Digitized by the Internet Archive in 2010 with funding from Boston Library Consortium IVIember Libraries http://www.archive.org/details/cyclopediaofarchamer 'A^y^s-.f^Ul i/> .-vi* Cyclopedia of Architecture, Carpentry, and Building A General Reference Work ON ARCHITECTURE, CARPENTRY, BUILDING, SUPERINTENDENCE, CONTRACTS, SPECIFICATIONS, BUILDING LAW, STAIR-BUILDING, ESTIMATING, MASONRY, REINFORCED CONCRETE, STRUCTURAL ENGINEER- ING, ARCHITECTURAL DRAWING, SHEET METAL WORK, HEATING, VENTILATING, ETC. Prepared by a Staff of ARCHITECTS, BUILDERS, ENGINEERS, AND EXPERTS OF THE HIGHEST PROFESSIONAL STANDING Illustrated with over Three Thousand Engravings TEN VOLUMES CHICAGO AMERICAN TECHNICAL SOCIETY 1914 BOSTON COi-J-EGE \^ PHYSICS DEPT. COFYRiGHT. 1907. 1909. 1912 BY AMERICAN SCHOOL OF CORRESPONDENCE Copyright. 1907. 1909. 1912 BY AMERICAN TECHNICAL SOCIETY Entered at Stationers' Hall. London All Rights Reserved Authors and Collaborators JAMES C. PLANT Superintendent of Computing Division, Office of Supervising Architect, Treasury, Washington, D. C. WALTER LORING WEBB, C, E. Consulting Civil Engineer Author of "Railroad Construction," "Economics of Railroad Construction," etc. J. R. COOLIDGE, Jr., A. M. Architect, Boston President, Boston Society of Architects Acting Director, Museum of Fine Arts, Boston H. V. VON HOLST, A. B., S. B. Architect, Chicago President, Chicago Architectural Club *^ FRED T. HODGSON Architect and Editor Member, Ontario Association of Architects Author of "Modern Carpentry," "Architectural Drawing, Self -Taught," "The Steel Square," "Modern Estimator," etc. GLENN M. HOBBS, Ph. D. Secretary, American School of Correspondence FRANK 0. DUFOUR, C. E. Assistant Professor of Structural Engineering, University of Illinois American Society of Civil Engineers SIDNEY T. STRICKLAND, S. B. Massachusetts Institute of Technology Ecole des Beaux Arts, Paris WM. H. LAWRENCE, S. B. Professor of Architectural Engineering, Massachusetts Institute of Technology Authors and Collaborators— Continued EDWARD NICHOLS Architect, Boston V» H. W. GARDNER, S. B. Associate Professor of Architecture, Massachusetts Institute of Technology ^« JESSIE M. SHEPHERD, A. B. Associate Editor, Textbook Department, American School of Correspondence GEORGE C. SHAAD, E. E. Professor of Electrical Engineering, University of Kansas MORRIS WILLIAMS Writer and Expert on Carpentry and Building HERBERT E. EVERETT Professor of the History of Art, University of Pennsylvania ^* ERNEST L. WALLACE, B. S. Assistant Examiner, United States Patent Office, Washington, D. C. Formerly Instructor in Electrical Engineering, American School of Correspondence OTIS W. RICHARDSON, LL. B. Of the Boston Bar WM, G. SNOW, S. B. steam Heating Specialist Author of "Furnace Heating," Joint Author of "Ventilation of Buildings" American Society of Mechanical Engineers W. HERBERT GIBSON, B. S., C. E. Civil Engineer and Designer of Reinforced Concrete ELIOT N. JONES, LL. B. Of the Boston Bar Authors and Collaborators — Continued R. T. MILLER, Jr., A. M., LL. B. President, American School of Correspondence ^« WM. NEUBECKER Instructor, Sheet Metal Department of New York Trade School WM. BEALL GRAY Sanitary Engineer Member, National Association of Master Plumbers EDWARD MAURER, B. C. E. Professor of Mechanics, University of Wisconsin EDWARD A. TUCKER, S. B. Architectural Engineer Member, American Society of Civil Engineers EDWARD B. WAITE Head of Instruction Department, American School of Correspondence American Society of Mechanical Engineers Western Society of Engineers ^* ALVAH HORTON SABIN, M. S. Lecturer in New York University Author of "Technology of Paint and Varnish," etc. American Society of Mechanical Engineers GEORGE R. METCALFE, M. E. Editor, American Institute of Electrical Engineers Formerly Head, Technical Publication Department, Westinghouse Electric & Manufac- turing Co. ^« HENRY M. HYDE Editor "Technical World Magazine" ^* CHAS. L. HUBBARD, S. B., M. E. Consulting Engineer on Heating, Ventilating, Lighting, and Power Formerly with S. Homer Woodbridge Co. Authors and Collaborators— Continued FRANK CHOUTEAU BROWN Architect, Boston Author of " Letters and Lettering" DAVID A. GREGG Teacher and Lecturer in Pen and Ink Rendering, Massachusetts Institute of Technology CHAS. B. BALL Chief Sanitary Inspector, City of Chicago American Society of Civil Engineers ERVIN KENISON, S. B. Assistant Professor of Mechanical Drawing, Massachusetts Institute of Technology CHAS. E. KNOX, E. E. Consulting Electrical Engineer American Institute of Electrical Engineers JOHN H. JALLINGS Mechanical Engineer FRANK A. BOURNE, S. M., A. A. I. A. Architect, Boston Special Librarian, Department of Fine Arts, Public Library Boston ALFRED S. JOHNSON, Ph. D. Formerly Editor "Technical World Magazine" GILBERT TOWNSEND, S. B. With Ross & McFarlane, Montreal HARRIS C. TROW, S. B., Managing Editor Editor-in-Chief, Textbook Department, American School of Correspondence Authorities Consulted THE editors have freely consulted the standard technical literature of America and Europe in the preparation of these volumes. They desire to express their indebtedness particularly to the following eminent authorities whose well-known works should be in the library of everyone connected with building. Grateful acknowledgment is here made also for the invaluable co- operation of the foremost architects, engineers, and builders in making these volumes thoroughly representative of the very best and latest prac- tice in the design and construction of buildings ; also for the valuable drawings and data, suggestions, criticisms, and other courtesies. J. B. JOHNSON, C. E. Formerly Dean, College of Mechanics and Engineering, University of Wisconsin Author of "Engineering Contracts and Specifications," "Materials of Construction," Joint Author of "Theory and Practice [in the Designing of Modern Framed Struc- tures" ^* JOHN CASSAN WAIT, M. C. E., LL. B. Counselor-at-Law and Consulting Engineer; Formerly Assistant Professor of Engineer- ing at Harvard University Author of "Engineering and Architectural Jurisprudence" T. M. CLARE Fellow of the American Institute of Architects Author of "Building Superintendence," "Architect, Builder, and Owner before the Law" FRANK E. KIDDER, C. E., Ph. D. Consulting Architect and Structural Engineer; Fellow of the American Institute of Architects Author of "Architects' and Builders' Pocket-Pook;" "Building Construction and Superintendence; Part I, Masons' Work; Part II, Carpenters' Work; Part III, Trussed Roofs and Roof Trusses; " "Churches and Chapels" AUSTIN T. BYRNE, C. E. Civil Engineer Author of "Inspection of Materials and Workmanship Employed in Construction,' "Highway Construction" W. R. WARE Formerly Professor of Architecture, Columbia University Author of "Modern Perspective" Authorities Consulted— Continued CLARENCE A. MARTIN Professor of Architecture at Cornell University Author of "Details of Building Construction" FRANK N. SNYDER Architect Author of "Building Details" CHARLES H. SNOW Author of "The Principal Species of Wood, Their Characteristic Properties" OWEN B. MAGINNIS ^ Author of "How to Frame a House, or House and Roof Framing ' HALBERT P. GILLETTE, C. E. ^ Author of "Handbook of Cost Data for Contractors and Engineers" OLIVER COLEMAN ^ Author of "Successful Houses" CHAS. E. GREENE, A. M., C. E. Formerly Professor of Civil Engineering, University of Michigan Author of "Structural Mechanics" LOUIS de C. BERG "^ Author of "Safe Building" *^ GAETANO LANZA, S. B., C. & M. E. Professor of Theoretical and Applied Mechanics, Massachusetts Institute of Technology Author of "Applied Mechanics" IRA O. BAKER ^ Professor of Civil Engineering, University of Illinois Author of "A Treatise on Masonry Construction" ^• GEORGE P. MERRILL Author of "Stones for Building and Decoration" FREDERICK W.TAYLOR, M. E., and SANFORD E.THOMPSON, S. B., C.E. Joint Authors of "A Treatise on Concrete, Plain and Reinforced" Authorities Consulted— Continued A, W. BUEL and C. S. HILL Joint Authors of "Reinforced Concrete" NEWTON HARRISON, E. E. Author of " Electric Wiring, Diagrams and Switchboards" FRANCIS B. CROCKER, E. M., Ph. D. Head of Department of Electrical Engineering, Columbia University; Past President, American Institute of Electrical Engineers Author of " Electric Lighting" *^ J. R. CRAVATH and V. R. LANSINGH Joint Authors of "Practical Illumination" JOSEPH KENDALL FREITAG, B. S., C. E. Authors of "Architectural Engineering," " Fireproofing of Steel Buildings" WILLIAM H. BIRKMIRE, C. E. Author of "Planning and Construction of High Office Buildings," "Architectural Iron and Steel, and Its Application in the Construction of Buildings," "Compound Riveted Girders," "Skeleton Structures," etc. EVERETT U. CROSBY and HENRY A. FISKE Joint Authors of "Handbook of Fire Protection for Improved Risk" CARNEGIE STEEL COMPANY Authors of " Pocket Companion, Containing Useful Information and Tables Appertain- ing to the Use of Steel" ^• J. C. TRAUTWINE, C. E. Author of "Civil Engineer's Pocket-Book" */» ALPHA PIERCE JAMISON, M. E. Assistant Professor of Mechanical Drawing, Purdue University Author of "Advanced Mechanical Drawing" ^« FRANK CHOUTEAU BROWN Architect, Boston Author of " Letters and Lettering" Authorities Consulted— Continued HENRY McGOODWIN Author of "Architectural Shades and Shadows" ^* VIGNOLA Author of "The Five Orders of Architecture," American Edition by Prof. Ware ^* CHAS. D. MAGINNIS Author of "Pen Drawing, An Illustrated Treatise" FRANZ S. MEYER Professor in the School of Industrial Art, Karlsruhe Author of " Handbook of Ornament," American Edition *»• RUSSELL STURGIS Author of "A Dictionary of Architecture and Building," and "How to Judge Archi- tecture" A. D. F. HAMLIN, A. M. Professor of Architecture at Columbia University Author of "A Textbook of the History of Architecture RALPH ADAMS CRAM Architect Author of "Church Building" C. H. MOORE Author of "Development and Character of Gothic Architecture" ROLLA C. CARPENTER, C. E., M. M. E. Professor of Experimental Engineering, Cornell University Author of "Heating and Ventilating Buildings" WILLIAM PAUL GERHARD Author of "A Guide to Sanitary House Inspection" •>• I. J. COSGROVE Author of " Principles and Practice of Plumbing" 2 m i-l ^ O o U <0 S X •< Pd S Q fe fn O » Fig. 9. Log Showing Checks Fig. 10. Finished Timber Showing Checks for the market they show themselves in the form of cracks which extend along the faces of solid squared timbers and boards, seriously impau-ing their strength. Fig. 10 shows checks as they would appear in a square post or column. Conversion of Timber into Lumber. Lumber may be found in lumber yards in certain shapes ready for use, having been cut from the logs and relieved of their outside covering of bark. The cutting up of the logs is done in the mills by machinery and there are various methods in use for transforming the logs into boards, planks, and 5 \ \ \ \ ^ \ \ \ Fig. 11. Economical Method of Cutting Logs Fig. 12. Another Method of Cutting Logs heavy timbers. The method of cutting the log determines the appearance of the wood when finished and also affects it in other ways. If the log is to be squared off so as to form only one heavy beam or post, a good rule to follow is to divide the diameter into three 84 CARPENTRY 15 equal parts and then to draw perpendiculars to this diameter at the division points one on each side of the center, as shown at A and B in Fig. 11. The points C and D in which these perpendiculars to the diameter cut the circumference of the log, together with the points E and F in which the diameter cuts the circumference of the log, will be the four corners of the timber. The lines joining these points will give an outline of the timber, which will be rectangular and will be found to be the largest and best timber which can be cut from the log. Another good rule is to divide the diameter of the log into four equal parts and to proceed in the same way as described above, using the outside quarter points from which to draw the perpendiculars as shown in Fig. 12. This method will give the outlines of a stiff er beam than the one described above, but there will be more waste from the log and the beam will not be on the whole as strong as the other. In Fig. 13 are shown several different methods of cutting planks from a log. First it is divided into quarters, and the planks are cut out as shown in the figure, there being four ways in which the work may be done. All of the four methods shown may be said to give what is called quarter-sawed lumber since the log is first cut into quarters, but that shown at A is the best. All of the planks are cut radiating from the center of the log and there will be no split- ting or warping, but the method is very expensive, as all of the planks have to be squared up afterward and there is much waste as a result. A fairly good method is that shown at B where the planks are nearly along radial lines and may be much more easily and cheaply cut out than can those shown at A. The method shown at C is a com- mon one and leads to fairly good results, although only the plank nearest the center is on a radial line. It is practically as good a method as that shown at B and is much more simple. The method shown at D is not so good as the others, as planks cut out in this way are very liable to warp and twist. If the silver grain, caused by cutting of the medullary rays is desired, the planks must be cut as shown at ^, B, or C. Fig. 13. Method of Cutting Planks from a Log 25 16 CARPENTRY Fig. 14. Method of Slicing a Log Planks are sometimes simply sliced from the log as shown in Fig. 14, without first dividing it into quarters, but this is the worst possible way of cutting them, as the natural tendency of the timber to shrink causes the planks to curl up as shown in Fig. 15. It is almost impossible to flatten them out again, and they can not be used in that condition. There is another method of cutting up a log which has been introduced more recently than the others, and which is known as the "rotary cut." It consists in placing the log on a movable carriage which keeps it whirling rapidly about its longitudinal axis, at the same time bring- ing it up against a long stationary knife which catches the log and peels off strips around the circumference of any desired thickness. This method is used extensively in the preparation of wood to be used as veneers, and in the case of many kinds of wood the figure is brought out to better advantage in this way than is possible with any other method. Waney Lumber. When a log of wood has been sawed up into boards, each board is apt to have along the edge a strip of the bark which was originally on the outside of the log, and the edges will not be square with the face of the board, owing to the cylindrical shape of the log. Such boards should be squared up by having the rough edges to which the bark adheres trimmed off. But sometimes the bark alone is stripped off, leaving the boards with the edges not square with the face. Such boards are said to be waney, and very often specifications state that no waney lumber shall be employed on the work. The pieces which are cut off when waney boards are trimmed in order to square them up are called "edging" and are used to make laths. Slabs. The pieces known as slabs are those which are left over after a log has been sawed up into boards. In cross section they are of the shape of a half moon, and are covered with bark. They are useless except for laths or fuel. Fig. 15. Slicing Logs Gives Warped Lumber 26 CARPENTRY 17 VARIETIES OF TIMBER Although there are a great many different kinds of trees growing in different parts of the world, only a comparatively small number of them yield wood which is used to any great extent in building work. These differ very much among themselves, each variety possessing certain characteristics which render it especially suitable for use in one part of a building, while the same peculiarities of growth or of texture may make it unfit for use in another. For use in places where the timber must be partly buried in the ground a wood is required which will be able to withstand the deteriorating effects of contact with the earth, and for this purpose chestnut, white cedar, cypress, redwood, or locust may be used. For light framing is needed a cheap, light wood, as free as possible from structural defects, such as knots and shakes, and one which can be readily obtained in fairly long, straight pieces. Spruce, yellow pine, white pine, and hemlock all satisfy these requirements fairly well, spruce being perhaps a little better than the others, and more popular. For heavy framing, such as trusses, girders, and posts, a timber is needed which is strong, and which can be obtained in large, long pieces. Georgia pine, Oregon pine, and white oak may all be used for such work, and also Norway pine and Canadian red pine. White oak is the timber which was always used for framing in the old days, but is too expensive to be used with profit for such work now. The timber most commonly used today is the Georgia pine. A wood which can be easily worked and which will also be able to withstand the deteriorating effects of the weather is in demand for the outside finish. White pine is usually selected for this purpose, although cypress and redwood are also suitable and are used to some extent. The same woods are used for shingles, clapboards, and siding, with the addition of cedar and spruce for shingles, and Oregon pine and spruce for siding. For the interior finish is chosen a wood which will give a pleasing appearance when finished and which will take a high polish, while for floors, hardness, and resistance to wear are the additional require- ments. For floors, oak, hard pine, maple, and birch are good, while for the remainder of the interior finish white pine, cypress, and red- 27 18 CARPENTRY wood for painting, or any of the hard woods such as ash, cherry, oak, walnut, or mahogany, may be selected. Some of the more important varieties of timber used in Car- pentry will now be mentioned, and a brief description of each variety will be given in order to convey an idea of their characteristics and the part of the world from which they come. Conifers or Needle=Leaved Trees. These trees are found mostly in the North, where they form large forests from which are taken the large quantities of timber of this kind used every year. The wood is very popular for use in rough building construction or for finished work which is to be painted, as it is very regular in structure and consequently easy to work; it can be obtained in large, long, straight pieces, and is light and strong. The demand for woods of this kind is considerably in excess of the demand for the harder woods. The trees are mostly but not all evergreen, and bear needles instead of leaves, together with the cones, from which they are called conifers. Cedar. The wood known as cedar has long been used in con- struction, as is illustrated by the references in the Bible to the "Cedars of Lebanon" from which the Temple of Solomon was constructed. The wood in use at the present day called cedar is, of course, not of exactly the same species as was that used in the famous temple, but it is of the same family and possesses the same general character- istics. There are two kinds, the red cedar, and the white cedar, which differ from each other principally in color, the white cedar being grayish brown, while the red cedar is reddish brown. There are several different kinds of white cedar in use, of which one is known as the canoe cedar. The wood is not very strong, but is light and soft, possessing considerable stiffness and a fine texture. In color it is as mentioned above, grayish brown, the sapwood being, however, of a lighter color than the heartwood. It seasons quickly, is remarkably durable, and does not shrink or check to any great extent. The wood is used in building construction, principally for shingles, for which purpose its durability in exposed positions makes it especially valuable. It is also used for posts and ties. The trees are usually scattered among others of different kinds, forming occasionally, however, forests of considerable size. They are 28 CARPENTRY 19 to be found all through the northern part of the United States and in Canada, also on the Pacific Coast in California, Oregon, and Wash- ington. They also grow to some extent in the southern states. Some of the trees are of small or medium size, while others are very large, especially the canoe cedar of the Northwest. In addition to the white cedars, there are the red cedars, which are similar to the white cedars but differ from them slightly in the color of the wood, which is reddish brown instead of grayish brown. The red cedars are also of somewhat finer texture than the white cedars. Red cedar is used but little in building con- struction, but is used extensively in cabinet work for chests and closets which this wood is supposed to render proof against moths. The wood is also used for the making of lead pencils and for cigar boxes, large quantities of timber being used for these purposes every year. Cedar trees are sometimes subject to a disease similar to wet rot, which attacks the growing tree. This disease does not, how- ever, render them unfit for use in every case, as the disease often disappears as soon as the tree has been cut down, and trees have been known to yield timber which has endured for long periods, although the living tree itself was diseased. Redwood. There is a wood which greatly resembles good red cedar and which is found only in the State of California. One species of this tree grows to an enormous size and is famous on this account, but this is not the one which yields the lumber used for building purposes, which is known as the common redwood. The wood is used for cheap interior finish and for shingles, also for use in heavy construction, thus serving nearly the same purposes as does hard pine in the eastern states. Redwood is light, and not very strong, but on the other hand, it is remarkably durable, resisting fire to a considerable extent. It is easy to work and will take a polish so that it is valuable for inside finish, and some of the wood has a wavy grain which adds greatly to its finished appearance. This wood is known as "curly" redwood. In color the heartwood is red, but the sapwood is nearly white, with the wood between them varying in color and averaging a rich reddish brown. The grain is usually straight and the wood is solid and dense in structure but the grain is more or less coarse in appearance. 29 20 CARPENTRY Cypress. This is a wood which is somewhat similar to white cedar in appearance, and which grows in quantities only in the southern states, where it may be seen in great swamps with the roots very often partially exposed. Although there are a great many varieties, they are similar in their general characteristics, differing only in quality. "Gulf Cypress," growing near the Gulf of Mexico, is the best. "Bald Cypress," is a name which has been applied to these trees on account of the fact that they show no leaves in winter and this gives them a peculiar appearance. When the wood is dark in color it is called "Black Cypress," and in some localities yellow and red cypress are spoken of. The growing trees are often affected by a disease which leaves the wood full of small holes which look as though they might have been made by driving pegs into the wood and then withdrawing them. Cypress wood affected in this way is called "peggy." Hemlock. There are two varieties of hemlock, one found in the northern states, from Maine to Minnesota, and along the Alle- ghenies southward to Georgia and Alabama, while the other is found in the west from Washington to California and eastward to Montana. The eastern tree is smaller than the western and its wood is lighter, softer, and generally inferior. The trees are evergreen and bear cones, with flat, blunt needles, and they usually grow alone or in small groups in the midst of forests of other trees. The timber is of a light, reddish-gray color, fairly durable, but shrinks and checks badly, and is coarse, brittle, and usually cross grained. It is hard to work but will hold nails very well. The wood is sometimes used for cheap framing, and has been used for cheap interior finish, but it is so liable to imperfections, such as windshakes and starshakes, that it is not the best wood to use for these purposes, although the increasing cost of the better woods will no doubt force it into more general use. Hemlock is most frequently used for rough boarding and sheathing. Spruce. Another evergreen and cone-bearing tree which lur- nishes great quantities of lumber to the market every year is the spruce. There are three kinds of spruce, white, black, and red, of which the white spruce and the red spruce are the varieties com- monly found on the market. The white spruce is scattered through- out all of the northern states, along the streams and lakes, the largest 30 CARPENTRY 21 varieties being found in Montana. The black spruce is found in Canada and in some of the northern states. It is distinguished from the other varieties by its leaves and bark only, the foliage being much darker in color than that of the white spruce, while the cones remain in place for several years, a much longer time than do those of the white spruce. The red spruce is sometimes known as Newfoundland red pine and is found in the northeastern part of North America. It is used very extensively in northern New Eng- land where it serves as a substitute for soft pine, and large quantities of it are used up every year for pulp wood. The leaves of the spruce are single and have sharp points at the ends. They are short and four-sided and are arranged on the stem so as to point in all directions. The cones hang downward, while those of the fir trees point upward. Spruce trees have many natural enemies and numbers of the trees are destroyed before they reach the market. Large quantities of fallen tree trunks are to be found in the forests, blown down by the wind alone during heavy wind storms, or so weakened by the ravages of insects that they have fallen from their own weight. There is a beetle which attacks these trees especially, and which causes great damage, while very often the same trees are attacked by various kinds of fungous growths. Spruce timber is of a light color, very nearly white except the heartwood which has a reddish tinge. It is very dense and compact in structure and straight grained. The wood is light and soft, fairly strong for a soft wood, but not very durable when exposed. It is very resonant and is frequently used for sounding boards on this account. It can not be obtained in large sizes, but it is considered by many to be the best framing timber available, except the pines. Pine. This is the timber which has been used in building con- struction to a greater extent than any other except perhaps oak. It is peculiarly fitted for the purpose as it has grown in great abund- ance all over the United States and possesses all of the most desirable characteristics of a good building material, being strong, but at the same time light in weight and easily worked, elastic, and very dur- able. The tree is almost always a large one with branches starting at a considerable distance from the ground. It has a smooth, straight trunk, evergreen, needle-shaped leaves, of varying length, and cones. SI 22 CARPENTRY There are two distinct classes of pines used in building work, the soft and the hard pines, both of which are found in large quanti- ties. The softer varieties are used for outside finish of all sorts, and the harder varieties for heavy framing and for flooring. The ease with which the soft-pine lumber can be cut and shipped to the market, makes it the most popular wood in use at the present time. It is of uniform texture and nails without splitting, seasons very well, and does not shrink so much as the harder pines, will take paint and is very durable. The wood is white in color, straight grained, and has few knots. The hard pines furnish the strongest timber in use for building, with the exception of oak, which is now almost too expensive to be used for heavy framing. The pieces can be obtained in large sizes and great lengths and the wood is very hard, heavy, and durable, at the same time being tough. In color it is yellow or orange, the sapwood being of lighter color than the heartwood. There are many different kinds of pines, which are recognized in different parts of the country under various names, but there are five general classes into which the species is commonly divided, though the same timber may be called by different names in two different localities, as will be seen. (1) The term "hard pine" is used to designate any pine which is not white pine, a classification which is very general, though it is often seen in works on Carpentry and in specifications. (2) "White pine," "soft pine," and "pumpkin pine," are terms which are used in the eastern states for the timber from the white-pine tree, while on the Pacific Coast the same terms refer to the wood of the sugar pine. (3) The name "yellow pine," when used in the northeastern part of the country, applies almost always to the pitch pine or to one of the southern pines, but in the West it refers to the bull pine. (4) "Georgia pine" or "longleaf yellow pine," is a term used to distinguish the southern hard pine which grows in the coast region from North Carolina to Texas, and which furnishes the strongest pine lumber on the market. (5) "Pitch pine" may refer to any of the southern pines, or to pitch pine proper, which is found along the coast from New York to Georgia and among the mountains of Kentucky. 82 CARPENTRY 23 Of the soft pines there are two kinds, the white pine and the sugar pine, the latter being a western tree found in Oregon and Cahfornia, while the former is found in all the northern states from Maine to Minnesota. There is also a smaller species of white pine found along the Rocky Mountain slopes from Montana to New Mexico. There are ten different varieties of hard pine, of which, however, only five are of practical importance in the building industry. These are the "long-leaf southern pine," the "short-leaf southern pine," the "yellow pine," the "loblolly pine," and the "Norway pine." The long-leaf pine, also known as the "Georgia pine" and the "long straw pine," is a large tree which forms extensive forests in the coast region from North Carolina to Texas. It yields very hard, strong timber, which can be obtained in long, straight pieces of very large size. The loblolly pine is also a large tree but has more sapwood than the long-leaf pine, and is coarser, lighter, and softer. It is the common lumber pine from Virginia to South Carolina, and is found as well in Texas and Arkansas. It is known also by the names of "slash pine," "old field pine," "rosemary pine," "sap pine," and "short straw pine," and in the West as "Texas pine." The short-leaf pine is much like the loblolly pine and is the chief lumber tree of Missouri and Arkansas. It is also found in North Carolina and Texas. The Norway pine is a Northern tree found in Canada and the northern states. It never forms forests, but is scattered among other trees, and forms small groves. The wood is fine grained and of a white color but is largely sapwood and is not durable. Fir. The fir tree yields timber very similar to spruce, and is often mixed in with spruce in the market. There are two kinds of fir trees, the western fir tree and the eastern fir tree, the first being known as the silver fir and the other as the balsam fir. All of the firs are evergreen, and bear cones which stand erect instead of hanging down. The wood is soft and not strong, being of a much coarser quality than ordinary spruce. It can be used in building work only for the roughest work in the case of the eastern fir, while the western fir is used more extensively but is not as good as spruce. Tamarack. This is a wood which is very much like spruce in 33 24 CARPENTRY structure, but is hard and very strong, resembling hard pine in this respect. The tree grows in the northern part of the United States and Canada, both in the East and in the West, and also in Europe. Its true name is larch, but it has come to be known as tamarack, tamarack pine, and hackmatack. In the East the tree grows in wet places called tamarack swamps, but the tree in the West and in Europe thrives best in dryer soil, and grows more quickly under these conditions than in a swamp. The wood is used mostly for long straight timbers such as posts, poles, and quite extensively for piles. It has also been used a great deal for railroad ties. It is supposed to be very durable, and is well suited for use as ties or as piles, but it can not always be obtained now. It has never been used to any extent as sawn lumber, because the demand for the trunks for use as posts and poles has been so great that it did not pay to saw them up. Broad=Leaved Trees. Ash. Ash is a wood which is frequently employed for interior finishing in public buildings, such as school houses, churches, and so forth, and also in the cheaper classes of dwelling houses. It is one of the cheapest of hard woods, and is used when it is desired to have a hard-wood finish and when the more expensive kinds of hard wood, such as oak, can not be afforded. The wood is somewhat like oak in texture and appearance, the difference being that ash is coarser, and the pith rays do not show. It is strong, straight grained, and tough, comparatively easy to work, elastic, and fairly durable. It shrinks moderately, seasons with little injury, and will take a good polish. The trees do not grow together in forests, but are scattered. They grow rapidly, and attain only medium height. Of the six different species found in the United States, only two, the "white ash," and the "black ash," are used extensively in building work. The first is most common in the basin of the Ohio River, but is also found in the North from Maine to Minnesota, and in the South, in Texas. The black ash is found from Maine to Minnesota, and southward to Virginia and Arkansas. There is very little difference between the two species. The black ash is also known as the "hoop ash," and the "ground ash." Beech. This wood is not used to any great extent in Carpentry except in Europe, but is made up into tool handles, shoe lasts, and 34 CARPENTRY 25 so forth, and is also used in wagon making and ship building. The tree grows freely in the eastern part of the United States and Canada and also in Europe. There are a number of different species and the tree is sometimes called by other names such as "ironwood," and "horn-beam." The wood is used for building work in the United States only occasionally for inside finish and is not a popular wood. It is heavy, hard, and strong, but of coarse texture like the ash. In color it is light brown, or white. It shrinks and checks during the process of drying out, and is not durable when placed in contact with the ground. It works fairly well, stands well, and will take a good polish. Birch. Birch is a very handsome wood of a brown or red color and with a satiny luster. There are two kinds, the red birch and the white birch, but they are both taken from the same kind of tree, the difference being that the red birch consists of more and older heartwood, while the white birch is the sapwood or the younger heartwood. The trees are of medium size and form large forests. They are found throughout the eastern part of the United States and Canada, and in the extreme north. The distinguishing feature of the tree is the bark, which is famous because of its beauty and its usefulness for a number of purposes. This bark is white in color with long dashes of a darker color running around the tree trunk in a horizontal direction. It is water-tight and pliable, which made it useful to the Indians for the covering of their canoes. It was also used in ancient times, before the manufacture of paper, as a material to write upon. The bark has been used for a number of other pur- poses. The wood is used quite extensively for inside finish and floors, and to imitate cherry and mahogany, as it has a grain which is very similar to the grain of these woods. It takes a good polish, works easily, and does not warp after it is in place, but it is not durable when exposed to the weather. Butternut. Butternut is really a branch of the family of wal- nuts, and differs from them only slightly. The wood is used to some extent for inside finish, and is cheaper than most of the other hard woods. It is light, but not strong, and is fairly soft. In color it is light brown. The trees, of medium size, are found in the eastern states from Maine to Georgia. Cherry. Cherry is a wood which is frequently used as a finishing 85 26 CARPENTRY wood for the interior of dwellings and of cars and steamers, but, owing to the fact that it can be obtained only in narrow boards, it is most suitable for molded work, and work which is much cut up. The wood is heavy, hard, strong, and of fine texture. The heartwood is of a reddish brown color, while the sapwood is yellowish white. It is very handsome and takes a good polish, works easily, and stands well. It shrinks considerably, however, in drying. The timber is cut from the wild black cherry tree, not from the cultivated cherry tree. This tree is of medium size, and is found scattered among the other broad-leaved trees along the western slopes of the Alle- ghenies, and as far west as Texas. The fruit of the wild cherry is of a dark purple color, about the size of a large pea. When ripe it tastes slightly bitter. The bark of the tree also tastes bitter. Cherry is often stained to resemble mahogany, and sometimes birch is stained to resemble cherry. Chestnut. The grain of chestnut somewhat resembles oak but it is much softer and coarser in texture and does not show the medullary rays which form the distinguishing feature of oak. Chest- nut is used for cabinet work, for interior finishing, and sometimes for heavy construction. It is light, fairly soft, but not strong. The wood has a rather coarse texture, works easily and stands well, but shrinks and checks in drying. It is very durable and can be safely used in exposed positions. The tree grows in the region of the AUeghenies, from Maine to Michigan, and southward to Alabama. The wood is dark brown in color, with the sapwood a little lighter. Elm. There are five species of elm trees in the United States, scattered throughout the eastern and central states. The trees are usually large and of rapid growth, and do not form forests. The timber is hard and tough, frequently cross-grained, hard to work, and shrinks and checks in drying. The wood has not been used very extensively in building, but has a beautiful figured grain, can take a high polish, and is well adapted to staining. The texture is coarse to fine, and the color is brown with shades of gray and red. Gum. The wood of the gum tree has been used extensively for cabinet work, furniture, and interior finish. It is of fine texture and handsome appearance, heavy, fairly soft, yet strong. Its color is reddish brown. The wood warps and checks badly, is not durable when exposed, and is hard to work. It has a close grain, and some 36 CARPENTRY 27 pieces are so regular that they have been stained to imitate black walnut and used as veneers for the manufacture of furniture and cabinet work. The species of gum tree, which yields timber of use in carpentry, is known as the sweet gum. It is of medium size, with a straight trunk. The trees do not form forests, though they are quite abundant east of the Mississippi River. The leaves have five lobes which are long and pointed, thus giving them a starlike appear- ance. The bark is very rough, and its resemblance in appearance to the skin of an alligator has caused the wood to be called "Alligator Wood" in some localities. Maple. Almost all of the maple used in building work comes from the hard sugar maple, which is most abundant in the region of the Great Lakes, but which is also found from Maine to Minnesota and southward to Florida. The trees are of medium to large size and form quite considerable forests. They are so abundant in Canada that the maple is the national tree, and the national emblem is a maple leaf. The wood when finished presents a very pleasing appearance, and ranks as one of the best of the hard woods in this respect. It is heavy and strong, of fine texture, and often has a fine wavy grain which gives the effect known as "curly." Other defects which add to the beauty of the grain occur in what is called "blister" and "bird's-eye" maple. These defects are the result of twisting of the fibers which make up the woody structure of the tree, and the maples seem to show them more frequently than any of the other trees, though they sometimes are to be found in birch and various other woods. The color of the sapwood is a creamy white while the heartwood is tinged with brown. The lumber shrinks moder- ately, stands well, is easy to work, and is tough, but not very durable when subjected to exposure. The finished wood takes an excellent polish. It is most commonly employed for floors, and in other posi- tions where a good wearing surface is required, as well as for ceiling and paneling, and other interior finish, OaJc. This is a wood which has probably been used more than any other kind in all classes of structures. In ancient times it was about the only wood in use both for the building of houses and for shipbuilding. Since the softer woods have become popular and oak has become somewhat less easy to get, its use has diminished to some extent, but it is still one of the most useful of woods. The trees 37 28 CARPENTRY grow freely all over the northern parts of Europe and America, extending as far south as the Equator, and have been particularly plentiful in the British Isles. There are about twenty different kinds of oaks to be found in various parts of the United States and Canada, but there are three distinctly different species, which are sold sepa- rately. These are the "white oak," the "red oak," and the "live oak." The red oak is usually more porous, less durable, and of coarser tex- ture than the white oak or the live oak. The trees are of medium size and form a large proportion of all the broad-leaved forests. Live oak was once very extensively used, but has become scarce and is now expensive. Both the red oak and the white oak are used for inside finishing, but they are liable to shrink and crack and must, therefore, be thoroughly seasoned. They are of slightly different color, the white oak having a straw color while the red oak has a red- dish tinge, so that they can not be used together where the work is to be finished by polishing. Oak is always best if quarter-sawed and it then shows what is known as the "silver grain." This is the result of the cutting of the medullary rays, and appears on the finished wood as a succession of splashes or blotches which are of lighter color than the rest of the wood and which glisten in the light. Poplar. This wood is also known in the market as "white wood," "tulip wood," and sometimes as "basswood." The poplar, the whitewood or tulip tree, and the basswood are, however, three distinct kinds of trees, but the wood of each so nearly resembles that of the others as to be indistinguishable in the market and so it is sold under any one of these various names. The lumber yielded by the tulip tree and known commercially as whitewood is the best. This tree is a native of North America and grows freely all over the United States and Canada. There are a number of different varieties grow- ing in various parts of the country. It is sometimes called "yellow poplar." The poplar or cottonwood is most common in the region of the Ohio basin, and grows in the western desert regions along the water courses. The tree is a large one and usually grows in small groups, not forming extensive forests. The basswood tree, also known as the linden, grows all over the eastern part of the United States and Canada, and in the middle west. The wood of all these trees is light, soft, free from knots, and of fine texture. In color it is white, or yellowish white, and frequently has a satiny luster. It 38 CARPENTRY 29 can be so finished as to retain its natural appearance, but it is often stained to imitate some of the more costly woods, such as cherry. It is used extensively for cheap inside finish and fittings, such as shelving, and sometimes for doors, but it warps badly if it is not thoroughly seasoned, and will not stand exposure. Sycamore. Sycamore is frequently used for finishing, and is a very handsome wood. It is heavy, hard, strong, of coarse texture, and is usually cross grained. It is hard to work, and shrinks, warps, and checks considerably. The tree is of large size and rapid growth, found in all parts of the eastern United States, and is most common along the Ohio and Mississippi rivers. Walnut There are a number of different kinds of walnut trees, of which only one or two, however, yield timber which is suitable for use in building construction. The best known trees are the "English walnut," the "black walnut," the "white walnut" or "butternut," and the "Circassian walnut." The English walnut grows in Europe, and is not very popular as a finishing wood, while it is too expensive to be used for rough lumber. Formerly great quantities of it were used in the manufacture of gun stocks, so much so as to create a demand for the entire supply. The black walnut is a native of North America, and until about thirty years ago it was used very extensively in the United States for interior finish and furniture, taking the place of oak for these purposes. During recent years, however, the wood has ceased to be popular, and is now very seldom used. This is partly due to the scarcity and consequent high price of the timber. It is a heavy hard wood of coarse texture and of a rich dark-brown color. Very handsome pieces having a beautiful figure may be selected for veneers for furniture and cabinet work. Although the wood shrinks somewhat in drying, it works easily, stands well, and will take a good polish. The tree is large and of rapid growth. It was formerly very abundant in the Allegheny region, and was found from New England to Texas and from Michigan to Florida. White walnut, or butternut, is somewhat like black walnut wood, but is of a lighter color and is not so pleasing when finished. Circassian walnut is beautifully figured, and is sometimes used for piano cases, and costly cabinet work, but it is very scarce and very expensive. Laurel. The tree of this name which is most extensively used in building work is the California laurel, which grows on the Pacific 89 30 CARPENTRY Coast and is seldom seen used in the eastern part of the country. The wood is hard, heavy, and strong, Hght brown in color, and of close grain. The sapwood is considerably lighter in color than the heartwood. The wood takes a very good polish and is quite gener- erally used on the Pacific Coast for cabinet work and interior finish- ing. Osage Orange. This is a southern wood, growing in the Gulf States and seldom seen in the North. The tree is of medium size, bears fruit somewhat resembling an orange, and is protected by large thorns. The wood ranges in color from bright yellow or orange to brown and is hard and strong, though at the same time very flex- ible. It is very durable in damp places, and in positions where it comes in contact with the earth. The wood shrinks somewhat, and checks, but will take a good polish, and it is, therefore, used to some extent for interior finish, but its principal use is for poles and posts, piles, ties, etc. Locust. The locust is a tree which yields wood valuable in con- struction on account of its great durability in exposed positions. The eastern tree is called the black locust, while the tree in the west- ern states is known as the mesquite. There is a slight difference between the two trees, but they belong to the same family. The wood is hard, heavy, and strong, reddish brown in color, and close- grained. It was largely used in the past for long wood pegs called tree-nails and is now used wherever great durability is required. Holly. This wood is very highly prized for use in inlaid work, both on account of its beautiful even grain, and on account of its clear white color. The American tree grows in all the eastern states where it attains to medium size. It is characterized by its ever- green foliage and its red berries. The wood is, cream white in color, and moderately strong. It is easily worked, but is not durable and can not be exposed. Imported Timber. Besides the woods which grow in the United States, a number of others are brought in from foreign lands for use in the best grade of public buildings and private residences. The most popular of these are the mahogany, rosewood, satinwood, French burl, and Circassian walnut. Mahogany comes from Cuba and Mexico, and formerly was obtained also from Santo Domingo and Honduras. Other kinds of 46 CARPENTRY 31 so-called mahogany are also obtained from Africa and India, and some come from South America. The wood is generally imported in the rough log and cut up by the purchasers as it is required. It is easy to work, will take an excellent polish, and stays in place very well if it is properly seasoned. The color varies from very light to deep red, which becomes darker and richer with age. There is also what is called white mahogany, which is golden yellow in color. The wood is very costly and can only be used for the best work. Gener- ally it is used in the form of veneers. Satinwood comes from both the East and the West Indies. It is hard and strong and very durable, but brittle and hard to work. It is so costly as not to be used for anything but the finest cabinet work, for which it is valued on account of its color, which is very light yellow, and its satiny luster. It takes a very good polish. French hurl comes from Persia, and Circassian walnut from near the Black Sea. Both of these woods are very expensive and can be used on that account only in veneers and only for the best work. GENERAL CHARACTERISTICS OF TIMBER In speaking of wood we are accustomed to use certain words to express our idea of its mechanical properties, or of its probable behav- ior under certain conditions. Thus we say that a wood is hard, or tough, or brittle, or flexible, and frequently we use these terms with- out having a clear understanding of just what they mean. A very brief discussion of some of these properties or characteristics of lum- ber will now be given in order that we may see what peculiarities of structure or of growth cause them. Hardness. If a block of wood is struck with a hammer when lying on a bench, the hammer-head will make an impression or dent in the wood, which will be deeper or shallower according as the wood is soft or hard. A wood is said to be very hard when it requires a pressure of about 3,000 pounds per square inch to make an impres- sion one-twentieth of an inch deep. A hard wood requires only about 2,500 pounds to produce the same effecto Fairly hard wood will be indented by a pressure of 1,500 pounds, and soft woods require even less. Maple, oak, elm, and hickory are very hard; ash, cherry, birch, and walnut are hard; the best qualities of pine and spruce are fairly hard, and hemlock, poplar, redwood, and butternut are soft. 41 32 CARPENTRY Toughness. "Toughness" is a word which is often used in rela- tion to timber, and impUes both strength and pHabihty, such as is found in the wood of the elm and the hickory. Such timber will withstand the effect of jars and shocks which would cause other woods like pine to be shattered. Flexibility. Timber is said to be flexible when it bends before breaking instead of breaking off short, or, in other words, a flexible wood is the opposite of one which is brittle. The harder woods, taken from the broad-leaved trees, are usually more flexible than the softer woods, taken from the cone-bearing trees. The wood of the main tree trunk is more flexible than that of the limbs and branches, and moist timber is more flexible than dry wood. Hickory is one of the most flexible woods. Cleavage. Most woods split very easily along the grain, espe- cially when the arrangement of the fibers is simple, as in the conifers. In splitting with an axe, the axe-head acts as a wedge and forces the fibers apart, so that usually the split runs along some distance ahead of the axe. Hard woods do not split so easily as do soft woods, and seasoned wood not so easily as green wood, while all timber splits most easily along radial lines. CARPENTERS' TOOLS Steel Square. It is not only important that the workman should know the character and usefulness of the various materials, but it is also essential that he should be familiar with the steel square, which is the universal tool used to lay out the material. Figs. 16 and 17 show one side of one of the common squares in general use. The three parts are distinguished by special names, the tongue, the blade, and the heel. The longer and wider arm is the blade, the shorter and narrower arm is the tongue, and the point where the two arms meet is called the heel. The numerous applications of the tool are given in detail under the subject of "The Steel Square." Saws. Another very important and much used tool, wherever wood working is done, is the saw, and so much depends upon its careful manipulation and intelligent use that it will not be out of place to devote a few pages to a consideration of the different kinds 42 d o O CO > US X .-3 H § I CARPENTRY 33 Fig. 16. Steel Square of saws and their respective possibilities, as well as to their care and the way in which they should be chosen. There are in general two kinds of saws, which differ from each other in the arrangement of the teeth, and which are intended, one for cutting wood in the direc- tion of the grain, and the other for cutting wood at right angles to the grain. In order to cut the wood in the direction of the grain it is not necessary to cut through very many of the fibers, as the cut is, in general, parallel to them, but it is necessary rather to force the fibers apart without tearing them. Of course it is impossible to cut the wood without tearing the fibers to some extent, but this is the best way to make clear the difference in principle between the two kinds of saws, and an understanding of this difference is necessary in order to appreciate their construction and the proper care of them. The cutting of the wood across the grain, on the other hand, requires a tool made especially with a view to cutting through the fibers as quickly and as easily as possible. Besides these two there are various other special saws de- signed for a particular kind of work, such as the cutting out of key holes, the cutting of dovetails, the cutting of miters, and other oper- ations required in joiners' or carpenters' work. Rip Saw. This saw is designed for cutting along the direction of the grain of the wood, and from this comes its name, which suggests very clearly its purpose. Fig. 18 shows one of these saws, but the shape of the blade varies a great deal with different makers, and some people prefer one shape while others prefer another. The distinguishing feature is the shape and ar- rangement of the teeth, which are shown in detail in Fig. 19 Fig. 17., Steel Square There are always a certain number of teeth to the inch length of the saw. 43 34 CARPENTRY In this kind of a saw, the number is usually four or five, and it will be noticed that one side of the tooth is vertical while the other slopes. The vertical side of the tooth is always toward the front or point of the saw, while the sloping side is always toward the handle. The amount of slope to be given to the teeth of a saw is a matter of Fig. 18. Side View of Rip Saw opinion and can be regulated when the saw is sharpened, or "filed," but the slope should always be a flat one in this kind of a saw, that is, it should make an angle of less than forty-five degrees with the horizontal, or with the line of the back of the saw. It is held by some that the teeth of a rip saw should be straight on the front edge, that Fig. 19. Slope of Teeth in Rip Saw is, that they should have the edge at right angles with the side of the blade, while others maintain that the edge of the tooth should be cut across obliquely, so as to be at an angle of about eighty-five degrees with the side of the blade. A saw may be filed either way, according to the opinion of the owner, the determining factor being usually the kind of wood to be cut and whether the grain is absolutely straight or more or less crooked. In the latter case the edges of the teeth should certainly have a slight bevel so as to give a cutting edge. The bevel should, however, be on alternate sides of adjacent teeth, that is, one tooth should be beveled toward the right and the next toward the left and so on. This arrangement helps to keep the saw straight while cutting, and prevents it from being forced over to one side or the other. Fig. 20 shows a view of the cutting edge of a rip saw, showing the way in which the teeth should be filed. z2n:3 Fig. 20. Blade of Rip Saw Edge-On 44 CARPENTRY 35 Fig. 21. Slope of Teeth of Cross-Cut Saw Lj^ I'gt: "\:sc~ ^^ ^"^C- Fig. 22. Cross-Cut Saw, Blade Edge-On Cross-Cut Saw. As has been already explained above, and as is clearly indicated by its name, this saw is intended for the cutting <)f wood across the grain at right angles to the direction of the fibers. It differs from the rip saw prin- cipally in the size and arrange- ment of the teeth, those of the cross-cut saw being smaller, usu- ally numbering about eight to the inch. The shape of the teeth in the two kinds of saws is also different, as the front of the tooth in the cross-cut saw, instead of being straight as in the rip saw, is inclined backward at an angle of about 115 degrees, while the back of the tooth slopes at an angle of about 125 degrees. The slope of the teeth should be varied according to the hardness of the wood to be sawed, those given above being suitable for soft wood. The bevel on the front of the tooth should also be varied according to the hardness of the wood, so as to give a more or less sharp cutting edge. In the saw described above this bevel should be about sixty degrees, while for harder wood it should be as much as seventy-five degrees. In general, the harder the wood to be cut, the smaller should be the teeth of the saw. Fig. 21 shows a cross-cut saw with the slope of the teeth indicated, and Fig. 22 shows how the teeth should be filed. The cross-cut saw is also known as the "panel saw." Hand Saw. There is a saw, which is much used for general work, which combines the qualities of the rip saw and the cross-cut saw. It is called the "hand saw," and is a cross between the other Fig. 23. Back Saw two. It may be used for either cutting with the grain or against it, but in any case does not do such good work as the special saw which is intended for the particular kind of work which is at hand. 45 36 CARPENTRY Fig. 24. Slope of Teeth of Back Saw Back Saw. Fig. 23 shows a saw which is known as a back saw, probably because of the extra piece on the back which Umits the depth to which the saw will cut. It is also called a "tenon saw." There are a number of different kinds, varying in the width of the blade and in the length of the saw, and they are used for various special purposes, usually in miter boxes and for sawing bevels on moldings. Fig. 24 shows the arrangement of the teeth on a back saw. It will be seen that the front of the tooth is nearly straight, and that the slope of the back is very sharp, making the number of teeth to the inch more than in the rip or cross-cut saws. Keyhole Saw. Fig. 25 shows a set of saw blades which are intended to be fastened in turn to the same handle and used for various purposes. These blades are very thin and can be used for cutting out small holes such as keyholes, and it is for this reason that such saws are called "keyhole saws." The teeth are in general similar to those of the back saw, but are usually smaller. Great care must be exercised in the filing of a saw, to give it the proper "set" to enable it to do the work required of it, and this work is better left to an expert. Most carpenters, however, like to know how to file their own saws and to keep them in good condi- tion. A great deal has been written on this subject both in books and in trade papers, but it is almost impossible to describe, in writing, the proper methods. It is a part of the carpenter's trade which must be learned by experiment and by watching the older workmen. Fig. 25. Keyhole Saw with Detachable Blades Planes. Timber comes from the mills rough from the saw, and before it can be used for any finished work it must be prepared to receive paint or other kinds of finish. This preparation consists in a smoothing or planing which can be carried to any extent, including 46 CARPENTRY 37 Fig. 26. Jack Plane sandpapering or even polishing. The instruments used for the rougher part of this work are called planes, after which, if more smoothing is required, come scrapers and sandpaper. There are a great many different kinds of planes, but the principle of all of them is the same. They consist of a sharp blade, or knife, in the form of a chisel, which is held in a large block of wood or iron by means of clamps, so that the knife can be kept steady and guided easily. The knife projects at the bottom of the back through a slot, and takes off a shaving which is larger or smaller according to the projection of the knife. For smoothing, the cutting edge of the knife must be absolutely straight and must be clamped into the block in such a way that the projection will be exactly the same all along the edge. Any imperfections in the edge of the knife will be repeated on the surface of the wood. Planes are in general of two kinds, namely, "jack planes" and "trying planes." Jack Planes. The jack plane is used for the rougher work to give the preliminary smoothing after the lumber comes from the mill. It is bigger and, as a rule, heavier than the finishing planes, and is almost always made of wood, while the others are often made Fig. 27. Wood-Bottom Smooth Plane with Handle Fig. 28. Wood-Bottom Smooth Plane without Handle of iron. Fig. 26 shows a view of a jack plane. The handle is neces- sary to push the block forward, and it is usually necessary to bear down heavily on the forward end of the block to keep the knife down into the wood. Trying and Smoothing Planes. The smoothing plane is usually much smaller than the jack plane, as it is not expected to take off 47 38 CARPENTRY so much material and there does not have to be so much leverage. In construction it is similar to the jack plane, and may be made of either wood or iron. Very often, however, it is without a handle, as no great force is required to oper- ate it. The trying plane is longer than the jack plane and is used after it so as to obtain a truer surface on the piece of timber than is possible with the jack plane. It is also used for edging boards, and it is narrower than either the jack plane oi- the smoothing plane. Figs. 27 and 28 show two wood- bottom smooth planes, one with a handle and one without, and Fig. 29 shows a smooth plane with , an iron bottom. In Fig. 30 is shown a sectional view of both a wood and an iron smooth plane, with the various pieces numbered, and in Fig. 31 are shown some of these same pieces separately with the same numbers attached to each. The names of the various parts are as follows: Fig. 29. Iron-Bottom Smooth Plane Fig. 30. Sections of Wood-Bottom and Iron-Bottom Smooth Planes No. 1 is the "plane iron," and should be made of steel, well tempered and ground. It should be throughout of the same thickness. No. 2 is the "plane iron cap," also of steel, the purpose of which is to protect the plane iron. No. 3 is the "plane iron screw," which fastens the plane iron cap to the plane iron. No. 4 is the "cap" (or "cap iron"), which holds the plane iron in place, and it is fastened to the "frog" by means of the "cap screw," No. 5. No. 6 is the "frog," which acts as a support for the plane irons, and which is fastened to the body of the plane by the "frog screw," No. 10. 48 CARPENTRY 39 No. 7 is the "Y" adjustment, the end of which fits into an opening in the plane iron cap, and makes possible the close adjustment of the position of the plane iron. The adjustment is made by means of the brass "adjusting nut," No. 8. No. 9 is the "lateral adjustment," by means of which the plane iron can be shifted very slightly sideways in the plane, if necessary, so as to bring it parallel with the edge of the bottom of the plane, where it passes through the slot. No. 11 is the "handle," which is fastened to the bottom of the plane by the "handle screw," No. 15, and by the "handle bolt and nut," No. 13. No. 12 is the "knob," fastened to the bottom by the "knob bolt and nut," No. 14. No. 16 is the "bottom" of the iron plane, while No. 18 is the "bottom" of the wood plane. No. 17 is called the "top. casting" and occurs only on the wood bottom plane o / Fig. 31. Details of Parts of Smooth Plane Nails. In general, nails are of two kinds, namely, cut nails and wire nails, the difference between the two kinds being in the material and the method of manufacture. Cut Nails. The cut nails, also called plate nails, are stamped out of a flat iron plate, in alternate, slightly wedge-shaped pieces, and the head is afterward formed on the large end of each piece. The cut nails are made in three classes, according to finish, and are called, respectively, "common," "casing," and "finish" nails. The nails known as "finishing nails," however, are far too rough for fine finished work. The length of the nail is regulated according to the "penny," which formerly had reference to the weight, but which now 49 40 CARPENTRY is purely arbitrary. Thus a three penny nail is 1| inches long; four penny, 1| inches; five penny, If inches; six penny, 2 inches; seven penny, 2| inches; eight penny, 2| inches; nine penny, 2f inches; ten penny, 3 inches; twelve penny, SJ inches; sixteen penny, 3^ inches; twenty penny, 4 inches; thirty penny, 4| inches; forty penny, 5 inches; fifty penny, 5| inches; and sixty penny, 6 inches. The speci- fications which have just been given for cut nails also hold good for wire nails. Wire Nails. Wire nails are rapidly replacing the cut nails in general use. They are now very nearly the same price and are very much stronger, so that they do not buckle up when driven into hard wood, and they are not nearly so liable to split the wood on account of their cylinder-shaped shaft, which is the same size through- out its entire length. They are made from wire, which is cut in lengths by machinery and pointed and headed. They can also be ribbed or barbed, if desired, which gives them a stronger hold on the wood. They are made with various kinds of heads, some being large and flat, so that the nail can be easily withdrawn, while others are very slightly larger than the shaft of the nail and can be made almost invisible in the finished work. For framing, large nails should be used, from 4 to 6 inches in length. For the rougher exterior and interior finish, such as sheathing and rough flooring, nails about 3 inches long are suitable, while for the finer inside finish smaller nails from 2| inches down to 1| inches should be used. Roofing should be put on with special galvanized or copper nails so as not to rust out. Screws. Screws are now used in building work to a much greater extent than was formerly the custom, largely on account of their decreased cost. They have the advantage over nails, as they do not split the wood, and they can be easily withdrawn when desired, without injuring the work materially. There are a great many different kinds of wood screws, which vary as to the shape of the head, the size of the shaft, and the length. They are made in about the same lengths as those given above for nails, and with both round and flat heads. Screws can be had in iron, steel, copper, bronze, and brass. They are also made with the heads silver- and gold-plated, or lacquered to match finishing hardware. 50 CARPENTRY 41 LAYING OUT Having now considered the material and tiie most important of the tools with which the carpenter performs his work, we shall pass to a consideration of the work itself, and see how a building of wood construction is put together. Ground Location. In undertaking the construction of any building, the first thing to do is to make a thoughtful examination of the piece of ground upon which the structure is to be placed. This is very important as the character of the soil upon which a dwelling is located will very largely determine its sanitary condition, and will influence to a great extent the health of the occupants. Very often a difference of a few yards in the location of a building will be enough to cause the difference between a perfectly dry cellar and one which is constantly flooded with water. Water is, indeed, the one thing above all others which must be- guarded against, since it is impossible to keep it out of a cellar which is sunk in damp ground, unless some elaborate system of waterproofing is employed. Ground Water. Below the surface of the earth there is always to be found what is known as "ground water." This stands prac- tically always at a level, and is not met with so near the surface on a slight knoll or other elevation as in a depression. If possible, a house should be located on comparatively high land, so that the floor of the cellar does not come below the ground-water level. Below the surface of a hill, however, there may be a stratum of rock which will hold the rain water and prevent it from sinking at once to the ground-water level. Such a ledge of rock causes the water to collect and then flow off in small subterranean streams, which will penetrate the walls of a cellar if they happen to be in their path. A good way to discover the depth of the ground-water level or the existence of rock ledges beneath the surface of the ground, is to dig a number of small, deep holes at various points of the site. These should be carried below the proposed level of the cellar bottom. A suitable location for the building may thus be chosen. If, however, it is not easy to make so thorough an examination of the site as this would allow, another method may be employed. This consists in the use of an instrument called an "auger," which 51 42 CARPENTRY is very much like an ordinary carpenter's auger or bit, though much larger. The auger generally used is about 2 inches in diameter. It is driven into the ground, and as it descends into the hole which it bores it brings to the surface small portions of all the different kinds of material through which it passes. This material may be preserved and examined at leisure. The character of the site may be determined in this way. Staking Out. When the approximate position of the structure has been decided upon, the next step is to "stake it out," that is, the position of the corners of the building must be located and marked in some way, so that when the excavation is begun the workmen may know what are the exact boundaries of the cellar. This "staking out" should always be carefully attended to, no matter how small the building may be. In works of importance it is best to have the work done by an engineer, but on small work it is customary for the contractor or the archi- tect to attend to it. It is well to have at hand some instrument with which angles can be accurately measured, such as a transit; but the work can be done very satisfactorily with a tape measure and a "mason's square." This simple instrument is composed of three sticks of timber nailed together as shown in Fig. 32, to form a right-angled triangle. It is important that the tape used should be accurate, a steel tape being always preferable, and that the mason's square should give an exact right angle. A mistake in the staking out may cause endless trouble when the erection of the building itself is begun, and it is then too late to remedy it. There are several different lines which must be located at some time during the construction, and they may as well be settled at the start. These are: The line of excavation, which is outside of all; the face of the basement wall, inside of the excavation line; and in the case of masonry building, the ashlar line, which indicates the outside of the brick or stone walls. In the case of a wood structure only the two outside lines need be located, and often only the line of the excavation is determined at the outset. Fig. 32. Mason's Square 52 CARPENTRY 43 The first thing to do is to lay out upon the ground the main rectangle of the building, after which the secondary rectangles, which indicate the position of ells, bay windows, etc., may be located. Starting at any point on the lot where it is desired to place one corner of the building, a stake should be driven into the ground and lines laid out parallel and perpendicular to the street upon which the structure is to face. At the ends of these lines, which form sides of our rectangle, the lengths of which are determined by the dimensions of the building, other stakes should be driven, which define the direction and the length of the building. The exact location of the ends of the line may be indicated by a nail driven into the top of each stake. After these lines have been thus laid out, others may be laid out perpendicular to them at the ends, with the aid of the mason's Fig. 33. Diagram Showing Wrong Ground Layout Fig. 34. Batter Board to Indicate Layout square and the tape measure. The accuracy of the right angle may be checked by the use of the "three-four-five" rule. This rule is based upon the fact that a triangle, whose three sides are, respectively, 3, 4, and 5 feet long, is an exact right-angled triangle, the right angle 53 44 CARPENTRY being always the angle between the 3-foot and the 4-foot sides. This fact may be proven by applying the well-known theorem, which states that the length of the hypotenuse of a right-angled triangle is equal to the square root of the sum of the squares of the other two sides. The rule may be used as follows: Lay off on one of the side lines already laid out on the ground any multiple of 3 feet, as 9 feet or 12 feet. On the other line, pre- sumably at right angles to the first one, lay off the same multiple of 4 feet, as 12 feet or 16 feet. Now a straight line measured between the points so obtained, should have a length equal to the same multiple of 5 feet, as 15 feet or 12 feet. If this is not found to be the case the angle laid out is not an exact right angle, and instead of a rectangle we have a parallelogram as shown in Fig. 33. This will not do at all, and the inaccuracy must be corrected. It is possible to lay out the right angle in the first place by this same method, using two flexible cords, respectively, 4 feet and 5 feet long. The end of the 4-foot cord should be fastened at the end of the side line of the building, and the end of the 5-foot cord should be fast- ened on this same side line, 3 feet away from the corner. When the loose ends of both cords are held together, and the cords are both drawn taut, the point ^^^- ^^' B?tterB^o*ard ^ '^^^'^ °^ whcrc the cuds meet will be a point on the side line of the building perpendic- ular to the first side line. It is evident that this point must be just 4 feet from the corner, and that the distance between it and the point on the other side line, 3 feet from the corner, must be 5 feet. After all the corners of the building have been located, their position should be indicated by the use of "batter boards." One of these is shown in Fig. 34. It will be seen that it consists of a post A, which is set up at the corner, together with two horizontal pieces BB, which extend outward for a short distance along the sides of the rectangle that has been laid out. The horizontal pieces may be braced securely as shown, and the whole will be a permanent indication of the position of the corner. Notches may be cut in the top of the horizontal pieces to indicate the position of the various lines, 54 CARPENTRY 45 and cords may then be stretched between the notches from batter board to batter board. These cords will give the exact location of the Hnes. Another way to indicate the position of the Hnes is by driving small nails into the tops of the batter boards instead of cutting notches in them; but nails may be withdrawn, while the notches when they are once cut, can not easily be obliterated. Batter boards should always be set up very securely, so that they will not be displaced during the building operations. If there is danger that the form of batter board shown in Fig. 34 may be displaced, because of the large size of the structure and the length of time during which they must be used, the form shown in Fig. 35 may be substituted. Two of these at right angles to each other must be placed at each corner. 55 ■ ^W! iwlliii liiiiAwiiii ■111 iiiiiiiiii II -^^** . '?.■»»! *W^i|^^^^^| •. - - HOUSE IN THE WHITE MOUNTAINS, NEW HAMPSHIRE The Boulders, which are Plentiful in this Region, have been Used to Good Advantage. HOUSE NEAR PHILADELPHIA, PA. It is Verily a Part of the Landscape, CARPENTRY PART II FRAMING After the building has been laid out, and the batter boards are in place, the next work which a carpenter is called upon to do is the framing. This consists in preparing a skeleton, as we may say, upon which a more or less ornamental covering is to be placed. Just as the skeleton is the most essential part of the human body, so is the frame the most essential part of a wood building; and upon the strength of this frame depends the strength and durability of the structure. When the carpenter comes to the work, he finds everything prepared for him; the cellar has been dug and the foun- dation walls and the underpinning have been built. It is his busi- ness to raise the framework on them. First is the wall, then the floors, and then the roof. Therefore, the subject may be subdivided, and considered under these three main headings. In connection with the walls we may consider the partitions as well as the outside walls, and in connection with the floors we may consider the stairs, while the roof may be taken as comprising the main roof and also subordinate roofs over piazzas, balconies, and ells. This covers all the framing that will be found in a wood building, except special framing. (See page 147, Part III.) Whatever framing there is in a brick or stone building is similar to that in a wood building, with the slight differences which may be noted as we come to them. JOINTS AND SPLICES IN CARPENTRY Before beginning a description of the framing, it will be well to consider the methods employed in joining pieces of timber together. The number of different kinds of connections is really very small, and the principles upon which they are based may be mastered very quickly. Copyright, 1912, by American School of Correspondence. 57 48 CARPENTRY All connections between pieces of timber may be classified as joints or as splices. By a "splice" we mean a connection between two pieces which extend in the same direction, as shown in Fig. 36, and each one of which is merely a continuation of the other. The only reason for the existence of such a connection is the fact that sticks of timber can be obtained only in limited lengths and must, therefore, very often be pieced. By a "joint" we mean any connec- tion between two pieces which come together at an angle, as shown in Fig. 37, and which are, therefore, not continuous. Such a connection may be required in a great many places, and especially at the corners of a building. Joints. The principal kinds of joints to be met with in car- pentry are the "butt joint," the "mortise-and-tenon joint," the "gained joint," the "halved joint," the "tenon-and-tusk joint," and the "double-tenon joint.". Butt Joint. This is the most simple of all the joints, and is made by merely placing the two pieces together with the end of Fig. 36. Example of a Splice Fig. 37. Example of Plain Joint Fig. 38. Square Butt Joint one piece against the side of the other and naihng them firmly to each other, after both have been trimmed square and true. Such a joint is shown in Fig. 38. The two pieces are perpendicular to each 58 < M fs K ri M o S tsi 3i r/1 o t" h BQ ^ S ^-M 1-1 aj N ^a « ^^-- iij oc::: 0*- 1 < « a «-^ i_ «,- 10 (/)- 1 :^ ■ & cS f- "3 cc: -" CARPENTRY 49 other and neither piece is cut. The nails are driven diagonally through both pieces, an operation which is known as "toe-nailing" and are driven home, if necessary, with a nail set. This is called a "square" butt joint. Fig. 39 shows two pieces which are not perpendicular to each other. They are trimmed to fit closely together, and are then nailed in place. Such a joint is called an "oblique" butt joint. The butt joint does not make a strong connection between the pieces, and should not be used if much strength is required. It depends entirely upon the nails for its strength, and these are very likely to pull out. This form of joint is sometimes modified by cutting away a part of one of the pieces, so that the other may set down into it as shown in Fig. 40, the square joint at A, and the oblique joint at B. Fig. 39. Oblique Butt Joint Fig. 40. Special Types of Square and Oblique Butt Joints This gives much additional strength to the joint, especially in the case shown at B, where there may be a tendency for one piece to slide along the other. 59 50 CARPENTRY Mortise-and- Tenon Joint. From the modified butt joint it is only a step to the "mortise-and-tenon" joint, which is formed by cutting a hole called a "mortise" in one of the pieces of timber, to receive a pro- jection called a "tenon" which is cut on the end of the other piece. This arrange- ment is shown in Fig. 41. The mortise is shown at A, and the tenon is shown at B. It will be noticed that there is a hole bored through the tenon at C, and that another hole is bored in the mortised piece at D. These holes are so placed that when the pieces are joined together, a wood pin may be driven through both holes, thus preventing the tenon from being with- The pin should always be inserted in a Ordinarily this pin is of hard wood, even Fig. 41. Square Mortise-and- Tenon Joint drawn from the mortise mortise-and-tenon joint. when the pieces to be joined together are themselves of soft wood, and it may be of any desired size. Round pins from | inch to | inch in diameter are ordinarily employed, although it may sometimes be found better to use a square pin. The form of mortise-and-tenon joint described above may be Fig. 42. Form of Bridge or Straddle Joint Fig. 43. Anothor Form of Bridge or Straddle Joint used wherever the pieces are perpendicular to each other. When, however, the pieces are inclined to each other, a modification of the 60 CARPENTRY 51 above joint known as the "bridge" or "straddle" joint is employed. This joint is shown in Figs. 42 and 43. It is similar to the square mortise-and-tenon joint, having a similar mortise and tenon, but these are cut in a slightly different way. In Fig. 42 the tenon A is cut in the end of the inclined piece and fits into the mortise B cut in the other piece. In Fig. 43 the mortise A is cut in the end of the inclined piece and the tenon B is cut in the other piece. Gained Joint. The joints which have so far been described are applicable only where the members are subjected to direct compres- sion, as in the case of posts or braces, or in certain cases where direct tension is the only force acting on the pieces. When bending and shearing are to be expected, as in the case of floor beams connecting to sills or girders, a slightly different sort of joint must be employed. One of the most common joints for such places is a modifi- cation of the mortise-and-tenon joint which is known as the "gained joint." An example of this form of connection is shown in Fig. 44, and it may be seen that .the end of one piece is tenoned in a peculiar way. The tenon proper is the part A-B-C and this tenon sets into a corre- sponding mortise cut in the other piece as shown. It is evident that the tenon can not be held in place by a pin, but it may be secured by naihng. The reason for this pecuHar form of tenon may be explained as follows: A floor beam, or any other timber, which is loaded trans- versely, has a tendency to fall to the ground, and must be supported at its ends either by resting directly on a wall or sill, or by being mortised into the latter member. Moreover, in order that the end of the piece resting on the support, may not be crushed or broken, a certain amount of bearing surface must be available. This same bearing surface must be provided in every case no matter whether the timber rests directly on the top of the sill or is mortised into it. Of course the simplest connection is obtained by resting the transverse piece directly on top of the sill without cutting either Fig. 44. Gained Joint 61 52 CARPENTRY Fig. 45. Tenon-and-Tusk Joint piece; but such a joint is not stiff and strong, and it is often neces- sary to bring the timbers flush with each other at the top or at the bottom. For this reason a mortised joint is used; and in order to obtain the required amount of bearing surface without cutting the piece too much, the form of tenon shown in Fig. 44 is em- ployed. The available bearing area here is furnished by the sur- faces D-A and B-C and it may easily be seen that this area is the same as would be available if the piece rested directly on top of the sill. The operation of cutting such a tenon and mortise is known as "gaining," and one piece is said to be "gained" into the other. Tenon-and-Tusk Joint. A joint in very common use in such situations as those which have just been mentioned is a develop- ment of the gained joint which is called the "tenon-and-tusk" or the "tusk-tenon" joint. This joint is shown in Fig. 45. The char- acteristic feature of this joint is to be found in the peculiar shape of ^^^^^ the tenon which is cut in the end of one of the pieces to be joined, as shown in the figure. It may be seen that there is a small square tenon B cut in the ex- treme end of the piece, and that in addition to this there are other cuts C which constitute the "tusk." The bearing area is furnished partly by the under side of the tenon and partly by the under side of the tusk. This joint makes a very good connection, and the cutting of the mortise does not weaken the piece of timber so much as does the mortise for a gained joint. It is especially applicable when it is desired to have the two pieces flush on top, although it may also be used in other positions. When the top of the tenoned piece must project above the top of the mortised piece, the tenon may be cut as shown in Fig. 46. Fig. 46. Tenon-and-Tusk Joint with Specially Cut Tenon Piece 62 CARPENTRY 53 There are several ways of securing the tenon in place. The simplest is that shown in Fig. 47, where the pin B is passed through the tenon A and the mortised piece so as to hold the tenon securely in place. Another scheme is to cut the square tenon a little longer, Fig. 47. Pinned Tenon-and-Tusk Joint Fig. 48. Pegged. Tenon-and-Tusk Joint as shown in Fig. 48, so as to pass clear through the mortised piece, and to fasten it with a peg B on the other side. The peg may be cut slightly tapering, as shown, so that when it is driven in place it will draw the pieces together. Still another plan is shown in Fig. 49. Here a small hole is cut in the header some distance back from the tenon and a nut C is placed in it, while a bolt B is passed through a hole bored lengthwise in the header to receive it. The bolt passes through the nut, which may be screwed up tight, thus drawing the pieces closely together and making the joint secure. In tightening this up, it is the bolt which must be turned, while the nut is held stationary inside of the square hole in which it is inserted and which is just large enough to receive the nut and a wrench. Double Tenon Joint. Fig. 50 shows a form of tenon joint called the "double tenon" joint, which is not very extensively used at the present time but which has some advantages. As may be readily seen, there are two small tenons A and B through which a pin may be passed if desired. Halved Joint. A form of joint which may be used to connect two pieces which meet at a corner of a building, is shown in Fig. 51. Fig. 49. Bolted Tenon-and-Tusk Joint 63 54 CARPENTRY This is known as the "halved" joint from the fact that both pieces are cut half way through and then placed together. The pieces are held in place by nails or spikes. If one piece meets the other near the center instead of at the end of the piece, and if there is danger that the two pieces may pull away from each other, a form of joint called the "dovetail" halved joint is used. This is shown in Fig. 52. Both the tenon and the mortise are cut in the shape of a fan, or dovetail, which prevents the two pieces from being pulled apart. This joint may also be cut as shown in Fig. 53, with the flare on only one side of the tenon, the other side being straight. Splices. As already explained, a splice is merely a joint between two pieces of timber which extend in the same direction, and is som^e- times necessary because one long piece can not be conveniently or cheaply obtained. The only object in view, then, is to fasten the two pieces of timber together in such a way that the finished piece Fig. 50. Double Tenon Joint Fig. 51. Halved Joint Fig. 52. Dovetail Halved Joint will be in all respects equivalent to a single unbroken piece, and will satisfy all of the requirements of the unbroken piece. This is really the only measure of the efiiciency of a splice. 64 CARPENTRY 55 Dovetail Halved Joint with One Flare There are three kinds of forces to which a piece may be sub- jected, namely: Compression, tension, and bending. A sphce which would be very effective in a timber acted upon by one of these forces might be absolutely worth- less in a piece which must resist one of the other forces. We have, therefore, three classes of splices, each designed to resist one of these three forces. Splices for Compression. The simplest splices are those in- tended to resist compression alone, and of these the most simple is that shown in Fig. 54. Fig. 53. This piece is said to be "fished"; the two parts are merely sawed off square and the ends placed together. A couple of short pieces A-A, called "fish plates," are nailed on opposite sides to keep the parts in fine. In the sphce shown in Fig. 54, the sphcing pieces are of wood, and ordinary nails are used to fasten them in place, but in more important work thin iron plates are used, the thickness being varied to suit the conditions. They are held in place by means of bolts with washers and nuts. If for any reason it is desired not to use plates of this kind, four small pieces called dowels may be used, as indicated in Fig. 55. These dowels may be set into the sides of the timbers to be spliced, so that they do not project at all beyond the faces of these pieces and a very neat job may thus be obtained. It is but a step to pass from this simple splice to the "halved" sphce shown in Fig. 56. It will be noticed that it is much like the halved joint described above, the only difference being that the pieces are continuous, instead of being perpendicular to each other. The nature of the splice will be easily understood from the figure without further Fig. 54. Fished Splice 65 56 CARPENTRY explanation. A modification of this which is somewhat more effec- tive, is shown in Fig. 57. The cuts are here made on a bevel in such a way that the parts fit accurately when placed together, and the splice is called a "beveled" splice. Fig. 55. Doweled Splice Fig. 56. Halved Splice The halved splice is perhaps the best that can be used to resist direct compression, and when it is combined with fish plates and bolts, as shown in Fig. 58, it may be used in cases where some tension is to be expected. It will be noticed that in Fig. 58 the ends of the timbers are cut with a small additional tongue A, but this does not materially strengthen the splice and it adds considerably to the labor of forming it. In general it may be said that the simplest splice is the most effective. Fig. 57. Beveled Splice Fig. 58. Halved Splice with Fish Plates and Bolts Whenever the pieces are cut to fit into one another, as they do in the halved and beveled splices, the splice is known as a "scarf' splice, and the operation of cutting and joining the parts is called 66 CARPENTRY 57 "scarfing." Scarf splices are used, as we have already seen, both alone and in combination with fish plates. The fished splice is always the stronger, but the splice where scarfing alone is resorted to has the neatest appearance. Splices for Tension. There are several common forms of splices for resisting direct tension. These differ from each other mainly in the amount of labor involved in making them. The simplest of them is shown in Fig. 59, and it will be seen that it is only a slight modification of the halved splice used for resisting compression. It is evident that the pieces can not pull apart in the direc- tion of their length until the timber crushes along the face marked A-B, or shears along the dotted line A-C. By varying the dimensions of the splice it may be made suit- able for any situation. The parts are held closely together by the light fish plate shown in the figure, which also incidentally adds something to the strength of the splice. Instead of cutting the ends of the beams square, as shown in Fig. 59, they frequently are cut on a bevel as shown in Fig. 60, and a further modification may be introduced by inserting a small "key" of hard wood between for the pieces to pull against, Fig. 61. This key is usually made of oak and may be in two parts, as shown in Fig. 59. Squared Splice for Tension Fig. 60. Beveled Splice for Tension Fig. 61. Beveled and Keyed Splice for Tension Fig. 62, each part in the shape of a wedge, so that when they are driven into place a tight joint may be obtained. The two wedge- shaped pieces may be driven in from opposite sides, the hole 67 58 CARPENTRY Fig. 62. Key for Beveled and Keyed Joint being a little smaller than the key. If the key is made much too large for the hole, however, a so-called "initial" stress is brought into the timbers, which uses up some of their strength even before any load is applied. This should be avoided. If it is desired, two or more keys may be employed in a splice, the only limiting condition being that they must be placed far enough apart so the wood will not shear out along the dotted line shown in Fig. 61. Another feature of the splice here shown is the way in which the pieces are cut with two bevels on the end instead of one. One bevel starts at the edge of the key and is very gradual, the other starts at the extreme end of the piece and is rather steep and sharp. These bevels can be used only in joints which resist tension alone. If such a splice were subjected to compression, the beveled ends would slide on each other and push by each other very easily, except as they are prevented from so doing by the fish plates, if these are used. Tension Splice with Fish Plates. The splices for tension which have so far been described have all been scarf joints, but there is a fished splice which is very commonly used for tension. This splice is shown in Fig. 63. The fish plates, in this case of wood, are cut into the two pieces to be spliced, so as to hold them firmly together. The pieces can not be pulled apart until one of the plates shears off along the dotted line A-B. The dis- tance C-D must also be made large enough so that the piece will not shear. This splice is very often used for the lower chords of the various forms of wood trusses, and it is considered one of the best that has been devised for resisting direct tension. Splices for Bending. It sometimes happens that a piece which Fig. 63. Tension Splice with Fish Plates 68 CARPENTRY 59 Fig. 64. Splice Designed for Bending Strains is subjected to a bending stress must be spliced, and in this case the spHce must be formed to suit the existing conditions. It is well known that in a timber which is resisting a bending stress the upper part of the piece is in compression, and the tendency is for the fibers to crush, while the lower part of the piece is in tension, and the tendency is for the fibers to pull apart. To provide for this, a form of splice must be selected which combines the features of the tension and compression splices. Fig. 64 shows such a splice. The parts are scarfed together, as is the case with other splices de- scribed, but in this case the end of the top piece is cut off square to offer the greatest possible resistance to crushing, while the under- neath piece is beveled on the end as there is no tendency for the timbers to crush. We have already seen that in the lower part of the splice, there is a tendency for the parts to be pulled away from each other. In order to prevent this, a fish plate. A, is used, which must be heavy enough to take care of all the tension, since it is evident that the wood can not take any of this. The plate must be securely bolted to both parts of the splice. There is no need of a fish plate on the top of the pieces because there is no tendency for the pieces to pull apart on top, and the bolts shown in the figure are sufficient to pre- vent them from being displaced. In any case where it is not desirable to scarf the pieces in a splice subjected to bending, the form of butt joint shown in Fig. 65 may be used. The plates, either of wood or iron, are in this case bolted to the sides of the pieces. If wood is used, of course the plates must be made very much heavier than if iron is used. In either case they must be large enough to take care of all the bending stress, and a sufficient number of bolts must be used to fasten them securely to both parts of the splice. Fig. 65. Butt Joint with Plates 6& 60 CARPENTRY JOINTS AND SPLICES IN JOINERY We have considered a few of the most important joints and spHces used in the putting together of rough framing, and we will now take up some of the methods used in the joining together of finished work, where more care is necessary and where the joint or splice must very often be concealed from view. Much ingenuity has been exercised in devising some of these concealed joints, and great credit is due to the workmen, unfortunately unknown, who first invented them. Splices. Plain Butt Splice. Fig. 66 shows the simplest kind of splice which can be used, similar in principle and construction to the butt joint already described. Here the pieces are simply planed off square and true on the ends and glued together with nothing but the glue to hold them. It is evidently not a very strong splice and should not be used where any tension or bending is likely to come at the point where the splice is made. Splice with Spline. Fig. 67 shows a splice which is a slight advance over the simple butt splice. It is formed by ploughing the ends of the pieces to be spliced after they have been finished square and true, and inserting into the slot thus formed a third piece, which is called a "spline" or a "tongue." The spline is about 1 inch or less in width, and about f or i^ inch thick. Its length is regulated by the width of the pieces to be spliced together. As will be explained later, this form of connection is made use of in the construction of the better class of doors. Tongued-and- Grooved Splice. This form of splice is somewhat similar to the splice with splines, the difference being that only one of the pieces is ploughed, and the other is rabbeted on both sides so as to leave a projecting portion called a "tongue" which fits into the groove formed by ploughing the other piece and is fastened there securely with glue. The tongue should be about f inch thick, and should project about the same distance from the main body of the piece. The groove in the other piece must, of course, be of corre- sponding dimensions. The figures given are for l|-inch stuff, which is the most common thickness of lumber used in cabinet work and interior finishing. For other thicknesses they should, of course, be varied. The tongued-and-grooved splice, Fig. 68, is used extensively in flooring. TO CARPENTRY 61 Rabbeted Splice. In Fig. 69 is shown what is known as a "rab- beted splice." It is similar to the halved splice described before but depends upon glue or small nails for its strength. It must be much more carefully made than the rough halved splice. As will be seen each piece is rabbeted on one side so that when put together they fit into each other perfectly. The tongue should here be about one half of the thickness of the piece and its projection from the main body of the piece should be about equal to its thickness. If this ig. 06. Plane Fig. 67. Splice Fig. 68. Tongued- Fig. 69. Rab Butt Splice with Spline and-Grooved Splice beted Splice tongue is made too thin and projects too much, it is liable to curl up, as the wood shrinks in drying, and make ugly ridges on the finished work besides leaving the splice open. Filleted Splice. A form of splice which is little used in this country, but which can occasionally be worked to advantage, is the filleted splice which is shown in Fig. 70. It is made by rabbeting the two pieces to be spliced, as in the case of the rabbeted splice, but this time both on the same side, and a third piece called a "fillet," which is somewhat like the spline in the splice with a spline, is inserted 71 62 CARPENTRY in the hollow space so as to join them together. This prevents any possibility of an open joint. The fillet is generally made somewhat less than one half the thickness of the pieces to be spliced, and is about 1 inch in width. Battened Splice. A variation of the filleted splice, which is quite generally made use of where great strength is not required and it is only necessary to cover up the butt splice neatly, is what is called a "battened splice," Fig. 71. As will be seen this con- Fig. 70. Filleted Splice ig. 71. Bat- Fig. 72. Tongued- Fig. 73. Tongued tened Splice Grooved-and- Grooved-and- Rabbeted Splayed Splice Splice sists simply in covering the butt splice with a small piece called a "batten." The batten should be about the same size as that given above for the size of the fillet, but can be made large or small as desired. Tongued-Grooved-and-Rabheted Splice. A combination splice, which combines the advantages of the tongued-and-grooved and the rabbeted splices, is shown in Fig. 72. The groove is ploughed in the end of one piece and the tongue is left projecting on the end of the other piece while, in addition to this, one of the pieces is rabbeted 72 CARPENTRY 63 against the other. The tongue should be about | inch thick and should project about the same amount from the end of the piece. One piece of the splice should be rabbeted against the other a distance of about f inch. This splice is considerably stronger than the simple tongued-and-grooved splice, but it is a great deal harder to make and is more expensive as regards both material and labor. Tongued-Grooved-and-Splayed Splice. Another variation of the tongued-and-grooved splice consists in the introduction of a splay on one side of the tongue and a corresponding splay on one side of the groove so that they fit into each other. Fig. 73 shows this arrangement. It makes a very neat form of splice and it looks well, but it is apt to be less strong than the simple tongued-and-grooved splice and much weaker than the tongued-grooved-and-rabbeted splice, though stronger than the simple rabbet. This is, of course, a very troublesome and expensive form of splice to make, and it is, in consequence, seldom used. Joints. Miters. A miter is a joint between two pieces which come together at a corner at an angle of ninety degrees with each other. Strictly such a joint can be called a mitered joint only when each piece is beveled off so that each will come to a sharp edge at the corner. There are, however, a number of Fig. 74. Miter Joint different methods of cutting the pieces so that they will come together in this way. The simplest method is to cut off each piece along the edge at a bevel of forty-five degrees, so that when they are put together they will make an angle of ninety de- grees with each other. This method is shown in Fig. 74. In practice, however, it is very difficult to make a perfect joint of this kind. The joint is very apt to open on the outside of the corner and leave an unsightly crack there, and great care must be exercised to see that the bevels are cut to exactly forty-five degrees, as the least variation will cause endless trouble. Miter with Spline. A simple mitered joint may be made stronger by the introduction of a spline, which is inserted at the joint in a direction perpendicular to it. This is shown in Fig. 75. The sphne 73 64 CARPENTRY Fig. 75. Miter with Spline used in this way Is also known as a "feather." It strengthens the joint very considerably, and a joint of this kind is a great improve- ment over the simple mitered joint. The spline or feather should be about f inch wide and about | inch thick. Its length, of course, varies with the width of the pieces which meet at the joint. Great care must be taken in plough- ing out the grooves into which the spline fits, for if they are not exactly the same distance from the corner on each of the pieces the finished joint will not be neat and true. Rabbeted Miter Joint. There are two or three vari- ations of the simple mitered joint made by rabbeting one piece on the other at the corner, so that the miter goes only part way through each piece. One of these joints is shown in Fig. 76, in which only one of the pieces is rabbeted and the other piece has a simple miter. This form of joint can only be used when one piece is some- what wider than the other, so that it can be rabbeted a little and still have a miter which will match the miter on the narrower piece. If both pieces are of the same width, this can not be done. Wher- ever it is possible, however, this joint is an excellent one to use. Fig. 77 shows another way of rabbeting a mitered joint which is much better than the method shown in Fig. 76. This can be done when both pieces are of the same width or when they are of different widths. It is much stronger than the other method but requires a little more material than the simple mitered joint, as some must be cut away from one piece to form the rabbet and thus much of the timber Is wasted. Very often, however, it happens that this timber would have to be used in any case and, when this occurs, the waste need not be con- sidered. The increased strength of the joint would seem to be worth much more than the small additional amount of material which is required. Rabbeted-Mitered-and-Splined Joint. In Fig. 78 is shown a joint which is mitered and at the same time rabbeted and spHned. In order Fig. 76. Rabbeted Miter Joint 74 GRACE MEMORIAL CHAPEL, CHICAGO, ILL. Cram, Goodhue & Ferguson, Architects, Boston and ITew York. interior of Gray Pressed Brick; Trimmings of Terra-Cotta of Color and Finish Similar to Bed ford Stone. Floor of 9-in. Welsh Quarries; Chancel Floor of Mercer Tile. For Exterior, See Vol. I, Page 10; for Detail of Wood-Carving, See Vol. II, Page 10. rnasT ncsyo, plan OOUm EIZXKTION STABLE FOR MR. J. S. HANNAH, LAKE FOREST, ILL. Shepley, Rutan & Coolidge, Architects, Chicago, 111. For Location, See Vol. I, Page 74; for Exterior and Plans of House, See Vol. I, Pages 74 and 90, CARPENTRY 65 to accomplish this, one piece is cut at an angle of forty-five degrees and then rabbeted to the thickness of the other piece. The other piece is then cut with a miter to natch that left on the first piece and also cut to match the rabbet. Both pieces are ploughed to take a spline, and thus a very strong joint is formed which combines the advantages of the mitered joint and the rabbeted or splined joints. The spline in this case should be about f inch thick and about 1| inches wide and should, in all cases be of hard wood. Fig. 77. . . Rabbeted Joint Mitered and Keyed. Another way of strengthening a mitered joint is by inserting what are known as "keys" into the pieces on the outside of the joint. These keys are thin slices of hard wood which are placed in slots prepared to receive them and held in place by means of glue. As the glue fastens them securely to each of the pieces at the joint, they hold them firmly together and prevent the joint from opening. Fig. 79 shows a joint of this kind. The keys, of course, show on the outside of the joint, but they can be cut very thin and only the edge of them can be seen. The keys give a great amount of additional strength to the connection and are more effective than is a spline for preventing the joint from opening, as they come right out to the edge of both pieces and can be placed as near together as seems to be necessary. Sometimes, instead of being placed horizontally, or in a plane perpendicular to the edge of the joint, they are inclined as shown in Fig. 80. This arrangement strengthens the joint still more. Tenon Joint with Haunch. In Fig. 81 is shown a form of joint called the "tenon joint," with the addition of a "haunch" which adds considerably to its strength. This joint is used extensively in the making of doors. One of the pieces to be joined is rabbeted on each side to about one-third of its depth, leaving a projecting part called the "tenon" about one-third the thickness of the This tenon is then rabbeted on either the top or the bottom, but instead of being cut entirely back to the body of the piece, the Fig. 78 Rabbeted Mitered-and- Splined Joint piece. 75 66 CARPENTRY rabbet is stopped a little short of this and a "haunch" is left. In Fig. 81, yl is the tenon and B is the haunch. The other of the two Fig. 79. Mitered Joint Keyed Square Fig. 80. Mitered Joint Keyed Diagonally pieces which are to be joined is cut with mortises to receive the tenon and the haunch. in the first piece. Fig. 81 shows the simplest form of simple tenon joint, but there are many variations of this, two of which are shown in Figs. 82 and 83. Fig. 82 shows a single tenon joint with two tenons, while Fig. 83 shows a double tenon joint which has four ten- ons. Both of these joints have haunches as well as tenons. The one most commonly used is that shown in Fig. 82. All of the splices so far con- sidered have been end to end splices, that is, they have been those kinds which would be used in fastening pieces together at the ends. It often becomes nee- Fig. 81. Tenon Joint with Haunch CSSary, hoWCVCr, tO fastcU picCCS 76 CARPENTRY 67 together side by side. Any of the methods already described for spHces will be applicable in such a case, but there are in addition a few others which are especially useful, two or three of which will now be described. Dovetail Key. This method consists in the use of a strip of wood which is applied to the back of the several pieces to be held together and prevented from slipping by means of glue. The strip, however, is let into the pieces a little way in a special manner known as dove- tailing, which prevents it from pulling out, and gives it an especially Fig. 82. Single Tenon Joint with Two Tenons Fig. 83. Double Tenon Joint with Four Tenons strong hold on them. Fig. 84 shows this arrangement both in ele- vation and in section. It is useful in making up large panels from narrow boards. In this method, only one of the pieces must be glued to the strip, the others being left free to move. Another method of accomplishing this same result is by the use of a strip which sets against the back of the pieces to be joined, but is not let into them at all. Fig. 85. It is held in place by means of screws which go through slotted holes in the strip. This is in order that the pieces may have a chance to swell or shrink without bulg- ing or splitting. It is usually customary to employ brass slots which are let into the wood. These resist much better the wear of 77 68 CARPENTRY the screws and prevent them from working loose. If, however, the strip is of very hard wood this is not always necessary. A third method is that shown in Fig. 86. This is sometimes called the "button" method on account of the use of the small side pieces or buttons which fit over the center strip and hold the pieces of board together, at the same time allowing them to swell or shrink freely. Only the small pieces are screwed to the boards, Fig. 84. Method of Binding Boards Together by Means of Dovetail Key the center strip being fastened to one of the pieces only. This arrangement takes up a little more room than the others and looks somewhat more clumsy but is quite satisfactory otherwise. In all three of the methods described, the strip should be from 3 to 4 inches wide. Dovetailing. There is another way of joining two pieces meet- ing at right angles, and it is better and stronger than any other but, on account of the work involved in the process of making the joint, 78 CARPENTRY 69 is seldom used except in the best work. This method is known as dovetaihng and there are three different ways of arranging the dove- tails as will be shown. The first is the simple dovetail which is illus- trated in Fig. 87. As will be seen, it consists in cutting tenons in the end of one piece and mortises in the end of the other piece, which are of such a shape as to form a sort of locking device, so that the pieces Fig. 85. Method of Binding Boards Together by Means of Strip with Slots and Screws can be separated only by a pull in one particular direction. The use of glue makes the joint still stronger. Of course, the forming of a joint of this kind requires a large amount of time and considerable skill. A variation of the simple dovetail joint which is much used in the manufacture of drawers and in any other position where it is desirable that the joint shall be concealed from one side only, is shown in Fig. 88. This is called a lap dovetail, its peculiarity 79 70 CARPENTRY consisting in the fact that in one of the pieces the mortises are not cut the full thickness, but only partly through the wood, so as to leave a covering or lap, which prevents the joint from being seen. A further development of the dovetail joint is shown in Fig. 89. In this case the work is so arranged that the joint can not be seen Fig. 86. Button Method of Binding Boards Together from any side of the finished product. This is accomplished by cutting the same tenons and mortises as in the case of the simple dovetail joint, but not directly on the end of the pieces. They are so cut as to project at an angle of forty-five degrees, and thus to form a combination of the mitered joint and the dovetail joint with the tenons and mortises entirely out of sight when the pieces have been put together. This joint is obviously not so strong as 80 CARPENTRY 71 Fig. 87. Simple Dovetail Fig. 88. Lap Dovetail Fig. 89. Further Development of Dovetail Joint 81 72 CARPENTRY are the other forms of dovetail joints because the tenons are not so large. WALL Let us next consider the framing of the walls of a wood or frame building. In this work there are two distinct methods of procedure, known, respectively, as "braced framing" and ''balloon framing," of which the first is the older and the stronger method, while the second is a modern development and claims to be more logical and at the same time more econom- ical than the other. Balloon fram- ing has come into use only since about the year 1850, and it is still regarded with disfavor by many architects, especially by those in the eastern states. Figs. 90 and 91 show the framing of one end of a small building by each of the two methods, the braced framing in Fig. 90 and the balloon framing in Fig. 91. Braced Frame. In a full-brac- ed frame all the pieces should be fastened together with mortise- and-tenon joints, but this re- quirement is much modified in common practice, a so-called "combination" frame being used in which some pieces are mortised together and others are fastened by means of spikes only. A framework is constructed consisting in each wall of the two "corner posts" AA, Fig. 90, the "sill" B, and the "plate" C, together with a horizontal "girt" D at each story to support the floors, and a diagonal "brace" E at each corner, which, by keeping the corner square, prevents the frame from being distorted. Balloon Frame. In a balloon frame there are no braces or girts, and the intermediate studs FFF, Fig. 91, are carried straight up Fig. 90. Example of Braced Framing 82 CARPENTRY 73 from the sill H to the plate K, with a light horizontal piece J, called a "ribbon" or "ledger board," set into them at each floor level to support the floor joists. This frame depends mainly upon the board- ing for its stiffness, but sometimes light diagonal braces are set into the studs at each corner to prevent distortion. The methods by which all these pieces are framed together will be explained in detail under the proper headings. Sill. The sill is the first part of the frame to be set in place. It rests directly on the underpinning and extends all around the building, being jointed at the corners and spliced where neces- sary; and since it is subject to much cutting and may be called upon to span quite considerable openings (for cellar windows, etc.) in the underpinning, it must be of a good size. Usually it is made of 6 X 6-inch square timber, but in good work it should be 6 X 8 inches and nothing lighter than 6X6 inches should be used ex- cept for piazza sills. For piazza sills a 4 X 6-inch timber may be used. The material is generally spruce, although sometimes it is Norway pine or native pine (depending upon the locality) . The sill should be placed on the wall far enough back from the outside face to allow for the water table, which is a part of the outside finish, which will be described later; and 1 inch should be regarded as the minimum distance between the outside face of the sill and the outside face of the underpinning. Fig. 92. A bed of mortar A, preferably of cement mortar, should be prepared on the top of the underpinning, in which the sill C should rest; and the under side of the sill should be painted with one or two coats of linseed oil to prevent it from absorbing moisture from the masonry. In many cases, at intervals of from 8 to 10 feet, long bolts B are f/ Fig. 91. Example of Balloon Framing 83 74 CARPENTRY set into the masonry. These bolts extend up through holes bored in the sill to receive them and are fastened at the top of the sill by a washer and a nut screwed down tight. They fasten the sill, and consequently the whole frame securely to the underpinning, and should always be provided in the case of light frames in exposed positions. The beams or "joists" D, which form the framework of the first floor, are supported at one or both ends by the sill and may be fastened to it in any one of several different ways. The ideal method is to nang the joist in a patent iron hanger fastened to the sill, as shown in Fig. 93, where A is the sill, B the joist, and C the hanger. In this case neither the sill nor the joist need be weakened by cutting, but it is too expensive a method for ordinary work, although the saving in labor largely offsets the cost of the hanger. The usual method is to cut a mortise in the sill to receive a tenon cut Fig. 92. Method of Setting Underpinning, Sill, and Joists Fig. 93. Joist Hung in Patent Hanger Fig. 94. Example of Sill Mortised to Receive Joist in the end of the joist, as shown at A in Fig. 94. The mortises are cut in the inside upper corner of the sill. They are about 4 inches deep and cut 2 inches into the width of the sill and are fixed in position by the spacing of the joists. 84 K u S H g 2 o O PU m cc o u PQ r'3 CH < s H rt o Q o '-'•s tn Ol >;« t/7 ^ g ^1 s < Tl > PL, rt » o S o H Ph C8 (0 o •^ b a « S" cS £ Ph CG ^ < CS H e3 O U ^ K M [X) u C3 ^ t r/1 o Z ^ "-< ^ s o o o o a frt a> o u S! > i-l CARPENTRY 75 Mortises are also cut in the sill to receive tenons cut in the lower ends of the studs, as shown at B in Fig. 95. They are cut the full thickness of the studding, about 1| inches in the width of the sill and about 2 inches deep. The position of these mortises is fixed by the spacing of the studding, and by the condition that the outer face of the studding must be flush with the outer face of the sill in order to leave a plain surface for the board- ing. - The sills are usually halved and pinned together at the corners, as shown in Fig. 96; but sometimes they are fastened together by means of a tenon A cut in one sill, which fits into a mortise cut in the other, as shown in Fig. 97. This method may be stronger than the other, but the advantage gained is not sufficient to compensate for the extra labor involved. Sills under 20 feet in length should be made in one piece, but in some cases splicing may be necessary. In such cases a scarf joint should always be used, the splice should be made strong, and the pieces should be well fitted together. Fig. 95. Sills Mortised to Receive Studs Fig, 96. Sills Halved and Pinned at Corners Fig. 97. Sills Sometimes Joined by Tenon Joint In some parts of the country it is customary to "build up" the sill from a number of planks 2 or 3 inches thick, which are spiked securely together. A 6 X 6-inch sill can be made in this way from 87 76 CARPENTRY Fig. 98. Built up Sill three planks 2 inches thick and 6 inches wide, as shown in Fig. 98. Planks of any length may be used, and may be so arranged as to break joints with each other in order that the sill may be continuous without splicing. It is often easier and cheaper to build up a sill in this way than it is to use a large, solid timber, and if the parts are well spiked together, such a sill is fully as good as the other. When a sill of this kind is used, however, it should always be placed on the wall in such a way that the planks of which it is composed will rest on their edges, and not lie flat. Corner Posts. After the sill is in place, the first floor is usually framed and roughly covered at once, to furnish a surface on which to work, and a sheltered place in the cellar for the storage of tools and materials, after which the framing of the wall is continued. The corner posts are first set up, then the girts and the plate are framed in between them, with the braces at the corners to keep everything in place; and lastly the frame is filled in with studding. The corner posts are pieces 4X8 inches, or sometimes two pieces 4X4 inches placed close together. Corner posts must be long enough to reach from the sill to the plate. The post is really a part of only one of the two walls which meet at the corner, and in the other wall a "furring stud" of 2 X 4-inch stuff is placed close up against the post so as to Fig. 99. Plan of Corner Post Details Fig. 100. Plan of Another Form of Corner Post form a solid corner, and give a firm nailing for the lathing in both walls. This arrangement is shown in plan in Fig. 99, A is the corner post, B the furring stud, C the plastering, and D the boarding and shingling on the outside. Sometimes a 4 X 4-inch piece is used for 88 CARPENTRY 77 the corner post and a 2 X 4-inch furring stud is set close against it in each wall to form the solid corner, as shown in plan in Fig. 100; but a 4X 4-inch stick is hardly large enough for the long corner post, and the best practice is to use a 4X8-inch piece although in very light framing a 4 X 6-inch piece might be used. A tenon is cut in the foot of the corner post to fit a mortise cut in the sill, and mortises CC, Fig. 101, are cut in the post at the proper level to receive the tenons cut in the girts. Holes must also be bored to receive the pins DD which fasten these members to the post. The braces are often only nailed in place, but it is much better to cut tenons on the braces for pins, as shown at A in Fig. 102. The plate is usually fastened to the posts by means of spikes only, but it may be mortised to receive a tenon cut in the top of the post. Fig. 101. Details of Tenon Joints for Corners Fig. 102. Corner Bracing with Mortise- and-Tenon Joints In the case of a balloon frame no mortises need be cut in the posts for the girts or braces, as they are omitted in this frame; but the post must be notched instead, as shown in Fig. 103, to receive the ledger board or ribbon and the light braces which are sometimes used. Girts. The girts are always made of the same width as the posts, being flush with the face of the post both outside and inside, and the depth is usually 8 inches, although sometimes a 6-inch timber may be used. The size is, therefore, usually 4X8 inches. A tenon at each end fits into the mortise cut in the post, and the whole is secured by means of a pin DD, as shown in Fig. 101. The pin should always be of hard wood and about | inch in diameter. 80 78 CARPENTRY It is evident that if the girts in two adjoining walls were framed into the corner post at the same level, the tenons on the two girts would conflict with each other. For this reason the girts A which run parallel with the floor joists are raised above the girts B on which these joints rest, and are called "raised girts" to distinguish them from the others which are called "dropped girts." The floor joists are carried by the dropped girts, and the raised girts are so placed that they are just flush on top with the joists which are parallel to them. Ledger Board. The heavy girts are used only in the braced frame. In the balloon frame, light pieces called "ledger boards" or "ribbons" are substituted for them. These are usually made about | inch thick and 6 or 7 inches deep, and are notched into the posts and intermediate studs instead of being framed into them as in the braced frame. This notching is shown in Fig. 104, on which A is the ledger board and B the stud. The ledger boards themselves are not cut at all, but the floor joists which they carry are notched over Fig. 103. Notched Post in Balloon Framing Fig. 104. Notched Stud with Ledger Board Fig. 105. Ledger Board with Notched Floor Joist in Place them, as shown in Fig. 105, and spiked to them and to the studding. In Fig. 105, A is the joist, B the ledger board, and C the stud. Even in the braced frame a ledger board is usually employed to support the joists of the attic floor, which carry little or no weight. The disadvantage of the ledger board is that, as a tie between the corner 90 a S O) O to O 2 be CARPENTRY 79 posts, It is less effective than the girt, and consequently a wall in which it has been substituted for the girt is not as stiff as one in which the girt is used. Plate. The plate serves two purposes : First, to tie the studding together at the top and form a finish for the wall; and second, to furnish a support for the lower ends of the rafters. See Fig. 106. It is thus a connect- ing link between the wall and the roof, just as the sill and the girts are connecting links between the floors and the wall. Some- times the plate is also made to support the attic floor joists, as shown in Fig. 107, in which ^ is a rafter, B the joist spiked to the rafter, C the plate built up from 2 X 4-inch pieces, and D the wall stud. It acts in this case like a girt, but this arrangement is not very common, the attic floor joists usually being supported on a ledger board, as shown in Fig. 105. The plate is merely spiked to the corner posts and to the top of the studding; but at the corner where the plates in two adjacent walls come together, they should be Fig. 106. Section of Plate between Studs and Rafters Fig. 107. Plate as Support for Attic Floor Fig. 108. Plates Used in Balloon Framing connected by a framed joint, usually halved together in the same way as the sill. In the braced frame, a fairly heavy piece, usually a 4X6 inch is used, although a 4 X 4 inch is probably sufficiently strong. In a balloon frame the usual practice is to use two 2 X 4-inch pieces placed one on top of the other and breaking joints, as shown at A 01 80 CARPENTRY in Fig. 108, in order to form a continuous piece. The corner joint is then formed, as shown at B. No cutting is done on the plate except at the corners, the rafters and the attic floor joists being cut over it, as shown in Figs. 107 and 109. Braces. Braces are used as permanent parts of the structure only in braced frames, and serve to stiffen the wall, to keep the cor- ners square and true, and to prevent the frame from being distorted by lateral forces, such as wind. In a full-braced frame, a brace is placed wherever a sill, girt, or plate makes an angle with a corner post, as shown at E in Fig. 90. Braces are placed so as to make an angle of forty-five degrees with the post^ and should be long enough to frame into the corner post at a height of from one-third to one-half the height of the story. This construction is often modified in practice, and the braces are placed as shown at A in Fig. 109. Such a frame is not quite so stiff and strong as the regular braced frame, but it is sufficiently strong in most cases. The braces are made the same width as the posts and girts, usually 4 inches, to be flush with these pieces both outside and inside, and are made of 3 X 4-inch or 4X 4-inch stuff. They are framed into the posts and girders or sills, by means of a tenon cut in the end of the brace, and a mortise cut in the post or girt, and are secured by a hardwood pin. or I inch in diameter. The connection is shown in Fig. 102. In a balloon frame there are no permanent braces, but light strips are nailed across the corners while the framework is being erected, and before the boarding has been put on, to keep the frame in place. As soon as the outside boarding is in place these are removed. This practice is also modified, and sometimes light braces are used as permanent parts of even a balloon frame. They are not framed into the other members, however, but are merely notched into them and spiked, as shown in Fig. 110. A is the brace, i? the Fig. 109. Braced Frame The pin should be 92 CARPENTRY 81 sill, C the corner post, and DD are studs. In such a case every stud must be notched to receive the brace, which is really the same as the temporary brace mentioned above, except that it is notched into the studs instead of being merely nailed to them, and is not removed when the boarding is put on. These braces are usually made of IX 3-inch stuff. Studding. When the sill, posts, girts, plates, and braces are in place, the only step that remains to complete the rough framing of the wall is the filling in of this framework with studding. The stud- ding is of two kinds, viz, the heavy pieces which form the frames for the door and window openings and the stops for the partitions; and the lighter pieces which are merely "filling-in" studs, and are known by that name, or as "intermediate" studding. Fig. 110. Light Braces for Balloon Frame Fig. 111. Construction of Door and Window Frames The frames for the door and window openings are usually made in a braced frame, from 4X 4-inch pieces. A vertical stud A A, Fig. Ill, is placed on each side of the opening, the proper distance being left between them, and horizontal pieces BB are framed into them at a proper level to form the top and the bottom of the opening. In all good work a small truss is formed above each opening by set- ting up two pieces of studding CC over the opening, in the form of a triangle. This is to receive any weight which comes from the studding directly above the opening, and to carry it to either side of the opening where it is received by the studding and in this way 93 82 CARPENTRY carried down to the sill. Such a truss is shown in Fig. 111. The pieces used are 3X4 inches or 4X4 inches, and may be either framed into the other members or merely spiked. There should be a space D of at least 1 inch between the piece B forming the top of the window frame and the piece E forming the bottom of the truss, so that if the truss sags at all it will not affect the window frame. This is a point that is not generally recognized. The piece B is usually made to serve both as the top of the window and bottom of the truss. Fig. 112. Framing Details of Window Opening in Balloon Frame Building Fig. 112 shows the framing for the top of a window opening in a balloon-framed building, where the ledger board is partly sup- ported by the studs directly over the opening. Since the floor joists rest on the ledger board, there may be considerable weight trans- ferred to these studs; and in order to prevent the bottom of the truss from sagging under this weight, an iron rod should be inserted as shown. In the balloon frame, the door and window studs are almost 04 CARPENTRY 83 Fig. 113. Door and Win- dow Studs in Balloon Frame always made of two 2 X 4-inch pieces placed close together, and in this case the connection of the piece forming the top and bottom of the frame with those forming the sides is made as shown at A in Fig. 113. It should be noticed that in a balloon frame all studding is carried clear up from the sill to the plate, so that if there is an opening in the wall of the first story, and no corresponding openings in those of the second or third story, the door and window studding must still be carried double, clear up to the plate, and material is thus wasted. In designing for balloon frames, therefore, it is well to take care that the window openings in the second story come directly above those in the first story wherever this is possible. The same difficulty does not occur in the case of a braced frame, because in such a frame the studding in each story is independent of that in the story above or below it; the window openings may, therefore, be arranged independently in the different stories according to the requirements of the design. Nailing Surfaces. Whenever a partition meets an outside wall, a stud wide enough to extend beyond the partition on both sides and to afford a solid nailing for the lathing must be inserted. A nailing surface must be provided for the lathing on both the outside wall and the partition, and the first stud in the partition wall is, therefore, set close up against the wall stud, forming a solid corner. This arrangement is shown in plan in Fig. 114. The large wall stud A is usually made of a 4X 8-inch piece set flatwise in the wall, so that if the partition is, say 4 inches wide, there is a clear nailing surface of 2 inches on each side of the partition. A 4 X 6-inch piece could also be used here, leaving a clear nailing surface of 1 inch on each side of the partition. Sometimes the same thing is accomplished by using two 3 Fig. 114. Plan of Studding in Outer Wall Opposite Partition OS 84 CARPENTRY 4X 4-inch pieces placed close together, as shown in plan in Fig, 115, instead of one 4X 8-inch piece. Sometimes two pieces, 2X4 inches or 3X4 inches, are used, placed far enough apart so that they afford a space for nailing on each side of the partition, as shown in plan in Fig. 115. Two-Piece Stud in Outer Wall Opposite Partition Fig. 116. Placing of 2 X4 Studs to Give Nailing Surface Fig. 116. Whenever this is done, small blocks A, Fig. 117, should be set in between the two studs at intervals of 2 to 3 feet throughout their entire height in order to give them added stiffness and make them act together. The end in view in every case is to obtain a solid corner on each side of the partition where it joins the wall, and any construction which accomplishes this is good. In the best work, however, the 4 X 8-inch solid piece is used, and this construction can always be depended upon. It makes no difference what the angle between the wall and the partition may be, but usually this angle is a right angle. Intermediate Studding. The pieces which make up the largest part of the wall frame are the "filKng-in" or "intermediate" studs. These, as the name im- plies, are used merely to fill up the frame made by the other heavier pieces, and afford a nail- ing surface for the boarding, which covers the frame on the outside, and the lathing, which covers it on the inside. The fill- ing-in studs are usually placed 16 inches apart, measured from the center of one stud to the center of the next. In especially good Fig. 117. Placing of Block Stiffeners in Con- struction Shown in Fig. 116 96 CARPENTRY 85 work they are sometimes placed only 12 inches apart on centers, but this is unusual. In no case should they be placed more than 16 inches apart, even in the lightest work. The studs are made the full width of the wall, usually 4 inches, but sometimes in large buildings (such as churches) 5 or even 6 inches and their thickness is almost always 2 inches, 2X4 inches being the more usual dimen- sions. The lengths of the intermediate studs are made to fit the rest of the frame. In the braced frame, there must necessarily be a great deal of cutting of the intermediate studding, because all the large pieces are made the full width of the wall and the intermediate stud- ding must be cut to fit between them. In the balloon frame, however, the intermediate studding in all cases extends clear up from the sill to the plate, and no cut- ting is necessary except the notching to receive the other parts of the frame. See Fig. 91. In a balloon frame it often happens that the studs are not long enough to reach from the sill to the plate and they must be pieced out with short pieces which are spliced onto the long stud. This splicing is called "fish- ing," and it is accomplished by nailing a short thin strip of wood A A on each side of the stud, as shown in Fig. 54, in order to join the two pieces firmly to- gether. The strips should be well nailed to each piece in order to give the required strength. All door and window studs should have a tenon cut at the foot of the piece to fit a mortise cut in the sill. Inter- mediate studs are merely spiked to the sill without being framed into it. The tenons are cut in two different ways, as shown in Figs. 118 and 119. They are always made the full thickness of the piece, and by the first method they are placed in the middle of the piece, as shown. The width of the tenon is about 1| inches, leaving 1^ inches on the outside and 1 inch on the inside of the stud. Another way is to make the tenon on the inside of the stud, as shown in Fig. 119, the tenon being 1| inches wide as before. There is • no choice between these methods, both being good. 97 86 CARPENTRY PARTITIONS The studding used in partition walls is usually of 2 X 4-inch stuff, although 2 X 3-inch studding may sometimes be used to advantage if the partition does not support any floor joists. Furring Walls. The partition walls are made 4 inches wide, the same as in the outer walls, except in the case of so-called "furring" Fig. 120. Plan Showing Plaster Applied Directlj- to Chimney Brickwork partitions. These are built around chimney breasts and serve to conceal the brickwork and furnish a surface for plastering. They are formed by placing the studding flatwise, in order to make a thin wall; and as it is usually specified that no woodwork shall come within 1 inch of any chimney, a 1-inch space is left between the brickwork and the furring wall. It is possible to apply the plaster directly to the brickwork, and this is sometimes done, but there is Fig. 121. Plan Showing Construction of Furring Wall always danger that cracks will appear in the plastering at the corner A, Fig. 120, between the chimney breasts and the outside wall. This cracking is due to the unequal settlement of the brickwork and the woodwork since the plastering goes with the wall to which it is 98 CARPENTRY 87 applied. The method of constructing a furring wall is shown in plan in Fig. 121. A A are the furring studs, B is the plastering, and CC the studding in the outside wall. The arrangement without the furring wall is shown in plan in Fig. 120. If there are any open- ings in the furring wall, such as fireplaces, or "thimbles" for stove pipes, it is necessary to frame around them in the same v/ay as was explained for door and window openings in the outside walls. See Fig. 122. AA are furring studs, BB are pieces forming the top and bottom of the opening. If the outside walls of the building are of brick or stone, a wood "furring" Fig. 122. Furring Strip Frames Around Openings Fig. 123. Furring strips on Outside Wall wall is usually built just inside of the outer wall; this furnishes a surface for plastering and for nailing the inside finish. The stud- ding for these walls is 2 X 4 inches or 2 X 3 inches or 1X2 inches set close up against the masonry wall and preferably spiked to it. See Fig. 123. Spikes are usually driven directly into the mortar between the bricks or stones of the wall, but sometimes wood blocks or wedges are inserted in the wall to afford a nailing surface. Wherever a wood partition wall meets a masonry exterior wall at an angle, the last stud of the partition wall should be securely spiked in the masonry wall, to prevent cracks in the plastering. 90 88 CARPENTRY ■- Cap and Sole. All partition walls are finished at the top and bottom by horizontal pieces, called, respectively, the "cap" and the "sole." The sole rests directly on the rough flooring whenever there is no partition under the one which is being built; but if there is a partition in the story below, the cap of the lower partition is used as the sole for the one above. The sole is made wider than the stud- ing forming the partition wall, so that it projects somewhat on each side and gives a nailing surface for the plasterer's grounds and for the inside finish. It is usually made about 2 inches thick and 5| inches wide, when the partition is composed of 4-inch studding, and this leaves a nailing surface of | of an inch on each side. The sole is shown at B in Fig. 124. The cap is usually made the same width as the studding, and 2 inches thick, so that a 2 X 4-inch piece may be used in most cases ; but if the partition is called upon to support the Fig. 124. Partition Framing Showing Sole Fig. 125. Partition Framing Showing Cap floor beams of the floor above, the cap may have to be made 3 or even 4 inches thick, and some architects favor the use of hard wood such as Georgia pine for the partition caps. The cap is shown at A, Fig. 125. Bridging. In order to stiffen the partitions, short pieces of studding are cut in between the regular studding in such a way as to connect each piece with the pieces on each side of it. Thus, if one piece of studding is for any reason excessively loaded, it will not have to carry the whole load alone but will be assisted by the other pieces. This operation is called "bridging," and there are two kinds, which 100 CARPENTRY 89 may be called "horizontal bridging" and "diagonal bridging." The horizontal bridging consists of pieces set in horizontally between the vertical studding to form a continuous horizontal line across the wall, every other piece, however, being a little above or below the next piece as shown in Fig. 126. The pieces are 2 inches thick and the full width of the studding; and in addition to strengthening the wall, they prevent fire or vermin from passing through, and also may be utilized as a nailing surface for any inside finish such as wainscoting or chair rails. The second method, which we have called diagonal bridging, is more effective in preventing the partition from sagging than is the Fig. 126. Horizontal Bridging Fig. 127. Diagonal Bridging straight bridging, but both methods may be used with equal pro- priety. In the diagonal bridging the short pieces are set in diagonally, as is shown in Fig. 127, instead of horizontally, between the vertical studding. This method is certainly more scientific than the other, since a continuous truss is formed across the wall. All partitions should be bridged by one of these methods, at least once in the height of each story, and the bridging pieces should be securely nailed to the vertical studding at both ends. It is cus- tomary to specify two tenpenny nails in each end of each piece. Bridging should be placed in the exterior walls as well as in the 101 90 CARPENTRY partition walls; and as a further precaution against fire, it is good practice to lay three or four courses of brickwork, in mortar, on the top of the bridging in all walls, to prevent the fire from gaining head- way in the wall before burning through and being discovered. This construction is shown in Fig. 128. Special Partitions. A partition in which there is a sliding door must be made double to provide a space into which the door may slide when it is open. This is done by building two walls far enough apart to allow the door to slide in between them, the studding being of 2 X 4-inch or 2 X 3-inch stuff, and placed either flatwise or edgewise in the wall. If the studding is placed flat- wise in the wall a thinner wall is possible, but the construction is not so good as in the case where the studs are placed edge- wise. If the partition is to sup- port a floor, one wall must be made at least 4 inches thick to support it, and the studs in the other wall may then be placed flatwise if desired, and the floor supported entirely on the thick wall. The general arrangement is shown in plan in Fig. 129. It is better to use 2 X 4-inch studding set edgewise in each wall so as to make two 3-inch walls with space enough between to allow the door to slide freely after the Fig. 128. Horizontal Bridging with Bricks Laid on Top pocket has been lined with sheathing. A piece of studding A, Fig. 130, should be cut in hori- zontally between each pair of studs B, 8 or 10 inches above the top of the door in order to keep the pocket true and square. The pocket should be lined on the inside with matched sheathing C. It is well known that ordinary partitions are very good con- ductors of sound; and in certain cases, as in tenement houses, this is Fig. 129. Section of Partition Construction for Sliding Doors 102 CARPENTRY 91 objectionable, so that special construction is required. If two walls are built entirely separate from each other and not touching at any- place, the transmission of sound is much retarded; and if heavy felt paper or other material is put in between the walls, the partition is made still more nearly soundproof. In order to decrease the thickness of such a wall as much as possible, the "staggered" partition is used, in which there are two sets of studding, one for each side of the wall, but ar- ranged alternately instead of in pairs as in the double partition. The arrange- ment is shown in plan in Fig. 131. The two walls are entirely separate from each other and the felt paper may be worked in between the studs as shown, or the whole space may be packed full of some soundproof and fireproof material such as mineral wool. There is a so-called "quilting paper" or "sheathing-quilt" manufactured from sea weed, which is much used for this pur- pose. The inside edges of the two sets of studs are usually placed on a line, making the whole wall 8 inches thick, where 4-inch studding is used, and the studs may be placed about 16 inches on cen- ters in each wall. Each set of studding should be bridged sep- arately. Another case where a double wall may be necessary, is where pipes from heaters or from plumbing fixtures are to be carried in the y..a.;M-- M^.:.. ' ...Lu,. . ....u.,u . ..K-,.,.,.... . ^ .. . ... ... .. . . . ... . ..... .. ... ... .. .. . , wall. In case of hot pipes, care must be taken to have the space large enough so that the woodwork will not come dangerously near the pipes. An important matter in connection with the framing of the partitions is the way in which they are supported; but this involves knowledge of the framing of a floor and, therefore, it will be left for the present. It will be taken up after we have considered the floor framing. Fig. 130. Details of Bracing in Double Partition Fig. 131. Partition with Stagger Studding to Prevent Sound Transmission 103 92 CARPENTRY SHRINKAGE AND SETTLEMENT An important point which must be considered in connection with the framing of the walls and partitions, is the settlement due to the shrinkage of timber as it seasons after being put in place. Tim- ber always shrinks considerably ACROSS the grain, but very little in the direction of the grain; so it is the horizontal members, such as the sills, girts, and joists, which cause trouble, and not the vertical members, such as posts and studding. Every inch of horizontal timber between the foundation wall or interior pier and the plate is sure to contract a certain amount, and as the walls and partitions are supported on these horizontal members, they, too, must settle somewhat. If the exterior and interior walls settle by exactly the same amount, no harm will be done, since the floors and ceilings will remain level and true, as at first; but if they settle unequally, all the levels in the building will be disturbed, and the result will be crack- ing of the plastering, binding of doors and windows, and a general distortion of the whole frame, a condition which must be avoided if possible. It is evident that one way to prevent unequal settlement, so far at least as it is due to the shrinkage of the timber, is to make the amount of horizontal timber in the exterior and interior walls equal. Thus, starting at the bottom, we have from the masonry of the founda- tion wall to the top of the first floor joists in the outside walls 10 inches, or the depth of the joists themselves, since these rest directly on the top of the wall. In the interior, we have, if the joints are framed flush into a girder of equal depth, the same amount, so that here the settlement will be equal. But the studding in the exterior wall rests, not on the top of the joists, but on the top of the 6-inch sill, while the interior studding rests on top of the 10-inch girder. Here is an inequality of 4 inches which must be equalized before the second floor level is reached. If the outer ends of the second- floor joists rest on the top of an 8-inch girt, and the inner ends on a 4-inch partition cap, this equalizes the horizontal timber inside and outside, and the second floor is safe against settlement. The same process of equalization of the horizontal timber may be continued for each floor up to the top of the building, and if this is carefully done it will go far toward preventing the evils resulting from settle- ment and shrinkage. 104 CARPENTRY 93 With a balloon frame this can not be done, because there are no girts in the outside wall, but only ledger boards which are so fas- tened that they can not shrink, while in the interior walls we have still the partition caps. All that can be done in this case, is to make the depth of the sills and interior girders as nearly equal as possible, and to make the partition caps as shallow as will be consistent with safety. FLOORS After the wall, the next important part of the house frame to be considered is the floors, which are usually framed while the wall is being put up and before it is finished. They must be made not only strong enough to carry any load which may come upon them, but also stiff enough so that they will not vibrate when a person walks across the floor, as is the case in some cheaply-built houses. The floors are formed of girders and beams, or "joists," the girders being large, heavy timbers which support the lighter joists when it is impos- sible to allow these to span the whole distance between the outside walls. Girders. Girders are generally needed only in the first floor, since in all the other floors the inner ends of the joists may be sup- ported by the partitions of the floor below. They are usually of wood, though it may sometimes be found economical to use steel beams in large buildings, and even in small buildings the use of steel for this purpose is increasing rapidly. Wrought iron was once used, but steel is now cheaper and has taken its place. However, when this is not found to be expedient, hard pine or spruce girders will answer very well. The connections used in the case of steel girders will be explained later. The girders may be of spruce or even of hemlock, but it is hard to get the hemlock in such large sizes as would be required, and spruce, too, is hardly strong enough for the purpose. Southern pine, therefore, is usually employed for girders in the best work. The size of tne girder depends on the span, that is, the distance between the supporting walls, and upon the loads which the floor is expected to carry. In general, the size of a beam or girder varies directly as the square of the length of the span, so that if we have two spans, one of which is twice as great as the other, the girder for 105 94 CARPENTRY the longer span should be four times as strong as the girder for the smaller span. In ordinary houses, however, all the girders are made about 8X12 inches in sections, although sometirhes an 8 X 8-inch timber would suf- fice, and sometimes perhaps a 12-inch piece would be required. It should be remembered in deciding upon the size of this piece, that any girder is increased in strength in direct propor- tion to the width of the timber (that is, a girder 12 inches wide is twice as strong as one 6 inches wide), but in direct proportion also to the square of the depth (that is, a girder 12 inches deep is four times as strong as one 6 inches deep) . Hence, the most economical girder is one which is deeper than it is wide, such as an 8X 12-inch stick; and the width may be decreased by any amount so long as it is wide enough to pro- vide sufficient stiffness, and the depth is sufficient to enable the piece to carry the load placed upon it. If the piece is made too narrow in Fig. 132. Buckled Girder Show- ing Too Light Construction V Fig. 133. Double Stirrup Joist Supporter Fig. 134. Single Joist Hanger proportion to its depth, however, it is likely to fail by "buckling," that is, it will bend as shown in Fig. 132. It has been found by experience that for safety the width should be at least equal to i of the depth. There are at least three ways in which the joists may be sup- ported by a girder. The best, but most expensive, method is to 106 FIRST AND SECOND STORY PLANS OF HOUSE FOR MR, C. M. THOMPSON, CAMBRIDGE, MASS. Cram, Goodhue & Ferguson, Architects, Boston and New York. CARPENTRY 95 support the ends of the joists in patent hangers or stirrup irons which connect with the girder. This method is the same as was described for the sill, except that with the girder a double stirrup iron, such as that shown in Fig. 133, may be used. These stirrup iron hangers are made of wrought iron, 2| or 3 inches wide, and about f inch thick, bent into the required shape. They usually fail by the crushing of the wood of the gir- ders, especially when a single hanger, like that shown in Fig. 134, is used. Fig. 135 shows a double stirrup iron hanger in use. Patent hangers as shown in Fig. 136 are the most suit- able. If hangers of any kind are used, there will be no cutting of the girder except at the ends, where it frames into the sill, and even there a hanger may be used. The girder may be placed so that the joists will be flush with it on top, or so that it is flush with the sill on top. Fig. 135. Double Hanger with Joists in Place Fig. 136. Patent Hanger with Joists in Place Fig. 137. Joists Framed into Sill If the joists are flush with the girder on top, and are framed into the sill in the ordinary way, as shown in Fig. 137, the girder can not be flush on top with the sill; while, on the other hand, if the girder is flush with the sill on top, it can not at the same time be flush with the joists on top. If joist hangers are used on the girder to support 107 96 CARPENTRY the joists, they will probably be used on the sill as well, as explained in connection with the sill; and in this case the girder can be made flush with the sill on top and the joists hung from both girder and sill with hangers, thus bringing both ends of a joist to the same level. Fig. 138. Method of Bringing Joist Level When Resting on Sill as shown in Fig. 138. If the girder were framed into the sill at all, it would almost always be made flush with the sill on top, and by the proper adjustment of the hangers the joists would be arranged so as to be level. For framing the girder into the sill, a tenon-and-tusk joint, as shown in Fig. 139, would be used if the girder is to be flush with the sill on top. Since the girder would in most cases be deeper than the sill, the latter having a depth of only 6 inches the wall would neces- sarily have to be cut away in order to make a place for the girder. This condition is clearly shown in Fig. 140. The girder itself should not be cut over the wall, as shown in Fig. 141, because this greatly weakens the girder. If this method is used, the joists should be framed into the girder in the same way as they are framed into the sill, a mortise being cut in the girder, and a tenon on the joist. This is called "gaining" and is shown in Fig. 137. The top of the girder thus comes several inches below the top of the floor. Another method is to make the top of the girder flush with the top of the joists. The joists are then framed into the girder with a Fig. 139. Tenon-and-Tusk Joint between Sill and Girder 108 CARPENTRY 97 Fig. 140. Wall Cut Away to Allow Girder and Sill to Join at Same Level tenon-and-tusk joint, as shown in Fig. 139, and the girder is "gained" into the sill, as shown in Fig. 137. Still another method in common use is simply to "size down" the joists on the girder about 1 inch, as shown in Fig. 142. In this case, of course, the girder is much lower than the sill, usually so low that it can not be framed into the sill at all, but must be supported by the walls independently. Holes are left in the wall where the girders come, the latter being run into the holes, and their ends resting directly on the wall, inde- pendent of the sill. This is not very good construction, however, because the floor is not tied together as it is when the girder frames into the sill. The first method is the best and is the one in most common use. The girders serve to sup- port the partitions as well as to support the floors, and should, therefore, be designed to come under the partitions whenever this is possible. When the distance between the outside walls is too great to be spanned by the girder, it is supported on brick piers or posts of hard wood or cast iron in the cellar. Such piers or posts should always be placed wherever girders running in different directions intersect. Girders are also often supported on brick cellar partitions. Fig. 1-11. Cutting Girder Bad Practice on Account of Weakening Effect 109 98 CARPENTRY Joists. Joists are the light pieces which make up the body of the floor frame and to which the flooring is nailed. They are almost always made of spruce, although other woods may be used, and may be found more economical in some localities. They are usually 2 or 3 inches thick, but the depth is varied to suit the conditions. Joists as small as 2X6 inches are sometimes used in very light buildings, but these are too small for any floor. They may sometimes be used for a ceiling where there are no rooms above, and, therefore, no weight on the floor. A very common size for joists is 2X8 inches, and these are probably large enough for any ordinary construction, but joists 2X10 inches make a stiff er floor, and are used in all the best work. Occasionally joists as large as 2X12 inches are used, especially in large city houses, and they make a very stiff floor, but this size is un- usual. If a joist deeper than 12 inches is used, the thickness should be increased to 2| or 3 inches, in order to prevent it from failing by buckling, as ex- Fig. 142. "Sizing Up" Joist on Girders plained for girdcrs, P. 95. The size of the joists depends in general upon the span and the spacing. The usual spacing is 16 or 20 inches between centers, and 16 inches makes a better spacing than 20 inches, because the joists can then be placed close against the studding in the outside walls and spiked to this studding. It is generally better to use light joists spaced 16 inches on centers than to use heavier ones spaced 20 inches on centers. The spacing is seldom less than 16 inches and should never be more than 20 inches. Supports and Partitions. In certain parts of the floor frame it may be necessary to double the joists or place two of them close together in order to support some very heavy concentrated load. This is the case whenever a partition runs parallel with the floor joists, unless there is another partition under it. Such partitions may be supported in several different ways. A very heavy joist, or two joists spiked together, may be placed under the partition, as shown at A in Fig. 143, C being the sole, B the under or rough 110 CARPENTRY 99 Fig. 143. Objectionable Construction for Parti- tion Support flooring, and DDB the studding. This method is objectionable for two reasons : It is often found convenient to run pipes up in the partition, and if the single joist is placed directly under the partition, this can not be done except by cutting the joist and thus weakening it. Moreover, if the single joist is used, there is no solid nailing for the finished upper flooring, unless the joist is large enough to project beyond the partition stud- ding on each side. The joist is seldom, if ever, large enough for this, and the finished flooring must, therefore, be nailed only to the under flooring at the end where it butts against the partition, so that a weak, insecure piece of work is the result. This may be seen by referring to the figure. A much better way is to use two joists far enough apart to project a little on each side of the partition, as shown at A A in Fig. 144, and thus afford a nail- ing for the finished flooring. These joists must be large enough to support the weight of the partition without sagging any more than do the other joists of the floor, and, therefore, joists 3 or even 4 inches thick should be used. They should be placed about 6 or 7 inches apart on cen- ters, and plank bridging should be cut in between them at inter- vals of from 14 to 20 inches, as shown at E in Fig. 144, in order to stiffen them and make them act together. This plank bridging should be made of pieces of joist 2 inches thick and of the same depth as the floor joists, and should be so placed that the grain will in every case be horizontal. Fig. 144. Approved Construction for Par- tition Support 111 100 CARPENTRY A partition, supported as described above, is bound to settle somewhat as the 10 or more inches of joist beneath it shrinks in seasoning, and the settlement may cause cracks in the plaster- ing at the corner between the partition and an outside wall. In order to prevent this settle- ment, partitions running par- allel with the floor joists are often supported on strips which are secured to the under side of the floor joists, as shown at A in Fig. 145. These strips can not be allowed to project into the room below, and so they must be made as thin as is consistent with safety. Strips of iron plate about I inch thick and wide enough to support the partition studs are, there- fore, used for this purpose, and are fastened to the joists by means of Fig. 145. Partition Supported by Strips Secured to Under Side of Joist Fig, 146. Header and Trimmer Construction bolts or lag screws. Partitions which run at right angles to the floor joists can also be supported in this way. If a partition runs at right angles to the joists near the center of their span, the tendency 118 CARPENTRY 101 for the joists to sag under it will be very great, and they must be strengthened either by using larger joists, or by placing them closer together. If the span of the floor joists is large and the partition is a heavy one, it may be necessary to put in a girder running at right angles to the joists to carry the partition. In this case the partition stud will set directly on the girder, which may be a large timber, or in some cases, a steel I-beam. Headers and Trimmers. Another case where a girder may be necessary in a floor above the first, is where an opening is to be left in the floor for a chimney or for a stair well. The timbers on each side of such an opening are called "trimmers," and must be made heavier than the ordinary joists; while a piece called a "header" must be framed in between them to receive the ends of the joists, as shown in Fig. 146. The trimmers may be made by simply doubling up the floor joists on each side of the opening, or, if necessary, I-beams or heavy wood girders may be used. In most cases these trimmers may be built up by spiking together two or three joists, and the header may be made in the same way. Joist Connection. With Sill. Joists are also "gained" into the sill, as shown in Fig. 94, in which case a mortise is cut in the sill and a corresponding tenon is cut in the end of the joist. The mortise was illustrated and described in connection with the sill, while the end of the joist is cut as shown in Fig. 94, the tenon being about 4 inches deep and gained into the sill about 2 inches. This brings the bottom of the joist flush with the bottom of the sill, and the top of the joist somewhat above the top of the sill, according to the depth of the joist. The top of a 10-inch joist would come 4 inches above the top of a 6-inch sill, and the joist would rest partly on the masonry wall, thus relieving the connection of a part of the stress due to the weight of the loaded joist. A common but very bad method of framing the joist to the sill is simply to "cut it over" the sill without mortising the latter, as shown in Fig. 147. This does not make a strong con- Fig. 147. Bad Method of Framing Joist to Sill lis 102 CARPENTRY Fig. 148. Crack in Joist Due to Bad Construction nection even when the joist rests partly on the masonry wall; and if it is not so supported it is almost sure to fail by splitting, as shown in Fig. 148, under a very small loading. In fact, it would be much stronger if the joists were turned upside down. Frequently the joist is cut as shown in Fig. 149, where the tenon is sunk into a mortise cut in the sill, thus bringing the top of the joists flush with the top of the sill; but in this case the bottom of the joists will almost invari- bly drop below the bottom of the sill and the wall must be cut away to make room for it, as shown in Fig. 140. It is also weak in the same way as is the connection shown in Fig. 148. With Girders. The framing of the joists into the girders may be accomplished in several ways, according to the position of the girder. The placing of the girder is quite an important point. The top of the floor, on which rest the sole-pieces of the cross-partitions, must remain always true and level, that is, the outside ends of the joists must be at the same level as the inside ends. Otherwise the doors in the cross-partitions will not fit their frames, and can not be opened or shut and the plastering is almost sure to crack. Both ends of the joists will sink somewhat, on account of the shrinkage of the timber in seasoning, and the only way to make sure that the shrinkage at the two ends will be the same is to see that there is the same amount of horizontal timber at each end between the top of the floor and the solid masonry. This is be- cause timber shrinks very much across the grain, but almost not at all along the grain. If the joists are framed properly into the sill, so that they are flush on the bottom with the sill, we have at the outer end of the joist a depth of horizontal timber equal to the depth of the joist itself, as shown in Fig. 138; and in order to have the same depth of timber at the inside, the bottom of the joist must be flush with the bottom of the girder, which usually rests on brick piers. Of Fig. 149. Another Bad Joist and Sill Construction 114 CARPENTRY 103 course the top of the girder must not in any case come above the top of the floor joists; therefore, in general, the girder must be equal in depth to the floor joists and flush with these joists on top and Fig. 150. Superior Floor Joist and Sill Construction bottom, as shown in Fig. 150. This method is not always followed, however, in spite of its evident superiority; and the girder is often sunk several inches below the tops of the floor joists, as shown in Fig. 151. Joist Sized Down on Girder Fig. 138, or even in some cases very much below, as shown in Fig. 151. Both of these methods cause an unsightly projection below the ceiling of the cellar. Where the joists are brought flush with Fig. 152. Fastening Joists by Iron Strap Fig. 153. Fastening Joists by Dog the girder top and bottom, they may be framed into it with a tenon- and-tusk joint, as are the girders, as shown in Fig. 139, and a hole bored through the tenon to receive a pin to hold the joist in place. 115 104 CARPENTRY Other methods of frammg tenon-and-tusk joints are shown in Figs. 47, 48, 49, and also a double-tenon joint in Fig. 50, which might be used in this case, although it is much inferior to the tenon-and- tusk joint. Two joists framing into a girder from opposite sides should be fastened strongly together on top either by an iron strap passing over the top of the girder and secured to each joist, as shown in Fig. 152, or by means of a"dog" of round bar iron, which is bent at the ends and sharpened so that it may be driven down into the abutting ends of the joists, as shown in Fig. 153. These bars should be used at every fifth or sixth joist, to form a series of continuous lines across the building from sill to sill. If the girder is sunk a little below the tops of the joists these may be gained into it in the same way as they are gained into the sill. Fig. 154. Joist Gained into Girder Fig. 155. Joist Fastened to Girder by Tenon Joint and Dowel In this case joists should be arranged as shown in Fig. 154, so that they will not conflict with one another; and the two adjacent joists may be spiked together, thus giving additional stiffness to the floor. If the tenon-and-tusk connection is used, the joists may be arranged exactly opposite each other, provided that the girder is sufficiently wide, but it is always much better to arrange them as shown in Fig. 155, even in this case. The tenon may then be carried clear through the girder and fastened by a dowel as shown. Very rarely a simple double-tenon joint, such as that shown in Fig. 50, might be used, but it is much inferior to either the gaining or the tenon-and-tusk joint. If the girder is sunk very much below the tops of the joists, as in Fig. 151, these will usually rest on top of it and be fastened by 116 CARPENTRY 105 spikes only, or will be "sized down" upon it about 1 inch, as shown. There is no mortising of the girder in either case. Joists are also thus sized down upon the girts and partition caps, and are notched over the ledger boards as shown in Fig. 105. In cutting the joists Fig. 156. Joist Supported by Brick Wall Fig. 157. Anchored Joist for sizing and notching, the measurements should be taken in every case from the top of the joists, since they may not be all of exactly the same depth, and the tops must be all on a level after they are in place. This is really the only reason why the joists should be sized down at all, because otherwise they might simply rest upon the top of the girder, or girt, and be fastened by nailing. With Brick Wall. When a joist or girder is supported at either end on a brick wall, there will either be a hole left in the wall to receive it, or the wall will be corbeled out to form a seat for the beam. If the beam enters the wall the end should be cut as shown in Fig. 156, so that in case of the failure of the beam from overload- ing or from fire, it may fall out without injuring the wall. Every fifth or sixth joist is held in place by an anchor, as shown in Fig. 157, of which there are several kinds on the market. Fig. 158 shows the result when a beam which is cut off square on the end, falls out of the wall. Fig. 158. Effect of Releasing Diagonal- and Square-Ended Joists 117 106 . CARPENTRY There must always be left around the end of a beam which is in the wall, a sufficient space to allow for proper ventilation to prevent dry rot, and the end should always be well painted to keep out the moisture. Patent wall-hangers and box anchors are often used to support the ends of joists in brick buildings, but only in case of heavy floors. The floor framing in a brick building is the same as that in a building of wood except that there is no girt to receive the ends of the floor boards, so that a joist must be placed close against the inside of the wall all around the building to give a firm nailing for the flooring. Crowning. In any floor, whether in a wood or brick building, if the span of the floor joists is very considerable so that there is any chance for deflection they must be "crowned" in order to offset the effect of such deflection. The operation called "crowning" con- sists in shaping the top of each joist to a slight curve, as shown in Fig. 159 A, so that it is 1 inch or so higher in the middle than it is at Fig. 159. Crowning Joist the ends. As the joist sags or deflects, the top becomes level while the convexity will show itself in the bottom, as shown in Fig. 159B. Joists need not be crowned unless the span is quite large and the loads heavy enough to cause a deflection of an inch or more at the center of the joist. Bridging. Floor frames are "bridged" in much the same way as was described for the walls, and for much the same purpose, namely, to stiffen the floor frame, to prevent unequal deflection of the joists and to enable an overloaded joist to get some assistance from the pieces on either side of it. Bridging is of two kinds, "plank bridging" and "cross bridging," of which the first has already been shown in connection with the partition supports. Plank bridging is not very effective for stiffening the floor, and cross bridging is always preferred. This bridging is somewhat like the diagonal bridging used in the walls, and consists of pieces of scantling, usually 1X3 inches or 2X3 inches in size, cut in diagonally between the 118 CARPENTRY 107 Fig. 160. Cross Bridging between Joists floor joists. Each piece is nailed to the top of one joist and to the bottom of the next; and two pieces which cross each other are set close together between the same two joists, forming a sort of St. Andrew's cross, whence we get the name "cross bridg- ing" or "herringbone bridging" as it is some- times called. The arrange- ment is shown in Fig. 160, and the bridging should be placed in straight lines at intervals of 8 or 10 feet across the whole length of the floor. Each piece should be well nailed with two eightpenny or tenpenny nails in each end. If this is well done there will be formed a continuous truss across the whole length of the floor which will prevent any overloaded joist from sagging below the others, and which will greatly stiffen the whole floor so as to prevent any vibration. The bridging, however, adds nothing to the strength of the floor. Porch Floors. A word might be appropriately inserted at this point in regard to floors of piazzas and porches. These may be supported either on brick piers or on wood posts, but prefer- ably on piers, as these are much more durable than posts. If piers are used, a sill, usually 4X6 inches in size, should be laid on the piers all around, and light girders should be inserted between the piers and the wall of the house, in order to divide the floor area into two or three panels. The joists may then be framed parallel to the walls of the house, and the floor boards laid at right angles to these walls. The whole frame should be so constructed that it will pitch outward, away from the house at the rate of 1 inch in 5 or 6 feet, thus bringing \ .^r-'*-«rj3y /? '-J^i.^^r ^ \ // \v- l//\\ V w c > I \ ■ 1 Fig. 161. Construction of Unsupported Corner 119 108 CARPENTRY the outside edge lower than the inside edge and giving an opportunity for the water to drain off. Stairs. The stairs are built on frames called "stringers" or "carriages," which may be considered as a part of the floor framing. They consist of pieces of plank 2 to 3 inches thick and 12 or more inches wide, which are cut to form the steps of the stairs and which are then set up in place. There are usually three of these stringers under each flight of stairs, one at each side and a third in the center, and they are fastened at the bottom to the floor and at the top to the joists which form the stair well. This subject is taken up more fully under "Stair Building." Unsupported Corners. An interesting place in a floor framing plan is where we have a corner without any support beneath it, as at the corner A in Fig. 161. This corner must be supported from the three points B, C, and D, and the figure shows how this is accom- plished. A piece of timber E is placed across from B to C, and another piece starts from D and rests on the piece B C, projecting beyond it to the corner A. This furnishes a sufficiently strong sup- port for the corner. 120 CARPENTRY PART III THE ROOF The framing of the roof is one of the most difficult problems with which the carpenter has to deal, not because of the number of complicated details, for these are few, but because of the many different bevels which must be cut in order to allow the rafters to frame into one another properly. There are many kinds of roofs, and before describing the different varieties it will be well to consider briefly the purpose of the roof and its development from simple forms to those which are more elaborate and perhaps more ornamental. The primary object of a roof in a temperate climate is, of course, to keep out the rain from the interior of the building and at the same time to keep out the cold. The roof must, therefore, be so constructed as to free itself from the falling water as soon as possible, that is, it must be built to shed water and, therefore, it must be sloped or inclined to the horizontal in some way. If it is necessary for the sake of economy or for any other reason to construct the roof practically flat, it must be made more secure against the passage of water than if it is made sloping, and some provision must be made to carry off the water, and to cause it to collect in one or two low places in the roof surface, from which pipes or down spouts may be provided to take it away to a safe place outside the building. The roof covering must be of some material through which water will not readily penetrate, such as tin, galvanized iron, lead, or zinc or copper among the metals, or a composition of tar and gravel, if metal is not suitable to the purpose. This would be for a roof which is nearly flat; for a roof which slopes, and will shed the water, thus getting rid of it more quickly, a covering of slates, tile, or wood shingles Copyright, 1912, by American School of Correspondence. 121 no CARPENTRY may be employed, as well as tin, copper, or other metals. The slates or shingles would have to be laid in several thicknesses, the number depending upon the steepness of the slope of the roof, and in order to accomplish this they would have to be laid overlapping each other. In order that the water escaping from the roofs may not run down the sides of the building, making unsightly streaks and wetting the windows and doors, it is necessary that the roofs should project somewhat beyond the face of the walls all around. The projecting part of the roof is called the eaves. ROOF CHARACTERISTICS Styles of Roofs. The different varieties of roofs, from the simple pitch roof to the most complicated combination of hips and valleys, are developments of a few simple forms. Lean-To Roof. The lean-to roof is the most simple of them all, and is usually employed for small sheds, piazzas, porches, ells, and Fig. 162. Lean-To Roof in many other situations where appearance is not a matter of great moment, and where the essential thing is to obtain the shelter as cheaply and as easily as possible. The lean-to roof is shown in Fig. 162. It consists of a plain surface, one end or one side of which is raised to a higher level than the other side or end, and supported in this position by means of walls or by means of posts at the four corners. The position in which the surface is supported enables the rain water to drain freely from it, and thus it fulfills the require- ments of a roof so long as it remains water-tight. 122 E3 -a ° 9. CARPENTRY 111 Pitch or Gable Roof. Next to the simple lean-to roof with a single sloping surface comes the ordinary pitch or gable roof, which has two sloping surfaces one on each side of the center line of the building, coming together at the ridge in the middle. This form of Fig. 163. Pitch or Gable Roof roof, which is shown in Fig. 163, is very common and is also quite simple in design, and economical in construction, so that it has been very popular indeed for all classes of buildings except very large structures and city buildings. The slope of the roof, that is, its "pitch" or its inclination to the horizontal, may be varied to an infinite extent, from a very flat slope to a very steep one, and these Fig. 164. Gambrel Roof variations have been made in different countries and in different climates to suit either the taste of the designers or the practical requirements of the climate. The roof may be used in combination with roofs of other kinds and, indeed, it is usually used in this way, 123 112 CARPENTRY so much so that, although the simple gable roof Is the base of almost all the roofs of ordinary structures, it is sometimes hard to distin- guish it from among the other roofs which have been added as additions to it. Gambrel Roof. The gambrel roof is a variation of the simple pitch or gable roof and was probably developed from it to meet a new condition, namely, the necessity for more space in the portion of the building immediately under the roof surface. This form of roof has a sort of gable at each end of the building, but the gable is not triangular in shape, as is shown in Fig. 164. It will be seen that the roof surface has been broken near the middle on both sides of the building and that the portion below the break has been Fig. 165. Mansard Roof made steeper, and the portion above the break flatter than would be the case in a simple roof surface for a building of the same size with a gable roof. This arrangement allows of considerably more space and much greater head room in the attic. The position of the break in the roof surface may be varied to suit the taste of the designer, and the slopes of both the upper and the lower parts of the roof may be arranged as desired. Gambrel roofs may be seen on many old houses built in the Colonial days, and they have lately come again into favor for the roofs of cottages and small suburban or country houses. Mansard Roof. The mansard roof, called by the name of the architect who introduced it, is like the gable roof except that it slopes very steeply from each wall toward the center, instead of 124 CARPENTRY 113 from two opposite walls only, and it has a nearly flat deck on top. This form of roof gives better rooms in the attic space than does either of the two forms already described. It was at one time very popular for large suburban and city houses, but it is now seldom Fig. 166. Hip Roof with Ridge employed. The mansard roof is shown in Fig. 165. It bears a close relation to the so-called hip roof. Hip Roof. The hip roof also slopes from all four walls toward the center, but not so steeply as does the mansard roof. It is usually brought to a point or a ridge at the top, as in Fig. 166, but sometimes it is finished with a small flat deck, as in Fig. 167. The hip roof as a rule allows of very Httle useful space in the attic, and Fig. 167. Hip Roof with Deck if this type of roof is employed it is usually done with the idea of sacrificing the attic space and doing without the rooms under the roof for the sake of the exterior appearance. Valley Roof. In Fig. 168 is shown a very simple form of what is known as a valley roof. It is a combination of two simple pitch 125 114 CARPENTRY roofs which intersect each other at right angles. In the figure both ridges are shown at the same height, but they are not always built in this way. Either ridge may rise above the other, and the two roofs may have the same pitch or different pitches. If the ridge of Fig. 168. Valley Roof the secondary roof rises above the ridge of the main roof, the end which projects above the main ridge is usually finished with a small gable a, or a small hip h, as shown in Fig. 169. This arrangement does not make a pleasing appearance, however, and should be avoided if possible. Almost all roofs are hip and valley roofs, as it is very seldom that a building of any considerable size can be cov- ered with a simple roof of any of the forms described above. There Fig. 169. Hip and Valley Roof Showing Dififerent Roof Levels Around Inner Court are usually wings or projecting portions of some kind which must be covered with a separate roof which must be joined to the main body of the roof with valleys, and it is these valleys which are the 126 CARPENTRY 115 cause of most of the leaky roofs, as a large quantity of water collects in them and it is no easy matter to make them waterproof. Rafters. In all roofs the pieces which make up the main body of the framework are called the rafters. They are for the roof what the joists are for the floor, and what the studs are for the wall. The rafters are inclined members, spaced from 16 to 24 inches apart on centers, which rest at the bottom on the plate, and are fastened at the top in various ways, according to the form of the roof. The plate, therefore, forms the connecting link between the wall and the roof and is really a part of both of them. The size of the rafters varies, depending upon their length and the distances at which they are spaced. It is sometimes allowable to use them as small as 2X4 inches, but this should be done only for the lightest work. The size of the rafters for an ordinary dwelling house is usually 2X8 inches. In larger buildings, such as school houses, it may be found necessary to use rafters as large as 2X 10 inches, when they are of a considerable length. The material usually employed for rafters is the same as that used for the joists and for the studding. This is generally spruce in the eastern states and yellow pine in the Mississippi Basin, but may be hemlock in very cheap work. The size and spacing of the rafters vary to some extent with the location of the building, as in the northern part of the country the roof is called upon to carry a considerable weight of snow and ice, while in the south there is little or no snow and the roof is not called upon to carry so much weight. If snow freezes and packs on the roof it may weigh as much as twenty-five or thirty pounds per square foot of roof surface. The wind pressure must also be considered, as well as the weight of the material composing the roof itself. The connection of the rafters to the wall is the same in all the types of roofs described above. They are not framed into the plate but are simply spiked to it, being cut at the bottom so as to rest on top of it. Usually they extend out for a considerable distance beyond the wall to form the eaves, as shown in Fig. 170, and they are then cut over the plate and allowed to continue beyond it. The usual length for the eaves is about 1 foot, but it may vary to suit the taste of the designer of the building. Sometimes the rafter itself is not extended beyond the plate, but is cut off just as though it 127 116 CARPENTRY was not intended that it should continue beyond the wall line, and a separate piece called a "false rafter" is nailed against it alongside to form the projection for the eaves, as shown in Fig. 171. This piece does not always continue in the same line with the real rafter. Fig. 170. Rafter Extended to Form the Eaves Fig. 171. False Rafter but may, and usually does, make an angle with it, as shown in the figure, so as to give a break in the roof hne near the line of the eaves. It is sometimes desired to form a concealed gutter around the eaves, and in order to do this the joists are allowed to extend 10 or Fig. 172. Roof Showing Concealed Gutter 12 inches and on the ends of these a 2X4 is laid flat and nailed, and the rafters rest on this piece. The scantling is nailed directly above the plate and the gutter is run in notches cut in the 128 CARPENTRY 117 overhanging joists which also support the cornice. The general appearance is shown in Fig. 172 and the details of the construction in Fig. 173. There are four different kinds of rafters used in framing roofs, all of which may sometimes be found in a single roof frame, if the roof is of complicated design, while ordinary roofs may be framed with only the more simple forms of rafters. In Fig. 174 is shown in plan the framing for a roof in which all four kinds of rafters are to be found. AAA are the common rafters, which extend from the plate to the ridge and which are not connected with or crossed by any of the other rafters. B B B are jack rafters, which are shorter than the common rafters and which do not extend from the plate to the ridge, but are connected at one end to a hi-p or valley rafter. C C are the valley rafters, which are needed at every corner between the main building and an ell or other projection, while the hip rafters B D are found at the outside corners. At the points where the valley rafters are situated there are troughs or valleys formed by the roof surfaces — as these pitch down- ward on both sides toward the valley rafter — while at the outside corners, where the hip rafters are found, the roof surfaces pitch upward on each side to the hip rafter. This may be seen and perhaps better understood by looking at any hip and valley roof as actually constructed, and as this type of roof is very common there will be no lack of examples. Pitch of Roof. The pitch of a roof is the term used to indicate the slope of the sides of the roof surface or the inclination of these sides with respect to a horizontal plane or a surface absolutely flat and parallel to the horizon. Evidently the pitch of any roof may vary to an almost infinite extent. It may be absolutely flat or it may be practically vertical, or it may be inclined at any angle between Fig. 173. Construction of Concealed Gutter of Fig. 172 129 118 CARPENTRY these two limits. Unfortunately there are several systems in use for indicating the pitch or the amount of the angle of the slope, so that there is likely to be some misunderstanding about it. Usually, however, some one system is in use in any one section of the country and there is a general understanding that this is the system intended when speaking of the pitch of a roof. The most simple way of indi- cating the pitch and at the same time the most accurate way, is to give the angle which the roof surface makes with a horizontal plane Thus the pitch of the roof may be thirty degrees, or forty-five degrees Fig. 174. Plan of Roof Framing Showing Use of Four Kinds of Rafters or sixty degrees. This system is much in use among civil engineers, by whom it is favored on account of its accuracy and the small probability of its being misunderstood, but it is not much in use among carpenters and architects, who generally prefer to use some other system. Another method of indicating the slope of the roof surfaces is to take the rise of the roof at the center of the span, or the vertical distance from the top of the plate to the under side of the rafters at the center of the span, and to divide this distance by the span itself or the distance between the inside edges of two rafters which 180 CARPENTRY 119 come opposite to each other in the roof frame, at the point where they intersect the top surface of the plate. The fraction thus obtained is used to express the degree of slope of the roof, or the angle that the roof surface makes with the horizontal plane, in the following way : If the span of the roof between the edges of the rafters at the level of the top of the plate is 20 feet and the rise of the roof at the center, measured vertically from the top of the plate to the under side of the rafters, is 10 feet, then the roof is of half pitch, since the fraction obtained by dividing the rise of the roof by the span of the roof is one over two or one half. The angle which this roof surface makes with the horizontal plane is forty-five degrees, since the rise at the center is equal to half the span, and the rise of the sloping rafter is equal to its run or its projection on the horizontal plane. This slope is also called a square slope or a square pitch for the reason that the rise of the rafter is equal to its run. In this roof also it will be seen that the rafter rises a distance of 12 inches for each foot of run, counting the rise always from the level of the top of the plate and the run from the point where the under side of the rafter intersects the top of the plate. Thus at the center of the roof the run is 10 feet and the rise is also 10 feet. If this same roof were one of full pitch, the rise at the center of the span would be 20 feet, equal to the span itself, and there would be 2 feet of rise for each foot of run. This would make a very steep roof, in fact it is very seldom so steep a roof is used in ordinary work. A two-thirds pitch would be a little less steep than the full pitch, between the full pitch and the half pitch, and this roof would have a rise of 16 inches for each foot of run, so that if the span were 20 feet, as in the case of the other roofs just mentioned, the rise at the center of the span would be 160 inches or 13 feet and 4 inches. The reason why this pitch is called a two- thirds pitch is that in the case of a full pitch roof the rise for each foot of run is 24 inches, in this case the rise for each foot of run is 16 inches, and 16 inches is just two-thirds of 24 inches. Also the rise at the center of the span, 13 feet and 4 inches, is just two-thirds of the span, which is 20 feet. Whenever it is desired to give a roof a steeper pitch than the half pitch, the two-thirds pitch is generally employed. If the roof is one of one-third pitch, the rise of the rafter for each foot of run will be one-third of the 24 inches which are required to make a full pitch, one-third of this being 8 inches. Thus a one-third 131 120 CARPENTRY pitch roof has a rise of 8 inches for each foot of run, and the rise at the center of the span is one-third of the entire span. If the span of the roof is 20 feet, the rise at the center of the span will be 6 feet and 8 inches, or just one-half as much as in the case of the roof of the same span and with a two-thirds pitch. If the roof has a "one-quarter pitch," this means that the rise of the rafter for each foot of run is one-quarter of 24 inches, which is 6 inches, and that the rise of the roof at the center of the span is one- quarter of the entire span. If the roof has a span of 20 feet, this will make the rise in the case of a one-quarter pitch equal to 5 feet. The pitches mentioned above are the most common pitches and those most generally used, though, of course, any pitch may be used as desired. The two-thirds pitch corresponds to an angle with the horizontal of about fifty-three degrees, and the one-half pitch corre- sponds exactly with an angle of forty-five degrees. The one-third pitch corresponds to an angle of thirty-three and three-quarters degrees and the one-quarter pitch corresponds with an angle twenty- six and one-half degrees. From this it will be seen that the names of the pitches, one-third, one-half, and one-quarter, do not express the relation of the angles which the various slopes make with the hori- zontal to the angle made by the roof of full pitch. There are several factors which enter into the problem of deter- mining the most suitable pitch to give a roof, and they must be carefully considered before arriving at a decision. In the first place there is to be considered the appearance of the finished roof when the building is completed. In this connection it may be said that personal preference and individual taste on the part of the designer are the determining factors, and that no hard and fast rules can be laid down. Another thing which must be thought of is the relative cost of the different slopes or pitches, as this is often of great impor- tance and, in the case of a large number of buildings, would make considerable difference in cost. It may be said that in general a roof with a comparatively low pitch, say about thirty degrees, corre- sponding to a rise of approximately 6| inches per foot of run, is the most economical so far as the roof framing alone is concerned. Of course such a roof gives no accommodation in the attic portion of the building. Consideration must also be given to the question of the climate in which the proposed building is to be erected, as this 132 CARPENTRY 121 will have a very decided influence upon the decision in regard to the most suitable pitch for the roof. In cold northern climates where the snowfall is great, it is best to have a roof with a steep pitch, so that it will shed the snow and rain, or melted snow as quickly and as thoroughly as is possible. In a warm southern climate where there is no snow and where the rain fall is not large, a roof of smaller pitch may safely be used and will be more economical of construction. The character of the material to be used for covering the roof sur- faces must also be remembered in determining the pitch, since if this roof covering is very impervious to water the roof may be given a low^er pitch than if the roof covering is more easily penetrated by rain and snow. In general it may be said that roofs covered with slates may be safely given a pitch of from 5 to 5| inches to the foot run, while a roof covered with shingles must not be flatter than thirty degrees or nearly 7 inches to the foot run. Flat roofs should be covered with some preparation of tar and gravel, or with metal, tin, copper, galvanized iron, or zinc. Any roof which has a rise of less than 3 inches to the foot may be considered to be flat. ROOF FRAME Layout of Roof Plan. The laying out of the roof plan for a building is a problem which requires some little thought and skill and it may be well to give a little space to a consideration of the best way in which to approach this problem. Suppose that we have a frame building whose general outhne in plan is as shown in Fig. 175, and on which we wish to plan a hip and valley roof. There are, we will say, two projections or wings on the front of the building, at A A, another wing on the right-hand side of the building at B, and another wing on the back of the building at C. The first thing to do is to draw the rectangle A B C D,'m Fig. 176, enclosing the main portion of the building, and leaving out the wings or projections. From each corner of the rectangle A B C D, may be drawn a hne at forty-five degrees, with the side of the rec- tangle, each pair of which will meet at the points R R, and these points R R may be connected by a line parallel to the long sides of the rectangle. This is a plan of a simple hip roof covering the main portion of the building, E being the ridge and G G G G being the hip fines. The projecting portions or wings are, however, not yet 133 122 CARPENTRY covered and in order to take care of them some further planning is necessary. Let us consider the two wings on the front of the build- ing, marked A A,in Fig. 175. We will decide to cover these with a simple gable roof and for 1 ^ this the first step is to draw in the ridges. These ridges will, of course, come exactly in the center of the wings and will be shown on the plan by a line in the center of the plan of the wings, perpendicular to the line of the front. These lines should be drawn in as shown in Fig. 177, where they are marked E E. The lines E E intersect the hip lines marked G G, in Fig. 176, at a point about half way between the corners D and C, and the peaks R R. In Fig. 175. Ground Plan of Building on Which Rafter Must Be Placed Fig. 176. Plan of Roof Covering Main Portions of the Building order to look well the slope or pitch of the sides of the pitch roof which covers the wings A A, must be the same as the slope or pitch of the end of the hip roof shown in Fig. 176 and there marked S S, 134 CARPENTRY 123 and thus the roof of the wing A will on that side become a part of the roof over the main portion of the building, and the lower portion of the hip hne G may be erased, leaving only the upper portion showing as a hip, as indicated in Fig. 177. The other side of the pitch roof over the wing A will, however, not correspond with any slope in the roof over the main portion of the building and must intersect it in some line. Since the ridge E is at the top of this roof surface and the wall line of the wing A is at the bottom of the roof surface, a line drawn from the corner in which the wall line of the wing intersects the wall line of the main portion of the building, to X.- - y B zr-- £ G/ / / \ / 1 V \ f/ C £ E Fig. 177. Added Development of Roof Plan Covering Wings A, B, and C of Fig. 175 the point in which the ridge line intersects the hip line of the main roof, will be the line of intersection of the two roofs. This Hne is shown in Fig. 177, where it is marked ¥. The Hne F in the plan. Fig. 177, will represent a valley. Thus we have the two wings, A A, Fig. 175, completely roofed over, and the small roofs connected to the large main roof. Suppose that we wish to cover the wing on the right-hand side of the building also with a simple gable roof. This wing is marked B in Fig. 175. We proceed in the same way as explained for the wings A A, drawing the ridge Hne E, in Fig. 177, until it intersects the 135 124 CARPENTRY hip line of the main roof G and then drawing the valley line F. The slope on the back side of the roof over the wing B should have the same slope as the back side of the main hip roof and, therefore, the lower part of the hip line G, starting at the point B, can be erased, leaving only the three lines, E, F, and G, shown in Fig. 177. Thus the wing B is completely roofed over and shown in plan. The line F in this case also represents a valley. Suppose that we wish to cover the wing on the back of the building, marked C in Fig. 175, with a hip roof instead of a gable roof. We will start at the outside corners, and from these points draw lines G G in Fig. 177 at forty-five degrees with the front and side wall lines of the wing, until they meet. The lines must meet exactly in the center of the wing between the two side wall lines, and from this point a line should be drawn at right angles to the front wall line of the wing, but away from this wall line instead of towards it. This line is marked E in Fig. 177. It will intersect the hip line G of the main roof and from this point of intersection a line F should be drawn at forty-five degrees, which will meet the side wall line of the wing in the point in which this side wall line meets the main wall of the building. One slope of the roof over the wing C, Fig. 175, will be the same as the slope of the end of the main hip roof, and so the lower part of the line G, starting at the point A, may be erased and the upper part only left to show as a hip line. The line F in this case also will be a valley line. Thus the wing C, Fig. 175, will be completely roofed over and shown on the plan. Our roof plan is now complete in outline, all the lines marked E being ridges, all the lines marked G being hip lines, and all the lines marked F being valley lines. The same method of procedure may be followed out in the case of any roof plan, and the final complete plan obtained by successive steps as explained above. The first step is always to lay out the roof over the main portion of the building and then to proceed with the roofing of the projecting portions or wings. Ridge. In the lean-to roof the rafters rest at the top against the wall of the building of which the ell, or porch, is a part; and the work of framing the roof consists simply in setting them up and securing them in place with spikes or nails. The pitch roof, how- ever, is formed on the principle that two pieces which are inclined against each other will hold each other up, and so the rafters must 136 CARPENTRY 125 rest against each other at the top in pairs, as shown in Fig. 178. It is customary to insert between the rafters, at the top, a piece of board about 1 inch in thickness and deep enough to receive the whole depth of the rafter, as shown at A in Fig. 179. This piece of Fig. 178. Construction of Ridge Fig. 179. Placing of Ridge Pole between Abutting Ends of Rafters board is called the ridge or the ridge pole and extends the whole length of the roof. It serves to keep the rafters from falling side- ways, and keeps the roof frame in place until the roof boarding is on. It is sometimes extended above the rafters, and forms a center for some form of metal finish for the ridge, as shown in Fig. 180. Interior Supports. In small roofs which have to cover only narrow buildings and in which the length of the rafters is short, there is no necessity for any interior support, and when the rafters Fig. 180. Ridge Pole Extended above Roof Fig. 181. System of Interior Supports for Rafters have been cut to the correct length, set up against the ridge, and secured in place, the roof framing is complete. In long spans, how- ever, the roof would sag in the middle if it were not strengthened in some way, so it is customary to support long rafters as near the 137 126 CARPENTRY center as possible. This support may be formed by placing a piece of studding under each rafter, somewhere between the plate and the ridge, and if this is done very much lighter rafters can be used than would otherwise be considered safe. It is claimed by some that it is cheaper to do this than to use the heavy rafters. A more common method is to use fewer upright pieces and to place a horizontal piece A on the top of them, running the whole length of the building and supporting each rafter. This is shown in Fig. 181. An upright piece B should be placed under every sixth or sev- enth rafter in order to give the necessary stiffness to the whole construction. For the uprights, pieces of ordinary studding 2X4 inches or 2X3 inches in size may be used. When there is to be a finished attic in the building, these upright studs may be made to Fig. 182. Collar or Tie Beams as Interior Support for Rafters Fig. 183. Example of Double Gable Roof form the side walls of, the attic rooms, and are then spaced about 16 inches on centers to receive the laths. Such walls are called dwarf walls. Another form of interior support is the collar beam or tie beam. This is a piece of timber which extends between the rafters on opposite 138 3 n •^ X a FIREPLACE IN DINING ROOM OF HOUSE FOR MR. C. M. THOMPSON, CAMBRIDGE, MASS. Cram, Goodhue & Ferguson, ArcMtects, Boston and New York. CARPENTRY 127 sides of the roof and ties them together, as shown at A in Fig. 182. It may be a piece of board about 1 inch thick and 8 or 10 inches wide, which is nailed onto the side of the rafter at each end. It is placed as near the center of the rafter as may be practicable, and in the case where a finished attic is required it forms the support for the ceiling. For this reason it must be at a considerable height from the attic floor, and can not always be placed very near the center of the rafter. The important point is to see that it is well nailed at each end. Fig. 184. Rafter and Wall Framing for Double Gable Roof Double Gable Roof. A very interesting form of gable roof is that in which there is a double gable with a valley between, which forms the roof of an ell when the main roof is a simple pitch roof. This form of roof is shown in Fig. 183. Fig. 184 shows how such a roof may be framed. The piece A is placed in the wall and supported by the studding so as to serve as a plate to receive the ends of the valley rafters B. These, together with the piece C, form the framing for the shallow valley between the two gables. The valley rafters on the outside, marked D in the figure, are similar to those used in 139 128 CARPENTRY the case of a single gable. The pieces E E are jack rafters and are very short. This form of roof is not common, but in some places it gives a good effect. Qambrel Roof. A gambrel roof is framed in very much the same way as is a pitch roof or a hip roof. The slope of the roof, however, is broken at a point between the plate and the ridge. The part of the roof above this break makes an angle with the hori- zontal plane of less than forty-five degrees usually, while the portion below the break makes an angle with the horizontal plane greater than forty-five degrees. This is shown in Fig. 185. The lower slope may almost be considered a part of the wall, and at the point where the slope changes there is a secondary plate from which the upper slope starts, as shown at A in Fig. 185. The secondary plate may be utilized as a support for the ends of the ceiling joists B, which should also be securely spiked to the rafters, as shown in the figure. The rafters C forming the upper slope, must be cut over the plate A, and firmly spiked to it, while at the top they rest against a ridge board B. The rafters E, form- ing the lower slope, are cut out at the top so as to form a seat for the plate A, and must be very securely fastened at the bottom to the main wall plate F. It is an excellent plan to have the floor joists G spiked to the lower rafters, so as to act like tie beams across the building and to counteract the outward thrust of the rafters. Sometimes these floor joists are dropped below the wall plate F, and are supported on a ledger board notched into the wall studding I. This construction is not so good as that shown in the figure, because the joist is not so effective as a tie across the building. If it is employed the floor joist must be securely nailed to the wall studding 7, and they must not in any case be dropped more than 2 or 3 feet below the plate. The plate must Fig. 185. Framing for Gambrel Roof 140 CARPENTRY 129 Fig. 186. Method of Finding Contour for Gambrel Roof always be firmly nailed to each stud to prevent it from being forced outward as it receives the thrust from the rafters E. A good rule for determining the point at which to place the secondary plate, and for determining the general shape of the roof, is illustrated in Fig. 186. Let the points A and B represent the main plates on each side of the building. Draw a line A B between them and bisect this line at C. With C as a center and C yl as a radius describe the semicircle A D E F B. At any distance G above A B draw a line D F parallel to A B, cutting the semicircle at the points D and F. Also bisect the arc at E. Then by joining the points A D E F and B by straight lines as shown, we will have the outline of a gambrel roof. The proportions of the roof may be varied by varying the distance G. Gambrel roofs are not very strong unless they are stiffened by cross partitions in the attic stories, and these should be provided whenever it is possible. No gambrel roof, unless it is well braced, should be used on a building which is exposed to high winds, or which is likely to receive a heavy weight of snow. Mansard Roof. A mansard roof is framed in very much the same way as is a gambrel roof, as may be seen in Fig. 187. Resting on the main wall plate A, we have a piece B which is inclined slightly inward, and which supports at its upper end a second- ary plate C. On the plate C rests the outer end of the deck rafter D which is nearly horizontal. The piece £ is a piece of stud- ding, 2X4 inches to 4X6 inches in size, depending upon the size of the roof. It sup- ports the whole weight from the rafters, carrying this weight to the main wall plate and thence into the walls of the building, always be straight, and the curved shape Fig. 187. Framing for Mansard Roof This member should which is usual on mansard roofs is obtained by the use of the furring piece E. The piece E is nailed to the upright member B at the top, and at the 141 130 CARPENTRY Fig. 188. One Form of Dormer Window bottom it is secured to the lookout F, which also forms a support for the projecting cornice. The floor joist G is supported on a ledger board H, or it may rest directly on the plate A. The piece of stud- ding I is merely a furring stud to form the wall of the attic room. It may be omitted en- tirely if desired, or if the attics are to be unfinished. The ceil- ing joist K may be supported on a ledger board as shown, or may be simply spiked to the studding I or to the upright B. The studding I may rest directly on the floor joist G with a sole piece L at the bottom as shown. The plate C should be of a good size, at least 4X6 inches, and should not be placed more than 2 or 3 feet above the ceiling joists K. The ceiling joists act as ties across the building and prevent the plates C from spreading apart, as they receive the thrust from the rafters D. For this reason it is better to have the ceil- ing joist K fastened to the upright B rather than to the furring stud /. Dormer Windows. In Figs. 188 and 189 are shown what are known as dormer windows, this name being applied to all windows in the roofs of buildings, whatever may be their size or shape. The figures show two different kinds of dormer windows which are in general use, the one shown in Fig. 188 resting entirely on the roof, while the one shown in Fig. 189 is merely a continuation of the wall of the building above the line of the eaves. The second type is often seen on buildings only one story in height, while the other kind is employed on larger structures. Fig. 189. Another Form of Dormer Window 142 CARPENTRY 131 In order to construct a dormer window an opening must be made in the roof surface, and the window must be built up over the opening. Headers are framed in between two of the rafters as shown at A and B in Fig. 190, and thus a rectangular opening is formed in the roof frame. The rafters C and D, which form the sides of the opening, are called trimmers and should be nmch stronger than the common rafters. Usually the trimmers are made by doubling the ordinary rafters. The headers receive the ends of the rafters which are cut by the opening, and must be large enough to carry the weight which comes from them besides supporting the walls of the dormer. Timbers 4X8 inches to 6X10 inches, according to the size of the dormer, are usually large enough for the headers and often smaller timbers may be safely used. The headers are shown in section at A and B in Fig. 191, and it will be noticed that they are not used in exactly the same way. The piece at the top A is so placed that its longer dimension is at right angles to the plane of the roof, while the piece at the bot- tom B has its longer dimen- sion vertical. In the case shown in Fig. 189, where the front wall of the dormer is merely an extension of the main wall of the build- ing, there is no need of the lower header B, the main Fig. lOO. RoorFraming for Dormer Window wall plate taking its plact and supporting the studding for the front wall of the dormer, as shown at the right-hand side of Fig. 191. Fig. 191 shows sections taken through two dormers of the types mentioned above, parallel to the direction of the main rafters and at right angles to the main wall plate of the building. At the left is a section taken through the type of dormer shown in Fig. 188, while at the right a section of the other type is shown. The studs C C which form the side walls of the dormer, are notched over the trimmer rafters and roof boarding about 1 inch, and allowed to con- 143 132 CARPENTRY tinue downward to the attic floor. This is shown at section D D. At E is a section of the trimmer rafter, C is the wall stud, G is the attic floor boarding, and iZ^ is a section of one of the attic floor joists. The studs C are in line with the studs forming the side walls of the attic room, so the studs I can not be carried down to the attic floor. They are stopped, at the bottom, against a 2X3 inch strip K which is nailed to the side of the trimmer rafter. At L is the ridge board, and M M M are the short rafters which form the pitch roof of the dormer. They may be very light, as they are short and carry little weight. They rest, at the foot, on a plate 0, and at the top bear Fig. 191. Framing Details for Both Types of Dormer Window? against the ridge board L. In the dormer shown on the right of the figure the rafters P are in planes parallel to the main rafters, and a furring piece S may be nailed to each of them so as to give the dormer roof any desired curve. Besides the openings in the roof frame for dormer windows there must be other openings for chimneys and skylights. These are formed in the same way as explained for the dormer openings, with headers and trimmer rafters. A plan of such an opening is shown at E in the roof framing plan in Fig. 174. 144 CARPENTRY 133 RAFTERS The ends of rafters are usually cut to fit accurately against one another and against the plates on which they rest. The cutting of these bevels is not at all difficult when the relation of the rafter to the roof surfaces is seen and the steel square is used to show this relation. Common Rafters. Method of Cuttiyig Bevel. Fig. 192 shows the bevels that are used on the common rafters in a simple gable roof such as is illustrated in Fig. 163. In Fig. 191, B is the plate and £ is a point midway between the two plates, and the distance D E is the run of the common rafter C. P is the point where the line drawn through D parallel to the edges of the rafters is directly above the point E. The distance DP is usually taken as the length of the rafter. The length is taken from these points because the dis- tance D E represents the exact run and E P represents the rise for this run. The first step to take in laying out the rafter is to locate the point D on the uncut piece as shown in Fig. 193. The point is chosen so that it is far enough from the end to form the eaves FD, and the distance T D is usually taken 2 inches on a 2 X 4 inch and 3 inches or more on larger size rafters. It is well to remember that the measurement is to be taken from the top in order that the roof surface may always be even and smooth throughout. Now the edge of the blade of the square must coin- cide with D, but the position which the square will take depends entirely on the pitch. In all cases 12 inches is used on the blade of the square and the figures on the tongue depend on the rise per foot of run. If the rise of the rafter is 12 inches to the foot, 12 inches should be used on the tongue also. If the rise is 10 inches, use 10 inches on the tongue. In Fig. 192. Diagram Showing Bevel Used on Common Rafters 145 134 CARPENTRY the cut the rise is 8 inches to the foot, the run is 12 feet, the rise is 8 feet, and 12 and 8 are the figures used. Fig. 193. Method of Laying Out Rafter with Steel Square Usually the carpenter uses the edge M N to obtain this bevel but the line F P may also be used. The line D is the heel or plate cut, as shown in Figs. 192 and 193. The rafter is sawed along the lines F D and D 0. Now the next step is to find the length D P. This may easily be determined by any one of four methods. The easiest of these is to turn to the rafter table on the square. Oppo- Fig. 194. Method of Cutting the Bevel at the Top of the Rafter site 12 — 8 — I and under the 12 (indicating feet run) the length is given as 14 feet 5 inches. By extracting the square root of the sums of the square of rise and run the same result is obtained. The third 146 CARPENTRY 135 way is to measure the distance in inches and twelfths from the 12 on the blade of the square to the 8. Each inch represents one foot of length and each twelfth rep- resents 1 inch. The distance is 14 feet 5 inches. The fourth method is to use the method illustrated in Fig. 194. First locate the line DP and then beginning £t D move the square along the edge of the rafter as many times as there are feet and fractions of feet in the run; thus the point P is determined. A little study of the figures will suffice to reveal to anyone the reason for this method of Fig. 195. procedure. Every time the square is moved into a new position it has advanced 12 inches or 1 foot along the run of the rafter, since the distance D E is 12 inches and is meas- ured horizontally. After the square has been moved twelve times it has advanced 12 feet along the run of the rafter or the distance required. This gives the position of the top bevel. It should be Fitting a Rafter against Ridge Board Fig. 196. Cutting Rafter for Con- cealed Gutter Fig. 197. Another Method of Cutting Rafters for Concealed Gutter noticed that for a run of 12 feet the square must be moved along twelve times; for a run of 8 feet, eight times; and so on. The run of the rafter may be easily obtained by subtracting one-half the 147 136 CARPENTRY thickness of the ridge board from one-half of the total span of the roof from outside to outside of wall plates. Fig. 194 shows the rafter in the position which it would occupy in a building, the plate and a part of the wall studding being indi- cated. When the rafter is cut along the line N S, Fig. 193, it is ready to be put on the building. In case, however, that a ridge board is used to hold the rafter in place, as shown by R in Fig. 192, the rafter is cut parallel to N S but shorter, as shown in Fig. 195, one-half of the thickness of the ridge board being cut away. The /? c\ ^c /I \ / \^ y' / / / y^ \ \ Fig. 198. Bevels for Valley and Hip Rafters cut at J) 0, Fig. 192, is horizontal, and the bevels at NS and VK, Figs. 193, 194, and 195, are plumb cuts. In case a concealed gutter is used and the rafter is set directly over the wall, the line DP coincides with the line M N, Fig. 193, and the rafter has only the horizontal cut at the bottom or a hori- zontal and vertical cut, as shown in Figs. 196 and 197. Valley and Hip Rafters. In Fig. 198 the rafters C C are valley rafters and, although the bevels for these rafters are not the same as the common rafter in either roof surface, yet the bevels depend upon the relation between the common rafters and the valley rafters. 148 CARPENTRY 137 It is best to consider the common rafter as the hypotenuse of a right triangle or as the diagonal of a rectangle whose length is the run of the rafter and whose width is the rise of the rafter. In studying the valley rafter it is evident that there are three dimensions to be considered. Rafter C extends to the right to the ridge of the main roof besides rising. It may, therefore, be considered as the diagonal of a rectangu- lar solid. For instance, if the run of the common rafter is 12 feet, the rise 10 feet, and the distance MR is 8 feet, the valley rafter will form the diagonal of a rectangular solid 12 inches X 10 inches X 8 inches, and its length and bevels can be found as shown in Figs. 199 and 200. In Fig. 198 we find the run which is the hypotenuse of the triangle C R M. That is, the run of the valley rafter is taken from the dis- tance between the 12 and 8 on the square. It is 14 A inches, showing that the run of the valley is 14 feet 5 inches. Now the rise is the same as the rise of the common rafter C R. That Fig. 199. Method of Finding Bevels for Various Runs of Rafters Fig. 200. Cutting Bevels on Common Rafter is, it is 10 feet and the bevel at the foot of the rafter is cut along the blade of the square when the figures read 14 A inches on the blade and 10 inches on the tongue. The plumb cut at the top of the rafter is made by holding the square in the same position and cutting along the tongue. 149 138 CARPENTRY The length of the rafter is determined either by measuring the distance from the 14tV and 10 on the square or by finding the square root of the sums of the squares of the three dimensions. The latter method gives l/l44+100 + 64=17.5=17 feet 6 inches (approx.). The layout of a hip rafter is the same in principle as the layout of a valley rafter. To find the run of a hip rafter, find the diagonal of a square whose sides are equal to the run of the common rafter. That is, if the run of the common rafter is 10 feet, the run of the hip rafter is the hypotenuse of a right triangle whose sides are 10 feet A F 1 / i 1 1 \ 1 \ 1 l\ 1 \ i \ / I. z c E Fig. 201. Plan for Valley Rafter Connecting Two Roofs of Unequal Pitch and Width and this distance is 14.14 feet or 14 feet 1^ inches. The rise of the hip is the same as the rise of the common rafter. If, then, the rise is 8 feet, use 14| inches on the blade and 8 inches on the tongue to lay off the horizontal and the plumb cuts. The length of the hip is the hypotenuse of the triangle between the 14|-inch mark on the blade and the 8-inch mark on the tongue. To compute this mathemat- ically we have 1^10H10H8'= V 264= 16.25 feet, or 16 feet 3 inches. When a valley rafter serves to connect two roofs of unequal pitch and width, the problem is more complex. In Fig. 201 a lOX 12 foot roof covers the main building and an 8X 12 foot roof covers the ell on the left. The rise of rafter ^ C is 13 feet 4 inches, the rise of 150 CARPENTRY 139 rafter C D is 7 feet 6 inches, and the ridge of the main roof is nearly 6 feet above the ridge of the ell. One of the valley rafters C F runs to the ridge of the main roof, its rise being 13 feet 4 inches. In extending to F the valley runs 16 feet toward the main ridge, and the distance AF is found by propor- tion or by drawing the plan to scale and measuring. In using proportion, take the run of the common rafters A C and C D. If the ridge of the ell roof coincides with the ridge of the main roof, the common rafter C M would be in proportion with C D, thus : Rise of CD: vise of CM:: run of C Z): run of C M Substitutmg ^,_g„ . ^g,_^„ . . 10' : run of C M 90:160 :: 120: run of CM Run of CM=213|" = 17'-9i" Now find the diagonal distance C Fhy mathematics or the use of the square. On the square use 17f inches on the blade and 16 inches on the tongue. The distance is 23 feet 11 inches. The rise is 13 feet 4 inches. Hence, use 23x1 inches on the blade and 13| inches on the tongue to give the horizontal and plumb bevels and length of the valley. To cut the side bevel at the top, use the distances C M and A C, cutting along the C M side. In order, however, that this cut car be made accurately, the rafter must be backed and the square laid on the backed surface. Few carpenters, if any, ever back a \aLey rafter and consequently a roundabout method is used to get this bevel. The common result is, that the bevel very rarely fits snugly against the ridge. Where the rafter is not more than 2 inches thick, the misfit is not so noticeable, but in 4-inch material the open joint must be "doctored" by gauging and resawing after it has been tried. When the rafter is cut properly and set in place it will be found that the plane of its top surface does not lie in either of the two roof surfaces. The surface of the top lies at an equal angle to each roof surface and one edge extends up above the other rafter in both roofs. Fig. 202A shows how the edges of a hip rafter extend over the plate at the bottom. To overcome this the rafter can be 151 140 CARPENTRY backed or cut shorter as shown in Figs. 202B and 202C. To back the rafter, lay the square on the bevel at the bottom end in the position the plate will occupy. Now mark the points C D and draw the lines D F and C G from these points parallel to the edge of the rafter and ^2? A B C Fig. 202. Cutting Rafter to Prevent Ends from Projecting Over Plate cut away the triangular part A B DFH and A EC GH. This is an expensive means of making the rafter conform to the roof surface and most carpenters merely shorten the rafter until the outside corners conform to the surfaces, as shown in Fig. 203. If the rafter is not to be backed, the effect of backing can be easily obtained by nailing a thin board on the top of the rafter and giving this board the proper bevel. The method is illustrated in Fig. 204. The clapboard A is nailed at the edges and the one side wedged up to the angle the backing would take, care being taken to allow the square to touch the edge B B of the rafter at C C. The square used on the side of the rafter gives the plumb cut CD, and CM over the clapboard gives the side bevel. In cutting, the saw is held at an angle to coincide with both CD and C M. In the rafter B E,F\g. 201, the horizontal cut at E is obtained by using 23 feet 11 inches and 13 feet 4 inches and is the same as the Fig. 203. Simpler Method of Avoiding Projecting Hip Rafter 152 CARPENTRY 141 cut at C. The length of the rafter can be found by using proportion or by finding the length oi B D. In using proportion, it is evident that B E, the run of the short valley, is to Ci^ as 7 feet 6 inches, or 90 inches, is to 13 feet 4 inches, or 160 inches. 5£;:287::90:160 287X90 BE= 160 = 161.5 inches=13 feet 5h inches The rise is 7 feet 6 inches and the run is 13 feet 5| inches. Another way in which the problem may be solved, is to find where the ridge of the ell intersects the main roof surface. The inter- section is at a height of 7 feet 6 inches which is tW of the run of Fig. 204. Method of Using Clapboard to Cut Bevel on Rafter A C, or ® of 16 feet and the distance B D is just 9 feet. Hence, the run of 5 E is Vl¥+9'= 13.45= 13 feet 5| inches, and this is -A of the run of the rafter C F. Hence, the run oiCFis 13.45X V = 23 feet 11 inches. The end cut at B on B E, that is, the bevel that fits against CF, to be cut accurately, must be handled like the side bevel at F. First cut the bevel at the plate and get the backing line that makes B E he in the main roof surface. Now, at B, either back the rafter a short distance, or use a clapboard as in Fig. 204. Jack Rafters. Fig. 205 shows the plan of the roof in which there are, in addition to hip and valley rafters, sets of jack rafters. A B and B D are hip rafters, C ^ is a valley rafter, and the other rafters are common and jack rafters. At B E and E H are shown 153 142 CARPENTRY the ridge boards. Of the jack rafters there are three different kinds: those Uke IJ which run from the valley rafter to. the ridge board; those like KL which run from hip rafter to plate; and those hke N T, which run between the hip and valley rafters. These jack rafters differ only in respect to the bevels which have to be cut on them. The rafter 7 J is a simple plumb cut at the top, similar to the cuts at the top of the common rafters, and at the bottom where the rafter meets the valley there are two cuts — a plumb cut and the D K \ - E ^ -- - .• --i 1 1 \ 1 1 \ 1 1 \ 1 -^ I 1 '^ 1 1 1 / T I / / \ \ \ k ! -\ 1 1 A // ) 1 A- ^J // - ^ r \ 1 / '/ *s V 1 1 -4 ■/ / / J y Y •^ / J{ "^^ A. c 1 1 1 1 1 1 1 Fig. 205. Roof Plan Showing Hip and Valley Rafters and Jack Rafters side cheek cut — which are similar to the cuts in a valley rafter where it comes against a ridge board. This cut has been previously explained. The rafter iCX has a simple horizontal cut at the bottom like that used on the common rafter, but at the top there are two cuts similar to those at the foot of rafter 1 J. The rafter N T has two cuts at both top and bottom. All these bevels are obtained just as the bevels for the hip and valley rafters. The length of a jack rafter is proportional to its distance from the ridge or plate to which it is parallel. The longest jack rafter is 154 ^ s ^ M O .« Q o n i* S ID « H O CARPENTRY 143 equal in length to a common rafter, and the length steadily decreases as the distance of the rafter from its first full length rafter. The exact difference in length between the first jack rafter and the next, is determined by finding how far apart the jack rafters are to be placed, and comparing this distance with the distance from the top of the first full length jack rafter to the point where the hip or valley rafter rests on the ridge board or plate. Suppose, for instance, that the rafters are to be spaced 2 feet apart, and the length of the com- mon rafter is 10 feet. If the distance from the top of this rafter to the point where the valley rafter is fitted against the ridge is 12 feet, it is evident that each rafter will be 2 feet shorter. That is, the B A Fig. 206. Roof Plan Showing Rafters Cut for Ogee Roof second rafter will be 8 feet and the next 6 feet and so on. We use six spaces, although there are only five rafters, there being no rafter used where the valley and ridge join. Curved Hip Rafters. A form of hip rafter which is sometimes a source of considerable trouble is one which occurs in a curved roof, such as an ogee roof over a bay window, or a curved tower roof. The slope of the curve to which the top edges of the common rafters must be cut, is determined from the shape of the section of the curved roof surface, but the curve at the top of the hip rafter is entirely dif- ferent and must be determined in another way. The principle used in finding this curve is the same as was employed in the case of the valley rafter, namely, that any line drawn in the roof surface parallel to the wall plate must be horizontal, or that it must be exactly the same elevation throughout its entire length. 155 144 CARPENTRY Fig. 206 shows how this may be appHed. At A is shown a plan of an ogee roof over a bay window with a hip rafter D E and com- mon rafters. At B is shown an elevation of one of the common rafters cut to coincide with the curve of the roof surface. The shape of the curve may be varied to suit the fancy of the designer. At C is shown an elevation of the hip rafter D E, showing the curve to which it must be cut in order to fit into the roof. To determine this curve we draw on the roof plan at A any num- ber of lines, parallel to the wall plate. These must be horizontal, so that any point in either of the lines is at the same height above the top of the plate as in every other point in the same line. The lines F G and H I in the elevation, shown at B and C, represent the level of the top of the plate. By projection we find that the line KO XL, for example, is at a distance MN above the top of the plate at the point where it crosses the common rafter shown at B. Every other point in this line is at the same elevation, including the point 0, in which it intersects the center line of the hip rafter D E. By projection we can locate the point in the elevation shown at C, making the distance OP equal to the distance MN. In the same way we can obtain as many points in the curve of the hip rafter as we have lines drawn on the roof plan. The lines may be drawn as close together as we wish, and the number of points obtained may thus be increased indefinitely. Wlien a sufficient number of points have been located, the curve can be drawn through them, and a pattern for the hip rafter is thus obtained. The shape of the curve for a valley rafter is found in the same way as explained for a hip rafter. ATTIC PARTITIONS It is often necessary to build partitions in the story directly beneath the roof, and such partitions must extend clear up to the under side of the rafters and be connected with them in some way. This makes it necessary to cut the tops of the studs on a bevel to correspond with the pitch of the rafters, and the cutting of this bevel is not always an easy task. Fig. 207 shows the framing plan of the roof of a small simple building. In this figure A B\s the ridge. The plate extends around the outside from C to I) to E to F, and back again to C; and G H I J KLare the rafters. A partition H M 156 CARPENTRY 145 is shown beneath the roof running diagonally across the building, making an angle with the direction of the rafters and an angle with the direction of the ridge. At iV is shown another partition running parallel to the ridge, and at P Q still another, running parallel to the rafters. Now since all the rafters slope upwards from the plate to the ridge, it is evident that the tops of all the studs must be cut on a bevel if they are to fit closely against the under sides of the rafters. This is illustrated in Fig. 208, where the stud A must fit against the rafter B. To take the simplest case first, let us consider the stud marked r 1 // / / t Q 1 1 ^ / AT // 1 1 1 1 1 1 1 ^ / / /y o \ p 'b\ 1 R 1 ^_ - — - — — — — — __J c >v ^ / 1 F Fig. 207, Framing Plan of Roof of Simple Building jR, Fig. 207. Since all the rafters have the same pitch or slope, all the studs in the partition iV will have the same bevel at the top, and if we find the bevel for one we can cut the bevel for all. Fig. 208 shows this stud drawn to a larger scale and separated from the rest; ^ 5 Z) C is a plan of the stud, and the rafter is shown at E F H G. We will take the distance F H, or the run of the part of the rafter shown, as one foot exactly. Now if A^ and B^ represent a side eleva- tion of the rafter and stud, the run of the part of the rafter shown is the distance J Q, and the distance Q should be equal to the rise of 157 146 CARPENTRY the rafter in one foot. Let the rise in this case be 9 inches. Then K N shows the bevel of the top of the stud. If the stud is a 2X4 Fig. 208. Studs Beveled to Fit Under Side ef Rafter Fig. 209. Cutting Studs for Partition Running Parallel to Rafter stick, the distance K Ris just 4 inches or one-third of the run of the rafter, and consequently the distance R N is just 3 inches, or one third of the rise of the rafter. In the case of the studs forming the partition P Q in Fig. 207, the bevel is found in the same way, the only difference being that the rafter now crosses the stud, as shown in Fig. 209, where A B C D is the stud and EFGH the rafter, both shown in plan. In the case of the partition H M, Fig. 207, we have to deal with a some- what more difficult problem because the rafter crosses the stud diagonally and the studs must be beveled diag- onally on top so that the bevel will run from corner to corner instead of straight across the stud from side to side. An enlarged plan of one stud with the rafter running across it is shown in Fig. 210. Let A B C D be the stud and EFGH the rafter; I J L K shows the rafter in elevation looking in the direction shown Fig. 210. Cutting Studs for Parti- tion Running Diagonally 158 CARPENTRY 147 z c \ by the arrow, and ^i-Bi C^D^ shows the stud as seen from this same direction. The edge D^ of the stud can not be seen from this side and is shown as a dotted Hne in the figure. The rafter runs across the stud, thus giving the bevel ylj 5^ Ci Di as shown in the figure. SPECIAL FRAMING We have, in the preceding pages, considered the framing which enters into a building of light construction, such as an ordinary dwell- ing house, but there are certain classes of structures which call for heavier framing, or framing of special character. Among these may be mentioned battered frames, or frames with inclined walls; trussed partitions; inclined and bowled floors; special forms of reinforced beams and girders ; the framing for balconies and galleries; tim- ber trusses, towers and spires, domes, pendentives and niches; and vaults and groins. These subjects will now be taken up and dis- cussed, and the methods employed in framing such structures will be explained. Battered Frames. Sometimes it is necessary to build a structure with the walls inclined inward, so that they approach each other at the top, and so that the top is smaller than the bottom. This is the case with the frames which support water tanks or windmills. An elevation of one side of a frame of this kind is shown in Fig. 211 with a plan in outline at C. It will be seen that the corner posts A A are inclined so as to approach each other at the top, and that they are not perpendicular to the sill at the bottom. This means that the foot of the post, where it is tenoned into the sill, must be cut on a bevel, and the bevel must be cut diagonally across the post, from corner to corner, since the post pitches diagonally toward the center, and is set so that its outside faces coincide approximately with the planes of the sides of the structure as indicated in the plan shown in Fig, 212. The girts B, Fig. 211, will also have Fig. 211. Battered Frame 159 148 CARPENTRY to have special bevels cut at their ends, where they are framed into the posts. After a corner post has been cut to the proper bevel to fit against the sill the section cut out at the foot will be diamond shaped, as shown at A B C D in Fig. 212, which shows a plan of one corner of the sill. It will be noticed that the faces A B and A D oi the post do not coincide with the edges of the sill A F and A G. If the struc- ture is merely a frame and is not to be covered over with the board- ing on the outside, it is not necessary that the outside faces of the post should coincide exactly with the planes of the sides of the structure, and in this case posts of square or rectangular section may be used, with no framing except the bevels and the mortises for A ^r J\J\ // K 1 Fig. 212. Battered Frame Detail Fig. 213. Method of Cutting Foot of Inclined Post by Steel Square the girts. If, however, the frame is to be covered in, the post must be backed in order that it may be prepared to receive the boarding. The backing consists in cutting the post to such a shape that when the bevel is cut at the foot, the section cut out wall be similar to that shown at E B C D in Fig. 212. The backed post must then be set on the sills so that the point E will come at the corner A. The face of the post E B will then coincide with the face of the sill A F. The post should be backed before the top bevel is cut because setting it back the distance A E may make a difference in the required length between bevels. ' If the post is of square section before backing it will have, after backing, a peculiar rhombus-shaped section, as is shown at A in Fig. 212. Here H I J K shows the original square section, and L I J K shows the section after backing. These sec- tions are taken square across the post perpendicular to the edges. 160 CARPENTRY 149 Fig. 213 shows how the foot cut for the inclined post may be obtained by using the steel square. In Fig. 211 it will be seen that the post A slopes toward the center in the elevation there shown, and it likewise slopes toward the center in the other elevations, either with the same pitch or with a different pitch. The result of the two slopes is to cause the post to slope diagonally. It is an easy matter to find the pitch in each elevation since it depends upon the size of the base and top, and the height between them. We then have the two pitches, the combination of which gives the true pitch diagonally; they can, however, be treated separately. The square may be applied to the post, as shown in Fig. 213, with the rise on the blade and the run on the tongue, and a line may be drawn along ! /^ yv I A yy B_ Fig. 214. Method of Finding Amount of Backing for Post the tongue. The post can then be turned over and the pitch shown in the other elevation may be laid off on the adjacent side in the same way, with the rise on the blade and the run on the tongue of the square. Thus a continuous line A B C D may be drawn around the post and it can be cut to this line. Fig. 214 shows how the amount of backing necessary in any particular case may be determined. Suppose that we have a case where the plan of the frame is not square, as shown in Fig. 211, but is rectangular, one side being much longer than the other. In this case the diagonal of the frame formed by the sills will not coincide with the diagonal section of the post. Fig. 214 shows at A a, plan of the post as it would appear if it were set up with one edge per- pendicular to the sill M, after the bottom bevel is cut. To cut the backing, lay the steel square along the side of the post parallel to 161 150 CARPENTRY M, and so as to coincide with the opposite corner 0. When the triangular piece R S is cut away, the backing is complete. At B is a plan where the post is set with corners T T, so as to coincide with the outside edges of the plate. To back the post in this position, place the square so as to coincide with the points T T, making the distance C T and C T proportional to the lengths of the sills M and N. In this case, the backing consists in cutting away the area STCT. Trussed Partitions. It is very often necessary to construct a partition in some story of a building above the first and in such a position that there can be no support beneath it such as another =15 — . — = — ■^ - Fig. 215. One Form of Trussed Partition partition. In this case the partition must be made self-supporting in some way. The usual method is to build what is known as a "trussed partition." This consists of a timber truss, light or heavy according as the distance to be spanned is small or large, which is built into the partition and covered over with lathing and plaster- ing or with sheathing. Figs. 215 and 216 show two forms of trussed partitions which are in common use. The one shown in Fig. 215 may be employed for a solid partition, or a partition with a door opening in the middle, while the one shown in Fig. 216 is applicable where the wall must be pierced by door openings in the sides. The truss must be so 162 CARPENTRY 151 designed that it will occupy as little space as possible in a lateral direction, so that the partition need not be abnormally thick. If possible, it is best to make the truss so that it will go into a 4-inch partition, but if necessary 5- or 6-inch studding may be used and the truss members may be increased in size accordingly. The faces of the truss members should be flush with the faces of the partition studding so as to receive lathing or sheathing. The size of the truss members depends entirely upon the weight which the partition is called upon to carry. Besides its own weight, a partition is often called upon to carry one end of a set of floor joists Fig. 216. Another Form of Trussed Partition and sometimes it supports columns which receive the whole weight of a story above. In any case, the pieces must be very strongly framed or spiked together, and sound material free from shakes and knot holes must be used. *• In Fig. 217 is shown another form of trussed partition spanning the space between two masonry walls. As will be seen this partition is constructed in a slightly different way from the others illustrated and described above. At the two sides of the opening, which is in this case in the center of the partition, are two uprights which are 163 152 CARPENTRY made considerably heavier and stronger than the ordinary studding of which the frame of the partition is composed. In the figure, the opening is marked A, and the uprights at the sides of the opening are marked B B. In the upright pieces shoulders are formed, as shown at C in the figure, and into the shoulders are fitted braces which go diagonally across the partition to the lower corners near the wall where they are notched into the lowest member of the trussed frame. These diagonal pieces are marked D in the figure, and the lowest member of the frame is marked E. The piece E Fig. 217. Trussed Partition Spanning Space between Two Brick Walls goes across from wall to wall and should run well into each wall as shown, so as to obtain a good bearing on the masonry and there should be a bearing plate or template of some kind under each end of it, as shown in the figure at the points F, to distribute the weight of the partition over a large surface of the masonry. For this pur- pose a thin iron plate will answer very well, or a large flat stone may be used. The piece E strengthens the floor construction and helps support the partition; the joists G rest on top of the piece E, or are notched over it, and the flooring H rests on these joists. 164 CARPENTRY 153 Above the door opening A there are two diagonal pieces I which come together at the top of the partition, forming a small truss over the opening and completing the trussing of the partition. The diagonal pieces / meet the uprights on each side of the door opening at the point where the horizontal piece M meets the uprights, and they should be notched into either one or the other or both. The topmost member of the trussed partition frame is marked in the figure, and on top of it rest the joists of the floor above, which either rest directly on it or are notched over it according to circum- stances. These joists are marked P in the figure. They support the flooring R of the floor above the partition. The main members of the partition frame are filled in with ordinary studding, 2X4 inches or 2X3 inches, spaced 1 foot or 16 inches apart. These studs are marked ^S in the figure. Inclined and Bowled Floors. In any large room which is to be used as a lecture hall the floor should not be perfectly level Fig. 218. Building an Inclined Floor throughout, but should be so constructed as to be higher at the back end of the room thar it is at the front. The fall of such a floor from back to front should be not more than f of an inch in 1 foot, and a fall of ^ an inch in 1 foot is much better. If the floor has a greater slope than this it becomes very noticeable when anyone attempts to walk over it. The simplest way to arrange for the slope is to construct what is known as an "inclined" floor, which rises steadily from front to back, so that a line drawn across it from side to side, parallel to the front or rear wall of the room, will be level from end to end. There are two methods of building an inclined floor, the difference between them being in the arrangement of the girders and floor joists. The two methods are shown in Figs. 218 and 219. 165 154 CARPENTRY Fig. 218 shows the arrangement when it is necessary to have the girders run from the back to the front of the room, parallel to the slope of the floor. In this case the girders A are set up on an incline and the joists B resting on top of them are level from end to end. Each line of joists is at a different elevation from the Hues of joists on each side of it. The floor laid on top of the joists will then have the required inclination. The slope of the girders must be the same as the slope required for the finished floor. Fig. 219 shows the arrangement when it is desired that the girder shall run from side to side of the room, at right angles in the direction of the slope of the floor. The joists A will then be parallel to the direction of the slope, and are inclined to the horizontal, while the girders B are level from end to end. Each line of girders Fig. 219. Building Inclined Floor When Girders Run at Right Angles of Slope' is at a different elevation from every other fine of girders, and these elevations must be so adjusted that the joists resting on top of the girders will slope steadily from end to end. When a simple inclined floor is employed, the seats must be arranged in straight rows, extending across the room from side to side, so that each fine of seats may be level from end to end. This arrangement is not always desirable, however, and it is often much better to have the seats arranged in rings facing the speaker's platform. In this case a bowled floor must be built. A bowled floor is so constructed that an arc, drawn on the floor from a center in the front of the room, on or near the speaker's platform, will be perfectly level throughout its length. This means that the floor must pitch upward in all directions from the speaker's platform, or, in other words, it must be bowled. There are two methods of constructing a floor of this kind. The simplest way is to build first an ordinary inclined floor, which slopes from the front to the back 166 CARPENTRY 155 of the room, and then to build up the bowled floor with furring pieces. This method should always be followed when it is necessary to keep the space beneath the lecture hall free from posts or columns. The second method is to arrange girders, as shown in the framing plan of a bowled floor in Fig. 220» These girders A are tangent to concentric circles which have their center at the speaker's platform, and each line of girders is at a different elevation. The elevations of the different lines of girders are so adjusted that the floor joists B Fig. 220. Framing Plan of a Bowled Floor Showing Arrangement of Girders which rest on them, will slope steadily upward as they recede from the platform. The girders may be supported on posts beneath the floor of the hall, and if the space under the floor is not to be used for another room, this is a very good method to employ. Immediately around the platform there will be a space D, the floor of which will be level, and the slope will start several feet away from the platform. If the floor is framed in this way it means that there will have to be a large number of posts in the space immediately beneath the floor, so many in fact as to make it practicafly impossible to make 167 156 CARPENTRY use of this space for another purpose. It would be necessary to put a post at each intersection of the girders which are arranged in concentric rings about the speaker's platform, so that the posts in Fig. 221. Framing Plan for Bowled Floor of Longer Type than Fig. 220 the space below would also appear in rings parallel to each other and only a comparatively small distance apart. It is not possible to do away with absolutely all of these posts except as explained 168 CARPENTRY 157 above, by building up on top of a plain inclined floor surface, but it is possible to do away with a large number of them if necessary, as will be explained. In Fig. 221 suppose that the space marked F is the flat space at the front of the room which we wish to floor with a bowled floor. We can place posts around this space under the floor as shown at the points marked A, and some more posts farther back from the front as shown at the points marked B. Between each set of points marked A and B we can run girders, resting at the front end on the post A, and at the other end on the post B. Other girders can be run from the posts A to the wall as, for example, the girder A C; and others again may be run from the points B to the walls as, for example, the girders B D. These girders can all be inclined so as to slope evenly toward the front from all directions, so that points on all the girders at a given distance from the center of the room at the front wall will be at the same level. The framing formed by the girders may now be filled in by joisting E, and the flooring laid on top of the joisting so as to form a solid floor surface on which the seats may be placed. The floor surface thus formed will slope towards a point in the center of the front wall and all the seats will face the platform in concentric rings, each ring being "level from end to end. In the space beneath the floor there will be only a comparatively small number of posts, arranged in such a way that the space can be utilized for rooms if desired. All the posts marked B will be in a straight line and can be covered by a partition, so that only the posts marked A will be troublesome, and these are clustered together at the front where they can be easily concealed. The room shown in Fig. 221 has been purposely made somewhat different from the room shown in Fig. 220. In Fig. 221 the room shown is longer than it is wide while in Fig. 220 the room shown is wider than it is long. This gives rise to a slight difference in the appearance of the framing, but the principle is the same in both cases, and the two methods of procedure apply equally well to both rooms. Heavy Beams and Girders. For ordinary framed buildings there will be no difficulty in obtaining timbers large enough for every purpose, but in large structures, or in any building where heavy loads must be carried, it is often impossible to get a single piece which is strong enough to do the work. In this case it becomes 160 158 CARPENTRY Fig. 222. Cutting Girders to Fit Steel Beams necessary to use either a steel beam or a trussed girder of wood, or to build up a compound wood girder out of a number of single pieces, fastened together in such a way that they will act like a single piece. Steel beams are very often employed for girders when a single timber will not suffice, and although they are expensive, the saving in labor helps to offset the extra cost of the material. Wherever wood joists or girders come in contact with a steel beam they must bt cut to fit against it. The steel shape most commonly employed is the I-beam and the wood members must be cut at the ends so as to fit between its flanges. This is shown in Fig. 222. The joist B is sup- ported on the lower flange of the I-beam C and the strap A prevents it from falling away from the steel member. The strap is bolted or spiked to the wood beam and is bent over the top flange of the steel beam as shown. 'If two wood beams frame into the steel beam opposite each other, a straight strap may be used passing over the top of the steel beam and fastened to both the wood beams, thus holding them together. If a better support is desired for the end of the wood beam, an angle may be riveted to the web of the steel I-beam, as shown in Fig. 223, and the end of the wood joist may be sup- ported on the angle. This is an expensive detail, however, and it is seldom necessary. If a timber is not strong enough to carry its load, and if it is not desirable to replace it with a steel beam, it may be strengthened by trussing. There are two methods of trussing beams: by the addition of compression members above the beam, and by the addition of ten- sion members below it. The first method should be employed when- ever, for any reason, it is desired that there be no projection below the bottom of the beam itself. The second method is the one most Fig. 223. Fitting Girders to Steel Beams by Means of Angles 170 CARPENTRY 159 commonly used, especially in warehouses, stables, and other build- ings where the appearance is not an important consideration. In Fig. 224 is shown a beam which is trussed by the first method with compression pieces A above the beam. All the parts are of Fig. 224. Trussing a Girder by Use of Compression Members wood excepting the rods B, which may be of wrought iron or steel. The beam itself is best made in two parts E E placed side by side, as shown in the section at A. This section is taken on the line C D. The depth of the girder may be varied to suit the conditions of each \ T ^ \ r^ SECT/ON £-F 'r Fig. 225. Trussing Girder by Use of Tension Member — King-Post Trussed Beam case. In general the deeper it is made the stronger it becomes, pro- vided that the joists are made sufficiently strong. Usually girders of this kind are made shallow enough so that the compression mem- ber will be contained in the thickness of the floor and will not pro- SECTION EF Fig. 226. Trussed Girder with Two Struts — Queen-Post Trussed Beam ject above it. A slight projection below the ceiling is not a serious disadvantage. The floor joists F may be supported on the pieces E, as shown at A. In Figs. 225 and 226 are shown examples of girders which are trussed by the second method with tension rods D below beam. These rods are of wrought iron or steel, and the struts A are of cast 171 160 CARPENTRY iron. The struts may be made of wood if they are short, or if the loads to be carried are not heavy. Sometimes the girders are made very shallow and the struts A are then merely wood blocks placed between the beams C and the rod D to keep them apart. The girder shown in Fig. 225 is known as a king-post trussed beam, while the one shown in Fig. 226, with two struts instead of one, is known as a queen-post trussed beam. The beam itself, C, may be made in two or three pieces side by side with the rods and the struts fitting in between them, or it may be a single piece, and the rods may be made in pairs, passing one on each side of the beam. The struts bear against the bottom of the beam, being fastened to it by bolts or spikes, as shown in the illustrations, so that they will not slip sidewise. It sometimes happens that a heavy girder is required in a situa- tion where trussing can not be resorted to, and where steel beams Fig. 227. Construction of Compound Beam Fig. 228. Flitch-Plate Girder can not be readily obtained. In this case the only resource is to build up a compound beam from two or more single pieces. A girder of this kind can be constructed without much difficulty, and can be so put together as to be able to carry from eighty to ninety per cent of the load which a solid piece of the same dimensions will bear. There are many ways of combining the single timbers to form compound beams, some of the most common of which will be described. The most simple combination is that shown in Fig. 227. The two single timbers are bolted together side by side, with sometimes a small space between them. The bolts should be spaced about 2 feet apart and staggered as shown, so that two will not come side by side. Usually bolts three-quarters of an inch in diameter are used. In Fig. 228 is shown a modification of this girder known as a "flitch-plate" girder. It has a plate of wrought iron or steel, inserted between the two timbers, and the whole is held firmly together by bolts. The size of the plate should be in proportion to the size of the 172 CARPENTRY 161 Action of Compound Girder Under Tension timbers, so as to make the most economical combination. In general the thickness of the iron plate should be about one-twelfth of the combined thickness of the tim- bers. If we have two pieces of timber out of which we wish to make a compound girder, it is Fig. 229. always possible to get a stronger combination by placing them one on top of the other, than by placing them side by side. This is because the strength of a beam varies as the square of its depth, but only directly as its width. For this reason most compound girders are composed of single pieces placed one above the other. The tendency is for each piece to bend independ- ently, and for the two parts to slide by each other, as shown in Fig. 229. This tendency must be overcome and the parts so fastened together that they will act like a single piece. There are several methods in common use by which this object is accomplished. Fig. 230 shows the most common method of building up a com- pound girder. The timbers are placed together, as shown, and narrow strips of wood, are nailed firmly to both parts. The strips are placed close against each other and have a slope of about forty- five degrees, sloping in opposite directions, however, on opposite sides of the girder. It has been claimed that a built-up girder of this kind has strength ninety-five per cent as great as the strength of a solid piece of the same size but it is very doubtful whether this is true in most cases. Actual tests seem to indicate that such girders have an efficiency of only about seventy-five per cent. They usually fail by the splitting of the side strips, or the pulling out and bending of the nails, but seldom by the breaking of the main pieces. It is, therefore, essential that the strips should be very securely nailed to each of the parts which make up the girder, and that they should also be carefully selected so that only those pieces which are free from all defects may be used. These girders are liable to considerable deflection, and should not be used in situations where such deflection would be harmful. Fig. 230. Method of Building up Com- pound Girder 173 162 CARPENTRY In Fig. 231 is shown another form of girder with the parts notched, as shown, so as to lock together. This prevents them from shpping by each other. BoHs are employed to hold the parts together, so that the surfaces will always be in close contact. While this form of girder is very easily constructed, it has many disadvantages. A great deal of timber is wasted in cutting out the notches, as these must be deep enough to prevent crushing of the wood at the bearing surfaces, and thus the full strength of the timbers is not utilized. Moreover, it is apt to deflect a good deal, and its efficiency is not so great as that of other forms. On the whole it is inferior to the form previously described. The compound beam which is almost universally considered the best is that shown in Fig. 232. This is known as a keyed beam, its characteristic feature being the use of keys to keep the parts from sliding on each other. The strength of a keyed beam has been found by actual experiment to be nearly ninety-five per cent of the !^ Fig. 231. Compound Girder with Notched Surfaces -&^ ' o- Fig. 232. Example of Keyed Beam strength of the solid timber, while the deflection when oak keys were used was only about one-quarter more than the deflection of the solid beam. By using keys of cast iron instead of wood this excess of deflection in the built-up girder can be reduced to a very small percentage. The keys should be made in two parts, each shaped like a wedge, as explained in connection with the keys for tension splices, and should be driven from opposite sides into the holes made to receive them, so as to fit tightly. They should be spaced from 8 to 16 inches apart, center to center, according to the size of the timbers, and should be spaced more closely near the ends of the beam than near the middle. In the center of the span there should be left a space of 4 or 5 feet without any keys. Balconies and Galleries. In churches and lecture halls it is almost always customary to have one or more balconies or galleries, extending sometimes around three sides of the main auditorium. 174 CARPENTRY lb3 but more often in the rear of the room only. These galleries are supported by the wall at the back and by posts or columns in front, and the framing for them is usually a simple matter. Fig. 233 shows a sectional view of a gallery frame, as they are commonly constructed. There is a girder A in front, which rests on top of the columns T, and supports the lower ends of the joists B, forming the gallery floor. The size of these pieces will depend upon the dimensions of the gallery, the spacing of the columns which support the girders in front, and various other considerations. Usually posts 2X 10 or 3X 12, and girders 8X 10 or lOX 12 will be found to be sufficiently strong. The joists should be spaced from 14 to 20 inches, center to center. Very often cast-iron columns are Fig. 233. Sectional View of Gallery Framing employed to support the girders. At the top, where the joists rest on the wall, they should be cut, as shown in the figure, so that they may have a horizontal bearing on the masonry, and at least every second joist must be securely anchored to the wall, as is the one shown. Usually galleries are made with straight fronts, but if it is desired that the seats should be arranged in concentric rings, all facing the speaker, the joists may be placed so as to radiate from the center from which the seats are to be laid out. The seats are arranged in steps, one above the other, and the framing for the steps must be built up on top of the joists, as shown in the figure. Horizontal pieces C, usually 2X4 or 3X4 in size, are nailed to the joists at one end, and at the other end they are supported by upright pieces D. The uprights are either 2X4 pieces resting on top of the joists, or strips of board, 1 inch to 1| inches thick, which 175 164 CARPENTRY are nailed to the sides of the joists and to the sides of the horizontal pieces. Both methods are shown in the figure. If boards are used, they should be placed on both sides of the joists. Great care should be taken to see that the horizontal pieces are truly horizontal. Balconies and galleries almost always project a considerable distance beyond the line of columns which support the lower ends of the joists. This projection varies from 3 feet to 10 or 12 feet. If the overhang is not more than 5 feet, it can be supported by extend- ing the joists beyond the girder, as is shown in Fig. 233. A strip of board E, about 1| inches thick, is nailed to the side of the joist, and Fig. 234. Section of Gallery Framing with Overhanging Portion a furring piece F is nailed on top of the joist at its lower end to make it horizontal. The railing at the front of the gallery should be about 2 feet high, and may be framed with 2X4 posts G having a cap H of the same size on top. If the overhang of a gallery is more than 5 feet it must usually be supported by a brace, as shown in Fig. 234. The brace A may be nailed to the post B and to the overhanging joist C, or framed into these pieces. If the construction is very light, the braces may consist of strips of board nailed to the sides of the joists, but in heavy work they must be timbers of a good size, well framed into both the post and the joists. The braces can only be placed at 176 CARPENTRY 165 points where there are posts, and to support the ends of the joists which come between the posts there must be a girder D, running along the front of the gallery and supported by the braced canti- levers at the points where posts are placed. The forms of balconies described above are all of such a sort as to require the presence of posts in the main floor below the bal- cony to support it, but it very often happens that such posts are very undesirable and must be avoided if it is possible to do so. In this case the balcony must be supported from above in some way, and the method most commonly employed is to hang the outer end of the main timbers from the ceiling of the main hall or room of which the balcony forms a part. Hangers made of round or square iron .y-»j.^->-.,.^w J.U.'J-.i O ^r' J / Fig, 235. Details of Balcony Construction Showing Hanger Attached to Roof or steel rods are used, and these are fastened at the upper end to some member of the floor construction of the floor above, or to some member of the roof construction in case there is no floor above. The most common arrangement is to fasten the upper end of the hanger to the lower chord of the roof truss. Fig. 235 shows a balcony constructed as described above. One side is supported by the masonry wall of the building, marked F in the figure, and the other side is hung from the roof or ceiling by means of the hangers marked E. At L is shown a section through the balcony from the wall to the inside edge, while at M is shown a view looking at the edge of the balcony from the inside of the hall or room in which the balcony is situated. The principal support- ing members of the construction are the pieces marked A which run well into the wall so as to obtain a sufficient support at this end. 177 166 CARPENTRY These pieces are made double or in pairs, as shown in the end view M, and are separated a Httle so as to allow the hanger rod to pass between the two pieces as shown. E is the hanger, which is a round or a square rod about 1 inch in diameter. The pieces A should continue a short distance beyond the point where the rod passes between them, and the ends may be cut to any shape desired in order to give them a pleasing appearance, as shown. They should be spaced 7 or 8 feet apart and on top of them may be placed ordinary joists of small size, marked B in the figure, which are spaced about 12 or 14 inches apart. On top of the joists B is laid a rough floor, marked C in the figure, and above this again is laid the finished floor- ing of the balcony D. A joist should be placed on each side of the hanger E, as shown at N and 0, and against the joist marked should be nailed the finished fascia piece G. This finished piece G should run up past the under boarding C and stop against the under side of the finished top flooring of the balcony. There should be a bed molding H at the juncture between the piece G and the top flooring D, so as to cover and conceal the joint. The hanger E should pass between the pieces A, and should be threaded at the bottom so as to receive a nut J by which it may be tightened up. There should be a washer I consisting of a square plate of iron, between the nut J and the wood of the pieces A, so the wood will not be crushed and so that the nut will not sink into the wood. The under side of the balcony seen from the floor of the main hall may be treated in any one of a variety of ways. The joists may be furred and lathed and plastered on the under side so that a plas- tered surface will be presented, or the under side of the joists may be covered with sheathing, V-jointed or beaded, or the joists may be more carefully chosen and left exposed to view from below. TIMBER TRUSSES In the discussion of roofs and roof framing which has already been given here, only those roofs have been considered which were of so short a span that they could easily be covered with a frame- work of ordinary rafters, spaced from 1 to 2 feet apart, between centers, but it is very often necessary to build roofs of larger span, for which ordinary rafters, even if supported by dwarf walls and 178 CARPENTRY 167 collar beams, are not sufBciently strong. In this case a different method of framing must be employed. Instead of a number of rafters- spaced fairly close together, and all of equal strength, we will have a few heavy "trusses," placed at intervals of 10 or more feet, and spanning the entire distance between the two side walls. On top of the trusses are laid "purlins," running parallel to the walls, which in their turn support the common rafters, running perpendicular to the side walls, as in the case of simple rafters in an ordinary roof. There may be one or more purlins in r — ... — — - ---- --- ... ■-- -.. .... — .... .... ; i =z zr — J = z= = =: = = = D = - — = = — = = = — = = 1 ] = :^ = =J = = = = = = — B ■C A- 1 — ^ =1 = = zz = = = = 1 1^ — 1 , Fig. 236. Roof Plan Showing Use of Heavy Trussed Purlins and Common Rafters each slope of the roof, depending upon the size of the span, since the purlins must be spaced near enough together so that a small rafter can span the distance between them. Usually there will be a pur- Un at each joint of the truss and the joints will be determined by the safe span for the rafters. This arrangement is shown in plan in Fig. 236, in which A are trusses, B are the purlins, C are the common rafters, and D is the ridge. There are many different kinds of trusses in common use for various kinds of buildings, which differ from each other chiefly in 179 168 CARPENTRY the arrangement of the tension and compression pieces of which every truss is built up. Some trusses are built entirely of timber, while in others timber is employed only for the compression pieces, and wrought iron and steel for the tension pieces. King=Post Truss. Fig. 237 shows what is known as a king-post truss. Its distinguishing feature is the member A called a king- post, B are the purlins, and E are the rafters resting on them. As will be seen by a study of the figure, the members of the truss are so arranged as to divide it up into a series of triangles, or rather into a series of triangular open spaces, bounded by the various members of the framework. This is an essential characteristic of a good and efficient truss. Such a framework may fail by overloading in such Fig. 237. Section Showing Design of King-Post Truss a way as to be crushed or broken, but it can not be distorted, that is, none of the triangular spaces can change their shape without some membei' of the truss being either lengthened or shortened, which means that some member of the framework must fail by either tension or compression before the truss can be distorted, or can fail to carry its load by reason of the failure of the joints. This prin- ciple does not hold true for a framework composed of spaces in the form of rectangles, of which the members of the framework form the sides, because it is possible for a rectangular framework to become distorted without any side being either lengthened or shortened, by the simple failure of some of the joints and the movement of the members around the joints. For this reason the first thing to con- 180 CARPENTRY 169 sider in designing a truss is the arrangement of the members and the position of the joints so all of the open spaces will be in the form of triangles. In Fig. 237 are shown two different methods of placing the pur- lins. As will be readily seen, some of them are set so that their Fig. 238. Section Showing Design of King-Post Truss for Wide Span longer dimension in cross section is vertical, while others are set so that their longer dimension is at right angles to the rafters. Both of these methods are commonly employed. The tension members C are merely for the support of the lower chord or tie-beam D. Fig. 239. Section Showing Design of Trussed Roof Using Iron Castings at Joints Fig. 238 shows a truss of the same general form as the one shown in Fig. 237, but of larger span. This truss is of such a span and has its joints and purlins arranged in such a way that it is similar to the trusses shown in plan in Fig. 236 and there marked A. In this truss 181 170 CARPENTRY also the vertical members are not iron rods, as in Fig. 237, but are composed of timber. The stresses in these members are, however, still tension stresses just the same as in Fig. 237, and for this reason it is a common practice to fasten them to the chords of the truss by means of iron straps, as shown at the points marked A in Fig. 238. In other respects this truss is constructed in a manner similar to that in which the truss shown in Fig. 237 is built. Fig. 239 shows a truss with the diagonal members running in a direction opposite to that in which run the diagonal members in the two trusses previously shown. This figure also illustrates the practice of placing an iron casting at each joint of the truss to receive the members which come together at that joint. This Fig. 240. Section Showing Design of Queen-Post Truss arrangement is, however, an expensive one on account of the cast- ings, and it is doubtful if the advantage gained by the use of them is sufficient to warrant the additional cost. Usually the castings can not be kept in stock and must be made to order for each truss. Queen=Post Truss. Fig. 240 shows a modification of the king- post truss, which is called the queen-post truss. Here there are two queen-posts instead of the single king-post. The queen-post truss is somewhat more popular in building work than is the king-post truss, but both are frequently employed in halls, warehouses, and stables, where an ornamental truss is not required, and also in churches and audience rooms, where they are to be concealed by other finish. Fig. 240 also shows how a floor or ceiling may be 182 CARPENTRY 171 supported on the lower chord or tie-beam of the truss. The joists C are hung from the chord by means of stirrup irons or patent hangers. This arrangement makes the tie-beam act as a beam as well as a tie and in this case it must be made sufficiently strong to carry the load from the joists without sagging. The queen-post truss, as will be seen, is not entirely composed of triangles, the center panel being in the form of a rectangle. In most cases this is not a serious disadvantage, since, when the truss is uniformly loaded, as it would be if it were an ordinary roof truss, there is no tendency to distort the center panel. It is almost always Fig. 241. Section Showing Design of a Fink Truss better, however, to introduce an additional diagonal member into this panel so as to divide it into two triangles. This obviates any danger of distortion of this panel. Fink Truss. In Fig. 241 is shown a Fink truss, which is a very popular form, especially for trusses built of steel. It has neither king-post nor queen-posts, and the tie-beam A is of iron or steel instead of timber. This is a simple and cheap form of truss for any situation where there is no floor or ceiling to be carried by the lower chord. The struts B may be of wood or of cast iron. It will be seen that the truss consists essentially of two trussed rafters set up against each other, with a tie-rod A to take up the horizontal thrust. Open Timber Trusses. Besides the forms of trusses described above, there are other forms which are used in churches and chapels. 183 172 CARPENTRY as well as in halls where open timber work is required, and where the trusses will not be concealed by other finish, but will be made ornamental in themselves. Among these the most com- mon forms are the so-called scissors truss and the ham- mer beam truss. Scissors Truss. The scissors truss is shown in Fig. 242. It has no tie- beam and, therefore, it will exert considerable thrust on the walls of the building, which thrust must be taken care of by buttresses built on the outside of the walls. This is perhaps the most simple form of truss which can be used when an open timber truss is required. Fig. 242. Design of a Scissors Truss Fig. 243. Hammer Beam Truss Used Particularly for Churches All the parts are of wood. If desired, an iron tie rod may be inserted between the two wall bearings of the truss, so as to eUmi- 184 CARPENTRY 173 nate the thrust on the walls, and this rod need not detract seriously from the appearance of the open timber work. Hammer Beam Truss. A very popular form of truss for use in churches is the hammer beam truss mentioned above. This is shown in Fig. 243. On the left is shown the framework for the truss, while on the right is shown the w^ay in which it may be finished. Its characteristic feature is the hammer beam A. The sizes of the pieces can only be determined by calculation or experience, and depend entirely upon the span of the truss and the loads to be carried, which are different for different locations. It is common practice to insert a tie rod between the points B and C to take up the thrust which would otherwise come on the walls. All parts of the framework must be securely bolted or spiked together so as to give a strong, rigid foundation for the decoration, which, should be regarded merely as decoration and should not be con- sidered as strengthening the truss in any way. Truss Details. There are several methods of supporting the purhns on wood trusses, but the method illustrated in Fig. 244 is one of the best as well as the most frequently employed. A block of wood A is set up against the lower side of the purlin, and prevents it from turn- ing about the corner B, which it has a tendency to do. The block is set into the chord of the truss to a depth suffi- cient to keep the purlin from sliding downward as it receives the weight from the rafters E. This figure also shows the most simple method of fram- ing a strut into the chord of a truss. The strut C is set into the chord D far enough to hold the strut in place. If it is perpendicular to the chord, it need not be so set into it, if the pieces are well nailed together, because in this case there is no tendency for the strut to slide along the chord. Care should be taken not to weaken the chord too much in cutting these mortises. In Fig. 245 are shown the most common methods of forming the joint between the top chord and the tie-beam of a truss. The Fig. 244. Method of Supporting the Purhns on Wood Trusses 185 174 CARPENTRY connection shown at A depends upon the bolts for its strength, while that shown at B depends upon the wrought-iron straps E, which are bent so as to engage notches cut in the tie-beam F. The piece C is very often added beneath the tie-beam, at the bearing, to strengthen it at this point, where the beam is subject to consider- able bending stress. The block D is merely for filling and to pro- tect the bolts where they pass between the chord and the tie-beam. It may be omitted in many cases. The plate G is placed between the nuts or bolt heads and the wood to prevent the crushing of the latter. Washers should be used with all bolts for this purpose. Fig. 246 shows how the joint at the center of the tie-beam of a king-post truss, or any joint between two struts, may be formed. Fig. 245. Two Methods of Forming a Joint between Both Chord and Tie-Beam of a Truss The tie-beam is shown at A, and B are the struts. The blocks C, set between the struts, receive the thrust from them. They should be notched into the tie-beam A deep enough to take care of any inequality between the thrusts from the two struts, which have a tendency to balance each other. The block is often made of cast iron. It may be omitted altogether, in which case the struts will come close together and bear against each other. The rod D is the king-post which supports the tie-beam A at this point. It is often made of wood and sometimes the struts B are framed into it instead of being framed into the tie-beam A. 186 HALL AND PARTIALLY ENCLOSED STAIRCASE IN LONG HALL, GREYROCKS, ROCKPORT, MASS. Frank Chouteau Brown, Architect, Boston, Mass. For Plans and Exteriors, See Vol. I, Pages 273, 283, and 399. HALL AND STAIRCASE IN HOUSE AT WOLLASTON, MASS. Frank Chouteau Brown, Architect CARPENTRY 175 Fig. 247 shows a form of connection for the peak of a truss, where the two top chords or principal rafters come together. The plate A acts as a tie to keep these members in place, as does the bent plate B, also. The plate B, moreover, prevents the crushing of the timber by the nut of the king-post tie rod. The purlin C supports the rafters and is hollowed out at the bottom to admit the nut D. The two principal rafters bear against each other and must be cut so that the bearing area between them will be sufficient to prevent the crushing of the timber. In light trusses the king-post E is often made of wood and is carried up between the principal rafters so that these members bear against it on each side. If this construction is adopted it must be remembered that the post is a tension member, and is held up by the principal rafters, and these pieces must be mortised into it in such a way as to accomplish this result. Fig. 246. Method of Forming Joint at Center of Tie-Beam in King-Post Truss Fig. 247. Construction of the Peak Connection of a Truss There are a great many different ways of arranging the details for wood trusses, each case usuall}^ requiring details peculiar to itself and unlike those for any other case. There are, therefore, no hard and fast rules which can be laid down to govern the design of these connections. A perfect understanding of the action of each piece and its relation to all of the other pieces is necessary in order to insure an economical and appropriate design. The aim should always be to arrange the details so that there will be as little cutting of the pieces as possible, and so that the stresses may pass from one piece to another without overstraining any part of the truss. TOWERS AND STEEPLES Towers are a very common feature in building construction, ranging in size from the small cupola used on barns to the high tapering spire which is the distinguishing mark of churches. 187 176 CARPENTRY They have roofs of various shapes, some in the form of pyramids, with four, eight, or twelve sides, some of conical form, and others bell-shaped or having a slightly concave surface. The construction of all these forms of towers is much the same, consisting of an arrangement of posts and braces, which becomes more elaborate as the tower or steeple becomes larger. The bracing is the most important consideration, because the towers will be exposed to the full force of the wind and must be designed to resist great strain. Cupola. Fig. 248 shows a section through the frame of a simple cupola. It has posts A at each corner, which rest at the bottom on the sills B. The sills are supported on extra heavy collar beams C, which are very securely spiked to the rafters of the main roof M. The corner posts ex- tend clear up to the main plate D, which supports the rafters E of the cupola. There are hip rafters at the corner of the roof, which bear at the top against a piece F placed in the center of the roof. This scant- ling extends above the roof surface far enough to receive some kind of metal finial which forms the finish at the extreme top of the cupola; and at the bottom it is firmly fastened to the tie G, which is cut in between the plates. The braces H stiffen the frames against the wind. Girts I are cut in between the corner posts and form the top and bottom of the slat frame opening R, besides tying the posts together. The sides of the opening for the slat frame are formed by the vertical studs A'. The rafters of the main roof M are placed close up against the corner posts on the outside, and the posts may be spiked to them. The pieces are of Fig. 248. Section Through Frame of a Simple Cupola 188 CARPENTRY 177 plank 2 inches thick, and are simply furring pieces placed at intervals of 1 1 to 2 feet all around the cupola to give the desired shape to the bottom part. The size of the pieces will depend on the size of the cupola. The posts may be 4X4 inches or 6X6 inches, and the braces, girts, and intermediate studding may be 3X4 inches or 4X6 inches. Miscellaneous Towers. Other towers are framed in a manner similar to that described for a cupola. There is always a base or drum, with posts at the corners and with the walls filled in with studding, which supports a plate at the top. The rafters forming the tower roof rest at the foot on this plate, and at the top they bear against a piece of scantling which is carried down into the body of the tower for a considerable distance and is there fastened to a tie passing between rafters on opposite sides. This is shown in Fig. 249. The tie A is securely nailed to the rafters at each end, and to a post in the mid- dle. The post is cut so as to have as many faces as the roof has sides, four for a square hip roof, eight for an octagon roof, and so on. Each face receives one of the hip rafters and the intermediate rafters are framed in between them. If the roof is conical or bell shape, as shown in the figure, the post at the top may be cylindrical in form. Although the roof shown is bell shaped the rafters are not cut to fit the curve. They are made straight and are filled out by furring pieces B. Pieces of plank C are cut in between the^furring pieces, as shown, so as to give a nailing for the boarding, and they are cut to the shape of segments of circles, so as to form complete circles around the tower when they have been put in place. If a tower of this shape is to be built, having a number of faces and hips, the curve of the hip rafters will not be the same as the curve shown by a section through one of the faces of the tower. In order to find the true curve for the hip rafter, the same method is followed Fig. 249. Section Showing Frame Construction of Small Tower 189 178 CARPENTRY as was explained for hip rafters in an ogee roof over a bay window, using the principle that any line drawn in the roof surface parallel to the plate is horizontal throughout its length. By this means any number of points in the curve of the hip rafter may be obtained and the curve for the hip may be drawn through them. Thus a pattern for the hip rafter may be obtained. Church Spire. Fig. 250 shows the method of framing a church spire, or other high tapering tower. The base of the drum N is square and is supported by the posts A, one at each corner, which rest on the sills B. The sills are supported by the roof trusses of the main roof. The corner posts extend the full height of the drum and are strongly braced in all four faces, with intermediate ver- tical studding C between them to form the framework for these faces. The spire itself may rest on top of this square drum or there may be another eight- or twelve-sided drum constructed on the top of the first drum, on which the spire may rest. This depends upon the design of the spire. The hip rafters D do not rest directly on top of the drum, how- ever, as this arrangement would not give sufficient anchorage for the spire. They are made so as to pass close inside the plate E at the top of the drum and are securely bolted to this plate with strong bolts. This is shown at L, which is a plan of the top of the drum, showing the hip rafters in place. The plate is shown at E, and the hip rafters at D. Tlie rafters extend down into the body of the drum as far as the girts H (shown in the elevation) to Fig. 250. Section Showing Contruction of Church Spire 190 CARPENTRY 179 which they are again securely spiked or bolted, being cut out at the foot so as to fit against the girt. In this way a strong anchorage for the spire is obtained. Horizontal pieces I are cut in between the hip rafters at inter- vals throughout the height of the spire, braces K, halved together at the center where they cross each other, are firmly nailed to the rafters at each end. These braces are needed only in lofty spires, which are likely to be exposed to high winds. At the top the hip rafters bear against a post M, the same as in the other towers. If a conical spire is called for in the design, the horizontal pieces I must be cut Fig. 251. Section Showing Frame Construction of a Dome to the shape of segments of circles, and in this case the rafters are no longer hip rafters. The horizontal pieces / will receive the boarding, which will form a smooth conical surface. The spire above the drum is usually framed on the ground before being raised to its final position. It then may be raised part way and supported by temporary staging while the top is finished and painted, after which it may be placed in position on the top of the drum. Domes. Timber domes have been built over many famous buildings, among which may be mentioned St. Paul's Cathedral at London, and the Hotel Des Invalides at Paris. While these structures 191 180 CARPENTRY are domical in shape they are not, strictly speaking, domes, because they do not depend for support upon the same principle which is implied in the construction of a dome. They are, correctly speak- ing, arrangements of trusses of such a shape as to give the required domical form to the exterior of the roof. Fig. 251 shows such a truss supported at either end on a masonry wall. Fig. 252, which is a plan of the framing of this roof, shows Fig. 252. Plan of Framing for Dome Shown in Section in Fig. 251 how the sections or bents may be arranged. There are two complete bents, A B and C D, like the one shown in the elevation. Fig. 251, which intersect each other at the center. A king-post A in the elevation is common to both bents and the tie-beams B are halved together where they cross. These two bents divide the roof surface into four quarters, which are filled in by shorter ribs, as indicated in the plan, Fig. 252. The posts C, in Fig. 251, carry all the weight of the roof to the walls and are braced by means of the pieces D. 192 CARPENTRY 181 The rounded shape is given to the exterior and interior of the bent by pieces of plank bent into position as shown. The whole is covered with boarding which is cut to a special shape so that it can be bent into place. The methods of applying the boarding to domical roofs will be explained in connection with other rough boarding. The arrangement of trusses described above is suitable for a plain domical roof without a lantern or cupola on top, but very Fig. 253. Framing Plan for Dome Having Cupola frequently this feature is present in the design, and the roof must be framed to allow for it. There are several different ways of arrang- ing the trusses so as to leave an opening in the center of the roof for the lantern. Fig. 253 shows a very good arrangement. Four trusses A span the entire distance between the walls, and are placed as shown in the figure, so as to leave the opening B in the center. Four half trusses C are inserted between them, as, shown, and eight shorter ribs D are employed to fill in the rest of the space. 193 182 CARPENTRY Fig. 254 shows another arrangement, providing for a lantern at the center. There are a number of ribs A, twelve in number, in the figure, all radiating from the center where there is a circular opening for the lantern or cupola. In Fig. 255 is shown a section through a domical roof framed in this way, showing an elevation of one of the ribs. The rib is so constructed as to be entirely contained in the restricted space between the lines of the exterior and interior of the roof. Fig. 254. Framing Plan for Dome Having Lantern at Center Pendentives. In the preceding paragraphs we have considered the subject of domical roofs covering buildings of circular plan, which is the simplest possible case, but unfortunately not the most usual one. It very often happens that a domical roof must be erected over a building which is square or rectangular in plan, in which case a new and difficult problem must be considered, namely, that of the pendentives. A horizontal section taken through a dome must in every case show a circle or possibly an ellipse. If, then, we con- 194 CARPENTRY 183 sider the horizontal section cut from a domical roof by the plane of the top of the wall, it must usually be a circle and can not exactly coincide with the section cut from the wall of the building by the Fig. 255. Section through Domical Roof Showing One of the Ribs same plane, unless the building is circular in plan. This is shown in Fig. 256 in which A B C D represents the section cut from the wall of the building by a horizontal plane, and the circle E F G H represents the section which would be cut from a domical roof cover- ing the building if the framing for the dome were carried down to meet this plane all the way around. In order to cover every part of the building, the dome must be large enough to include the corners, and if made sufSciently large for this it must overhang the side walls of the building, by an amount A E B on each side, if the framing is carried down to the same horizontal plane all the way around. Horizontal sec- tions taken through the dome at intervals throughout its height, however, show smaller and smaller circles as they are taken nearer and Fig. 256. Diagram Showing Relation of Dome Roof of Square Building 195 184 CARPENTRY Fig. 257. Perspective Outline of Pendentives nearer to the top of the dome. Some one of these sections will cut out from the dome a circle which will appear in plan as though it were inscribed in the square formed by the walls of the building. Such a circle is shownat / Jii"!/ in Fig. 256. A dome built up with this circle as a base would not cover the corners of the build- ing, so that the triangular spaces like AIL would be kept open. These triangular spaces, or rather the coverings over them, are called the pendentives. Fig. 257 shows in perspective the outline of four pendentives E D H, H C G, etc. We have seen that a dome built up on the circumscribed circle as a base is too large, while a dome built up on the inscribed circle is too small and mil not completely cover the building. To over- come this difficulty it is customary to erect a dome on the smaller or inscribed circle, as a base, and to extend the ribs so as to fill up the corners and form a framework for the pendentives. This is shown in Fig. 258 which is a plan of the framework for a domical roof. The ribs will be of different lengths and will in- tersect the inside face of the wall at different heights, be- cause as they are extended outward they must also be ex- tended downward. Each one will be curved if the dome is spherical, and straight if the dome is conical. The upper ends of the ribs bear against the curb A, leaving a circular open- ing for a lantern or cupola. The lower ends may be supported on a masonry wall, or may rest on curved wood plates, as shown in Fig. 259. This is an elevation of a conical dome, and shows the straight ribs A. Fig. 258. Dome Framing with Extended Rafters 196 CARPENTRY 185 Fig. 259. Diagram Showing Construction for Apex of Conical Dome Fig. 260 shows an elevation of a spherical dome which has curved ribs A, as shown. Each of these ribs must be bent or shaped to the segment of a circle, in order that the edges may lie in a spherical surface. If the design calls for a domical ceiling and the exterior may be of some other form, then only the inside edges of the ribs need be dressed to corre- spond with a spherical or conical surface, in order that they may receive the lathing or furring, and the outside may be left rough so that a false roof of any desired shape may be built. If the exterior must be of domical form, while on the interior there is a suspended or false ceiling of some kind, then only the outside edges of the ribs must lie in the conical or spherical surface, so as to receive the roof board- ing, while the inside edges may be left rough or shaped to any other form. If both the exterior and interior must be domical, then both the inside and outside edges of the ribs must be dressed so as to lie in the domical surface. Conical domes are very uncommon, but they are sometimes used. A conical dome is much easier to frame than a spherical dome because the ribs are straight. The shape of the curved plate which sup- ports the lower ends of the ribs may be easily determined, since it must conform to the line of intersection between the conical or spherical surface of the dome, and the plane of the face of the wall. Fig. 260. Diagram Showing Construction for Spherical Dome 197 186 CARPENTRY Fig. 261. Plan of Vertical Studding for a Niche Niches. Niches are of common occurrence in building work, especially in churches, halls, and other important structures. Some- times they are simply recesses in the wall with straight corners and a square head, but more often they are semicircular in form, with spherical heads, in which case the framing becomes a matter of some difficulty. The fram- ing of the wall for a semicircular niche is the easiest part of the work, since all the pieces may be straight, but for the framing of the head the ribs must be bent or shaped to conform to the surface of a sphere. Fig. 261 shows in plan the way in which the vertical studding of the walls must be placed. The inside edges must he in a cylindri- cal surface, and will receive the lathing and plastering. There must be a curved sole piece for them to rest upon at the bottom and a cap at the top. The cap is shown at A B, in Fig. 262, which is an eleva- tion of the cradling or framing for the niche. This figure shows how the ribs for the head of the niche must be bent. The ribs and vertical studs must be spaced not more than 12 inches apart, center to center. The form of niches described above is the most common one for large niches intended to hold full size casts or other pieces of statuary, but smaller ones for holding busts and vases are quite common. These are often made in the form of a quarter sphere or some smaller segment of a sphere, with a flat base or floor and a spherical head, as is shown in section in Fig. 263. They are framed with curved ribs in the same way as described above, and finished with lathing and plastering. Vaults and Groins. Although vaulted roofs are an outgrowth of masonry construction, and are almost always built of brick or Fig. 262. Elevation of Framing for the Niche 198 CARPENTRY 187 Fig. 263. Section of Small Niche stone, they are occasionally built of timber, and in any case a timber centering must be built for them. A vault may be described as the surface generated by a curved line, as the line moves through space, and in accordance with this definition there are vaults of all kinds, semicircular or barrel vaults, elliptical vaults, conical vaults, and many others. In Fig. 264 is shown in outline a simple semicircular or barrel vault, known as well by the name cylindrical vault. The point A where the straight vertical wall ends and the curved surface begins is called the springing point. The point B is the crown of the vault. The distance between the springing points on opposite sides of the vault is the span, and the vertical dis- tance between the springing point and the crown B is the rise. It may easily be seen how a barrel vault, or a vault of any kind, may be framed with curved ribs spaced from 1 foot to 18 inches apart on centers, and following the outline of a section of the vault. If the framework is intended to be permanent and to form the body of the vault itself, then the inner edges of the ribs must lie in the surface of the vault and must be covered with lathing and plastering. If only a center- ing is being built, on which it is intended that a masonry vault shall be supported temporarily, then the outer edges of the ribs must conform to the vaulted sur- face and must be covered with rough boarding to receive the masonry. When two vaults intersect each other, as in the case of a main vault, and the vault over a transept, the ceil- ing at the place where vaults come together is said to be groined. The two vaults may be of the same height or of different heights. If they are of different heights the intersection is known as a Welsh groin. Welsh groins are of common occurrence in masonry construction, but in carpentry work the vaults are Fig. 264. Diagram of Cylin- drical Vault 199 188 CARPENTRY almost always made equal in height and often they are of equal span as well. The framework for each vault is composed of ribs spaced com- paratively close together, and resting on the side walls at the spring- ing line. When, however, the two vaults intersect each other, the side walls must stop at the points where they meet, and a square or rectangular area is left which has no vertical walls around it The covering for this area must be supported at the four corner points in which the side walls intersect. This is shown in plan in Fig. 265 where A B C D are the points of intersection of the walls of the vaults. The method of covering the area common to both vaults is also shown Fig. 265. One Method of Cradling for Groined Ceiling Fig. 266. Another Method of Cradling for Groined Ceiling in the figure. Diagonal ribs A D and C B are put in place so as to span the distance from corner to corner and these form the basis for the rest of the framing. They must be bent to such a shape that they will coincide exactly with the line of intersection of the two vaulted surfaces. The ribs which form the framing for the groined ceiling over the area are supported on the diagonal ribs as shown in the figure. They are arranged symmetrically with respect to the center, and are bent or shaped to the form of segments of circles or ellipses. Fig. 265 shows one method of forming the cradling for a groined ceiling, but there is another which is also in common use. This is shown in Fig. 266. There are four curved ribs AB, B D, CD, and 200 CARPENTRY 189 C A, which span the distances from corner to corner around the space to be covered. The diagonal ribs A D and C B are also employed as in the first method. Straight horizontal purlins are supported on these ribs, running parallel to the direction of the vaults, as shown in the figure. They are spaced about 16 inches apart and form the framework for the ceiling. The only difficult problem in connection with groined ceilings is to find the shape of the diagonal rib. This rib, as has been explained above, must coincide with the line of intersection of the vaults. The problem, then, is to find the true shape of the diagonal rib. Let us consider the two vaults shown in plan in Fig. 267. They are not of the same span, but they will be of the same height if we wish to have a common groin and not a Welsh groin; so if one is semicircular the other must be elliptical. Elevations of ribs in each vault are shown at A and B and the diagonal ribs are shown in plan at ^ C and B D. It is easy to find the plan of these ribs because they must pass from corner to corner diagonally. To find the eleva- tion we must use the same prin- ciple that was employed in finding the position of the valley rafter and the shape of the curved hip rafter for an ogee roof, namely, any line drawn in the roof or ceiling surface parallel to the plate or side walls must be horizontal, and all points in it must be at the same elevation. We start with the assumption that one of the vaults is semicir- cular, as shown in elevation at A, Fig. 267. Taking any line in the vaulted surface, shown in plan, as the line S P 0,we produce it until it intersects the plan of the diagonal rib AC at the point 0. This point must be the plan of one point in the line of intersection of the vaulted surfaces. The elevation of the point above the springing line of the vaults is shown by the distance P S, since the line *S P is exactly horizontal throughout. This distance is laid off at EF with the Fig. 267. Diagram Showing Method of Finding Diagonal Rib for Two Intersecting Vaults 201 190 CARPENTRY line G E H representing the horizontal plane which contains the springing Hnes of the vaults. The point H is the point from which the diagonal rib starts. The point F, as we have seen, is another point in the curve, and we can by a similar process locate as many points as we need. This will enable us to draw the complete curve GFH of the line in which the vaults intersect, and to which the diagonal rib must conform. By continuing the line from the point at right angles to its former direction and parallel to the wall line, we may obtain the point K, which is a plan of one point in the surface of the elliptical vault. The elevation of this point also above the springing lines must be the same as for the point S and may be laid off, as shown at K M. By finding other points in a similar way the curve N M R oi the elliptical vault may be readily determined. 202 STAIR HALL OF HOUSE IN URBANA, ILL. LIVING ROOM OF HOUSE IN URBANA, ILL. Wbite & Temple, Architects, University of lUinoi For Exterior and Plans, See Page 366. CARPENTRY PART IV EXTERIOR AND INTERIOR FINISH In the preceding pages we have considered, the most important of the methods in use for the construction of the rough framework of buildings. We will now take up the general subject of finish, both outside and inside. There are two things which the outside finish of a building is intended to do : first, to protect the vital part of the structure — the framework; and second, to decorate this frame- work and to make it as pleasing in appearance as may be possible. Both of these purposes must be borne in mind when designing or erecting any outside finish, as both are equally important and neither should be neglected. Material. The material used for the finish varies under different conditions and in different parts of the country. Of course, it must first of all be durable when exposed to the weather, and it must be a wood which can be easily worked. The best kinds of wood for the purpose are white pine and cypress, and one of these woods is generally used. Spruce and Georgia pine are sometimes used on cheap work, but they are much inferior to white pine. Poplar is very good but scarce. Pine should be employed wherever it is obtainable. OUTSIDE WALL FINISH Sheathing. The first operation in connection with the applica- tion of the finish is that of covering the framework with sheathing, which should be about 1 inch in thickness, and for the best work, dressed on one side. The sheathing should be placed diagonally across the studding when the frame is of the "balloon" variety, but in case of a braced frame the boards need not be so placed. With the braced frame, the boarding may be started at any time, but with the balloon frame it is necessary to wait until all of the studding Copyright, 1912, hy American School of Correspondence. 203 192 CARPENTRY has been placed in the outside walls. The sheathing should be laid close, but need not be matched, and the boards should be fairly wide, say 8 to 10 inches, or even more. The most common material for this work is yellow pine free from loose knots. Spruce or hemlock may be used with good results. The roof sheathing is nailed on at right angles to the rafters but the boards are narrower. The width is usually 4 inches but 6 inches works well also. They are laid with spaces between them for the passage of air. These spaces are from 2 to 3 inches wide and are left for the ventilation of the shingling or other roof covering, and for the cheapening of the rough boarding. Roof boarding is not laid diagonally, but if this were done a very much stronger building would be obtained than where the boarding is laid horizontally, because each board acts somewhat as a truss to bind the whole framework more securely together. Some curved work, such as round towers and bay windows, is boarded diagonally, when it is impossible to place the boards in any other way. Each piece should be well nailed to each stud or rafter with large nails, ten-penny or eight-penny, and no serious defects such as large knots should be allowed in the sheathing. Building Paper. Building or sheathing paper is now placed over the sheathing to keep out the weather and to cover more com- pletely the joints in the boarding. This paper must be tough and strong as well as waterproof, and must be easy to handle and put in place. It must not be easily broken as any hole in it is an opening through which the weather will surely find its way. There are a number of different kinds of building paper on the market, prepared with various chemicals to render them as nearly waterproof as possible. Tar is sometimes used for this purpose. The paper should always be laid with a good lap. Water Table. Starting at the bottom of a wood structure, at the point where the masonry foundation wall stops and the timber framework begins, the first part of the outside finish which meets the eye is the water table, so called because its purpose is to protect the foundation wall from the injury which would result if water were allowed to run down the side of the building directly onto the masonry. The water table, therefore, must be so constructed as to direct the water away from the top of the foundation wall. This 204 CARPENTRY 193 may be accomplished in several ways, some of which will be described and illustrated. Fig. 268 shows the simplest form which can be used for a building which is to be covered with clapboards (this covering will be discussed later). A is the foundation wall, which may be stone, brick, or concrete. B is the sill, which may be in one or two pieces and which has already been described at length. C is the boarding or sheathing, which in this case is flush with the outside of the foundation wall, the sill being set back about 1 inch from this face. The board D is nailed so as to cover the joint between the top of the foundation wall and the sheathing, and on the top of it is fastened the water Fig. 268. Simple Form of Water Table Fig. 269. Type of Water Table with Rabbeted Table Board table E, which is inclined downward and outward so as to shed the water. The piece F is inserted to set out the first and lowest piece of siding G. Fig. 269 shows another way of constructing a water table for a building with clapboards. In this case the sill B is set back about 4 inches from the face of the wall A, and a block is nailed to the bottom of the boarding, as at C, so as to set the piece D out clear of the face of the wall. This piece is rabbeted at the top as shown, so as to take in the bottom of the last clapboard. Another method is shown in Fig. 270. In this case the board D is again made use of to cover the joint between the sheathing and the foundation wall, and it is nailed directly to the boarding 205 194 CARPENTRY as before, but the piece E is much larger and is blocked out by means of the addition of the piece G, so as to throw the water well away from the masonry. In many respects this detail is the best of the three, as the joint is well covered and at the same time there is pro- vided a very good wash for the rain water. There are many other ways of building this part of the finish, but only one more will be shown for use when the walls are to be covered with shingles. Fig. 271 shows how this should be done. Two or even three blocks, as at F, are nailed to the boarding and the shingles G are bent over them so as to shed the water free of the masonry without further help. They may be finished at the bottom with a molded piece D to hide the joint. The water table is sometimes omitted entirely and the clap- boarding is started directly from the foundation wall, but this is not considered good practice and will surely be found to be unsatisfactory. Fig 270 Water Table of Simple Construction Fig. 271. Water Table Used with Shingled Walls Clapboards for Wall Covering. The clapboards used for cover- ing walls are usually of white pine or spruce, though they are some- times made from a cheaper timber such as hemlock or fir. They are about 5 or 6 inches wide and about 4 feet long and are thicker on one side or edge than on the other. The thicker edge measures about I inch while the thin edge is only about | inch thick. Each clapboard, therefore, is as shown in Fig. 272, where A is an elevation and 5 is a section of the board. The tapering section is obtained by sawing the boards from a log, cutting each time from the circum- 206 CARPENTRY 195 ference Inward. The boards are thus all quarter sawed and shrink evenly, if at all, when they are exposed. When laid up on the side of a building, the clapboards should lap over each other at least 1| inches, as shown in Fig. 273. Here, A is the clapboarding, B is the sheathing, and C is the studding. As will be seen, the clapboards lap over each other, leaving a certain amount of each board exposed to the weather. This term "to the weather" is made use of in many specifications to indicate the amount of board which is to be exposed. Thus, "4 inches to the weather" means that 4 inches will be exposed. Building paper should be placed between the clapboarding and the sheathing, as shown at D, to keep out the weather. Fig. 272. Side and End Views of a Clapboard Fig. 273. Section Showing Method of Laying Clapboards Fig. 274. Section Showing Method of Laying Siding Siding. The only difference between common siding and clap- boards is in the length of the pieces, the siding coming in lengths of from 6 to 16 feet, while the clapboards are in short lengths as explained above. Common siding is put on in the same way as clapboards, but there is manufactured a rabbeted siding which is laid up as shown in Fig. 274. Here the rabbet takes the place of the lap, and is made 207 196 CARPENTRY about f inch deep. This siding is also furnished molded to a number of other patterns besides the simple beveled pattern, and is of various widths up to about 12 inches. Sometimes it is nailed directly to the studding, no building paper or outside boarding being used, but this construction, although it is cheap, is not suitable for any but temporary buildings. Corner Boards. It is cus- tomary, whenever the walls of a building are covered with clapboards, to make a special finish at the corners. This finish usually takes the form of two boards, one about 5 inches wdde, the other 3| inches wide by about 1| inches thick, placed vertically at each side of the corner so as to project 1| inches — the thickness of the board — beyond the face of the sheathing. Thus they form something for the clapboards to be fitted against. The corner boards may be mitered at the corner, but this is not desirable, as it is hard to make such a joint so that it will not open up under Fig. 275. Simple Corner Board in Place Fig. 276. Section Showing Corner Board Construction Fig. 277. Shingled Wall Requires no Corner Board the influence of the weather. The corner boards are, therefore, usually finished at the corner with a simple butt joint, the two pieces being securely nailed together. In some styles of work it may be 208 CARPENTRY 197 well to give the corner boards a special character, and this can be done by crowning them at the top with a capital, so that they will form a sort of pilaster at each corner of the house. A base may also be added if desired, though it is hard to make a base finish well on top of the water table. Fig. 275 shows a view of a simple corner board in place on the outside corner of a house. A is the corner board, B is the clapboarding, C is the water table, and D is the foun- dation wall. Fig. 276 shows a section taken horizontally through the corner of a building with corner boards and clapboards, showing how the corner boards are applied to the outside boarding. In this figure, A is one of the corner boards, B is the outside sheathing, C is the studding at the corner built up of 2 X 4-inch pieces, and D are the clapboards. The width of the boards may, of course, be varied to suit the taste of the designer. When the walls of the building are to be covered with shingles it is not necessary to have corner boards, as the shingles can be brought together at the corner and made to finish nicely against each other. The usual method is to allow the shingles on the adjacent sides to lap over each other alternately as shown at A in Fig. 277. Shingles. Instead of clapboards, shingles may be used for covering the walls of a building, though this method is more expen- sive than the other. The advantages are in the appearance of the work, the variety of effects which may be obtained, and also in the fact that the shingles may be more easily dipped in some stain and a greater variety of colors thus obtained. Wall shingles should be laid with not more than 6 inches to the weather, and an exposure of 5 inches is better, but even if 6 inches are exposed, there will be a greater thickness of wood covering any particular spot of the wall with the shingling than there is with the clapboarding, and thus a greater protection from the weather is obtained. The arrangement of shingling on a wall is shown in Fig. 278. It will be seen that the shingles are in all cases two layers and in some cases even three Fig. 278. Section of Shingled Wall 209 198 CARPENTRY layers thick. The width of ordinary shingles varies from about 3 inches to about 12 inches, and for rough work these widths may be used at random, but shingles which are called "dimension shingles," find are cut to a uniform width of 6 inches, may be had and these should be used for any careful work. Also shingles may be obtained which have their lower ends cut to a great variety of special and stock patterns, which may be worked into the wall so as to yield any desired effect. A shingled wall is shown in elevation in Fig. 277. Building paper should be used under shingles in the same way as under clapboards. Belt Courses. It is often desir- able, for the sake of effect or for the purpose of protecting the lower part of the walls of a building, to arrange a horizontal projecting band or "course," as it is called, which will slightly overhang the lower part of the wall. This is called a "belt course" and usually occurs at or near a floor level or across the gable end of a building at the level of the eaves. A belt course is formed by placing blocks or brackets at intervals against the face of the outside boarding, these blocks being cut to the required shape to support thin pieces of molding. This arrangement is shown in section in Fig. 279. Here, A is the studding, B is the boarding, C is the block or bracket, D is the finish under the block, E is the wall shingling, F shows where the shingles come down over the belt course and the furring G supports the finish and provides nailing surface for the first course of shingles. A similar belt course may be placed on a building with any other kind of wall covering, the prin- ciple being the same in every case, and the purpose being always to form a projecting ridge from which the water will drip without injur- ing the wall surface beneath. Sometimes the wall covering is not brought out over the top of the belt course, but is stopped immedi- ately above it, and in this case care must be taken to see that the top Fig. 279. Details of Belt Course 210 CARPENTRY 199 of the course is well flashed with galvanized iron or copper so that the water can not get through the wall around it. It is best to cover the entire top of the belt course with the flashing and to run it up onto the vertical wall 4 or 5 inches with counter flashing over it. The method of flashing will be explained later. Fig. 280. Simple Gutter OUTSIDE ROOF FINISH Finish at the Eaves. The point, or rather the line, in which the sloping roof meets the vertical wall is called the "eaves" and this point must always be finished in some way. This finish, however, may be varied to almost any extent, and it may be very simple, so as to barely fulfill the necessary requirements, or it may be very elaborate and ornamental as well as practically useful. Gutters. Practical considerations require that at the eaves some kind of a gutter must be provided, to catch the water which falls on the roof and streams off from it. This gutter, of whatever kind, must be supported far enough away from the straight vertical wall of the building so that the water dripping from it in the case of a possible overflow will fall free of the walls and not injure them. Usually, the gutter ought not to be nearer to the wall than one foot. There are a number of different kinds of gutters in use, and per- haps it will be as well to describe some of them at this time, as they form a part of the eave flnish. The simplest kind is made of wood, and is generally kept in stock in several different sizes by lumber dealers. It is of the general shape shown in section in Fig. 280, and the most common sizes are 4X6, 5X7, and 5X8 inches. They are usually made of white pine, but may better be made of cypress or redwood. Spruce is hardly durable enough for use as a material for gutters. Besides the wood gutter just described, there are in use a number of different forms of metal gutters, some of which are carried in stock by dealers in roof supplies and others which must be made to order by the roofer for each particular job. The metal gutters are made of galvanized iron, copper, or of a tin Fig. 281. Metal Gutter 211 200 CARPENTRY lining in a wood form. Any wood gutter may be improved by lining it with tin or zinc, and there should always be a piece of one of these metals used to cover the joint between the two pieces of a wood gutter where they meet. Wood gutters can be had only in lengths of about 16 feet at the most, so usually there must be some joints to be covered with metal. The simplest metal gutter takes the form of a trough, as shown in Fig. 281, and is fastened in place by hangers placed at frequent intervals or by brackets which answer the same purpose. Either the hangers or the brackets may be spiked to the ends of the rafters, and thus a cheap and simple gutter may be obtained. Open Cornice. The most simple way of sup- porting the gutter is to let the main rafters of the roof framing extend out over the wall as far as necessary and cut a rabbet in the end of each of them into which the gutter will fit. In this case the wood gutter should be used. It is fastened into the notches left in the ends of the rafters, as shown in Fig. 282. In this figure, A is the gutter, B is the rafter, C is the studding of the building, D is the plate with the rafters cut over it and the ceiling joists E resting on top of it, F is the outside boarding, which may be covered with clapboards or shingles, and G is the roof boarding, which may be covered with shingles or slates. In one respect the construction shown in this figure is faulty, because the gutter being placed in the position shown, snow sliding off the roof would catch the outer edge of it and perhaps tear it off. The gutter should, wherever possible, be placed low enough so that the line of the finished roof, H in the figure, will clear the edge of it. In order to improve the appearance of the eaves it is well to place a board J along the edge of the rafters so as to hide them and present a plain surface to the eye and to finish the joint between this board Fig. 282. Section Showing OiDen Cornice Construction 212 CARPENTRY 201 and the under side of the gutter with a small bed molding K. Under- neath the rafters where they cross the plate and come through the outside boarding is placed a board L which forms a stop for the sid- ing or shingles, and another board should be inserted between this and the under side of the roof boarding, as shown at M. It is also a good plan to cover the ends of the ceiling joists with a strip of board- ing to keep the wind out of the roof space. This is shown at 0. The finish shown in Fig. 282 is of the simplest and barest kind and can be used only for buildings of an unimportant character such as stables and outhouses or for cheap country houses. Boxed Cornice. A better finished form of cornice is shown in Fig. 283. Here an extra piece P is placed just above the gutter so as to cover the spaces between the rafters, and the entire under side of the rafters outside of the wall of the building is covered with boarding as at Q, so that the rafters will not be seen at all. The piece L must still be used, however, to stop the wall covering. This figure also shows the roof shing- ling at R. This should be laid about 4 to 4| inches to the weather. It is laid closer than the wall shingling because it is nearer flat and the water will stand on it longer than it will stay on the wall. The water thus has a greater chance to leak through and the shingling must be laid closer and thicker on the roof. It is wise to insert blocks on top of the plate as showm at *S, and to continue the side sheathing up to the roof sheathing so as to make everything tight. All of the boarding used for the boxing in of the rafters at Q, together with the pieces L and J, may be of |-inch stuff. The piece P should usually be of l|-inch stufjf so as to allow of rabbeting it to receive the gutter. The piece K may be molded as desired by the designer of the building. J is called the fascia and Q is called the planceer. Fig. 283. Section Showing Boxed Cornice Construction 313 202 CARPENTRY The principal objection to the boxed cornice is that, if snow melts in the gutter and freezes afterward so as to fill up the gutter during the winter, it is very likely to work its way up under the shingles when it finally melts in the spring, and in this way find its way into the attic. In the case of the open cornice there is little chance of this happening, as the water can get away at the back of the gutter between the rafters. Sometimes it is desired to place the planceer in such a position that it will be horizontal instead of following up along the under side of the rafters. This is accom- plished by fastening a piece of furring to the end of the rafter and to the wall in such a way that the bottom of it will be hori- zontal, and spacing these pieces 12 or 16 inches apart all around the eaves of the building. To this furring can be nailed the planceer which will box in the cornice and be in a horizontal position. The gutter can still be fastened to the end of the rafters as before. This construction is shown in Fig. 284. The pieces of furring referred to are shown at A. Fake Rafter Construction. It is often desirable, for the sake of architectural effect, to break the surface of the roof just above the eave line, and in order to do this it is necessary to make use of small pieces of rafters, called false rafters, or sometimes "jack" rafters, which are nailed to the ends of the regular rafters. In this case the regular rafters stop at the point where they rest on the plate, and the false rafters project out over the wall line as far as may be desired. These false rafters are cut into various shapes and are usually left exposed on the under side, being in this case made of a better and harder class of wood than that used for the regular rafters. The gutter may be placed on the ends of these false rafters if desired, but Fig. 2S4. Boxed Cornice with Horizontal Planceer 214 CARPENTRY 203 it is more usual to make use of a construction such as is shown in Fig. 285. Here it will be seen that the gutter is merely formed up on top of the roof shingling by a piece A which is held in place by the bracket B, The brackets occur at intervals of about 2 feet, while the piece A is continuous. The shingles which cover the roof are stopped on the strip C and the inside of the gutter thus built up is covered with galvanized iron, or copper, to make it water-tight. The ends of the rafters may be finished with a fascia as shown at D, and the space between the false rafters along the wall may be finished as shown at E. In place of the wood piece A which forms the out- side member of the gutter, a piece of metal rnay be used to accom- plish the same purpose, and this is often done. Fig. 285. False Rafter Construction with Shallow Gutter Concealed Gutters. Another common form of eave provides a concealed gutter. In this construction the ceiling joists are extended beyond the outside walls and the rafters are cut to set over the plate. The cornice and gutter are illustrated in Fig. 286. The studding is shown at S, and on the plate P the joists C extend over from 8 to 14 inches, depending on the effect desired. Around the joist the planceer Q, the fascia J, the bed N, and crown molding M are fixed. A notch is cut in the joist and a |-inch piece is nailed in this notch and on the outer end. Tin or copper is used to conduct the water over the gutter. The tin should be nailed to the crown molding and should be run 8 or 10 inches above the shingle line. Finish for Brick Walls. Any of the forms of eave finish described 215 204 CARPENTRY above may be used equally well in cases where the wall is of brick instead of wood. In this case a wood plate is placed on top of the brick wall and the rafters are brought down over it, and are either extended out over the wall or are fitted with false rafters. The joint between the brick wall and the wood rafters is finished with a wood frieze. Cornices may be much more elaborate than any of those illus- trated above, indeed those shown here are suitable only for the plain- est and cheapest kind of work, but the principles of construction are the same in all cases, the difference being in the amount of orna- ment applied to the building. The ornament takes the form of molded pieces of timber which are supported by rougher furring pieces placed behind them. Economy demands that the finished pieces be so arranged as to be cut out of boarding of medium thickness, and as much space as possible should be occupied by the rough, concealed furring. As a rule, all of the eave finish can be taken out of |-inch stuff. Care must be taken always to give the gutters the proper slope to the outlets called "down- spouts" and they should be made large enough so as not to over- flow. A gutter should slope f inch to the foot. Ridge Finish. At the ridge of a roof where the two slopes meet there must be some special provision made for the proper finish of the roof covering, something for the shingling or slating to finish on as well as some adequate means of covering and making water- tight the joint which occurs at this place. There is usually a ridge board, or ridgepole, which receives the rafters, and this piece may be cut off just at the ridge so that the rough roof boarding goes over it, or it may be made wider and may project up above the roof boarding, so that this boarding stops against it instead of going over it. The two different methods call for different kinds of ridge Fig. 286. Cornice Construction for Con- cealed Gutter 216 CARPENTRY 205 finish. The first is the most simple, and may be taken care of as shown in Fig. 287. Here, A is the ridge board, BB are the rafters, E is the roof boarding, C is the shinghng, and D is the finish at the ridge, consisting of two pieces of board about 6 or 8 inches wide and I inch thick, which are nailed on top of the shingling to form a finish. In case the ridge board is carried up above the roof boarding, it is customary to make the ridge finish of galvanized iron or of copper or other metal. This may be done very simply, as shown in Fig. 288. Here the ridge board is extended above the roof board- ing and around it is shaped a strip of galvanized iron or copper or zinc, which is continued down over the shingles of the roof so as to form a flashing. This makes a good ridge finish and one which is water-tight if it is properly put on. The galvanized iron should be Fig. 287. Simple Ridge Finish Fig. 288. Galvanized Iron Ridge Finish flashed down over the shingles for a distance of at least 6 inches. It is not necessary that the ridge board should be extended above the roof boarding. The same result may be accomplished by nail- ing a separate piece of 2 X 4-inch or 2 X 5-inch scantling to the top of the shingles, running lengthwise of the roof, to form a ridge over which the metal may be shaped. This method may perhaps make a tighter job than the other. Skylight Openings. It is sometimes necessary to make an opening in a roof surface for the admission of light to the rooms under the roof. This is usually done by the formation of what is known as a dormer window, the method of framing for which has been already described, but often it is desired to admit light when the attic space is not of sufficient importance to justify the introduc- tion of a dormer window in the roof. In this case recourse is had 217 206 CARPENTRY to a skylight. A skylight may be obtained by the use of glass in the roof surface in place of other roof covering such as roof boarding and shingles, but it is almost always necessary to provide some sort of frame for this glass, and the glass surface is usually raised about 6 or 8 inches above the shingled roof surface so as to keep the glass as free as possible from snow and to separate it from the general roof surface. The first thing to be done is to frame an opening, in the rough, between the rafters, by introducing trimmer pieces just above and below the place where the opening is to be. This Fig. 289. Section Showing Skylight Construction completes the rough framing for the skylight. The finished open- ing may be formed in a number of ways, one of which is shown in Fig. 289. Here, AA are the pieces referred to above, which frame between the rafters BB and form the rough opening. D is the roof boarding which is brought up to the edge of the opening as framed, and sawed off flush with it as shown. CC are pieces of rough stuff which are nailed on top of the boarding to raise the skylight above the roof surface. They may be of any size desired, but in this case they are 4X6 inches. EE is sheathing about | inch thick which forms a finish for the inside of the skylight opening. It is con- 218 tH 1-1 < H a. Ul U Z M 1-1 (« u O a 1-1 u s * o M C) o l-l l-l o «J ri K