to shorten them, and, being opposed by the rigidity of
the fibres, the pith-ray is greatly strained. But this is not the only strain
it has to bear. Since the fibers from a to b (Fig. 104) shrink as much
again as the pith-ray in this, its longitudinal direction, the fibres tend to
shorten the ray, and the latter, in opposing this, prevents the former from
shrinking as much as they otherwise would. Thus the structure is sub-
jected to severe strains at right angles to each other; and herein lies the
greatest difficulty of wood-seasoning, for whenever the wood dries rapidly
these fibres have not the chance to ' ' give " or accommodate themselves,
and hence fibres and pith-rays separate and checks result which, whether
visible or not, are detrimental in the use of the wood.
The contraction of the pith-rays parallel to the length of the board is
probably one of the causes of the small amount of longitudinal shrinkage
which has been observed in boards.* The smaller shrinkage of the pith-
rays along the radius of the log (the length of the pith-ray) opposing the
shrinkage of the fibres in this direction becomes one of the causes of the
second great trouble in wood-seasoning, namely, the difference in the
FIG. 104. Small Pith-ray m Oak. a,
6, wood-fibres; c, d, cells of pith-
rav
* In addition to this all fibres having an oblique position, as those at pith-rays and
knots, also the oblique, tapering ends^all fibers, contribute to this longitudinal shrink-
age, since one component of their^rrrm) shrinkage is longitudinal.
TIMBER.
231
FIG. 105. Effects
of Shrinkage.
amount of the shrinkage along the radius and that along the rings or
tangent.
This greater tangential shrinkage appears to be due,
in part, to the cause just mentioned, but also to the fact
that the greatly shrinking bands of summer wood are
interrupted, along the radius, by as many bands of porous
spring wood, while they are continuous in the tangential
direction. In this direction, therefore, each such band
tends to shrink, as if the entire piece were composed of
summer wood; and since the summer wood represents the
greater part of the wood-substance, this tendency of
greater tangential shrinkage prevails.
The effect of this greater tangential shrinkage affects
every phase of woodworking. It lead to permanent
checks, and causes the log to split open on drying.
Sawed in two, the flat sides of the log become convex,
as in Fig. 105; sawed into a timber, it checks along the
median line of the four faces, and if converted into
boards the latter take on the forms shown in Fig. 105,
all owing to the greater tangential shrinkage of the wood.
Briefly, then, shrinkage of wood is due to the fact
that the celL walls grow thinner on drying. The thicker cell-walls and there-
fore the heavier wood shrinks most, while the water in the cell-cavities does
not influence the volume of the wood. Owing to the great difference of
cells in shape, size, and thickness of walls, and still more in their arrange-
ment, shrinkage is not uniform in any kind of wood. This irregular defor-
mation produces stresses which grow with the difference between adjoining
cells and are greatest at the pith-rays. These deformations cause warping
and checking, and exist even when no outward signs are visible; they are
greater if the wood is dried rapidly than if dried slowly, but* can never be
entirely avoided.
Temporary checks are caused by the more rapid drying of the outer parts
of any stick; permanent checks are due to the greater shrinkage, tangen-
tially, along the rings than that along the radius. This, too, is the cause of
most of the ordinary phenomena of shrinkage, such as the difference in
behavior of entire and quartered logs, "bastard" (tangent) and "rift"
(radial) boards, etc., and explains many of the phenomena erroneously attrib-
uted to the influence of the bark, or of the greater shrinkage of outer and
inner parts of any log.
Once dry, wood may be swelled again to its original size by soaking in
water, boiling, or steaming. Soaked pieces, on drying, shrink again as
before; boiled and steamed pieces do the same, but to a slightly less degree.
Neither hygroscopicity, i.e., the capacity of taking up water, nor shrinkage of
Wood can be overcome by drying at temperatures below 200-F. Higher
232 THE MATERIALS OF CONSTRUCTION.
temperatures, however, reduce these qualities, but nothing short of a coaling
heat robs wood of the capacity to shrink and
swell. Rapidly dried in the kiln, the wood of
oak and other hard woods " caseharden," that
is, the outer part dries and shrinks before the
interior has a chance to do the Fame, and thus
FIG. 106. " Honeycombed " f orms a fi rm s } ie n or case O f shrunken, corn-
Board The checks or cracks mo]lly checked wood around the interioi , This
form along the pith-rays. in-, -i , <.
shell does not prevent the interior irorn dry-
ing, but when this drying occurs, the interior is commonly checked along
the medullary rays, as shown in Fig. 106. In practice this occurrence can be
prevented by steaming the lumber in the kiln, and still better by drying the
wood in the open air or in a shed before placing in the kiln. Since only
the first shrinking is apt to check the wood, any kind of lumber which has
once been air dried (three to six months for 1-inch stuff) may be subjected
to kiln-heat without any danger. Kept in a bent or warped condition during
the first shrinking, the wood retains the shape to which it was bent and
firmly opposes any attempt at subsequent straightening.
196. Amount of Shrinkage in Timber. Sapwood, as a rule, shrinks more
than heartwood of the same weight, but very heavy heartwood may shrink
more than lighter sapwood. The amount of water in wood is no criterion of
its shrinkage, since in wet wood most of the water is held in the cavities,
where it has no effect on the volume.
The wood of pine, spruce, cypress, etc., with its very regular structure, dries
and shrinks evenly and suffers much less in seasoning than the wood of
broad-leaved trees. Among the latter, oak is the most difficult to dry with-
out injury. Small-sized split ware and "rift" boards season better than
ordinary boards and planks.
To avoid "working" or warping and checking, all high-grade stock is
carefully seasoned, preferably in a' kiln, before manufacture. Thicker pieces
may be made of several parts glued together; larger surfaces are made in
panels or of smaller pieces covered with veneer. Boring is sometimes resorted
to to prevent the checking of wooden columns.
Since repeated swelling increases the injuries due to seasoning, wood
should be protected against moisture when once it is dry.
Since the shrinkage of our woods has never been carefully studied, and
since wood, even from the same tree, varies within considerable limits, the
figures given in the following table are to be regarded as mere approximations.
The shrinkage along the radius and that along the tangent (parallel to the
rings) are not stated separately in the following table, and the figures repre-
sent an average of the shrinkage in the two directions. Thus, if the shrink-
age of soft pine is given at 3 inches per hundred, it means that the sum of
radial and tangential shrinkage is about 6 inches, of which about 4 inches fall
TIMBER. 233
to the tangent and 2 inches to the radius, the ratio between these varying
from 3 to 2, a ratio which practically prevails in most of our woods.
Since only an insignificant longitudinal shrinkage takes place (being
commonly less than 0. 1 inch per hundred, though in oak it is much more), the
change in volume during drying is about equal to the sum of the radial and
tangential shrinkage, or twice the amount of linear shrinkage indicated in
the table.
Thus, if the linear average shrinkage of soft pine is 3 inches per hun-
dred, the shrinkage in volume is about 6 cubic inches for each 100 cubic
inches of fresh wood, or 6 per cent of the volume.
APPROXIMATE SHRINKAGE OF A BOARD, OR SET OF BOARDS, 100 INCHES
WIDE, DRYIXG IN THE OPEN AIR.
Lateral
Common Names of Species. Shrinkage
Inches.
(1) All light conifers (soft pine, spruce, cedar, cypress) 3
(2) Heavy conifers (hard pine, tamarack, yew), honey-locust, box-elder, wood
of old oaks 4
(3) Ash, elm, walnut, poplar, maple, beech, sycamore, cherry, black locust.. 5
(4) Basswood, birch, chestnut, horse chestnut, blue beech, young locust 6
(5) Hickory, young oak, especially red oak Up to 10
MECHANICAL PROPERTIES OF WOOD.*
197. General View. Every joist and studding, every rafter, sash, and
door, the chair we sit on, the floor we walk on, the wood of the wagon or
boat we ride in, are all continually tested as to their stiffness and strength,
their hardness and toughness. Every step from the simple splitting of a
shingle or stave to the construction of the most elegant carriage or side-
board involves a knowledge not only of one, but of several, of the mechani-
cal properties of the material.
In the shop the fitness of the wood for a given purpose never depends on
any one quality alone, but invariably upon a combination of several quali-
ties. A spoke must not only be strong, it must be stiff to hold its shape, it
must be tough to avoid shattering to pieces, and it must also be hard or else
its tenons will become loose in their mortises.
Selecting wood in this way, the woodworker has learned almost all that
is at present known about his material; but in many cases the great diffi-
culty which always attends the judgment of complex phenomena has led to
erroneous conclusions, and not a few well-established beliefs have their origin
more in accidental errors of observation than in fact.
The experimenter endeavors to avoid this complexity by testing the
* This section of the Bulletin was made very simple for popular comprehension. It
seems somewhat out of place, therefore, in a scientific work, but is retained here for the
sake of completeness. J. B. J.
234 THE MATERIALS OF CONSTRUCTION.
wood for each kind of resistance separately; when tested as to their stiff-
ness, the pieces are all shaped, placed, and loaded alike. The wood is
selected with a definite object in view; it is green or dry, clear or knotty,
straight or cross-grained, according as he wishes to find out the influence of
each of these conditions. If pine and oak are to be compared, the pieces
are from the same position in the tree and are tried under exactly the same
'conditions, and thus the case is simplified.
But even results thus arrived at cannot be used indiscriminate!}^ and
the figures on the strength of oak given in any book must not be supposed
to apply to all oak if tested in the given manner. This is due to the fact
that a piece of wood is not simply a material but a structure, just as much
as a railroad-bridge or a balloon frame, and as such varies greatly even in
the wood of the same tree, nay, more than that, even in the same year's
growth of the same cross-section of a log.
A scantling resists bending; it is stiff. On removal of the load it
straightens; it is elastic. A column, a prop, or the spoke of a wagon-wheel
resists being crushed endwise. So does the upper side of a joist or beam
when loaded, while the under side of the beam or of an axe-handle suffers in
tension. The tenons of a window sash or of a door tend to break out their
mortises, the wood has to resist shearing along the fibres; the steel edge of
the eye tends to cut into the hammer-handle, it tries to shear it across- the
grain, and every nail, screw, bore-hole, or mortise tends to split the board
and tries the wood as to its cleavability, while all "bent" ware, from the
wicker basket to the one-piece felly or ship's knee, involves its flexibility.
198. Stiffness. If 100 pounds placed in the middle of a stick 2 by
inches and 4 feet long, supported at both ends, bend or "deflect" this
stick one eighth of an inch (in the middle), then 200 pounds will bend it
about one-fourth inch, 300 pounds three-eighths inch, the deflection varying-
directly as the load. Soon, however, a point is reached where an additional
100 pounds adds more than one-eighth inch to the deflection the limit of
elasticity has been reached. Taking another piece from the straight-grained
and perfectly clear plank of the same depth and width, but 8 feet long, the
load of 100 pounds will cause it to bend not only one-eighth inch, but will
deflect it by about 1 inch. Doubling the length reduces the stiffness eight-
fold. Stiffness then decreases as the cube of the length.
Cutting out a piece 2 by 4 inches and 4 feet long, placing it flatwise so
that it is double the width of the former stick, and loading it with 100
pounds, we find it bending only one-sixteenth inch: doubling the width
doubles the stiffness.
Setting the same 2 X 4-inch piece on edge, so that it is 2 inches wide
and 4 inches deep, the load of 100 pounds bends it only about one sixty-
fourth inch: doubling the thickness increases the stiffness about eightfold.
It follows that if we double the length and wish to retain the same
stiffness we must also double the thickness of the piece.
TIMBER. 235
A piece of wood is usually stiffer with the annual rings set vertically
than if the rings are placed horizontally to the load.
Cross-grained and knotty wood, to be sure, is not as stiff as clear lumber;
a knot on the upper side of a joist, which must resist in compression, is,
however, not so detrimental as a knot on
the lower side, where it is tried in tension,
Every large timber which comes from A
--~'
3
the central part of the tree contains knots,
and much of its wood is cut more or less ^^ ^
,.,. FIG. 107. Bending a Beam,
obliquely across the grain, both conditions
rendering such material comparatively less stiff than small clear pieces.
The same stick of pine green or wet is only about two thirds as stiff as
when dry. A heavy piece of long-leaf pine is stiffer than a light piece;
heavy pine in general is stiffer than light pine, but a piece of hickory,
although heavier than the pine, may not be as stiff as the piece of long-leaf
pine; and a good piece of larch exceeds in stiffness any oak of the same
weight.
( In the same tree stiffness varies with the weight, the heavier wood being
the stiffer; thus the heavier wood of the butt log is stiffer than that of the
top; timber with much of the heavy summer wood is stiffer than timber of
the same kind with less summer wood. In old trees (of pine) the centre of
the tree and the sap are the least stiff; in thrifty young pine the centre is
the least stiff, but in young second-growth hard woods it is the stiffest.
Since it is desirable, and for many purposes essential, to know before-
hand that a given piece with a given load will bend only a given amount,
the stiffness of wood is usually stated in a uniform manner and under the
term " modulus (measure) of elasticity."
If AB, Fig. 107, is a piece of wood, and d the deflection produced by a
weight or load, the stiffness of the wood, as usually stated, is found by the
formula
wr
Modulus of elasticity E = . , a ,[
where W is the weight, I the length, b and li the breadth and depth (height)
of the stick, and d the deflection for the load W In the following table
the woods are grouped according to their stiffness. The figures are only
rough approximations which are based on the data given in Vol. IX of the
Tenth Census. The first column contains the above modulus, the second
shows how many pounds will produce a deflection of 1 inch in a stick 1 by 1
by 12 inches, assuming that it could endure such bending within the limits
of elasticity, and the third column gives the number of pounds which will
bend a stick 2 by 2 inches and 10 feet long through 1 inch.
The stick is assumed to rest on both ends; if it is a cantilever, i.e , fas-
tened at one end and loaded at the other, it bears but half as much load
at its end for the same deflection.
236
TEE MATERIALS OF CONSTRUCTION.
Prom the third column it is easy to find how many pounds would bend
a piece of the same kind of other dimensions. A 2 X 4-inch bears eight,
a 2 X 6-inch twenty-seven times as much as the 2 X 2-inch; a piece 8 feet
long is about twice as stiff as a 10-foot piece; a piece 12 feet only about
three fifths, 14 feet one third, 16 feet two ninths, 18 feet one sixth, and 20
feet one eighth as stiff.
The number of pounds which will bend any piece of sawed timber by 1
inch may be found by using the formula
. , .
JNecessary weight =
~ ,
where E is the figure in the first column, and b, li, I, the breadth, depth, and
length of the timber in inches. If the deflection is not to exceed one-half
inch, only one half this load, and if one-fourth inch, only one fourth this loadj
is permissible; or, in general, W = ; ^ , where d is the deflection in inches.
CvO
TABLE OF STIFFNESS (MODULUS OF ELASTICITY) OF DRY WOOD.
GENERAL AVERAGES.
Species.
Modulus of
Elasticity
Wl*
Approximate Weight
which deflects by
Inch a Piece
4d6/i 3
per Square
Inch.
1 by 1 Inch
and
12 In. long.
2 by 2 In.
and
10 Ft. long.
Pounds.
62
51
40
37
(1) Live oak, good tamarack, long-leaf, Cuban, and short-
leaf pine, good Douglas spruce, Western hemlock,
yellow and cherry birch, hard maple, beech, locust,
and the best of oak and hickory ...
Pounds.
1,680,000
1,400,000
1,100,000
1,000,000
Pounds
3,900
3,200
2,500
2,300
(2) Birch, common oak, hickory, white and black spruce,
loblolly and red pine, cypress, best of ash, elm, and
poplar and black walutit
(3) Maples, cherry, ash, elm, sycamore, sweet gum, but-
ternut, poplar, basswood, white, sugar, and bull
pine, cedars, scrub pine, hemlock, and fir
(4) Box-elder, horse-chestnut, a number of Western soft
pines inferior grades of hard woods
199. Cross-breaking or Bending Strength. When the addition of 100
pounds to the load on our 2 X 2-inch piece begins to add more than one
eighth inch to the deflection, that is, when the stick has been- bent beyond
its "elastic limit/' it still requires an increase of 30 to 50 per cent to the
load before the stick breaks. The load which is borne before the limit of
elasticity is reached indicates the strength of the wood up to this important
point; the load which causes it to break represents its absolute strength, or
the "cross-breaking or bending strength" as it is commonly called.
In long-leaf pine the former (modulus of strength at the elastic limit)*
* The ;: elastic limit " in this case is somewhat of an arbitrary quantity, namely, the
point where 100 pounds produces a deflection 50 par cent greater than the first 100
pounds.
xx, 237
is commonly about three fourths of the latter. If left loaded for a consid-
time, a load even less than that which brings the stick to its elastic limit
will cause it to break, and this load should therefore not be reached in prac-
tice.
Unlike the stiffness, the strength of a timber varies approximately with
the squares of the thickness and decreases directly with increasing length and
not with the cube of this latter dimension. Thus if our piece 2 by 2 inches
and 4 feet long can bear 1000 pounds before it breaks, a 2 X 4-inch laid
flat will break with about 2000 pounds, and if set edgewise it requires about
4000 pounds to break it, while a piece of the same kind 2 by 2 inches and
double the length (8 feet) breaks with half the original load, or only 500
pounds.
All conditions of the material which influence the stiffness also influ-
ence the bending strength. Seasoning increases, moisture decreases, the
strength; knots and cross-grain depress it, and both are more dangerous on
the lower than on the upper side. But while the conifers with,their sim-
ple cell-structure excel in stiffness, the better hard woods develop the
greater strength in bending. Like elasticity and stiffness, the strength is
expressed in a uniform manner by the so-called "modulus of rupture," to
permit ready estimation of the strength of any given piece. This modulus
refers to the resistance per square inch which the parts most strained, " the
extreme fibre," offer. The figures usually tabulated are obtained by the
formula
3 TT7
Strength per square inch of extreme fibre = / = ^TTTJ
where W is the breaking-load, I the length, ~b and li the breadth and depth
of the tested piece of wood.
The following table presents our common woods grouped as to their
strength in bending. The load, as before, is supposed to act altogether in
the middle. Column 1 gives the strength of the extreme fibre, as explained
above; column 2, the number of pounds which will break a piece 1 by 1 inch
and 12 inches long; and column 3, the strength of a stick 2 by 2 inches and
10 feet long: from which the strength of any given piece can readily be esti-
mated, allowing, however, for defects, which increase with the size. Tims,
if a good piece of pine 2 by 2 inches and 10 feet long breaks with 400 pounds,
a 2 X 4-inch set on edge requires 1600 pounds, a 2 X 6-inch, 3600 pounds,
a 2 x 8-inch piece 6400 pounds to break it. If a piece 2 by 4 inches and
10 feet long breaks with 1600 pounds, a 2 X 4-inch and 12 feet long piece
breaks with about 1300 pounds, one 16 feet with 1000 pounds, etc. ; and if
a factor of safety of 10 is allowed, only one tenth of the above loads are
permissible.
A, board \ inch by 12 indies and 10 feet long contains as much wood
as a 3 x 2-inch of the same length, and if placed edgewise should offer
four times as much resistance to breaking. Owing to its small breadth,
238
THE MATERIALS OF CONSTRUCTION.
however, it "twists" when loaded, and in most cases, therefore, bears
less than the 2 x 3-inch. To prevent this twisting, joists are braced, and
the depth of timbers is made not to exceed four times their thickness.
Short deep pieces shear out or split before their strength in bending can
fully be called into play.
To allow for normal irregularities in the structure of wood itself, as well
as in the aggregate structure of timbers, an allowance is made on the num-
bers which have been found by experiment; this allowance is called the " fac-
tor of safety." Where the selection of the wood is not very perfect, the load
is a variable one, and the safety of human life depends on the structure, the
factor is usually taken quite high, as much as G or 10; i.e., only one sixth or
one tenth of the figures given iu the tables is considered safe, and the beam
is made six to ten times as heavy as the calculation requires.
STRENGTH
CROSS-BREAKING OF WELL-SEASONED SELECT PIECES.
Common Names of Species.
Strength of
the Extreme
Fibre
3 Wl
' ~ 2bli*
per Square
Inch.
Approximate Weight
which breaks a Stick
1 by 1 Inch
and 12 Inches
long.
2 by 2 Inches
and 10 Feet
long.
(1) Robinia (locust), hard maple, hickory, oak, birch,
best ash and elm, long-leaf, short- leaf, and
Pounds.
13 000
Pounds.
720
550
350
Pounds.
570
440
280
(2) Soft maple, cherry, ash, elm, walnut, iuferioi
oak and birch, best poplar, Norway, loblolly,
and pitch pines, black and white spruce, hem-
lock and tjood cedar . . . . .
10,000
6,500
(3) Tulip, basswood, sycamore, butternut, poplars,
white and other soft pines, firs, and cedars. . .
200. Tension and Compression. When'a piece of wood is pulled length-
wise, in the manner shown in Fig. 108, part of the fibres are torn asunder
or broken, but many are merely pulled or shredded out from between their
neighbors. Since failure in tension thus involves lateral adhesion as well
as strength of fibres, it is affected not only by the nature and dimensions
of the fibres, but also by their arrangement. Owing to their transverse
position the medullary rays (a large part of all woods) offer but one tenth
to one twentieth as much resistance as the main body of fibres, and more-
over weaken the timber by disturbing the straight course of the fibres and
the regularity of the entire structure.
The resistance is also much affected by the position of the grain. The
perfectly cross-grained piece a (Fig. 109) sustains but about one tenth to
one twentieth of the load which is supported by the straight-grained piece
c, and it is evident that the piece b, which represents an excessive degree
of cross -grain, is likewise weakened by the oblique position of the grain.
TIMBER.
239
This explains the detrimental influence of a knot on the under side of a
board, as in Fig. 110. Since the lower side of the board, in bending, is
stretched, the upper side being compressed, the fibres of the lower side are
subjected to tension, and the wood of the knot, like the piece of cross-grained
FIG. 109. Straight and Cross-
grained Wood.
FJG. 108.
Specimen in Tension Test.
FIG. 110. Effect of Knots and their
Position.
wood, offers but little resistance. Commonly the defect is greatly increased
by a season-check in the knot itself, so that the knot affects the strength
of the board like a saw-cut of equal depth, but to a less degree.
Tested in compression endwise (Fig. Ill), the fibres act as so many
hollow columns firmly grown together; and when the load becomes too
great the piece fails in the manner illustrated in Fig. 113.
This failure is a very complex phenomenon; in wood like
pine the fibres of the plane in which failure occurs become
separated into small bodies; they tear apart and cease to
behave as one solid body, but act as a large number of very
small independent pieces. Like the strands of a rope these
small bodies offer but little resistance to compression; they
bend over, and the piece " buckles."
It is evident that a vertical position and a regular ar-
rangement of the fibres increase the resistance, and that
therefore the medullary rays and oblique position of fibres
in cross-grained and knotty timber tend to reduce the
strength in compression.
From the following table of strength in tension and compression it will
be seen that these two are not always proportional, the stiffer conifers ex-
celling in the latter, the tougher hard woods in the former.
pressioii
wise.
end-
240
TEE MATERIALS OF CONSTRUCTION.
KATIO OF STRENGTH IX TENSION" AND COMPRESSION, SHOWING THE DIF-
FERENCE BETWEEN RIGID CONIFERS AND TOUGH HARD WOODS.
Name of Species.
Ratio :
tensile strength
K ; r
A Stick 1 Square Inch in Cross-
section.
Weight required to
compressive strength
Pull apart.
Crush endwise.
3.7
3.8
2.3
2.2
Pounds.
32,000
29,000
19,400
17,300
Pounds.
8,500
7,500
8,600
7,400
Elm
Larch
Long-leaf pine. ...
STRENGTH IN COMPRESSION OF COMMON AMERICAN WOODS IN WELL-
SEASONED SELECT PIECES.
(Approximate weight per square inch of cross-section requisite to crush a piece of wood endwise.)
(1) Black locust, yellow and cherry birch, hard maple, best hickory, long-leaf
and Cuban pines, and tamarack
(2) Common hickory, oak, birch, soft maple, walnut, good elm, best ash, short-
leaf and loblolly pines, Western hemlock, and Douglas fir
(3) Ash, sycamore, beech, inferior oak, Pacific white cedar, canoe cedar, Law-
sou's cypress, common red cedar, cypress, Norway and superior spruces,
and fir
(4) Tulip, basswood, butternut, chestnut, good poplar, white and other common
soft pines, hemlock spruce, and fir
(5) Soft poplar, white cedar, and some Western soft pines, and firs
Pounds.
9,000-h
7,000-H
6,000+>
5,000+
4,000-f-
201. Shearing. "When, in a structure like that shown in Fig. 11*2, a
weight is placed on / and the tenon T by downward pressure breaks out
the piece A BCD, this is said to shear out along the fibre. In the same
manner, if the shoulder A BCD in Fig. 112 is pushed off along BD, it is
sheared, and if BD and CE are each 1 inch, the surface thus sheared
FIG. 112. Longitudinal Shearing.
off is 1 square inch, and the weight necessary to do this represents the*
shearing- strength per square inch of the particular kind of wood. This
resistance is small when compared to that of tension and compression.
TIMBER.
241
116
1
! 4700
B
116
4
FIG. 113. Various Forms of Failure. A and B, compression endwise ; (7, shearing (the
bolt of a stirrup passed through the mortise and sheared out the end); D, tension.
The lower figure indicates the number of pounds per square inch which produced
the failure in tests by the Division of Forestry. No. 116 (upper figure on each
piece) is white pine. Nos. 1, 2, and 5 are long-leaf pine, about one fifth natural size.
242 THE MATERIALS OF CONSTRUCTION.
In general wet or green wood shears about one third more easily than
dry wood; a surface parallel to the rings (tangent) shears more easily than
one parallel to the medullary rays. The lighter conifers and hard woods
offer less resistance than the heavier kinds, but the best of pine shears one
third to one half more readily than oak or hickory, indicating that great
shearing strengh is characteristic of " tough " woods.
RESISTANCE TO SHEARING ALONG THE FIBRE.
Pounds per
Square Inch
(1) Locust, oak, hickory, elm, maple, ash, birch 1000 *
(2) Sycamore, long-leaf, Cuban, and short-leaf pine, and tam-
arack GOO
(3) Tulip, basswood, better class of poplar, Norway, loblolly
and white pine, spruce, red cedar . 400
(4) Softer poplar, hemlock, white cedar, fir . . 400 \
NOTE. Resistance to shearing, although a most important quality in wood, has not
been satisfactorily studied. The values in the above table, taken from various authors,
lack a reliable experimental basis and can be considered as only a little better than guess-
work. See Results of Forestry Division Tests in Chapter XXXII.
202. Influence of Weight and Moisture on Strength. It has been stated
that heavy wood is stronger than lighter wood of the same kind, and that
seasoning increases all forms of resistance. Let us examine why this is so.
Since the weight of dry wood depends on the number of fibres and the
thickness of their walls, there must be more fibres per square inch of cross-
section in the heavy than in the light piece of the same kind,J and it is but
natural that the greater number of fibres should also offer greater resistance,
i.e., have the greater strength.
The beneficial influence of drying and consequent shrinking is twofold:
(1) In dry wood a greater number of fibres occur per square inch, and (2)
the wood-substance itself, i.e., the cell-walls, become firmer. A piece of
green long-leaf pine, 1 by 1 inch and 2 inches long, is only about 0.94 by
0.96 inch and 2 inches long when dry; its cross-section is 10 per cent
smaller than before, but it still contains the same number of fibres. A
dry piece 1 by 1 inch, therefore, contains 10 per cent more fibres than a
green piece of the same size, and it is but fair to suppose that its resistance
or strength is also about 10 per cent greater.
The influence of the second factor, though unquestionably the more
important one, is less readily measured. In 100 cubic inches of wood-sub-
stance the material of the cell-walls takes up about 50 cubic inches of water
aud thereby swells up, becoming about 150 cubic inches in volume.. In
keeping with this swelling the substance becomes softer and less resistant.
* Over. f Less than.
^ This imperfect assumption is used only for comparison.
TIMBER. 243
In pine wood this diminution of resistance, according to experiments*
seems to be about 50 per cent, and the strength of the substance therefore
is inversely as the degree of saturation or solution.
203. Hardness. Heavy wood is harder than lighter wood; the wood of
the butt, therefore, is harder than that of the top, the darker summer wood
harder than the light-colored spring wood. Moisture softens, and season-
ing, therefore, hardens wood. Wood is much harder when pressed longi-
tudinally than when pressed transversely to the fibres, and it is somewhat
stronger tangentially than radially. Though harder wood resists saw and
chisel more than softer wood, the working quality of the wood is not always
a safe criterion of its hardness.
The following indicates the hardness of our common w r oods:
1. Very hard woods requiring over 3200 pounds per square inch to pro-
duce an indentation of one-twentieth inch: Hickory, hard maple, osage
orange, black locust, persimmon, and the best of oak, elm, and hackberry.
2. Hard woods requiring over 2400 pounds per square inch to produce
an indentation of one-twentieth inch: Oak, elm, ash, cherry, birch, black
walnut, beech, blue beech, mulberry, soft maple, holly, sour gum, honey-
locust, coffee-tree, and sycamore.
3. Moderately hard woods, requiring over 1600 pounds per square inch
to produce an indentation of one-twentieth inch: The better qualities of
Southern and Western hard pine, tamarack and Douglas spruce, sweet gum,
and the lighter qualities of birch.
4. Soft woods requiring less than 1600 pounds per square inch to pro-
duce an indentation of one-twentieth inch V The greater mass of conifer-
ous woods; pine, spruce, fir, hemlock, cedar, cypress, and redwood; poplar,
tulip, basswood, butternut, chestnut, buckeye, and catalpa.
204. Cleavability. When an axe is struck into a piece of wood as shown
in Fig. 114, the cleft projects beyond the blade of the axe and the process is
not one of cutting, but of tension across the grain.
The axe presses on a lever, a~b, while the surface in
which the transverse tension takes place is reduced
almost to a line across the stick at b. If the wood is
very stiff, the cleft runs far ahead of the axe, the
lever-arm ab is long, and the resistance to splitting
proportionately small. A high modulus of elasticity,
therefore, helps splitting, while- great shearing strength,
a good measure for transverse tension and hardness,
hinder it.
Wood splits naturally along two normal planes, the
most readily along the radius, because the arrange-
114. Cleavage, ment of fibres and pith-rays is radial, and next along
the tangent, or with the annual rings, because the softer spring wood forms
continuous planes in this direction. Cleavage along the radius, however,
244 THE MATERIALS OF CONSTRUCTION.
is from 50 to 100 per cent easier, and only in case of cross-grain, etc., the
cleavage along the ring becomes the easier. In the wood of conifers, wood-
fibres and pith-rays are very regular, the former in perfect radial series or
rows, and cleavage is, therefore, very easy in this direction. The same is
brought about in the oak by the very wide pith-rays, but where they are
thick and narrow, as in sycamore, and generally in the butt cuts and about
knots, they impede cleavage by causing a greater irregularity in the course
of the wood-fibres. The greater the contrast of spring and summer wood,
the easier the cleavage tangentially or in the direction of the rings. This is
especially marked in conifers and also in woods like oak, ash, and elm, where
the spring wood appears as a continous series of large pores. Very slow
growth influences tangential cleavage, narrow-ringed oak breaks out and
splits less regularly even in a radial direction ; in conifers, however, this
difference scarcely exists. Weight of wood affects the cleavage but little ; in
heavy wood the entrance of the axe, to be sure, is resisted with more force,
but the greater rigidity of the wood, on the other hand, counterbalances this
resistance. Irregularities in the course of the fibres, whether spiral growth,
cross-grain, or in form of knots, all aid in resisting cleavage. Knotty sticks
.are split more easily from the upper end, since the cleft then runs around the
knots (see Fig. 95). Moisture softens the wood and reduces lateral adhe-
sion, and therefore wood splits more easily when green than when dry.
205. Flexibility.* Pine is brittle, hickory is flexible ; the former
breaks, the latter bends. Being the opposite of stiffness, want of stiffness
would seem to indicate flexibility. This, however, is only partly true; hick-
ory and ash are stiff and yet among the most flexible of woods. Their
small dimensions cause shavings and thin strands of most woods to appear
pliable. For this reason the pliable, twisted wicker-willow is not a fair
measure of the flexibility of the wood of this species. Generally hard
woods are more flexible than conifers, wood of the butt surpassing in this
respect that of the main part of the stem, the latter being usually superior
to that of the limbs. Moisture softens wood and thereby increases its flexi-
bility. Knots and cross-grain diminish flexibility, but the irregular struc-
ture of elm, ash, etc. (particularly the arrangement of bodies of extremely
firm fibres, like so many strands, among the softer tissue, as well as the
interlacement of fibres due to post-cambial growth), favorably influences'
the flexibility of these woods.
206. Toughness, f So far the load by which the exhibition of the vari-
ous kinds of strength in compression, tension, cross-bending, etc., was pro-
* The writer here uses " flexibility " as Rankine uses "toughness," that is, the ability
to withstand great deformation before rupture. Flexibility, as the opposite of stiff-
ness or rigidity, would signify the readiness to deflect under a given load, which is
mathematically shown by a small modulus of elasticity. J. 13. J.
f The writer here uses toughness as indicating resilience, when this term is made to
upply to the whole period of deformation and not simply to the elastic field. J. B. J.
TIMBER. 245
duced has always been assumed as applied slowly and gradually. When a
wagon goes lumbering along a cobble pavement, the load on the spokes is
not thus applied. Every stone deals the wheel a blow, and a mile's-
journey means many thousand blows to every wheel-rim and spoke. In
chopping, the axe-handle is jarred, and a handle made of pine wood, which
shears easily along the fibre, would soon be shattered to pieces. Loads
thus applied are " shocks," and resistance to this form of loading requires
a combination of various kinds of strength possessed only by " tough "
woods. Toughness is a familiar word to woodworkers, and yet is rarely
defined. Tough wood must be both strong and pliable. Thus a willow is
not tough when dry; it is weak and brittle, and requires, notwithstanding
its small lateral dimensions, to be moistened and twisted or sheared inta
still smaller strands so that its fibres are subjected almost exclusively to ten-
sion if great deflection and great strength are to be combined (handles of
wicker baskets). Hickory is both strong and pliable ; in the dimensions
of a willow twig it can be used almost like a rope. The term " tough,"
therefore, is properly applied to woods like hickory and elm, and improp-
erly to willow.
Judging from the behavior of elm and hickory, wood may be pro-
nounced " tough " if it offers great resistance to
(1) Longitudinal shearing over 1000 pounds per square inch,
(2) Tension over 1G,000 pounds per square inch,
and permits, when tested dry, of an aggregate combined distortion in.
compression and tension amounting to not less than 3 per cent.
For instance, of a piece of dry hickory (//. alba) we may expect
Strength in shearing pounds 1,200
Strength in tension do. 25,000
Distortion in tension percent 2.03
Distortion in compression do. * 1.55
Total distortion do. 3.58
207. Practical Conclusions. From the foregoing considerations a few
valuable facts, mostly familiar to the thoughtful woodworker, may be
deduced
In framing, where light and stiff timber is wanted, the conifers excel;
where heavy but steady loads are to be supported, the heavier conifers, hard
pine, spruce, Douglas spruce, etc., answer as well as hard woods, which are
costlier and heavier for the same amount of stiffness. On the other hand,
if small dimensions must be used, and especially if moving loads are to be
sustained, hard woods are safest, and in all cases where the load is applied
in form of "shocks" or jars, only the tougher hard woods should be em-
ployed. The heavier wood surpasses the lighter of the same species in all
kinds of strength, so that the weight of dry wood and the structural fea-
246 THE MATERIALS OF CONSTRUCTION.
tures indicative of weight may be used as safe signs in selecting timber for
strength.
In shaping wood it is better, though more wasteful, to split than to saw,
because it insures straight grain and enables a more perfect seasoning.
For sawed stock the method of "rift" or "quarter" sawing, which has
so rapidly gained favor during the last decade, deserves every encourage-
ment. It permits of better selection and of more advantageous disposition
of the wood; rift-sawed lumber is stronger, wears better, seasons well, and
is least subject to "working" or warping.
All hardwood material which checks or warps badly during seasoning
should be reduced to the smallest practicable size before drying, to avoid
the injuries involved in this process; and wood once seasoned should never
again be exposed to the weather, since all injuries due to seasoning are
thereby aggravated. Seasoning increases the strength of wood in every
respect, and it is therefore of great importance to protect wooden struc-
tures bearing heavy weights against moisture.
Knots, like cross-grain and other defects, reduce the strength of timber.
Where choice exists, the knotty side of the joist should be placed upper-
most, i.e., should be used in compression.
Season-checks in timber are always a source of weakness; they are more
injurious on the vertical than on the horizontal faces of a stringer or joist,
and their effect continues even when they have closed up, as many do, and
are no longer visible.
Rafted timber, kiln-dried or steamed lumber are, as far as our present
knowledge extends, as strong as other kinds; and wherever any of these pro-
cesses aids in a more uniform or perfect seasoning, it increases the strength
of the material.
Pine ''bled" for turpentine is as strong as "unbled."
Time of felling, whether season of the year or phase of the moon, does
not influence strength, except that summer-felled hard wood rarely seasons
as perfectly as that felled in the fall, and to this extent an indirect influence
may be observed, as well as by the fact that fungi and insects have a better
opportunity for developing.
Warm countries and sunny exposures generally produce heavier and
stronger timber, and conditions favorable to the growth of the species also
improve its quality. But exceptions occur; neither fast nor slow growth is
an infallible sign -of strong wood, and it is the character of the annual ring,
rather than its width, and particularly the proportion of summer wood,
which determines the quality of the material.
CHEMICAL PROPERTIES AND TECHNOLOGICAL PRODUCTS OF WOOD.
208. Chemical Composition. Wood dried at 300 F. is composed of over
99 per cent of organic and less than 1 per cent of inorganic matter; the
latter remains as ashes when wood is burned.
TIMBER. 247
Wood consists of a skeleton of cellulose, permeate^ by a mixture of
other organic substances, collectively designated by the word lignin, and
particles of mineral matter or ashes.
Cellulose is the common substance of which plant-cells form their cases
or walls; in flax, the entire fibre is almost pure cellulose, but the amount
' of cellulose obtained from wood, by the common processes, rarely exceeds
' one half of its dry weight. Cellulose is identical in composition with
starch, but unlike the latter it resists alcoholic fermentation, though the
plants themselves, as well as decay-producing fungi, are able to reconvert
it into starch, from which it seems originally derived, and also to change it
into various forms of sugar.* Lignin is as yet a chemical puzzle. The
| substances forming it are carbohydrates like cellulose itself, but of slightly
j different proportions and distinguished by greater solubility in acids, and
by other chemical properties.
In 100 pounds of wood (dried at 300 F.) and of cellulose the following
proportions are found :
Wood, Cellulose,
Pounds. Pounds.
Carbon 49 44.4
Hydrogen G C. 1
Oxygen 44 49.3
This composition of wood is fairly uniform for different species.
At ordinary temperatures wood is a very stable compound; both in air
and under water it remains the same for centuries, and only when living
organisms attack it with their strono" solvents and convertants do change
and decay set in.
209. Wood as a Fuel. Heated to 300 F. wood gives off only water,
though some slight chemical changes are noticeable even at this tempera-
ture. If the heat is increased, gases of pungent odor and taste are evolved;
and if the temperature is sufficiently raised, the gases may be ignited, form-
ling the flame of the fire, while the remaining solid part glows like an ignited
: charcoal, giving much heat, but no flame. The amount of heat produced
by wood varies. If first dried at 300 F., 100 pounds of poplar wood should
give as much heat as 100 pounds of hickory. In the natural state, however,
this is not the case.
The beneficial effect of thorough seasoning for firewood appears from
the following consideration :
One hundred pounds of wood as sold in the wood-yards contains in round
numbers 25 pounds of water, 74 pounds of wood, and 1 pound of ashes.
The 74 pounds of wood are composed of 37 pounds of carbon, 4.4 pounds
of hydrogen, and 32 pounds of oxygen.
. f
* Chemists have succeeded in producing reconversion into grape-sugar; and though
the methods thus far employed are expensive, it is to be expected that in the near future
Nvood will become the principal source of both vinegar and alcohol.
248 THE MATERIALS OF CONSTRUCTION.
In burning (which is a process of oxidation) 4 pounds of hydrogen are
already combined with 32 pounds of oxygen, and there are only the 37 pounds
of carbon and 0.4 pound of hydrogen available in heat-production. Thus
only about one half the weight of the wood-substance itself is heat-produc-
ing, while every pound of water combined in the wood requires about 600
units of heat to evaporate it, and thus diminishes the value of the wood as
fuel. Hence under the most favorable circumstances 100 pounds of green
wood (50 percent moisture) furnishes about 270,000 units * of heat; 100
pounds of half-dry (30 per cent moisture) about 410,000 units; 100 pounds
of air-dry (20 per cent moisture) about 500,000 units; 100 pounds of air-
dry (10 per cent moisture) about 580,000 units; 100 pounds of kiln-dry (
per cent moisture) about 630,000 units.
In the ordinary stove or other small apparatus the evil effect of moisture
in the wood is very much increased, since combustion is materially interfered
with.
One hundred pounds of ordinary charcoal furnishes 1, 200,000 units of
heat, but the same quantity of charcoal produced at a temperature of 2000
F. furnishes 140,000 units of heat.
Conifers and the lighter hard woods produce more flame, while the heavy
hard woods furnish a good bed of live coal and exceed the former by 25 to
30 per cent in production of heat with ordinary appliances.
210. Charcoal. Heated in a closed chamber or covered with earth, as in
charcoal-pits, the wood is prevented from burning and a variety of changes
occur, depending on the rate of heating. If the temperature is raised grad-J
ually so that the wood is heated several hours before a temperature of 600
F. is reached, the process is called dry distillation. In this process the wood
is destroyed. It forms at first "red" or "brown" coal, still resembling
wood, and finally charcoal proper. This coal is darker, heavier, conducts
heat and electricity better, requires a greater heat to ignite, and produces
more heat per pound in burning the higher the temperature under which iti
is formed.
One hundred pounds of wood (dried at 300 F.) leaves only about 3C
pounds of charcoal. In common practice much less charcoal (18 to 20 peil
cent) is produced. In this change from wood to coal the volume is dimin-
ished by one half, so that a cord of wood which contains about 100 cubic ;
feet of wood solid would be converted into 50 cubic feet at best.
211. Products of Wood-distillation. Of the 70 pounds of gaseous prod
nets which 100 pounds of wood lose, during coaling, in being heated up t<|
700 F., about 63 pounds become volatile before the temperature of 550 F
is reached.
If condensed in a cooler, about three fourths of the 63 pounds of vol
* A unit of heat in this case is the amount of heat which raises the temperature c
1 pound of water 1 F.
TIMBER. 249
atile matter first evolved is found to be wood -vinegar, from which about
4 pounds of pure acetic acid, the only source of perfectly pure vinegar, is
obtained. Besides acetic acid, the liquid contains wood-spirits and a quan-
tity of various allied substances.
After the first stage of dry distillation, a large part of the products devel-
oped cannot be liquefied in the ordinary cooler. They are gases like the
illuminating-gas, mostly belonging to the marsh-gas series; they lack oxy-
gen and thus show ftiat the available oxygen has been nearly exhausted in
the preceding part of the process. Products of the latter stages are tars
and heavy oils, volatile only at high temperatures. Here also belong the
substances, known collectively as wood-creosote, employed as antiseptics in
wood -impregnation.
212. Cellulose. Warmed in dilute nitric acid with a little chlorate of
potash, the cells of a piece of wood may be separated. Each cell remains
intact, but its wall is reduced in thickness and material ; the lignin substances
being dissolved out, only the cellulose is left. In commercial-cellulose
manufacture, soda, sulphates, and of late chiefly sulphites are substituted
for the nitric acid. The wood is chipped, boiled in the respective solution
under high pressures, the residue is washed, and the remaining cellulose
bleached and ready for use. As a matter of economy the residual liquid is
evaporated and the soda used over again.
213. Resin, Turpentine, and Lampblack. When resinous wood, "fat
pine," "lightwood," such as the knots and stumps of long-leaf, pitch, and
other pines, is heated in a kiln or retort, the resins ooze out, are collected,
and in distillation with steam yield turpentine and resin. The resins and
their components vary with the species; the balsam of fir is limpid, its tur-
pentine remains clear on exposure; the resin of pines is very viscid, their
turpentines readily oxidize and darken when brought in contact with air.
Resins are gathered more commonly either from cracks, such as " wind " and
"ring shakes," as in the case of larch and fir (Venetian turpentine), or else
from wounds made especially for this purpose, as in the case of naval stores
gathered from pines. This latter process is known as " bleeding," " tapping,"
or " orcharding," and is at present the principal method of obtaining tur-
pentines 'and resins.
On burning resinous wood, wood-tar, etc., in a smouldering fire, soot is
deposited on the walls and partitions of the specially constructed soot-pit.
It is then collected, but must be freed of various products of dry distilla-
tion, by carefully heating to red heat, before it becomes the lampblack used
in printer's ink and otherwise much employed in the arts.
214. Tannin. Many kinds of wood and the bark of most trees contain
tannin. To serve in tanning the bark must contain at least 3 per cent of
tannin; the kind's mostly used vary from 5 to 15 per cent, and even the best
probably never furnish over 20 per cent in the average. The use of tan-
bark involves many disadvantages. It is difficult to dry and preserve, very
250 THK MATERIALS OF CONSTRUCTION.
liable to mould, it is bulky and therefore expensive to ship and store, and
very variable in the amount of tannin which it contains.
To avoid these difficulties the tannic compounds are, in recent times,
leached out of the finely ground bark and wood, condensed by evaporation,
and shipped as extracts containing 20 to 40 per cent of tannin.
The manufacture of pulp, and the production of fibre capable of being
spun and woven, are also technological uses of wood Which rely partly upon
chemical reactions.
DURABILITY AND DECAY OF WOOD.
215. All Decay Produced by a Fungus-growth. All wood is equally
durable under certain conditions. Kept dry or submerged, it lasts indefi-
nitely. Pieces of pine have been unearthed in Illinois which have lain
buried 60 or more feet deep for many centuries. Deposits of sound logs of
oak, buried for unknown ages, have been unearthed in Bavaria; parts of the
piles of the lake-dwellers, driven more than two thousand years ago, are
still intact.
On the radial section of a piece of pine timber, with one of the shelf -like
fungus-growths, as shown in Fig. 115, both bark and wood are seen to be
affected. A small particle of the half-decayed wood presents pictures like
that of Fig. 116. Slender, branching threads are seen to attach themselves
closely to the walls of the cells, and to pierce these in all directions. Thus
these little threads of fungus mycelium soon form a perfect network in the
wood, and as they increase in number they dissolve the walls, and convert
the wood-substance and cell-contents into sugar-like food for their own con-
sumption. In some cases it is the woody cell-wall alone that is attacked.
In other cases they confine themselves to eating up the starch found in the
cells, as shown in Fig. 117, and merely leave a stain (bluing of lumber). In
all cases of decay we find the vegetative bodies, these slender threads of
fungi, responsible for the mischief. These fine threads are the vegetative
body of the fungus; the little shelf is its frniting-body, on which it pro-
duces myriads of little spores (the seeds of fungi). Some fungi attack only
conifers, others hard woods ; many are*confined to one species of tree, and
perhaps no one attacks all kinds of wood. One kind produces " red rot,"
others " bluing." In one case the decayed tracts are tabular, and in the
direction of the fibres the wood is " peggy." In other cases no particular
shapes are discernible.
Cutting off a disk of loblolly pine, washing it, and then laying it in a
clean, shady place in the sawmill, its sap wood will be foumd stained in a
few days. Nor is this mischief confined to the surface; it penetrates the
sapwood of the entire disk. From this it appears that the spores must have
been in the air about the mill, and also that their germination and the
growth of the threads or mycelium are exceedingly rapid. (Watching the
TIMBER.
progress of mould on a piece of bread teaches the same thing.) Placing a
fresh piece of sapwood on ice, another into a dry kiln, and soaking a few
others in solutions of corrosive sublimate (mercuric chloride) and other sim-
ilar salts, we learn that the fungus-growth is retarded by cold, prevented
and killed by temperatures over 150 F., and that salts of mercury, etc.,
have the same effect. The fact that seasoned pieces if exposed are not so
FIG. 115.
PIG. 115. "Shelf" Fungus on the Stem of a Pine. (Hartig.) a, sound wood; b,
resinous "light" wood; c, partly decayed wood or punk; d, layer of living spore-
tubes; e, old flllcd-up spore-tubes; /, fluted upper surface of the fruiting-body of
the fungus, which gets its food through a great number of fine threads (the my-
celium), its vegetative tissue penetrating the wood and causing its decay.
FIG. 116. Fungus-threads iu Pine Wood. (Hartig.) a, cell-wall of the wood-fibres ;
b, bordered pits of these fibres; c, thread of mycelium of the fungus; d, holes in
the cell-walls made by the fungus-lhreads, which gradually dissolve the walls as
shown at e, and thus break down the wood-structure.
readily attacked by fungi shows that the moisture in air-dried wood is insuf-
ficient for fungus-growth.
From this it appears that warmth, preferably between 60 and 100 F.,
combined with abundance of moisture (but not immersion), is the most
important condition favoring decay, and that the defence lies in the proper
regulation or avoidance of these conditions, or else in the use of poisonous
salts, which prevent the propagation of fungi.
252
THE MATERIALS OF CONSTRUCTION.
It is also apparent, therefore, why wood decays faster in Alabama than
in Wisconsin, faster in the swamps than on the plains, and why the presence
of large quantities of decaying wood about the yard, constantly producing
fresh supplies of spores, stimulates decay. Covering with tar or impregnat-
ing with creosote, salts of mercury, cop-
per, etc., enables even sapwood to last
under the most trying conditions. Con-
tact with the ground assures most favor-
able moisture conditions for fungus-
growth, and the higher temperatures
near the surface of the ground, together
with the ever-present supply of spores,
cause rot in a post to start at the sur-
face more readily than 30 inches below.
216. Prevention of Decay. The
use of means to prevent decay is there-
fore desirable where timber is placed in
positions favorable to fungus-growth,
as in railway ties ; and all joists and tim-
ber in contact with damp brick walls, as
also all building material whose perfect
seasoning is prevented by the absence
of proper circulation of air, should be
specially protected. In the former cases
it is economy to apply preservative proc-
esses; in the latter a sanitary necessity.
Wood covered with paint, etc., before
it is perfectly seasoned falls a prey to
"dry-rot " ; the fungus finds abundance
of moisture, and the protection intended
FHJ. 117. Cells of Maple-wood attacked for the wood protects its enemy, the
by Fungus-threads (Neciria cinna- fungus. Since charcoal resists the sol-
bnrina Mayer). Section of three vents of fungi, charring the outer parts
.wood-fibres showing the threads of the of pogts ma k es? jf we n done, namely, so
fungus branching in their cavities and ^ uot t() checks .^ ^ ^^
consuming the starch stored in these
cells, a, interior or cavity of cells ; b, of tlie wood > a ?J fine protection,
threads of the fungus ; c, partly de- Under ordinary circumstances, only
stroyed starch-grains; d, dead por- the second great factor of decay, i.e.,
tions of the fungus-thread together the mo isture condition, can be con-
witli debris ; e, holes bored by the
fungus through the cell-walls ; S,
starch-grains just being attacked.
trolled.
Perfect seasoning, preferably kiln-
drying, before using, and protection
against the entrance of moisture by tar, paints, and other covers, when put
in place, prolong the life of wooden structures. Where such a covering is
too expensive, good ventilation at least is necessary. Contact-surfaces,
TIMBER. 253
where timber rests on timber or brick, should in all cases be especially pro-
tected.
Different species differ in their resistance to decay. Cedar is more
durable than pine, and oak better than beech; but in most cases the condi-
tions of warmth and moisture in particular locations have so much to do
with durability that often an oak post outlasts one of cedar, even in the same
line of fence, and predictions of durability become mere guesswork.
Containing more ready-made food, and in forms acceptable to a great
number of different kinds of fungi, the sapwood is more subject to decay
than the heartwood, doubly so where the latter is protected by resinous sub-
stances, as in pine and cedar. Several months of immersion improves the
durability of sapwood, but only impregnation with preservative salts seems
to render it perfectly secure. Once attacked by fungi, wood becomes pre-
disposed to further decay.
Wood cut in the fall is more durable than that cut in summer, only
because the low temperature of the winter season prevents the attack of the
fungi, and the wood is thus given a fair chance to dry. Usually summer-
felled wood, on account of prevalent high temperature and exposure to sun,
checks more than winter-felled wood ; and since all season-checks favor the
entrance of both moisture and fungus, they facilitate destruction. Where
summer-felled wood is worked up at once and protected by kiln-drying, no
difference exists. (The phases of the moon have no influence whatever on
durability!)
In sawing timber much of the wood is bastard-cut; at these places water
enters much more readily, and for this reason split and hewn timber and
ties generally resist decay perhaps better than if sawed.
The attacks of beetles, as well as those of the shipworm, cannot here be
considered ; like chisel or saw they are mechanical injuries against which
none of our woods are proof, except by impregnation of creosote or other
chemical.
RANGE OF DURABILITY IN RAILROAD-TIES.
Years.
Redwood 12
Cypress and red cedar 10
Tamarack 7 to 8
Long-leaf pine 6
Hemlock 4 to 6
Spruce 5
Years.
White oak and chestnut oak.
Chestnut...
Black locust 10
Cherry, black walnut, locust 7
Elm.., ,; 6 to V
Red and black oaks 4 to 5
Ash, beech, maple 4
The durability of wood exposed to the changes of the weather and
where painting, after thorough seasoning, is impracticable, is increased by
impregnating it with various salts or other chemicals which prevent the
fungus from feeding on the wood. The wood is first steamed, to open the
pores and remove the hardened surface coating of sap and dirt, and a liquid
solution of the preservative material is then injected with the assistance of
heat and pressure.
254 TEE MATERIALS OF CONSTRUCTION.
The most efficient fluids used on a large scale are bichloride of zinc and
creosote, or both combined. The " life " of railroad-ties is thereby increased
to twice and three times its natural duration.
HOW TO DISTINGUISH THE DIFFERENT KINDS OF WOOD.*,
217. An Examination of the Structure Essential to Identification. The
carpenter or other artisan who handles different woods becomes familiar with
those he employs frequently, and learns to distinguish them through this
familiarity, without usually being able to state the characteristic differences.
If a wood comes before him with which he is not familiar, he has, of course,
no means of determining what it is, and it is possible to select pieces even of
those with which he is well acquainted, different in appearance from the
average, that will make him doubtful as to their identification. Further-
more, he may distinguish between hard and soft pines, between oak and
ash, or between maple and birch, which are characteristically different; but
when it comes to distinguishing between the several species of pine or oak
or ash or birch, the absence of readily recognizable characteristics is such
that but few practitioners can be relied upon to do it. Hence in the
markets we find many species mixed and sold indiscriminately.
To identify the different woods it is necessary to have a knowledge of
the definite, invariable differences in their structure, besides that of the
often variable differences in their appearance. These structural differences
may either be readily visible to the naked eye or with a magnifier, or they
may require a microsocpical examination. In some cases such an examina-
tion cannot be dispensed with if we would make absolutely sure. There are
instances, as in the pines, where even our knowledge of the minute anatom-
ical structure is not yet sufficient to make a sure identification.
218. A Structural Key to Species. In the following key an attempt has
been made the first, so far as we know, in English literature to give a
synoptical view of the distinctive features of the commoner woods of the
United States which are found in the markets or are used in the arts. It
will be observed that the distinction has been carried in most instances no
further than to genera or classes of wood's, since the distinction of species
can hardly be accomplished without elaborate microscopic study, and also
that, as far as possible, reliance has been placed only on such characteristics
as can be distinguished with the naked eye or a simple magnify ing-glass, in
order to make the key useful to the largest number. Recourse has also-
been taken for the same reason to the less reliable and more variable general
external appearance, color, taste, smell, weight, etc.
The user of the key must, however, realize that external appearance,
such, for example, as color, is not only very variable, but also very difficult
to describe, individual observers differing especially in seeing and describing
* The matter in the remainder of this chapter is mostly the joint product of Dr.
B. E. Fernow and Mr. Filibert Roth.
TIMBER. 255
shades of color. The same is true of statements of size when relative and
not accurately measured, while weight and hardness can perhaps be more
readily approximated. Whether any feature is distinctly or only indistinctly
seen will also depend somewhat on individual eyesight, opinion, or practice.
In some cases the resemblance of different species is so close that only one
other expedient will make distinction possible, namely, a knowledge of the
region from which the wood has come. We know, for instance, that no
long-leaf pine grows in Missouri or Arkansas, and that no white pine can
come from Alabama, and we can separate the white cedar, giant arbor vit^e
of the West and the arbor \itse of the Northeast only by the difference of
the locality from which the specimen comes. With all these limitations
properly appreciated, the key will be found helpful toward greater familiar-
ity with the woods which are more commonly met with.
219. Characteristic Structural Features. The features which have been
utilized in the key and with which (their names as well as their appearance),
therefore, the reader must familiarize himself before attempting to use
the key, are mostly described as they appear in cross-section. They are:
(1) Sapwood and heart wood (see Art. 180), the former being the wood
from the outer, and the latter from the inner, part of the tree. In some
- A--
B C
FIG. 118. " Nou-porous" Woods. A, fir; B, " Lmrd " pine; C, soft pine, ar, annual
ring, o. e., outer edge of ring; i. e., inner edge of ring; s. w., summer wood; sp. w. t
spring wood; rd, resin-ducts.
cases they differ only in shade, and in others in kind of color, the heartwood
exhibiting either a darker shade or a, pronounced color. Since one cannot
always have the two together, or be certain whether he has sapwood or
heartwood, reliance upon this feature is, to be sure, unsatisfactory, yet
sometimes it is the only general characteristic that can be relied upon. If
farther assurance is desired, microscopic structure must be examined; in
such cases reference has been made to the presence or absence of tracheids in
pith-rays and the structure of their Avails, especially projections and spirals.
(2) Annual rings, their formation having been described in Art. 181.
(See also Figs. 118, 120.) They are more or less distinctly marked, and by
means of such marking a classification of three great groups of wood is
Dossible.
-*.
(3) Spring wood and' Summer wood, the former being the interior (first-
formed wood of the year), the latter the exterior (last-formed) part of the
256
THE MATERIALS OF CONSTRUCTION.
ring. The proportion of each and the manner in which the one merges into
the other are sometimes used, but more frequently the manner in which the
pores appear distributed in either.
FIG. 119. "Ring-porous" Woods White Oak and hickory, a. r., annual ring; su. w.
summer wood; sp. w., spring wood; v, vessels or pores; c. L, " concentric" lines; rt t
darker tracts of bard fibres forming tbe firm part of oak wood; pr, pitli-rays.
(4) Pores, which are vessels cut through, appearing as holes in cross-
section, in longitudinal section as channels, scratches, or indentations. (See
p. 213 and Figs. 119 and 120.) They appear only in the broad-leaved,
so-called, hard woods; their relative size (large, medium, small, minute, and
indistinct, when they cease to be visible individually by the naked eye) and
manner of distribution in the ring being of much importance, and especially
in the summer wood, where they appear singly, in groups, or short broken
lines, in continuous concentric, often wavy, lines, or in radial branching
lines.
(5) Resin-ducts (see p. 210 and Fig. 118), which appear very much like
pores in cross-section, namely, as holes or lighter or darker colored dots, but
much more scattered. They occur only in coniferous woods, and their
presence or absence, size, number, and distribution are an important dis-
tinction in these woods.
i Beech ' Sycamore ! RirftK i
FIG. 120. " Diffuse porous " Woods, ar, annual ring; pr, pith rays which are " broad "
at a, "fine " at b, " indistinct " at d.
(6) Pith-rays (see Art. 184 and Figs. 119 and 120), which in cross-
section appear as radial lines, and in radial section as interrupted bands of
varying breadth, impart a peculiar lustre to that section in some woods.
TIMBER. 257
They are most readily visible with the naked eye or with a magnifier in the
"broad-leaved woods. In coniferous woods they are usually so fine and
closely packed that to the casual observer they do not appear. Their breadth
and their greater or less distinctness are used as distinguishing marks, being
styled fine, broad, distinct, very distinct, conspicuous, and indistinct when
no longer visible by the naked (strong) eye.
(7) Concentric lines, appearing in the summer wood of certain species
more or less distinct, resembling distantly the lines of pores, but much finer
and not consisting of pores. (See Fig. 119.)
Of microscopic features, the following only have been referred to:
(8) Tracheids, a description of which is to be found in Art. 185.
(9) Pits, simple and bordered, especially the number of simple pits in
the cells of the pith-rays, which lead into each of the adjoining tracheids.
For standards of weight, consult table in Art. 191; for standards of
hardness, the classification in Art. 203.
Unless otherwise stated the color refers always to the fresh cross-section
of a piece of dry wood; sometimes distinct kinds of color, sometimes only
shades, and often only general color effects appear.
220. The Use of the Key. Nobody need expect to be able to use success-
fully any key for the distinction of woods or of any other, class of natural
objects without some practice. This is especially true with regard to woods,
which are apt to vary much, and when the key is based on such meagre
general data as the present. The best course to adopt is to supply one's self
with a small sample collection of woods accurately named.* Small, polished
tablets are of little use for this purpose. The pieces should be large enough,
if possible, to include pith and bark, and of -sufficient width to permit ready
inspection of the cross-section. By examining these with the aid of the
key, beginning with the better-known woods, one will soon learn to see the
features described and to form an idea of the relative standards which the
maker of the key had in mind. To aid in this, the accompanying illustra-
tions will be of advantage. When the reader becomes familiar with the key,
the work of identifying any given piece will be comparatively easy. The
material to be examined mast, of course, be suitably prepared. It should be
moistened ; all cuts should be made with a very sharp knife or razor and be
clean and smooth, for a bruised surface reveals but little structure. The
most useful cut may be made along one of the edges. Instructive, thin,
small sections may be made with a sharp penknife or razor, and when placed
on a piece of thin glass, moistened and covered with another piece of glass,
they may be examined by holding them toward the light.
Finding, on examination with the magnifier, that it contains pores, we
know it is not coniferous or nonporous. Finding no pores collected in the
spring-wood portion of the annual ring, but all scattered (diffused) through
* Hough's Wood Sections will be found both helpful and pleasing.- About one hun-
dred and fifty species of American woods are now so prepared by Mr. Romeyn Hough,
Lowville, N. Y. J. B. J.
258 THE MATERIALS OF CONSTRUCTION.
the ring, we turn at once to the class of " diffuse-porous woods." We now
note the size and manner in which the pores are distributed through th<=
ring. Finding them very small and neither conspicuously grouped, nor
larger nor more abundant in the spring wood, we turn to the third group of
this class. We now note the pith -rays, and finding them neither broad nor
conspicuous, but difficult to distinguish even with the magnifier, we at
once exclude the wood from the first two sections of this group and place it
in the third, which is represented by only one kind, cottonwood. Finding
the wood very soft, white, and on the longitudinal section with a silky lustre,
we are further assured that our determination is correct. We may now turn
to the list of woods and obtain further information regarding the occurrence,
qualities, and uses of the wood.
Sometimes our progress is not so easy; we may waver in what group or
section to place the wood before us. In such cases we may try each of the
doubtful roads until we reach a point where we find ourselves entirely wrong
and then return and take up another line; or we may anticipate some of the
later-mentioned features and, finding them apply to our specimen, gain
additional assurance of the direction we ought to travel. Color will often
help us to arrive at a speedy decision. In many cases, especially with con-
ifers, which are rather difficult to distinguish, a knowledge of the locality
from which the specimen comes is at once decisive. Thus, Northern white
cedar, and bald cypress, and the cedar of the Pacific will be identified even
without the somewhat indefinite criteria given in the key.
Engineers and architects can, in the case of the two leading kinds of
Southern pine (long-leaf, P. palustns, and short-leaf, P. ecliinata), usually
determine the species by learning with certainty where the lumber was
sawed. This is the more easy with large orders as these are filled directly
from the mills, and the shipping bills for the particular cars on which it is
delivered may be demanded. The two maps shown in Plates V and VI*
will serve to identify the species when the locality is known. It will be
seen at once that these two species do not often occupy the same territory.
221. KEY TO THE MORE IMPORTANT WOODS OF NORTH AMERICA.
[The numbers preceding names refer to the List of Woods following the Key.]
I. Non-porous Woods. Pores not visible or conspicuous on cross-section even
with magnifier. Annual rings distinct by denser (dark-colored) bunds of summer
wood (Fig. 118).
II. Ring-porous Woods. Pores numerous, usually visible on cross-section with-
out magnifier. Annual rings'distinct by a zone of large pores collected in the spring
wood, alternating with the denser summer wood (Fig. 119).
III. Diffuse-porous Woods. Pores numerous, usually not plainly visible on
cross-section without magnifier. Annual rings distinct by a fine line of denser
summer-wood cells, often quite indistinct ; pores scattered through annual ring, no
zone of collected pores in spring wood (Fig. 120).
* These maps are reduced from similar ones published by the Forestry Division of
the U. S. Agr. Dept. Washington, as Bulletin No. 13.
TIMBER.
NOTE. The above-described three groups are exogenous, i.e., they grow by add-
ing annually wood on their circumference. A fourth group is formed by the endog-
enous woods, like yuccas and palms, which do not grow by such additions.
I. NON-POROUS WOODS.
Includes all coniferous woods.
A. Resin-ducts wanting.*
1. No distinct heartwood.
a. Color effect yellowish white ; summer wood darker yellowish (under
microscope pith-ray without tracheids) (Nos. 9-13) FIRS.
&. Color effect reddish (roseate) (under microscope pith-ray with tra-
cheids) (Nos. 14 and 15) HEMLOCK.
2. Heartwood present, color decidedly different in kind from sapwood.
a. Heartwood light orange-red ; sapwood pale lemon ; wood heavy and
hard (No. 38) YEW.
6 Heartwood purplish to brownish red ; sapwood yellowish white ; wood
soft to medium hard light, usually with aromatic odor.
(No. 6) RED CEDAR.
c. Heartwood maroon to terra cotta or deep brownish red ; sapwood light
orange to dark amber, very soft and light, no odor ; pith-rays very
distinct, specially pronounced on radial section . . . (No. 7) REDWOOD.
3. Heartwood present, color only different in shade from sapwood, dingy-
yellowish brown.
a. Odorless and tasteless (No. 8) BALD CYPRESS.
6. Wood with mild resinous odor, but tasteless. .(Nos. 1-4) WHITE CEDAR.
c. Wood with strong resinous odor and peppery taste when freshly cut.
(No. 5) INCENSE-CEDAR.
ADDITIONAL NOTES FOR DISTINCTIONS IN THE GROUP.
Spruce is hardly distinguishable from fir, except by the existence of the resin-
ducts, and microscopically by the presence of tracheids in the medullary rays.
Spruce may also be confounded with soft pine, except for the hetirtwood color of the
latter and the larger, more frequent, and more readily visible resin-ducts.
In the lumber-yard hemlock is usually recognized by color and the slivery char-
acter of its surface. Western hemlocks partake of this last character to a less
degree.
Microscopically the white pine can be distinguished by having usually only one
large pit, while spruce shows three to five very small pits in the parenchyma-cells of
the pith-ray communicating with the tracheid.
The distinction of the pines is possible only by microscopic examination. The
following distinctive features may assist in recognizing, when in the log or lumber-
pile, those usually found in the market :
The light, straw color, combined with great lightness and softness, distinguishes
the white pines (white pine and sugar-pine) from the hard pines (all others in the
market), which may also be recognized by the gradual change of spring wood into
summer wood. This change in hard pines is abrupt, making the summer wood
appear as a sharply defined and more or less broad band.
The Norway pine, which may be confounded with the short-leaf pine, can be dis-
* To discover the resin-ducts a very smooth surface is necessary, since resin-ducts are frequently
seen only with difficulty, appearing on the cross-section as fine whiter or darker spots normally scat-
tered singly, rarely in groups, usually in the summer wood of the annual ring. They are often much
more easily seen on radial, and still more so on tangential, sections, appearing there as fine lines or
dots of open structure of different color, or as indentations or pin-scratches in a longitudinal direction.
260 THE MATERIALS OF CONSTRUCTION.
B. Resin-ducts present.
1. No distinct heartwood ; color white, resin-ducts very small, not numerous.
(Nos. 33-36) SPRUCE.
2. Distinct heartwood present.
a. Resin-ducts numerous, evenly scattered through the ring.
a'. Transition from spring wood to summer wood gradual ; annual
ring distinguished by a fine line of dense summer-wood cells ;
color white to yellowish red : wood soft and light.
(Nos. 18-21) SOFT PINES.*
&'. Transition from 1 spring wood to summer wood more or less
abrupt ; broad bands of dark-colored summer wood ; color
from light to deep orange ; wood medium hard and heavy.
(Nos. 22-32) HARD PINES.*
#. Resin-ducts not numerous nor evenly distributed.
a'. Color of heartwood orange-reddish, sapwood yellowish (same as
hard pine); resin-ducts frequently combined in groups of 8 to
30, forming lines on the cross-section (tracheids with spirals).
(No. 37) DOUGLAS SPRUCE.
V. Color of heartwood light russett-brown ; of sapwood yellowish
brown ; resin-ducts very few, irregularly scattered (tracheids
without spirals) (Nos. 16 and 17) TAMARACK.
II. KING-POROUS WOODS.
[Some of Group D and cedar-elm imperfectly ring-porous. J
A. Pores in the summer wood minute, scattered singly or in groups, or in short
broken lines, the course of which is never radial.
tinguished by being much lighter and softer. It may also, but more rarely, be
confounded with heavier white pine, but for the sharper definition of the annual
ring, weight, and hardness.
The long-leaf pine is strikingly heavy, hard, and resinous, and usually very reg-
ular and narrow-ringed, showing little sapwood, and differing in this respect from
the short-leaf pine and loblolly pine, which usually have wider rings and more sap-
wood, the latter excelling in that respect.
The following convenient and useful classification of pines into four groups,
proposed by Dr. H. Mayr, is based on the appearance of the pith-ray as seen in a
radial section of the spring wood of any ring :
Section I. Walls of the tracheids of the pith-ray with dentate projections.
a. One to two large, simple pits to each tracheid on the radial walls of the
cells of the pith-ray. Group 1. Represented in this country only by P.
resinosa.
b. Three to six simple pits to each tracheid, on the walls of the cells of the
pith-ray. Group 2. P. tceda, palustris, etc., including most of our
"hard " and " yellow " pines.
Section II. Walls of tracheids of pith-ray smooth, without dentate projections.
a. One or two large pits to each tracheid on the radial walls of each cell of
the pith-ray. Group 3. P. strobus, lainbertiana, and other true white
pines.
6. Three to six small pits on the radial walls of each cell of the pith-ray.
Group 4. P. parryana and other nut-pines, including also P. bal-
fouriana.
* Soft and hard pines are arbitrary distinctions, and the two are not distinguishable at the common
limit.
TIMBER.
261
1. Pith-rays minute, scarcely distinct.
a. Wood heavy and hard; pores in the summer wood not in clusters.
a'. Color of radial section not yellow (Nos. 39-44) ASH.
I'. Color of radial section light yellow; by which, together with its
hardness and weight, this species is easily recognized.
(No. 103) OSAGE ORANGE.
6. Wood light and soft; pores in the summer wood in clusters of 10 to 30,
(No. 56) CATALPA.
2. Pith-rays very fine, yet distinct; pores in summer wood usually single or
in short lines; color of heartwood reddish brown; of sapwood yellowish
white: peculiar odor on fresh section (No. Ill) SASSAFRAS.
3. Pith-rays fine, but distinct.
a. Very heavy and hard; heartwood yellowish brown.
(No. 77) BLACK LOCUST.
I. Heavy; medium hard to hard.
a. Pores in summer wood very minute, usually in small clusters of
3 to 8; heartwood light orange-brown.
(No. 83) RED MULBERRY.
b'. Pores in summer wood small to minute, usually isolated; heart-
wood cherry-red (No. 61) COFFEE-TREE.
4. Pith-rays fine, but very conspicuous, even without magnifier. Color of
heartwood red; of sapwood pale lemon (No. 78) HONEY-LOCUST.
ADDITIONAL NOTES FOR DISTINCTIONS IN THE GROUP.
Sassafras and mulberry may be confounded but for the greater weight and hard-
ness and the absence of odor in the mulberry; the radial section of mulberry also
shows the pith-rays conspicuously.
Honey-locust, coffee-tree, and black-locust are also very similar in appearance.
The honey-locust stands out by the conspicuousness of the pith-rays, especially on
radial sections, on account of their height, while the black locust is distinguished by
the extremely great weight and hardness, together with its darker brown color.
FIG. 121. Wood of Coffee-tree.
The ashes, elms, hickories, and oaks may, on casual observation, appear to
resemble one another on account of the pronounced zone of porous spring wood.
The sharply defined large pith-rays of the oak exclude these at once; the wavy lines
of pores in the summer wood, appearing as conspicuous finely-feathered hatchings
on tangential section, distinguish the elms; while the ashes differ from the hickory
by the very conspicuously defined zone of spring-wood pores, which in hickory
appear more or less interrupted. The reddish hue of the hickory and the more or
less brown hue of the ash may also aid in ready recognition. The smooth, radial
surface of split hickory will readily separate it from the rest.
262
THE MATERIALS OF CONSTRUCTION.
B. Pores of summer wood minute or small, in concentric wavy and sometimes
branching lines, appearing as finely-feathered hatchings on tangential
section.
1. Pith-rays fine, but very distinct; color greenish white. Heartwood absent
or imperfectly developed (No. 70) HACKBERRY.
2. Pith-rays indistinct; color of heartwood reddish brown; sapwood grayish
to reddish white (Nos. 62-66) ELMS.
C. Pores of summer wood arranged in radial branching lines (when very crowded
radial arrangement somewhat obscured).
1. Pith-rays very minute, hardly visible (Nos. 58-60) CHESTNUT.
2. Pith-rays very broad and conspicuous (Nos. 84-102) OAK.
D. Pores of summer wood mostly but little smaller than those of the spring wood,
isolated and scattered; very heavy and hard woods. The pores of the
spring wood sometimes form but an imperfect zone. (Some diffuse-porous
woods of groups A and B may seem to belong here.)
1. Fine concentric lines (not of pores) as distinct, or nearly so, as the very fine
pith-rays; outer summer wood with a tinge of red; heartwood light reddish
brown (Nos. 71-75) HICKORY.
2. Fine concentric lines, much finer than the pith-rays; no reddish tinge in
summer wood; sapwood white; heartwood blackish.. (No. 105) PERSIMMON.
ADDITIONAL NOTES FOR DISTINCTIONS IN THE GROUP.
A
B
FIG. 122. .1, Black Ash; .B, White Ash; G, Greeu Ash.
The different species of ash may be identified as follows:
1. Pores in the summer wood more or less united into lines.
a. The lines short and broken, occurring mostly near the limit of the ring.
(No. 39,) WHITE ASH.
b. The lines quite long and conspicuous in most parts of the summer
wood (No. 43) GREEN ASH.
2. Pores in the summer wood not united into lines, or rarely so.
a. Heartwood reddish brown and very firm. (No. 40) RED ASH.
b. Heartwood grayish brown and much more porous. .(No. 41) BLACK ASH.
TIMBER.
263
ADDITIONAL NOTES continued.
In the oaks two groups can be readily distinguished by the manner in which the
pores are distributed in the summer -wood. In the white oaks the pores are very fine
and numerous and crowded in the outer part of the summer wood, while in the black
or red oaks the pores are larger, few in number, and mostly isolated. The live oaks,
as far as structure is concerned, belong to the black oaks, but are much less porous,
and are exceedingly heavy and hard.
FJG. 123. Wood of Red Oak. (For White Oak see Fig. 119.)
FJG. 124. Wood of Chestnut.
FIG. 125. Wood of Hickory.
264
TEE MATERIALS OF CONSTRUCTION.
III. 1>J ^FUSE-POROUS WOODS.
[A few indistinctly ring-porous woo i of Group II, D, and cedar-elm may seem to belong here.]
A. Pores varying in size from large to minute; largest in spring wood, thereby giving
sometimes the appearance of a ring-porous arrangement.
1, Heavy and hard; color of heartwood (especially on longitudinal section)
chocolate-brown, (No. 116) BLACK WALNUT.
2. Light and soft; color of heartwood light reddish brown..(No. 55) BUTTERNUT.
B. Pores all minute and indistinct; most numerous in spring wood, giving rise to a
lighter-colored zone or line (especially on longitudinal section), thereby
appearing sometimes ring-porous; wood hard, heartwood vinous-reddish;
pith-rays very fine, but very distinct. (See also the sometimes indistinct
ring-porous cedar-elm, and occasionally winged elm, which are readily
distinguished by the concentric wavy lines of pores in the summer wood.)
(No. 57j CHERRY.
C. Pores minute or indistinct, neither conspicuously larger nor more numerous in
the spring wood and evenly distributed.
1. Broad pith-rays present.
a. All or most pith-rays broad, numerous, and crowded, especially on tan-
gential sections, medium heavy and hard, difficult to split.
(Nos. 112 and 113) SYCAMORE.
6. Only part of the pith-rays broad.
a'. Broad pith-rays well defined, quite numerous; wood reddish white
to reddish (No. 47) BEECH.
I'. Broad pith-rays not sharply defined, made up of many small rays,
not numerous. Stem furrowed, and therefore the periphery
of section, and with it the annual rings, sinuous, bending in
and out, and the large pith-rays generally limited to the fur-
rows or concave portions. Wood white, not reddish.
(No. 52) BLUE BEECH.
DDITIONAL NOTES FOR DISTINCTIONS IN THE GROUP.
Cherry and birch are sometimes confounded. The high pith-rays on the cherry on
radial sections readily distinguish it; distinct pores on birch and spring-wood zone
in cherry, as well as the darker vinous-brown color of the latter, will prove helpful.
Two groups of birches can be readily distinguished, though specific distinction is
not always possible.
1. Pith-rays fairly distinct, the pores rather few and not more abundant in the
spring wood; wood heavy, usually darker.
(No. 48) CHERRY-BIRCH and (No. 49) YELLOW BIRCH.
H
pr,.-
Beech ! Sycamore ! Birch -_-j
FIG. 126. Wood of Beech, Sycamore, aud Birch.
TIMBER.
265
2, No broad pith-rays present.
a. Pith-rays small to very small, but quite distinct.
a'. Wood hard.
a". Color reddish white, with dark reddish tinge in outer sum-
mer wood ........................ (Nos. 79-82) MAPLE.
I". Color white, without reddish tinge ........ (No. 76) HOLLY.
&'. Wood soft to very soft.
a". Pores crowded, occupying nearly all the space between
pith-rays.
a'". Color yellowish white, often with a greenish tinge
in heartwood ........... (No. 115) TULIP-POPLAR.
(No. 116) CUCUMBER-TREE.
V". Color of sapwood grayish, of heartwood light to
dark reddish brown ........ (No. 69) SWEET GUM.
6". Pores not crowded, occupying not over one third the pith-
rays; heartwood brownish white to very light brown.
(Nos. 45 and 46) BASSWOOD.
5. Pith-rays scarcely distinct, yet if viewed with ordinary magnifier
plainly visible.
a'. Pores indistinct to the naked eye.
a". Color uniform pale yellow; pith-rays not conspicuous even
on the radial section ......... (Nos. 53 and 54) BUCKEYE.
6". Sapwood yellowish gray, heartwood grayish brown; pith-
rays conspicuous on the radial section.
(Nos. 67, 68) SOUR GUM.
&'. Pores scarcely distinct, but mostly visible as grayish specks on the
cross-section ; sapwood whitish, heartwood reddish.
(Nos. 48-51) BIRCH.
3. Pith-rays not visible or else indistinct, even if viewed with magnifier.
1. W T ood very soft, white, or in shades of brown, usually with a silky lustre.
(NOS. 105-110) COTTONWOOD (POPLAR).
2. Pith-rays barely distinct, pores more numerous and commonly forming a
more porous spring-wood zone; wood of medium weight.
(No. 51) CANOE- OR PAPER-BIRCH.
The species of maple may be distinguished as follows:
1. Most of the pith-rays broader than the pores and very conspicuous.
(No. 79) SUGAR MAPLE.
FIG. 127.- Wood of Maple.
266
THE MATERIALS OF CONSTRUCTION.
ADDITIONAL NOTES continued.
2. Pith-rays not or rarely broader than the pores, fine but conspicuous.
a. Wood heavy and hard, usually of darker reddish color and commonly
spotted on cross-section (No. 80) RED MAPLE.
b. Wood of medium weight and hardness, usually light-colored.
(No. 82) SILVER MAPLE.
Red maple is not always safely distinguished from soft maple. In box-elder the
pores are finer and more numerous than in soft maple.
The various species of elm may be distinguished as follows:
1. Pores of spring wood form a broad band of several rows; easy splitting, dark
brown heart (No. 64) RED ELM.
2. Pores of spring wood usually in a single row, or nearly so.
a. Pores of spring wood large, conspicuously so (No 62) WHITE ELM.
b. Pores of spring wood small to minute.
a . Lines of pores in summer wood fine, not as wide as the interme-
diate spaces, giving rise to very compact grain,
(No. 63) ROCK-ELM.
V. Lines of pores broad, commonly as wide as the intermediate spaces.
(No. 66) WINGED ELM.
c. Pores in spring wood indistinct, and therefore hardly a ring-porous
wood (No. 65) CEDAR-ELM.
FIG. 128. Wood of Elm. , Red Elm; b, White Elm; c, Winged Elm.
FIG. 129. Walnut, p r., pith-rays;
c. L, concentric Hues; D, vessels or
pores; su. w., summer wood; sp. w.
spring wood.
FIG. 130. Wood of Cherry.
TIMBER.
267
LIST OF THE MORE IMPORTANT WOODS OF THE UNITED STATES.*
[Arranged alphabetically.]
NOTE. In the following descriptions the terms expressing size have been used
with the following meanings :
Small = trees of 50 feet high or less.
Medium = " " 50 to 100 feet high.
Large = " " over 100 feet in height.
All these terms must be understood as having been used as approximate estimates only.
A. CONIFEROUS WOODS.
Woods of simple and uniform structure, generally light, soft but stiff;
abundant in suitable dimensions and forming by far the greatest part of all
the lumber used.
222. Cedar. Light, soft, stiff, not strong, of fine texture; sap and heart-
wood distinct, the former lighter, the latter a dull grayish brown or red.
The wood seasons rapidly, shrinks and checks but little, and is very dur-
able. Used like soft pine, but owing to its great durability preferred for
shingles, etc. Small sizes used for posts, ties, etc. Cedars usually occur
scattered, but they form, in certain localities, forests of considerable extent.
a. White Cedars. Heartwood a light grayish brown.
1, WHITE CEDAR (Thuya occidentalis) (Arbor-
vitse): Scattered along streams and lakes, frequently
covering extensive swamps; rarely large enough for
lumber, but commonly used for posts, ties, etc.
Maine to Minnesota and northward.
FIG. 131. T. occidentalis.
2. CANOE-OEDAK (Thuya gigantea) (red cedar of
the West) : In Oregon and Washington a very large
tree, covering extensive swamps; in the mountains
much smaller, skirting the watercourses; an impor-
tant lumber tree. Washington to northern California
and eastward to Montana.
FIG. 182. T. gigantea.
* The text here is from U. S. Forestry Bulletin No. 10, while many of the cuts are
from Apgar's Trees of the Northern States. The remaining cuts have been specially
drawn for this work, under the direction of Dr. William Treiease, Director of the Mis-
souri Botanical Garden, St. Louis.
268
THE MATERIALS OF CONSTRUCTION
3. WHITE CEDAR (Chamcecyparis thy aides):
Medium-sized tree, wood very light and soft. Along
the coast from Maine to Mississippi.
FIG. 133. C. thy aides.
4. WHITE CEDAR (ChamcBcyparis lawsomana)
(Port Orford cedar, Oregon cedar, Lawson's cypress,
ginger-pine): A very large tree, extensively cut for
lumber; heavier and stronger than the preceding.
Along the coast-line of Oregon.
Fie*. 131. C. lawsoniana.
5. WHITE CEDAR (Libocedrus decur-
rens] (incense-cedar) : A large tree, abun-
dantly scattered among pine and fir>
wood fine-grained. Cascades and Sierra
Nevada of Oregon and California.
S K
FIG. 135. L. decurrens.
TIMBER.
269
b. Red Cedars. Heartwood red.
6. RED CEDAR (Juniperus virginiana) (Savin juniper) : Similar to
white cedar, but of somewhat finer texture. Used in cabinet work in
cooperage, for veneers, and especially for lead-pencils,
for which purpose alone several million feet are cut
each year. A small to medium-sized tree scattered
through the forests, or, in the West, sparsely covering
extensive areas (cedar-brakes). The red cedar is the
most widely distributed conifer of the United States,
occurring from the Atlantic to the Pacific and from _
Florida to Minnesota, but attains a suitable size for T
lumber only in the Southern, and more especially the FlG " ' 36 -~ t
Gulf, States.
ana.
7. REDWOOD (Sequoia sempervirens) : Wood in
its quality and uses like white cedar; the narrow
sapwood whitish; the heartwood light red, soon
turning to brownish red when exposed. A very
large tree, limited to the coast ranges of Cali-
fornia, and forming considerable forests, which
are rapidly being converted into lumber.
FIG. 137. 8. sempervirens.
223. Cypress.
8. CYPRESS (Taxodium disticlmm) (bald
cypress ; black, white, and red cypress) : Wood in
appearance, quality, and uses similar to white
cedar. " Black cypress " and " white cypress" are
heavy and light forms of the same species. The
cypress is a large deciduous tree, occupying much
of the swamp and overflow land along the coast
and rivers of the Southern States.
FIG. 138. T. disticlium.
224. Fir. This name is frequently applied to wood and to trees which
are not fir; most commonly to spruce, but also, especially in English mar-
kets, to pine. It resembles spruce, but is easily distinguished from it, as
well as from pine and larch, by the absence of resin-ducts. Quality, uses,
and habits similar to spruce.
70
THE MATERIALS OF CONSTRUCTION.
9. BALSAM-FIR (Abiesbalsamea) : A medium-sized
tree scattered throughout the northern pineries; cut,
in lumber operations, whenever of sufficient size, and
sold with pine or spruce. Minnesota to Maine and
northward.
FIG. 139. A. balsamea.
10. WHITE FIE (Abies grandis and Abies cdncolor
large-sized tree, forming
an important part of most
of the Western mountain-
forests, and furnishing
much of the lumber of the
respective regions. The
former occurs from Van-
couver to central Cali-
fornia and eastward to
Montana; the latter from
Oregon to Arizona and
eastward to Colorado and
New Mexico.
Medium to very
&
/IG. 140. A grandis.
FIG. 141. A. concolor.
11. WHITE FIR (Abies amabilis):
Good-sized tree, often forming exten-
sive mountain-forests. Cascade
Mountains of Washington and Oregon.
FIG. 142. A. amabilis.
TIMBER.
271
12. RED FIR (Abies nobilis) (not to be con-
founded with Douglas fir; see No. 37): Large to
very large tree, forming with A. amabilis extensive
forests on the slope of the mountains between 3000
and 4000 feet elevation. Cascade Mountains of
Oregon.
CO
FIG. 143. A nobilis.
13. BED FIR (Abies magnified): Very
large tree, forming forests about the base
of Mount Shasta. Sierra Nevada of Cali-
fornia, from Mount Shasta southward.
FIG. 144. A. magnified.
225. Hemlock. Light to medium weight, soft, stiff but brittle, com-
monly cross-grained, rough and splintery; sapwood and heartwood not well
defined ; the wood of a light, reddish-gray color, free from resin-ducts,
moderately durable, shrinks and warps considerably, wears rough, retains
nails firmly. Used principally for dimension stuff and timbers. Hemlocks
are medium to large-sized trees, commonly scattered among broad-leaved
trees and conifers, but often forming forests of almost pure growth.
14. HEMLOCK (Tsuga canadensis): Medium-
sized tree, furnishes almost all the hemlock of the
Eastern market. Maine to Wisconsin.; also following
the Alleghanies southward to Georgia and Alabama.
FIG. 145. 7\ canaden-
272
THE MATERIALS OF CONSTRUCTION.
-'
15. HEMLOCK (Tsuga mertensiand): Large-
sized tree; wood claimed to be heavier and harder
than the Eastern form and of superior quality.
Washington to California and eastward to Mon-
tana.
FIG. 146. T. mertensiana.
226. Larch or Tamarack. Wood like the best of hard pine both in ap-
pearance, quality, and uses, and, owing to its great durability, somewhat pre-
ferred in ship-building, for telegraph-poles and railroad-ties. In its struc-
ture it resembles spruce. The larches are deciduous trees, occasionally
covering considerable areas, but usually scattered among other conifers.
10. TAMARACK (Larix americana) (Hackma-
tack) : Medium-sized tree, often covering swamps,
in which case it is smaller and of poor quality.
Maine to Minnesota, and southward to Pennsyl-
vania.
FIG. 147. L. amtricana.
17. TAMARACK (L. occidentalis): Large-sized trees,
scattered, locally abundant. Washington and Oregon to
Montana.
FIG. 148. L. occi-
dentalis.
TIMBER. 273
227. Pine. Very variable, very light and soft in "soft" pine, such as
white pine; of medium weight to heavy and quite hard in " hard " pine, of
which long-leaf or Georgia pine is the extreme form. Usually it is stiff,
quite strong, of even texture, and more or less resinous. The sapwood is
yellowish white; the heartwood, orange-brown. Pine shrinks moderately,
seasons rapidly and without much injury; it works easily; is never too hard
to nail (unlike oak or hickory); it is mostly quite durable, and if well sea-
soned is not subject to the attacks of boring-insects. The heavier the wood,
the darker, stronger, and harder it is, and the more it shrinks and checks.
Pine is used more extensively than any other kind of wood. It is the prin-
cipal wood in common carpentry, as well as in all heavy construction,
bridges, trestles, etc. It is also used in almost every other wood iudusti^
for spars, masts, planks, and timbers in ship-building, in car and wagon con-
struction, in cooperage, for crates and boxes, in furniture work, for toys and
patterns, railway-ties, water-pipes, excelsior, etc. Pines are usually large
trees with few branches, the straight, cylindrical, useful stem forming by
far the greatest part of the tree; they occur gregariously, forming vast
forests, a fact which greatly facilitates their exploitation. Of the many
special terms applied to pine as lumber, denoting sometimes differences in
quality, the following deserve attention:
"White pine," '''pumpkin-pine," "soft pine," in the Eastern markets re-
fer to the wood of the white pine (Pi mis strobus), and on the Pacific Coast
to that of the sugar-pine (Pinus lambertiana).
(i Yellow pine " is applied in the trade to all the Southern lumber pines;
in the Northeast it is also applied to the pitch-pine (P. rigidci); in the West
it refers mostly to bull-pine (P. ponderosa).
" Yellow long-leaf pine," " Georgia pine," are terms which refer to long-
leaf pine (P. palustris).
" Hard pine " is a common term in carpentry, and applies to everything
except white pine.
" Pitch-pine" includes all Southern pines and also the true pitch-pine
(P. rigida), but is mostly applied, especially in foreign markets, to the
wood of the long-leaf pine (P. .palustris).
For the great variety of confusing local names applied to the Southern
pines in their homes, part of which have been adopted in the markets of
the Atlantic seaboard, see report of Chief of Division of Forestry for 1891,
page 212, etc., and also the list below.
274
THE MATERIALS OF CONSTRUCTION.
a. Soft Pines.
18. WHITE PINE (Pinus strobus) : Large to very
large-sized tree; for the last fifty years the most im-
portant timber tree of the Union, furnishing the
best quality of soft pine. Minnesota, Wisconsin
Michigan, New England, along the Alleghanies to
FIG. 149. P. strobus.
\19. SUGAR-PINE (Pinus lambertiana): A very
^ large tree, together with Abies concolor, forming
extensive forests; important lumber tree,
and California.
Oregon
FIG. 150. P. lambertiana.
20. WHITE PINE (Pinus monticola} : A large
tree, at home in Montana, Idaho, and the Pacific
States; most common and locally used in northern
Idaho.
FIG. 151.- P. monticola.
21. WHITE PIKE (Pinus ftexilis) : A small tree,
forming mountain-forests of considerable extent
and locally used; eastern Rocky Mountain slopes;
Montana to New Mexico.
FIG. 152. P. flexilis.
TIMBER.
215
b. Hard Pines.
22. LONG-LEAF PINE (Pinus palustris)
(Georgia pine, yellow pine, long straw-pine, etc.) :
Large tree; forms extensive forests and furnishes
the hardest and strongest pine lumber in the mar-
ket. Coast region from North Carolina to Texas.
FIG. 153. P. palustris.
23. BULL-PINE (Pinus ponderosa) (yellow
pine): Medium to very large-sized tree, forming
extensive forests in Pacific and Rocky Mountain
regions; furnishes most of the hard pine of the
West; sapwood wide; wood very variable.
FIG. 154. P. ponderosa.
24. LOBLOLLY PINE (Pinus tceda) (slash-pine,
old field-pine, rosemary-pine, sap-pine, short straw-
pine, etc.) : Large-sized tree, forms extensive
forests; wider-ringed, coarser, lighter, softer, with
more sapwood than the long-leaf pine, but the two
often confounded. This is the common lumber
pine from Virginia to South Carolina, and is
found extensively in Arkansas and Texas.
Southern States; Virginia to Texas and Arkansas.
FIG. 155. P. tceda.
25. NORWAY PINE (Pinus resinosa): Large-
sized tree, never forming forests, usually scattered
or in small groves, together with white pine; largely
sapwood and hence not durable. Minnesota to-
Michigan; also in New England to Pennsylvania.
FIG. 156. P. resinosa.
276
THE MATERIALS OF CONSTRUCTION.
26. SHORT-LEAF PINE (Pinus ecMnata]
(slash-pine,, Carolina pine, yellow pine, old field-
pine, etc.): Resembles loblolly pine; often ap-
proaches in its wood the Norway pine. The
common lumber pine of Missouri and Arkansas.
North Carolina to Texas and Missouri.
FIG. 157. P. ecMnata.
27. CUBAN PINE (Pinus cubensis] (slash-pine, swamp-
pine, bastard-pine, meadow-pine) : Resembles long-leaf pine,
but commonly has wider sapwood and coarser grain; does
not enter the markets to any great extent. Along the
coast from South Carolina to Louisiana.
FIG. 158. P. cu-
bensis.
28. BULL-PINE (Pinus jeffreyi)
(black pine) : Large-sized tree, wood re-
sembling bull-pine (P. ponderosa); used
locally in California, replacing P.
derosa at high altitudes.
FIG. 159. P. jeffreyi.
The following are small to medium-sized pines,
not commonly offered as lumber in the market;
used locally for timber, ties, etc. :
29. BLACK PINE (Pinus murrayand) (lodge-
pole pine, tamarack): Rocky Mountains and
Pacific regions.
FIG. 160. P. murrayana.
TIMBER.
277
30. PITCH-PINE (Pinus rigida): Along the
coast from New York to Georgia, and along the
mountains to Kentucky.
FIG. 161. P. rig ida.
31. JERSEY PINE (Pinus inops) (scrub-pine)
As before.
FIG. 163. P. banksiana.
FIG. 162. P. inops.
32. GRAY PIXE (Pinus banksiana) (scrub-
pine) : Maine, Vermont, and Michigan to Minne-
sota.
Redwood. See CEDAR.
228. Spruce. Resembles soft pine, is light,
very soft, stiff, moderately strong, less resinous than
pine; has no distinct, heartwood, and is of whitish color. Used like soft
pine, but also employed as resonance-wood and preferred for paper pulp.
Spruces, like pines, form extensive forests; they are more 'frugal, thrive on
thinner soils, and bear more shade, but usually require a more humid cli-
mate. "Black" and "white" spruce, as applied by lumbermen, usually
refer to narrow- and wide-ringed forms of the black spruce (Picea nigra).
33. BLACK SPRUCE (Picea nigra) : Medium-
sized tree, forms extensive forests in northeastern
United States and in British America; occurs
scattered or in groves, especially in low lands
throughout the Northern pineries. Important
lumber tree in Eastern United States. Maine to
Minnesota, British America, and on the Allegha-
nies to North Carolina.
FIG. 164. P. nigra.
278
THE MATERIALS OF CONSTRUCTION.
34. WHITE SPRUCE (Picea alba) : Generally
associated with the preceding; most abundant
along streams and lakes, grows largest in Mon-
tana, and forms the most important tree of the
subarctic forest of British America. Northern
United States, from Maine to Minnesota, also
from Montana to Pacific, British America.
FIG. 165. P. alba.
35. WHITE SPRUCE (Picea engelmanni):
Medium to large-sized tree, forming extensive
forests at elevations from 5000 to 10,000 feet
above sea-level; resembles the preceding, but occu-
pies a different station. A very important timber
tree in the central and southern parts of the
Rocky Mountains. Rocky Mountains from Mexico
to Montana.
FIG. 166. P. engelmanni.
36. TIDE-LAND SPRUCE (Picea sitchensis): A
large-sized tree, forming an extensive coast-belt
forest. Along the seacoast from Alaska to central
California.
Bastard Spruce. Spruce or fir in name,
but resembling hard pine or larch in the appear-
ance, quality, and uses of its wood.
FIG. 167. P. sitchensis.
37. DOUGLAS SPRUCE (Pseudotsuga doug-
lasii) (yellow fir, red fir, Oregon pine) : One of
the most important trees of the Western United
States; grows very large in the Pacific States,
to fair size in all parts of the mountains, in
Colorado up to about 10,000 feet above sea-
level; forms extensive forests, often of pure
growth. Wood very variable, usually coarse-
grained and heavy, with very pronounced sum-
mer wood, hard and strong ("red" fir), but
often fine-grained and light (" yellow " fir). It
replaces hard pine and is especially suited to
heavy construction. From the plains to the
FIG. 168. P. douglasii. Pacific Ocean ; from Mexico to British America.
TIMBER.
279
Tamarack. See LARCH.
229, Yew. Wood heavy, hard, extremely stiff and strong, of fine tex-
ture, with a pale yellow sapwood and an orange-
red heart; seasons well and is quite durable.
Yew is extensively used for archery, bows, turn-
er's ware, etc. The yews form no forests, but
occur scattered with other conifers.
38. YEW (Taxus brevifolia): A small to
medium-sized tree of the Pacific region.
FIG. 169. T. brevifolia.
B. BROAD-LEAVED WOODS, (HARDWOODS.)
Woods of complex and very variable structure and therefore differing
widely in quality, behavior, and consequently in applicability to the arts.
230. Ash. Wood heavy, hard, strong, stiff, quite tough, not durable in
contact with soil, straight-grained, rough on the split surface and coarse in
texture. The wood shrinks moderately, seasons with little injury, stands
well and takes a good polish. In carpentry ash is used for finishing lumber,
stairways, panels, etc.; it is used in ship-building, in the construction of
cars, wagons, carriages, etc., in the manufacture of farm-implements,
machinery, and especially of furniture of all kinds, and also for harness
work; for barrels, baskets, oars, tool-handles, hoops, clothespins, and toys.
The trees of the several species of ash are rapid growers, of small to medium
height with stout trunks; they form no forests, but occur scattered in
almost all our broad-leaved forests.
39. WHITE ASH (Fraxinus americana):
Medium, sometimes large-sized tree. Basin of
the Ohio, but found from Maine to Minnesota
and Texas.
FIG. 170 F. americana.
280
THE MATERIALS OF CONSTRUCTION.
40. RED ASH (Fraxinus pulescens) : Small-sized
tree. North Atlantic States, but extends to the
Mississippi.
FIG. 171. F. pubesccns.
41. BLACK ASH (Fraxinus sambucifolia) (hoop-
ash, ground-ash): Medium-sized tree, very common.
Maine to Minnesota, and southward to Virginia and
Arkansas.
FIG. 172. F. sambuci-
folia.
42. BLUE ASH (Fraxinus quadrangulata):
Small to medium-sized. Indiana and Illinois;
occurs from Michigan to Minnesota and southward
to Alabama.
FIG. 173. F. quadrangu-
lata.
43. GKEEN ASH (Fraxinus viridis): Small-
sized tree. New York to the Rocky Mountains,
and southward to Florida and Arizona.
FIG. 174. # viridis
TIMBER
281
44. OREGON ASH (Fraxinus oregana):
Medium-sized tree. Western Washington
to California.
FIG. 175. F. oregana.
Aspen. See POPLAR.
231. Basswood.
45. BASSWOOD (Tilia americana) (lime-tree,
American linden, lin, bee-tree): Wood light, soft,
stiff but not strong, of fine texture, and white to
light brown color. The wood shrinks considerably
in drying, works and stands well; it is used in car-
pentry, in the manufacture of furniture and wood-
enware, both turned and carved, ih cooperage, for
toys, also for panelling of car and carriage bodies.
Medium to large-sized tree, common in all' Northern
broad-leaved forests; found throughout the Eastern
FIG. 176.-T. americana. United States.
46. WHITE BASSWOOD (Tilia heterophylla) : A
small-sized tree most abundant in the Alleghauy
region.
FIG. 177. T. lietero-
phylla.
282
THE MATERIALS OF CONSTRUCTION.
232. Beech.
47. BEECH (Fagus ferruginea): Wood heavy, hard, stiff,' strong, of
rather coarse texture, white to light brown, not dura-
ble in the ground, and subject to the inroads of
boring-insects; it shrinks and checks considerably
in drying, works and stands well, and takes a good
polish. Used for furniture, in turnery, for handles,
lasts, etc. Abroad it is very extensively employed
by the carpenter, millwright, and wagon-maker, in
turnery as well as wood-carving. The beech is a
medium-sized tree, common, sometimes forming for-
ests; most abundant in the Ohio and the Mississippi
basin, but found from Maine to Wisconsin and
southward to Florida.
233. "Birch. Wood heavy, hard, strong, of fine texture ; sap wood whit-
ish, heartwood in shades of brown with red and yellow; very handsome,
with satiny lustre, equalling cherry. The wood shrinks considerably in dry-
ing, works and stands well and takes a good polish, but is not durable if
exposed. Birch is used for finishing-lumber in building, in the manufac-
ture of furniture, in wood-turnery for spools, boxes, wooden shoes, etc., for
shoe lasts and pegs, for wagon-hubs, ox-yokes, etc., also in wood-carving.
The birches are medium-sized trees, form extensive forests northward, and
occur scattered in all broad-leaved forests of the Eastern United States.
FIG. 178.^. ferruginea.
48. CHERRY-BIRCH (Betula lento) (black birch,
sweet birch, mahogany-birch): Medium-sized tree;
very common. Maine to Michigan and to Tennessee.
FIG. 179. B. lentd.
49. YELLOW BIRCH (Betula luted) (gray
birch): Medium-sized tree; common. Maine to
Minnesota and southwest to Tennessee.
T /
FIG. 180. B. lutea.
TIMBER.
283
50. RED BIRCH (Betula nigra) (river-birch):
Small to medium-sized tree; very common; lighter
and less valuable than the preceding. New England
to Texas and Missouri
XJU,
Fm. 181. B. nigra.
51. CANOE-BIRCH (Betula papyri/era) (white
birch, paper-birch) : Generally a small tree; common,
forming forests; wood of good quality, but relatively
light. All along the northern boundary of United
States and northward, from the Atlantic to the Pacific.
FIG. 182. B. papy-
r if era.
Black Walnut. See WALNUT.
234. Blue Beech.
52. BLUE BEECH (Carpinus caroliniana) (horn-
beam, water-beech, ironwood) : Wood very heavy,
hard, strong, very stiff, of rather fine texture and
white color; not durable in the ground; shrinks
und checks greatly, but works and stands well.
Used chiefly in turnery for tool-handles, etc.
Abroad much used by millwrights and wheel-
wrights. A small tree, largest in the Southwest,
but found in nearly all parts of the Eastern United
States.
FIG. 183. (7. caroliniana.
Bois d'Arc. See OSAGE ORANGE.
235. Buckeye Horse-Chestnut. Wood light, soft, not strong, often
quite tough, of fine and uniform texture and creamy-white color. It shrinks
considerably, but works and stands well. Used for wooden ware, artificial
limbs, paper-pulp, and locally also for building-lumber. Small-sized trees,
scattered.
284
THE MATERIALS OF CONSTRUCTION.
53. OHIO BUCKEYE (JSsculus glabra) (fetid buckeye)
Alleghanies, Pennsylvania to Indian Territory.
FIG 184.
E. glabra.
54. SWEET BUCKEYE (^Esculus flava): Alle-
ghanies, Pennsylvania to Texas.
236. Butternut.
FIG. 1SQ.J. tinerea.
237. Catalpa.
FIG 185. ^0. flaxa.
55. BUTTERNUT (Juglans cinerea) (white
walnut): Wood very similar to black walnut,
but light, quite soft, not strong, and of light-
brown color. Used chiefly for finishing lumber,
cabinetwork, and cooperage. Medium-sized
tree, largest and most common in the Ohio
basin; Maine to Minnesota and southward to
Georgia and Alabama.
56. CATALPA (Catalpa speciosa): Wood light,,
soft, not strong, brittle, durable, of coarse texture
and brown color; used forties and posts, but well
suited for a great variety of uses. Medium-sized
trees; lower basin of the Ohio Eiver, locally com-
mon. Extensively planted, and therefore promising
to become of some importance.
FIG. 187. C. speciosa.
TIMBER.
285
238. Cherry.
57. CHERRY (Prunus serotina) : Wood heavy, hard, strong, of fine tex-
ture; sapwood 'yellowish white, heartwood reddish to brown. The wood
shrinks considerably in drying, works and stands well, takes a good polish,
and is much esteemed for its beauty. Cherry is
chiefly used as a decorative finishing-lumber for
buildings, cars, and boats, also for furniture and in
turnery. It is becoming too costly for many ^purposes
for which it is naturally well suited. The lumber-
furnishing cherry of this country, the wild black
cherry (Prunus serotina), is a small to medium-sized
tree, scattered through many of the broad-leaved
woods of the western slope of the Alleghanies, but FlG - 188 -~ p - 8e tina.
found from Michigan to Florida and west to Texas. Other species of this
genus as well as the hawthorns (Cratcegas) and wild apple (Pyrus] are not
commonly offered in the market. Their wood is of the same character as
cherry, often even finer, but in small dimensions.
239. Chestnut.
FIG. 189. a vulgaris.
58. CHESTXUT (Castanea vulgaris var. ameri-
cana) : Wood light, moderately soft, stiff, not strong,
of coarse texture; the sapwood light, the heartwood
darker brown. It shrinks and checks considerably
in drying, works easily, stands well, and is very dura-
ble. Used in cabinetwork, cooperage, for railway-ties,
telegraph-poles, and locally in heaVy construction.
Medium-sized tree, very common in the Alleghanies,
occurs from Maine to Michigan and southward to
Alabama.
59. CHINQUAPIN (Castanea pumila): A small-sized
tree, with wood slightly heavier than, but otherwise sim-
ilar to, the preceding; most common in Arkansas, but
with nearly the same range as the chestnut.
FIG. 190. C. pumila.
286
THE MATERIALS OF CONSTRUCTION.
60. CHINQUAPIN ( Castanopsis chryso-
pliylld] : A medium-sized tree of the west-
ern ranges of California and Oregon.
240. Coffee-tree.
FIG. 191. C. chrysophylla.
61. COFFEE-TREE (Gymnodadus oanadensis)
(coffee-nut) : Wood heavy, hard, strong, very stiff, of
coarse texture; durable; the sapwood yellow, the
heartwood reddish brown; shrinks and checks con-
siderably in drying; works and stands well and takes
a good polish. It is used to a limited extent in cab-
inetwork. A medium to large-sized tree; not com-
mon. Pennsylvania to Minnesota and Arkansas.
FIG. 192. G. canadensis.
Cottonwood. See POPLAR.
Cucumber-tree. See TULIP.
241. Elm. Wood heavy, hard, strong, very tough: moderately durable in
contact with the soil; commonly cross-grained, difficult to split and shape,
warps, and checks considerably in drying, but stands well if properly han-
dled. The broad sapwood whitish, heart brown, both with shades of gray
and red; on split surface rough; texture coarse to fine; capable of high
polish. Elm is used in the construction of cars, wagons, etc., in boat- and
ship-building, for agricultural implements and machinery; in rough cooper-
age, saddlery and harness work, but particularly in the manufacture of all
kinds of furniture, where the beautiful figures, especially those of the tan-
gential or bastard section, are just beginning to be duly appreciated. The
elms are medium to large-sized trees, of fairly rapid growth, with stout trunk,
form no forests of pure growth, but are found scattered in all the broad-
leaved woods of our country, sometimes forming a considerable portion of
the arborescent growth.
TIMBER.
287
62. WHITE-ELM (Ulmus americana) (American
elm, water-elm) : Medium to large-sized tree, com-
mon. Maine to Minnesota, southward to Florida
and Texas.
FIG. 193. U. americana.
63. ROCK-ELM (Ulmus racemosa) (cork-elm,
hickory-elm, white elm, cliff-elm) : Medium to large-
sized tree. Michigan, Ohio, from Vermont to Iowa,
southward to Kentucky.
FIG. 194. U. racemosa.
64. RED ELM (Ulmus fulva) (slippery elm,
moose-elm) : Small-sized, tree, found chiefly along
watercourses. New York to Minnesota, and south-
ward to Florida and Texas.
FIG. 195. U. fulva.
65. CEDAR-ELM ( Ulmus crassifolia] :
Small-sized tree, quite common. Arkansas
and Texas.
FIG. 196. U. crassifolia.
288
THE MATERIALS OF CONSTRUCTION.
66. WINGED ELM (Ulmus alata) (Wahoo):
Small-sized tree, locally quite common. Arkan-
sas, Missouri, and eastern Virginia.
FIG. 197. U. alata.
242. Gum. This general term refers to two kinds of wood usually dis-
tinguished as sweet or red gum, and sour, black, or tupelo gum, the former
being a relative of the witch-hazel, the latter belonging to the dogwood
family.
ica.
67. TUPELO (Nyssa sylvatica) (sour gum, black
gum) : Maine to Michigan, and southward to Flor-
ida and Texas. Wood heavy, hard, strong, tough,
of fine texture, frequently cross-grained, of yellowish
or grayish-white color, hard to split and work,
troublesome in seasoning, warps and checks consid-
erably, and is not durable if exposed; used for wagon-
hubs, wooden ware, handles, wooden shoes, etc.
Medium to large-sized trees, with straight, clear
trunks; locally quite abundant, but never forming
forests of pure growth.
68. TUPELO GUM (Nyssa uniflora) (cotton-
gum) : Lower Mississippi basin, northward to Illi-
nois and eastward to Virginia; otherwise like pre-
ceding species.
FIG. 199. N. uniflora.
TIMBER.
289
69. SWEET GUM (Liquidambar styraciflua) (red gum, liquidambar,
bilsted) : Wood rather heavy, rather soft, quite stiff
and strong, tough, commonly cross-grained, of fine
texture; the broad sapwood whitish, the heart-
wood reddish brown; the wood shrinks and warps
considerably, but does not check badly, stands well
when fully seasoned, and takes good polish. Sweet
gum is used in carpentry, in the manufacture of fur-
niture, for cut veneer, for wooden plates, plaques,
baskets, etc., also for wagon-hubs, hat-blocks, etc.
A large-sized tree, very abundant, often the princi-
pal tree in the swampy parts of the bottoms of the
FIG. 200. L. styraciflua. Lower Mississippi Valley; occurs from New York
to Texas, and from Indiana to Florida.
243. Hackberry.
70. HACKBERRY (Cettis occidentalis) (sugar-berry):
The handsome wood heavy, hard, strong, quite tough,
of moderately fine texture, and greenish- or yellowish-
white color; shrinks moderately, works well, and takes
a good polish. So far but little used in the manufac-
ture of furniture. Medium to large-sized tree, locally
quite common, largest in the Lower Mississippi Valley;
occurs in nearly all parts of the Eastern United States.
244. Hickory. Wood very heavy, hard, and strong, proverbially tough,
of rather coarse texture, smooth and of straight grain. The broad sapwood
white, the heart reddish nut-brown. It dries slowly, shrinks and checks
considerably; is not durable in the ground, or if exposed, and especially the
sapwood, is always subject to the inroads of boring-insects.. Hickory excels
as carriage and wagon stock, but is also extensively used in the manufacture
of implements and machinery, for tool-handles, timber-pins, for harness
work and cooperage. The hickories are tall trees with slender stems, never
form forests, occasionally small grove's, but usually occur scattered among
other broad-leaved trees in suitable localities. The following species all
contribute more or less to the hickory of the markets:
71. SHAGBARK HICKORY (Hicoria ovata) (shell-
bark hickory) : A medium to large-sized tree, quite
common; the favorite among hickories; best devel-
oped in the Ohio and Mississippi basins ; from Lake
Ontario to Texas, Minnesota to Florida.
FIG. 201. G. occiden-
talis.
FIG 202. H. ovata.
290
THE MATERIALS OF CONSTRUCTION.
77. MOCKERNUT HICKORY (Hicoria alia) (black
hickory, bull- and black-nut, big-bud, and white-heart
hickory): A medium to large-sized tree, with the same
range as the foregoing; common, especially in the
South.
FIG. 203. H alba.
73. PIGNUT HICKORY (Hicoria glair a)
(brown hickory, black hickory, switch-bud hick-
ory) : Medium to large-sized tree, abundant; all
Eastern United States.
FIG. 204. H. glabra.
74. BITTER-NUT HICKORY (Hicoria minima)
(swamp hickory) : A medium-sized tree, favoring
wet localities, with the same range as the preceding.
FIG. 205. H. minima
75. PECAN (Hicoria pecan) (Illinois nut): A
large tree, very common in the fertile bottoms of the
Western streams. Indiana to Nebraska and south-
ward to Louisana and Texas.
FIG. 206. //. pecan.
TIMBER.
291
245. Holly.
FIG. 207. I. opaca.
76. HOLLY (Ilex opaca): Wood of medium weight,,
hard, strong, tough, of fine texture and white color;
works and stands well, used for cabinetwork and turnery.
A small tree, most abundant in the Lower Mississippi
Valley and Gulf States, but occurring eastward to Massa-
chusetts and north to Indiana.
Horse-chestnut. See BUCKEYE.
Ironwood. See BLUE BEECH.
246. Locust. This name applies to both of the following:
77. BLACK LOCUST (Robinia pseudacacia) (black locust, yellow locust) :
Wood very heavy, hard, strong, and tough, of
coarse texture, very durable in contact with the
soil, shrinks considerably, and suffers in season-
ing; the very narrow sapwood yellowish, the
heartwood brown, with shades of red and green.
Used for wagon-hubs, treenails or pins, but espe-
cially for ties, posts, etc. Abroad it is much used
for- furniture and farm-implements, and also in
turnery. Small to medium-sized tree, at home in
the Alleghanies, extensively planted, especially
in the West.
FIG. 208. R. pseudacacia.
78. HONEY-LOCUST (Gleditschia triacantlws)
(black locust, sweet locust, three-thorned acacia) :
Wood heavy, hard, strong, tough, of coarse texture,
susceptible of a good polish, the narrow sapwood yel-
low, the heartwood brownish red. So far but little
appreciated except for fencing and fuel; used to some
extent for wagon-hubs and in rough construction.
A medium-sized tree, found from Pennsylvania to *^ 5"
Nebraska, and southward to Florida and Texas; lo- FIG. 209. G. triacanthos.
cally quite abundant.
Magnolia. See TULIP.
247. Maple. Wood heavy, hard, strong, stiff, and tough, of fine texture,,
frequently wavy-grained, this giving rise to " curly " and " blister" figures;
not durable in the ground or otherwise exposed. Maple is creamy white,
with shades of light brown in the heart; shrinks moderately, seasons, works
and stands well, wears smoothly, and takes a fine polish. The wood is used
for ceiling, flooring, panelling, stairway, arid other finishing-lumber in house,
ship, and car construction; it is used for the keels of boats and ships, in the
manufacture of implements and machinery, but especially for furniture,
where entire chamber sets of maple rival those of oak. Maple is also used
S92
THE MATERIALS OF CONSTRUCTION.
for shoe-lasts and other form-blocks, for shoe-pegs, for piano actions, school
apparatus, for wood type in show-bill printing, tool-handles, in wood-carv-
ing, turnery, arid scrollwork. The maples are medium-sized trees, of fairly
rapid growth; sometimes form forests and frequently constitute a large pro-
portion of the arborescent growth.
79. SUGAR-MAPLE (Acer saccharum) (hard
maple, rock-maple) : Medium to large-sized tree,
very common, forms considerable forests. Maine
to Minnesota, abundant, with birch, in parts of
the pineries; southward to northern Florida;
most abundant in the region of the Great Lakes.
FIG. 210. A. saccharum.
80. RED MAPLE (Acer rubrum) (swamp- or water-
maple) : Medium-sized tree. Like the preceding, but
scattered along watercourses and other moist localities.
FIG. 211. A. rubrum.
81. SILVER MAPLE (Acer saccliarinum) (soft
maple, silver maple): Medium-sized, common; wood
lighter, softer, inferior to hard maple, and usually
offered in small quantities and held separate in the
market. Valley of the Ohio, but occurs from Maine
o Dakota, and southward to Florida.
FIG. 212. A. sacchari-
num.
82. BROAD-LEAVED MAPLE (Acer macropliyl-
lum): Medium-sized tree, forms considerable
forests, and like the preceding has a lighter,
softer, and less valuable wood. Pacific Coast.
PIG. 213. A. macrophyllum.
TIMBER.
293
248. Mulberry.
83. RED MULBERRY (Morus rubra): Wood moderately
heavy, hard, strong, rather tough, of coarse texture, dur-
able; sapwood whitish, hard yellow to orange-brown;
shrinks and checks considerably in drying; works and
stands well. Used in cooperage and locally in ship-build-
ing and in the manufacture of farm-implements. A small-
sized tree, common in the Ohio and Mississippi valleys, FIG gl4 _ M
but widely distributed in the Eastern United States. bra
249. Oak. Wood very variable, usually very heavy and hard, very strong-
and tough, porous, and of coarse texture; the sapwood whitish, the heart
" oak " brown to reddish brown. It shrinks and checks badly, giving trouble
in seasoning, but stands well, is durable, and little subject to attacks of in-
sects. Oak is used for many purposes: in ship-building, for heavy construc-
tion, in common carpentry, in furniture, car, and wagon work, cooperage,
turnery, and even in wood-carving; also in the manufacture of all kinds of
farm-implements, wooden mill machinery, for piles and wharves, railway-
ties, etc. The oaks are medium to large-sized trees, forming the predomi-
nant part of a large portion of our broad-leaved forests, so that these are
.generally " oak forests," though they always contain a considerable propor-
tion of other kinds of trees. Three well-marked kinds, white, red, and live
oak, are distinguished and kept separate in the market, Of the two princi-
pal kinds white oak is the stronger, tougher, less porous, and more durable.
Ked oak is usually of coarser texture, more porous, often brittle, less dura-
ble, and even more troublesome in seasoning than white oak. In carpen-
try and furniture work red oak brings about the same price at present as
white oak. The red oaks everywhere accompany the white oaks, and, like
the latter, are usually represented by several species in any given locality.
Live-oak, once largely employed in ship-building, possesses all the good
qualities (except that of size) of white oak even to a greater degree. It is
one of the heaviest, hardest, and most desirable building-timbers of this
country; in structure it resembles the red oaks, but is much less porous.
84. WHITE OAK (Quercus alba): Medium to
large-sized tree, common in the Eastern States,
Ohio and Mississippi valleys; occurs throughout
Eastern United States.
FIG. 215. . alba.
294
THE MATERIALS OF CONSTRUCTION.
85. BUR -OAK (Quercus macrocarpa)
(mossy-cup oak, over-cup oak): Large-sized
tree, locally abundant, common. Bottoms
west of Mississippi; range farther west than
preceding.
FIG. 216. Q macrocarpa.
86. SWAMP WHITE OAK (Quercus bicolor):
Large-sized tree, common. Most abundant in the
Lake States, but with range as in white oak.
FIG. 217. Q. bicolor.
87. YELLOW OAK (Quercus prinoides) (chestnut-oak, chin-
quapin oak) : Medium-sized tree. Southern Alleghanies, east- g-
ward to Massachusetts.
FIG. 218.
Q. prinoides.
88. BASKET-OAK (Quercus michauxii) (cow-
oak): Large-sized tree, locally abundant; lower
Mississippi and eastward to Delaware.
FIG. 219. Q. michauxii.
89. OVER-CUT OAK (Quercus lyrata) (swamp white
oak, swamp post-oak): Medium to large-sized tree,
rather restricted; ranges as in the preceding.
FIG. 220. Q. lyrata.
TIMBER.
295
90. POST-OAK (Quercus obtusilola) (iron-
oak): Medium to large-sized tree. Arkansas
to Texas, eastward to New England, and
northward to Michigan.
FIG. 221. Q. oUmiloba.
91. WHITE OAK (Quercus durandii): Medium to
small-sized tree. Texas, eastward to Alabama.
FIG. 222. Q. du-
randii.
92. WHITE OAK (Quercus garryana): Medium to
large-sized tree. Washington to California.
FIG. 223. Q.
garryana.
93. WHITE OAK (Quercus lobata): Medium to
large-sized tree; largest oak on the Pacific coast, r
California.
FIG. 224, Q. lobata.
296
THE MATERIALS OF CONSTRUCTION.
94. RED OAK (Quercus rubra) (black oak):
Medium to large-sized tree; common in all parts of
its range. Maine to Minnesota, and southward to
the Gulf.
FIG. 225.$. rubra.
95. BLACK OAK (Quercus tinctoria) (yellow
oak) : Medium to large-sized tree; very common in
the Southern States, but occurring north as far as
Minnesota, and eastward to Maine.
FIG. 226. Q. tinctoria.
96. SPANISH OAK (Quercus falcata) (red oak):
Medium-sized tree; common in the South Atlantic
and Gulf region, but found from Texas to New
York, and north to Missouri and Kentucky.
FIG. 227. Q. falcata.
97. SCARLET OAK (Quercus coccinea): Me-
dium to large-sized tree; best developed in the
lower basin of the Ohio, but found from Maine
to Missouri, and from Minnesota to Florida.
FIG. 228. Q. coccinea,
98. PIN-OAK (Quercus palustris] (swamp Spanish
oak, water-oak): Medium to large-sized tree, common
along borders of streams and swamps. Arkansas to Wis-
consin, and eastward to the Alleghanies.
FIG. 229.$. palus-
iris.
TIMBER
297
99. WILLOW-OAK (Quercus pliellos) (peach-oak):
Small to medium-sized tree. New York to Texas,
and northward to Kentucky.
Q. phellos.
100. WATER-OAK (Quercus aquatica) (duck-oak,
possum-oak, punk-oak) : Medium to large-sized tree,
of extremely rapid growth. Eastern Gulf States, east-
ward to Delaware, and northward to Missouri and
Kentucky.
FIG. 231. Q. aquatica.
101. LIVE-OAK (Quercus virens): Small-sized tree,
scattered along the coast from Virginia to Texas.
FIG. 232. Q. virens.
102. LIVE-OAK (Quercus
cJirysolepis] (m a u l-o a k,
Valparaiso oak): Medium-
sized tree. California.
FIG. 233. Q. cJirysolepis.
298
THE MATERIALS OF CONSTRUCTION.
250. Osage Orange.
103. OSAGE ORAXGE (Madura aurantiacd) (Bois d'Arc): Wood very
heavy, exceedingly hard, strong, not tough, of
moderately coarse texture, and very durable; sap-
wood yellow, heart brown on the end, yellow on
longitudinal faces, soon turning grayish brown if
exposed; it shrinks considerably in drying, but
once dry it stands unusually well. Formerly much
used for wheel stock in the dry regions of Texas;
otherwise employed for posts, railway-ties, etc.
Seems too little appreciated ; it is well suited for
turned ware and especially for wood-carving. A
small-sized tree, of fairly rapid growth, scattered
through the rich bottoms of Arkansas and Texas.
251. Persimmon.
FIG. 234. M. aurantiaca
104, PERSIHMOK (Diospyros virginiana) : Wood
very heavy and hard, strong and tough; resembles
hickory, but is of finer texture; the broad sapwood
cream-color, the heart black; used in turnery for
shuttles, plane-stocks, shoe-lasts, etc. Small to
medium-sized tree, common and best developed in
the Lower Ohio Valley, but occurs from New York
to Texas and Missouri.
FIG. 235. D. virginiana,
252. Poplar and Cottonwood. (See also TULIP-WOOD.) Wood light,
very soft, not strong, of fine texture and whitish, grayish, to yellowish color,
usually with a satiny lustre. The wood shrinks moderately (some cross-
grained forms warp excessively), but checks little; is easily worked, but is
not durable. Used as building- and furniture-lumber, in cooperage for
sugar- and flour-barrels, for crates and boxes (especially cracker-boxes), for
woodenware and paper-pulp.
105. COTTONWOOD (Populuo monilifera): Large-
sized tree; forms considerable forests along many of
the Western streams, and furnishes most of the
cottonwood of the market. Mississippi Valley and
west; New England to the Rocky Mountains.
FIG. 236. P. monilifera.
TIMBER.
299
106. BALSAM (Populus balsamifera) (balm of
Gilead): Medium to large-sized tree; common all
along the northern boundary of the United States.
FIG. 237. P. balsamifera.
107. BLACK COTTON-WOOD (Populus trichocarpa) \
The largest deciduous tree of Washington ; very common.
Northern Rocky Mountains and Pacific region.
FIG. 238.
P. trichocarpa.
108. COTTONWOOD (Populus fre-
montii var. ivislizeni) : Medium to large-
sized tree, common. Texas to Cali-
fornia.
FIG. 239. P. wislizeni.
109. POPLAR (Populus grandidentata) : Medium-
sized tree, chiefly used for pulp. Maine to Minne-
sota and southward along the Alleghanies.
FIG. 240.
P. grandidentata.
300
THE MATERIALS OF CONSTRUCTION.
110. ASPEX (Populus tremuloides): Small to
medium-sized tree, often forming extensive forests
and covering burned areas. Maine to Washington
and
California and New Mexico.
See GUM.
See GUM.
th ward, south in the Western mountains to
FIG. 242. 8. sassafras.
FIG 241
P. tremuloides.
Sour Gum.
Red Gum.
253. Sassafras.
111. SASSAFRAS (Sassafras sassafras): Wood
light, soft, not strong, brittle, of coarse texture,
durable; sapwood yellow, heart orange - brown.
Used in cooperage, for skiffs, fencing, etc.
Medium-sized tree, largest in the Lower Mississippi
Valley, from New England to Texas, and from
Michigan to Florida.
Sweet Gum. See GUM.
254. Sycamore.
112. SYCAMORE (Platanus occidentalis} (button-wood, buttonball-tree,
water-beech) : Wood moderately heavy, quite hard,
stiff, strong, tough, usually cross-grained, of coarse
texture, and white to light-brown color; the wood
is hard to split and work, shrinks moderately,
warps and checks considerably, but stands well.
It is used extensively for drawers, backs, bottoms,
etc., in cabinetwork, for tobacco-boxes, in cooper-
age, and also for finishing lumber, where it has
too long been underrated. A large tree, of rapid
growth, common and largest in the Ohio and Mis-
sissippi valleys, at home in nearly all parts of the
Eastern United States. The California species
FIG. 243. P. occidentalis. -, 10 rtl , ...
113. Platanus racemosa resembles in its
wood the Eastern form.
255. Tulip-wood.
114. TULIP - TREE (Liriodendron tulipifera)
(yellow poplar, white wood) : Wood quite varia-
ble in weight, usually light, soft, stiff but not
strong, of fine texture and yellowish color; the
wood shrinks considerably, but seasons without
much injury; works and stands remarkably well.
Used for siding, for panelling and finishing-lum-
ber in house-, car-, and ship-building, for side-
boards and panels of wagons and carriages; also
in the manufacture of furniture, implements, and
machinery, for pump-logs, and almost every kind
FIG. 244. L. tulipifera.
TIMBER.
301
of common wooden ware, boxes, shelving, drawers, etc. An ideal wood for
the carver and toyman. A large tree, does not form forests, but is quite
common, especially in the Ohio Basin; occurs from New England to
Missouri and southward to Florida.
115. CUCUMBER - TREE (Magnolia acumi-
nata) : A medium-sized tree, most common in
the southern Alleghanies, but distributed from
New York to Arkansas, southward to Alabama,
and northward to Illinois. Resembling, and
probably confounded with, tulip-wood in the
markets.
FIG. 245. M. acuminata.
Tupelo. See GUM.
256. Walnut.
116. BLACK WALNUT (Juglans nigra): Wood heavy, hard, strong, of
coarse texture; the narrow sapwood whitish, the
heartwood chocolate-brown. The wood shrinks
moderately in drying, works and stands well,
takes a good polish, is quite handsome, and has
been for a long time the favorite cabinet-wood in
this country. Walnut, formerly used even for
fencing, has become too costly for ordinary uses,
and is to-day employed largely as a veneer, for
inside finish and cabinetwork; also in turnery, for
gunstocks, etc. Black walnut is a large tree, with
stout trunk, of rapid growth, and was formerly
quite abundant throughout the Alleghany region,
occurring from New England to Texas, and from
Michigan to Florida.
White Walnut. See BUTTERNUT
White Wood. See TULIP, and also BASSWOOD.
Yellow Poplar. See TULIP.
FIG. 246.
nigra.
PART III.
TESTING-MACHINES AND METHODS OF TESTING
MATERIALS OF CONSTRUCTION.
CHAPTER XIV.
MECHANICAL TESTS IN GENERAL.
257. General Observations. Mechanical tests are those most commonly
used to discover the working qualities of the materials of construction.
Since these materials nearly always have to resist the action of external
forces, it follows that the suitableness of such a material to resist the action
of these forces is best determined by tests approximating as nearly as may
be to the conditions of actual practice.
Mechanical tests, therefore, are of supreme importance in the study of
any building material. By standardizing the conditions under which these
tests are carried out, the results become comparable wherever or by whom-
soever they are made, and they also become authoritative in all countries
and for all purposes. If such results can be made wholly independent of
the means employed in making the tests, and hence to furnish a knowledge
of the true characteristics of the material, they can be used safely in theo-
retical generalizations on the one hand, and in the practical designing of
structures on the other. With many kinds of tests this ideal divorcement
of the results from the conditions of the tests can certainly never be
attained, as in the case of tests by impact, but it doubtless can be practically
attained in some of the more simple tests, as in tension and compression.
In the former case the most that can be accomplished is to prescribe uniform
conditions in order that the results obtained by different experimenters may
be comparable, although they may not serve for accurate scientific general-
izations. They might also serve to give a relative value to the various
materials or samples so tested, and to grade them with some degree of ap-
proximation to their true relative merits for a proposed purpose. Such
tests, therefore, may serve fully their immediate object even though the
302
MECHANICAL TESTS IN GENERAL. 303
results can be given no absolute significance whatever. If, however, the
conditions of such tests are allowed to vary, they would lose even this rela-
tive significance, and would therefore be quite worthless. The standardiz-
ing of any particular kind of test evidently depends on the state of the science
at the time ; and as our knowledge of any particular property of a material
increases, it is probable that our standard methods of testing will also have
to change. No such standards, therefore, can be fixed permanently, but
certain methods can be agreed on and followed for a time, and when a change
is made let all change together. To attain to this kind of unity of action
it is necessary to have a world's representative body which will command
the confidence and allegiance of both the theoretical and the practical users
of materials in all civilized countries to decide such questions. A beginning
has been made in this direction in the International Commission on the
Standardization of Methods of Testing the Materials of Construction, which
has had several meetings in Europe at intervals of about three years, the
last one being at Zurich in September, 1895, where a permanent organiza-
tion was effected. The French Government, also (in 1891, as a result of
action taken by engineers at their centennial exposition in 1889), appointed
a national French Commission of over one hundred of the leading authorities
in France to report on this subject. Their report, printed in four quarto
volumes (1895), is to-day (1897) by far the best single source of information
on these subjects. They have proposed what appeared to them practicable
standard tests for nearly all kinds of structural materials.
Evidently no complete standardization can be effected for tests on entire
structural forms, since these vary in shape, size, and disposition of parts,
but specimen tests can be standardized since all significant conditions can be
made uniform.
258. Mechanical Tests Classified. In a general way we may divide
mechanical tests of building materials into the following classes :
With reference to the method of applying the loads we have
(1) Static Tests, or those made with gradually increasing loads, such as
the ordinary tests in tension, compression, cross-bending, torsion, and shear-
ing.
(2) Dynamic Tests, or those made with suddenly applied loads, as by a
falling weight.
(3) Wearing Tests, or those made for determining resistance to abrasion
and impact, as in the case of paving-materials.
With reference to the character of the test specimen we have
(1) Specimen Tests, or those made upon specimens of the material
specially prepared and given standard forms and dimensions.
(2) Structural Tests, or those made on full-sized structural forms, as
bridge members, brick piers, pipes, wire ropes, chains, riveted joints, etc.,
or on the structure as a whole, such as boilers, simple trusses, frames, and
various parts of machines.
Complete standard rules for making tests of structural materials can be
304 THE MATERIALS OF CONSTRUCTION.
adopted for making all kinds of tests on specially prepared specimens, but
they can be only partially prescribed for tests of structural forms.
259. General Remarks on Testing-machines. ffhe following considera-
tions apply to testing-machines and testing-appliances in general :
1. The weighing apparatus should be quite independent of the loading
apparatus, the former usually being fixed and the latter movable.
. 2. In lever machines the length of the knife-edges must be proportioned
to the maximum loads in order not to be crushed down, and they should be
so placed that all will receive their share of the load. They must also be so
mounted as not to change the leverage by any reaction displacement which
may occur To insure this, the knife-edges must be attached to the levers,
and the bearings to the platform.
3. The knife-edges and bearings of any beam must lie in the same
straight line, and this line should lie in the gravity axis of the beam and its
rigid attachments. This is especially necessary for the weighing-beam itself,
so that its vertical angular movement may not disturb the counterbalancing.
If the poise is moved by a cord over a pair of pulleys, this cord should be
attached to the poise-hanger in this same axial line, so that the pulling of
the poise may not supply a leverage on the beam to raise or lower it.
4. Manometer machines have many peculiar -errors. For example, any
air-bubble in the indicating liquid vitiates the results by its own change in
volume under pressure. Again, the exact area of surface subjected to pres-
sure is always uncertain.
5. The weighing apparatus should be so constructed as to be readily
verified by the imposition of known weights, and the parts should be open
to inspection and easily repaired and kept in order.
6. A precision of 1 in'250 has been considered sufficient.* This is a
proportional error of 0.4 of 1 per cent.
7. The loading should proceed gradually and uniformly, and not by
sudden increments as by large pump-pulsations, or by the adding of over-
weights by hand to the weighing-beam. The rate of loading should also be
under perfect control.
-8. The machine should be so constructed as to permit the free use of
appliances for measuring distortion of the specimen by some suitable device.
9. If used for compression tests, one of the bearing-surfaces should be
slightly adjustable to accommodate the machine to the non-parallel faces of
the test-block, and these bearing-surfaces should be harder than any material
tested by them. The neutral axis of these bearing-plates- should coincide
with the axis of symmetry of the applied forces as transmitted by the
machine to the specimen. For these tests the machine should have a very
slow movement.
10. If used for cross-breaking tests, it should be furnished with means for
the deflection. To do this properly a rigid connection must be
* This standard is given by the French Commission.
MECHANICAL TESTS IN GENERAL. 305
established between the two end bearings to the middle bearing (or bear-
ings), through parts not under stress, in order that the loading of the
specimen may not disturb this rigid relation.
11. Torsion testing-machines should apply the torsion movement as a
true couple and without developing any tensile or bending stress in the
specimen.
12. Impact testing-machines should as far as possible satisfy the condi-
tions imposed in Art. 292. That is, as far as possible, the entire energy of
the blow should pass into the specimen. The falling weight should be held
to its course either by vertical guides, in the case of a falling weight, or by
n pendulum mounted on a transverse axis resting on knife-edge bearings.
The former method is to be preferred. The falling weight should be sym-
metrical in form, with suitable guiding attachments, to be formed (cast) in
one piece, of hard metal, with its centre of gravity as low as possible. The
height of the weight should be greater than the width between the guides,
which latter should be quite rigid, true, and vertical, and should offer no
frictional resistance to the falling weight. The supporting mass should be
very great as compared to that of the falling weight. The French Commis-
sion recommend that it be at least 15 or 20 times that of the striking body.
Impact tests can only be standardized by using exactly similar appliances in
all respects, including the supporting blocks and the foundation on which
these supports rest.
260. The Effect of the Rate of Loading on the Results of the Test, The
French Commission quote M. A. Le Chatelier on this subject as follows:
" Metals do not respond instantly to the deforming action of external forces.
These deformations, both elastic and permanent, continue to increase with
time, and the termination of the instant when the deformation correspond-
ing to a given load has been fully completed depends only on the exactness
of the measuring instruments employed. Speaking absolutely, this condition
of equilibrium is never attained, and we may say the deformation increases
indefinitely. It approaches, however, a limiting value (as an asymptote),
especially in the case of elastic deformations, and even for permanent defor-
mations the time may be found beyond which the remaining deformation
will not exceed a given amount."
It is admitted, however, that for metals at ordinary temperatures a ten-
sion test (for instance), extended over a few minutes' time, gives practically
the same results, in every respect, that would be obtained by any slower
imposition of the load. This has been thoroughly established by Bauschin-
ger, as well as by Considere and Le Chatelier. Zinc and tin are exceptions
to this law, comparatively small external forces causing final rupture if these
forces continue active. Copper and aluminum also fail under a somewhat
smaller permanent load than is required to produce rupture in ordinary
tests.* It is well known that timber yields continually under about one half
* Report of the French Commission, vol. i. p. 93.
306 THE MATERIALS OF CONSTRUCTION.
the breaking-load, and this half -load, permanently placed, may ultimately
cause failure. For all kinds of test specimens of wrought iron and steel, and
other structural metals, at ordinary temperatures, a test extending to one
minute or more may be considered as giving normal results.
For very rapid tests, or where rapture occurs in less than a minute, the
breaking-load increases and so does the ultimate elongation. In the case of
soft steel, however, which has a great local reduction of area, the elongation
diminishes as the test period decreases, reaching a minimum for a period of
about one minute, and then the elongation rapidly increases again for very
quick tests, because it then reduces in cross-section more uniformly through-
out its entire length.
It has been shown by M. Considere * that the. stress-diagram giving the
simultaneous relation between stress and deformation is very different under
very quick impositions of load, as in case of a shock, or impact, from the
diagram for ordinary laboratory static tests of more than one minute dura-
tion. His results are shown in Fig. 52, p. 79. This subject is fully dis-
cussed in Art. 292, where impact tests are described.
In conclusion it may be said that in all ordinary tests of metals the test
period should fall between one and six minutes.
261. Significant Limits of Deformation. Ever since the properties of
building materials have been studied, " the elastic limit " has been defined
both as the greatest load which will not produce a permanent set, and as the
greatest load at which deformation remains proportional to the load, or
stress. It has commonly been supposed that these two limits were one and
the same, and in the commercial testing, which has probably been carried on
in America to as great an extent as in any country, this so-called " elastic
limit" has been observed as the point at which the deformation increases
rapdly under a constant load, or, as it has been called, the " yield-point."
The French Commission has studied this subject with care, and after mature
deliberation a majority have agreed to adopt three critical points, as follows:
1. The elastic limit (" la limite (Telasticite"\ or the unit stress beyond
which a portion of the deformation remains as a permanent set. (Point E
of the stress-diagram.)
2. The proportional elastic limit (la limite d'elasticite proportioned, or
limite des deformations proportionelles), corresponding to the point where the
deformation ceases to be proportional to the loads. (Point P of the stress-
diagram.)
3. 'I he apparent elastic limit (la limite d'elasticite apparente, or origine
des deformations sous charge constante), corresponding to the point where the
deformations increase rapidly without any increase in the force exerted.
(Point F of the stress- diagram.)
While in a scientific study of metals it may be important to determine all
these three limits whenever they occur, in practical or commercial testing
* French Commission Report, vol. n. p. 344.
MECHANICAL TESTS IN GENERAL.
307
it will be found sufficient to observe the third one only, or the second only
in case the third limit does not obtain for that material. The first of these
limits has seldom been determined, since it involves a release of the stress
by removing the load. This involves a great loss of time, since to determine
this limit accurately the load would have to be released for each small incre-
ment of, let us say, 1000 pounds per square inch for iron and steel. The
second and third limits can be determined after the test has been completed,
provided simultaneous readings of the load and deformation were taken
during the test at frequent intervals, or provided an autographic stress-
diagram was made by suitable attachments to the machine and test specimen.
The third limit is commonly, but not accurately, determined in commercial
testing in America, by " the drop of the weighing-beam," this marking the
point where the deformation increases for very ductile metals under a con-
stant load.
With unrolled (or unforged) castings of the various metals any load pro-
duces some appreciable permanent set, so that the first elastic limit does not
exist for such materials. It is probable that this is also true to an inappre-
FIG. 247. Average Curve of Four Tests of |-in. Wrought-iron Rods.
Rep., 1888.)
(Wat. ATS.
ciable degree of all the rolled metals, so that this is a very unreliable and
unsatisfactory test of any important property of materials. The law of
proportionality is much better defined, but, owing to the mixed character of
the elementary forms entering into the composition of all metal products,
even the purest, this law is often found to fail when the most refined means
of measurement are employed, since there then seems to be no strictly con-
stant ratio between the load-increments and the corresponding increments
of the deformation, and hence the exact point where this ratio begins to
308
THE MATERIALS OF CONSTRUCTION.
change is in these cases difficult to fix. In such cases, by plotting the
deformation to a very large scale with the loads, one can determine graphi-
cally, as in Fig. 2-47, about where the deformation increments begin to
increase. The second and third elastic limits are marked on this diagram
" true elastic limit " and " yield-point " respectively. ^The " apparent
elastic limit," or " yield-point," is the most important and significant of the
three, as well as the most easily determined, but in high carbon-steel,
especially in hard-steel wire and in all cast metals, and often in wrought
iron, this point does not appear, in which case the " proportional elastic
limit" takes its place, as shown in Fig. 249, where it is marked as the
SQ000
- Y/f/./TPtf/WT
A/ (9 A , r / A/
P ft
C'L
T
/r
02
u .W/' J02
FIG. 248. Typical Stress-diagram of Hard-drawn Brass. ( Wat. Ars. Rep., 1886.) (The
" U. S. Elastic Limit " is that given in the published report.)
" true elastic limit." Sometimes, also, when point ^is very marked, point
P is found above it, as in Figs. 7 and 8, pp. 15 and 16.
262. All these Absolute Elastic Limits Unsatisfactory. It is proposed
now to show that no one of the three definitions of elastic limit given in
Art. 261 can be used in practice. They all will be shown to be either abso-
lutely indeterminate or wholly dependent on the delicacy of the measuring
apparatus, rather than on the qualities of the material tested.
Thus the first two definitions undertake to fix a limit, and evidently the
position of this limit is simply the point where either the permanent set or
the deviation from a linear relation between load and deformation becomes
measurable. If one can measure accurately to 0.0001 inch, he will discover
these limits earlier, or at a lower stress, than if he can only measure to 0.001
inch. The French Commission recommend that measurements be made to
MECHANICAL TESTS IN GENERAL.
309
the nearest 0.001 mm. or to ^-^or inch. To discover a permanent set,
furthermore, requires a constant release of the load, which is liable to disturb
the deformation-measuring apparatus, and in any case the load at which a
permanent set occurs can only be said to lie between two particular loads,
the greater of which has produced the first permanent set observed. The
time and trouble involved in releasing the load so often will act to remove
this test from nearly all scientific and commercial work.
It has generally been assumed that the first two definitions of elastic
limit given in Art. 261 locate identical points in the stress-diagram, but
#0000
\L
m
tit
/
r
yyt'si
/
"V" Lr JH / /
I
'0 .0K .0/0 .0/S .020
FIG. 249. Tension Tests of Steel Piano- wire. Gauged length 6 in., Diam. 0.04 in.
(Wat. ATS. Rep., 1894.)
with the most delicate measuring appliances it is found that these two defi-
nitions may locate points very far apart.
The third definition is also indefinite, since it remains a question as to
which load is to be taken, the higher load at which the first great permanent
elongation occurred, or the lower load under which this elongation continues
to spread throughout the entire length of the bar. These often differ as
much as from 3000 to GOOO pounds per square inch. (See Figs. 7 and 8.)
Again, when the most delicate apparatus is employed, several specimens
from the same bar of the most uniform material may give elastic limits of
either of the first two kinds which differ widely from each other and hence
become mutually contradictory. In other words, such delicate tests are quite
worthless for all practical purposes, i.e., the results are not characteristic.-
263. The Apparent Elastic Limit. The term " relative elastic limit "
was coined by the author in 1891, to be used in his work of testing timber
for the Forestry Division of the U. S. Agricultural Department. He then
310 THE MATERIALS OF CONSTRUCTION.
defined it as the point on the stress-diagram (of tests in cross-bending)
where the rate of deformation is fifty per cent greater than it is at the origin
(see Figs. 247, 248, and 249). To find it draw a tangent to the stress-diagram
at the origin, and then lay a parallel ruler on a line making with the load
line an angle whose tangent is fifty per cent greater than that of the original
tangent line, and then move the ruler until its edge becomes tangent to the
stress-diagram, and draw such tangent line. The " relative elastic limit"
is then located by eye as this point of tangency. In the tests of wooden
beams this point usually falls on the diagram where its curvature is about
the most rapid (radius of curvature a minimum), and in all tension stress-
diagrams of the various metals it will be found to mark a well-defined point,
whose coordinates are practically fixed and constant for the same material.
While this point is certainly beyond all " true elastic limits " when defined
as limits, yet in the metals it will be found to fall very little beyond such true
limits. In fact it commonly falls below the " elastic limit " recorded in all
the Watertown Arsenal Tests, where the deformations were measured and
recorded to the nearest T o~Vr?r ^ an i ncn > as shown in Figs. 9 and 248,
and many others in Chapter XXIV.
The French Commission make use of the term "apparent elastic limit "
to indicate what in England is called the " yield-point " or " break-down
point," and in Germany is called the " beginning of great elongations."
But with such materials as cast iron, high carbon-steel, and often with
wrought iron and other metals there is no " yield-point," or no point where
the material deforms under a constant load, unless it be at the point of
maximum load, to which of course it is not intended to apply. The term
" apparent elastic limit " in this sense, therefore, has not a universal appli-
cation, and hence cannot be used as the commercial elastic limit to be em-
ployed in practical tests.
The term " apparent elastic limit " has not as yet come into use in
English; and if it now be defined arbitrarily as the term " relative elastic
limit " is defined above, it could have this specific meaning and would be of
universal application.
When so located it will be found at practically the same point on all tests
of like kind on similar materials. It is therefore characteristic of the
material.
It does not require the use of expensive and troublesome appliances for
its accurate location. Relatively crude appliances can be used to measure
the deformations ; and though these may not fall on a smooth curve, a mean
line drawn through them, as the stress-diagram, furnishes a satisfactory
means of locating this " apparent elastic limit. " It is therefore a practically
determinate function.
It can readily be determined from an automatically registered stress-
diagram, of any description, thus admitting of a continuous and uninter-
rupted progress of loading, so that the conditions of a test can be exactly
duplicated as to speed, which cannot be done when the test is stopped to
MECHANICAL TESTS IN GENERAL. 311
take deformation readings. It is therefore a true relative limit of the elastic
field.
Although this point is slightly beyond the true elastic limit, it will mark
a point corresponding to a permanent set much less than can be measured on
any scale by the naked eye,* and hence it may be regarded as the true elastic
limit for commercial purposes.
It serves perfectly to classify materials as to the maxim am loads they can
resist without receiving deformation which would injure them for continued
service, these being the real " ultimate loads" for all practical purposes.
It marks, therefore, the most valuable and important property of all engineer-
ing or building material. In other words, it is the most essential character-
istic point on the stress-diagram.
In the opinion of the author of this work, therefore, an " apparent elastic
limit," defined and determined as here described, is the best if not the only
satisfactory solution of this troublesome question. The fifty per cent increase
in the rate of deformation was chosen as being about the least which would
mark a well-defined point of tangency on the stress-diagram. Since it
always marks a point which corresponds to an extremely small permanent
set, there would seem to be no objection to its use.
* Seldom more than one one-thousandth of one per cent of the measured length. It
is a maximum in the case of the high-grade, hardened steel wire, shown in Fig. 249,
where it reaches about ^ of one per cent.
CHAPTER XV.
TENSION TESTS.
264. Significance of Tension Tests. Tension tests are at once more com-
mon, more readily made, and more useful in revealing the true character of
a metal than any other kind of mechanical test. In fact, when other kinds
of tests are made it would commonly be well to accompany them with a few
tensile tests for the purpose of being able better to coordinate the results
with those obtained on other materials by similar tests, or on like materials
by different tests. In this connection, however, it is well to remember that
all the metals are wanting in strict homogeneity, and that they may be
regarded as aggregations of more or less dissimilar elements embedded in a
common matrix, somewhat like granite. (See Arts. 105 and 108.) For
instance, the planes of rupture will be different for different kinds of tests
on the same specimen, and hence the strength developed will be that of a
different combination of elements in each case. Also, the strength to resist
various kinds of stress may lie 4n entirely different elements of the aggrega-
tion, as, for instance, in the case of cast iron the strength to resist tension is
the strength of the graphitic carbon matrix in which the iron crystals are
embedded, while the strength in compression is largely the strength of the
iron crystals themselves.*
What we call the maximum strength of the material, therefore, or its
strength at rupture, is not usually the sum of the maximum resistances of the
several elementary portions of the cross-section, since they do not all distort
equally. It is often the case that actual rupture occurs successively over
many elementary portions of the broken section before the final failure
occurs. More especially is this true of the elastic limits of the material,
while with iron and steel castings this failure in detail is so prominent as to
cause the stress-diagram to be a curve almost from the beginning of the load-
ing. Here, too, the irregular shrinkage often leaves very great internal
stresses in the body, which causes some portions to come to their elastic
limits and ultimate strength much earlier than others, again giving rise to a
curved stress-diagram.
For these reasons we find, when the most delicate means are employed to
measure deformations under increasing loads, that in almost no case is the
* M. Osmond.
312
TENSION TESTS. 313
deformation strictly proportional to the load, and that even very small loads
will produce some little permanent deformation or set. This is why the
definitions given in Art. 261 must depend on certain arbitrary limits of
deformation and set, and are not true absolutely as they have hitherto
commonly been defined.
The tension test is especially well calculated to show what local irregulari-
ties may be found in a finished product, and to indicate to what extent the
work of forging (rolling or hammering) has produced that degree of homo-
genity expected of it.
The tension test is more readily standardized than any other so as to be
independent of " personal equation " and of variations in the testing-machines
employed. It also demands the least amount of preparation of the test
specimen, if tests are to be made only for commercial purposes. Except for
the inherent want of uniformity or of homogeneity mentioned above, there-
fore, the tension test may be made to give typical and uniform results, and it
should be considered as the best single test to make on any of the metals.
265. Selection of the Test Specimens. Test specimens may be taken
either from the finished product or from the material when poured if it is
derived from a fluid condition. In American steel mills it is common to
roll (or hammer) a test rod from a small ingot poured from each heat,
whether it be of the Bessemer or the open-hearth process, and the maker
depends on the tests made on this bar to guide him in the further use of the
ingots poured from that heat. The user also is often satisfied with these
tests, especially when his requirements are not very rigid.
In making iron and steel castings it is common to have test samples
cast in the same moulds with important castings, and joined thereto, so that
they shall represent, of necessity, the identical metal of which the structural
form is composed. If special moulds are used for these test specimens, they
should be of dry sand, under a head of at least eight inches, and with an
inclination of at least one in five to allow the escape of the gases. With
very heavy castings (over three inches in thickness) test specimens may be
cut from the head itself which is cut off from the upper end of the casting.
If the specimens are taken from rolled structural forms, they should be
taken from the thicker parts, which have received the least work in the
rolls. The thinner parts are always harder, have higher elastic limits and
greater ultimate strength.
With wrought iron a great difference will be found in specimens cut with
and across the direction of the rolling, the former having much higher
strength and a greater ductility. In steel plates there is little difference,
and in rolled brass and copper plates there is no difference. In the case of
the bronzes it is necessary to have test samples poured from different parts of
the same melting, as the mixture changes its characteristics rapidly when in
a melted state.
266. The Preparation of the Test Specimen. In order that the test
specimen may fairly represent the material under examination, or the par-
314 .THE MATERIALS OF CONSTRUCTION.
ticular plate, or bar, or rolled form from which it is to be taken, it is neces-
sary to observe a number of rigid requirements.
The specimen must be obtained by cutting it out in a way that will leave
it perfectly straight. If it is bent in getting it out, it should be heated to
straighten it; but this may often change the original molecular arrangement,
and should be avoided if possible. When the specimen is cut from a larger
portion of a plate or rolled form by shearing, it will invariably take a curved
form. In this case the plate, or form, should be sheared aivay from the
specimen, in narrow slices, so as to leave the test specimen unbent. If the
specimen is bent and then straightened, it raises the elastic limit and hardens
the metal, the same as any other kind of cold working. Instead of shearing,
some milder process, such as planing or drilling or sawing, should be resorted
to to obtain the test specimen. For, besides the bending action on the bar
as a whole, the effect of the shearing or punching is to seriously injure the
metal for about an eighth of an inch beyond the sheared surface, leaving it
so non-ductile, or brittle, that it will not elongate appreciably, and hence
under a tensile test these surfaces will be severed very early in the test, and
the cracks so started may cause the remainder of the cross-section to tear
asunder in detail. To prevent this action on sheared or punched specimens,
at least an eighth of an inch of thickness should be removed from all
punched or sheared faces, by reaming, planing, or filing. The effect of not
doing this is shown in various figures in Chapter XXVI, where both punched
and drilled test specimens of one-fourth inch iron and steel plates had been
grooved for testing, leaving varying widths of metal between the bottoms
of the grooves. The effect of this variation in width is also here shown.
Various bending tests there shown also exhibit the weakness resulting from
shearing and punching.
The ordinary lathe, planer, and milling-machine tools are not suitable
for i\iQ final finishing of the specimen, as they tear and bruise the remaining
metal, giving rise to a condition favorable to the starting of incipient
fractures at the surface of the specimen. These tools may be used for
roughing out the shape desired, but it should be finished with the file, and
in case of the softer metals, like copper, the file should be followed with fine
emery-paper.
Castings should be cut down about an eighth or a tenth of an inch from
the rough exterior, on the reduced section, and they should also be trued-up
on the ends which are to be gripped.* Rectangular edges should always be
taken off by a file to remove any incipient cracks or irregularities which may
be left here by a kind of crushing-down action of the tools operating on one
or both of the plane faces meeting on these lines. Test specimens of the
softer metals should never be beaten with a steel hammer, but with wooden
or copper mallets, if it is necessary to use such means to straighten them.
Standard shapes of test specimens are shown in Fig. 251.
* For commercial purposes a cheap form of specimen which is tested without turn-
ing down at all is shown in Chapter XXIV.
TENSION TESTS.
315
KULES OF THE FRENCH COMMISSION FOR TENSION TESTS.
1. The test should be continuously progressive.
2. The duration of the test should in a general way increase with the volume
of the specimen.
3. For standard tests on specimens of ordinary dimensions of which the sectional
area is not more than one square inch (600 sq. mm.) and the measured length not
more than eight inches (20 cm.) it seems that the duration of the test should be
included between one and six minutes.
4. For test specimens having a thickness less than 0.2 in. (5 mm.) the duration
of the test should be less than thirty seconds.
5. It is necessary to avoid, especially with soft metals, producing a sensible
heating of the test bars.
267. Standard Dimensions of Tension-test Specimens. It has been
shown * that so long as the test specimens of a given material maintain the
FIG. 250. Showing the Constancy of the Strength and of the Elongation when
= = a constant (8 in this case). Each result is the mean of five tests on the same
material. (French Com. Rep., 1894, vol. in. p. 72.)
* By MM. Lebasteur, Marie, and Barba.
23.
See Rep. French Commission, vol. in.
316 THE MATERIALS OF CONSTRUCTION.
same relative dimensions, or are geometrically similar in form, the strength
and the percentage of elongation remain constant, as shown in Fig. 250.
The French Commission have, therefore, adopted the relation f = 6G.67^4,
or for cylindrical specimens I 7.2d, where I is the measured length on
which percentage of elongation is computed. An eight-inch specimen would
then be 1.11 inch in diameter, or nearly 1 sq. in. in area of cross-section.
Since this relation between I and A was chosen for convenience (for I = 200
mm., A = 600 sq. mm.), persons using inch units might well choose the
relation I = 8^7, or
F = 8IA (1)
For square sections, therefore,
I - 96, (2)
while for round sections
(3)
these being intermediate between the French and the German standard
dimensions.*
Equations (1), (2), and (3), therefore, may be employed in finding what
length of specimen to use to give comparable and consistent percentages of
elongation, when the excessive elongation near the broken section is included.
If it is practicable to so prepare the specimen as to make the area of the
cross-section nearly constant, then a fixed length of specimen could be used
for all tests. Thus for the standard length of eight inches the diameter of
round specimens would be one inch. Where the cross-section varies from
this the lengths should vary in accordance with equation (3). Thus
For diameters of inch make I = 7 inches.
" f "
" Z= 6
" "
" 1 = 5
"* "
" 7 = 4
< t 3 a
8
Z = 3
u i u
' tf ? = 2
For plate tests we have, from (2),
F = SIA = Slbt, or I = 9 VU (4)
Since it is common to prepare several of these together in a milling-
machine, it is desirable to have a common width for these tests, and a width
of one inch has been usually employed in America. This may be done if
the lengths are varied to give comparable results. Thus, from eq. (4),
I = 9 Vbt, the following scheme of lengths and thicknesses is derived, the
width being one inch in all cases:
* The German Commissions have agreed on I = 11.3 V'A=- IQd for round section
and 11.36 for square sections.
TENSION TESTS. 317
For plates inch thick make I = 4^ inches.
" " f " " " I 5i
u I if I- 6 1
ff fi 5 ff if a 7 __ r~
fi if 3 fi if ii 7 __ /J'S
r
if if 7 if if if 7 _ Q J_ . .'
For thin sheet metal make the measured length always four inches, and
use three standard widths as follows :
For thicknesses from 0.1 inch to 0.2 inch make width = inch.
" 0.05 " " 0.1 " " " = f "
less than 0.05 inch " li =
All these relative dimensions agree closely with those recommended by
the French Commission (1895).
The measured portion of the reduced section (called I in the above
discussion) must be removed from the shoulders by a distance at least as
1
. L jt
, ^ j , M
f 1.
d ct>
FIG. 251. Standard Dimensions for Rectangular and Cylindrical Test Specimens.
great as the diameter or thickness of the test bar, in order to avoid the effect
of these enlarged portions in reducing the elongation. The* reduction of the
percentage of elongation near the ends is well shown in Fig. 252, where the
bar had a total reduced length of 15 inches and a diameter of 1^ inches.
Steel bars of this shape will break near the centre, while wrought-iron bars
will break at various distances from the centre, with a more uniform distri-
bution of the elongation.
268. Tetmajer's Analysis of the Elongation of Tension-test Specimens.*
The typical forms of tension-test specimens are shown in Fig. 251, where both round
and rectangular sections are given. It has been shown by numerous experiments
that the strength and the reduction of area are somewhat dependent on the form of
the test specimen, while the elongation is very greatly dependent on these relative
dimensions. This is well illustrated in the reproduced photographs shown in
Fig. 10, and also in the elongation-diagrams Figs. 252 and 253. Here are shown first
the original specimens, then the specimen stretched to its maximum loading, the
elongation being nearly uniformly distributed over the length, and finally the speci-
men greatly reduced at one point where rupture is about to occur. It is evident,
* Tetmajer's Communications, vol. iv.
318
THE MATERIALS OF CONSTRUCTION.
PIG. 252. Showing Distribution of the Elongation over
2 in. in Diameter. (Rep. Fr. Com., vol.
-i ----- 1
a Steel Bar 15 iu. Long and
in, PI. n.)
FIG. 253. Showing the Variation in the Distribution of the Elongation of the Several
Inch-spaces of Six-inch Test Bars of Steel and Wrought Iron 0.56 in. in Diameter.
(Wat. Ars. Rep., 1890.)
TENSION TESTS.
319
from a study of these specimens, that the total elongation of tension-test specimen
may be divided into two distinct parts, namely:
1. The general elongation.
2. The local elongation.
It will further appear that the local elongation is nearly the same in all cases,
and is practically independent of the length, while the general elongation, having
occurred uniformly along the bar, is directly proportional to the length.
If I = measured length of specimen,
Al total elongation,
A = proportional distributed elongation
distributed elongation
~~T~ ~'
Ah = total local elongation
then A may be found for any given test by measuring Al for two lengths of the
same specimen, each to include the broken section. Since the standard length
of specimen is 8 inches (200 mm.), it usually will be found convenient to use 8
inches and 4 inches for these two lengths. If the specimen be marked originally
every inch, then after it has broken two sets of these marks may be chosen, each
pair to include the fracture, and to be points originally 8 inches and 4 inches apart
respectively. Then we may have
Ah = 8A + A1 9
Ah = 4/1 + Ah
Ah - Ah = 4A
or
A \(Ah Ah] (1)
Til is function, A, obtained in this way, is independent of the length of the speci-
men, and is the true characteristic elongation of the material.
It would seem that this function is the one to be generally adopted as the true
index of the ductility of the material. Unfortunately custom has established -j as
20 40
FIG. 254. Elongation of a Specimen of Copper 200 mm. Long for the Loads as given.
(Fr. Com. Rep., vol. in. PI. vn.)
the ductility function, or "the percentage of elongation, 1 ' and this varies greatly
with the ratio of length to form and area of cross-section.
Professor Tetmajer shows that the following relative dimensions give practically
equal percentages of total elongation :
(A) CYLINDRICAL SPECIMENS.
Diameters . . ...
0.4 in
0.6 in.
0.8 in.
1.0 in.
Observed length
4.0 in.
6. 4 in.
8.0 in.
10.0 in.
Mean observed elongation, " Phoenix" steel. ..
30.1*
30. 4#
30. 5#
30. 6#
320 THE MATERIALS OF CONSTRUCTION.
(B) RECTANGULAR SPECIMENS, ALL 0.4 IN. THICK.
Ratio : b - =
t
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
Observed length, iiiches. .
4.0
4.8
6.0
7 2
8.0
8.0
8.0
8.4
9.2
Mean observed percentage
of elongation
270
27.2
27 2
26 8
26.1
25.7
26.1
26.7
27.0
Each of the above observed elongations in Table (A) is the mean of five tests and
in Table (B) of ten tests, and the results indicate that equivalent lengths were used.
When the length was taken as 8 inches in each case, the percentage of elonga-
tion in Table (A) ranged from 26.5 to 32.4, while in Table (B) it ranged from 21.3
to 28.6. These results are consistent with the rules laid down by the French Com-
mission and which the author has interpreted approximately in English measures in
the previous article.
269. The Time Function of Tension Tests is not an important one.
Bauschinger has shown that within the ranges of practicability the time
element is of no consequence. This is also shown in Fig. 255, where results
~\70
PEf.D /A
CO
40
30
/
234-
FIG. 255. Showing the Effect of Pulling Speed of the Testing-machine on the Recorded
Results with Structural Steel. (Campbell's Structural Steel, p. 253.)
obtained with the greatest rapidity obtainable in the American standard
testing-machines are compared with results from very slow tests, with very
little difference in the results. The greatest recorded differences are found
in the elastic limit; but as these were all observed by the "drop of the
beam," it is likely that the speed had more to do with the obtaining of the
reading than it had with the real action of the material itself.
270. Tension-test Machines. There are three general types of universal
(tension, compression, and cross-bending) testing-machines on the American
market, viz., hydraulic machines, screw-gear machines, and the Emery
machines. The hydraulic-power machines have, however, been practically
abandoned in favor of the screw-gear for all experimental and commercial
TENSION TESTS.
321
purposes, while for high scientific accuracy and an incredible delicacy
nothing ever made can compare with the Emery machines.* The first two
varieties of machine owe their present high state of development largely to
Mr. Tinius Olsen, and the Emery machine was originally designed by Mr.
A. H. Emery, but has since been greatly simplified and improved by Wm.
Sellers & Co., the present manufacturers.
The hydraulic machines have some advantages when operated by hand,
but they all have the disadvantage that it is impossible to maintain a given
* The hydraulic and screw-gear machines are made by Tinius Olsen and by Riehle
Bros., and the Emery machines by "Win. Sellers & Co., all of Philadelphia, Pa.
322 THE MATERIALS OF CONSTRUCTION.
load without continuous pumping to supply the small leakage which always
occurs.
The Olsen screw-gear 100,000-lb. testing-machine, shown in Fig. 256,
will be taken as a type of the testing-machines which are almost exclusively
employed in America. The power is applied to the pulleys 26 and 27, the
former used for direct (downward) and the latter for reverse (upward)
motion of the moving cross-head, 5. Extremely slow speed is obtained by
throwing into contact the small friction-gear 35, operating upon the large
wheel 34, which is rigidly attached to the driving-shaft. This is effected by
drawing on the chain 37 by turning the hand-wheel 39, which tightens the
band 42 and starts 35 to revolving. The band-wheels 2G, 27, and 43 all
revolve freely on the driving-shaft, except as 2G or 27 is made fast to it by
the friction-clutch 28, 30, through the hand-lever 33.'
A medium speed of the moving cross-head is obtained by simply throwing
26 in gear by the hand-lever 33, and a high (upward or downward) speed is
attained by changing the gears by means of lever 25, the upward speed
greatly exceeding the downward because of the different band connections
on the direct and on the reversing-pulleys 26 and 27 respectively.
The moving cross-head, 5, is brought down by the turning of four screws,
one at each corner, only two of which are visible in Fig. 256. A tension-
test specimen is placed between this moving head-piece and the fixed cross-
head above, being gripped in each by means of hardened steel wedges with
grooved faces. The pull on the specimen is thus transmitted through the
four cast-iron columns, 2, to the weighing-table, 3, which rests by means of
fixed spurs upon the three weighing-levers, 117. Between these and the
weighing-beam, 118, there is one intermediate multiplying-lever (not num-
bered). The large poise, 106, in this machine is supposed to be moved by
means of the screw 105, which is under automatic electric control. When
the beam lifts, the screw is put in motion; and when it leaves its upper con-
tact the screw stops its motion. The weighing-beam is thus automatically
maintained in constant balance. If operated by hand, the large poise, 106,
is set forward by full revolutions of the screw by means of the handle shown,
and the intermediate loads indicated by balancing with the small poise shown
at the right-hand end of its scale, 118'. When operated automatically this
poise is not used, and the fractional part of the total load is read on a grad-
uated disk attached to the screw at the left-hand end, but not clearly shown
in the figure.
Compression tests are made by attaching a compression-block to the lower
side of the moving cross-head, and inserting the specimen between.it and the
weighing-table, 3.
Cross-breaking tests are made by placing the end bearings on the weigh-
ing-table (or on an I-beam resting on this table if the specimen is long), and
attaching a knife-edge bearing to the lower side of the moving cross-head.
A machine in nearly all respects quite similar to the above is that shown
in Fig. 257, made by Riehle Bros. The poise here is moved by a chain
TENSION TESTS.
323
passing over a driving-pulley, which pulley is operated either by hand or by
power under electrical control, the same as the screw in the Olsen machine.
Only two screws are here used for moving the pulling cross-head, instead of
four as in the former case. Both forms of machines are made in the highest
style of the art, both being the survival of the fittest in a long succession of
types of testing-machines. They are by far the most useful and convenient
testing-machines made,* and are not likely to undergo much change in the
FIG 2.37.
future. (They are now, 1896, being sold in Europe.) The speeds at which
the machine shown in Fig. 257 may be driven directly are as follows: T V in.
per min. ; in. per min. ; f in. per min. ; 1-j- in. per min. ; and 8 in. per min.
For tests of low ultimate strength, speeds of in. per min. and of 4 in. per
min. can also be used. The higher speeds, down to f in. per min., can also
be used in raising the moving head. By changing the speed of the main
shaft from which power is obtained all these speeds can be increased or
diminished at pleasure. The speeds as given above are for 150 revolutions
per minute of the driving-pulleys on the testing-machine.
In Fig. 258 is shown a small screw-gear power machine, made by Eiehle
* The author offers no apology for not giving descriptions of any of the scores of
styles of machines which have been built and which are still in use mostly in Europe.
They will never be built, bought, or used in this country, and most of them can be
found illustrated in the Report of the French Commission, 1895, vols. n. and in.
324
THE MATERIALS OF CONSTRUCTION.
Bros, in capacities of 20,000 Ibs., 30,000 Ibs., 40,000 Ibs., 50,000 Ibs., and
60,000 Ibs., with hand-power attachments, and automatic weighing appli-
.ances if desired. With the 20,000-lb. machine the hand-power does very
well,* although steam or electric power is always preferable. Autographic
FIG. 258.
recording attachments (see Art. 275) are attached to these the same as with
the larger machines. They are too small, however, for general commercial
purposes.
271. Gripping Devices. A great variety of grinning devices have been
employed, such as eyes and pins, shoulders and snlit Beeves, screws-treads
and nuts, and plain bars with wedge-grips T his last *orm has now replaced
all others in America except such as mav still be employed on some of the
older machines. For round specimens notched grins are used, while with
square or flat specimens the plain wedges are employed. The Eiehle plain
grips are swelled in the centre so as to grip the specimen hardest along its
* These small machines are the best patterns for students' use. It is better to have
several of these than one larger machine. They serve almost every purpose in a course
of study on the strength of materials.
TENSION TESTS.
325
FIG. 259.
326
TEE MATERIALS OF CONSTRUCTION.
axis of symmetry. The Olsen grips are swivelled on spherical bearings at
the back to enable them to more readily adjust themselves to the specimen.
FIG. 260.
The wedge-grips can be nsed with plain bars and plates, without any reduc-
tion of section, as well as with specimens specially prepared by turning down
or by a milling-machine.
TENSION TESTS.
327
272. Similar Machines for Various Special Purposes. In Fig. 259 is
shown an Olsen machine designed for testing full-sized structural specimens,
whether tension or compression members, or beams, and made in capacities
of 200,000, 300,000, and 400,000 Ibs. These machines have all the rates of
motion, and the automatic weighing and autographic recording appliances
described with the smaller machines. The height may be made almost any-
thing which may be desired.
In Fig. 260 is shown a similar machine of 300,000 Ibs. capacity, in which
the four cast-iron corner-posts are replaced by two large compression-screws.
Fig. 261 illustrates a convenient and cheap machine for testing hoop-iron
FIG. 261.
and round bars, in tension only, of 20,000 Ibs. capacity, suitable for office
use.*
In Fig. 262 is shown one form of a wire-testing machine, made by Olsen,
in capacities of 10,000, 15,000, and 20,000 Ibs. It is also adapted for test-
ing band-iron and other forms, and could also be used for compression and
cross-breaking. (See also other wire-testing machines in Chap. XXXIII.)
Fig. 263 shows a form of cloth- and paper- testing machine, j having a
* Made by Riehle.
f Made by both Olsen and Riehle.
TEE MATERIALS OF CONSTRUCTION.
capacity of 200 Ibs. on one inch in width of material, while Fig. 264 shows
another form, for paper only, of 100 Ibs. capacity. The stress is indicated
on the face of the dial.
Tension-testing machines for breaking cement briquettes are shown
and described in Art. 324.
273. The Emery Testing-machine. This is without doubt by far the
most perfect weighing-machine ever devised. It was originally invented
FIG. 262.
and constructed by Mr. A. H. Emery, C.E., for the use of the U. S. Test
Board in 1879, and this first machine, having a capacity of 800,000 Ibs.
in both tension and compression, is still in daily use at the U. S. Arsenal
TENSION TESTS.
329
at Watertown, Mass.
It is now manufactured in various sizes by Wm.
Sellers & Co. of Philadelphia, who have modi-
fied and improved the original plans in many
respects. It is not too much to say of this ma-
chine that it operates absolutely without fric-
tional resistance, tests with equal accuracy large
and small specimens, is easily and quickly
operated, and is practically indestructible by
any amount of legitimate use. The marvellous
character of this machine merits a careful study
FIG. 263.
FIG. 264.
by all students in engineering, and an attempt, is here made to adequately
describe it.*
The essential principle of this machine consists in a means of transmit-
ting a definite percentage of the force applied to the specimen to the scale-
beams, and there weighing it accurately, without any friction whatever in
the receiving, transmitting, or weighing parts. Hence any very small
increment of the force applied is weighed with equal accuracy, whether this
increment is added to a great or to a small previous load.f This is accom-
* The author has been greatly aided in this by drawings and descriptions made by
Mr. Carl G. Barth, M.E., who is one of the joint inventors and patentees of the various
improvements made on it by Wm. Sellers & Co.
t When the first machine was tested, a steel bar, 5 in. in diameter, was first broken
under a load of 722,800 Ibs., and then a single horse-hair was tested', and the machine
gave the strength (16 oz.) of this as accurately as a small spring-balance which was used
for a check. Rep. U. S. Test Board, vol. n. p. 1.
330
THE MATERIALS OF CONSTRUCTION.
plished by means of two connected metallic sacks, or bags, of different sizes,
the larger one, called the hydraulic support, receiving the full force trans-
mitted through the specimen, while the other is rigidly held upon the
primary weighing-beam. Thus in Fig. 265, which is merely a schematic
drawing, the load is received on the hydraulic support shown at A, and the
pressure is transmitted by means of the enclosed liquid through the pipe e
to the smaller sack at B, which being rigidly supported by the heavy cast-
iron frame, G, shown in section above and below, the bearing-plate c is
forced downward, causing the block H to press upon the primary lever C.
TENSION TESTS. 331
This is then transmitted through D to the weighing-beam E and the indi-
cator-arm F.
In place of knife-edges on the weighing-beams, thin plates of steel are
employed, these being rigidly fastened in the two attached beams. These
are so proportioned as to bring the combined bending and compression
stresses produced in them well within their elastic limits. While these offer
absolutely no friction, they do offer some elastic resistance, but this is all
allowed for in calibrating or standardizing the machines. The weighing-
beam is kept in balance by lowering upon it various weights which are placed
upon the several poise-frames which are suspended from this weighing-beam.
The particular manner of doing this, though ingenious and peculiar to this
machine, is not essential and will not be further described here.* This part
of the apparatus is entirely closed by a glass front, and it is never necessary
to open it, the weights being imposed and removed from outside, and the
load continuously indicated on a counter-cylinder shown in Fig. 265 at the
left end of the indicator-arm F.
Passing now to a study of the latest form of the machine itself and its
essential details, we have in Fig. 266 a view of the 200,000-lb. testing-
machine made for Sibley College of Cornell University. To the left is seen
the fixed weighing-head, containing the hydraulic support, and to the right
the movable straining -head, \ or hydraulic cylinder, by means of which the
load is applied to the specimen and its deformation taken up. Both of these
are supported and kept in alignment on a substantial wrought-iron girder-
bed. In the background is also clearly seen the wooden case containing the
scale, with some of the levers and poise-frames visible through its glass front,
and also part of the hydraulic pump which supplies the straining-cylinder at
either end, according as the machine is being used for a tension or a com-
pression test. The supply- and exhaust-pipes are seen coining up through
the floor at the end of the bed, and connecting with the* straining-head by
means of jointed pipes as shown. Rigidly secured to the weighing-head by
means of nuts at each end of long bearings, one on each side, are seen the
two horizontal reaction-bars, on which are cut continuous screw-threads.
These form the rigid connection between the fixed weighing-head and the
movable straining-head, through which they pass in smooth bearings placed
sufficiently far apart to allow room with several inches of clearance-space
for two large abutment-nuts^ one on each screw. These nuts may be revolved
simultaneously, by suitable mechanism, to produce motion for adjustment of
the straining-head in either direction for different lengths of specimen.
The cut shows the machine ready for a compression test, the two com-
pression-platforms, the one on the end of the draw-bar of the weighing-head,
* For a paper by J. Sellers Bancroft, on this machine, with full illustrations, see
Engr. News and Am. Machinist, both of March 22, 1894.
fin this term the word "strain" is used as meaning deformation under stress.
Thus the elongation (or compression) is all produced by the moving of this head to the
right or left, thus producing the straining, or deforming, of the specimen.
334
TEE MATERIALS OF CONSTRUCTION.
TENSION TESTS. 335
A and B, are respectively the cylinder and plunger (more properly the bear-
ing-plates), or blocks, of the hydraulic support, which is thus seen to be
annular in the actual machine, and not circular as the one shown and
described in connection with Fig. 205. The parts marked ^Vand are two
annular washers that clamp the edges of the thin brass diaphragms composing
the sack to the bearing-block L, the liquid being confined between these
diaphragms, as also already described in connection with Fig. 265. The two
plates forming this closed sack are not soldered together. They are simply
held by the great stress in the bolts which bind the washers N and so as
to maintain a tight joint under the maximum pressure. The pipe forming
the communication between the hydraulic support and the reducing-chamber
of the scale is marked R.
The draw-bar heads E and F are provided with a number of external
projections or spur-ribs, 7, extending into cavities in the main castings A
and J3, as may be seen in the case of the head E and the main casting A.
The cavities in A and B leave between them an equal number of projections,
F, which extend back into the cavities between the ribs U on the draw-bar
heads, as may be seen in the case of the head F and the main casting B.
The arrangement of these ribs and cavities, which enables a machine with a
single hydraulic support to be used both for tension and compression tests,
will be better understood from Fig. 208, in the right-hand view of which
this is made perfectly plain.
The manner in which the pull on the draw-bar is transferred by the ribs
on the head E to the bearing-block Z, from this through the liquid-sack over
to the other bearing-block J/, and finally from this to the ribs on the main
casting B, will now readily be understood, and it will also be seen that a
push on the draw-bar, due to a compression test, would in a similar manner
be transmitted by the ribs on the head F over to the annular bearing J/,
from this through the liquid over to the corresponding bearing L, and finally
from this on to the ribs on the main casting A.
Owing to supposed internal strains and stresses in the diaphragms of the
hydraulic support, it has been found necessary to put about 50 or CO Ibs. per
square inch initial pressure on the same, the levers of the scale having then
first to be balanced by means of a suitable weight that can be slid along a rod
attached to the poise-frame lever before beginning a test. This initial
pressure on the hydraulic support is obtained in the smaller machines by the
device shown in Fig. 207, while in larger machines a very complex device is
used to get the requisite pressure, the principle of this, however, being essen-
tially the same. The device as here shown consists of a large flat spring S,
attached, in a manner clearly seen in the figure, to the projecting ends of the
reaction-bars, and forming the bearing for a screw, T, secured against end
motion in this bearing by a collar attached to it on the inside of the spring,
and by the capstan-head W fastened to it by a key and a nut in the cus-
tomary manner on the outside of the spring. This screw fits into a tapped
hole in the end of the draw-bar; and it will be understood that if the capstan-
336
THE MATERIALS OF CONSTRUCTION.
TENSION TESTS. 337
head be turned in the direction of the arrow, the tendency is to move the
draw-bar, with it its heads and the whole hydraulic support, away from
the spring, until the bearing-block M brings up against the ribs on the main
casting B, further turning producing a bending of the spring in the opposite
direction. The resistance of the spring to this bending is then transmitted
through the screw T, the end of the draw-bar, the ribs on the head E, the
bearing-block Z, the liquid, and finally the bearing M over to the casting B.
This condition of affairs is exactly what is brought about before beginning a
tension test, for which the machine is supposed to be represented in this
figure. A sufficient turning of the capstan-head in the direction opposite to
that indicated by the arrow will similarly pull the draw-bar, etc., towards
the spring 8 until the cylinder L is brought up against the ribs of the main
casting A', a further turning bending the spring towards the end of the
draw-bar. This, with its consequences, is the condition of affairs brought
about before beginning a compression test.
It will be seen that the spring S is provided with a stop, to match a
similar stop on the capstan-head to bring up against, for the purpose of
limiting the motion of this latter in either direction, the pitch of the screw
T being so selected and the whole device so fitted up that the determined
amount of bending of the spring in the one or the other direction will pro-
duce the desired amount of initial pressure on the diaphragm of the hydraulic
support. The total possible movement of the hydraulic support in its cham-
ber, between the attached heads A and B, is only 0.006 in., so that the
maximum movement from a mean position is only 0.003 in.
Fig. 268 is a vertical section of a weighing-head, in which the hydraulic
support is shown in its most improved and complete form, the initial pressure
device being, however, of the same type as that already described. It will
be seen that the annular bearing-block M is here not supported on the draw-
bar D directly, but that it is in the first place fixed in its p'roper relation to
the bearing-block Z, independently of the draw-bar, by two annular steel
plates, rigidly clamped in the usual manner; and that it is also bolted to
another large annular casting F, which is supported on the draw-bar in the
same manner as the bearing-block L. In Addition to this casting V there
is also seen another annular casting X, a small arc of the circumference of
which is provided with gear-teeth meshing with the teeth of the pinion Y
on the end of the hand-lever shaft Z. It may thus to a limited extent be
rotated in the one or the other direction, but it is otherwise confined between
the main casting A and B, and also to be centred on A by a circular tongue
and groove. The plates X and V are on the figure shown to be in contact,
and the contact surfaces are helicoidal or screw-shaped, so that rotation of
JTin one direction tends to force them apart, and to bring Fup against the
ribs of the casting B, while rotation of X in the opposite direction leaves V
with a little play between X and B. The latter of these conditions is
brought about by the operator before beginning a compression test, the
former before beginning a tension test. The purpose of this arrangement is
THE MATERIALS OF CONSTRUCTION.
to protect the diaphragm of the hydraulic support from the great shock it
would otherwise receive by the sudden release of the stresses on the various
parts of the weighing-head, on the sudden breaking of large specimens when
a powerful recoil of the draw-bar occurs. With the annular anvil V set up
tight against B by the annular wedge X, the energy of this recoil is trans-
mitted directly over to the main casting A without passing through the liquid
support and the diaphragm, on which an exceedingly high pressure would
otherwise be produced, which if frequently repeated would finally destroy it.
In Fig. 269 is shown on a larger scale a section of the annular hydraulic
support on one side of the draw-bar only, which, without any further explana-
TENSION TESTS.
339
tion, will serve to give a clearer idea of the detailed arrangement of the
diaphragms and the surrounding parts, and also the manner in which the
pipe 7?, which forms the communication between the hydraulic support and
the corresponding reducing-chamber on the scale, is attached. to the former.
The two plates which form the sack lie flat against the bearing-blocks,
underlaid, however, by connected grooves, their edges being held so tight by
340
THE MATERIALS OF CONSTRUCTION.
the bolts B as to prevent all leakage. The plate on the left side of this
sack is cut away and spun into an annular pocket in the auxiliary block PT,
and it is made tight to it by running in a solder. To this auxiliary block
is now attached the connecting pipe R as shown. This pipe is first soldered
to a screw-plug which has a spherical front which bears on the conical bottom
FIG. 271. The Riehle-Marshall Extensometer.
of the opening in W> so that by screwing up hard a perfect joint is made by
the elastic deformation of the metal at the surface of contact.
In Fig. 270 is shown the form given to this machine when made with a
capacity of 100,000 Ibs.
274. Extensometers. There are three general types of extensometers
in common use, viz., Double Micrometer-screws with Electric Contact;
TENSION TESTS.
341
Friction-rollers with Dial Indicators; and Bauschinger's Mirror Apparatus.
The micrometer-screw extensometer was developed and perfected by Mr.
C. A. Marshall, M. Am. Soc. C. E.,* and it is now manufactured, with some
improvements, by Riehle Bros, as shown in Fig. 271. The essential features
of all these extensometers are :
1. Measurements are taken on opposite sides of the specimen, between
symmetrically placed points on rigid collars which are attached to the speci-
men by screw-points or knife-edges lying in two transverse planes a known
distance apart.
2. The measurements must be taken to the nearest T o-jhro- inch.
3. The apparatus must be removable with-
out releasing the load on the specimen.
In the Marshall Instrument the collars
are open, thus enabling them to be removed,
and also giving to them a sufficient spring to
take up the reduction in the diameter of the
specimen as it elongates. The elongations
are read on micrometer-screws to 0.0001
inch, the contact being determined by the
ringing of an electric bell on the closing of
a circuit by the contact. With a low but
constant current this contact- distance is
found to be constant within the limit of read-
ing given above, f Both screws are read after
each increment of loading, and the average
movement taken as the stretch of the speci-
men. This is a most excellent and delicate
instrument and is very largely used. It has
the advantage of a more positive and direct
measurement of the deformation of the speci-
men than either of the other forms.
The extensometer having friction-rollers
with, dial-indicators, shown in Fig. 272, is a
modification of one of Bauschinger's forms,
as made and used by the author. J It operates
by means of two axles, having friction-rollers
at one end and a vernier-needle at the other
which moves over a graduated dial. The
friction-roller is just 0.5 inch in circumference, and the dial is graduated to
500 divisions. The vernier on the end of the needle reads readily to 0.1 of
* A brilliant young engineer, a friend and college-mate of the author's, who lost his
life in the great Johnstown flood, 1889.
f By using a resistance relay in the circuit a strong current may be made to pass
through the bell, and a weak one through the contact points.
t Manufactured by Mahn & Co. of St. Louis, Mo.
FIG. 272, The Author's Exten-
someter.
342 THE MATERIALS OF CONSTRUCTION.
a division, thus giving readings to 0.0001 inch. The collars are attached
by three screws, and removed by opening them by a hinge movement.
One of the screws has a spring bearing in the collar to take up the shrink-
age of the specimen when under stress.
The friction-rollers are actuated by means of two arms which have a spring
bearing upon them, the opposite ends being rigidly attached to the opposite
collar in each case. Thus while the needles move in the same direction any
bending of the specimen, or angular movement of the collars with respect
to each other, is eliminated in the mean of the two readings, the same as
with the Marshall apparatus. The needles are delicately mounted so as to
have very little friction, and experience shows that the friction-contact is
entirely reliable.
The advantages of this form of extensometer are:
1. It shows by its movements the deforming action of the specimen,
which is a great advantage for students.
2. It is equally suited to measure large deformations beyond the elastic
limit as it is to measure the extremely small movements inside that limit.
3. It is equally adapted to compressive and tensile tests.
kl
^M&
FIG. 273. Diagrammatic Plan of Bauscbinger's Mirror Extensometer.
4. It is adapted to all lengths of specimens by simply changing the side
arms, several pairs of which go with the instrument.*
5. On releasing the load it shows the permanent set without any manip-
ulation.
6. It never needs to be touched by the observer during a test, which is a
great advantage in making such delicate measurements.
7. After passing the elastic limit, and under a given constant load, the
continued movement of the needle indicates the time-effect of such loads
and when such cold-flowing has practically ceased.
If the specimen should unexpectedly break with the apparatus on, no
great harm results. At most some of the clamping-screws may be bent, but
these cost little to renew.
* The author has used it successfully for observing the effects of moving loads on
bridge members, with arms five feet long, covered with thin rubber at their roller ends,
and specially made U-shaped clamps instead of collars.
TENSION TESTS.
343
Bauschingers Mirror Apparatus is shown in Figs. 273 and 274. The
specimen is clamped at two points, as shown at (#), Fig. 274, and at a and #,
Fig. 273. The stretch of the specimen is fully represented by the turning
of the friction-rollers r l and r a , these being rigidly attached to the mirrors
w, and m n through the arms a l and a . The screws set back of the mirrors
FIG. 274. Buuschingcr's Mirror Exteusometer Apparatus. FIG. 275.
are used to adjust them to a zero-reading on the scale, which is reflected into
telescopes as shown in Fig. 273.
For simply observing stretch for the purpose of detecting the elastic limit
with reasonable accuracy, the Paine extensometer, Fig. 275,* may be used
* Designed by W. II. Paine, M. Am. Soc. C. E., and used by him for finding the
elastic limit of the steel wire used on the New York- Brooklyn suspension bridge.
Made now by Richie Bros.
342 THE MATERIALS OF CONSTRUCTION.
a division, thus giving readings to 0.0001 inch. The collars are attached
by three screws, and removed by opening them by a hinge movement.
One of the screws has a spring bearing in the collar to take up the shrink-
age of the specimen when under stress.
The friction-rollers are actuated by means of two arms which have a spring
bearing upon them, the opposite ends being rigidly attached to the opposite
collar in each case. Thus while the needles move in the same direction any
bending of the specimen, or angular movement of the collars with respect
to each other, is eliminated in the mean of the two readings, the same as
with the Marshall apparatus. The needles are delicately mounted so as to
have very little friction, and experience shows that the friction-contact is
entirely reliable.
The advantages of this form of extensometer are :
1. It shows by its movements the deforming action of the specimen,
which is a great advantage for students.
2. It is equally suited to measure large deformations beyond the elastic
limit as it is to measure the extremely small movements inside that limit.
3. It is equally adapted to compressive and tensile tests.
^
FIG. 273. Diagrammatic Plan of Bauschinger's Mirror Extensometer.
4. It is adapted to all lengths of specimens by simply changing the side
arms, several pairs of which go with the instrument.*
5. On releasing the load it shows the permanent set without any manip-
ulation.
6. It never needs to be touched by the observer during a test, which is a
great advantage in making such delicate measurements.
7. After passing the elastic limit, and under a given constant load, the
continued movement of the needle indicates the time-effect of such loads
and when such cold-flowing has practically ceased.
If the specimen should unexpectedly break with the apparatus on, no
great harm results. At most some of the clamping-screws may be bent, but
these cost little to renew.
* The author has used it successfully for observing the effects of moving loads on
bridge members, with arms five feet long, covered with thin rubber at their roller ends,
and specially made U-shaped clamps instead of collars.
TENSION TESTS.
3413
Bauschinger's Mirror Apparatus is shown in Figs. 273 and 274. The
specimen is clamped at two points, as shown at (), Fig. 274, and at a and b,
Fig. 273. The stretch of the specimen is fully represented by the turning
of the friction-rollers r, and r a , these being rigidly attached to the mirrors
m v and m, through the arms a l and a a . The screws set back of the mirrors
FIG. 274. Bauschinger's Mirror Exteusometer Apparatus. FIG. 275.
are used to adjust them to a zero-reading on the scale, which is reflected into
telescopes as shown in Fig. 273.
For simply observing stretch for the purpose of detecting the elastic limit
with reasonable accuracy, the Paine extensometer, Fig. 275,* may be used
* Designed by W. II. Paine, M. Am. Soc. C. E., and used by him for finding the
elastic limit of the steel wire used on the New York-Brooklyn suspension bridge.
Made now by Richie Bros.
344
THE MATERIALS OF CONSTRUCTION.
with advantage. Its multiplication is usually made about 20 to 1, and it
may be read by vernier to 0.0001 inch, though it is usually made to read
only to 0.001 inch. This instrument has also been used to obtain the stretch
of bridge members under moving loads. As it measures stretch on only one
side of the specimen, its indications must not be accepted as absolute,
especially inside the elastic limit, while its capacity is very small beyond the-
elastic limit.
A very simple and inexpensive apparatus which may be made to give
excellent results is shown in Fig. 276. If care be taken to secure a practi-
FIG. 276. Extensometer used in Yorkshire College, Leeds. (From Engineering, Sept.
11, 1896.)
cally constant length of the short arm of the indicator, and if the legs of the
main frame which carry the graduated arc are equally flexible, fairly accu-
rate readings can be obtained.
275. Autographic Stress-diagram Appliances. These fall into two gen-
eral classes:
1. Those in which the load coordinate is recorded through a movement
of the poise on the weighing-beam.
2. Those in which the load coordinate is recorded through the lifting of
the weighing-beam against the increasing resistance of a calibrated spring
attached to its free end.
The deformation coordinate is in all cases multiplied either by levers or
by the principle of the cone-pulley. The paper is usually attached to a
cylinder, although it has sometimes been attached to a plane board. The
pencil usually moves in a straight line, indicating one of the two coordinates,
while the cylinder (or board) moves to register the other function, and it
matters not which of the two movements is made by the deformation of the
TENSION TESTS.
345
specimen and which by the increasing load. The location of the paper and
its mounting is a matter of convenience simply. The cords (or wires) which
are to transmit the stretch of the specimen must form a pair, symmetrically
placed on opposite sides of the specimen; they must be attached to one collar
and pass through pulleys similarly placed on the other. They should then
pass off in a plane at right angles to the specimen * and connect with the
ends of an " evener " (lever), to the centre of which is attached the single
cord which passes either to the pencil-holder or to the cylinder which carries
the paper. If cords are used, they should be such as do not stretch appreci-
ably for such changes of stress as occur in them during the test.
One method of mounting these parts is shown in Fig. 256 (Olsen's), and
another form in Fig. 258 (Richie's). In both cases the pencil is moved by
the poise by a reducing-gear, and the cylinder is moved by the deformation
of the specimen by a multiply ing-gear.
Figs. 277 and 278 show two improvised forms of autographic diagram
Automatic Strain Diagram
Apparatus
for
Tension Tests,
FIG. 277.
apparatus which the author devised and had attached to his 100,000-lb.
hydraulic testing-machine, for tension and punching tests respectively,
* This is necessary in order that the stretch of the specimen may be fully repre-
sented in the shortening up of the cord. The eoi^s should therefore be attached to the
moving end of the specimen.
346
THE MATERIALS OF CONSTRUCTION.
Here the .pencil is moved by means of a fine wire which coils over the small
spindle to which the poise driving-pulley is attached, while the cylinder is
Automatic Strain Diagram
Apparatus
for
Shearing Tests .
FIG. 278.
turned by the cord from the specimen runs upon a small dram at top.*
This cord is kept taut by means of a small plumb-bob. The clamp-screws
in the collars are set so as to take hold of a square or rectangular specimen
as well as a round one. The second form was made for a series of punching
tests of steel plates.
A great defect of all the autographic appliances given above is that they
do not readily give the last end of the curve after passing the maximum
load, although they would do this if the poise should be run back so as to
keep the beam in balance at all times. This is a difficult feat with both the
hand- and the electrically-controlled movement of the poise, and the result is
that this part of the diagram is usually worthless. ^
In, the Gray Extensometer Apparatus^ however, Fig. 279, this part of
the curve is obtained perfectly, for the weighing-beam is at all times in
perfect balance, since it pulls upon a calibrated spring. Here the pencils
are moved at two rates of speed by the deformation of the specimen, and the
cylinder is turned by the lifting or dropping of the weighing-beam. The
* This cord should lead off from the specimen at the upper (fixed) cross-head instead
of from the lower one. The drawing is erroneous in this particular,
f Designed by Prof. Thos. Gray, and manufactured by Riehle Bros.
TENSION TESTS.
347
348
THE MATERIALS OF CONSTRUCTION.
ong trussed lever at top gives a movement to one of the pencils of from one
to two times the actual deformation between the collars, while the other
trussed lever may give to the other pencil a movement from one hundred to
five hundred times the actual relative movement of the collars, depending
in each case on the link connections which are made preparatory to start-
ing the test. Furthermore, both pencils operate inside the elastic limit and
some distance beyond, when the one moving more rapidly is automatically
thrown out of gear, while the other pencil proceeds to record the complete
diagram on the smaller scale. The result is a double stress diagram, such
as shown in Fig. 5. The diagrams shown in Fig. 280 have been photo-
graphically reduced directly from autographic diagrams made by this appli-
Fro. 280. Tension-stress Diagrams of Low Carbon-steel automatically recorded by the
Gray Apparatus. (Made by Prof. Gray for the author.)
ance for the author b}^ Prof. Gray himself. They do not extend beyond the
yield-point stage of the test.
Mr. Olserfs new Micrometer Autographic Attachment forms a supplement
to his general stress-diagram apparatus, for the purpose of making a diagram
inside of and somewhat beyond the elastic limit, in which the stretch of the
specimen is magnified five hundred times. He accomplishes this by revolving
the drum one hundred times as fast inside the elastic limit as is done beyond
that limit, the stretch of the specimen here being greatly multiplied by a
micrometer-screw and its accompanying gearing shown in Fig. 281. When-
TENSION TESTS.
349
ever the collars separate, the lower pair of fingers (being weighted) drop
with the lower collar and so break an electric spring-contact shown on the
left of Fig. 281, which sets in motion both the drum carrying the paper (not
shown here) and the micrometer-screw and its gearing, which at once closes
the circuit by raising the outer end of the lever to which the spring-contact
is attached. A very large circumferential motion can thus be obtained for
a very small movement of collars (500 to 1), and this can be conveyed by a
positive connection to the dram of the autographic apparatus. When the
elastic limit has been passed this part of the apparatus is thrown out of gear,
the drum set back to its proper position under the pencil for the small-scale
350
THE MATERIALS OF CONSTRUCTION.
diagram (deformation 5 to 1), and the test proceeds to its completion at the
final rupture of the specimen. The poise is moved either forward or back-
ward at pleasure by making the proper electric connections, or both these
connections can be made at once, in which oase the poise moves forward when
the beam is up, and backwards when it is down. The last part of the diagram
can thus be obtained. The entire machine, with both the large- and the
small-scale diagram apparatus, is shown in Fig. 282.
276. Micrometer-callipers. In Figs. 283, 284, and 285 are shown three
forms of micrometer-callipers, one or more of which are necessary for accu-
TENSION TESTS.
351
352
TEE MATERIALS OF CONSTRUCTION.
rately measuring the dimensions of test specimens, The form shown in Fig.
283* spans 8 inches; that shown in Fig. 284 spans 2 inches, and that in
Fig. 285 spans 2 inches, all of them reading to 0.0001 inch. The advan-
tage of the last form is that one may be set for the width and the other for
the thickness of plates, thus saving much running back and forth of the
micrometer-screw.
277. Gauging-implements. It is common to lay off the test specimen
into 1-inch divisions either by scratch-awl or centre-punch marks. For the
FIG. 286.
former (used on plate specimens) the laying-off gauge shown in Fig. 286 is
used, and for the latter (used on round specimens) the double-pointed
FIG. 287.
centre-punch shown in Fig. 287 is most convenient. Some such instru-
ments are essential to accurate work, and they also are great time-savers.
* Designed by Prof. Sweet and made by the Syracuse Twist Drill Co., Syracuse,
N. Y.
CHAPTER XVI.
COMPRESSION TESTS.
278. Objects of Compression Tests. While tension tests are made for the
purpose of determining many of the more significant mechanical properties
of the malleable metals, compression tests are made to determine resistance
to compression alone. In Chapter III it was shown that the materials of
construction divide themselves into two general classes with respect to their
manner of failure in compressive tests, these two classes being called plastic
or viscous materials, such as the malleable metals, and brittle or comminuible
materials, such as cast iron, stone, brick, etc.
When testing plastic materials in compression the " apparent elastic
limit" must be regarded as the ultimate strength; and since this limit in
compression is in nearly all cases the same as it is in tension, it is commonly
taken as the same, and no compression tests are made on such materials except
when made up into fall-sized columns. The compression tests of these are
known as " tests of columns," rather than " compression tests " of that
material.
Brittle materials are tested in compression to determine their resistance
to crushing.
279. Compression-test Specimens. In the case of metals the test speci-
mens can be turned or shaped accurately, but in the case of stone, cement,
concrete, brick, etc., it is not practicable to obtain perfectly true specimens,
and hence some suitable provision must be made for these when placing them
in the testing-machine. For such materials the form of specimen hitherto
almost universally employed has been that of the cube. In Chapter III it
was shown that this form is too short to give a normal failure; that the
length in the direction of the applied load should be at least 1 times the
least .lateral dimension. It is probable, however, that compression tests will
continue to be made on cubical forms, for the reason, that the results may
thus be comparable with those hitherto obtained and published. The
general relation between the strength of cubes and of prisms of various ratios
of height to least breadth, for sandstone, is shown in Fig. 17, Chapter III.
While perfectly true and parallel surfaces cannot usually be obtained,
they should be made as nearly so as possible. This can be done at a small
cost if stone-grinding works are at hand, or if such a special grinding-machine
is available as that shown in Fig. 288.
353
354
THE MATERIALS OF CONSTRUCTION.
The test specimen should be very nearly prismatic, since when the sides
protrude much beyond the bearing-surfaces the specimen is strengthened as
shown in Fig. 18.
/0000
1
7000
FIG. 288. An Abrasion Testing-
machine.
1
3000
FIG. 289. Showing the Effect of Bed-
ding on the Strength of Sandstone.
(Inst. Civ. Engrs., vol. cvn.)
280. Bedding the Specimen in the Testing-machine. If the specimen
has not true and parallel beds, it is necessary to embed the specimen in
plaster of paris. This is done by inserting sized paper between the plaster of
paris and the specimen to prevent the absorption of water by the specimen,
which invariably weakens it if it has a high absorbing capacity. A small
load is brought upon the specimen while the plaster beds are soft, and this
is left on for some ten minutes or longer, till the plaster has hardened, when
the test proceeds to failure. Great care must always be taken to put the test
specimen accurately in the axis of the testing -mad Line. Compression tests
probably more often give erroneous results from not having done this than
from any other cause.
If the specimen has true and parallel beds, then it may be placed directly
between steel plates, or between the machine cross-heads, if these are true
and smooth. Or single thicknesses of tar-board may be employed. In any
case no bedding material must be used which will flow, like lead, or spread,
like wood, when the load comes on. This causes the specimen to split up
and to fail in detail. (See Fig. 289.)
An infallible test of x>roper bedding and placing in the testing-machine
is the manner of failure of the specimen. If it spalls off on the sides
(especially if it spalls mostly on one side) before final crushing down, some-
COMPRESSION TES
355
thing is wrong. It should spall very little, and should crush down suddenly,
with a great explosive sound, and fly over the room.
An adjustable bearing-plate at one or at both ends of the specimen is
desirable, bat not strictly necessary if care be taken to secure in other ways
a true initial bearing.
281. Compression-test Machines. The universal machines shown in Figs.
25(5, 257, 258, 259, 2GO, 266, and 270 are all adapted to the making of
compression tests as well as tests in tension. In Fig. 377 is shown a machine
for compression (cement) tests only, and the author has had constructed a
machine for testing timber columns, with a capacity of 1,000,000 Ibs., which
works in compression only, bat in general all the compression machines used
in America are of the universal type.
282. Compressometers. Since compression-test specimens are generally
very short, the ordinary appliances used in tension tests for measuring
FIG. 290.
deformations cannot be employed. In Fig. 290 a very convenient com-
pressometer is shown, which is adjustable to varying heights of specimen by
moving the geared pair of screws, and to specimens of excessive height by
introducing new sets of screw-stems. The bearing-points are in pairs,
mounted on rockers, so that any unsymmetrical movement is provided for and
eliminated. It reads to y^/nro inch by electric contact at the right end,
under the set-screw there. The deformation of the specimen breaks this
contact, and by turning the micrometer-screw the contact is made, this being
indicated by the ringing of a magneto-bell.
For large specimens a form like that of Prof. C. Bach, Fig. 291, may
be used.* This measuring device consists of two rings, AA (on top) and BB
(below), each of which is fastened to the specimen by means of four screws,
* The description here given is taken from Zeits. d. Ver. Deutscher Ingenieure, April
27, 1895.
356
THE MATERIALS OF CONSTRUCTION.
PIG. 291. Bach's Apparatus for Compression Tests of Concrete Columns 10 in. diam-
eter and 40 in. long.
COMPRESSION TESTS. 357
being at right angles to each other at any convenient distance apart. (The
apparatus shown in the figure was used on concrete cylinders 10 inches in
diameter and 40 inches long, the rings being 30 inches apart.)
The measuring apparatus is shown on the right. If a compression of the
test specimen occurs, the upper terminal point of the rod, C\ which length
remains the same, will move upwards a distance equal to the distortion of
the specimen, thus causing the lever DEF to turn around its axis at E, and
to carry along with it, by the small metallic thin ribbon fastened on its seg-
mental end F, the axis 6r, on which the indicator is fastened, which latter
runs along the arc graduation. The indicator is not pointed at the end, but
flat, and upon it is an index-line, as may be seen in the drawing. The ratio
is made so that 1 mm. compression of the test specimen equals 300 mm.
distance on the arc scale. Since this can be read to -^ mm., the distortion
of the measured distance can be read to ^F mm - The only new feature
of this instrument is the use of the thin metallic ribbon in place of gearing.
The disadvantage of employing a rack and pinion was that the loads had
to be varied by loading and unloading; the least lost motion would produce
serious errors. Furthermore, the transmission proportion is dependent on
the form of the tooth, for, being obliged to make the teeth so very small, we
cannot depend on forming them sharp enough to maintain a constant trans-
mission proportion.*
There should always be two such measuring instruments, set opposite
each other, as shown in the general view on the left. By this method the
measurement of the deformation takes place at two diametrically opposite
points, the mean of the two readings being used.f
For measuring given percentages of compression deformation of wood
blocks of varying thicknesses, for instance, the author devised the apparatus
shown in Fig. 292. Here a metal point, attached to a sleeve, moves on an
adjustable inclined arm, so bent that the point moves on a fine through the
hinge in the plane of the flat base of the apparatus. The making of the
contact at the point rings an electric bell, and the free movement of this
point is interrupted by spring stops at such percentages of distortion as are
to be observed (with the U. S. timber tests, in compression across the grain,
these observed deformations were 3 per cent and 15 per cent of the thick-
ness of the block). For a specimen of any height it is only necessary to
move the point to its outer limit, raise it into contact with the upper com-
pression-head of the machine, and tighten the thumb-nut. Then slide the
point back to thb first stop and proceed to load the specimen. When the
bell rings note the load for that limit, and slip the point to the next stop,
etc.
* Where simple friction of a bar on a rolling pinion is relied on to move the indicator-
needle, great care must be exercised, especially when loads are applied or released sud-
denly.
f These instruments are made by C. Klebe of Munich, Germany.
358
THE MATERIALS OF CONSTRUCTION.
Weighing Table of Testing Machine .
FIG. 292. Compressometer designed by the Author to Indicate two Conventional
Limits of Deformation (3$ and 15$) of Wood Blocks when Tested Across the Grain.
FIG. 293. Tetmajer's Compressometer for very short Specimens Tested in a Horizontal
Machine. (Zurich Laboratory Communications, vol. iv.)
COMPRESSION TESTS. 359
283. Tetmajer's Apparatus for Short Specimens. In Fig. 293 is shown
Prof. Tetmajer's apparatus for measuring the deformation of very short
specimens.* It is used in a horizontal (Werder) machine, and stands
upright as in the figure. The compression of the specimen is taken up on a
micrometer-screw which operates on the short arm of the indicator, the long
arm of which actuates the upper, balanced, horizontal lever, thus bringing it
to its zero position as shown by graduation lines on its left-hand end, and on
the adjoining fixed portion of the frame. Because of the measurements
being taken on one side only, and that a long distance from the specimen,
the readings, while giving true relative motion, would probably not be true
absolutely. Tetmajer used it mainly to determine elastic limits. His read-
ings were taken to 0.0001 inch, and great care was taken in centring the
specimen in the testing-machine.
284. Compression Tests of Columns, f Since the strength of a long
column consists in its resistance to bending, rather than in its resistance to
crushing, it follows that the strength of a straight column is a function of
1. The elastic rigidity (modulus of elasticity) of the material, E.
2. The ratio of its length, ?, to the rigidity function of its cross-section,
which is the radius of gyration, r, that is, .
3. The character of its end bearings as to their tendency to hold the
column to its original position, and
4. The eccentricity of the loading.
For a straight column, symmetrically loaded, supported at its gravity
axis, and so as to be perfectly free to bend, and for a ratio of sufficiently
large, it is shown in works on mechanics (and in the author's " Modern
Framed Structures ") that Eider's Formula gives the strength of the
column. This formula is
where p = ultimate strength of the column, in pounds per square inch;
E modulus of elasticity, in pounds per square inch;
I = length of the column between the pivot-bearings, in inches;
* Described in Tetmajer's Communications, vol iv. (1890).
f It does not fall within the province of this work to enter into a general discussion
of the strength of columns. The following is given as supplementary to what is usually
found in the works on applied mechanics, and on framed structures, on this subject.
\ "While this is the only purely theoretical column formula which is true in practice,
II I \
it is only applicable to very long columns ( - > 150 for pin ends, and t > 200 for flat ends j ,
such as are seldom or never used in actual structures, and hence it is of little practical
value. This form-da must never be used for the ordinary lengths.
360
THE MATERIALS OF CONSTRUCTION.
r least radius of gyration of the cross-section of the column, in
inches.
a/ #.2 as #4 tar 0s #7 #0- #3 /0
FIG. 294. Variation of Moduli with increasing Percentage of Carbon in Steel. (Wat.
Ars. Rep., 1886.)
In Fig. 296 the locus of this equation is shown for E = 30,000,000, the
coordinates being p and .
Theoretically, for a perfect column, centrally loaded, the strength is con-
stant for increasing lengths, this strength being the " apparent elastic limit "
or "yield-point" of the material (see Fig. 294) until the critical length is
reached under ivhich the column bends indefinitely under its maximum load,
when for any further increase in length the load which will produce this
bending regularly diminishes in accordance with the law of Eider's curve.
The theoretical locus, therefore, for the value of p plotted to would be a
horizontal line at the apparent elastic limit of the material, extended to an
intersection with Euler's curve, and then down along this curve indefinitely.
But because no column is perfectly straight, nor, perfectly free to turn, nor
COMPRESSION TESTS.
361
loaded and supported exactly in its gravity axis, nor has the same modulus
of elasticity in all its parts, nor is of exactly uniform cross-section, etc., it
follows that any locus derived from experiment would usually fall below this
theoretical locus, and could never rise above it except from a higher modulus
of elasticity, or from a higher elastic limit, or from more fixed end condi-
tions than had been assumed.
In making tests of metal columns there are but two conditions of end
supports to which any theory can be adapted, these being rigidly fixed in
direction and absolutely free to revolve. As it is impossible to satisfy the
former condition, the latter becomes the only one to which any theoretical
formula should be expected to conform. It seems remarkable, therefore,
that, so far as the author is aware, there lias been but one set of observations
made which has fairly satisfied this requirement, these having been made by
M. Considere, Ingenieur-en-chef des Fonts et Chaussees,* France. Both
Prof. Bauschinger and Prof. Tetmajer attempted to satisfy this condition,
but they mounted their columns with cone or knife-edge bearings at the
computed gravity axes, while M. Considere mounted his with lateral-screw
O!
B
FIG. 295. Considere 's Mounting for Column Tests. (Fr. Com. Rep., vol. in. p. 124.)
adjustments, as shown in Fig. 295, and arranged a very delicate electric
contact at the side so as to indicate a lateral deflection as small as 0.001 mm.
* First reported in 1889, and described in vol. i. p. 1*28 and vol. in. p. 124 of the
Report of tJie French Commission on the Methods of Testing Engineering Materials, 1895.
362
THE MATERIALS OF CONSTRUCTION.
He then applied moderate loads to the columns and adjusted the end-bear-
ings until they stood under such loads rigidly vertical, with no lateral move-
ment whatever.* Then with his double knife-edge bearings at each end, as
shown in Fig. 295, the columns were perfectly centred and absolutely free
to move or turn about their end bearings, as the theory demands. With
these conditions perfected he made 155 tests of columns, of various lengths
from = 40 to = 3-40, and on various forms of cross-section.
80M0
FIG. 296 One Series of Results of M. Considered Column Tests, with Material having
different " Apparent Elastic Limits." (Rep. Fr. Com., vol. in. p. 124.)
In Fig. 296 the author of this work has plotted the tests made on solid
rectangular steel bars, 10 mm. by 17 mm. in cross-section, of six degrees of
* This precaution is essential to a perfect test of the material of which the column is
composed. Only in this way can other sources of weakness be eliminated. It is to the
interest of the contractor, therefore, to provide these appliances.
COMPRESSION TESTS. 363
hardness. He has also fitted to these six sets of observations parabolic loci
which cut the axis of loads at the respective " apparent elastic limits," or
" yield-points," of the material,* and which are made to become tangent to
Euler's theoretical curve drawn for E 30,000,000. The close agreement
of these loci with their respective sets of observed ultimate strengths would
seem to indicate that they cannot well be improved upon, and that therefore
this parabolic law may fairly be assumed to fit the actual facts in an ideal set
of experiments as closely as it is possible to do.f
It further appears from these curves that the coefficient of the subtrac-
/ IV
tive term (-) follows a very definite law, as shown in " Modern Framed
Structures," p. 150. Using this theoretical value of this coefficient, we
have, as the maximum strength of any pivoted wrought-iroii or steel column,
in pounds per square inch for
= 30,000,000, ;-
To show that the strength of a column is no function of the ultimate
strength of the material either in tension or compression, M. Considere cold-
rolled the medium hard steel in No. 5, which had an apparent elastic limit
of 47,000 Ibs. per square inch, until it had elongated ten per cent of its
original length. This raised its elastic limit to 71,000 Ibs. per square inch,
while its ultimate tensile strength was raised only from 83,000 to 88,500 Ibs.
per square inch. Thus metal No. 0, with an elastic limit of 71,000 Ibs. and
an ultimate strength of 88,500 Ibs. per square inch, was over 10 per cent
stronger in columns than metal No. 8, which had an elastic limit of 64,000
Ibs. and an ultimate strength of 98,000 Ibs. per square inch.
While the parabolic curves hers given are purely empirical in form,
theory dictates :
1. That this locus shall start horizontally from the vertical axis at the
" apparent elastic limit ";
2. That it shall become tangent to Euler's curve; and
3. That it shall have no points of inflection other than the point of tan-
gency with Euler's curve.
While there may be an infinite number of curves which would satisfy
these requirements, the parabola is the simplest of all, and it also seems to fit
the observations as well as any, whether these observations be made under
ideal conditions, as in Fig. 296, or under the nearly ideal conditions of Prof.
von Tetmajer, Figs. 297 and 298, or under the conditions of practice, as in
Fig. 208 of " Modern Framed Structures."
* As computed from the ultimate strengths which aloue were given in the original
communication, the yield-points not having been observed.
f The author had already developed this curve as best representing -the strength of
ordinary columns too short for Euler's formula to apply to, in " Modern Framed Struc-
tures," p. 148 (1892).
364
THE MATERIALS OF CONSTRUCTION.
COMPRESSION TESTS.
365
1
" / T%r?"'i
W6
S^
*-5x
S?M
^1-
~*TNJ ^ S3
S'
M
V
*N. i
^
fe
^
b
^i
i
I
3
^3
366 THE MATERIALS OF CONSTRUCTION.
Prof, von Tetmajer's tests cover a great variety of forms, simple and com-
posite, on both wrought-iron and steel, and the results of these tests are all
plotted, with characteristic symbols, in Figs. 297 and 298. While these
results scatter somewhat, owing to varying elastic limits of the specimens and
to the fact that the knife-edge bearings were placed at the computed gravity
axes, and were not adjusted to the true centres by lateral adjustment under
small loads, as were those of M. Considere, still the parabolic curve fits the
average position of the plotted points as well as could be desired. See also
the similar diagram for wooden columns in Chapter XXXII.
The following are the author's parabolic column formulae as given in
" Modern Framed Structures " :
ULTIMATE STRENGTH OF COLUMNS, IN POUNDS PER SQUARE INCH.
For Wrought-iron Columns, Pin Ends, ( - < 170,1
p = 34,000 - -67^-) (2)
For Wrought-iron Columns, Flat Ends, ( < 210,1
p = 34,000 - .43 (j^) (3)
For Mild-steel Columns, Pin Ends, (- < 150,1
= 42,000 - .97 ...... . . . (4)
For Mild-steel Columns, Flat Ends, (- < 190,J
p
= 42,000 - .62 (IV. ... (5)
-
For Cast-iron Columns, Round Ends, (- < 70 j
p = 60,000 (-) . (6)
4- \/* I
11 ~
For Cast-iron Columns, Flat Ends, -^120,
\r i
(?)
COMPRESSION TESTS. 367
For White-pine Columns, Flat Ends, (- = 60, )
(8)
For Short-leaf Yelloio-pine Columns, Flat Ends, \-j <60,1
p = 3300 -.7 (t\ (9)
For Long-leaf Yellow-pine Columns, Flat Ends, (-= <60,J
p = 4000 - .8 g)' (10)
\U/ I
For White-oak Columns, Flat Ends, (y<60j
p = 3500 - .8 (1)' (11)
To obtain from the above working formulae for designing, divide both
terms of the right-hand members of these equations by the factor of safety
chosen for the work in hand and for the material used. The smallest factors
would be used with rolled mild steel, and the largest with timber and cast
iron.
285. Spring Testing-machines. Fig. 299 shows a form of spring testing-
machine adapted for both compression and tension tests, the former being
made at A and the latter at B. It is made in two sizes, of 2500 and 4000
Ibs. capacity respectively.
In Fig. 300 is shown a spring testing-machine of 65,000 Ibs. capacity,
for compression only, and not requiring the use of over- weights, although
such are furnished for one half the total load, if desired.
368
THE MATERIALS OF CONSTRUCTION.
FIG. 299. Spring Testing-machine for both Tension and Compression Tests,
Fiu. 800. Spring Testing-machine.
CHAPTER XVII.
CROSS-BENDING TESTS.
286. Objects of Cross-bending Tests. Brittle materials like cast-iron,
stone, brick, and concrete are tested in cross-bending to determine their
ultimate strength, and perhaps also their resilience. Timber is so tested also
to determine its ultimate strength and its modulus of elasticity. Springs and
spring-steel are tested in this way to obtain their elastic limits and their
deflections under given loads, and railroad rails are sometimes tested for
elastic limit and ultimate strength. Cross-bending tests are also made for
scientific purposes to test the correctness of the ordinary formula for the
strength and the deflection of beams.
Since three kinds of stress, tension, compression, and shearing, are devel-
oped when a beam is bent under the action of external forces, the problem
is more complex than those considered in the two previous chapters Usually
the shearing stresses are left out of account in designing both for strength
and stiffness, but the conditions under which this stress should be recognized
and taken account of are given in Article 38, for strength, and Article 46,
for deflection.
287. Essential Considerations in Cross-bending Tests. The essential con-
ditions which must be satisfied in making cross-bending tests are:
1. The loads should be applied centrally in the direction of the greatest
or of the least moment of inertia of the beam, in order to prevent torsion.
2. The supports must be rounded knife-edges, bearing on auxiliary
plates if necessary to prevent indentation.
3. The loads should be continuously progressive, without shock, and in
the case of timber they must increase at a fixed rate with no stopping when
readings are taken.
4. The deflections must be measured by observing the movement of the
neutral plane at the loaded point with reference to the neutral pLine at
the two end supports. That is to say, the deflection apparatus must be
attached directly to the specimen, or rest on the end bearings, and be self-
contained with the test specimen, and independent of all deforming move-
ments of the machine itself.
The fourth condition is seldom properly satisfied. It is common to
measure deflections with reference to some part of the frame of the testing-
369
370
THE MATERIALS OF CONSTRUCTION.
machine, assuming this to be rigid, or by means of a deflection apparatus
attached to this framework and moving with it.
FIG. 301. The Author's Beam-testing Machine.
Cross-lending Testing -machines. In the author's machine, shown in
Fig. 301, the deflection is measured by means of a micrometer-screw,
reading to 0.001, inch held in place by a bar which is attached directly to
the knife-edge bearings through parts which are not under stress. The
micrometer-screw bears upon the top of the power screw which presses
on the centre knife-edge. As all these bearings have steel plates intervening
between them and the specimen (if this be timber), the movement of the
centre bearing, with reference to the end bearings, is registered on the
micrometer-screw, and this is the deflection of the specimen. One half the
load is weighed on any ordinary form of platform-scales.
In the large beam-testing machine of the author's, shown in Fig. 302,
used mostly for testing large wooden beams, the deflection is measured by
means of a fine thread attached (at one end by a rubber band) to two nails
driven into the stick in the neutral plane over the end supports. At the
centre a nickel-plated scale, graduated to 0.1 in. and polished to act as a
mirror, is fastened to one or both sides of the beam. The thread is then
read on this scale by bringing it and its image into coincidence and esti-
mating its position to the nearest 0.01 in. The load is applied by pumping
oil into the cylinder below, thus depressing the screws and the cross-head
carried by them, and one hal f the load is weighed on the 50,000-lb. plat-
form-scales under one end. The base of this machine consists of two long-
leaf yellow-pine sticks, 6x18 inches in section and 24 feet long, with a
f X 18-inch steel plate inserted between them. Its capacity is 100,000 Ibs.
In the machine shown in Fig. 303, specially designed for cast-iron tests,
the deflection is correctly indicated on the graduated arc by means of an
CROSS-BENDING TESTS.
371
372 THE MATERIALS OF CONSTRUCTION.
ingenious arrangement of levers underneath, not shown in the figure. Its
capacity is 4000 Ibs.
In Fig. 304 is shown Keep's autographic recording transverse test appa-
FIG 303. Cross-bending Testing-machine for Cast Iron. Deflection
correctly measured. *
i-atiis for his standard form of specimen inch square by 12 inches long.
Ic makes an autographic record like those shown in Chapter XXIV, the de-
flection of the test specimen being taken up by the gradual falling of the
weighing-beam. The movement of the poise also moves the paper, while
the deflection of the specimen moves the pencil. This is a valuable machine
for tests on this size of specimen.
The universal machines shown in Figs. 256 to 260, and in Figs. 266 and
270, are all adapted to making transverse tests by inserting an I beam or
other rigid base on the weighing-table when the specimen is longer than
this, and supporting the specimen on it.
288. Importance of Measuring the Deflection in Transverse Tests of Cast
Iron. The importance of measuring the deflection of transverse-test speci-
mens of cast iron, as well as the breaking strength, is now generally recog-
nized. The resistance of the metal to shock is measured by the product of
CROSS-BENDING TESTS.
373
the ultimate load into the final deflection, divided by 2, this being
approximately the total area of the stress-
diagram in cross-bending. If this be now di-
vided by the volume of the metal between the
end bearings, it gives resistance to shock in
inch-pounds per cubic inch of metal.* Since
it is more convenient, however, to weigh the
bar than to compute its volume, the resistance
to shock is commonly computed per pound of
metal (between end bearings).
From many experiments made by the author,
he has recommended the following require-
ments t for test-bars about 1 inch square:
Inch-pounds per - c
pound of Cast Iron, g
For the lower grades of castings .... 20-30 \*
For good machine castings ......... 40-50 fc
For stove-castings, and for impact 4
machinery ...................... 60-70
In this way both the strength and the de- Ef.
flection are properly allowed for; and since jj
these results usually vary inversely with each ~.
other, both may vary greatly without showing W
where f r modulus of rupture in pounds per square inch;
W= breaking-load at centre in pounds;
I = length between end bearings in inches ;
l> =. breadth in inches;
Ji = height in inches;
y } = distance from neutral axis to outside fibre which failed under
the breaking-load, in inches;
/ = moment of inertia of the cross-section about its neutral axis;
M bending moment on the beam at the section of rupture.
Because these formulas are strictly true inside the elastic limit, it must
not be inferred that this so-called " modulus of rupture in cross-bearing "
represents any actual stress on any outside fibre, either in tension or com-
pression, ut the time of rupture. (See a discussion of this question in Arts.
35 and 36, p. 50.) In general this modulus is about twice the tensile
strength, in the case of cast iron, while in timber it is somewhat below an
average of the tensile and the compressive strength of the wood. With this
understanding this modulus is a convenient, though conventional, method
of stating the strength of any material under cross-breaking stress, and for
comparative purposes it is very useful. For the purposes of the designer,
however, these formulas are strictly correct, since he always works with loads
and stresses inside the elastic limits of the materials he uses.
290. The Modulus of Elasticity (Stiffness) is preferably found from a
transverse test, since it is mostly used for computing the deflection of beams.
Since this quantity is of necessity computed from loads and their correspond-
ing deflections inside the elastic limit, it follows that this modulus is found
to be practically the same whether it is computed from tensile, compressive,
or transverse tests. Thus Prof. Tetmajer tested fourteen rolled wrought-iron
I beams, from 4 to 16 inches in depth, and obtained from them an average
value of E = 27,840,000, while on thirty-one tension tests on specimens cut
from shape iron from the same metal he obtained a value of E = 28,110,000;
ior wrought-iron riveted plate girders, from 16 to 28 inches deep, he ob-
tained a value of E = 26,000,000.
In mild steel he found
For the tension specimens E = 30,550,000;
" " riveted plate girders E = 28,160,000.
CHAPTER XVIII.
IMPACT AND HARDNESS TESTS.
MPACT TESTS.
291. Object of Impact Tests. As explained in Art. 53, impact tests
cannot give absolute results, like those obtained from tension, compression,
and transverse tests, and hence they are properly used only where other
methods of testing are not available. They are commonly employed on cast-
iron car-wheels, on cast-steel and malleable-iron car-couplers, and on car-
axles and sometimes on rails and rail-joints. Axles and rails, however, can
be tested statically in cross-bending, and more can be learned by testing them
in this manner, by bending them back and forth, and plotting their bending-
stress diagrams, than by the drop tests.*
Because of the extreme difficulty of arranging an impact test so as to give
to the specimen a plain tensile stress, without allowing a large and uncertain
part of the energy of the blow to be absorbed in the auxiliary appliances,
this has seldom been attempted, and certainly it never has succeeded in giv-
ing any valuable results.
Impact tests in compression are seldom employed except to produce pene-
tration of a standard form, to determine hardness, which will be described
later under the head of hardness tests. The impact test given to car-couplers
might be called a compression test, perhaps, since the blow is given " end-on" ;
but as failure here occurs by breaking off portions of the enlarged head, by
developing in it excessive transverse stresses, it is really a transverse test.
In general, therefore, impact tests are all transverse, or cross-bending,
tests.
The author has also introduced a species of impact test for street-paving
brick in place of the abrasion test hitherto employed. This was done from
the fact that paving-brick do not wear out by abrasion, but by being broken
down by the blows from horse's shoes, and from the wheels of vehicles.
* Au exception may have to be made in the matter of brittleness of metals induced
by very low temperatures, which cau, it is said, only be determined by impact or drop
tests.
375
376 THE MATERIALS OF CONSTRUCTION.
Impact tests, therefore, are usually made to determine the resistance te
shock of structural forms which cannot readily be tested in any other way.
292. Essential Conditions of Impact Tests. Since the force of a blow
depends as much on the resistance offered by the body struck as it does on
the striking body, it follows that the anvil, or bed, of an impact machine is
quite as important as the weight of the ram and the height of its fall. A
standard impact test, therefore, involves a standard size of anvil and a
standard froundation for it, quite as much as a standard weight of hammer
and standard fall of same.
The pendulum machine would seem to offer one advantage, however,
which cannot be realized in drop machines. The pendulum machine can be
so designed as to allow the pendulum weight to pass the specimen when it
breaks, and by automatically recording its extreme movement, and deducting
this vertical component from the original total fall, the actual energy absorbed
by the specimen, up to rupture, would be determined proivded the anvil is
rigid.* In this way the specimen could be broken on the first blow, and the
energy spent upon it (and absorbed by it) exactly determined. This would
seem to be the only proper way to make comparable impact tests.
The Pennsylvania Railroad Company has standardized the impact test
of cast-iron car-wheels and of car-axles, and the American National Car-
builders' Association has now (1896) standardized the test f or car-axles as
indicated in Art. 29-t; but with these exceptions it can hardly be said that
impact tests made in different places in this country can be considered as at
all comparable, because of a want of identity in the foundation portion. If
the impact machine be of the pendulum form, it must strike the specimen
at the. centre of percussion of the entire pendulum in order to prevent a
portion of the energy from spending itself by bending the pendulum.
While pendulum machines are more convenient, drop machines are more
certain to deliver the full theoretical force of the blow. In a pendulum
machine the energy of the blow is, of course, the total weight of the pendulum
into the distance through which its centre of gravity falls.
293. The Energy of the Blow. The unit of measure in impact tests is
the foot-pound (or kilogram-meter). This energy cannot be measured in
pounds, and no scheme of equivalents can be devised between the foot-pound
units of an impact test and the pound units of a static test, although this has
often been attempted. There is 110 relation between the resistance to shock
and the resistance to a static load, since there is no relation between the total
area of a stress-diagram and its stress coordinate. The attempt which is
often made, therefore, to equate these two kinds of resistance is as foolish as
the ancient practice of estimating the discharge of a stream, or aqueduct, or
pipe from its cross-section alone.
From the law of the conservation of energy we have:
* Some experiments along this line have recently been made by Mr. S. B. Russell,
M. Am. Soc. C. E., in the St. Louis Water- works Department.
IMPACT AND HARDNESS TESTS.
377
The work which gravity does on the falling weight, and which is wholly
represented by the energy of the hammer at the time it strikes,
must be absorbed by the resisting body. This energy is equal, in
foot-pounds, to the weight of the ram in pounds multiplied by
its total vertical fall (including the vertical deflection of the
specimen) in feet.
In the case of a pendulum impact machine the entire weight
of the swinging parts must be divided into two parts, and these
parts concentrated at the axis of rotation and at the centre of
percussion. The latter part, only, multiplied by its vertical
drop is the measure of the energy of the blow (but this is the
same as the total weight into the fall of its centre of gravity).
To find the centre of percussion and the equivalent weight
to be considered as concentrated at this point, a graphical solu-
tion may be employed, as follows:
Let AG extended, Fig. 305, be the pendulum, with its axis
of rotation at A. Let G be the centre of gravity of the entire
pendulum, with all its rigidly connected parts (to be found by
trial). -Let GD, drawn perpendicular to AG at G, be made (to
the given scale) equal to the radius of gyration of the entire
pendulum about its centre of gravity G, (to be computed).
Then draw AD, and DC perpendicular to AD, cutting AG
extended in G. Then is C the centre of percussion of the
pendulum.* If the graduated arc (see Fig. 308) have a radius equal to AC,
and the vertical components of the pendulum's motion (versed sines) be laid
off on this arc, then the equivalent weight to be concentrated at G is to be
used for computing the energy of the blows, and this equivalent weight is
FIG. 805.
Graphical
Method of
Finding the
Centre of
Percussion.
or equivalent weight at
W c : W::AG:AC',
W.AG.
- Hc - ~
(1)
where TFis the total weight of the pendulum used in finding the centre of
gravity G.
If the graduated arc has a radius equal to AG, then the total weight W
is to be used with the versed sine to compute the energy of the blow.
If a conventional radius, R, has been used by the maker of the machine,
then a corresponding W r must be employed which will satisfy the equation
In every case, however, the point of impact of the pendulum should be
at C, the centre of percussion.
Rankine's Applied Mechanics, Art. 581.
378
THE MATERIALS OF CONSTRUCTION.
FIG.
-Most Approved Form of Impact-testing Apparatus. (Rep. French Com-
mission. )
IMPACT AND HARDNESS TESTS.
379
294. Impact-testing" Machines. In Fig. 306 is shown the impact-
machine designed and used by Prof. A. Martens in his testing laboratory at
Charlottenburg, Germany.* Its dimensions are given in meters. It admits
of an extreme fall of 4.5 meters (about 15 feet), and of a weight of ram of
200 kilograms (440 Ibs.), although he has used weights of 36 and 56 kilograms
only. It is intended for specimen tests only. The anvil weighs 1250 kilo-
grams (2750 Ibs.), or 22.5 times the heaviest ram usually employed. This in
turn is set on a strong cement-masonry foundation, separate from that of the
building, f as should always be done with drop machines.
The National Car-builders' Association of America has (1896) adopted a
spring support to the anvil, as shown in Fig. 307, in order to insure perfect
FIG. 307. Standard Impact-testing Machine with the Anvil-block on Springs, as
recommended by the Master Car-builders of America, 1896. Previously used by
the Penn. Ry. Co.
identity of reactions in different machines. This is perhaps the only way to
eliminate the varying effects of different foundations of the anvil-block.
Fig. 308 gives a view of Keep's autographic recording pendulum impact-
* See Mittheilungen am den Koniglichen Tcchnischen Versuchsanstalten zu Berlin,
1891, p. 2, and Plate I. The machine was made by E. Becker, machinist, Berlin.
f At first it was set on the floor, but it was found necessary to put it on a more solid
foundation.
380
THE MATERIALS OF CONSTRUCTION.
machine. It is used only for testing his standard cast-iron bars $ inch
square and 12 inches long. The hammer weighs 25 Ibs., and swings on a
FIG. 308. Keep's Impact- testing Machine.
IMPACT AND HARDNESS TESTS.
381
radius of 6 feet. The weight of the anvil is admitted to be too light, and
the designer recommends that it be set against a brick wall ! * The d,rc is
graduated to vertical drops of J inch, with a total fall of 6 inches. The
paper on which the deflection is recorded at each blow is automatically
moved T 3 g- inch after each blow, so that the record consists of a series of
parallel lines, each being the deflection for that blow, magnified four or five
times by the leverage of the recording apparatus.
TESTS FOR HARDNESS.
295. Hardness Defined. The term hardness is used in two senses, as
applied to metals, minerals, and other solids. It is used to signify
(a) Resistance to indentation (permanency of form) ;
(b) Resistance to abrasion or scratching (permanency of substance). f
These two kinds of hardness are more or less related, and are often con-
fused. In practice the demands for these two kinds of hardness are quite
distinct, and hence two very distinct kinds of tests are employed to determine
them.
296. Hardness Test for Permanency of Form or Resistance to Indenta-
tion. The only test of this kind which has ever been standardized is the
indentation test by means of a pyramidal steel punch, attached to a falling
FIG. 309. The Rodman Steel Punch for Hardness Tests. Dimensions in millimeters.
(Rep. Fr. Com.^ vol. in. p. 261.)
weight. The form most favorable to exact results is that chosen by Lieut. -
Col. T. J. Rodman (IT. S. A.) before 18GO,J; and shown in Fig. 309 in
metric measurements. These are used because this test has been standard-
ized in France, and the degree of hardness is given in metric units. Such
a steel point is rigidly attached to the base (or striking side) of the ram in an
impact-testing machine, such as shown in Fig. 300 or 307 or 308. The sur-
* Evidently the particular character of this setting will greatly affect tho force of the
blow on the specimen.
f After Osmond.
% See his Report of Experiments on Metals for Cannon and Cannon-powder, 1861.
While the author of this work has, as a rule, expressed quantities in English units
only, he makes an exception in this case.
382 THE MATERIALS OF CONSTRUCTION.
face of the substance to be tested is planed or filed flat and polished. The
steel point is then made to fall normally upon the surface from any desired
height, and the observation consists in noting
The weight of the ram = irin kilograms;
The height of the fall = h in millimeters;
The length of the indentation = / in millimeters.
The work done upon the body tested, or in producing the indentation, will
be Wli kilogram-millimeters, provided the anvil, or body struck, was very
massive and firm in comparison with the weight of the ram. This is, of
course, essential to the correctness of the assumption that the energy of the
falling body spends itself wholly in producing the indentation; and this must
be assumed and secured for a perfect accordance of results.
The volume of displaced material resulting from the indentation will be
ml 2 , where m will vary with different forms of pyramids, but will be a con-
stant for any one pyramid, or punch. It has been shown most conclusively
by Lieut. -Col. Martel * that when the essential conditions of the test are
satisfied,
For all forms of pyramids, for all weights of ram, and for all heights of
fall, the volume of the displaced material of a given quality is equal to the
energy of the blow (Wh) divided by a constant, D,\ which constant is the
work or energy necessary to displace (by deformation] a unit-volume of that
material. This constant is therefore characteristic of that material and may
be taken as its index of hardness, or of its resistance to indentation.
Since the kilogram-millimeter anits have already been used in France,
and the hardness of many kinds of materials has been found and published
011 this scale, it would lead to unnecessary confusion to change the units,
since this would change the numerical index of hardness.
To find the volume of the pyramidal displacement of the Rodman punch
(Fig. 309) from the measured length, multiply the cube of the length by
0.0009413 (log 1.97375), or vol. = 0.0009413Z 3 .
For any other -form of punch the volume would be readily computed, but
approximately this form is best, because it gives u very large length to be
measured for a very small volume. In other words, we argue from a longer
base, thus making the percentage error of . observation correspondingly less.
Furthermore, the indentation is shallow and hence injures the material
(which may be a finished or unfinished final form) to a less degree. Any
other form would, however, give strictly comparable results, so that there is
no real necessity of adopting this particular form. It goes without saying
that the punch should itself be so hard as not to suffer any permanent de-
formation in service.
The truth of the theory that V = - r has been established by Martel
* Commission des Methodes d'Essai des Materiaux de Construction, vol. in. p. 261.
f For durete, hardness.
IMPACT AND HARDNESS TESTS. 383
within the limits of the errors of observation, and hence can be accepted
with confidence, as giving an absolute standard by which to measure hard-
ness when this implies resistance to indentation. The following table con-
tains values of D (degrees of hardness On the Martel scale) in kilogram-
millimeter units for various metals.
DEGREES OF HARDXESS OX THE MARTEL SCALE.
Metals Tested. De - ree of Har dness.
Kilogram-millimeter Units.*
High carbon (" diamond'') steel, hardened in oil. . 613
" " " " not hardened 460
Medium steel (for cannons), hardened in oil 455 to 300
Hoop-steel (for large guns), hardened in water 330 to 295
Rolled wrought iron 226
Hammered wrought iron 238
Cast iron (for guns) 300 to 208
Bronze (cast in shells) 154
" after cold-hammering 238
" after drawing down 12$ on a mandrel 310
" cast in sand (C 88, Sn 12, Z 2) 137
" " " " without zinc. . 115
Copper, rolled . , 156
" reheated and cooled in water. . , 64
Zinc, rolled 77
Tin, cast 33
Lead, cast , . . . 9
297. Hardness Test for Permanency of Substance or Resistance to Abra-
sion. While a great many tests of this property have been devised and used,
none of them has given such satisfactory measurable results as the one just
described for resistance to induction. The scratch test has long been in use
for classifying minerals as to their hardness, and ten grades of hardness are
recognized under this test. It is purely relative, and is entirely inadequate
to the requirements of the user of metals. By this test the body A is harder
than 1> when a point or sharp corner or edge of A will scratch the surface
of 13, and when the converse will not hold.
Mr. Thomas Turner has devised the instrument shown in Fig. 310, and
this has been largely used by both Turner and W. J. Keep for the grading
of cast irons for hardness. In this a diamond point is fixed at the base of
a vertical pencil which is carried by a perfectly balanced arm. Provision is
made for loading the pencil by weights (in grams), and the hardness is
indicated by the number of grams required to make a standard scratch on
the surface tested. Evidently the standardizing of the scratch offers great
To change these figures to pound-inch units, multiply by 1422.
384
THE MATERIALS OF CONSTRUCTION.
difficulties, so that results obtained by this instrument in the hands of
different persons would probably not be strictly comparable.
Prof. Martens (Berlin) has undertaken to standardize this instrument by
making the load on the pencil a constant and measuring the width of the
FIG. 310. Turner's Apparatus for Testing Hardness.
scratch with a micrometer-microscope, and this is the method employed by
the German artillery officers.
Standard abrasion-machines have also been used (see Fig. 288), but it is
almost impossible to duplicate the conditions exactly.
In general, therefore, it may be said that there is now no absolute test
for hardness as meaning resistance to scratching or abrasion. (For a brief
account, without illustrations, by Osmond, of the many devices which have
been tried, see Report of the French Commission) vol. in. p. 279.)
CHAPTER XIX.
SHEARING AND TORSION TESTS.
SHEARING TESTS.
298. Essential Conditions of a Shearing Test. In order to obtain the
true shearing strength of any substance it is necessary to develop in it, along
a given plane, shearing stress only, unaccompanied by the bending stresses of
tension and compression. To accomplish this it is necessary to concentrate
the external forces of action and reaction on planes an infinitely small dis-
tance (dx) apart. Any finite distance between these planes will develop a
cross-bending action and its resultant direct stresses across the plane of
shear. As it is impossible to so concentrate the external shearing forces, it
is necessary to overcome the bending stresses set up by the non-concurrence
of the external forces by preventing the bending of the specimen subjected
to these forces. This can only be done by reinforcing the specimen between
the shearing planes. This may be done by grooving the specimen in the
planes of shear, or by supporting it by auxiliary clamps. As neither of
these expedients has usually been resorted to in shearing tests, it follows
that very few such, tests have ever been made in which shearing stress has
been unaccompanied by large direct stresses.*
299. The Occurrence of Shearing Stress in Practice. Shearing stress is
present in nearly all cases where there is cross-bending (see Art. 37), and in
rivets, bolts, bridge-pins, crank-pins, etc., shearing stress becomes of practi-
cal interest. In none of these cases, however, is it found acting alone, but
it is always combined with bending stress. In the case of rivets it is always
combined with a very great tensile stress, caused by the contraction of the
rivet in cooling after the heads have been made, this stress from contraction
in good work always exceeding the tensile stress of the rivet at its elastic
limit. In fact a riveted joint, if well made, always acts by frictional re-
sistance alone, since this is always more than the working stress on the joint.
(See a discussion of this subject in Chap. XXVI.) While rivets are computed
for shear, therefore, as a matter of fact they are seldon subjected to a shear-
ing stress.
* Both Dr. Kennedy and Mr. Barba grooved their specimens for double shear, and
also held them in rigid forms. See Rep. French Commission, vol. m, Plate XIX.
385
386
THE MATERIALS OF CONSTRUCTION.
For these reasons a knowledge of the true shearing strength of any of
the metals is of little value, except for purely scientific purposes, and for
computing resistance to torsion, where the stress developed is that of pure
shear.
In the case of timber, however, which more often fails in shearing along
the grain than in any other way, the strength in shearing is of great interest.
In general the shearing strength of the metals may be taken as 80 per
cent of the tensile strength.
300. Shearing-test Appliances. Shearing tests can be made in an ordi-
nary tension or compression machine, if suitable appliances be used for
holding the specimen. In Fig. 31 1 Dr. Kennedy's appliances for single and
FIG. 311. Dr. Kennedy's Appliances for Single and for Double Shear.
double shear are shown. For single shear the specimen is held by two half-
rounds, all enclosed in a cylindrical sleeve. The shearing-faces are rein-
forced by steel rings. The plane of shear thus lies- in the line of the axes
of the two compression shear-blocks. For double shear the specimen is
grooved on the shearing-planes, and it is also fitted closely into the eyes of
the steel links through which the forces are applied.*
In Fig. 312 is shown the shearing-apparatus designed by the author
for finding the shearing strength of cast iron. Here the specimen is
gripped firmly at both ends and in the centre, and all bending distortion
prevented. By preventing this kind of deformation the bending stresses are
of necessity avoided. The bearing shear-plates at top and bottom are of
hardened steel.
For shearing tests on wood the apparatus shown in Fig. 313 has been
extensively employed by the author. Blocks about 2 inches square and 8
inches long are slotted one inch from each end, in planes at right angles to
each other, and also bored at the centre for the fixed hold. A rectangular
steel pin is inserted in the slot, and the stick is prevented from splitting by
attaching a clamp with an initial pressure just sufficient to hold it in place.
The steel pin is pulled by means of bronze stirrups which are held in the
* This apparatus will not serve for cast iron. See Trans. Inst. Civ. Enyrs., vol. xc.
p. 391.
SHEARING AND TORSION TESTS. 387
regular wedge-grips of the testing-machine. After shearing out one end of
the specimen it is turned over, the lower stirrup revolved 90, and the other
end pulled.
FIG. 312. The Author's Shearing-test Apparatus.
TORSION TESTS.
301. Contrasted with Shearing Tests. While torsional stress is a purs
shearing stress (and about the only means of obtaining a pure shearing"
stress), yet a torsion test differs from a shearing test in that the deformation,
acts over any length of bar, taken at pleasure, and in that it is not uniformly
distributed across the section, but is zero at the centre and increases uni-
formly towards the circumference. This enables the modulus of shearing
elasticity to be determined by noting the angular distortion over a given
length of bar, and it also makes possible the obtaining of autographic (or
plotted) stress-diagrams for shearing stress. The elastic limit and ultimate
strength in torsion have a value in the designing of shafting of all sorts
which serve to transmit power.
302. Torsion-testing Machines. in Fig. 314 is shown a simple attach-
ment to an ordinary " universal ' testing-machine. The power is applied to>
the specimen A by the screw-gear //, and the torsion is resisted by a couple>
388
THE MATERIALS OF CONSTRUCTION.
Shearing
Test
Apparatus
for
Wood.
4mrm
FIG. 313. Sliearing-tLSt Apparatus for Wood.
SHEARING AND TORSION TESTS.
389
FIG. 314. Torsion-test Attachment.
FIG. 315. Torsion Machine for Short Specimens.
390
THE MATERIALS OF CONSTRUCTION.
one arm of which, G, bears on the weighing-table. The angular deforma-
tion is observed by means of the two collars E and D, the latter holding
rigidly a bar which moves a pointer over the graduated circle on the former.
In Fig. 315 is shown a torsion machine for testing large specimens in
short lengths, while in Fig 316 is shown a large machine for bars of any
desired length. The former machine is self-contained, while in the latter
SHEARING AND TORSION TESTS.
FTG. 317- A 3|-iu. square Bessemer-steel Bar Twisted Hot. (Cassicr's Mag., vol. x. p.
44,3, 1896.)
392
THE MATERIALS OF CONSTRUCTION.
the lifting side of the weighing end is held down to the track by bolts, and
the downward-bearing end of the conple-arm bears upon a system of weigh-
FIG. 318. Tet major's Torsion-testing Machine for Wires, giving Autographic Records,
ing-levers, it is made in three sizes suited to steel shafts 1 in., 2 in., and
3-J in. in diameter, and for 16 feet in length or less. In this machine the
specimen is free to contract longitudinally while under test.
FIG. 319.
A very perfect machine for testing wires from 0.05 in. to 0.18 in. in
diameter (No. 18 to No. 7 13. W. G.) and for giving (a) the breaking moment,
(b) the number of turns, and (c) the complete stress-diagram, is shown in
Fig. 318.* This machine is used by Prof. Tetmajer and described by him
* Made by Messrs. Amsler-Laffoii & Sous, Schafl'hausen, Switzerland
SHEARING AND TORSION TESTS. 393
in vol. iv. of his Communications. The specimen is kept in tension
during the test by a weight suspended by a cord connected to the carriage
at the resisting (and recording) end of the specimen. The resisting moment
is developed by means of two weights suspended by cords which run in
symmetrically arranged spiral grooves.
A simple machine without recording apparatus is shown in Fig. 319.
CHAPTER XX.
COLD-BENDING AND DRIFTING TESTS.
COLD-BENDING TESTS.
303. Their Character and Significance. The test of the ductility of a
malleable metal by bending it cold is the most common and perhaps the most
useful of all the tests which can be applied to it. For wrought iron and
structural steel this test approaches more nearly to the severe usages of actual
practice than does the tension test with its elastic limit, ultimate strength,
elongation, and reduction of area. It is not so easily standardized, however,
and it is employed less in America than in Europe, partly because no stand-
ard methods and results have been agreed upon here. If a sample of wrought
iron or steel will, when cold, fold upon itself absolutely, as shown in Figs.
Thickness of
FIG. 320. Cold-bending Test of a 51-lb. 15-iu. Steel Channel-bar,
web = 0.78 in. (Engr. News, vol. xxxm. p. 212.)
320 and 322, or make the double fold as shown in Fig. 321, there can be
no doubt of its high quality. When it fractures, however, at intermediate
stages of this process, the question of its quality is left in doubt, and some
standard limit is required if this test is to be made the basis of acceptance.
The great advantage of this test is that it can be made at any time in the
shop, without the expense attaching to tension tests, and by the man who
uses or makes up the material. No standard method of making this test,
therefore, should remove it beyond the range of ordinary shop appliances.
In Europe a number of special machines are in use for making these tests,
394
COLD-BENDING AND DRIFTING TESTS.
395
but only siiop-tools will here be assumed as available. With these methods,
and in the hands of the same operator, uniform and comparable results may-
be obtained.
304. Methods of Making Cold-bending Tests. If the specimen is not
too large, a strong vise may be employed. If the bend is to be a true fold
(radius of curvature = 0), the specimen should be bent about the sharp edge
of the vise. If it is to be bent to a given radius, an auxiliary plate, dressed
to this radius, mast be clamped with the specimen in the vise. In either
FIG. 321. Double Cold licuds on f-iu. Su-el Plates. (Eng. News, vol. xxxin. p. 272.)
case the specimen must be damped fast to a long steel bar, or lever, so as to>
prevent all bending beyond the curved section. For this purpose two clamps :lc
are required, one of which must be close down to the vise.* The specimen,
is then bent to 90 by hand. Striking the specimen with a hammer should,
be avoided, as this kind of action cannot be standardized. If the specimen
is to be folded flat upon itself, it may be removed from the vise after it has'-
been bent to a right angle, and a second bar clamped to the other leg, and
these two bars can now be drawn together by hand. The final closing down
of the specimen may be done in a vise or under the hammer, a steam-
hammer always preferred.
The French Commission have adopted the interior angle as the index of
the ductility. Thus if a straight bar bends through an angle of GO before
rupture, it leaves an angle of 120, and this is the angle of record. A record
of signifies that the bar has bent through 180, and that it has been
either closed down Hat or bent to a given radius, according as the radius of
the bend is given as zero or something greater.
* Specially devised stirrups or clevises should be made up for clamping the specimen-
to the bending bar.
396
THE MATERIALS OF CONSTRUCTION.
If the specimen is too large to bend by hand as described above, it may
be bent under a steam- hammer (or in a hydraulic or screw press, or in a test-
ing-machine, or even by a heavy sledge), by resting it on supports as a
be-im and striking it at the centre. After bending it in this way through an
.
.
FIG. 322. Soft Bessemer-steel Bars, 3 in. by 2 in. in cross-section, Bent Cold. (Cassier's
Mag., vol. x. p. 442, 1896.)
angle of about 60, it may be set on end and struck by the hammer (or placed
in a press or testing-machine), as a bent column, and so brought down to
any desired angle or radius of curvature. This required radius of curvature
will ha.ye to be reached by flattening down the bent bar, nfter the zero angle
COLD-BENDING AND DRIFTING TESTS.
39T
(180 of curvature) has been attained. The ideal appliance here is a press
of some sort, but this requires a special machine (perhaps an ordinary " bull-
dozer," used for straightening, or curving members, might serve), and in the
absence of these a steam-hammer answers very well. A sledge is not good,
as it is too light and requires blows having too high a velocity, which spend
their energy in deforming the specimen at the point of impact and may
produce its rupture earlier than the other methods would.
Prof. Tetmajer has a machine for making bending tests without the use 1
of a mandrel, and whereby a uniform bending action is given to the bar.
This develops the distributed elongation of the specimen, whereas a bend,
concentrated at one point develops the " reduction of area " quality. Thus
a high-grade steel wire which will not elongate over two or three per cent
may, on failure in tension, show a reduction of area of GO per cent. Such
a wire would fold over to a much sharper curve (smaller radius) than it
could be bent to through a full circle.
Preparation of the Specimen. If the specimen has been cut from a-
plate or from a structural form, and it is to be tested in comparison with or
on the same basis as rolled bars, either round or rectangular, then the
sheared edges should be removed by planing or filing where the bending is
to be effected, in order to remove the brittle material resulting from the
shearing action.
On the other hand, if it is desired to learn the action of the metal after
it has been punched or sheared or threaded, then the specimen is purposely
so prepared and tested without removing these hardened and serrated sur-
faces. The cold-bending test of such prepared specimens develops the
injurious effects of these shop processes (punching, shearing, and threading)
as nothing else can, and it is therefore necessary to use it for such purposes.
The French Commission have recommended a length of 10 inches and a,
width (of plate specimens) of 1.6 inches, the thickness to be that of the plate
or bar.
Sometimes specimens are nicked or grooved on one side and then broken
in cross-bending, under the hammer, to test relative brittleness. This is-
not a test that can be relied on to give, absolute results, but Prof. Tetmajer
used it to good effect to disprove the commonly accepted theory that even
low steel is more brittle than wrought iron when subjected to shocks-
Photographic views of the results of these tests are shown in Fig. 323. The-
depths of these specimens are shown in the figure. The six upper ones were
0.8 inch thick, and the four lower ones 1.2 inches thick. The tension tests-
on these specimens gave the following average results (eighteen tests on
each material) :
Material.
Modulus of
Elasticity.
True Elastic
Limit.
Apparent
Elastic Limit.
Ultimate
Strength.
Per cent
of Elon-
gation.
Reduc-
tion of
Area
Low steol
Wrought iron . ,
f,bs. per sq. in.
81.000,000
28,600,000
Lbs. per sq. in.
28,500
21,800
Lbs. per sq. in.
36,600
33,000
Lbs. per sq. in.
61,000
52,200
jgin 8
" 27.8
160
59.3
2.1.4
398
THE MATERIALS OF CONSTRUCTION.
FIG. 323. Showing Relative Resistance to Shock of Wrought iron and Basic Bessemer
Steel Bars which hail been Grooved on the Tension Side to a depth of 0.06 in.
(Tetmajer, vol. iv. PI. X.) W = wrought iron, S = steel. The hammer weighed
660 Ibs. and fell 9 in., 19 in., "29 in., 51 in., uud 79 in., for the five sizes respectively,
the span being always 40 in. The number of blows were 2, 3, 2, 2, 1, 4, 1, 7, 1,
and 6 for the ten specimens as shown above, respectively.
_j
o a
oTS" *
ET.^- H : *
COLD-BENDING AND DRIFTING TESTS.
399
The great superiority of the steel in the impact tests is evident from the
figure.
In the cold-bending tests on the same forms, flatwise, the steel specimens
0.8 inch thick folded flat without sign of rupture, while those 1.2 inches
thick cracked in flattening down. The wrought-iron specimens 0.8 inch
thick cracked after bending through an angle of 120, and the 1.2-inch
specimens after bending through an angle of 60.
305. Comparison of Results of Cold-bending Tests with the Tensile
Strength and the Percentage of Elongation. In Bacle's report on Various
Cold Tests of Materials, in the French Commission Report, vol. in. p. 311,
are given the mean results of several thousand tests in tension and cold bend-
ing on wrought iron and soft steel, received from many different sources,
made by M. Hallopeau, a member of that Commission, and given in the fol-
lowing tables. These results are not only of great value for the information
itself, but they will serve to enable one to prepare standard specifications for
cold-bending tests from the relation of the results of these to the results of
tension tests which have long been the standard tests of acceptance in this
country.
TABLE XVIII. RELATIVE RESULTS OF TESTSIOX AXD COLD-BEN'DIXG TESTS.
Tension Test.
Bending Test.
Material.
Elastic Limit.
Ultimate
Strength.
Elonga-
. tion.
Cracked
at Angle
of
Ruptured
at Angle
of
Wrought iron with grain
Lbs. per sq. in.
32 000
Lbs. per sq. in.
48 500
Percentage
14
Degrees.
80
Degrees.
30
" " across "
30,000
42,000
6.5
130
115
Low steel (57,000 to 60,000 T. S.)
29,500
60,000
27.0
Medium steel (64,000 T. S.)
34,000
65,000
25.0
High steel (08,500 to 71, SOOT. S.)
35,500
72,000
24.0
The tension-test bars were 4 in. long, 1.6 in. wide, and 0.4 in. thick.
The bending-test bars were 6 in. " 1.6 in. " " 0.4 in. "
The bending-test angle of record is the angle formed by the two ends of the bar after
bending, or it is the supplement of the angle through which bending has taken place.
It will be seen from the above table that while the bending test would
serve to distinguish varying qualities of wrought iron, it would riot serve to
distinguish these three grades of steel, since they all bent through 180 and
folded flat without even cracking. Since these grades require distinction in
practice because of the different degrees of injury produced on them by
punching, the specimens can be punched and then bent, and the shades of
hardness indicated by the greater angles at whicli cracks appear. This has
been done by M. Hallopeau, and the results are shown in Table XIX. The
plates were all the same size as those given in the previous table, the punched
and drilled holes being 0.8 inch in diameter, or one half the "width of the
bar, the thickness being 0.4 inch. The die side of the plates was made the
tension, or convex, side.
400
THE MATERIALS OF CONSTRUCTION.
TABLE XIX. EFFECTS OF PUNCHING AND DRILLING ON WROUGHT IRON
AND STEEL, AS DETERMINED BY TENSION AND COLD-BENDING TESTS
ON BARS 0.4 INCH THICK.
Material.
Tension Test.
Bending Test.
Elastic
Limit.
Ultimate
Strength.
Elonga-
tion.
Cracked
at An-
gle of
Rupt'd
at Angle
of
WROUGHT IRON.
Plain.
, ( With the "-niiii .
Lbs.
per sq. in.
32,000
30,000
20,000
26,500
33,500
39,000
28,500
28,500
29,000
26,000
42,000
29,000
Lbs.
per sq. in.
48,500
42,000
47,000
41,000
53,000
46.000
47,000
41,500-
38,500
32,000
46,000
34,000
56,000
38,000
43,500
37,500
42,000
36,000
44.500
33,000
54,000
44,000
41,000
35,000
60,500
59,000
65.500
59,000
60,000
55,000
67,500
58,000
60,200
59,700
67,000
59,700
Percent-
age.
14.0
6.5
19.8
8.3
15.0
3.3
16.7
7.7
1.0
1.0
2.4
1.0
l'.6
1.5
2.7
1.5
3.3
1.8
1.8
0.8
3.2
1.5
27.0
27.8
22.4
30.0
4.5
4.6
4.1
5.0
6.8
7.4
68
7.1
Degrees.
80
130
65
115
80
135
100
105
175
180
170
175
175
178
170
175
172
177
176
177
175
177
172
178
100
90
150
90
45
40
110
20
Degrees.
30
115
20
85
40
100
50'
80
150
175
150
165
160
175
150
165
165
172
172
173
168
173
165
175
60
59
135
60
10
100
10
Natural \ Across tof grain. ......... .....
( W"ith the 'raiu
Annealed -j Acrogs lhc r ffrain
Hardened in w a ter|S^|-::;:.
Hardened and annealed | SsstL^lin.
Punched.
, ( With the o-raiu. .
Natural j ^TOSB the e grain. . . . .... ..'.'.
t ^ ( With the 12: rain ..
AnDealed | Across the graii .. .
( With the crain
Hardened in water J Aci . ()ss ,^ 8 gl ,, in ; ...
Hardened and an nealed|S^|^ ;
Drilled.
, ( With the erram
33,500
29^500
28,500
25 000
Natura M Across the grain
29,000
24,500
36,500
30,000
29,500
27,000
29 500
Amiealed | Across thl grain
Hardened in water ) ^'- ifc ain. n ......
Hardened and annea,ed|Wit 1 , s the eg n ...
LOW STEEL.
Plain.
Natural
An nealed
28 500
Hardened in water
34,000
29,000
36,200
36,500
45,000
36,400
36,000
34,200
39,200
33,500
Punched.
Natural
Hardened in water
Hardened and annealed
Drilled.
Nat u ral
Annealed
Hardened in water
Hardened and annealed
COLD-BENDING AND DRIFTING TESTS. 401
EFFECTS OF PUNCHING AND DRILLING ON WROUGHT IRON AND
STEEL continued.
Material.
Tension Test. Bending Test.
Elastic
Limit.
Ultimate
Strength.
Finn, i , Cracked | Rupt'd
S atAn - : t Angle
lon - gle of of
MEDIUM STEEL.
Plain.
Lbs.
per sq. in.
34,000
32,800
37,000
33,800
36,500
35,500
51,000
37,000
36,500
35,500
51,000
36,000
35,500
36,200
43,500
35,500
47,000
47,500
71,000
47,000
39,700
38,600
56,700
39,000
Lbs.
per sq. in.
65,500
64,000
73,500
66,000
67,000
64,000
77,000
64,500
66,000
64,500
73,000
65,000
72,000
70,000
80,000
71,300
72,000
71,000
80,000
71.000
71,800
69,000
82,000
71,600
Percent-
age.
25.0
26.8 :
23.8 !
26.0
4.5
4.7
2.2
4.7
6.2
6.5
6.3
6.5
24.0
25.0
20.0
25.0
3.9
3.4
3.4
3.4
5.5
5.6
5.8
5.7
Degrees.
100
95
145
100
70
65
100
60
20 to 60
140
120
160
135
80
70
140
85
Degrees,
80
75
130
75
50
40
80
40
105
95
145
105
60
55
125
60
Hardened in water . . . .
Hardened and annealed
Punched.
Natural
Annealed
Hardened in water
H'li'deued and annealed .
Drilled.
Natural .
Hardened in water
Hardened and annealed . .
HIGH STEEL.
Plain.
Natural
Annealed . . . . . .
Hardened in water
Hardened and annealed
Punched.
Natural . . .
Hardened in water .
Hardened and annealed
Drilled.
Natural <
Annealed . ........
Hardened in water
Hardened and annealed
NOTE. All the bars were 1.6 in. wide and D.4 in. thick. The elongation in the
tension test was measured on a length of 4 in. with both the plain and the punched or
drilled specimens. In the former case this elongation occurred throughout this entire
distance (4 in.), while in the latter it occurred only in the vicinity of the hole, but was
credited to the entire distance of 4 inches in computing the percentage of elongation.
These percentages are therefore not comparable as showing loss of ductility (as has been
assumed in the Rep. French Com.).
The bending-test angle of record is the angle formed by the two ends of the bar after
bending.
402 TEE MATERIALS OF CONSTRUCTION.
The following conclusions may be drawn from Table XIX :
1. The reduced ductility of wrought iron across the grain is fully brought
out both in the elongation of the tension tests and in the angles of rapture
in the cold-bending tests,
2. The weakening effect of both punching and drilling is very much
greater with the wrought iron than with the soft and mild steels, and some-
what greater than it is on the medium steel.
3. The annealing of the punched specimens in no case appreciably
increased their ductility, as shown by both the tension and the bending
tests. It increased the strength of the wrought-iron specimens somewhat,
but it lowered the strength of the steel specimens.
4. The drilled specimens of wrought iron do not differ appreciably from
the punched either in strength or ductility, while with the steel of all
grades the ductility of the drilled specimens is far greater than that of the
punched specimens, although the ultimate strength is the same.
5. The change from 65,000- to 72,000-lb. steel is very clearly indicated
by the bending test, where in the punched specimen, " natural," the angle
through which the specimen bent before cracking is 100$ greater with the
former than with the latter. No such difference appears as between the
60,000-lb. and the 65,000-lb. steel, showing that they are about equally well
adapted to such work. In the " plain " specimens all three grades of steel
closed down entire (angle 0) without sign of failure, thus manifesting no
difference in hardness. The bending test on punched specimens, therefore,
develops clearly this difference in fitness for riveted construction, and it
might well be used as a shop-criterion of acceptance. Thus the 60,000-lb.
and the 65,000-lb. steels bent through an angle of 80 after punching before
a crack appeared, while the 72,000-lb. steel bent through an angle of only
40 before cracking. If an angle of 60 were specified on this test for plates
0.4 in. thick (leaving an angle of 120 formed by the two ends of the bar)
before a crack should appear, it would seem to rule out the higher carbon-
steels, which are injured by punching and shearing. This angle would be
different, however, for different thicknesses of plate.
306. Combined Specified Requirements in Tension and Cold Bending.
The combined requirements given in Table XX are reproduced from the
Report of the French Commission, vol. in. pp. 342-353. While the joint
requirements of many French government bureaus are there given, only
those of the Ar tiller ie de terre are here given.
The Committee of the American Society of Civil Engineers has recom-
mended (1896) the following cold-bending tests of plain specimens:
Wrought-iron specimens should bend through 90 without fracture, with inner
radius not exceeding twice the thickness of the test specimen for bar-iron nor three
times that thickness for plate and shape iron.
Rivet-iron and Rivet-steel bars, when heated to a low cherry-red and quenched
in water (this for the steel bars only), must bend through 180 to a close contact
(radius = 0) without sign of fracture.
COLD-BENDING AND DRIFTING TESTS.
403
Low Steel (60,000 Ibs. T. S.), when treated in the same manner, must bend to a
2ero angle (through 180), with an inner radius equal to the thickness of the speci-
men, without sign of fracture.
Medium Steel (65,000 Ibs. T. S.) specimens, cut from bars, plates, or structural
forms, in their natural state, must bend through 180, with an inner radius equal to
one and one-half times the thickness of the specimen, without sign of fracture.
Hiyh Steel (70,000 Ibs. T. S.) specimens, cut from plates and forms, in their natu-
ral state, must bend through 180 to an inner radius equal to twice the thickness of
the specimen without showing sign of fracture.
TABLE XX. COMBINED REQUIREMENTS IN TENSION AND COLD BENDING.
Material.
Tension.
Thick-
ness
of
Speci-
men.
t
Cold Bending.
Ultimate
Strength.
Elonga-
tion.
Angle
before
Cracking.
Radius
of Bend.
0.5Z
0.5Z
1.5 to 2.0*
2t
5t
0.5Z
to 1.5*
1.5*
t
<*
t
WKOTJGHT IRON.
Rolled Forms (Round and
Rectangular).
First-class charcoal iron, threaded 1
First-class puddled iron, threaded
Good " " plain ....
Common " .
Lbs. per sq. in.
48,500
48,500
48,' 500
5o!66o
50,000
48,000 to 60,000
48,000 to 64,000
57,000 to 68,500
60,000 to 71,000
57 000 to 68 500
Per cent.
25
25
'25'
io'
10
....
26
23 to 25
22
21
21 to 23
In inches
t < 1.6
t < 1.6
t < 1.6 4
t < 6.6
t^ 0.6
t < 0.4
t < 0.4
0.2
>0.2
r <0.4
.08
y > .08
'<0.2
,> 0.2
1 <0.4
any t
any t
t <0.2
t> 0.2
t < 0.2
t > 0.2
Degrees.
0*
O 3
90
90
j- 90
\
| so
Iron for bolts threaded
Plate Iron. 5
WITH THE GRAIN.
First-class charcoal iron
ACROSS THE GRAIN.
Refined puddled iron . . -
a n
STRUCTURAL STEEL.
Low and Medium.
Rolled forms hardened
Plates, 6
High Steel.
Rolled forms, hardened
Plates, 5
1. Screw-threads cut on bar where the bending occurs, ns shown in Fig. 324.
2. This is the angle formed by the two ends of the bar after bending.
3. Some cracks are allowed here at the bottoms of the threads.
4. When the iron has a greater thickness than 1.6 in. it is to be cut down to this
thickness.
5. Specimens sheared off and filed up smooth.
307. Comparison of Tension, Impact, and Cold-bending Tests. It will be
seen from the following tables that the loss of ductility in punching iron
404
THE MATERIALS OF CONSTRUCTION.
and steel plates, and the benefit of subsequent annealing, are best developed
by impact tests. Also, the benefits of enlarging punched holes by boring
and reaming. The tables are compiled from M. Hallopeau's experiments
described above, and are given in Eep. Fr. Com., pp. 356-7.
These test-bars were 8 in. long, 2.4 in. wide, and 0.32 in. thick. The
punched and drilled holes were 0.8 in. in diameter, or one third the width
of the plates. The hammer used in the impact test weighed 88 Ibs., and it
had a constant fall of 16 in. . The average sums of all the heights of fall
before cracks appeared are given in the table. The figures given are the
average results of many tests.
TABLE XXI. COMPARISON OF RESULTS BY TENSION, IMPACT, AND COLD-
BENDING TESTS ON. PUNCHED AND DRILLED PLATES.
Tension Tests
Impact Tests.
Cold-bending Tests.
on
the Plain Specimens.
Total Height of
Drops.
Angles when
Cracks appeared.
Holes Punched.
Holes Punched.
Material.
i
0.04 in.
0.04 in.
1
i
a
i
small.*
I
small.
3
_o
a
99
N
'Z
|
d
MN
&0
03
CQ
"3 .i
B~
CO
*&
c^
a
OJ
C 3
3
g
w
O
M
1
I 5
So
o
ffi
i
1
go
Ibs. sq. in.
Ibs. sq. in.
%
in.
in.
in,
in.
deg.
deg.
deg.
deg.
Wrought iron, natural. .
" annealed
39,000
55,000
14.0
32
37
16
37
27
43
32
37
173
173
177
173
174
172
174
173
Steel natural. ,
43,500
59 500
99 5
108
6?
85
77
164
168
165
166
" annealed
130
107
128
112
159
161
160
161
* This is too small an enlargement to remove the material injured by punching, and
hence these results do not fully develop the differences of treatment. J. B. J.
TABLE XXII. COMPARISON OF RESULTS OF THE IMPACT TESTS.
Relative Treatment of Specimens.
Wrought Iron.
Steel.
Annealed.
Not
nnealed.
O3 ^
0) iX
II
1
T3
V
If
>
4OO
/ f
\ /
x/*
.--'-
ses-
/
/
T^
^
7*t2
-'-'
^
2OO
if
,rr
&
?7/?l
I
FIG. 340. Tensile Strength of Natural Cement-mortars, mixed with Different Percent-
ages of Water. (Wheeler, Rep. Ckf. Engrs., U. S. A., 1894, p. 2332.)
TESTING OF CEMENT.
26 3O 3S
423
430
2OO
/OO
PERCENTAGE QF WATER
FIG. 341. Effect of Varying Percentages of Water on Time of Setting of Xeat Cement.
(Wheeler, Rep. Clif. Engrs., 1895, p. 2935.)
24 28 <32 <3G 40
FIG. 342. Effect on the Strength of Louisville (Natural) Cement, Neat, of a Varying
Percent of Water in Gauging. (Jour. West. Soc. Engrs., vol. i. p. 82, Table XVI.)
used, this volume being that necessary to satisfy the conditions described in (2).
The water to be either fresh or salt, as may be specified. The whole is then stirred
and turned rapidly with a trowel for five minutes, counting from the instant the
water was added.
(2) With a portion of this gauged cement fill a vessel having an interior form of
a truncated cone, 8 cm. (3J inches) in diameter at bottom, 9 cm. (3f inches) in
diameter at top, and 4 cm. (If inches) deep, smoothing it off quickly on top with
the trowel.
Upon the centre of this top surface bring to bear normally and slowly a cylinder
of polished metal 1 cm. (f inch) diameter and weighing 0.3 kilogram (11 oz.), having
a full, flat, transverse sectional base. The apparatus to be constructed so as to
424
TEE MATERIALS OF CONSTRUCTION.
indicate the thickness of the film of mortar remaining below the cylinder when it
ceases to settle under its own weight. Two tests to be made on the same cake.
The consistency to be considered as normal when the cylinder stops just J inch
from the bottom of the cake.
For quick-setting cements use one half the amount of dry cement, and mix one
minute instead of five.
FIG. 343. Effect of Varying Percentages of Water iu Gaugiug Utica (Natural) Cement-
mortar, 1 C. : 1 S. (Jour. West. Soc. Engrs., vol. I. p. 82, Table XV.)
317. Normal or Standard Sand. That the quality of the sand exerts a
marked influence on the strength of cement-mortar is shown by Figs. 344
and 345. These tests show the great superiority of calcareous over siliceous
FIG. 344. Effect of the Quality of the Sand on Strength of Cement-mortar, 1 C.: 3 S.
(Wheeler, Rep. Chf. Engrs., U. S. A., 1894, vol. iv. p. 2321.)
sands in giving strength to the mortar. Evidently, however, sands con-
taining small shells containing air-spaces should be excluded.
To find the composition of a sand immerse it in cold hydrochloric acid,
which will dissolve the siliceous portion. The residuum may then be sepa-
rated into the insoluble calcareous sand and the clay, by rubbing and wash-
TESTING OF CEMENT.
425
ing, and thus its three significant constituents determined with sufficient
accuracy for commercial purposes.
Not only the strength but the permeability of mortar depends on the
size of the sand-grains; and as the resistance to the decomposing action of
800
700
00
00
FIG. 345. Comparative Value of Different Sands in Portland-cement Mortar 18
months old (in water). (Wheeler, Rep. Chf. Engrs., 1895, vol. iv. p. 2953.)
frost and sea-water depends almost wholly on its impermeability, the life
of the mortar, in exposed situations, is largely dependent on the character
of the sand used.
In order that tests of the strength may be comparable, therefore, it is
necessary to choose a normal, or standard, sand. In Germany and in France
natural sands are chosen, while in the United States a committee of the
American Society of Civil Engineers recommended in 1885 the use of
crushed quartz, such as is used in making sand-paper, of a size which passes
a No. 20 sieve and is stopped on a No. 30 sieve. This leaves all the grains
with maximum dimensions of from 1 mm. to 1.5 mm. When the grains are
so nearly of the same size and very angular or "splintery" the proportion
of voids is very great, so that a mixture of 1 cement to 3 sand by weight will
not be solid without a great amount of pounding on the briquette, which
must be mixed dry to enable it to receive such treatment.
426 THE MATERIALS OF CONSTRUCTION.
225
2OO
/SO
7 /4 2J 28
FIG. 346. Comparative Value of Three Kinds of Sand for Portland- cement Mortar,
1 0. : 3 S. (St. Louis Water Dept., 1895.)
300
20-4O
/MSSEO/0
FIG. 347. Effect .of Fineness of Sand on Strength of Portland-cement Mortar, 1 : 1 and
1 : 2. Age 6 months. (Wheeler, Rep. Clif. Engrs., 1895, p. 2972.)
TESTING OF CEMENT.
427
The French Commission have adopted the following:
Normal or standard sand consists of sand found on the beach at Leucate, and is
of three sizes:
No. 1, passing a sieve of 1 mm. and retained on one of 0.5 mm. mesh.
No. 2, passing a sieve of 1.5 mm. and retained on one of 1 mm. mesh.
No. 3, passing a sieve of 2 mm. and retained on one of 1.5 mm. mesh.
Simple normal sand is construed as meaning No. 2. Composite or
mixed normal sand is construed as meaning all three sizes in equal parts,
this mixture approaching closely the sand ordinarily employed in engineer-
ing works. The finest grade, No. 1, corresponds to such fine sand as is
found in the sand-dunes along our sea and lake shores, while the coarsest,
No. 3, corresponds to the very coarse sand taken from the bed of a rapidly-
flowing river. The composite or mixed sand is exclusively used in standard
tests of mortar. This is to be commended, as it gives fewer voids, and a
mixture of 1 cement to 3 sand readily makes a perfectly solid test specimen.
The above sieves, having meshes of 0.5, 1.0, 1.5, and 2 millimeters, would be
found to have approximately 35, 20, 15, and 11 meshes per inch respectively.
FIG. 348. Showing Effect of Varying Fineness of Clean River-sand in Cement-mortar,
1 C. : 3 S. (Wheeler, Rep. Chf. Engrs., vol. iv, 1894.)
In place, therefore, of using a 20-30 sand in making up standard mortar-
test specimens, as has become customary in America, in accordance with
the American Society of Civil Engineers Committee's recommendation, the
French are using a sand composed equally of three grades, which are respec-
tively 11-15, 15-20, and 20-30 sieve samples. This gives sand-grains vary-
ing from 0.5 mm. to 2.0 mm. in size, or a variation in size of 300$ of the
smallest, while the American Society of Civil Engineers' standard allows
428
THE MATERIALS OF CONSTRUCTION.
& S0 & 25
FIG. 349. Effect of Size of Limestone Screenings when used as Sand in Portland-
cement Mortar, 1 C. : 3 S. (Wheeler, Rep. Chf. Engrs., 1894, vol. iv.)
/) The minimum section should be small enough to insure rupture on
this portion of the briquette, but the reduction should be by gentle curves.
The English form (Fig. 354) is very defective in the first requirement, as
434
THE MATERIALS OF CONSTRUCTION.
FIG. 355. The European Continental Form of Briquette.
__H_ -
/// " I / or r a = 1.35r 6 . That is to say,
the form shown in Fig. 358 should be 35$ stronger than that shown in Fig. 360.
TESTING OF CEMENT.
439
briquette by the concentration of the load on a line or on a few points.
Very hard rubber pieces should be dovetailed into the metal clips at the
bearing-surfaces, and allowed to project a little beyond the metal.
When the second requirement is satisfied it is advisable to use an adjust-
ing frame for placing the clips symmetrically on the briquette.
FIG. 362. Apparent Varying Tensile Strength of Cement for Different Areas of Cross-
section of Briquette. (Grant and Whittemore, Engr. News, Dec. 14, 1893, p. 468.)
All these demands are well satisfied in the form of clip and adjusting-
frame shown in Fig. 363, which were devised by the author to be used with
his new form of briquette (Fig. 358).*
These clips are suspended from a steel point, the same as in the
German (Michaelis) machine. The hard-rubber (gutta-percha) pieces R
bear directly on the tangent surfaces of the briquette, and they are brought
to a symmetrical position (after the briquette is placed) by means of the
adjusting-frame, which is set over the screw-heads at top and over the
raised guides at bottom and then slipped downwards to a bearing. The
screw-clips at bottom are then turned, when both the briquette-clips are
held rigidly against the adjusting-frame and in their true position. The
movable clip is then screwed down to a hard bearing on the briquette, and
the adjusting-frame removed. It might, however, remain on throughout
the test if preferred, and in fact it could be permanently attached to the
rear sides of the clips.
* Both the moulds and the clips with their adjusting-frame are made by Mahn &
Co.. instrument-makers, St. Louis, Mo.
440
THE MATERIALS OF CONSTRUCTION.
These clips, used on briquettes of the form shown in Fig. 356, greatly
increase the breaking strength of neat Portland cement.
FIG. 363. The Author's Form of Clip and Adjusting-frame. Half-size,
used with the form of briquette shown in Fig. 356
To be
324. The Testing-machine. On the Continent Michaelis' machine, shown
in Fig. 3G4, is almost universally employed. The leverage is 50 to 1, and
the load is imposed by means of small shot which escapes from a reservoir
into the weight-pan. The dropping of this pan when the specimen breaks
shuts off the flow of shot. The pan is then weighed, and its weight, multi-
plied by 50, (this may be done in the graduation of the scales) gives the
strength of the briquette.
A very neat modification of this machine is that made by the Fairbanks
Scale Co. and shown in Fig. 365. It is entirely self-contained and dis-
penses with both the auxiliary reservoir and frame and the weighing-scales.
The shot-pan is moved from the end of the weighing-lever and hung from
TESTING OF CEMENT.
441
the hook at the left, and a weight-hook hung in its place on the weighing-
beam. The poise is then moved out on this beam, its extreme movement
corresponding to a load of 200 pounds. For greater loads " 200-pound "
weights are placed on the weight-hook and the poise moved out again to
balance. In both the machines the load comes on gradually and without
IBP
FIG. 364. Standard Form of German Cement-testing Machine for Tension Tests.
FIG. 365. The Fairbanks Cement- testing Machine.
shock by the flow of the shot, which is automatically shut off by the drop-
ping of the beam, and the rate of imposition of the load can be regulated
by varying the size of the gate. The tightening-screw P is first turned till
the weighing-beam moves to its highest limit, and then any further amount
to put an initial stress in the specimen short of rupture, after which the
442
THE MATERIALS OF CONSTRUCTION.
free movement of the weighing-beam is sufficient to break the specimen
without any further turning of the tightening-screw. The clips contain
adjustable bearings intended to prevent the breaking of the briquettes in
the clips.
77M
lW
20 40 #0 30 J00
FIG. 366. Effect of Varying the Rate of Loading on the Tensile Strength of Neat
Portland-cement Briquettes. (Faija, in Trans. Inst. C. E., vol. 75.)
600
00 $00
FIG. 367. Effect on Tensile Strength of Rate of Applying Load. (Wheeler, Rep
Chf. Engrs., 1895, p. 2951.)
The machines shown in Figs. 368 and 369 are made by Messrs. Riehle
Bros, and by Tinius Olsen, respectively, both of Philadelphia. They both
require the imposition of the load by hand ; but as this is through a screw-
gear and is very slowly applied, there would seem to be no appreciable un-
steadiness in it. The beam is kept in balance at the same time by the same
attendant, by moving out the poise till rupture occurs. Evidently the
speed here is entirely under the control of the operator, and both these
machin.es give entire satisfaction.
TESTING OF CEMENT.
443
In all these testing-machines the grips or clips are swivelled and
mounted in such a way as to allow of a free universal movement, or adjust-
ment of these to the specimen.
FIG. 368 Riehle Cement-testing Machine.
Fig. 370 shows the construction of a cement-testing machine designed
and used by Prof. J. M. Porter.*
"The load is applied hy water flowing into a tank suspended from the
long arm of a very sensitive 15-to-l lever. The weight of the lever and
tank is counterbalanced by an adjustable weight shown on the left. Water
is admitted to the tank from a large reservoir on the roof under a practi-
cally constant head of 90ft., so that there is no sensible variation of pressure-
in the stream admitted through a carefully fitted gate-valve in the supply-
pipe. The position of this valve at " on," " off," and all intermediate points
is shown by an index attached to the stem of the valve and registering on a
*The following description by the inventor is taken from the Engineering News of
March 5. 1896.
444
THE MATERIALS OF CONSTRUCTION.
dial marked off with the number of pounds per minute applied to the
specimen as determined and verified by previous experiment.
" When the briquettes break, the lever drops a few inches, then the
plunger at the right end of the lever enters the pneumatic stop, and the
lever and tank are gradually brought to rest. During the fall of the tank,
and before it comes to rest, a chain attached to the end of the valve-
stem in the tank is brought into tension and arrests the descent of the
FIG. 369. Olsen Cement-testing Machine.
valve before its seat stops descending. The opening of this valve allows
the contents of the tank to be quickly discharged into a hopper placed
upon the floor, and is then carried off through a waste-pipe to the sewer.
As soon as the tank has discharged its contents, the weight ou the left end
of the lever brings the lever and tank into the position shown in the illus-
tration, the valve taking its seat during this movement, and the machine is
ready for another break. The actual load can be applied at from to 80
Ibs. per minute, thus giving an increase of stress of from to 1200 Ibs.
TESTING OF CEMENT
445
per minute. The speed generally used is 400 Ibs. per minute, and with
the valve set for this speed the needle-beam will float every time within
I second of the proper time.
" The stress on the specimen is measured by a poise travelling on a gradu-
ated scale-beam, which can be read by means of a vernier to 1 Ib. and can
be moved automatically or by hand at the wish of the operator. The auto-
matic movement is accomplished by the following-described device:
" The horizontal disk and its engaged friction-wheel are driven continu-
ously by the pulley placed at the lower end of the vertical shaft and belted
to overhead shafting. This friction-wheel is feathered to a sleeve that runs
446 THE MATERIALS OF CONSTRUCTION.
loose on its shaft and carries a coned clutch that is nominally disengaged
from its cone, which is also feathered to the shaft, and can be moved
slightly longitudinally on the shaft into contact with the clutch by the
action of the vertical lever.
" When the needle-beam rises, it makes contact through a vertical pin in
the top of the frame, which completes an electric circuit and sends a cur-
rent through the electromagnet and causes it to attract its armature at the
lower end of the vertical lever, which moving to the right engages the fric-
tion-clutch and causes the shaft to revolve. This shaft operates the
sprocket-wheel and chain which draw out the poise on the scale-bearn until
the needle-beam drops, breaking the electric circuit. Breaking the elec-
tric circuit releases the armature and allows the friction-clutch to disen-
gage, and the poise comes to rest. The friction-wheel may be set at a
greater or less distance from the centre of the disk by turning the capstan-
head nut, and the chain is overhauled faster or slower, causing the poise to
move accordingly. If desired, the poise may be operated by the hand-
wheel without interfering with the automatic device other than cutting out
the circuit. The chain is attached to the poise in line with the three
knife-edges of the scale-beam; hence the tension in the chain has no ten-
dency to lift up or pull down the poise. This point is often overlooked
in designing this detail, not only in cement machines but in testing-ma-
chines in general. The writer [Prof Porter] has a cement machine in
which the error due to this cause is over 15 Ibs.
" This machine as described has been in almost constant use for eigh-
teen months and has given entire satisfaction. The operator has simply to
place the briquette in the clips, open the supply- valve, wait until the bri-
quette breaks, and then note the reading on the scale-beam. The objection
to this machine is the space it occupies, requiring a floor-area of 7 X 2 ft.,
and the necessity of a constant head of water."
325. Importance of an Exact Central Position in the Clips. It was shown
in Art. 26 that if h width of specimen and a eccentricity of load-
ing, the percentage of increase in the stress from this cause is given by the
fraction -^. Thus if a cement briquette 1 inch thick be placed in the clips
0.01 inch out of centre, its strength will be reduced by 6 per cent. This
assumes perfect freedom of motion of the clips at the surfaces of contact,
which they do not have. Experiments made at the Massachusetts Institute of
Technology * have shown that a displacement of T V inch decreased the ten-
sile strength by from 15 to 20 per cent (see Fig. 371). V y
326. Compression Tests of Cement have not been common in America,
though long practised in Europe. The excellent relation indicated in Fig.
337, p. 419, between the tensile and the compressive strength of Portland-
* Trans. Am. Soc. Mech. Engrs., vol. ix. p. 181.
TESTING OF CEMENT.
447
cement mortar (1 C. to 3 S.) would seem to show that both tensile and com-
pressive tests are not required, and American engineers have always acted
on this assumption.
The French Commission recommend compression tests, however, in addi-
tion to the tension tests, but they do not advise the making of separate test-
specimens. With the form of briquette shown in Figs. 355 and 356 the line
of rupture is definitely fixed (with very few breaks outside the grooved sec-
tion), and hence the two halves of the broken briquette will be nearly equal to
each other and to all other broken parts. These ends are then to be tested
FIG. 371. Showing Effect of Eccentric Position of Briquette in Clips. (Assoc. Eng. Soc.,
vol. vn. p. 207).
by crushing, the force to be applied normally to its bed, and the sum of the test
loads on the two ends of one briquette to be the crushing strength of that
specimen. In the absence of such broken briquettes to serve for this test,
cylinders of the same area and height are to be made up anil tested.
Since this height is but 22 mm., while the diameter of the equal cylinder
is 45 mm. (Fig. 355), the specimen has a height of but one half its lateral
dimension, and hence the compressive strength of such a form of specimen
is 20 per cent greater than that of a cubical form as shown in Art. 22, Fig. 17.
The broken briquettes are chosen to avoid making up additional specimens,
and also because this insures identical material for both the tension and
the compression tests. For instituting a comparison with the compressive
tests on otner material, where the cubical form has been almost universally
used, the correction coefficient of 0.83 can be employed, as stated above,
or else cubical specimens can be prepared and tested.
It ^s of course necessary to prepare all compress! ve-test specimens with
care, by reducing the two bearing-surfaces to true planes. It would also be
wise to provide a universal joint back of one of the bearing-plates.*
* See method employed at the Massachusetts Institute of Technology, Am. Soc.
Mech. Engrs., vol. ix. p. 172.
448
THE MATERIALS OF CONSTRUCTION.
In America compression tests of cement are made on the universal test-
ing-machines so common in this country. In Europe many special machines
are made for this purpose, one of the most recent of which is shown in Fig.
372. Here the load is indicated by the position of the radial arm moving
FIG. 372. Machine for Making Tests of Cement in Compression. (Manufactured by
Amsler-Laffon & Son, Scbnffbausen, Switzerland.)
over the graduated arc, the actual movement of the upper head being thus
multiplied a known number of times. As this movement is resisted by a
powerful helical spring, when this spring has been standardized its compres-
sion is a true index of the load.
327. Cross-bending Tests of Cement have been advocated occasionally,
but they have not come into genera] nee anywhere. The French Commission
TESTING OF CEMENT. 449
have also undertaken to standardize this test. They recommend a specimen
5 inches (120 mm ) long and 0.8 inch (20 mm.) square in cross-section,
and they show how this specimen may be broken on the Michaelis machine,
Fig. 3G4, by attaching the centre-bearing, upward-pulling stirrup to the
small hook at the left end of the lower lever.
M. Du rand -Clave has shown by very extended series of tests in tension
and in cross-bending, on identical samples of neat Portland cement, that
the average ratio of the modulus of rupture in cross-bending to the tensile
strength, as determined upon standard forms of briquettes, is 1.92 at 7 days
and 1.86 for 28 days, or an average of 1.89.* This relation was found to
subsist between averages made up from the means of the three tests in each
set of six, in both tension and cross-bending. The mean error of a single
test at 28 days was found to be 2.10 per cent for the tension tests and 2.13
per cent for the tests in cross-bending, thus showing that the two methods of
testing were equally accordant.
It would seem, therefore, that tests in cross-bending may be employed
with assurance as a means of determining both relative and absolute values
of cements and cement-mortars, their principal disadvantage lying in the fact
that there are few records extant with which tc compare the results of such
tests.
The principal recommendation for the use of transverse tests would seem
to lie in the economy of a testing outfit. It has been estimated that a
suitable machine for testing cement transversely could be constructed for
about $12, while a set of moulds for sixteen prisms would cost not to exceed
$3, or if these latter be made of cast iron the cost need not exceed $5 per
set of twelve after the patterns are made.f
It is further claimed that since all transverse breaks are fair, while with
the forms of briquettes and clips hitherto used in America nearly fifty per
cent of the breaks occur outside of the minimum section, the results of
transverse tests must be more reliable. If, however, a form of briquette
and clip can be devised which will always give fair breaks, this claim of
advantage will no longer stand. There seems to be now in this country
no inclination to change from tension to transverse tests of cement.
328. Standard Tests to Determine the Adhesion of Cement-mortars to
Various Substances. While the tensile strength of briquettes shows the
cohesion of the mortar, it has been found by experiment that its adhesion
either to other mortars or to the same mixture which has already hardened,
or to brick or stone or metal, is very much less than its cohesion. It is
important, therefore, to have a standard test of adhesion, as well as of
* Messrs. Abbott and Morrison, in their thesis published in Engineering News, Dec.
14, 1893, show that for neat cement this ratio was 1.8 on prisms one inch square and
broken on a span of four inches.
f See Engineering News, vol. xxx. p. 469, where complete detail drawings are
given of both the machine and of the moulds.
450
THE MATERIALS OF CONSTRUCTION.
strength. Because tests of this kind are comparatively new, no general
custom has been established in America on the subject; but the following
recommendations have been made by the French Commission:
(1) For tests of adhesion of cements and cement-mortars use will be
made of a special form of briquette, moulded in two parts, these two parts
consisting of the two materials whose adhe-
sion is to be tested, provided both cai\ be
moulded, or containing between them a
prism of the solid body to which the adhe-
sion of the mortar is to be determined The
form of this briquette, as modified for Eng-
lish units, with one square inch of area on
the surface of adhesion, is shown in Fig.
373. This mould is formed in two parts,
and is used to form in succession the two
halves of the complete briquette.
(2) To compare the force of adhesion of
different cements to a given material, nor-
mal adhesion-Nocks will be prepared as fol-
lows: Use for these always one kind *of
standard Portland cement which has passed
a sieve of eighty meshes to the linear inch,
mixed with the standard sand No. 3 (see
FIG. 373. Form of Briquette for . , oim . ,, ,. ,, ,,
A ,, m , Art. 317) m the proportion 01 one 01 cement
Adhesion Test of Cement as
- ,1
\
C"
-/" -^
/ ^
adopted by the French Commis-
to two of sand - These normal adhesion-
sion and adapted to English blocks will be moulded in the form of one
Units. half of the briquette shown in Fig. 373.*
It will be gauged with 9$ of water and rammed into the mould. At the
end of 2-i hours in air it will be placed in fresh water for a period of at
least twenty-eight days. When it is to be used, it will first be dried and its
adhesion-surface polished with emery-paper.
(3) The cement to be tested for adhesion with these standard blocks
prepared as above will be mixed as a normal plastic mortar, one of cement
to three of sand (see Art. 316), which will be introduced into the mould
with a trowel, this mould now being placed with a normal adhesion-block at
the bottom in place of the movable metallic disk. The mould will remain
upon this completed block until it is ready for testing, and the block will
be allowed to harden either in air or water, and for such period as the test
requires. It is recommended that the number of tests, the periods of time,
the methods of hardening, and the recording of the results should comply
with the conditions given for tension tests in Art. 319.
* The detail drawings of these moulds are given in the Fr. Com. Rep., vol. iv. p. 284.
TESTING OF CEMENT. 451
(4) To compare the force of adhesion of a given cement to different mate-
rials. For this purpose the test-specimens will be prepared as described
above, except that in place of the normal adhesion-blocks similar blocks of
the various materials to be tested will be prepared and allowed to harden,
provided these are such as can be moulded in this manner. If such mate-
rials are solid, small disks, about three-eighths of an inch thick, will be pre-
pared, and these will be used in the bottom of the mould in place of the
metallic disk, the adhesion-block to be completed by using neat Portland-
cement mortar. After this has hardened the briquette will be completed
by making the other half of a standard plastic mortar, one cement to three
of sand, using the particular kind of cement whose adhesion to these
various substances is to be tested.
If the normal plastic mortar is not used in adhesion tests, a full descrip-
tion of its composition should be indicated on the records.
These adhesion-briquettes to be broken on a standard tension-testing
machine, using the regular tension-clips.
329. Normal Variations in Volume of Cement-mortars in Air and in
Water. From elaborate tests on the swelling and shrinking of cement-mor-
tars hardening in air and under water made at the Massachusetts Institute
of Technology, Boston, and by Professor Bauschinger at Munich, it may be
stated :
1. Cement-mortar hardening in air shrinks almost uniformly for a period
of more than three months, the linear shrinkage in that time being, for neat
cement, from 0.12 to 0.3-1 of one per cent, and for cement-mortar, one of
cement to one of sand, from 0.08 to 0.17 of one per cent. The change in
volume is of course three times the above percentages.
2. Cement-mortars hardening under water increase in linear dimen-
sions from 0.04 to 0.25 of one per cent in three months for neat cement, and
from 0.00 to 0.08 of one per cent for a mortar composed of one part cement
to one of sand; the volumetric expansion being three times these
amounts.
Professor Bauschinger found for nine German Portland cements, neat,
an expansion, when hardened under water, of 0.05 of one per cent in sixteen
weeks; Mr. Grant found for English Portland cement an expansion of 0.08
of one per cent in three months, this latter figure agreeing with the average
results of the tests made on eight kinds of Portland cement at the Massa-
chusetts Institute.*
For mortars composed of one part of cement to three of sand the varia-
tions in volume are very much less than those given for mortars of equal
parts of sand and cement.
* See progress reports of committee of the Am. Soc. C. E. on the compressive strength
of cements, vol. xvn. p. 215, and vol. xvni. p. 264.
452 THE MATERIALS OF CONSTRUCTION.
330. Recommendations of the French Commission for Testing Perma-
nency of Volume. Tests of permanency of volume may be of two general classes
cold and hot.
Cold Tests will be made upon thin cakes of neat-cement paste made up on glass,
about six inches in diameter and f inch thick at the centre, with thin edges, and
placed immediately in water or in air, along with the briquettes which are hardening
in these two media. These cakes to be examined at periods of 7 days, 28 days, 3
months, 6 months, 1 year, 2 years, etc., corresponding to the like periods for the tests
of strength.
To measure the amount of the linear change of volume of neat cement immersed
in cold water, a small cement form, 32 inches long and ^ inch square in section, may
be moulded and placed vertically in a glass tube 1 inch in diameter filled with water.
The expansion would be indicated by the movement of the long arm of a lever over
a graduated scale, which is actuated by a pin embedded in the upper end of the
specimen when made.*
Evidently it may require years to assure one of the permanency of volume, or
soundness, of a cement by the use of the cold-water and air test.
Hot Tests to be made on cylinders of neat cement 1J inches (30 mm.) high and
1 J inches in diameter, made up and left in metal moulds composed of sheet metal 0.02
inch (0.5 mm.) thick (No. 25 gauge). This mould to be entirely severed on one
>|
FIG. 374. Apparatus for Testing Permanency of Volume of Cement. (Recommended
by the French Commission.)
element, and to have soldered to it on the opposite side two arms, six inches long,
forming an angle with each other as shown in Fig. 374. The decreasing distance be-
tween the extremities of these arms to be a measure of the swelling of the cement.
These moulds to be immersed in cold water as soon as filled, and allowed to set
for 24 hours, or for a shorter period if it is a quick-setting cement. The mould will
then be placed on a grating in a vessel of water and its temperature raised to the
boiling-point in from 15 to 30 minutes. This temperature is to be maintained for
six hours, when the water will be allowed to cool down before removing the speci-
men for remeasuring the distance between the six-inch arms.
This hot test not to be applied to natural cements, or to any cement which sets
very rapidly.
The consistency of the cement used in both the hot and the cold tests to be of the
normal consistency described in Art. 316.
331. The Permeability of Cement-mortar is often a very important mat-
ter, as in the case of reservoir wells and linings, and often in foundation-
walls placed below the level of the ground-water. Neat cement-mortar
is absolutely impervious when it has hardened and has not cracked,
and so also is a mixture of one to one, or even of two of sand to one of
* See Fig. 25, p. 302, vol. i, Report of the French Commission, 1895.
TESTING OF CEMENT.
453
cement, by weight, if well mixed. The normal mixture of three of sand
to one of cement may also be made practically impervious with the most
thorough mixing of the dry ingredients
and a compacting of the mortar by hard
ramming.
Professor Tetmajer has used the appara-
tus shown in Fig. 375 to obtain a modulus of
permeability. Here a cylinder of the mortar
is made and allowed to harden under water
for a specified time. It is then mounted
in the apparatus by means of annular rubber-
cushion or packing disks, and the water let
on below under a known pressure. The
permeability of the mortar is indicated by
the rate at which the water passes the disk
and rises in the glass tube above, which
is graduated to cubic centimeters. The
author has also used this apparatus with
satisfactory results, a convenient pressure
to use being that of the city water-mains.
The French Commission
recommend a standard per-
meability test as follows:
(I) The permeability of
cement-mortars will be indi-
cated by the number of liters of water passing per hour
through a cubical block of 7 cm. (say 2| inches) on a side,
under the following conditions.
The water will be brought to the top face of the specimen,
laid edgewise (what was the horizontal plane in the forma-
tion of the cube now becoming a vertical plane), through a
glass tube, 35 mm. internal diameter and about 4 or 5 inches
high, which is sealed to the top face of the cube by neat
cement-mortar as shown in Fig. 376. A rubber tube con-
the upper end of the glass tube with the reservoir
FIG. 375. Tetmajer's Apparatus
for Testing the Permeability of
Cement - mortar. (Communica-
tions, vol. vi. 1
Fm. 376.
lestmg the
placed at a height (from the surface of the water of immer-
Permeability s * on ^ ^ ie sur f ace ^ the water in the reservoir) of 4 inches,
of Cement-mor- 40 inches, or 400 inches (0.1 m., 1.0 m., or 10.0 m.).
tar. (Recom- Before beginning the experiment the cube of mortar to
mended by the b e immersed in water for 48 hours, and during the test the
nch ^ om- block is to remain immersed to prevent the formation of an
mission.) . . ,. , , , . ,
impervious coating on the outside irom the evaporation of
the exuding water.
The volume of water passing will be given for the standard periods of
454 THE MATERIALS OF CONSTRUCTION. '
24 hours, 7 days, 28 days, and 3 months. For very porous mortar a
shorter period than 24 hours may be employed, and at the same time the
head of water used must be stated.
Tests will be made on three similar specimens, the mean of the two most
accordant results to be used.
(2) The normal test of permeability will be made on cubes made up of
normal plastic mortar (3 sand to 1 cement, by weight) as described in Art.
319, and the specimen cubes must harden in water under the normal con-
ditions for 28 days before testing.
For tests on other mixtures, and for other periods of hardening, they
recommend that mixtures of 2 sand to 1 cement, and 5 sand to 1 cement, by
weight, and hardening periods of 7 days, 28 days, and 3 months be chosen.
In all cases the composition, age, and conditions of hardening must be
stated, as well as the amount of water passed and the pressure-head used.
332. Tests for the Decomposing Action of Sea-water. As a result of the
porosity of cement-mortars and concretes and of the resulting action of sea-
water on the interior of the mass, producing therein certain changes partly
by solution and partly by the formation of new chemical compounds, cement-
mortars and concretes are often disintegrated when subjected to the action of
sea-water. The French Commission have carefully studied this question,
and have recommended the following test, which they regard as of value in
determining the comparative resistance of cement to this action:
(1) Standard tension briquettes of normal plastic mortar (one cement to three of
sand) will be made, and after 24 hours in air will be placed in sea-water which is to
be renewed every two days during the first week', and every week thereafter. During
the first week the volume of this sea-water to be at least four times that of the bri-
quettes immersed in it.
An equal number of duplicate briquettes to be exposed in a similar manner to the
action of fresh water. Tension tests on these duplicate briquettes to be made at the
standard periods of 28 days, 3 months, 6 months, 1 year, etc., and the effect of the
sea-water to be shown by a comparison of the results.
(2) Filtration tests will be made on specimens having a cubical form, as described
in Art. 331, and to be exposed to the action of sea-water both in the bath and in the
filtration reservoir there described. The head will be 4 inches, 40 inches, or 400
inches, according to the permeability of the specimen. Two sets of duplicate test-
specimens will be subjected to this test, one having hardened in sea-water and the
other having hardened in air for the several standard periods chosen.
A third duplicate set of exactly similar cubical blocks to be preserved and hard-
ened in fresh water for the same periods of time.
In the absence of actual sea- water, artificial sea- water will be prepared with the
following formula:
Chloride of sodium (NaCl) 30 g.
Sulphate of magnesia, crystallized (MgOSO 3 ,7H 2 O) 5
Chloride of magnesium, crystallized (MgCl,6H 2 O) 6
Sulphate of lime, hydrated (CaOSO 3 ,2H 2 O) 1.5
Bicarbonate of potassium (KOH 2 O,2CO 3 ) 0.2
Distilled water 1000
All the above cubes to be superficially examined and tested in compression at the
standard periods chosen.
The following observations will be taken:
TESTING OF CEMENT. 455
(a) The comparative appearance of the specimen subjected to the several kinds
of treatment.
(6) The tensile strength of the two sets of briquettes which had hardened in
salt and in fresh water.
(c) The compressive strength of the three sets of cubical blocks which had hard-
ened in salt water and in air, and which had been subjected to the filtration tests,
and the blocks which had hardened in fresh water.
(d) Chemical composition of the cubical blocks subjected to the several kinds of
treatment.
For other compositions, mortars composed of one cement to two of sand, and one
cement to five of sand, to be chosen and tested at the standard periods of 7 days, 28
days, 3 months, etc.
CHAPTER XXII.
TESTS OF THE STRENGTH OF STONE AND BRICK.
TESTS OF STONE.
333. Tests of the Strength of Stone Limited to the Crushing Test.
Since stone can readily be prepared for crushing tests, these have been
almost exclusively employed in determining its strength. In view of the data
obtained from comparative tests of cement in tension and compression shown
in Fig. 337, p. 419, it might be inferred that the crushing test would show
also the relative strength in tension. Since failure in crushing is a failure
by shearing, it might be supposed that the true relative shearing strength
would also be shown by the compression test. When stone fails in cross-
bending it breaks first on the tension side of the beam, and hence this is a
failure in tension, and therefore the crushing test has been thought to give
correct relative values of cross-breaking strength.
These assumptions prove not to be correct, however, as has been very
conclusively shown by Bauschinger in volumes x, xvn, and xix of his
Communications, where the results of tests on more than a thousand speci-
mens of building-stones of the various kinds found in Bavaria are given.
These tests were made in compression, in tension, in cross-bending, and in
shearing, and no fixed relation can be given to these several kinds of
strength. Probably this is largely due to the fact that stones are not amor-
phous bodies, but are usually either sedimentary or crystalline or both, with
definite planes of cleavage and of weakness. Since stone is used, however,
almost exclusively in compression, it is usually considered sufficient to test
its strength in compression only.
The conditions to be fulfilled in the crushing test of stone are sufficiently
elucidated in Chapters III and XVI. While the test-specimens should have
heights greater than their least lateral dimensions, yet in order to make the
results comparable with those hitherto obtained and recorded it is necessary
to continue to make these tests on cubical forms.
TESTS OF PAVING-BBICK.
334. Kinds of Tests Required. The use of brick for the wearing sur-
face of street -pavements is now so universal that this new product, " vitri-
456
TESTS OF THE STRENGTH OF STONE AND BRICK. 457
fied paving-brick," has become one of the most important of the materials
employed by the civil engineer. Since appearances in this material are
entirely untrustworthy, and since these products vary greatly not only as
between the output of different manufacturers, but also as between different
kilns of the same factory or even in different parts of the same kiln, a
thorough system of mechanical tests to determine the probable wearing
qualities is absolutely essential. To develop these qualities four tests are
now commonly accepted as essential, namely * :
1. Cross-breaking.
2. Crushing.
3. Impact (the rattler test).
4. Absorption.
335. The Cross-breaking Test. This is made on single whole bricks by
setting them edgewise on two rounded knife-edge bearings about 7 inches
apart and loading them at the centre. In order to insure a true bearing of
the knife-edges the brick should be ground to true parallel surfaces, or else
the lower bearings should be rounded longitudinally sufficiently to prevent
a twisting or torsional action. The cross-breaking modulus of rupture is
found by applying the formula
3TP7
336. The Crushing Test is usually made on a half -brick, set edgewise,
and one or both of the ends of the brick previously used in the cross-break-
ing test may be used. As these faces are very rough in paving-brick (from
their having been reduced to a semi-plastic condition in the kilns, these
being the bearing-surfaces), it is impossible to make fair crushing tests on
these forms without grinding them to true parallel planes. This can readily
be done on a regular stone- or marble-grinding table operated by steam-
power, such being available in all large cities. When this is done they may
be bedded on single thicknesses of tar-board, or placed directly between
steel plates in the testing-machine. Great care must be exercised to place
the specimen centrally in the machine, and to see that the bearing-plates
fit evenly upon the specimen. One of these plates should have a spherical
base to make it adjustable and so insure an even and true bearing on the
test-specimen. The specimen should fail all at once, with a loud report,
with little or no previous spalling. The load must be increased very slowly
and uniformly, and the weighing-beam automatically balanced if practicable.
337. The Rattler Test. Formerly this test was made as an abrasion
test by using a great quantity of small castings. The author has long in-
sisted that this test should partake of the character of an impact test, and
this view now generally prevails. Paving-brick are broken to pieces in
service rather than worn or ground down, and the property of resilience
is the one sought rather than hardness. The standardizing of this test
* See Recommendations of a Committee of the National Association of Brick Maun-
facturers, in Art. 338.
458
THE MATERIALS OF CONSTRUCTION.
has proved a difficult task, but it has been fairly accomplished by Mr.
F. F. Harrington, a former student of the author's, now in charge of the
Testing Laboratory of the Board of Public Improvements of the City of
St. Louis. Mr. Harrington's rattler is a cast-iron barrel, polygonal in form
and having fifteen staves, similar to that shown in Fig. 377. Its length
is 42 inches and its diameter is 24 inches, and it revolves on trunnions at
FIG. 377. Rattler for Testing Paving-brick.
'flftfl
S% /
^ i A/ G T tf
jffffli
K wre/#Tj0f #f m f/M0fMfti wwt
of
7T//t//V/T/?/M=/S%
4 /? ft Z
/2 ?/ 30 33
FIG. 381. Showing Effect of Length of Ban-el in the Rattler Test of Paving-brick.
(Harrington.)
ience, therefore, the rattler should be made double, or so as to contain two
apartments having the dimensions described above.*
The first effect observed in the rattler test is one of chipping on all
edges. After these have rounded off somewhat the effect is more evenly
distributed over the outer surface, but it still remains principally an impact
action. The dust and small pieces fall through the spaces left between
staves for this purpose, so that the rattler remains comparatively clear of
debris. If absorption-tests are made, they should be made on bricks which
have passed the rattler test, because their glazed surfaces are then broken,
if not largely removed.
Accordant results could not be obtained when different kinds of brick
were put in the rattler at one time or when other materials, such as stone,
granite, or cast-iron blocks, were employed; hence the requirement that
the full complement of material in the rattler should consist of the brick
tested.
No specified requirements were named by the committee, as sufficient
data had not yet been obtained.
* Standard drawings and patterns for such a rattler have been prepared by Mr. M.
L. Hoi man, Water Commissioner of St. Louis, from which the author has had a
machine constructed.
CHAPTER XXIIL
TESTS OF THE STRENGTH OF TIMBER.
339. The Variable Strength of Timber. As shown in Chapter XIII,
sound timber of a given species varies in its strength from two general
causes, its structure and its moisture condition. Neither of these sources
of strength (or weakness) has hitherto received proper study and analysis,
and hence the known variations in the strength of timber has been attrib-
uted either to its inherent and undiscoverable variations, or to variations
iii the size of the sticks tested. Some of the most important conclusions
to be drawn from the U. S. Timber Tests are:
1. The strength of timber is about twice as great ivhen it is dry as when
it is green or wet.*
2. The strength of a given species of timber at a given percentage of
moisture is governed by the ratio of the summer (solid) to the spring (open)
wood, or in other words by its specific gravity or solidity.
3. The strength per square inch of a large stick in every kind of test is
fully equal to that of a smaller stick cut from it ivhen both are similarly
proportioned and similarly free from faults.
4. It is very highly probable that the strength of all kinds of wood-fibre
of like structural arrangement (thus putting the oaks into a separate class)
increases directly with the specific gravity (or weight per cubic foot) of the
dry wood. (See a discussion of this subject in Chapter XXXII.)
The moisture state is the great and governing cause of variation in
strength. When reduced to the same moisture condition it may be said, as
a result of about 40,000 tests of timber made by the author, that in crushing
endwise 90 per cent of all tests fall within 25 per cent of the mean, and
55 per cent of all tests fall within 10 per cent of the mean-value for that
species, and this is about as much as can be said for other kinds of building
materials when all have been subjected to a reasonable inspection.
340. " The United States Timber Tests," f so called, were inaugurated in
1891 by Dr. B. E. Fernow, Chief of the Forestry Division of the U. S.
Agricultural Department, and have been carried on with frequent inter-
ruptions ever since. They consist of a very complete series of investiga-
* Sec the curve showing variation of strength with moisture in Chapter XXXII.
f See Bulletins 6, 8, 10, 13, and others to be issued from time to time by the U. S.
Agricultural Department, Forestry Division, and to be had on application to the chief of
that division.
462
TESTS OF THE STRENGTH OF TIMBER. 463
tions into the habitat, conditions, and laws of growth, structure, strength,
and other properties, seasoning, preservation, and decay, and finally the arti-
ficial cultivation of the useful timbers of the United States. The great im-
portance of the timber industry (being second only to that of agriculture in
this country), and the universal absence of accurate (scientific) knowledge
on these subjects, have seemed to warrant the undertaking and prosecution
of this the greatest series of physical investigations ever carried out. The
field-studies and the collection of the material have been done by Dr.
Charles Mohr; the structural investigations have been made by Mr. Filibert
Eoth in Washington; and the mechanical tests have been made under the
direction of the author in his testing laboratory at Washington University,
St. Louis, Mo. The species examined and tested to date (December, 1896)
are given in tabular form in Chapter XXXII. It there appears that there
have been selected for these tests
1. Sixty-eight trees of Long-leaf Pine * from South Carolina, Alabama,
Mississippi, Louisiana, and Texas.
2. Twelve trees of Cuban Pine from South Carolina, Georgia, and Ala-
bama.
3. Twenty-two trees of Short-leaf Pine from Alabama, Missouri, Arkan-
sas, and Texas.
4. Thirty-two trees of Loblolly Pine from South Carolina, Georgia, Ala-
bama, and Arkansas.
5. Seventeen trees of White Pine from Michigan and Wisconsin.
6. Eight trees of Red (Norway) Pine from Michigan and Wisconsin.
7. Four trees of Spruce Pine from Alabama.
8. Twenty trees of Bald Cypress from South Carolina, Mississippi, and
Louisiana.
9. Four trees of White Cedar from Mississippi.
10. The test specimens of Douglas Spruce were not taken from selected
trees, but were obtained from lumber shipments to the St. Louis markets.
11-20. Eighty-three trees of ten species of Oak from Alabama, Missis-
sippi, and Arkansas.
21-27. Twenty-four trees of seven species of Hickory from Mississippi.
28-29 Five trees of two species of. Elm from Mississippi and Arkansas.
30-31. Four trees of two species of Ash from Mississippi.
32. Seven trees of Sweet Gum from Mississippi and Arkansas.
Besides these a great many small trees were taken, from which disks were
cut at frequent intervals, from butt to top, and sent to Washington for the
physical and structural studies Complete field notes were taken, also, of
the geographical position, the immediate surroundings as to forest growth,
character of soil, moisture conditions, etc. The diameter of the stump, the
age of the tree, and the distance to the first limb were also noted.
* Sixteen of these were " bled " trees to determine the effects of " boxing."
464
THE MATERIALS OF CONSTRUCTION.
All this material has come from the Southern States except the white and
Norway pines. The trees have been selected and cut by Dr. Mohr; the logs
cut from them were shipped to the author at St. Louis in car-load lots.
Disks 8 inches long have been taken from all these trees, at a number of
heights, and also from many trees of the same species too small for timber-
test specimens, and sent to Mr. Roth at Washington. The logs have
been cut into test-timbers in various ways as shown in Fig. 352. The
largest forms are full-sized beams (from the 18-foot logs) and columns (from
the 12-foot logs), the size depending on the size of the log. The intermedi-
ate forms are 4 inches square, the standard size of the " small " sticks. The
No. 1.
No. 2.
No. 3.
No. 4. No. 5. No. 6.
FIG. 382. Showing Methods of Sawing U. S. Timber Test Logs.
smallest size is 2 inches square, which has been employed only in "special
investigations." The form in No. 6 indicates that two or three large
sticks are to be cut and tested to failure, from the uninjured portions of
which are afterwards cut smaller sticks which are also tested for the pur-
pose of comparing the strength of large and small sizes. Form No. 5 has
been adopted as the standard method of cutting when only 4-inch
sticks are taken. Only logs over 24 inches in diameter at the small end.
could furnish this entire system of sticks, smaller logs giving the five interior
ones only. The logs are always laid out on their upper ends, taking the
pith as the centre of the diagram regardless of how unsymmetrical this may
TESTS OF THE STRENGTH OF TIMBER. 465
lie in the cross-section of the log. The logs are always 12 or 18 feet in
length, and the 4-inch test-sticks are cut to 6-foot lengths, thus getting
two or three such lengths from each 4-inch stick shown in the log diagram.
341. The Mechanical Tests. As a rule the following tests have been
applied to every 4-inch stick:
1. Cross-bending.
2. Crushing endwise of the grain.
3. Crushing across the grain.
4. Shearing along the grain.
5. Tension.
Since June 1895 no tests have been made in tension, as it was thought
this kind of strength was so great as to remove it from the category of pos-
sible methods of failure. It was thought timber would never fail in pure
tension in practice.
For each and every test a section of the stick about T 3 F inch thick is cut
from near the point of failure, and used for determining the percentage of
moisture as described in Art. 343.
The test-sticks were subject to a system of inspection and rejection
which it was thought would correspond to such a system in actual practice
where the timber was to be used in structures where the parts are propor-
tioned to their loads, and hence it is thought the average of all the results
fairly corresponds to such an average strength in practice at similar stages
of dry ness.
In order to make the results comparable it was of course necessary to
reduce them all to equivalent values at a standard percentage of moisture.
For all reductions made previous to May 1896 this standard had been 15 per
cent moisture, computed on the dry weight. After that date 12 percent was
chosen as better representing the condition of the roughly seasoned timber,
whether in or out of doors. In a dry, heated building the moisture falls as
low as 8 or 10 per cent.
342. The Cross-bending Test on the 4-inch sticks is made on a small
8000-lb. testing-machine designed by the author, shown in Fig. 301, page
370, while the large beams are tested on the 100,000-lb. machine shown in
Fig. 302, page 371. In both cases the loads are applied so as to produce a
uniform rate of deflection (with the 4-inch sticks it is always at the rate of
1 inch per minute, while with the larger sticks it is | inch per minute) in
order to eliminate the time-effect, which is very large with timber especially
under the higher loads.
With the small sticks two central bearing-points are used, 12 inches
apart, thus putting that length of stick under the maximum bending-stress,
and so really testing 12 inches in length of the stick instead of about 1 or
2 inches with a single bearing.* Of course in all cases the bearings are
* This was done as the effect of a paper by Prof. J. Burkett Webb before the Am.
Assoc. Adv. Science, Section D, at Rochester, N. Y., 1892.
466 THE MATERIALS OF CONSTRUCTION.
spread over a considerable area by means of steel plates to prevent the de-
struction of the fibres by crushing across the grain. With the large beams,
an oak saddle some 30 inches long, of the full width of the beam, and
rounded slightly on the bottom in a longitudinal direction, is used for the
centre bearing under the knife-edge of the machine.
The deflections of the small beams are measured by means of a microm-
eter-screw bearing on the head of the power-screw, as shown in Fig. 301,
while in the case of the large beams a thread was stretched (by a rubber band)
along one or both sides from nails in the neutral plane above the end bear-
ings, and readings taken on a scale tacked to the t}eam at the centre. The
scale was nickel-plated and kept polished to act as a mirror, and the par-
allax of the thread on the scale was obviated by Bringing the thread and its
image into coincidence when the readings were taken. The readings were
made to 0.001 inch with the small beams and 'to 0.01 inch with the large
beams.
The formulas of reduction for strength, modulus of elasticity, and resil-
ience were adapted to the particular method of test employed; but since the
resilience in inch-pounds per cubic inch is different for the two cases (single
and double bearings at centre), the larger results* obtained with a double
bearing have been reduced to their equivalent for a* single bearing to make
them all comparable with each other and with the results usually obtained,
which would be with a single bearing. The resilience of a beam loaded at
the centre, in inch-pounds per cubic inch, is from eq. 6, p. 84.
_L /!
" 18 E
This is also the measure of the resilience of the outer ends of the beams
tested with double bearings at the centre, while for the part between the two
centre bearings, where the bending moment is uniform, it is, from eq. 10,
p. 85.
!_/!
" 12 E'
The / being the same in the two cases, namely, the " apparent elastic
limit" of the material, found by fixing a point of the bending-stress diagram
where the rate of deformation is 50^ greater than at the origin, as explained
in Art. 13, p. 18.
The results obtained from each cross-bending test of timber, therefore,
are:
1. Modulus of strength at the "apparent elastic limit."
2. Modulus of strength at rupture.
3. Modulus of elasticity.
4. Modulus of resilience, or springiness.
When large beam shave been tested to failure, two smaller (4-inch) beams
6 feet long have been cut from the upper side of the larger beam at one end,
TESTS OF THE STRENGTH OF TIMBER.
467
and two from the lower side at the other end, and these four small beams
have been subjected to the same test to discover whether or not the ordinary
formulae are correct, or, what is the same thing, to discover whether the same
values of the moduli named above would be obtained from both sizes. All
the tests which have been made of this kind go to show that when both sizes
are equally free from faults this is true. This proves that the strength of
large sizes may safely be computed from tests on smaller sizes, other things
being equal.
343. The Crushing-endwise Test. For this test a section about 8 inches
long of the uninjured portion of a 4-inch beam which had been tested in
cross-bending, is taken by means of a circular cutting-off saw, and tested to-
failure in compression endwise. The stress-diagrams in this test are fairly
indicated in Fig. 383. Failure occurs by a buckling down of the fibres as
shown at B, Fig. 113, page 241. After this buckling action of the fibres
across the entire section has occurred the strength of the specimen is only
about 0.8 what it had been originally, as is shown in Figs. 29 and 383,
o a/0 0.20 a<3# &f0 aw 0.70
FIG. 383. Typical Stress-diagrams of Timber when subjected to Compression Endwise.
this residual strength remaining about constant for largely increasing
deformations.
This is the most valuable and characteristic single test to which timber
can be subjected. It is the only one in which a relatively large stick can be
evenly and simultaneously tested to failure throughout its entire cross-
section. It should be expected, therefore, to give more uniform results than
468 THE MATERIALS OF CONSTRUCTION.
any other, and such proves to be the case. It is also the simplest and easiest
to make. For commercial purposes, therefore, this test alone would serve
nearly every purpose, all the other various kinds of strength and stiffness
being inferred from this one test.
344. Crushing Across the Grain. Since timber is very weak in crush-
ing across the grain, as compared to crushing endwise, this is found to be
one of the most common methods of failure in practice. It is common to
rest a timber column on a sill of the same wood, and to design the column
for its maximum working load, paying no attention to the utter inability of
the sill to carry this load without crushing. Many failures of timber
structures are due to this cause alone.
As there is no definite point of failure in crushing across the grain, two
limits of deformation have been arbitrarily chosen at which the load has
been recorded, namely, at three per cent compression, as a working limit
allowable, and at fifteen per cent compression, as an extreme limit, or as
failure. The apparatus used to indicate these two limits, for heights (thick-
nesses) of specimen from 2 inches to 4 inches, is shown in Fig. 292 and ex-
plained in the text of Art. 282, p. 357. With such timber as oak, which
has large medullary or pith rays, the crushing strength in a radial direction
is greater than in a tangential direction.
345. The Shearing Test. This is intended to develop the strength of
timber to resist shearing along the grain. This strength is very small in
nearly all kinds of wood, and may be reduced almost to zero by seasoning
checks. It is a very common method of failure in timber framework, and
hence it is important to test for it. The apparatus used is illustrated and
described in Art. 300, p. 386.
346. The Tension Test. The tensile strength of timber is so great (often
over 30,000 Ibs. per square inch) that it is difficult to make a fair test of
timber in this way. Simple shouldering is out of the question, since the
specimen shears out or the shoulders crush down. The author, after trying
various methods, adopted the simple forms of specimens shown in Fig.
113, p. 241. This figure does not show the reduction of the cross-section
at the centre, which was done by cutting out two segments of circles on the
two sides by a band-saw, these segments having about an 18-inch radius.
The reduced section left at the centre of the specimen was about 3 inches
by f inch, making something over a square inch of net section. These
specimens were then gripped by flat, grooved, cast-iron wedges, in the
100,000-lb. universal testing-machine, and pulled to failure. Of course
nearly straight-grained timber must be used or the failure is partly or
wholly one of shearing. This test was finally abandoned altogether, as it
was assumed that timber would never fail in practice in this way.
PART IV.
THE MECHANICAL PROPERTIES OF THE MATERIALS OF
CONSTRUCTION AS REVEALED BY ACTUAL TESTS.
CHAPTER XXIV.
THE STRENGTH OF CAST IRON.
347. The Tensile Strength of Cast Iron varies from 15,000 to 35,000 Ibs.
per square inch, while ordinary foundry irons run from 18,000 to 22,000'
Ibs.* This strength depends greatly on the size of the specimen as well as
on the composition, and on its freedom from internal stress from a too
rapid cooling.
The general characteristics of cast iron when tested in tension are
shown in Figs. 384 and 385. It will be seen that there is here no well-
FIG. 384. Two Stress-diagrams of Cast Iron in Tension, each the average of eleven tests.
Average tensile strength = 33,500 Ibs. per square inch. (Wat. Ars. Rep. 1894.)
defined " elastic limit," and if there be such a point it is very low in com-
parison with the ultimate strength of the iron. The "apparent elastic
limit " falls at about 15,000 Ibs. per square inch (Fig. 385), or at about
* When annealed in the malleable process its strength is raised to from 30,000 to
50,000 Ibs. per square inch, as shown in Chapter VII, p. 115.
469
470
THE MATERIALS OF CONSTRUCTION.
of the ultimate strength, as is found to be the case with wrought iron and
rolled steel. The permanent set at this point, while it looks large in Fig.
Or
CQ
PWP0A
///7Z
o
FIG. 385. Typical Stress-diagram of Cast Iron in Tension, with Location of the "Ap-
parent Elastic Limit," corresponding to a permanent set of 0.0001 of the length.
(Wat. Ars. Tests, 1892.)
385, is in reality only about T ^ of one per cent, or entirely inappreciable.
It is safe to specify 25,000. Ibs. tensile
strength, if great strength is required.
Mr. AY. J. Keep has shown (Fig. 57, p. 96)
that the cross -breaking strength is very
largely a function of the size of the test-
specimen. As tension-test specimens are
usually cast about 1 in. to 1^ in. in diameter,
this variation with size is not so important
for tension-test purposes as it might appear
from this diagram. The test-specimens are
usually turned down, both at the gripped
ends and on the reduced portion. Fig. 386
shows a form of cast-iron test -specimen
which is intended to dispense with turning
altogether; but even here it would be better
to turn the gripped ends, to avoid all bend-
ing stresses in the grips.
FIG. 386. Form of Cast-iron
Specimen \vhich does not re-
quire turning down.
THE STRENGTH OF CAS? IRON.
471
TABLE XXIII. COMPOSITION AND STRENGTH OF HIGH-GRADE CAST IRONS
MADE AT THE FOUNDRY AT THE U. S. ARSENAL AT WATERTOWN, MASS.
TEST-SPECIMENS GROOVED. (Rep. 1894, p. 247.)
Composition of Charge.
Kind of Furnace.
Carbon.
Manganese.
Silicon.
Sulphur.
Phosphorus.
!i
go
a
d
'&
A
03
O
Combined.
Tensile Stre
per square
Muirkirk pig 35. 3 "1
cupola
do
do
do
do
air-furnace
cupola
do
air-furnace
cuoola
2.440
2.391
2.487
3.558
2.279
2.492
2.393
2.727
2.058
2.255
0.900
0.960
0.744
0.608
0.366
0.739
0.432
0.299
0.778
0.731
0.335
0.342
0.461
0.451
0.353
0.448
0.450
0.462
0.464
0.458
1.137
1.081
1.511
1.212
1.024
1.231
1.090
1.363
1.560
1.297
0.113
0.134
0.118
0.125
0.118
0.125
0.140
0.125
0.115
0.114
0.572
0.505
0.521
0.655
0.496
0.816
0.497
0.477
0.619
0.491
27,700
27,990
31,980
32,400
34,453
32,980
31,110
31,810
29,100
30,750
16.07
15.20
17.35
15.83
20.47
18.09
Old 8-inch shell 29.4 |
Heads 29.4 !
Scrap . 5 9 |
100 J
Muirkirk pig 35 31
Shell 29 4
Heads 29.4 !
Scrap 5. 9 f
100" j
Richmond pig No. 1 10 1
Richmond pig No. 2 10
Salisbury pig No. 4, high 15 V
Scrap 50
100 J
Salisbury pig No. 4 27.51
Salisbury pig No. 4, high 275
Scrap 45.0 >
100 j
Salisbury pig No. 4, high 50 "|
Salisbury pig No 4 . 50 !
100 j
Salisbury pig No. 4 33.31
Salisbury pig No. 4, high 11 . 1 |
Soft pig 22.21
Remelted pig 33.3 f
100 J
Salisbury pig No. 4 25 1
Salisbury pig No. 4, high 25
Scrap . . . 50 >
100 J
Richmond pig No. 1 11.1")
Richmond pig No. 2 11.1
Salisbury pig No. 4 16.7
Salisbury pig No. 4, high 16.7
Scrap . 44 4
100 J
Salisbury pig No. 4. 33.3")
Salisbury pig No. 4, high 11.1
Soft pig 22 2 !
Remelted pig 33.3 [
}00 J
Salisbury pig No. 4 20 |
Salisbury pig No. 4, high 20
Soft pig 20
Scrap . 40
100 ,
472 THE MATERIALS OF CONSTRUCTION.
COMPOSITION AND STRENGTH OF HIGH-GRADE CAST IRONS continued.
g
Carbon .
-
c .5
Composition of Charge.
&
3
1
%
o>
o
^
3
3
CO cS
I
o
Ic
IS
&
c
^
p
- CT*
a
O
p.
E
s
B 1
I
'Z f- 1
"g
a
o
6
cS
55
's
to
oi 53
-1 P<
aj
K
Richmond pig No. 1 9.41
Richmond pig No. 2 9.41
Salisbury pig No. 4 9.4 I
Salisbury pig No. 4, high 9.4 }
Scrap 62.5 1
cupola
2.890
0.458
0.388
1.645
0.105
0.487
27,320
100 j
Muirkirk pig ... .38.51
Soft pig 23.0 |
air-furnace
2,538
0,979
0.348
1.316
0.130
0.642
26,480
15. 6T
Remelted pig 38.51-
100 J
Salisbury pig No. 4 9.61
Salisbury pig No. 4, high 9.6
Richmond pig No 1 96
Richmond pig No. 2 9.61
Soft pig 23.1 f
do
2,770
0,256
0.470
2.444
0,110
0,587
28,010
Remelted pig. 38 5
100 J
Salisbury pig No 4 . 8.31
Salisbury pig No. 4, high 8.3
Richmond pig No. 1 8.3
Richmond pig No. 2 .... 8.3 !
Soft pig 20. 0|
do
2.751
0.357
0.435
1.908
0.095
0.420
29,120
11.08
Remelted pig 46.7
100 J
Salisbury pig No. 4... 33.31
Salisbury pig No. 4, high 11.1
Soft pig 22.2
Remelted pig 33 3 f
do
2.538
0.634
0.355
1.222
0.090
0.766
28,520
21.04
100 J
Salisbury pig No. 4 33.31
Salisbury pig No. 4, high 11.1
Soft pig 22.2,
Remelted pig 33.3 f
do
2.577
0.185
0.361
1.146
0.115
0.762
31,020
100~j
Salisbury pig No. 4 33 . 3 1
Salisbury pig No. 4, high 11.1
Soft pig 22.2 ,
Remelted pig 33.3 f
do
2.116
0.640
0.450
1.419
0.125
0.678
31,140
17.44
100 J
Salisbury pig No. 4 22.5 ]
Salisbury pig No. 4, high 22.5 |
Scrap 55 }
fMTnnln
2 825
047Q
0.361
1.062
0.076
0.238
32,010
16.82
100~J
* ^< y
Salisbury pig No. 4 8.31
Salisbury pig No. 4, high 8.3
Richmond pig No. 1 8.3
Richmond pig No. 2 8.3 '
Soft pig 20 r
air-furnace
2.481
0.687
0.454
1.175
0.120
0.673
31,990
Remelted pig 46.7
100 J
THE STRENGTH OF CAST IRON.
473
The tensile strength and the relative hardness of various high-grade
compositions are given in Table XXIII.
348. The Compressive Strength of Cast Iron varies from 60,000 to 200,000
pounds per square inch as shown in Fig. 55, p. 94. In Fig. 387 is shown a
3M00
OPC,
flT/L
MAT\
$S/tf/
o -402 m 006 m M
FIG. 387. Average Results of Twenty-two Tests of Cast Iron in Compression, from
B. L. 12-iu. Rifle-mortars. (Wat. Ars. Rep. 1894, p. 105.)
stress-diagram in compression, plotted from the average results from twenty-
two tests of gun-iron. As these were made on specimens 10.5 inches long,
having a sectional area of 1 sq. in., they all failed by triple flexure, or as
columns, at an average value of 63,000 pounds per square inch, the actual
crushing strength not having been found. (The tensile strength averaged
33,500 pounds per square inch.)
In Fig. 388 are shown Bauschinger's results on four kinds of cast iron,
as follows:
No. 1 is composed wholly of coke pig iron.
" 2 " " " " charcoal pig iron.
" 3 " 90$ coke-pig and 10$ steel.
" 4 " 80$ coke-pig and 20$ steel.
The tests were made to determine the time-effect in testing cast iron,
there being two specimens of each kind in both tension and compression,
with one of which a rest of one minute was allowed after each increment
of load, and with the other a rest of five minutes. No time-effect was
discovered, however, and here the mean' curves only are shown.
In all these cases it will be observed that no well-marked elastic limit
can be identified, although the curves are plotted to a very large scale of
deformations.
474
THE MATERIALS OF CONSTRUCTION.
FlG. 388. Stress-diagrams on Four Kinds of Cast Iron. Each Curve the mean of two
tests. (Bauschinger's Communications, vol. xx, Plate XII.)
THE STRENGTH OF CAST IRON.
475
e00&
4000
349. The Cross-breaking Strength of Cast Iron is in general from one
and one half to two and a quarter times its strength in tension on solid
rectangular sections. The cause of this was discussed in Chapter V. In
Fig. 389 are shown autographic stress-
diagrams of four kinds of cast iron
made on Keep's standard test-bars
-J- in. square and 12 in. long. They all
showed the same strength of 450 Ibs.
at the centre, giving a computed mod-
ulus of rupture of 64,800 pounds per
square inch. It will be noted that
their deformations under like loads
are very different, thus giving rise to
enormous differences in their strength
to resist shock. Thus their resistances
to shock, as determined by the total
areas of their stress-diagrams, are re-
spectively 10.0, 21.5, 28.9, and 35.1
inch -pounds per cubic inch.* These
four tests were selected to show the
necessity of observing the deflections
as well as the loads, if resistance to
shock is to be found. These diagrams
also show that no great error is made
in the case of cast-iron if the area of
the stress-diagram in cross-bending
be assumed to be equal to one half the
product of the breaking load into the
total deflection. This is the common FIG.
rule for cast iron. This half-product,
divided by the volume or weight of
the specimen between the supports,
gives the shock-resisting modulus in
inch-pounds per cubic inch or per pound of metal as the case may be, and
is independent of the particular dimensions except that the smaller the cross-
section the greater the strength-modulus, as is conclusively shown in Keep's
curves in Fig. 57, p. 96. Higher shock-resisting moduli will be obtained,
therefore, on small (thin) sections than on large ones, and the only safe
rule is to learn by trial what products to expect and to demand for given
sizes of test-specimens and given grades of iron. (See Fig. 57, p. 96.)
The committee of the American Society of Civil Engineers recommended
(1896) a cross-breaking test of cast iron in which /= 36,000 Ibs. on a bar
d/ 0.2 A3
389. Cross-breaking Autographic
Stress-diagrams of Four Kinds of Cast
Iron all having the Same Static Strength.
(Keep, Tr. Am. Soc. Mech. Engrs., vol.
xvn, 1896.)
* These can be taken out per pound of metal if preferred.
476 THE MATERIALS OF CONSTRUCTION.
2 in. by 1 in. tested flatwise on a 24-in. base, with a deflection v of 0.3 in.
This corresponds to only 6 inch-pounds per cubic inqh.
Mr. Vf. J. Keep has shown * that the 1-in. square bar gives more uniform
results than any other size, because at this size the effects of varying per-
centages of silicon are not appreciable, as is shown by Fig. 57, p. 96. In
reducing the great number of tests on cast iron (some 500 in all) in sizes
from 4- in. square to 4 in. square made for the Committee on Methods of
Testing Materials of the American Society of Mechanical Engineers, Prof.
Benjamin has found that the strength of all sizes of cast-iron bars can be
related to that of the 1-in. -square bar 12 in. long by the following formula: f
0.83^1.89
" ~ *' 1.058 > ....... ' ' (1)
where W= breaking load on the centre of the bar;
k = breaking load for a bar 1 in. square and 12 in. long;
1} breadth of the bar in inches;
h = height of the bar in inches;
I length of the bar in feet (multiples of 12 in., which is the length
of the standard bar.)
The theoretical relation is W = h~.
L
350. The Modulus of Elasticity of Cast Iron varies quite as much as its
strength. In this respect it is anomalous, since for all rolled iron and steel
the modulus of elasticity varies only about five per cent from its mean value
for strength variations, in the case of steel, of several hundred per cent.
There is probably a pretty definite relation between the strength of cast
iron and its modulus of elasticity, the stronger iron having the higher
modulus of elasticity, that is, it is stiffer. At least this relation is very clearly
shown in the curves in Fig. 55, p. 94. In general this modulus varies from
10,000,000 to 30,000,000, but for ordinary foundry iron it may be taken at
from 12,000,000 to 15,000,000, or about one half that of wrought iron and
rolled steel.
It may be well here to call attention again to a ready method of reading the
modulus of elasticity from any of the tension- or compression-stress diagrams
given in this book. Where the loads are given in pounds per square inch
and the deformations in percentages, or proportionate parts of the length,
then by observing where the tangent to the curve at the origin (or the curve
itself if it is straight so far) crosses the ordinate which marks a deformation
of 0.001, one has only to read the corresponding stress per square inch and
multiply it by 1000. The modulus of elasticity of cast iron is approxi-
mately the same in tension, in compression, and in cross-bending.
* Trans. Am. Soc. MecJi. Engrs., vol. xvn. (1896) p. 681.
f Ibid , p. 692.
THE STRENGTH OF CAST IRON. 477
351. Kirkaldy's Results. In Table XXIV are given the results of 469
tests of cast iron in each of the three ways, tension, compression, and cross-
bending, all on identical material in each case. The averages of all are :
Tensile strength 25,000 Ibs. per square inch
Compressive strength 121,000 " " " "
Cross-bending modulus >... 38,000 " " " "
Quality-coefficient 6.5 in. -Ibs. per cubic inch.
The mere fact that these specimens were submitted to Mr. Kirkaldy for
testing implies that the mixtures were better than are commonly used in
foundry practice, and yet these average results could readily be reached in
any good foundry.
The small ratio of the cross-breaking modulus of rupture to the tensile
strength, this having an average value of but 1.52, is due to the great depth
(2 in.) of the transverse test-bar. The small value of the "quality-coeffi-
cient" is doubtless due to the same cause. Otherwise all these irons would
be classed as comparatively brittle.
As this " quality-coefficient" was not taken out by Kirkaldy, and as it is
probable that only breaking strength was specified, it is quite probable that
high strength has been attained in this material at the expense of resilience
or shock-resistance. The ratio of the cross-breaking modulus to that in
tension is also lower with less flexibility. The modulus of elasticity was not
computed for any of these tests, and as only the final deflections are given
in the published report, it cannot now be computed from the tabular matter.
352. Shrinkage Stresses. The shrinkage of cast iron after it crystallizes
is so great that, if not provided for, it causes excessive deformations which
may develop very great stresses, even to rupture. The heavier or the thicker
the casting the greater are these shrinkage stresses. These have been
studied in the case of cast-iron guns, and one such analysis is shown in Fig.
390. Here the metal was over 11 inches thick. The outer and inner
surfaces cooled first, and the subsequent shrinkage of the interior put these
parts in compression. But since the total internal stress across any diametral
section must be zero, there being no external force acting, it follows that
the total tensile stress must equal the total compressive stress. These were
all found directly by cutting off a zone included between two transverse
sections, and by cutting this up into a series of concentric rings as shown by
the dashed lines in Fig. 390. Before cutting these, four diameters of each
ring were carefully measured, and these same diameters were again measured
after cutting out. An increase in mean diameter indicated an initial com-
pression, and vice versa, the initial stresses being found from the equation
f = UB,
where/ = stress in pounds per square inch;
A = proportionate change in circumference;
E = modulus of elasticity of the material.
478
THE MATERIALS OF CONSTRUCTION.
TABLE XXIV. SUMMARY OF RESULTS OF TESTS ON CAST IRON IN TENSION,
COMPRESSION, AND CROSS-BENDING, ON IDENTICAL MATERIAL.
From Kirkaldy's Report, 1891 (Report T T).
Total Number of Tests I
from One Foundry.
Grade.
Tension Strength in Pounds
per Square Inch.
Compression.
Cross-bending.
Strength in Pounds
per Square Inch.
Ultimate Deforma-
tion, Per Cent.
c c o>
'" 5
co 3
W i-, CT*
gJoEf
^
8 ?
~ -^ .
&2S.S
fill
o
1
8^
^ D
" o
3
0, hH
rt.2
i
* Resistance to Shock
or "Quality-factor"
in Inch-pounds per
Cubie Inch,
151
74
Highest ...
32,821
26,165
141,632
122,279
6.65
9 26
47,710
38,020
25,820
.32
.27
.21
8.84
5.94
Mean . .
Lowest
16,250
103,165
12.20
3.14
Highest, . . .
27,614
24,303
124,251
117,242
37,390
33,840
.32
.26
6.22
5 09
Lowest ...
19,311
109,682
27,260
.21
3.31
58
Highest
28,740
24,148
17,698
131,912
115,572
13.30
9.98
4.45
43,540
39,550
.40
.36
.33
1.00
8.24
6.46
Mean
Lowest
93/759
33,840
46
Highest
30,630
23,339
12,688
137,165
105.918
11.80
11.95
46,460
36,190
.33
.36
8.87
7.54
Mean
Lowest
66,363
12.70
34,240
.38
5.33
15
15
Highest ...
29 782
138,496
116,833
10.40
1066
46,650
44,460
42,860
.36
35
.32
9.72
899
Mean
22,727
Lowest
15 580
83,307
7.90
7.94
Highest
26,040
23.925
132,857
123,044
36,100
34,760
31,360
.28
.26
.24
5.85
5.22
4.35
Mean
Lowest
22,711
113,233
15
Highest
25,708
23,129
17,617
123.531
116,356
9.20
9 12
7.35
40,660
39,700
.36
.36
8.47
8.27
7.56
Mean
Lowest
105,258
38,400
.34
13
Highest
27,644
24,321
123,708
116,538
8.55
8.64
39.290
32.780
.39
.29
8.19
5.50
Mean
Lowest
19,188
104,281
7.00
26,930
.21
3.27
* This ' ' quality-coefficient
area of the stress-diagram in
" is a measure of the resistance to shock, or it is the
inch-pounds per cubic inch, found by multiplying the
THE STRENGTH OF CAST IRON. 479
SUMMARY OF RESULTS OF TESTS ON CAST IRON continued.
Total Number of Tests 1
from One Foundry.
Grade,
Tension Strength in Pounds
per Square Inch.
Compression.
Cross-bending.
Strength in Pounds
per Square Inch.
Ultimate Deforma-
tion, Per Cent.
Computed Stress on
Outer Fibre in
Pounds per Square
|
"o
13 .
<%$
0|
H)M
.5
5
*Resistance to Shock
or ''Quality-factor"
in Inch-pounds 'per
Cubic Inch.
13
Highest
26,502
21,711
16,090
175,950
123,336
103,859
4.05
5.49
6.45
54,140
40,750
27,070
36,050
34,750
31,870
40,460
38,020
.45
.34
23
14.10
8,02
3.60
Lowest
10
Highest .
25,176
24298
23,511
127.988
125,962
120,874
7.55
7.03
6.40
.32
.29
.25
6.68
5.83
4.61
jVIeau .
10
Highest
30,31(1
29,268
28,436
136,266
133,682
12.25
12.61
.30
.27
.21
7.02
5.94
4.14
Lowest
129,524
13.90
34,080
10
10
Highest
30,018
29,472
130,145
127,714
1210
11 90
39, 550
38,400
.34
.32
7.78 I
7.11
jVIean . .
27,501
28,416
27,763
122,155
11.75
37,050
1 .29
6.22
Highest
140,542
138,054
6.60
6.54
38,780
37,250
.31
.28
6.96
6.03
Mean
Lowest. . . .
26,851
135,577
6.60
35,380 .27
5.52
10
25 520
127,703
121,593
7.80
7.71
34,960
33,020.
. 32
.28
.26
6.46
5.35
4.54
Mean
24,803
23,435
Lowest
110,405
6.60
30,190
10
Highest
28,518
27 914
27,276
129,603
126,701
124,415
9.10
9.70
9.15
40 370
38,060
35,470
.35
.32
.26
8.18
7.05
]\Ieau
Lowest
5.34
g
Hio-hest ...
33,616
29,202
143,939
137,804
46,990
43,490
.41
.39
11.15
9.81
Mean
Lowest
19,046
124,410
43,490
.31
6.82
breaking load by the final deflection and dividing by twice the volume of the bar. It
was not computed by Kirkaldy. All these bars were 2 in. x 1 in. X 36 in. long, tested
edgewise, the tension and compression specimens being cast on the same pattern. The
great depth of these bars makes this quality-factor and also the modulus of rupture run
low as compared to tests on thinner specimens.
480
THE MATERIALS OF CONSTRUCTION.
In this way the stress-diagrams shown in Fig. 390 were computed and
drawn by the author from the data furnished in the original report. It
indicates that the interior surface was under an initial compressive stress of
some 7000 Ibs. per square inch, the outer surface of some 13,500 Ibs. per
square inch ; while the interior was under a tensile stress of some 2000 Ibs.
pei- square inch. Evidently the tension and compression areas on these
diagrams must equal each other. This is a very simple illustration of such
shrinkage stresses, because of its simple and symmetrical form. In complex
forms it would be impossible to study or predict the character of these
FIG. 390. Showing the Shrinkage Stresses in Cast-iron Cannon 11 in. thick.
(Wat. Are. Rep.)
stresses. They are evidently less when all parts are made of approximately
the same thickness.
353. Strength of Cast Iron Increased by Shocks. Mr. A. E. Outerbridge
has shown* that castings which have been subjected to a great number of
shocks or blows are from 10 to 15 per cent stronger under a static load and
over 20 per cent stronger under impact than they are before receiving such
treatment. He attributes this result to a sort of molecular rearrangement
by which the cooling stresses are relieved. In other words, such treatment
is equivalent to an annealing process. This is probably a general fact and
true for all kinds of castings, although this remains to be proved.
354. Strength of Malleable Cast Iron. For the experimental results of
tests of the strength of malleable cast iron see Art. 82, p. 114.
355. Cast-iron Pipes and Columns. In estimating the actual strength of
cast-iron pipes subjected to internal pressure, a great source of uncertainty
* Trans. Am. Inst. Mm. Engrs., Pittsburg Meeting, 1896.
THE STRENGTH OF CAST IRON. 481
lies in the probable unequal thickness of the metal in the upper and lower
sides when cast. There is a great tendency of the core to rise from the
buoyancy of the liquid iron, and since no core, however strong, can be abso-
lutely rigid, it must be assumed that it does always lift somewhat at the
centre, however strongly it is held at the ends. As a matter of fact, the
cores are not usually very rigid, so that such pipeL are apt to be very unequal
in thickness on the opposite sides, as shown in Fig. 391. The regular bell-
and-spigot water- and gas-pipes are now all cast vertically, and hence this
FIG. 391. Actual Section of an 8-iu. Cast-iron Steam-pipe. (From The Locomotive,
Oct., 1896.)
danger is largely obviated, but flange-pipes, such as are* used for steam
purposes, are cast horizontally. Any great inequality in thickness can be
found by rolling the pipe down inclined ways and noting the irregularity
of motion
Cast-iron columns, such as are used in buildings, are also cast horizon-
tally, and are subject to this same contingency. These may be bored to
determine thickness, but pipes cannot be examined in this way. It is such
undiscoverable faults as this which form the greatest objection to the use of
cast iron for these purposes. For other defects in cast-iron columns see
Plate III.
CHAPTEK XXV.
THE STRENGTH OF WROUGHT IRON.
356. The Tensile Strength of wrought iron along the grain varies from
45,000 to 55,000 pounds per square inch. It is greater in small rods and
thin plates than in large bars and thick plates, the material remaining
the same. .This is shown in Fig. 392, where the same material has been
rolled into bars from f in. to 2 in. in diameter, the tensile strength varying
from 52,000 in the smaller to 47,500 pounds per square inch in the larger
sizes.
The Elastic Limit is more dependent on the thinness of the final
section than on the tensile strength, as is well brought out in Fig. 392. Here
the apparent elastic limit varies from 40,000 pounds per square inch in the
f-in. rods to 23,000 pounds per square inch in the 2-in. rods, and is almost
identical with the "yield-point." This increase in the elastic limit with
increased reduction in the rolls ahv'ays occurs with both wrought iron and
steel, but it is much more pronounced with wrought iron. The true elastic
limit of wrought iron is nearly always much lower than the apparent
elastic limit. In Fig. 392 it is found from 5000 to 7000 Ibs. lower in every
case. In mild steel these two limits are almost identical.
The Percentage of Elongation in 8 in. varies from 5$ to 25$ when tested
in the direction of the fibres, depending on the quality of the material, the
reduction of area averaging about 50^ more than the elongation. The elon-
gations recorded in Fig. 392 were all taken on a length of 20 in., which
somewhat reduces the percentage, especially for the smaller sections.
357. The Tensile Strength across the Grain is always much less than
along the grain in the case of wrought iron, while with steel there is no
appreciable difference. Very few tests of wrought iron across the fibres
are to be found on record, but the author has often observed in his own
practice that it is very much less in this direction than parallel to the
direction of rolling. In Prof. Bauschinger's Communications, vol. n, we
find an elaborate study of this subject on wrought-iron boiler-plates from
eight different sources, some of them having been taken from boilers which
had exploded. From this report we have
1. From eight tensile tests along and eight across the grain, from an
exploded boiler, the ratio of lateral to longitudinal strength was 0.74.
482
THE STRENGTH OF WROUGHI IRON. 483
. From a wrought-iron plate from another exploded boiler eight test-
FIG. 392. Stress-diagrams (in tension) of Wrought-iron Bars of varying Diameters,
all rolled from the same Material. All Elongations measured on a Length of 20 in.
The " Apparent Elastic Limit " falls from 2# to 10# lower than the indicated "Elas-
tic Limit " in the original Report ; it varies from 23,000 in tbe 2-in. to 40,000 Ibs. in
the 3-in. specimens ; and it marks a point where the permanent set is less than 0.0001
of the length of the specimen. Each diagram is the average of from 3 to 6 tests.
(Wat. Ars. Rep. 1888.)
specimens were cut in each direction, giving a mean ratio of lateral to
longitudinal strength of 0.71.
484 THE MATERIALS OF CONSTRUCTION.
3. On six other new plates from as many different sources he obtained
ratios of 0.76, 0.62, 0.92, 0.90, 0.76, and 0.83.
The average value of all these is 0.78.
In short, we may fairly affirm that the ultimate tensile strength of
wrought iron transverse to the direction of the rolling is only about three
fourths of its strength parallel to this direction.*
The author is credibly informed that the best English Yorkshire
(Lowmoor) iron plates are always rolled from "puddled lumps" 12 in.
square, which correspond to muck-bars, these being piled so as to cross
their grain, and in this way the final plates are nearly as strong transversely
as they are longitudinally. It is claimed that for an ultimate strength of
51,500 pounds per square inch, with an elongation of 16$ longitudinally,
it shows an ultimate strength of 45,000 pounds per square inch and an
elongation of 12$ transversely. As this material costs about four times
as much as the best mild-steel plates, its use for all purposes where forging
and welding are not required is rapidly declining.
358. Tensile Strength of Wrought Iron as Affected by Pulling Speed and
by Length of Reduced Section. In Table XXV are given the results of a
series of very careful tests to determine the effects of speed and of the length
of the reduced section on the tensile strength of wrought iron. It will be
observed that the ultimate strength is somewhat increased by very rapid
testing, while the elongation and reduction are not appreciably affected by
this average range of speed from 15 sec. to 8.5 min. The strength is also
much greater for very short reduced sections than for longer ones. Similar
results on steel bars are shown in Fig. 426.
359. The Compressive Strength of Wrought Iron, like that of any of the
ductile metals, must be regarded as the "apparent elastic limit" or "yield-
point." Here the material buckles out of shape, and if the specimen has
appreciable length, failure at once follows. If this be allowed, then it may
be seen at once, from Fig. 392, that the same puddle-ball, rolled to different
sections, will show compressive strengths anywhere from 26,000 to 40,000
Ibs. per square inch. Since wrought-iron columns are built up of structural
forms which have been rolled to thin sections, from J- to f in. in thickness,
it follows that such material will have a "yield-point" or "apparent elastic
limit " of from 30,000 to 40,000 Ibs. per square inch. Since, also, the amount
of reduction in the rolls has a less effect on the elastic limit of mild steel, it
follows that the yield-point, and hence the compressive strength, of a
wrought-iron column may be about equal to that of a steel member built
up of similar sections, although the ultimate strength of tne steel in tension
may be 25 per cent higher than that of the wrought iron. This variation
of the compressive strength (yield-point) of wrougnt-iron columns with the
* The author has been unable to find the data for plotting a stress-diagram of
wrought iron across the grain.
THE STRENGTH OF WROUGHT IRON.
485
TABLE XXV. TESTS OF WROUGHT IRON TO SHOW EFFECT OF SLOW AND
OF RAPID FRACTURES ON SPECIMENS OF VARYING LENGTH.
Sixteen specimens taken from the same bar of iron If in. square, all reduced to a
diameter of 1.008 in. or to a sectional area of 0.80 sq. in. Specimens marked f rom A
to P consecutively, as cut from the bar. (From U. S. Wat. Ars. Rep. 1887, p. 924.)
Marks.
Length.
Elastic
Limit per
Square Inch.
Ultimate
Strength
pei-
Square Inch.
Duration
of
Test.
Gauged
Length.
Elonga-
tion in
Gauged
Length.
Contrac-
tion.
Final Load
per
Square Inch
on Rup-
tured Sec-
tion.
Inches.
Pounds.
Pounds.
Inches.
Per cent.
Per cent.
Pounds.
A
Grooved ....
57,200
6 miu.
32.4
72,090
B
Grooved
59,250
6 sec.
32.4
.
C
0.80
27', 375
49,380
6 uiin.
49.1
71,260
D
0.80
....
50,750
8 sec.
37.1
....
E
1.60
28', 500
47,730
10 min.
1
56.0
47.6
78,280
F
1.60
....
49,130
13 sec.
1
50.0
46.2
G
2.40
80,125
47.980
8 min.
2
39.0
43.2
81,680
H
2.40
49.000
14 sec.
2
41.5
49.1
I
3.20
29,625
47,070
10 min
3
39 .
49.1
77,150
J
3.20
....
48,120
15 sec.
3
36.0
47.6
....
K
4.80
29,875
46,860
10 miu
4
32.0
43.2
70,260
L
4.80
48,000
18 sec.
4
32 8
47.6
M
6.40
29,750
47,450
9 miu.
6
25.0
49.1
80,590
N
6.40
48,250
20 sec.
6
32.0
47.6
....
8.00
28.500
45,650
9 miu.
8
25.9
41.7
66,950
P
8.00
....
46,000
30 sec.
8
25.9
41.7
....
Mean of slow tests =
48,670
8.5 miu.
36.1
44.4
73,500
3Iean of quick tests =
49,810
15.5 sec.
d6.4
43.7
thickness of the sections of which the member is composed accounts for a
large portion of the disc^emncies found in the results of tests on wrought-
iron coiiimns in commercial sizes (see Fig. 297, p. 364). Thus Tetmajer
found for short columns composed of four angle-irons, 2.4 in. X 2.4 in. X
-] in. in thickness, me average strength of the wrought-iron columns was
38,000 Ibs. per square men of net section, while that of similar mild-steel
columns was but 36,000 Ibs. per square inch. The ultimate tensile strength
of the wrought iron was 50,000 Ibs. per square inch, while that of the
steel was 61,000 Ibs. per square inch. With angles 0.4 in. thick the steel
columns had an ultimate strength of 37,"500 Ibs. per square inch, while the
wrought-iron columns showed only 32,300 Ibs.* With sections in. thick
and less, therefore, there is probably little difference between the strength of
wrought-iron and soft-steel columns.
360. The Shearing Strength of Wrought Iron. The most elaborate
investigation ever made on wrought iron, so far as the author is aware, was
that made by Bauschinger and reported in vol. n of his Communications.
He made several hundred tests of the shearing strength of wrought-iron
plates from seven different sources, finding the shearing resistance in two
* Communications, vol. iv. pp. 141, 149, and 155.
486
THE MATERIALS OF CONSTRUCTION.
directions on each of the three of the principal planes, as shown in Fig. 393.
As there was a general agreement in the relative strength on these planes, only
the averages of a portion of the tests are given in the diagram. In general,
we may say that the shearing strength across the thickness of the plate, either
with or across the grain, is about 80 per cent of the tensile strength, while
if the external forces lie in the plane of the plate, and be applied on the
j. JLL MI At y
FIG. 393. Shearing Strength of Wrought Iron on the Six Principal Planes, as compared
to its Tensile Strength. The numbers indicate the number of results averaged.
The direction of rolling is indicated by the large arrow. (Bauschinger's Communi-
cations, vol. ii.)
planes of shear perpendicular to the plane of the plate, the shearing strength
is about the same as the tensile strength. The shearing resistance on a
plane parallel to the plane of the plate is less than 45 per cent of the tensile
strength.
361. The Effects of Stressing Wrought Iron beyond its Elastic Limit is to
raise this limit, and also to greatly increase the ultimate strength after a
period of rest. Thus in Fig. 394 are shown the results of tests on three
wrought-iron bars 3 in. X 1 in. Here the apparent elastic limit on the
second test (computed on the original cross-section) is much greater than
the original ultimate strength, and almost equal to the ultimate strength on
THE STRENGTH OF WROUGHT IRON.
487
flff/
FIG. 394. Wrought-irou Bars, 3 in. by 1 in. retested, First test gave El. lim. = 30,000;
ult. str. = 53,700; % elong. = 16 on 100 in. Ends of broken bars retested and loads
computed per sq. in. of original section. Second Elongation taken on 50 in. and per
cent elongation computed on the new gauged length. (Rep. Wat. Ars. 1882.)
30
SO
W9/W
//V 61W
rf-
FIG. 395. Increase in the Tensile Strength of Wrought Iron after having been Stressed
to the Tensile Limit. Points plotted are averages from 5 to 15 tests each. (Rtp.
U. S. Test Board, 1881, vol. i. pp. Ill and 115.)
488
THE MATERIALS OF CONSTRUCTION.
the second test. After annealing, however, the material snowed a much
lower elastic limit and ultimate strength than it had before it was stressed
at all. Stretching the material somewhat beyond the elastic limit, but not
to failure, would show less marked results. If little or no time is allowed
to elapse between the tests, there is no permanent increase in the strength,
but the rate of increase in strength with time after stretching to its tensile
limit is well shown in Fig. 395.
When any ductile metal is stressed beyond its normal elastic limit in
either tension or compression it loses its perfect elasticity under the opposite
kind of stress. Thus for wrought iron, Fig. 396, has been constructed by
~80jOOO\
FIG. 396. Showing that a small Permanent Set of Wrought Iron from either Ten-
sion or Compression greatly reduces the Elastic Limit under the Opposite Stress.
(Gray, in The Digest of Physical Tests, vol. i. p. 232 (1896).)
plotting consecutively a series of autographic diagrams taken by Prof. Gray.*
Although at first the permanent set given to the specimen in tension was
less than one per cent of its length, yet when this was relieved and a com-
pressive stress applied it was shown that the specimen was no longer per-
fectly elastic in compression, but gave a continuously curving stress-dia-
gram. It was then compressed back to its original length, and a tensile
stress applied, when it was found to be no longer perfectly elastic in tension.
It was now permanently elongated about one per cent of its length, and the
load removed and again imposed, and it was found that the specimen was
now perfectly elastic in tension up to the limit of its previous loading and
somewhat higher, but when it was now elongated 18 per cent it was no
* Reported in The Digest of Physical Tests, vol. i. p. 206.
THE STRENGTH OF WROUGHT IRON.
489
longer perfectly ela'stic in either tension or compression. Similar results
are shown on steel in Figs. 43G and 437.
362. The Strength of Wrought-iron Chains. The hundreds of tests on
wrought-iron chains given in the Report
of the U. S. Test Board, 1881, vol. i,
show that the ultimate strength of
chains may be taken at 1.6 that of the
iron from which the links are made. It
also appears from these tests that open
links are somewhat stronger than studded
links, though the open-link chains take
a permanent set earlier than the studded
links. It is thought, however, that open-
link cables would foul more readily than
the studded cables. The elastic proper-
ties of open-link chains made of 1-in.
and -J-in. iron are shown in Fig. 397,
where the tests have been carried to the
proof -load only, this being such as to
give to the chains a permanent set of
> i
about two per cent of their length. They FIG 397- _ Proof Tests of Cba j us with
now become perfectly elastic to 20,000 Open Liuks< Tbe 118 _ in chain
and 15,000 Ibs. respectively, and are also
about five times more stiff, or rigid, than
they were at first. All chains are im-
proved by this treatment, while it also discovers any very poor welds the
chain may have.
made of 1-in. iron, and the 90-in.
chain of f-in. iron. (Wat. Ars.
Rep. 1894.)
CHAPTER XXVI.
THE STRENGTH OF STEEL.
356. The Tensile and the Compressive Strength of Steel of various per-
centages of carbon are well shown in Figs. 399 to 404.* A study of these
...-
-1
Y
Aff0)t0ftT/{M//i7'
MS
000S
0./0
00/0
\
f/cfc
FIG. 398. Compression and Tension Tests on Midvale Steel Bars. Compression tests
on bars 1 in. diam. and 5 in. long, tension tests on bars 0.56 in. diam. and 5 in. long.
(Wat. Ars. Rep. 1889.)
* In all these figures the loads are given in pounds per square inch, though this is not
always so stated. '
490
THE STRENGTH OF STEEL.
^
"
::,
,:
V /Y/
/ 7 /7
V//?
// 1/
03,
.02 .04
7
;t
>
-cc
L
7/c
7 C
'7%
V
.02 .04 .06 .08
" .00/ .002 ^ ,00/ .002 .003
FIG. 399. Compression and Tension Stress-diagrams of Steel Bars of Varying Percentages
of Carbon. Compression specimens 12 in. long and 1 in. in diam.j- = 33\ with flat
ends. Tension specimens same diam. with a gauged Length of 30 in. All speci-
mens turned from open-hearth steel bars 1 in. in diam. ( Wat. Ars. Rep 1886 and
1887.)
492
TEE MATERIALS OF CONSTRUCTION.
figures, all of which are typical and characteristic, will lead to the following
conclusions:
4400
60000
40000
Ea
N
04
& f.
FfiL
w
1
7"
^
/7 /
0.4?
/v
.02
.00/
FIG. 400. Steel Tests supplemeiilary to those of Fig. 399.
1. The tensile strength varies from 55,000 Ibs. for 0.1 per cent carbon
to 150,000 Ibs. per square inch for 1.0 per cent carbon.
THE STRENGTH OF STEEL.
493
2. The " apparent elastic limit " is found between 60 and 70 per cent
of the ultimate strength.
3. The " apparent elastic limit " in compression is practically the same
as that in tension.
FIG. 401. Tension Stress-diagrams of Gautier Steel Bars which were used in the En-
durance Tests on Rotating Shafting subjected to Reversals of Stress. Percentages
of carbon and areas of the stress-diagrams in square inches are given on the curves.
(Rep. Wat. Ars. 1894.)
4. The modulus of elasticity in compression is slightly greater than that
in tension, and in both cases it is practically independent of the percentage
of carbon and of the ultimate strength.
5. The ultimate strength in compression is practically equal to the
" apparent elastic limit " (Fig. 294, p. 360).
494
THE MATERIALS OF CONSTRUCTION.
6. In the mild and medium steels (carbon 0.2 to 0.6 per cent) there is
a very decided drop in the stress-diagram after reaching the "apparent
elastic limit " often as much as 5000 or 6000 Ibs. per square inch.
FIG. 402. Tension Stress-diagrams (incomplete) of Steel Bars used for Endurance
Tests of Rotating Shafting. Average product of ultimate strength by percentage
of elongation is 2,180,000. (Wat. Ars. Reps. 1889 and 1891.)
7. The coefficient of expansion decreases with the increase in carbon, its
average value being about 0.0000065 per degree F. (Fig. 294).
8. The high carbon-steels are greatly softened, the tensile strength
lowered, and the ductility increased by annealing (Fig. 404).
THE STRENGTH OF STEEL.
495
357. The Effect of Thickness on the Mechanical Properties of Structural
Steel. In Figs. 400, 407, and 408 are shown the effect of thickness (bars
and angles) on the mechanical properties of structural steel. Thus from
.02
FIG. 403. Tension Tests of Steel Bars used for Endurance Tests of Rotating Shafting.
(Wat. Ars. Rep. 1889 and 1891. Supplementary to Fig. 402.)
Fig. 40G, where the thickness ranges by eighths of an inch from f to J inch,
we may see
1. The ultimate strength is nearly constant,
2. The apparent elastic limit varies from 41,000 at f to 37,800 Ibs. per
square inch at the f-in. thickness.
496
THE MATERIALS OF CONSTRUCTION.
/77
FIG. 404. Tension Stress-diagrams of Three Grades of High-carbon steel. (Wat. Ars.
Rep. 1894.)
mo.
20.000
7
WVRT/MATl
JO .20 .30
FIG. 405. Autographic Stress-diagram of
Rivet-steel in "Tension, showing Effect
of Removing the Load. (M. Dupuy, in
An. d. Fonts et Chaussees, PI. I. 1895.)
FIG. 406. Effect of Thickness on the
Mechanical Properties of Acid Open-
hearth Steel Angles. (Campbell's
Structural Steel, p. 202.)
THE STRENGTH OF STEEL.
497
3. The percentage of elongation in 8 in. is nearly constant.
4. The reduction of area varies from 58 per cent at f in. to 50 per cent
at I in. thickness.
Elastic limit
.0. Ihe elastic ntio: Ultimate strength vanes from 68 per cent at f in.
to 01.5 per cent at -| in. thickness.
emo
wooo
JWOO
7m.
-f
&
^
s
60.000
40 50.000
20 30,000
flV/VM
0M
OAAM&L
3
30
20
FIG. 408. Effects of Thickness on Bessemer Steel
Angles. Each point is the mean of fifty re-
sults. (Campbell's Structural Steel, p. 199.)
FIG. 407. Effect of Thickness on
the Mechanical Properties of Mild
Steel, Natural and Annealed.
(Campbell's Structural Steel.}
From Fig. 409 it may be seen that the variation in ultimate strength and
in the elastic limit for different thicknesses is much greater when the metal
leaves the rolls at a dull-red heat. Here the thickness ranges from ^ in. to
| in., and the elastic limit for normal rolling varies from about 50,000 Ibs.
per square inch at the ^-in. thickness to 39,000 Ibs. at a f-in. thickness.
When leaving the rolls at a dull-red heat, however, the elastic limit for the
-in. thickness was 57,500 Ibs., while for the f-in. thickness it was only
42,000 Ibs. per square inch.
In general the apparent elastic limit rises as the thickness of section
diminishes. Since wrought-iron and steel columns are built up from com-
paratively thin sections of metal (generally from ^ to i in. in thickness), and
as the ultimate strength of these is dependent wholly on the apparent elastic
limit, and not at all on the ultimate strength, it is necessary to evaluate this
elastic limit for the particular thicknesses of sections used a rather than from
498
THE MATERIALS OF CONSTRUCTION.
7Q000 i
6MOO
40000
r*
CT ! "^
BAR
70000
60030
50000
3DOOO
20000
8&
70
60
50
4-0
50
fe* % ABA BAB
FIG. 409. FIG. 410.
FIG. 409. Influence of Thickness on Mechanical Properties when the Percentage of
Reduction iii Rolling is constant, and when the Last Passage in the Rolls was at the
Normal and at a Dull Red Heat respectively. (Campbell's Structural Steel.}
FIG. 410. Showing the Effect of finishing Three Grades of Open-hearth Steel Bars at a
Low Red Heat. (Campbell's Structural Steel, Table 70.)
70000
60.000\
3QOOO
?oono
50
*J
30
20
/V A /V AN AN
FIG. 411. Showing Effect of Annealing Open-hearth Steel Bars 2in. X% in. when
Rolled Originally at a Normal and at a Low Red Heat. (Campbell's Structural Steel,
p. 213.)
THE STRENGTH OF STEEL.
499
special test-bars, which are usually not less than f in. in thickness. (See
Table XXV, p. 503. for the comparison between the results obtained on the
preliminary f-in. billet test-specimen and those from the specimens cut from
rolled bars and plates of various thicknesses from the same ingot.)
The comparatively small variation in the elastic limit (and other proper-
ties) shown in Fig. 408 is due to the fact that all were rolled from the same
sized ingot, and thus the thinner sections had more work done upon them.
When the proportionate reductions are the same in each case the differences
are very much greater, as shown in Fig. 409. These differences almost
wholly disappear on annealing, as shown by Fig. 407.
358. Effects of Finishing at a Low Red Heat. As shown by Figs. 409
and 410, the effect of finishing at a low red heat is to somewhat increase
the ultimate strength and the elongation, and to greatly increase the elastic
limit. This last increase is as much as from 8 to 10 per cent. From Fig.
411 it appears that while annealing lowers both ultimate strength and elastic
08
0.6
05
.
1
1
. '
.
^
..
kd
'
J
'
.
^
;
'.V
1
-
'
^
x
.
QAT.
'0:
SHEA
ff/A/G
wen
' Pft
1 SQ./
V.
TE/\
WL
MM
'A/GTf
/ Pf-h
r ML /
I/
60% 65% 70% 2S% &% 85% 30% Sit,
C.2
FIG. 411a. Showing thejAbsence of any Law of Relationship between Shearing Strength
and Tensile Strength of Steel Plates when the Shearing Strength is determined by
Punching Tests. (Experiments made at Washington University by Messrs Condron,
Harrington, and Norton. See a full account of these in Engr. News, vol. xxxn
(1894), p. 164.)
limit, it does not appreciably increase the elongation, neither does it bring
the "normal" and "dull red" specimens much nearer together in the
matter of the elastic limit. It might fairly have been assumed that annealing
500
THE MATERIALS OF CONSTRUCTION.
would have removed the effects of rolling at a low heat, as it always does the
effects of cold working, but it has not done so in this instance.,
358a. Punching Tests of Steel Plates. It has been claimed * that
structural steel may be tested by punching in place of the ordinary tension
tests. This subject was very fully investigated at the Washington University
Testing Laboratory, as the subject for a thesis, by Messrs. Harrington and
Norton, in 1894, assisted by Mr. T. L. Condron, C.E., and it was concluded
that no definite relationship could be established between the results of
punching and of tension tests on steel plates of varying quality and thick-
The tests were made with autographic stress-diagrams of every test,
ness.
and hundreds of tests made on steel plates of known chemical composition
and tensile properties. Fig. 41 la is here given as one of many such studies
made of the results, all going to show that the punching tests could not be
employed as tests of acceptance under any set of specifications.!
359. Effects of Quenching and of Annealing on Structural Steel. It
is not generally known that quenching from a bright cherry-red heat has a
64000
43000
20000
,20
-$K
k
FIG. 412. FIG. 413.
FIG. 412. Average Effects of Quenching Soft Steel (0.10 C.) from a Dull Red Heat.
Each point the mean of six tests. (Campbell's Structural Steel, p. 52.)
FIG. 413. Effects of Quenching a Very Low Carbon (0.057 C.) Steel at Different Tem-
peratures. (Campbell's Structural Steel, p. 53.)
marked effect on the softest grades of open-hearth steel. That it has is
shown by Figs. 412 and 413, where a steel having a normal tensile strength
* By Alfred E. Hunt before the World's Engineering Congress, Chicago, 1893.
Trans. Am. Soc. Civ. Engrs., vol. xxx. p. 181.
t See an illustrated article by Mr. Condron in Engr. News, vol. xxxii. p. 164.
THE STRENGTH OF STEEL.
501
of 48,500 Ibs. per square inch (Fig. 412) is raised to 63,500 Ibs. per square
inch by quenching. The apparent elastic limit is raised from 32,000 to
40,000 Ibs. per square inch; the elongation is lowered from 32 to 20 per
cent in 8 in. ; and the reduction of area is raised 55 to 63 per cent.* This
reveals also the difference between the elongation and the reduction of area
as characteristics of steel. The highest grade of steel wire, having a
strength of 200,000 Ibs. per square inch and upward, and having an elonga-
tion of but one or two per cent, will commonly show a reduction of area of
over 60 per cent. That is, the stretch largely occurs at the necked-down
portion only. The cold-bending test is a test of the reduction of area rather
than of the elongation. In Fig. 413 the effects are shown of quenching soft
open-hearth steel from various temperatures from a dull red to a bright red
heat. These two figures both show that the softest steel is greatly changed
by quenching, but since the reduction of area, has been raised its capacity
for making short bends has been increased.
When the cold-bending test is specified, after quenching, therefore, it
would appear from Fig. 412 that this soft material is better able to stand
SO
3^?%/%
s
m\
^ L,
S
50
/ xt/ ^ /K X7
FIG. 414. Effect of Annealing 2 in. X f in. Open-hearth Steel Bars of Different De-
grees of Hardness. (Campbell's Structural Steel, p. 210.)
this test than it was in its normal condition. This matter should be proved
by direct bending tests, however, before being accepted as true. If this
proves to be true, then quenching greatly improves the steel for all purposes.
The annealing process seems always to reduce the ultimate strength and
the elastic limit for all grades of steel, the lowering of the latter, as shown
The chemical composition was C = 0.105; Mu = 0.343, P = 0.009; S = 0.024.
502 THE MATERIALS OF CONSTRUCTION.
in Fig. 414, being about 25 per cent. The elongation is very slightly
increased and the reduction of area somewhat decreased. By the annealing
process, therefore, about one fourth of the effective strength of a steel mem-
ber is sacrificed, while its cold-bending capacity is probably reduced. It
would seem, therefore, that it should be practised only when necessary to
remove severe internal stresses produced by cold working.
360. The Billet Test is Characteristic of the Final Rolled Bars and
Pl a t es } r . G US C. Ilenning has shown * that specimen billets rolled or
forged from the sample of the steel dipped from an open-hearth furnace to
test its quality give results which are truly characteristic of the products
rolled from that heat. The tests of these sample billets are commonly called
" heat tests," since one such is made for each heat. If the results' of these
tests agree closely with the tests on specimens cut from the structural forms,
bars, and plates which are rolled from this material, then these latter may
be omitted and reliance placed wholly upon the heat tests. The results of
221 such corresponding sets of tests are given in Table \XV.
This table also contains results of tests of specimens cut from annealed
bars and plates. In all the specimen tests on rolled and annealed bars and
plates an extensometer was used, reading to 0.0001 in. by micrometer-screw
with electric contact, as shown in Fig. 271. The true elastic limits were there-
fore carefully and accurately determined. This was not done with the heat
or billet tests, so that no comparison can be made on this score. The yield-
points were observed by the " drop of the beam," which is a quite accurate
method with this quality of material if the test is not made too rapidly.
361. The Distribution of the Elongation over an 8-inch specimen 1 inches
in diameter is shown in Fig. 415. This shows that the stretch is nearly
uniform until the maximum load is reached, after which it begins to neck
down. From this on to final rupture the stretch is almost wholly confined
to the necked-down portion. Thus while the elongation near the ends was
only 21 per cent and the average elongation was about 28 per cent, the
elongation at the plane of rupture was 75 per cent.
The Reduction of Area of rectangular specimens is difficult to obtain
accurately, on account of the resulting curved outlines arising from the
greater contraction in the middle portions. This action is well shown
in Fig. 410. If the reduced section is calipered at about one fourth the
width from each side, the area resulting from taking the product of these two
dimensions will be insignificant.
362. The Compressive Strength is the Elastic Limit. Table XXVI con-
tains the results of some of the most careful experiments ever made on the
compressive strength of steel bars (see also Fig. 294). These were made by
Mr. Chas. A. Marshall, C.E.f These results show a practical identity in the
apparent elastic limit, or yield-point, in tension and compression, and also
* In Trans. Am. Soc. Mech. Engrg., vol. xin.
f See foot note p. 341.
THE STRENGTH OF STEEL.
503
TABLE XXV. COMPARISON OF KESULTS OF TESTS OF SPECIMEN" BILLETS
AND OF SPECIMENS CUT FROM FINAL ROLLED FORMS.
TESTS OF BARS.
Size of
Bar.
Kind of
Test.
Billet
Rolled
Annealed
Elastic
Limit.
35.867
37,850
Yield-point
Tenacity.
Per cent
Elonga-
tion in 8"
Per cent
Reduc-
tion in 8"
Modulus of
Elasticity.
30,191,000
29,799,000
31,290,000
No. of
Tests
Averaged
7" x 7/8" A
47,267
39,006
39,083
73.440
71,540
69,990
23.6
25.1
25.3
39.5
51.6
54.2
3
3
3
7"xH" -j
Billet
Rolled
Annealed
33,622
35,060
45,666
39.334
40,140
70,845
71,102
67,930
23.3
23.2
26.3
41.7
44.7
57.0
29,498,000
30,837,000
31,567,000
4
5
2
7"xlA"-j
Billet
Rolled
Annealed
29,650
33,110
44,498
36.620
40,035
70,392
69,750
70,860
23.3
24.1
25.2
41.3
48.9
53.7
30,793,000
29,939,000
31,410,000
3
2
2
7"xli" j
Billet
Rolled
Anaealed
34,725
38,125
45,010
38.255
40,725
71 ,035
72,108
69,800
22.6
22.8
24.6
37.1
40.2
42 .4
29,150,000
29,889.000
31,444,000
2
4
2
7"xlA"j
Billet
Rolled
Annealed
31,550
35,688
46,103
36,635
39,860
69.820
67,795
68,154
24.0
26.0
25.8
41.8
55.0
54.8
29,900.000
30,921,000
31,451,000
3
2
4
7"xl|" |
Billet
Rolled
Annealed
32,890
35,470
45,508
37,480
39,579
70,320
71,561
70,450
24.2
21.6
23.4
42.5
38.5
44.7
30,127,000
30,166.000
30,934,000
14
19
11
r'xi^'j
Billet
Rolled
Annealed
29.606
34,162
46,319
34,882
37,834
71,395
67.560
68.070
23.1
24.2
24.6
38.3
46.0
48.0
29,850,000
30,430,000
30,510,000
10
12
11
,,.{
Billet
Rolled
Annealed
33,730
34,620
47,350
38,000
38,500
71,595
70.090
69,840
22.0
26.2
27.5
37.9
48.8
57.5
29,840.000
30,528,000
30,528,000
1
1
7"xlH"j
Billet
Rolled
Annealed
32,082
37,170
45,698
36,315
39,152
70,665
71,269
69,688
23 7
23.3
25.4
41.5
43.5
50.1
29,500.000
30,759,000
31,268,000
7
11
10
7"xij" |
Billet
Rolled
Annealed
38,100
35,000
44,950
41,040
37,550
63,770
74,060
65,420
25.0
19.8
22.9
40.0
42.0
35.0
28,285,000
31,647,000
30,746,000
1
1
1
7"xlfi"j
Billet
Rolled
Annealed
29,080
32,100
45,740
31.510
36,640
72,330
66,180
73,670
23.6
26.2
23.7
38.1
54.9
50.4
29,455,000
32,479,000
31,302,000
1
1
1
-H
Billet
Rolled
Annealed
29,000
33,650
46,320
32,750
39,060
71,630
71,370
71,280
20.8
25.0
24.5
34.9
52.9
53.7
28,535,000
28,420,000
31.078000
1
1
1
Average -j
Billet
Rolled
Annealed
33.327
35,1(57
45,869
36,819
39,013
71.020
70,365
69,596
23.3
23.9
24.9
39.5
47.5
46.0
29,594,000
30,484,000
31,127,000
Total No.
161
TESTS OF PLATES.
I
Billet
47,700
80.966
20.5
35 8
30.601,000
5
1/2"
Rolled
47,048
48,120
86,572
19.7
36.3
30.061.000
5
1
Annealed
39,523
43,698
75,155
24.7
47.3
29,901,000
9
I
Billet
47,737
82,113
20.6
33.9
29,690,000
6
5/8" 1
Rolled
45.895
48,008
81,753
19.9
40.7
30,763,000
6
\
Annealed
39,792
42,066
74,624
22.1
47.2
30,591,000
9
(
Billet
51 ,936
81,818
21.1
38.8
29,612,000
5
3/4" ]
Rolled
44,692
47,643
79,127
21.8
42.3
30,879.000
5
I
Annealed
38,267
40,067
72,961
22.9
47.8
31,167,000
7
I
Billet
50,130
84,970
19.7
30.8
29,220,000
1
13/16" ^
Rolled
45.580
48,280
88,230
19.0
31.2
29,188,000
1
1
Annealed
41,730
43,590
77,010
25.8
50.8
30,528,000
1
Average -j
Billet
Rolled
Annealed
40,S04
39,828
49,376
48,013
42,355
S2.467
83,920
74,938
20.5
20 1
23 9
34.8
37.6
48.3
29,782,000
30.223.000
30,722,000
Total No.
60
004
THK MATKRIALS OF CONSTRUCTION.
that the ultimate strength of short burs is tho apparent elastic limit. They
also show the lessened elastic limit ami elongation for the larger sixes, all
being rolled from the same billet, the ultimate strength not varying much.
Thus for a reduction of ultimate strength from 3 " to XU in. diameter of
:
U&OL
/ J o 4 S 7 8 9 M //
Fio. 415. Showing tho Distribution of tho tilonpition on (00,000-UO Stool Specimen
Sin. loiii* niul U in. in ilitunotor. (Fr. Com. /&';>., vol. ui. PI. 111.)
A.-l per cent, we have a reduction in the elastic limit of ^0 per cent, and in
the percentage of elongation of HO per cent, the ratio of length to diameter
remaining constant.
363. The Elastic Limit in Compression marks the beginning of lateral
llowing of the metal, and the stress under which this action begins depends
T1IK STltKNUTll OF 8TKKL.
505
TAHLK XXVI. MILD STKKL IN TKNSION AND COMIMtKSSKW.
Comparison of Tensile ami Compressive Results with Results of Tests on Short Columns
of Round ami Square Huts from J in. to '.". in. in diameter, all rolled from one blow
of Hessemer Steel. Elongations ineaau red on a length equal to ten diameters, by
means of the Marshall Kxtensometer shown in Fi#. 271. (From Marshall's Kx-
pcriments, Trans. Am, Soc. C. K., vol. xvn, Tables I and II.)
1 l.i'.! u- Limit in rounds
|T Sijii;iri IlU'll
l liim.n,- si i,-n:-.ili in
Pounds per s.|ii.ni- liu'li
KloiiKUtion.
KV.III.-II..II
Spocimon
in Compi*.
in Cnmpr.
of A roil.
I'eraentage.
iii Tension.
for^.i.
in Tension.
,,;,.,,,
l.cll^lll of
Hpeclnien.
of
Klongullon,
8/4
45,181
45,000
08,711
44,1)70
8
20.4
45.8
48,880
45,855
08,240
48,540
10
25 .
81). 8
Ij.
40,1)08
42,880
07,500
40,455
12
20.4
48.0
11
89,795
42,015
00,51)8
40,150
15
25.4
81). 8
u
89.105
41,225
66,866
81), 700
18
21.8
88.8
2
88,207
89,170
05,008
40,800
20
28 . 1)
27.8
^i
87.055
80,542
05,400
88,080
22
18.7
17.2
2i
80,100
80,840
85,050
25
10.2
Means
40,103
41,129
66,935
40,350
21.9
35.0
Fro. 410. Showini; the Manner in which Rectangular Steel Test-spceimens reduce fn
Cross-section. (Kinjr. ,AVyr.s, vol. xxxm. p. 272.)
on tho freedom with whicli the nietn,! can How l.-itonilly. Thus in Kifj. 417
we luive it column compressed over its full cross-section with freedom to flow
laterally in every direction. 'This is the usual condition under which tho
elastic limit in compression is found.
In Kiij. 1 IS the specimen is compressed uniformly over a portion onlv of
its surface, and when the elastic limit is exceeded the metal finds escape hy
flowing laterally against the resistance of a, ring of unstressed metal. This
is a condition of restricted ilow, and evidently the elastic limit now is much
higher than before.
506
THE MATERIALS OF CONSTRUCTION.
In Fig. 419 only the metal towards the centre of the compressed surface
is constrained to flow under the direct stress, but in attempting to move
laterally it is held by a ring of metal which is confined and compressed ver-
tically, though inside its elastic limit. To find an escape the metal at the
I I
V,
I I
FIG. 417.
Free Flow.
FIG. 418.
Restricted Flow.
FIG. 419.
Confined Flow.
centre must force its way against a much wider ring of metal than in the
second case, and hence the elastic limit now is. very much higher than when
pressed by a flat disk.
The elastic limit in compression, therefore, is a meaningless expression
unless the conditions of lateral flow are also stated.
364. The Author's Tests of Areas of Contact between Car-wheels and
Rails. In Figs. 420 and 421 are shown a series of actual areas of contact
obtained by pressing sections of a cast-iron car-wheel and of a locomotive
steel driving-wheel upon the cylindrical top surface of a steel rail. This
was done in a testing-machine in such a way that there was no rocking
motion and the area of contact was clearly distinguished.*
The areas of these surfaces of contact were determined by a planimeter,
and these are plotted to their corresponding loads in Fig. 422. It will be
seen that these plot in nearly a straight line through the origin. If such a
law be assumed, it follows
1. That the area of contact increases directly with the load.
2. That the mean intensity of pressure is a constant for all loads,
3. That in these experiments this mean intensity of compressive stress,
for all loads, was about 82,000 Ibs. per square inch.
4. Since the maximum deformation (at the centres of these areas) is
twice the average deformation (assuming the volumetric deformation to be
that of a segment of a paraboloid of revolution), then the maximum com-
pressive-stress intensity for all loads is about 164,000 Ibs. per square inch.
5. Since no measurable permanent set was produced by any of these
loads on either wheels or rail, it follows that the "apparent clastic limits " of
the materials had not been reached for this condition of contact, although
the ordinary elastic limit of the rail material, for a free flow as in Fig. 417,
was about 50,000 Ibs. per square inch.
* See a full account of these tests, showing other areas of contact, in Trans. Am. 8oc.
Civ. Engrs., vol. xxxn. p. 270. 1894.
THE STRENGTH OF STEEL.
FIG. 420. Steel Driver, 44 in
diam. Flat tread
FIG. 421. Chilled Wheel,
33 in. diam. New.
508
THE MATERIALS OF CONSTRUCTION.
BQOOO
WflOO
/ V
;/
OF .CIA/TACT tlf SGj. //V
'0-4 &G
FIG. 422. Showing the Relation between the Total Load and the Area of Contact be-
tween "Wheels and Rails. (Johnson, in Trans. Am. Soc. C. E., vol. xxxii.)
/4000
/0,000
4M0&
r\
2,000
'A ML
'V
/A
/6
2 4 6 8 JO tf
FIG. 423. The Elastic-limit Loads per Lineal Inch of Rollers of Various Diameters.
(Crandall and Marston, in Trans. Am. Soc. Civ. Engrs., vol. xxxii. p. 120 (1894).)
THE STRENGTH OF STEEL.
509
These are important conclusions, and should be supplemented and
supported by further observations of this character.
In Fig. 423 are shown the results of tests made by Profs. Crandall and
Marston to find the elastic-limit loads on steel cylinders resting on or
between steel plates. These results show that the elastic loads vary directly
with the diameters, these loads per lineal inch of rollers, for mild structural
steel, being
p = 8SOd, (1)
where p = elastic-limit load in pounds per lineal inch, and d diameter of
roller in inches.
365. The Moduli of Elasticity in Tension and Compression for various
sizes and qualities of steel and wrought iron are given in Table XXVII,
TABLE XXVII. COMPARISON OF MODULI OF ELASTICITY IN TENSION AND
COMPRESSION.*
All results given in one-thousaud-pound units, identical material.
STEEL Tensile Strength less than 100,000
Ibs. per Square Inch.
SPRING-STEEL. Tensile Strength 144,000
Pounds per Square Inch.
Size
of
Bar.
Tension.
Compression.
Size
of
Bar.
Tension.
Compression.
E 1 ,
First
Loading.
E*
Second
Loading.
E l
First
Loading.
E. t
Second
Loading.
EI
First
Loading.
E*
Second
Loading.
E l
First
Loading.
EI
Second
Loading.
1 rd.
& sq.
Ird.
|rd.
Asq.
1 rd.
Ird.
Ird.
Ird.
mean
30,190
29,850
29,280
29,830
29,420
29,550
29,240
29,400
30,000
34,420
29,850
29,500
29,150
29,640
29,630
29.960
30,420
30,370
29,450
28,070
28,780
28,580
28,380
28,680
30,070
28,980
29,260
29,740
29,010
29,420
29,420
28,670
28,830
30,490
29,790
29,810
Ird.
Ird.
T 8 iysq.
T 8 i> s q-
29,480
29,390
28,880
29,200
29,760
29,580
29,420
29,200
28,880
28,880
29,090
29,090
29,300
29,200
29,220
29,350
mean
29,237
29,490
28,985
29,267
29,529
30,371
28,884
29,464
WROUGHT IRON.
Size
of
Bar.
|rd.
Ird.
Ird.
1 rd.
f sq.
| sq.
Isq.
Tension.
Compression.
Size
of
Bar.
Tension.
Compression.
E l
First
Loading.
Et
Second
Loading.
E l
First
Loading.
EI
Second
Loading.
E\
First
Loading.
EV
Second
Loading.
First
Loading.
#2
Second
Loading.
26,800
26,980
27', 540
28,990
27,790
27,800
27.500
27,410
26,700
27,540
29.180
27,900
25.840
25,920
25,670
26,020
27,420
25,650
26,490
26,160
26,240
26,440
26,350
27,790
27,300
IS
1 rd.
Ird.
frd.
28,290
27,590
28,290
26,580
30,190
28,290
28,570
28,480
28,480
30,190
27,100
27,250
27,430
26,500
29,520
26,734
27,990
28,570
28,570
28,180
29,910
mean
27,894
28,203
27.590
* From experiments by Charles A. Marshall, M. Am. Soc. C. E., reported in Trans.
Am. Soc. Civ. Engrs.,vo\. xvn. pp. 62-3.
510
TEE MATERIALS OF CONSTRUCTION.
these results also being from Marshall's experiments. They show the rela-
tion between the modulus of elasticity as obtained from the first and from
the second loading, the first loading not having been carried beyond the
elastic limit. As the modulus from the second loading is always a little
larger than that obtained from the first loading, it shows that on the first
loading there is always a small permanent set, and that moduli of elasticity
should be observed only after a load has been imposed and removed. The
4Q000
8QM0
O .0005 .OO /
.0005 .00/ i00/5
FIG. 424. Showing Variations, in the Modulus of Elasticity of Steel Eye-bars, 66,000
Ibs. T. S., after the Elastic Limit has been Passed. ( Wat. ATS. Rep. 1883.)
failure to do this may explain some of the low values of this modulus which
are often given.
If the specimen be stretched much beyond its elastic limit, however, the
elastic limit is lowered after each such higher loading, as indicated in
Fig. 424.
366. Modulus of Elasticity Independent of the Other Mechanical Prop-
erties. In Table XXVIII are given the average values of the moduli of
elasticity from 262 determinations on steels of five degrees of hardness.
The mean values for these five classes do not in any case differ from the
mean of all by more than six tenths of one per cent. As the mean of all is
THE STRENGTH OF STEEL.
511
TABLE XXVIII. MODULI OF ELASTICITY OF STEEL ON FIRST LOADINGS,
WITH VARYING PERCENTAGES OF CARBON, ONE SPECIMEN FROM
EACH HEAT.*
Number
of Heats
and Tests.
Average
Percentage
of Carbon.
Moduli of Elasticity E, in Pounds per Square Inch.
Kind of Steel.
Lowest Value.
Highest Value.
Average Value.
33
8
107
89
25
9
11
24
34
72
28,750,000
29,210,000
28,310,000
28,140,000
28,680,000
31,540,000
30,670,000
31,180,000
30,910,000
30,860,000
29,924,000
30,020,000
29,996,000
29,672,000
29,919,000
29,866,000
Bessemer
Open -hearth
Bessemer
Open-hearth
Weighted mean value =
* From Marshall's Experiments, 2'rans. Am. Soc. C. E., vol. xvn. p. 64.
TABLE XXIX. TENSILE TESTS ON ROUND STEEL RODS FROM 1 TO 3
INCHES IN DIAMETER, ANNEALED AND UNANNEALED.
Each recorded result is the mean of three tests. All the results in one horizontal line
are for tests on material cut from the same three bars.f
Size.
Elastic Limit $ in Pounds per
Square Inch.
Ultimate Strength in Pounds
per Square Inch.
Ratio of Elastic Limit to
Ultimate Strength.
Unannealed.
Annealed.
Unannealed.
Annealed.
Unannealed.
Annealed.
Diam.
in in.
1
?
?
Rods
100 in.
long.
Rods
10 in.
long.
Rods
10 in.
long.
Rods
100 in.
long.
Rods
10 in.
long.
Rods
10 in.
long.
Rods
100 in.
long.
Rods
10 in.
long.
Rods
10 in.
long.
72.7
67.5
62.0
60.3
58.3
43,330
42,400
36,520
34,130
35,700
46,970
43,300
39,570
37,230
36,530
45,130
42,170
36,270
34,300
33,500
63,810
62,507
61,320
58,950
58,550
66,050
65,020
61,420
60,300
59,830
62,010
62,460
58,490
56,790
57,370
67.8
67.8
59.5
57.8
60.9*
71.1
66.5
64.4
61.7
61.0
Size.
Diameter
in inches.
1
i*
f
Percentage of Elongation.
Percentage of Reduction.
Modulus of
Elasticity of the
Unannealed.
Annealed.
Unannealed.
Annealed.
100-inch Bars Un-
annealed in
Pounds per
Rods 100 in.
Rods 10 in.
Rods 10 in.
Rods 100 in.
Rods 10 in.
Rods 10 in.
Square Inch,
long
long.
long.
long.
long.
long.
First Loading.
19.21
25.6
22.1
60.2
61.1
65.7
27,300,000
21.42
26.9
24.9
55.3
55.0
58.4
21,100,000
25.62
30.9
30.4
58.8
59.4
62.5
30,000,000
23.50
31.4
32.5
56.9
54.9
62.7
30,600,000
17.34
30.6
33.9
54.6
49.8
61.2
28,400,000
From Kirkaldy's Report, 1891, reports M and HH.
This is the true elastic limit on the first loading ; it was about 5 per cent below the
yield-point, or the "apparent elastic limit."
512
THE MATERIALS OF CONSTRUCTION
THE STRENGTH OF STEEL.
513
29,866,000, and this from first loadings, it will not be appreciably in error
to call the modulus of elasticity of steel 30,000,000. (See also Figs. 398 to
403, where tensile stress indicated for an elongation of 0.001 is uniformly
about 30,000 Ibs. per square inch, thus giving a modulus of elasticity of
30,000,000.)
367. The Effect of Annealing on Steel Before and After Overstraining.
In Table XXIX are given results of annealing steel bars which have never
been stressed or worked cold. In Table XXX are given the results of tests
FIG. 426. Effect of Length of the Reduced Section on the Strength and Ductility of
Steel. James E. Howard (in charge of Tests at the Watertown Arsenal) before Inter.
Eag. Cong. 1893. Section of Nau, Eng. and Marine Arch., vol. n, J. Wiley & Sous,
New York.
on overstrained steel which was afterwards annealed. It will be observed
that, while it has little effect on soft steel in the normal state, the anneal-
ing largely restores the original qualities to overstrained soft steel, though
it does not fully do so. The high moduli of elasticity in compression given
in Table XXX for both the annealed and the unannealed specimens should
be accepted with caution as being probably erroneous.
514
THE MATERIALS OF CONSTRUCTION,
368 The Effect of Varying the Length of the Reduced Section.* " The
following results of the tests of six specimens from the same 1^-in. steel
bar illustrate the apparent elevation of elastic limit and the changes in other
properties due to changes in the lengths of stems which were turned down
in each specimen to 0.798 in. diameter. (See also Fig. 426.)
Description of Stem.
Elastic Limit,
Pounds per
Square Inch.
Teasile Strength,
Pounds pei-
Square Inch.
Contraction of
Area,
Per Cent.
1 00 ' long
64 900
94,400
49.0
*50 "
65,320
97,800
43.4
25 . ....
68,000
102 420
39.6
Semicircular srroove 4" radius
75 000
116 380
31 6
Semicircular srroove 1/8'' radius
86 000 about
134,960
23
90,000, about
117,000
Indeterminate
'
" These tests show the progressive elevation of the elastic limit as the
stems of the specimens were shortened, and the corresponding effect upon
the tensile strength. The contraction of area, of course, diminishes as the
other two features increase in value.
" The lower tensile strength of the specimen having the V-shaped groove
was probably due to the excessive concentration of stress at the bottom of
the groove from inability to elongate or contract, fracturing the metal more
in detail than happened to the other specimens/'
In Fig. 427 are shown the results of similar tests made by M. Duguet on
hard steel and by M. Barba on soft steel bars. In both of these sets of ex-
periments the very short reduced sections have a greatly increased breaking
strength.
In this connection it must be remembered that in ductile metals, where
the reduced 'section has appreciable length, there is a great reduction of
area, so that the stress per square inch at rupture on the actual section at
that time is about twice the tensile strength as computed on the original
cross-section. In the very short or grooved reduced sections, however, the
material has no opportunity to reduce in area, and hence the actual ruptur-
ing stress is developed over the full original area. In the case ot the sharp
V-shaped groove the material is likely to tear apart by failing first at the
outer edges- In other words, the stress is not uniformly distributed over
the cross section.
In Fig 428 are plotted the results of tests of the same grade of steel
(54,000 Ibs tensile strength), when tested in the standard form, with par-
allel sides, and when grooved as has long been required for the U. S. Marine
Service The effect of the groove is to raise the tensile strength from 7000
* Quoted paragraphs and table taken from a paper by James E. Howard read before
the World's Engineering Congress, 1893.
THE STRENGTH OF STEEL.
515
Ibs. per square inch on the f inch plate to over 12,000 Ibs. per square inch
on the 1-inch plate. The grooved specimen, furthermore, gives little or no
4000
80000
7000
S0.000
/T0S-ff0t/M0 SfCf/OA/S*
.
33
fff
FIG. 427. Showing the Effect of the Form of the Reduced Section on the Tensile
Strength of Two Kinds of Steel. (Fr. Com. Rep,, vol. in, p, 40.)
indication of the elastic limit, and no indication of the percentage of elonga-
tion. The requirement of grooved specimens on this service will probably
soon be abandoned.
369. Nickel-steel,* being an alloy of mild steel with about 3|- % of nickel,
has a very high elastic limit and ultimate strength, combined with great
ductility, as shown in Fig. 429. This alloy is doubtless destined to play a
* First made by Marbeau in 1885, and used for armor-plate in 1890. The price of
nickel steel was 40 to 45 cents a pound in 1894. For a complete study of the influence
of nickel on pure iron, in all proportions, see Berlin Testing Laboratory Communica-
tions, 1896, vol. iv. p. 222.
516
THE MATERIALS OF CONSTRUCTION.
5000
5500
\
es
60
55
50
FIG. 428. Showing Relative Results from Grooved and Parallel-sided Specimens
(Campbell's Structural Steel, p. 223.)
4004 &00S
FIG, 429. Tension Tests on Nickel-steel (Wat. Ars. Rep. 1894, pp. 199 ml VQO )
THE STRENGTH OF STEEL.
517
leading part wherever great elastic strength and a reasonable ductility are re-
quired. It would seem to be especially fitted for bicycle tubing and spokes,
aerial experimentation, the reciprocating parts of locomotive engines, motor
carriages, etc., as well as for armor-plates.
370. The Mechanical Properties of Steel as Affected by Forging and
Rolling. In Fig. 430 is shown the cross-section of a steel shaft 16 inches in
diameter (which broke soon after being put in service) from which eight test-
0006
0./S
0:30 0.35
0.20 0.2S
FIG. 430. Showing the Varying Character in the Material in Different Parts of the
Cross-section of a Large Steel Shaft when forged under a Ten-ton Hammer. ( Wat.
Ars. Rep., 1885.)
specimens were cut, lying symmetrically in a diametral section as shown.
Four of these were tested as cut from the shaft. The other four were forged
down after cutting out. The plotted results show*
1. The elongation of the unforged specimens varied from 21 per cent in the
518
THE MATERIALS OF CONSTRUCTION.
specimen taken from near the surface of the shaft to % per cent in the specimen
U-- 80- ____ 4 coming from near the centre. In the forged specimens,
however, taken from the opposite side of the disk, the
elongation varied from 28 per cent near the surface of
the shaft to 24 per cent near the centre, thus showing
that the material was identical throughout when it had
been similarly worked. In other words, the material
near the centre of the shaft was in its primitive con-
dition when first cast, while that near the surface was
that of well-rolled steel. This shows the necessity of
forging large shafts under enormously heavy hammers,
or, better, the necessity of using only hollow-forged
shafts for such service.
Steel car-axles are now rolled and then finished by
drawing through a die. In Fig. 431 are shown the
effects of a series of 16 blows upon a steel car-axle, 3f
in. in diameter, turned over after each second blow, of
a drop weighing 1640 pounds and falling from 18.8 to
28 ft. The deflections from each blow varied from 8-J-
to 13 in., but the -axle remained unbroken after this
severe treatment.
371. Steel- welded Tubes. In Tables VII and VIII,
pp. 130 and 131, it was shown that steel may be welded
as securely as wrought iron, but that the temperature
at which this material welds perfectly lies within com-
paratively narrow limits. If heated above the upper
limit the steel melts and oxidizes, and it is then said to
have been burned. If not heated up to the lower limit
= r- "7=3 an i m P ei 'f ect weld is formed. To effect a good weld
FALL OF 28 FT in a common blacksmith's forge therefore requires
FIG. 431. Showing the great ^^\\ an( j care> Qn the other hand, where steel
plates USed f r tube ' makin g ai ' e iformly heated in
amtee
Car-axle when tested a f urnace i ]1 which the temperature can be main-
by Impact without tained constant and of a given degree, these may then
Rupture. Wt. of be welded perfectly, especially if this be done by ma-
drop == 1640 Ibs. chinery. The evidence of such perfect welds is fur-
nished in Figs. 432 and 433, where welded steel tubes
are shown to have been subjected to the most severe abuse, by cold crushing
and twisting, and without any failure appearing in any of the welded joints.
372. Wrought-iron and Steel I Beams and Plate Girders. When rolled
into I beams, or when plates and angles are riveted together to form a plate
.girder, the true elastic limit of the beams and girders is below that of the
specimen test-pieces cut from the webs and flanges. The ultimate strength
of the wrought-iron beams and girders is higher than that of the specimens,
THE STRENGTH OF STEEL.
B ^
<-* O
520
THE MATERIALS OF CONSTRUCTION.
FIG. 4:53. P]ximples of Welded Steel Tubes twisted Cold, made by the National Tube
Works. (From The Iron Age, Sept. 17, 1896.)
0000
S0000
0.2S 0JO 0.75 /.00
FIG. 434. Bending Tests on Steel and Wrought-iron I Beams, 7.6 in. high, on 60-iu.
span. (Tetmajer, vol. in, PL IV.)
FIG. 435. Showing Variation in Moduli of Strength and Stiffness of Steel I Beams with
Varying Depth. (Tetmajer's Communications, vol. in, PI. V.)
THE STRENGTH OF STEEL.
521
but with the steel beams and girders the reverse is the case. All these
relations are shown by the following tables, which are the results of a series
of very careful tests made by Professor Tetmajer. The moduli of elas-
ticity of the rolled or riveted forms are not appreciably lower than those of
the materials of which they are composed; but this modulus and also the
moduli of strength seem to decrease with increasing heights of beam, as
shown in Fig. 435.
TABLE XXXI. ELASTIC LIMIT AND ULTIMATE STRENGTH OF WROUGHT-IRON
I BEAMS AS COMPARED WITH RESULTS OF TESTS OF SPECIMENS CUT
FROM THE WEBS AND FLANGES OF THE SAME.
(Each result is the mean of two tests. From Prof, von Tetmajer's Communications,
vol. iv.)
Depth
of
Elastic Limit in Pounds per
Square Inch.
Ultimate Strength in Pounds per
Square Inch.
Per Cent of
Elongation.
Modulus of
Elasticity of
Beam
I Beams in
in
Inches.
Web
Speci-
men.
Flange
Speci-
men.
Beam,
Extreme
Fibre.
Web
Speci-
men.
Flange
Specimen.
Beam,
Extreme
Fibre.
Web
Speci-
men.
Flange
Speci-
men.
Pounds per
Square Inch.
4
40,520
42,230
35,120
54,740
57,300
62,850
10.4
19.1
28,600,000
6
38,680
36,400
33,270
51,330
53,750
56,450
7.8
25.5
28,300,000
8
35,690
34,840
33,840
50.900
51,900
53,890
11.7
18.5
28,300,000
10
36,120
34,410
31,000
48,490
51,330
51,620
13.1
158
27,500,000
12
31,000
32,280
31,570
44,930
52,610
53,180
9.1
19.7
26,500,000
14
33,840
33,700
27,370
51,470
53,460
53,890
17.9
20.5
27,700,000
16
33,130
31,140
29,720
50,480
50,190
52,470
11.9
22.9
27,600,000
Means
35,572
35,000
31,770
50,334
52,934
53,478
11.7
20.2
27,800,000
TABLE XXXII. ELASTIC AND ULTIMATE STRENGTH OP WROUGHT-IRON
PLATE GIRDERS COMPOSED OF A SOLID WEB, FOUR ANGLES, AND TWO
COVER-PLATES, AS COMPARED WITH THE TENSILE STRENGTH OF THE
PARTS COMPOSING THEM.
(Each result is the mean of tests on two beams or on four tension specimens. From
von Tetmajer's Communications, vol. iv.) *
Test-specimen.
Elastic
Limit in
Pounds per
Square
Inch.
Ultimate
Strength
in Pounds
per Square
Inch.
Percentage
of Elonga-
tion in
8 Inches.
Percentage
of
Reduction
of Area.
Modulus of
Elasticity of
Girders in
Pounds per
Square Inch.
Web-plate lengthwise
crosswise .
40,950
36,120
35,120
34,690
53,320
37,400
51,040
48,350
14.1
0.5
13.9
8.4
16.2
0.4
17.0
14.6
Cover-plates lengthwise
Angles lengthwise
Plate girders 16 in. high
25,590
29,860
52,330
52,040
50,620
47,210
....
i
25,990.000
25,250,000
26,430.000
26,220,000
20 in. high
" " 24 in high . .
" " 28 in. hifh
27,580
Mean of the parts
38,720
27,680
47,050
50,550
9,2
11.8
Mean of the girders
25,970,000
522
THE MATERIALS OF CONSTRUCTION.
TABLE XXXIII. ELASTIC AND ULTIMATE STRENGTH OF MILD-STEEL PLATE
GIRDERS COMPOSED OF A SOLID WEB, FOUR ANGLES, AND TWO COVER-
PLATES, AS COMPARED WITH THE TENSILE STRENGTH OF THE PARTS
COMPOSING THEM.
(Each result is the mean of tests on two beams or on four tension specimens. From
von Tetmajer's Communications, vol. iv.)
Test-specimen.
Elastic
Limit in
Pounds per
Square
Inch.
Ultimate
Strength
in Pounds
per Square
Inch.
Percentage
of
Elongation
in 8 Inches.
Percentage
of
Reduction
of Area.
Modulus of
Elasticity of
Girders in
Pounds per
Square Inch.
52,330
53,040
51,190
39,960
32,560
35,690
33,700
32,280
64,560
65,980
64,840
53,750
24.1
20.0
25.0
31.0
59.5
46.2
562
66.1
" crosswise
Plate girders 16 iu hi *!! ....
55,310
54,320
55,030
53,750
27,700,000
28,310,000
28,510,000
28,140,000
" 20 in high
" " 24 in high
" < 28 in hi'h
Mean of the parts
49,130
33,550
62,280
54,600
25.0
57.0
^Vlean of the girders
28,165,000
373. The Effect on Mild Steel of Stressing it Beyond its Elastic Limit.
Both wrought iron and rolled steel, in their normal state, have " apparent
elastic limits" in tension and in compression numerically about equal. If
this material be stressed much beyond these limits, however, in either direc-
tion, its elastic limit in this direction is numerically raised to about the limit
of its greatest stress, while the elastic limit in the opposite direction is
greatly lowered or even reduced to zero. This is well shown in Figs. 436
and 437. Thus in Fig. 436 a steel specimen was stressed several times in
tension between zero and 45,000 Ibs. per square inch, when it was put in
compression to 50,000 Ibs. per square inch, at which load it passed its elastic
limit and began to flow. After compressing it 0.007 of its length the load
was removed and simultaneous readings of load and deformation taken
while the load was coming off, as shown in the figure. The specimen was
then worked between 10,000 Ibs. tensile and a like compressive stress, then
between zero and 20,000 Ibs. tension, and zero and 40,000 Ibs. tension, and
finally between zero and 50,000 Ibs. per square inch tension, when it had
lengthened 0.004 beyond its original length as shown in Fig. 436. It was
then put in compression under a load of 50,000 Ibs. per square inch, which
compressed it 0.004 below its original length, while a subsequent tensile
stress of 50,000 Ibs. brought it nearly back to its previous deformed length
under this tensile stress. The diagram shows the following remarkable facts:
1. A permanent deformation of one half of one per cent in either tension
or compression entirely destroys the perfect elasticity of the material under
the opposite kind of stress.
THE STRENGTH OF STEEL.
523
This is shown by the fact that the stress-diagram becomes a curved line
under all stresses of one kind after having been given a small permanent set
in the opposite direction. Hence we have :
2. The elastic field which is symmetrically placed about the line of zero
stress in the normal specimen becomes wholly limited to that side of this axis
on which the stress has exceeded the elastic limit.
A similar set of experiments, plotted in Fig. 437, was followed by the
annealing of the bar, after which it showed again its normal elastic limits in
524
TEE MATERIALS OF CONSTRUCTION.
both tension and compression, which were in turn again destroyed by
deforming the annealed bar 0.003 beyond its elastic limits, as before.
Hence we may say :
3. Annealing an overstressed liar restores it fully to its normal condition
of perfect elasticity in both tension and compression.
Both of these diagrams are very instructive and will bear close study.
Many more such could be plotted from the tabulated results found in the
BEfQ/? ANNEAL/ A/G
JFTEf? AMNEALM&
FIG. 437. Alternate Tensile and Compressive Distortions of a Steel Bar before and after
Annealing. (Wat. Ars. Rep. 1889.)
Reports of the Watertown Arsenal, from which the data for these were
obtained.
The effect on the ultimate strength of 60,000-lb. steel of pulling it
nearly to final rupture is shown in Fig. 43 7 a. This also gives some idea of
the homogeneity of the steel. Here, after the specimen had necked down
under the tensile load, but before it broke, it was removed, a continuous
screw-thread cut on it, and a series of grooves were cut in it as shown. The
bar was then broken in tension at all these grooves in succession, with the
results as shown in Fig. 43 la. The original tensile strength was about
57,000 Ibs. per square inch, the final breaking stress on the grooved sections
was about 100,000 Ibs. per square inch, while the final stress on the groove
placed at the centre of the necked-down portion was 155,000 Ibs. per square
inch. A portion of this increase in strength is due to the normal difference
between the strength of a grooved and of a parallel-sided specimen, as
shown in Fig. 427. After allowing for this difference there still remains a
great increase of strength due to the previous drawing out and the interven-
ing rest the specimen had experienced.
THE STRENGTH OF STEEL.
525
The results of similar tests on unstressed or normal bars are shown in the
original plate from which Fig. 437# was taken, which go to show that grooved
sections on the same steel bar may develop breaking tensile stresses which
differ from each other by as much as 20 to 25 per cent. No such differences
ffffoaf Tfsr/A/0
I
t
<^r
"*5
69
FIG. 437.
would be observed on specimens with parallel sides cut from the same bar,
and hence we must conclude that results of tests on grooved sections of steel
are very erratic and unreliable.
374. Shearing Resistance of Steel. Prof. A. B. W. Kennedy, by means
of the apparatus shown in Fig. 311, obtained values of tensile and shearing
resistances of various steels given in Table XXXIV.
TABLE XXXIV. SHEARING RESISTANCE OF STEEL.*
Tensio
a Test.
Shearing
Ratio of
Kind of Steel.
Number
of Tests
Averaged.
Elastic Limit,
Pounds pei-
Square Inch.
Ult. Strength,
Pounds per
Square Inch.
Strength,
Pounds per
Square Inch.
Shearing to
Tensile
Strength.
Landore Siemens steel..
Weardale Bessemer "
Bessemer steel . .
2
2
3
6
6
4
37,500
40,000
37,000
40,600
44,000
5 1 500
57,000
63,500
64,000
69,000
71,000
78 000
47,500
51,000
52,000
56.000
51,000
64 000
0829
0800
0.811
0.807
0.715
823
Crucible " ..!!..!..
4
2
62,000
69 500
82,000
116 000
59,000
74 000
0.721
632
Bessemer '' ...
2
70 000
118 000
79 000
670
* From Proc. Inst. Meek, Engrs. 1885, p. 262.
From the above table, which is simply corroborative of a vast amount of
similar data, we may reasonably use 0.8 as the ratio of shearing to tensile
strength of mild or structural steel.
526
THE MATERIALS OF CONSTRUCTION.
375. The Frictional Resistance of Riveted Joints.* The contraction of
rivets in cooling is always much more than their elastic stretch. Thus if the
modulus of elasticity be taken at 30,000,000, and the elastic limit of rivet-
steel at 30,000 Ibs. per square inch, then the elastic stretch is 0.001 of the
length. But as the contraction per degree F. is O.OOOOOG5, it follows that
( -\ = 154 F. change of temperature would bring rivets to their
VO.OOOOOG5J
elastic limit if they were not allowed to contract. Evidently, therefore, all
well-driven rivets in plates which are tightly clamped together when the rivet
o
'0L02 0.04
FIG. 438. ^ FIG. 439.
FIG. 438. Showing the Successive Stages of the Slipping of Riveted Plates.
(M. Dupuy in An, d. Ponts et Chaussees, 1895.)
FIG. 439. Autographic Stress-diagram of a Double-strap Butt-riveted Joint with oue
rivet on each side of joint, showing slips at aa' and bb'. (An. d. Ponts el Chaussees,
vol. ix, 1895.)
is driven, and so held till the rivet cools, are left in a state of tension exceed-
ing their elastic limits. If the coefficient of starting friction be taken at
0.4, and the elastic limit of steel rivets be taken at 30,000 Ibs. per square
inch, and of iron rivets at 25,000 Ibs. per square inch, it follows that the
frictional resistance would be 12,000 Ibs. per square inch of rivet section for
* Riveted joints are a kind of structure the strength and design of which do not fall
within the scope of this work. The elements of the strength of such structures, however,
are properly treated here.
THE STRENGTH OF STEEL, 527
steel rivets, and 10,000 Ibs. per square inch of rivet section for iron rivets,
in lap-joints, and twice these amounts for butt-joints with two cover-plates,
since in that case there are two frictional surfaces on each of which the full
tensile stress in the rivets acts. Theoretically, therefore, we might expect a
frictional resistance of 12,000 or 24,000 Ibs. per square inch of rivet-surface
for lap and butt joints respectively when the rivets are of steel, and of
10,000 or 20,000 Ibs. when the rivets are of iron.
Since the plates are clamped together much more firmly when steam or
hydraulic riveting-machines are used than when the rivets are driven by
hand, so the experiments show a much higher frictional resistance for
machine-driven rivets. To secure the greatest frictional efficiency the
machine pressure should remain on until the rivet has cooled, but in ordinary
commercial work this is seldom done.
M. Dupuy has carefully and fully investigated this question.* He shows
that after cooling the rivet does not fully fill the hole, as shown in Fig.
438 (1). The first slip, therefore, when a butt-joint has one rivet on each
side as shown in Fig. 438 (2), is that which brings the centre plate against
the rivet. This is shown at a and a' in Fig. 439, this being a reduced auto-
graphic stress-diagram for the joint shown in Fig. 438. The slip at a
occurred under a load of 17,500 Ibs. per square inch of rivet area when the
centre plate came up against the rivet on one side of the joint, and the slip
at a' marks a similar movement on the other side at 18,500 Ibs. per square inch
of rivet area. After these movements had occurred the load was increased to
28,000 Ibs. per square inch of rivet area, when the rivet-heads slipped on the
cover-plates, as shown in Fig. 438 (3), and on the stress-diagram in Fig. 439
at b. Soon after the same action occurred at the other rivet, marking the
deformation at V in Fig. 439. All these four slips were sudden, and were
accompanied by a sharp report like a pistol-shot. After the spaces had all
closed up in this way the deformation was gradual, and the rivet would
then be acting as a bolt, aud be subjected to a shearing stress and deforma-
tion as shown in Fig. 438 (4).
376. The Stresses per Square Inch of Rivet Section at which the First
Slipping Occurs, as determined by M. Dupuy, with an extensometer
(called by him an elasticimeter), are given in Fig. 440. They are presented
in this form in order that the relative frictional efficiencies of different
methods of riveting may be read at a glance. These results were obtained
by cutting the plate along the centre line of the rivets and then pulling out
the two halves of these rivets as indicated in the figure. In this way the
frictional resistance of the rivet-heads was correctly obtained without any
complication with bearing or shearing resistance, as must always be the case
when pulling actual riveted joints.
From tests on riveted joints made at the Watertown Arsenal (1882) we
find the following values of frictional resistance on plates with elongated
holes, hand-riveting:
* ILI An. d. Fonts et Cliaussees, 7th series, vol. ix, 1895
528
THE MATERIALS OF CONSTRUCTION.
Frictional Resistance
on one Surface in
Pounds per Square
Inch of Rivet Area.
Steel plates, f-in. iron rivets, If -in. grip, lap-joint, 4 tests 14,550
Iron " |-in. " " 1-in. ' " " 4 " 14,100
f-in, " " 1-in. " butt- joint, 2 cover-plates, 2 tests 9,000
It thus appears that the frictional resistance is not twice as much on a
butt-joint having two cover-plates as it is on a lap-joint. The reason may
/000\
P/?L~SS. C#A/T'0. X/VET Wtf/TE //Pf
PEL/EVED
CONT'D.
RED
tf/VET/NG
WH/TE
ff/l/ET COLO
FIG. 440. The Slipping Resistance of Steel Rivets in Pounds per Square Inch of Rivet
Cross-section for Various Conditions of Driving. Each result is the overage of 25
tests on single rivets from inch to 1.2 inches diameter. Plates and rivets cut on
the diametral lines and each half of rivet pulled out as shown. (M. Dupuy in An.
d. Fonts et Chaussees, 1895, p. 105.)
Jt
.02" .04" "36" .Of
FIG. 441. Stress-diagram of Test of Double-butt-strap Riveted Joint. Steel plates 0.662
in. thick and 20 in. wide. Thirteen f-in. steel rivets, machine-driven, on each side
of joint. Drilled holes f in. diameter. (Wat. ATS. Rep. 1887, p. 892.)
be that the distortion of the lap-joint increases the frictional resistance by
putting an additional tensile stress on the rivet from the bending of the
plates.
THE STRENGTH OF STEEL.
529
Since the frictional resistance is thus seen to depend directly upon the
total shearing area of the rivets, whether these be in single or in double
shear (although in double shear the frictional resistance on each bearing
surface seems to be less than in single shear), there would seem to be no
advantage in designing riveted joints for frictional resistance. It seems
/QOOO
.08 M .06" .08"
FIG. 442. Stress-diagram of a Test of a Double-butt-strap Riveted Joint. Steel plates
in. thick and 16.5 in. wide. Eleven 1-in. steel rivets, machine-driven, in three
rows on each side of joint. Straps | in. thick. Drilled holes ly 1 ^ in. diameter.
(Wat. Ars. Rep. 1887, p. 901.)
probable, however, that all riveted joints in practice do their work through
friction alone, and that in no case are the rivets subjected to either shearing
or bearing stress. But when the joint is dimensioned for shear, it is likely
to be also properly designed for frictional resistance. In the case of double
shear, riveted joints are usually proportioned for bearing stress, and here it
FIG. 443. Showing Manner of Failure of a Triple-riveted Steel Plate. Figures indicate
thickness of plate at that place. (Wat. Ars. Rep. 1887.)
would seem to be proper to make due allowance for the frictionai resistance
which, for all working loads, will at least greatly reduce the bearing stress.
The frictional resistance of joints containing double or triple rows of
rivets cannot be observed, because all the rivets do not draw with the same
tensile force, and hence the slipping is progressive and without any sudden
manifestation. This is well shown by Figs. 441 and 442, which are charac-
teristic of a great many such tests made at the Watertown Arsenal. Evi-
530
THE MATERIALS OF CONSTRUCTION.
dently it would be impossible here to locate the point of initial slipping.
This explains Prof. Kennedy's discrepant results on this class of joints, as
recorded in the Proceedings of the Institution of Mechanical Engineers
(London) for the years 1885 and 1888. He here records the loads for
which " visible slip began"; but as he used only a hand magnifying-glass,
and the movement was a gradually progressive one, it would be quite
impossible to obtain consistent or rational results. In fact, for such joints
no such stage of the test exists, since the slipping does not occur over the
entire joint at any one time.
377. The Bearing Resistance of Steel and Iron Plates is shown in Fig.
444. This is seen to increase directly with the distance of the hole from
4000A /$>
&
M
.pW/cf /A
f
1 /ffYlfJ O.
S.00
2:00
FIG. 444. Bearing Resistance on Rivet-holes at Rupture by Tearing Out of Hole. The
steel plates were of 60,000 Ibs. tensile strength. ( Wat. Ars. Rep. 1882.)
the edge of the plate. When this distance agrees with ordinary practice
the resistance is so high that it would seem a working bearing stress of
THE STRENGTH OF STEEL.
531
16,000 Ibs. per square inch might be employed for iron, and of 24,000 Ibs.
per square inch for steel, plates. The stresses here plotted were the bearing
stresses at rupture, where the plates had so reduced in thickness as to
:
. CD : o j
. CO = O C7I
1
TABLE XXXV. ^
VAVA
5"p
LBSOLUT
SOS OS OS
O JO JO
,^?
1
CO
M
>
fc-4
Or OS OS OS OS
OO t 1 O H- i i >
>" t-i o CO or
GO OS tO O GO
o o o o o
^
cr
d
"H<
w
&
t- 1
~3 JO j- _CD
OS CO GO CO
rf*. CO O tO
^
c
t**
H
M
II
rf. io bs to
^
II
II
a
O
H
c
<
^-
8
I
j
o'
o
cr
O
^j
H
W
II
CD CO . CO
CD GO GO
Lj bs co co
3
to
W
>
d
cr
1
1 GO -5 GO
GO OS tO O
05 CD CO
H,^
Ms
i
c
W
o
a
O5 --3 OS --3
-5 P J^ 1 J" 1
to to o >- 1
8
B X9
&
1
CO
1 I
*2j
OS *-3 OS OS *^?
i CO J30 CO O
CS -7 IO O O
0000
>.
g
5'
^
S
d
\~*
JO tO CC JO
-^ w? GO 3
to tO GO tO
o o o o
?
c
D.
o
o
0!
H^
-
to co co jt>.
b bs rf^ CD
s
1
ameter
c/:
rC
(1
h- 1 h- ' 1 1 I '
-J- ,1 ^ H- ^ 1 I
o o to *-
U ^ to b
i
II
II
K>
R
SCO CD
H-* ^ OT J-
^ CT tO k^
^
i
o
a
o
5-
!z5
o
W
O CO O CD
OS OS O CD
Hfx CO CO tO
M^
II
p
3
P c $ 8
bo to co io
s 1"*
i
w
tzi
CO
536 THE MATERIALS OF CONSTRUCTION.
the corresponding screw-threads, leaving the same diameter at the bottom
of the groove as obtained at the base of the threads. The actual and com-
parative average results of all of these tests are given in the following table,
from which the following conclusions may be drawn:
1. When subjected to plain tension both the screw-threads and the
grooved sections were stronger than the plain bars of the same net area of
cross-section, this excess of strength having an average value of about 14 per
cent. This excess of, strength is due to the re-enforcing action of the
shoulder in the case of the groove, and of the threads themselves in the case
of the screw.
2. There is no very marked difference in the average strength of the
bolts on which the several styles of thread were cat, the perfectly sharp
groove shown at A being slightly stronger than the others.
3. The weakening effect of the turning of the nut under stress at rup-
ture is much less than might have been predicted, when the distortion of
the screw below the nut by permanent elongation is taken into considera-
tion. The tests indicate for this case a strength of the one-inch bolts about
20 per cent less than that of the plain bars, and of the one-half-inch bolts
about 15 per cent less than that of the plain bars.
4. In general it may be said that the turning of the nut upon the bolt at
rupture reduces the strength of the net section of the bolt by about 30 per
cent.
5. It is very probable that the four forms of screw-threads here shown
would show, very different results under fatigue tests from repeated stresses,
and also for static loads on high-carbon steel. Under repeated loads and
under shock it is probable that the sharp re-entrant angle shown in Fig. 450 A
would develop incipient cracks much earlier than either of the other forms,
and that probably the Whitworth thread, shown in B, would be the last
to develop this kind of weakness, either with soft metal under repeated loads
or with high-carbon steel under static loads. No such tests have as yet been
made. It is to be hoped that this subject' will soon be investigated, as it is
of far more importance than the mere matter of static strength.
CHAPTER XXVII.
THE FATIGUE OF METALS.
383. Fatigue Defined. It has been found from experiment that metals
will fail under loads much less than their ultimate strength when such loads
are repeated or reversed many thousands or perhaps millions of times. It
has been commonly supposed that these repetitions or reversals caused a
general deterioration of the metal so stressed, so far as its cohesion is con-
cerned, which deterioration has been known by the term fatigue. It is now
known, however, that no such general deterioration takes place, but that
some of the millions of incipient defects or "micro-flaws" in the spec-
imen gradually extend their weakening influence, in an irregular plane of
cross-section, which ultimately becomes the plane of rupture, while the metal
immediately adjacent to this plane remains perhaps wholly uninjured. In
fact no tests of metal, on specimens as closely adjacent to such planes of
rupture as it is possible to procure them, have ever shown any deteriorating
effects of the repetitions or reversals of stress to which this metal had been
subjected. The word " fatigue," therefore, is scarcely the proper term to
apply to this class of failures. The gradual fracture of metals would be a
more truly descriptive term to use.
384. The Micro-flaws in Steel have been studied exhaustively by Mr.
Thos. Andrews, F.R.S., M. lust. C.E. of Sheffield, England, and described
in Engineering of July 10, 17, and 24, 189G. Some of his illustrations are
here reproduced in Fig. 451. The large flaws in Nos. 3, 4, 5, and 6 are
due to small blowholes, while the dark intercellular spaces in Nos. 1 and 2
are largely composed of the sulphide of iron, which, so far as it destroys the
continuity of the crystals, makes' the iron weak and brittle. These and
similar incipient faults, of which there are probably scores in every square
inch of any iron or steel cross-section, are doubtless the initial cause of the
weakness developed by repeated loadings. These breaks in the continuity
of the metal cause the stress to be concentrated at their edges, and the con-
stant variation of this stress, near or at the elastic limit, with its accompany-
ing molecular movements, gradually extends the fracture. Evidently there
can be no regularity of action of such causes, and hence no very rigid rule or
law for such failures. Even two specimens cut from the same bar may act
very differently, to say nothing of specimens made by the same processes at
537
538
THE MATERIALS OF CONSTRUCTION.
1. Micro-flaws x 250 ; Sulphur
= 0.25 percent.
2. Micro-flaws X 200 ; Sulphur
= 2.00 per cent.
3. Micro-flaws X 250. Siemens-
Steel Boiler-plate.
4. Micro-flaws x *J50. Siemens -
steel Propeller-shaft
5. Micro-flaws x 400. Bessemer- 6. Micro-flaws X 250. Bessemer-
steel Railway-axle. steel Rail.
FIG. 451. Views of Internal Micro-flaws in Steel. (Andrews in Engineering, July 1CV
1896.)
THE FATIGUE OF METALS.
539
different times and at different works, or of specimens made by different
processes and having different chemical compositions. Evidently the results
of fatigue tests would be extremely various, and this is the experience of all
the experimenters in this field of investigation.
385. Wohler's Tests. The first systematic study of the fatigue of metals
was made by Wohler from 1849 to 1870 for the German government, and
L
U
FIG. 452. Wohler's Machine for Repetitions of Tensile Stress.
these were continued after his death by Spangenberg. As Wohler's tests
have become historically famous, his appliances are here described.
For Repeated Tensile Stresses Wohler used the apparatus shown in Fig.
<52. Here the specimen A is stressed through the lever L and spring s
FIG. 453. Wohler's Machine for Repetition of Bending Stress.
acting on the auxiliary lever m. The pull of the spring s is measured by
the starting of the adjusted calibrated spring s' through the terminal lever g.
The nut at the rod d is adjusted to give the minimum load on the spring s
by starting the spring s' when adjusted tp a particular tension, and the cam-
movement of d to its extreme downward position is made to give the requi-
site maximum stress in the specimen by adjusting the spring s' so as just
540
THE MATERIALS OF CONSTRUCTION.
to lift at this position of d. The rod is adjustable by means of a turn-
buckle. In this way .the bar A can be stressed in tension between any chosen
limits.
For Repeated Bending Stresses Wohler employed the machine shown in
Fig. 453. Here the specimen A is bent downwards by the adjustable rod q
attached to the rocking lever below. If the load is not to be wholly removed
each time, a residual deflection is maintained by means of the abutting screw
in the lever m. Both the maximum and the minimum loads are fixed by
means of the calibrated spring s acting on the attached lever g.
For Reversals of Bending Stress Wohler made use of the apparatus
shown in Fig. 454. Here two test-bars AA are attached by a driving fit to
FIG. 454. Wohler's Machine for Reversals of Bending Stress.
the central axle, which is rotated by a belt and pulley. The ends of the
test-bars are held down by the calibrated springs ss, so that the bending
stresses are reversed at every revolution. Of course the test-specimens are
trued up to run truly after driving and before loading.
For Repetitions of Torsional Stress Wohler devised the machine shown
in Fig. 455. Here the specimen A is fastened to the moving lever L at one
FIG. 455. Wohler's Machine for Itepetition of Torsional Stress.
end and to the resisting levers hh at the other. The lever L is actuated by
the connecting-rod I and the lever 0, which in turn is moved by the recip-
THE FATIGUE OF METALS.
541
rocating-bar C. If the bar is stressed in opposite directions, then both the
levers g and g' are in use, and the calibrated springs ss act to limit the tor-
sional moment to the required amount as before.
386. The Results of Fatigue Tests. The most careful and complete set
of fatigue tests under repeated stresses was made by Bauschinger. His
results on mild-steel plates are shown in Fig. 456. For this material, which
had an ultimate strength of 64,000 Ibs. per square inch the repetition limit
was found to be about 35,000 Ibs. per square inch, or about the elastic limit
1
7777^/K
/N A//<
/ 2 3 4 5 ..... (3)
Working reversed stress = fa. J
To find the equations of the total working stress in terms of the maximum
and minimum total stresses on any member:
Let L = total live-load stress on any member;
D= " dead-load " " " "
A = area of cross-section of the member;
p = maximum stress in the member per square inch for both dead and
live loads;
a = working stress for live loads.
546 THE MATERIALS OF CONSTRUCTION.
Then we have, from Fig. 460,
nl = dead-load stress per square inch = ;
nm = live-load stress per square inch = ;
ml = total stress per square inch
COMPRESSION
/A/ c$Mf/?SS/o/v= -
FIG. 460.
And since rs = 2a, we have, from the figure,
p = Op = Oa -\- Jim and hm = -= (rs Oa) ;
rs
D
D
THE FATIGUE OF METALS. 547
But A - ; hence we have
'(-
-L)
and finally
min. stress
2 max. stress
This formula may be used in place of both Launhardt's and Weyrauch's
equations ((1) and (2) ), since it applies equally well to stresses of the same
or of opposite kinds, by paying attention to the sign of the minimum stress.
When the minimum stress becomes negative the sign of the second term in
the denominator changes to plus, thus reducing p below a.
Another argument in favor of this formula lies in the fact that it is the
same as the old rule of using twice the factor of safety for live as for dead
loads, as will now be shown.
With the same notation as above, we have
L D _ 2L + D
^7t + 2a~ ~ 2a '
also
Substituting the value of A, we have
_ _ _ a __ _ a
2L + D D _ min. stress '
~ 2(L + /)) ~ 2 max. stress
We find, therefore, that the past practice founded on experience, and the
fatigue experiments, all agree and are all expressed in this one formula which
is universal in its application to stresses of the same and of opposite signs.
Its use is more laborious than those hitherto used, as given in equations (1)
and (2), only in requiring a division in place of a multiplication; but as such
work is now done wholly by the slide-rule, even this objection is removed.
CHAPTER XXVIII.
STRENGTH OF THE COPPER-ZINC-TIN ALLOYS.
COPPER.
390. Strength of Copper. The first and most general error to guard
against in the matter of the strength of copper and its alloys is that of ignor-
ing the mechanical treatment to which the material has been subjected.
Thus in the case of copper plate, as shown by Fig. 461, a hot-rolled plate
has an elastic limit of only some 7000 or 8000 Ibs. per square inch, with an
elongation of 50 per cent, while the same plate, cold-hammered, has an
elastic limit of over 20,000 Ibs. per square inch, with an elongation of 30 per
cent. Both have an ultimate strength of about 33,000 Ibs. per square
5 /0 /5 2O 25 30 35
FIG. 461. Typical Stress diagrams of Copper Plate in. thick.
(Martens, Berlin Testing Lab. Communications, 1894.)
inch. When simply cast, without rolling or forging, both the elastic limit
and the ultimate strength are much less, but copper is seldom or never used
in this way.
Drawn copper wire has an elastic limit of about 25,000, with an ultimate
strength of some 35,000 Ibs. per square inch, as shown in Fig. 462, with an
elongation of about 30 per cent.
If the strength of copper be computed on the actual section at all stages
of the test, and if the strength so computed be plotted to the diminishing
cross-section, the results will plot in a straight line, as shown in Fig. 463.
548
STRENGTH OF THE COPPER- ZINC- TIN ALLOYS.
549
FIG. 462. Typical Stress-diagram of Drawn Copper. ( Wat. Ars. Rep. 1886,
vol. ii. p. 1673.)
70,000
SQ000
~24 22 2O /8 '/
\'~;
i
^
1
e
|
-^
k
'A
I
1
T&
yp\
'Rt
TU,
?l
^ 200 400 600' M
FIG. 472. Strength of Aluminum Bronze at Various Temperatures and for Various
Percentages of Elongation.
o56 THE MATERIALS OF CONSTRUCTION.
From an examination of this chart it is at once evident that only those
alloys near the copper apex are of any value, the strongest being, however,
near the copper- zinc side, where the composition is about 59 per cent copper,
39 per cent zinc, and 2 per cent tin. The tensile strength of such a casting,
if properly made, is about 00,000 Ibs. It is too brittle, however, to be of
much value. The most valuable alloys are those having an ultimate strength
of from 35,000 to 40,000 Ibs. tensile strength, this having from 20 to 30
per cent elongation. This is found in the vicinity of 75 to 85 per cent
copper, 1? to 5 per cent zinc, and 8 to 10 per cent tin.
Tobm Bronze is simply such a composition as the above hot-rolled after
casting. The effect of this rolling is to greatly increase both the strength
and the ductility, as shown by Fig. 469.* This material is in almost every
respect similar to soft steel, so far as its mechanical qualities are concerned.
It has the further advantage of not corroding under ordinary conditions,
hence its extraordinary value as a structural material. The author has seen,
however, some very remarkable and unexplained fractures of this material
which leads him to suspect its reliability.
Phosphor-bronze has no special mechanical properties other than marks
all good bronzes (see Fig. 470). Phosphorus is used to destroy the effects
of oxidation in melting rather than to add to the strength or ductility other-
wise. In destroying these oxides it does improve the product; but if the
melting is performed in such a way as to prevent oxidation, there is no need
of the phosphorus.
394. Aluminum-bronze may have great strength in Doth the cast and
the rolled forms, as shown in Fig. 471, where many tests on different
compositions and from different kinds of moulds are plotted from the same
origin. The effect of the rolling in increasing the strength and the ductility
is evident. A small percentage of aluminum is thus seen to greatly improve
the bronze although its strength alone is very small, as is also shown in this
figure.
* The total elongation, which was over 30 per cent, is not shown in this figure.
CHAPTER XXIX.
THE EFFECTS OF TEMPERATURE ON THE MECHANICAL PROPERTIES
OF METALS.
EFFECTS ON IRON AND STEEL.
395. As Shown by Stress-diagrams. This subject has been very fully
and carefully investigated at the testing laboratory of the U. S. Arsenal at
Watertown, Mass., and a full series of diagrams, similar to that shown in
60000
70000
02 4 6 8/0/2/4/3/8
FIG. 473.- Stress-diagrams of Steel Bars, 0.20$ Carbon, 0.45# Manganese, at Various
Temperatures. (Wat. Ars. Rep. 1888.)
Fig. 473, is given in the report for 1888. The curves in this figure exhibit
the action of 0.20 per cent carbon steel, having a normal tensile strength at
70 F. of 70,000 Ibs. per square inch, with a normal elastic limit of some 37,000
bis. per square inch. Fig. 473 reveals both the elastic limit and the ultimate
557
558
THE MATERIALS OF CONSTRUCTION.
strength, both of which are above the normal at F., and below the normal
at 210 F. The ultimate strength then increases with a rising temperature,
reaching a maximum at about 600 F., from which temperature it regularly
200 400 6W 80O
FIG. 474. Variation of Tensile Strength with Temperature. (Wat. Ars. Pep. 1888.
/00 00 <30O 400 300 600 700 <%X?
FIG. 475. Variation of Tensile Strength of Wrought Iron
v^
V*S0#/&.
/w
vest
/&ff
a /
W