STATE GEOLOGICAL SURVEY. Governor C. S. Deneen, T. C. Chamberlin, E. J. James, Commissioners H. Foster Bain, Director. Qualities of Clays Suitable for Making Paving Brick and Pryo-Physical and Chemical Properties of Paving Brick Clays BY ROSS C. PURDY [From Bulletin No. 9, pp. 133-278.] URBANA University of Illinois 1908 4. (c 134 PAYING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 The science of ceramics is making rapid progress in the solution of these problems; but today the question of what physical and chemical properties raw clay must possess that it might be suitable for use in the manufacture of paving brick is still unanswerable. Chemical analysis alone is not a safe criterion by which to decide this and similar ques¬ tions, for, as can be shown in nearly, if not all, geological survey re¬ ports on clays, the analyses of clays that are known to be suitable for paving brick have their counterparts in analyses of building brick clays. Indeed in range of variation in chemical constituents, these two types of clays overlap one another to a very large extent. There are possibly one or two characteristics in the chemical composition of paving brick clays that are not common to those used for building brick, and yet no fixed rule has been, or so far as the writer can perceive, can be, laid down at present, by which to identify paving brick clay by chemical analysis. Physical tests on green or unburned clay, so far as is now known, would not lead one any nearer the possibility of fairly judging a pav¬ ing brick clay than would chemical analysis. Possibly an exception should be made of determinations of fineness of grain. Plasticity, tensile strength, bonding power, slaking properties, etc., are .found to vary widely in different paving brick clays, so that no dependence can be placed upon any of them, taken alone. The determination of fine¬ ness of grain, however, does give a negative test that seems to be of some value. Fine grained clays, as will be seen later, have not proved to be good paving brick clays. It cannot be said, however, that all coarse grained clays are good paving brick clays. Indeed, although evidence is lack¬ ing, there is no obvious reason for believing that any hard and fast rule can at present be laid down in regard to either fine or coarse grained clays. When the history of a few paving brick plants in various parts of this country reveals the fact that experienced paving brick manufactur¬ ers have so misjudged a deposit of clay as to erect an extensive plant upon a particular site and soon find that they must abandon the idea of attempting to make any other than a building brick, it must be in¬ ferred that even a burning test as ordinarily conducted by ceramic engineers, surveys and brick machine manufacturers likewise often gives evidence that is untrustworthy. By what means then can the suitability of a clay for paving brick purposes be ascertained? It was with hopes of obtaining evidence upon* this problem that the Survey undertook a study of the properties of the clays and burned bricks of several of the leading paving brick manufactories in the middle west, together with several samples' of clays from various parts of this state, that are not now being used for paving brick manufacture. For many years scientists have been devising methods with which’ to determine the cause and effect of the various properties of clay, but they have not made much progress. For instance, the reason why kaolin and a ball clay, having similar chemical composition and size and apparently character of grain, should differ so widely in plasticityj, purdy] QUALITIES OF CLAYS FOE MAKING PAYING BKICK. 135 is still an open question. The refractoriness of a clay is still incal¬ culable from analytical data, although exhaustive researches have been made to determine the pyro-chemical effect of inorganic acids and bases, singly, collectively, and in mixtures, with standard clays and com-, pounds. While from these pyro-chemical studies it has been shown that the fluxing power of the bases is roughly proportional to their mole¬ cular weight, and that the several acids operate in a definite manner, so that synthetical mixtures can be made with assurance that each compon¬ ent will operate in a given manner, and that the resultant effect of the mixtures will in general be as presupposed, similar natural mixtures, known as clays, exhibit properties that are in the large majority of cases entirely contradictory to those of synthetical mixtures, due no doubt to differences in the physical properties of the minerals as well as to variation in mineral content. Many theories have been advanced concerning the geological history of clays, and general statements can be made as to the probable conditions that cause the breaking down of the parent rock, the character of the resi¬ dual debris, the agencies sorting and transporting this debris, and the conditions under which it can be deposited in different grades of fineness and purity. Geologists can state with considerable accuracy, the effect of vegetable growth and of ground water, the cause for the precipitation of salts from solutions, the cementing value of various compounds under different conditions, etc. They can establish the faot that there is a cycle of rock decomposition, residual deposition, and rock formation going on constantly en masse, as well as in the small grains of which clay and soils are composed. Yet, after all, neither geologists nor chemists are able to determine the exact stage of breaking down or building up, nor the exact combination of several ingredients existing in a clay at the time of examination. It certainly seems patent that until we can determine the exact mineralogical condition and chemical aggregation of a given clay it will be impossible to use the analytical data obtained by ordinary phy¬ sical and chemical tests as ground for predictions concerning its probable pyro-chemical behavior. In the process of any chemical analysis known to the writer, the character and exact identity of the clay as a whole, as well as its con¬ stituent parts, are destroyed by the disintegration or unlocking of the natural combinations, making an exact or complete determination of the chemical conditions originally present, a mere supposition. In fact all we know or can learn from a study of the origin and mode of forma¬ tion of clays, and of the alterations in their composition constantly go¬ ing on under varying conditions, as well as by attempts to unlock the combinations or separate the ingredients by chemical methods is, that we are arresting the changes of transition in the clay from one state to an¬ other, but are not able to ascertain the forms or conditions existing at that time. From these considerations it should be plain that two samples of clays having similar origin and chemical constitution, may differ rad¬ ically in their mineralogical make-up. The kind, size and composition of the several minerals affect so materially the pyro-chemical properties of the clay as a whole, that until mineralogists can find means of determin- 136 PAYING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 ing the kind, quantity and relative size of the several mineral ingredi¬ ents, ceramists cannot predict, even with a fair degree of accuracy, the behavior of a clay in burning. Several of the so called physical properties of raw clay may, however, be measured and their effect on the behavior of the clay described. In this, physical conditions of the clay, such as hardness, fineness of grain, plasticity, etc., are by custom regarded as properties. The properties of clays may be classified under three general heads: Physical, chemical, and pyro-chemical. PHYSICAL PROPERTIES. Introduction. Ceramists have tested clays by all the means that have been suggested to them. Many of the tests have proven fruitless, and not a few now in use are of doubtful value. The following physical tests on raw clays were made by the State Survey: 1. Specific gravity of the clay, or mineral aggregate. 2. Porosity of a dry, unburned brick made from “stiff mud.” 3. Drying behavior. 4. Shrink¬ age. 5. Tensile strength. 6 Fineness of grain. 7. Water of plasticity. 8. Plasticity. Specific Gravity. REAL AND APPARENT SPECIFIC GRAVITY. The determination of the specific gravity of a clay, if made at all, should be so conducted that the result would be a composite of the speci¬ fic gravities of the several minerals that make up the clay mass. If the specific gravity of a lump be taken as a whole, unless the mass be so thoroughly saturated that each grain becomes surrounded by the saturat¬ ing medium, it would vary with all kinds of irregularities incident to the processes of formation or manufacture. The first method would give what is known as true specific gravity, while the second would give only an apparent value. The writer can see no value in finding the apparent specific gravity, for, as a means of detection of any working property , it is absolutely valueless. The true specific gravity may have direct value as indicating some working property of the clay, but if it has, the fact has not been demonstrated. The data for the true specific gravity can, however, be used, as will be demonstrated later, in the analysis of some of the changes that take place in drying and burning, and serves as a check on the accuracy of some of the other data. METHODS OF DETERMINATION. The true specific gravity of the clays included in this report was ob¬ tained by three methods: By Seger volumeter, using unburned bricks; by pycnometer; by chemical balance, using unburned bricks. Determination by Seger s Volumeter. —Seger’s volumeter was used in the determination of the volume shrinkage, porosity and specific gravity on the several clays, as noted in Table I . Table I —Results of Physical Tests on Green Bricks. PURDY] QUALITIES OF CLAYS FOE MAKING PAYING BRICK. 137 Per cent Porosity. Percent variation. .. 0'#ooMNO)HHoooooainioocDot-N®aHooootDifl+Nt'C'ini-«o.o Average. oSSS^SosNH^NNMifliOO^OOOCWOa^OOQNiOaOOWOSWOSlO a g g d g g n s a si s' a a s’ s' a s' s' s' s' s' s' s' s' s' s' s' s' s s' s' Minimum ... 00[-00!D»NOrt(»NScONO«Oit'^iON«HOOOi05H01tOfflNOOtOH» s' s' s' s' s' s' s' s' s' s' s' s' s' s' s' s s' s' s s' s s s' s' s' s' s' s' s' s' s' s' s' s' s' Maximum.. .. t'H00NNSiHO!00'nSininriHMD-00'#NinCC05Hl£IMt'M0iOO00OlOC' s' s' s' s s' s' s' s' s' s' s' a s' s' s' s' s' s' s' s' s' s' s' s' s' s' s' s' s' s' s' s' s' s' s' .1■at' Apparent Specific Gravity. Percent variation. . . Average. Minimum ... 2 °°. ® os us c- o o os th ?o cm oo us co -<* co co io os ec oo co id o cm i-h o» c- os cm co io£ioioSDcDg33gio§cococo3cDio3co8coiocD2io82£ioSg3i®2 NNNNNNNNNMNNWNNNNNNNNNNMNNNNNNNNNNN S 2 S3 S © 2 S3 g 5; 2 S3 5 g 2 3 S3 2 8 3 S3 8 8 8 g 2 2 S 3 S3 S S3 8 S S 2 CM CM CM CM CM CM CM CM CM CM CM CM CM CM (M CM CM CM CM CM CM* CM CM CM CM CM CM CM CM CM CM CM CM CM CM Maximum.... 3 g g 8 3 8 P £ g 8 g 8 g 3 5 5 g 8 5 3 8 g 8 g 2 8 g 3 8 g 3 8 8 8 3 CM CM CM CM Cl CM CM Cl Cl CM CM Cl Cl Cl Cl Cl Cl CM Cl Cl Cl Cl Cl CM Cl Cl CM CM CM CM CM CM CM CM CM Per cent Linear Shrinkage. Percent variation... Average . CO 1 CD +05 CMCMOOOOO^^OO ©oococo©eM©©©'^coc-+iocMTHin©eM© 133. 70. 68 73 129 tl 95 75 37 73 44 48 93 §2©CMioi-l05r-ia5aOWCOC010 t-CM10CO©CMCO©©lOCMCOt-00©C-aOCM'-i<© HOONCOH^CONOWCOCOCOH (N^^CvOCOCOCOlOHlO^^CON^WCOI>I>Tt- Minimum ... OOClCOOOOOn-OCDOOOOOOCM O CM t— l CM O OO CM t— 1 © ID CM CM CM t— 1 WOMOONONNXNOONCO^NCDCDCDCO CMCOCOCMCMCMCM'^OCO'^^COrHiO'^lOiOiOCO Maximum.. .. 00C0O'*<0000"*i<©00CM-**''#-«tlC0 OOOCO©COOOCOOO©aO-^ICOCMOOaO-*<-^'COOO-*t< CM^CO^CM^^COrHt^^^^frJ ^'^lO^COCO^iOCOCOlOlOlCCOC^t'-tr-OOOOiO Per cent Volume. Shrinkage. Percent variation. .. OO-^O-OOC^COC^iOrHlOOOOiiOiO ©coc-co • t— .ooin®roiH»®cD -oc-ao*o 050100 ;© jC-CMC-N^CD©^ ;«HOO Average . cmS2SSoo3SS»oS3S 051-H©1-I -05 : kO W CO OO T* c- 00 ^ : iOO'<*'cd CD HHH i °2 ait>W HHHH CO PSS°> ;SS8P Minimum ... s ® 04 05 ® s' d s' ® " 11.44 12.6 13.5 8.00 5.76 11.0 6.98 13.7 13.6 14.1 9.54 7.17 20.7 16.2 16.8 19.9 10.07 Maximum.... co 8 co 2 co t- g eo oi S 05 g g +d d s *°* s' s* °°" °°'s s' s' s' ® 13.71 14.0 14.3 10.31 6.16 11.9 7.9 14.8 14.0 14.7 10.20 8.70 21.7 16.7 20.0 21.1 11.9 Per cent Hygroscopic Water. Percent variation... 13.9 9.2 4.1 13.0 19.5 25.9 4.7 20.2 13.4 8.0 11.5 8.7 21.3 10C-CO'#lOCM©OlOCOi-ICMlO©©00©OOlOC- ^'s'g's's'^'^'ds'd^'^'^'s'g's'ss's'd Average 3SS :lsSS23Sgsl NHN jdrtHHONHWNd -TiiCMi-ii-ii-ieMCMT-*a5coa5'SiSoe-©eMcoa5©co-'#©c-a5©oocOi-HCOio©eococoooeo s' d d s’ s' d d d s' § s' d s' d d d d d s’ s' s' d d d s’ d d s' s' d s’ s' s’ d s’ Sample number. Sc-w^iD^Voo'^S^SS^S^^dcM'co’SSV^+GY^SSSasa^ Kiln Letter. :0 L ' r O£KNJ t 138 PAVING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 The accuracy of the tests in the volumeter is shown in Table I by the percentage of variation volume shrinkage, specific gravity and porosity. These variations are very small considering the conditions under which the determinations were made. * The specific gravity of any substance is the ratio of the weight of that substance to the weight of an equal volume of some substance taken as a standard. In the metric system distilled water at 4°C. is taken as the standard. At this temperature a cubic centimeter of distilled water weighs one gram. Therefore, when using this system, volume and weight of water may be interchanged, i. e., 1 c.c.=l gram. With this understanding the formula by which the specific gravity of a clay can be obtained could be expressed as; W „ — == bp. gr. where W=dry weight in grams and V=volume in cubic centimeters. If the porosity of the brick has been determined the formula for the specific gravity could be written: w Y (100—P) bp. gr. Where W=dry weight as before, Y=volume of the brick in cubic centimeters and P=percentage porosity. It will be noted by comparing the specific gravities in Tables I and II, that those obtained by the volumeter are lower than those obtained by the pycnometer. This can be accounted for perhaps by the operator’s in¬ ability completely to saturate a brick, that is, to fill all the pore spaces with oil without resorting to the use of a suction or vacuum pump to re¬ move all the air from the pores so that oil could enter. If the air is not entirely exhausted it will pass through the oil very slowly, requiring a period extending over several weeks in which to esape. In ordinary labor¬ atory practice sufficient time can not be given to permit the complete es¬ cape of the included air. In the porosity and specific gravity tests here reported, no attempt was made to fill the pores completely. The bricks were simply soaked in coal-oil for 48 hours, with one face exposed at the level of the surface of the oil. This incompleteness of saturation under these conditions is shown by the difference in the specific gravity as de¬ termined by the volumeter and pycnometer. Determination by Pycnometer .—A pycnometer, or specific gravity bottle, as it is often called, is a small flask of known capacity, usually 25 to 100 c.c. When filled up to a given mark with air-free water at normal room temperature, its weight is noted. The flask is then partly emptied, a known weight of clay added, and the whole carefully boiled to exclude all the entrapped air, then cooled, filled up to the mark and weighed. By the formula, weight of dry sample (a) plus weight of bottle filled with cold air-free water (b) minus weight of bottle filled with sample and water (c), or a+b—c, will give the weight of water having the same volume as the sample or true total volume of the clay particles. Knowing the dry weight and true volume of the grains, their composite specific gravity is readily calculated by the formula (dry weight -T- volume). PURDY] QUALITIES OF CLAYS FOR MAKING PAYING BRICK. 139 This may be illustrated by the following calculation: . Weight of bottle filled with water = 143.22. Dry weight of sample = 3.41. Weight of bottle + sample + water required to fill to mark = 145.35. 143.22 + 3.41 — 145.35 = 1.28 total volume of the particles. 3.41 -f- 1.28 = 2.68 specific gravity of the sample. In the following table will be found specific gravity of the clays by the pycnometer method. TABLE II. I II Average K 1 Alton, Ill . 2.666 2.664 2.665 K 2 St. Louis, Mo. 2.602 2 527 2.564 K 3 Albion, Ill. 2.688 2.684 2.686 K 4 Springfield, Ill. 2.667 2.668 2.667 K 5 Edwardsville, Ill. K 6 Galesburg, Ill. 2.676 2.661 2.626 2.664 2.651 2.663 K 7 Streator, Ill. 2.643 2.63 2.636 K 8 Veedersburg, Ind. 2.693 2.685 2.689 K 9 Crawfordsville, Ind . 2.701 2.703 2.702 K 10 Terre Haute, Ind.. 2.683 2.689 2.686 K 11 Brazil, shale. K 12 Brazil, fire clay. 2.667 2.671 2.669 K 13 Clinton, Ind. 2.682 2.708 2.695 K 14 Western Brick Co. 2.633 2.646 2.639 K 15 Barr Clay Co., Streator, 111. 2.719 2.713 2.716 R l Nelsonville, O. 2.633 2.632 2.633 R 2 Portsmouth, O. 2.719 2.712 2.715 R 3 Canton Imperial. 2.655 2.656 2.655 R 4 Canton Royal. 2.720 2.722 2.721 S 1 Moberly, Mo. 2.643 2.646 2.643 S 2 Kansas Citv, Mo. 2.717 2.716 2.717 F 1 Danville Brick Co. 2.708 2.710 2.709 H 24 Carbon Cliff, fire clay. 2.660 2.654 2.657 H 17 LaSalle, Ill. 2.608 2.591 2.599 H 16 Peoria, Ill. 2 700 2.690 2.690 2 695 H 18 Sterling, Ill. 2.653 2.671 H 23 Carbon Cliff, shale. 2.628 2.624 2.626 H 21 Galena, Ill. 2.718 2.715 2.717 H 20 Savanna, Ill. 2.718 2.715 2.717 H-II ToDeka. Kan. 2 683 2.685 2.684 L-II Lawrence, Kan. 2.702 2.707 2.705 I-II Casey, Kan. 2.668 2.676 2.672 J-II Pittsburg, Kan. 2.699 2.697 2.698 B-II Atchison, Kan. 2.666 2.668 2.667 G-II Coffeyville, Kan. 2.704 2.707 2.706 Determination with Chemical Balance .—The dry, saturated and im T mersed weights of briquettes were determined by using a chemical bal¬ ance. For this it was found that briquettes of the size %"x%"x2%" could be used. Obviously the larger the briquette the more nearly true will be the determined specific gravity. Sizes larger than that given, however, cannot be used to advantage on the ordinary chemical balance. This method was used for but a small number of samples. The briquettes were dried to constant weight in an air bath at 120°C. cooled in a dessicator and their dry weight obtained as rapidly as possible. After weighing, the briquettes were immersed in clarified coal-oil with one face above the level of the oil. After standing thus for 20 to 24 hours, they were placed under a bell jar and the air kept exhausted for fifteen minutes, it having been found in previous work that this treat¬ ment was sufficient to attain nearly complete saturation. The briquette was then suspended by a silk thread from the beam of a chemical bal- 140 PAYING BRICK AND PAYING BRICK CLAYS. [BULL. NO. 9 ance and its saturated weight noted. A breaker partially filled with oil was then so placed that the briquette could swing clear and be com¬ pletely immersed. In this manner the immersed weight of the briquette was obtained. By the formula then of dry weight (D) divided by (dry weight (D) minus suspended weight (S) ) orD-f (D—S), the specific gravity of the material in the briquette was readily obtained. The comparative accuracy attained in the determination of the speci¬ fic gravity of clay by these three methods may be seen in the table fol¬ lowing. Table III. Volumeter- Average. Pycnometer- Average. Chemical Balance. Average of three determina¬ tions. Max. Min. K 1. 2.54 2.66 2.614 2.616 2.615 K 14. 2.60 2.64 2.69 2.57 2.65 K 4. 2.58 2.67 2.63 2.58 * 2.61 It is evident from this table that the essential fault in the first and third method of determining specific gravities lies in the fact that the brick was not completely saturated, and therefore, gave low specific gravities. The closeness in agreement, however, suggests that by the use of extra precaution in the saturation of the bricks the specific gravities of the clays could be made quite accurately by either of these two methods. Porosity. DEFINITION". The percentage of porosity expresses the relation between the volume of pore space and the combined volume of the particles ‘of which the clay is composed. It is the ratio, in terms of volumes, of void spaces to solid particles. If determined on an unburned brick it would measure the degree of consolidation of the mass. METHOD OP DETERMINATION. The porosity of an unburned clay mass may be determined directly by two methods: first, by use of the Seger volumeter; second, by the use of a chemical balance, and indirectly, or by calculation on basis of the pycnometer specific gravity determination. To obtain percentage of porosity by either of the direct methods, the briquette or lump must be dried to constant weight, and the dry weight obtained, then saturated in kerosene and the saturated weight obtained. The difference between the saturated and dry weights is obviously the weight of petroleum that is required to fill the pores. This weight divided by 0.8, the density of the oil, gives the equivalence of oil in PURDY] QUALITIES OF CLAYS FOE MAKING PAYING BRICK. 141 terms of water. Thus far, therefore, the actual amount of pore space in the brick in terms of water by weight is known. If the metric system has been used throughout, this amount of water by weight is equivalent to its amount by volume, since one cubic centimeter of water at room temperature weighs practically one gram. Complete saturation of a lump of unburned clay even with kerosene, for which clay seems to have a peculiar physical attraction, cannot be obtained without resorting to the use of a vacuum pump. Standing in oil for 48 hours is not sufficient to cause complete saturation, as has been shown on proceeding pages by the specific gravity so obtained, as well as by the discrepancy between the directly measured and the calculated porosity as given in table IV, page 144. In obtaining the dry and saturated weights, the two direct methods are alike. The data for actual volume of the pores thus obtained are, however, of no value in themselves, and cannot become of value until calculated to parts of 100 unit volumes of the whole brick. For this, it is more practical to determine the volume of the whole mass, i. e., pores plus solid particles. It is in the determination of the volume of the mass that the two direct methods above mentioned are differen¬ tiated. First method, Volumeter .—After complete saturation, the brick is placed in the volumeter and its volume determined in cubic centimeters. W—S By the formula 100 (-) where W=weight of oil taken up by the Y brick, S=the specific gravity of the oil, and V=the volume of the brick, there is expressed the part of 100 unit volumes of the brick as a whole which consists of open pore. In other words, it is the percentage por¬ osity. By referring to Table I it will be noted that the percentage of varia¬ tion in the porosity determination was relatively small. Since the data given in Table I represents determinations made on 60 bricks of each clay 47.5 as the maximum, 0.6 per cent as the' minimum, and 11.5 as the average percentage variation, is considered as being excellent. These percentages of variation in results are not surprising in view of the fact that an error of lcc. in determination of volume, or an error of 1 gram in obtaining either the dry or saturated weight, makes a dif¬ ference of 0.3 in the porosity. It is obvious, therefore, that when the dry weight of the bricks are obtained they must be absolutely dry, i. e., oven dried at 120°C. so as to expel all of the hygroscopic water. This was not done in obtaining the data given in Table I. Second method, Chemical Balance .—When the porosity of the brick is determined on a chemical balance the volume of the briquette is found by the apparent loss of weight of the briquette when suspended in the oil. The briquette appears to lose weight when thus suspended, and this loss of weight is equivalent to the weight of a quantity of oil 142 PAYING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 equal to that of the briquette. This method was used by Dr. E. R. Buckley in the test on the Wisconsin clays, and the porosity calculated by him using the formula: looi ^ — w D ) Sp. gr. —_ == r cen ^ p 0r08 ity. I(W-D) Sp. gr.+D In this formula W=saturated weight; D=dry weight; and Sp. gr. the composite specific gravity of the clay particles, as calculated from dry, saturated, and suspended weights of the briquette. This formula, however, can be simplified by substituting for the value of D in the denominator its value in terms of the Sp. gr. and suspended weight (S) as given in the formula for specific gravity where D=D (Sp. gr.)—S (Sp. gr.). rw— d) The Buckley formula then simplifies to the expression 100 k-1= [w-sj porosity. This formula holds true no matter what liquid is used in the saturation of the brick, so long as the same liquid is employed in ob¬ taining the suspended weight. The method is accurate but very slow and tedious, unless it is carried out with small pieces on the jolly balance. If a jolly balance is used in this determination, the weight of the bri¬ quette or piece must not exceed that which would stretch the spring be¬ yond its elastic limit. If any other than a light weight spring is used the difference between the several readings will not be sufficient to permit of very accurate determination. This method was used in the determina¬ tion of the rate of vitrification, which will be described under the gen¬ eral heading of “Pyro-Chemical Tests/ 5 so will not be discussed in de¬ tail at this time. Third Method, Calculation .—It has been noted that the specific grav¬ ity of the powdered clay by the pycnometer method is uniformly higher than that calculated from data obtained on the green bricks. It has also been noted that this difference between the specific -gravities is due to the incomplete saturation of the brick. Since the formula for speci¬ fic gravity is: Dry weight (W) divided by the combined volume of the W particles (V) or —=Sp. Gr., the true volume of the particles in the Y brick can be obtained bv the formula: Dry weight divided by the pycno- D meter specific gravity, or-=Yol. Then the volume of the whole Sp. Gr. brick (V b ) minus the volume of the clay particles (V c ) would give the volume of the pore spaces (V p ), or V b —Y C ==Y P ’. To obtain the frac¬ tional amount of pore space in a brick, the volume of the pores (Y p ) PURDY] QUALITIES OF CLAYS FOR MAKING PAYING BRICK. 143 yp yb_yc must be divided by the volume of the brick: or —. But since-= yb yb f V c ] . W — we have 100 ] 1-J=per cent pore space where V c = - V b l V b J Sp. Gr. The economy and accuracy in determining porosity by this method lies in the fact that it is not necessary to saturate the brick and obtain the saturated weight. It is obvious, therefore, that the bricks would either have to be' partially saturated or covered with a thin coating of paraffin and their volume determined in a volumeter. Without a volu¬ meter this method cannot be used. If the specific gravity has been determined by the pycnometer method and a volumeter is not accessible, the porosity is best calculated by the Buckley formula. In this, however, complete saturation of the brick must be assured, and the true specific gravity of the clay particles used. Neither the Buckley method nor the indirect method here proposed is usable on any other than a green or unburned lump of clay. For the [W-Dl porosity of a burned lump or briquette, the formula 100 ■{ -V is [W—SJ the only one that will give accurate results, as will be shown under the discussion of Pyro-Chemical and Physical Properties of Clays. In the following table are given porosity data obtained, first, by the usual volumeter method without taking into account the hygroscopic water; second, by the indirect method described above, without taking into account the hygroscopic water, and third, by the indirect method on a basis of absolute dryness of the bricks. The percentage of increase of porosity obtained in the second and third instance, over that obtained in the first is also shown. 144 PAYING BRICK AND PAYING BRICK CLAYS. [bull. no. 9 Table IV — Showing the percentage of errors in the Seger Volumeter method of determining porosity as customarily executed. Calculated results are by the indirect method. Sample and Brick Number. Dry weight in grains— Volume in cubic centimeter. Porosity by volumeter on air dried briquettes. Porosity by calculation on air dried briquettes. Percentage increase in poro¬ sity by calculations over that by volumeter. Porosity by calculations on oven dried briquettes. Percentage increase by cal¬ culation on overdried briquettes over that by volumeter on air dried briquettes. Air dried. Oven dried. K 3—34. 571.5 556.4 300.2 23.0 29.12 26.61 31.0 34.78 K 3-37. 572.5 558.0 297.8 24.0 28.43 18.45 30.24 26.00 K 3-39. 601.3 587.7 311.5 23.8 28.13 18.19 29.76 25.04 K 5-13. 644.9 639.3 329.5 25.6 26.17 2.22 26.81 4.73 K 5—15. 647.1 640.7 330.6 25.4 26.17 3.03 26.89 5.87 K 5-17. 658.4 652.7 337.4 25.5 26.39 3.49 27.03 6.00 K 7- 1. 607.4 596.6 327.7 27.3 29.69 8.76 30.94 13.34 K 7-3. 586.5 574.0 312.1 26.3 •28.71 9.16 30.23 14.94 K 7-5. 600.4 587.5 323.4 26.9 29.57 9.93 30.08 11.82 K 8—19. 663.7 652.4 336.6 24.3 26.67 9.75 27.92 14.89 K 8—24. 631.2 620.7 320.7 24.1 26.81 11.24 28.03 16.31 K 8—31. 621.5 611.7 316.3 24.1 26.93 11.75 28.08 16.51 K10— 1. 545.5 533.2 284.6 25.0 28.64 14.56 30.25 21.49 K10— 2. 539.9 523.5 281.6 24.7 28.62 15.87 30.79 24.66 K10— 4. 532.2 520.0 278.8 25.7 .28.94 12.60 30.56 18.91 K13-44. 622.4 607 9 323.6 26.6 28 63 7.63 32.10 20.67 K13-46. 600.7 587.8 310.3 26.6 28.17 5.90 29.71 11.69 K13-53. 619.5 605.9 322.0 28.4 28.61 6.27 30.18 7.40 H18— 1. 639.2 625.3 315.6 21.8 24.17 10.87 25 82 18.44 H18- 3 .. 637.0 624.0 314.3 22.1 24.12 9.14 25.67 16.16 H18— 5. 642.9 629.6 317.1 21.7 24.09 11.01 25.67 18.30 H20— 1. 583.0 567.4 300.1 23.0 28.49 23.87 30.40 32.18 H20— 3. 588.2 576.3 303.2 23.0 28.59 24.30 30.03 30.56 H20— 4. 579.4 563.5 296.6 23.0 28.09 22.13 30.06 30.70 H24— 1. 601.0 591.4 291.4 17.5 22.38 27.88 23.62 34.97 H24— 3.. 650.5 640.0 314.4 17.6 22.13 25.74 23.38 32.84 H24— 5. 600.5 591.5 292.9 18.3 22.84 24.81 23.99 31.09 R 3—1. 687.1 672.9 344.3 23.2 24.83 7.03 26.39 13.75 R 3—3. 667.0 653.9 336.3 23.0 25.30 10.00 26.76 16.35 R 3-5. 690.9 678.7 348.5 23.1 25.33 9.65 26.65 15.37 R 4—26. 705.6 689.2 348.7 20 3 25.64 26.31 27.36 34.78 R 4-29. 678.1 663.0 337.0 21.2 26.05 22.88 27.70 30.66 R 4-36. 716.1 697.8 352.0 20.3 25.24 24.34 27.15 33.74 S 1-1. 575.5 545.7 309.5 21.2 29.70 40.09 33.34 52.55 S 1—3. 573.0 541.5 307.2 20.0 29 48 47.40 33.36 66.80 S 1-5.;.... 565.5 540.0 303.8 21.5 29.62 37.77 33.80 57.20 L-II-11. 599.0 584.2 322.6 26.0 31.35 20.57 33.03 27.04 L-11-13. 612.2 591.4 326.7 23 9 30.72 28.53 33.10 38.40 L-II—15. 602.3 584.4 324.8 25.6 31.44 22.81 33.47 30.74 G-II-55. 708.7 702.6 349.7 22.5 24.67 9.64 25.32 12.53 G-II-57. 705.3 696.8 347.8 22.7 25.03 10.26 25.93 14.23 G-IT-58. 712.3 704.0 349 6 22.7 24.68 8.27 25.55 12.56 The data in table IV shows the inaccuracy of the nsual method of de¬ termining the porosity in dried clay wares. It has been stated 1 that three to six honrs is sufficient to saturate with oil unburned briquettes that measure 3x1 inches. Forty-eight hours was therefore considered ample time in which to saturate a brick that cubically was about eight times as large. From the fairly close agreement in the specific gravities as determined by the pycnometer and the volumeter, it was thought that the briquettes had been fairly well saturated. Such, however, was evi¬ dently not the case. Iowa Geological Survey, Vol. 14, p. 18. PURDY] QUALITIES OF CLAYS FOR MAKING PAVING BRICK. 145 RELATION OF RATE OF ABSORPTION TO POROSITY. Aside from exposing the irregularities in our method of analysis, this data gives evidence of the lack of relation of total porosity and rate of absorption in the green or unburned bricks. Since all the bricks were subjected to the same oil immersion treatment, it must follow that clays differ in the rate at which they can be saturated, and that this rate is not wholly a function of porosity. The writer is not aware of tests ever having been made to investigate this property, which we may call “absorption ratio," but its significance in connection with the drying behavior of clays is obvious. Johnson 1 has said, “Obviously, too, the quantity of liquid in a given volume of soil affects not only the rapidity, but also the duration of evaporation. The following table, by Schubler, illustrates the peculiar¬ ities of different soils in these respects. The first column gives the per¬ centages of water absorbed by the completely dry soil. In these experi¬ ments the soils were thoroughly' wet with water, the excess allowed to drip off, and the increase of weight determined. In the second column are given the percentages of water that evaporated during the space of four hours from the saturated soil spread over a given surface." TABLE V. Percent. 1 Per cent. Quartz sand. . 25 88.4 Gypsum. 27 71.7 Fine sand. 29 75.9 Slaty marl. 34 68 0 Clay soil (60 % clay). 40 52.0 Loam. 51 45.7 Plough land. 52 32.0 Heavy clay (80 1 clay). 61 34.9 Pure gray clay. 70 31.9 Fine carbonate of lime. 85 28.0 Garden mould. 89 24.3 Humus. (81 25.5 Fine carbonate of magnesia. 256 10.8 “It is obvious that these two columns express nearly the same thing in different ways. The amount of water retained increases from quartz sand to magnesia. The rapidity of drying in the air diminishes in the same direction." Johnson affirms 2 that “these differences—(in the imbibing power of clays)—are dependent mainly on the mechanical texture or porosity of the material." That Johnson’s statement, when applied to unburned bricks, is incorrect, is shown by the data in table IY. That there are other factors affecting the difference in rate of absorption and evapora- - tion in different clays is quite evident. Value of the Porosity Determination on Raw Clay Lump .—It has been contended at various times in ceramic literature that a porosity de¬ termination on a raw lump of clay would give evidence, concerning such properties, as slaking, weathering, amount of water required to de- 1 Loc. Cit., p. 18. 2 How Crops Feed, p; 175. —10 Gr 146 PAVING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 velop plasticity, etc., and thus indirectly the shrinkage. Such claims have‘never been based on data, nor are they substantiated by the data secured by this Survey. As will be shown in later paragraphs, neither data nor sound reason would warrant such statements. Value of the Porosity Determination on Green Brick .—Before the value of knowing the porosity of a green brick can be discussed it is neces¬ sary to show the correlation of that property with those which it affects. If porosity, fineness of grain, and drying behavior are in any degree re¬ lated functions, curves plotted from data should show such relations. Such a relation is shown in Fig. 4 where there seems to be an inverse ratio between fineness of grain in surface or loose grained clays, and the porosity of the green or unburned brick. The data from which this Fig. 4. Curve showing relation between porosity and fineness of grain. (From data of Beyer and Williams, Iowa Geol. Surv., Vol. 14, p. 123.) curve was plotted was obtained from the work of Beyer and Williams 1 . The surface factor was calculated from their data by the method given on page 113. The porosity data and calculated surface factor are as follows: TABLE IV. Porosity. Surface factor. Palp Rrirt Cn . 18.14 136.40 Gethmaun Bros.■. 22.43 114.38 L. C. Besley (bottom). 24.03 119.98 L. C. Besley (middle). L. C. Besley (top). 25.30 29.77 100.48 74.77 This reciprocal relation between fineness of grain and porosity could be taken as evidence in proof of the close relation of fineness of grain and porosity of the green brick to drying, shrinkage and other properties that are peculiar to wares manufactured from fine but loose-grained clays or mixtures. The writer hesitates, however, to affirm the truth of 1 Iowa Geol. Surv., Vol. XIV, p. 123. PURDY] QUALITIES OF CLAYS FOR MAKING PAYING BRICK. 147 such a relation from evidence obtained on a few samples of a single type of clay. Observation of the working behavior of boulder clays in build¬ ing 1 brick manufacture does not lead one to believe in such an exact rela- tion. In the manufacture of bricks by the ordinary dry process—where there is present only from 6 to 12 per cent of mechanical water, and the grains of the clay are not surrounded by slippery media that permit the particles to slide easily and freely upon one another—the clay cannot be formed into as compact a mass as when there is sufficient water present to permit of manufacture by the stiff mud process. If dry pressed bricks be formed in a hammer machine, or press, where the brick is subjected to repeated blows by a heavy hammer, the clay particles, even though nearly dry, would be forced over one another until the mass assumes a much closer or denser structure than is possible by the ordinary dry press process. The difference in structure, and its consequent effect on the burning properties of dry press bricks manufactured by these two methods in the St. Louis district is more evident than the difference between the structure of the dry and stiff mud or stiff mud and soft mud bricks. When the grains are made to lie closer together, either by strong force, or by a lighter force supplemented by a floating medium, better opportunity is offered for the grains to fuse with one another. This is shown nicely by the fact that hammered brick can be burned in the colder parts of a kiln to. a degree of hardness that is equal to and often exceeds the hardness of a brick made from the same clay by-the dry press method and burned in the hotter portions of the kiln. In the illustrations just cited, the difference in the burning properties of bricks made from the same clay by different processes, has been used as a means of noting that porosity of green brick is not wholly a func¬ tion of size of grain. It is clear that it is also very largely a function of process of manufacture. While there may be some relation between the size of grain of loose or soft clays and the porosity of the brick manufactured from them, it Fig. 5. Diagram showing the relation between porosity of green shale brick and the absolute fineness of grain. (From data given in Table I.) 148 PAYING BRICK AND PAYING BRICK CLAYS. [BULL. NO. 9 is still doubtful if a similar relation can be observed in the hard rock¬ like fossil clays, such as shales, where the mineral particles are so cement¬ ed as to very stubbornly resist separation by the crushing force of dry pan mullers as well as the disintegrating influence of the water used in pugging. In Fig. 5 are shown curves plotted from data obtained with shales in the same manner as the data for surface clays in Fig. 4. The porosity is taken from Table I, and the surface factor from Table VIII. The clay from which these shale brick were made had been crushed to pass through a dry pan and then screened. In the laboratory they were known as dry pan samples. These “dry pan samples” were then in the same state of mechanical subdivision as the clay used by the manu¬ facturer. In making the bricks from which the data in Table I were obtained, considerable time was expended in pugging or wedging the clays by hand, first in a large bulk, and later in quantities just sufficient for one brick. The operator batted a quantity of clay that would make approxi¬ mately 60 bricks tyz” on a plaster top table until it was as compact as he could make it. Then by use of a trowel in some instances and a wire in others, he cut off from .the large mass a quantity sufficient to fill the die of the press. This smaller piece was again thoroughly wedged by hand until all air blebs had been worked out and the whole took on the shape of a compact loaf. This loaf was then placed in the die, using care to see that it cleared the sides so as to prevent a shearing off of any portion of the loaf on the edge of the die when the plunger descended. The loaves were pressed into bricks on a slow screw tile press, so that the clay did not receive much compression, but yet sufficient to cause it to flow in shreds up around one side or another of the plunger. From this flowage of the clay past the plunger, together with the unusual amount of wedging by hand, it was considered that the clay had been sub¬ jected to a treatment that was approximately comparable to the pugging it would have received in the factory, so that the data as given in Table I show approximately the physical structure except for lamination that would be developed on a regular manufacturing basis. It is commonly known by paving brick manufacturers that some shales require inordinate pugging before they develop sufficient plasticity to permit the production of a perfect bar in the die of the brick machine. In fact it is not uncommon to see a battery of two pug-mills through which the clay must pass before it enters the brick machine proper. In the brick machine the clay receives further pugging before it issues as a bar from the die. It is also not uncommon to hear the manufacturers claim that they cannot pug clay sufficiently unless they use hot water. Not all manufacturers have to resort to this extra care in the pugging process, for some shales develop plasticity with sufficient readiness to allow them to emerge from the first pug-mill in a workable condition. This same difference in the working property of the various shales was perhaps more noticeable in the laboratory than in the factories. This difference in the working properties of shales is considered to be due to the fact that the grains of clay are cemented by substances that differ in their solubility in water. It is now well, known that soils and purdy] QUALITIES OF CLAYS FOR MAKING PAYING BRICKS. 149 clays contain soluble salts that are adsorbed by, or, to use a more homely expression, smeared over the particles, and are not easily extracted by water. It has been learned by experiment that clays can take on or adsorb soluble salts from solutions and so retain these' salts in their sub- microscopic pores that they cannot again under ordinary conditions be dissolved from the clay. The amount of water used in the pugging of shales is not sufficient to dissolve or loosen all of the cementing salts in a clay even by continued pugging, so that at best, only a portion of the clay particles are separated from one another, but the manufacturers must continue the pugging until a sufficiently large number of grains are separated to form a slippery medium, by virtue of which the unslaked or undisintegrated bundles of particles can slip past one another freely enough to permit a flowage of the mass under pressure. The difficulty encountered by manufacturers in breaking down the cementing bond in shales is increased many-fold when an attempt is made to disintegrate a clay into its ultimate grains, as is done in mechanical analysis. The data for texture or size of grain used in plotting the curves in Figs. 4 and 5 were obtained by mechanical analysis and are supposed to represent the subdivision of the clays into their ultimate particles. While it is comparatively easy to obtain separa¬ tion of the particles in loose-grained clays in the laboratory and in the factory, it is obvious that it is not possible to obtain a similar separation of the particles of the hard rock-like clays in the factory, and very diffi¬ cult to obtain much more than an approximation of ultimate subdivision in the laboratory. It is owing to this indefinite degree of solution of the natural bond in pugging that we have, in the case of shale bricks, a dis¬ cordant relation between the porosity of the brick and the fineness of grain, shown in Figs. 2 and 5, as contrasted with the semingly concord¬ ant relation in the case of the loess bricks, as shown in Fig. 4. Although a porosity determination on a green brick may not be of value as direct evidence of the so-called “working properties” of a clay, it can be shown that it is of indirect value, in that the data can be used as the basis of many interesting and valuable calculations. For practical demonstration of the commercial possibilities, or exposition of the work¬ ing properties of a clay, the writer has failed to find wherein porosity data on unburned bricks separately considered are of use. Fineness of Grain. DEFINITION. By fineness of grain or texture of a clay is meant the size of its min¬ eral particles. Experimental evidence indicates that variation in grain controls many of the physical and pyro-chemical properties exhibited by clays. Plasticity, shrinkage in drying and burning, tensile strength, drying properties, rate of oxidation, rate of vitrification, toughness of burned ware, and finally, to some extent pyrometric value of the clay, are all influenced by fineness of grain. 150 PAYING BRICK AND PAYING BRICK CLAYS. [BULL. NO. 9 The grains of many clays are so cemented that they resist separation in the ordinary png-mill or blnnger. When two or more particles are thus cemented they operate as a unit in their influence upon plasticity, tensile strength, drying behavior, etc. This accounts, in part, for many of the apparent exceptions to the general rules deduced from experiment¬ al evidence, for, in the usual methods applied for determining fineness of grain, special effort is made to separate the particles completely. This raises the question whether separation of the particles should be carried to such extremes when attempting to trace direct relations be¬ tween fineness of grain and the physical properties developed in the process of manufacture of clay into wares. On this point, however, we have no direct evidence, except perhaps as shown in Figs. 4 and 5, so the question will have to remain unanswered for the time being. It is known, however,, that it would be well-nigh impossible to determ¬ ine how far a mechanical separation of the particles should be carried in the laboratory to make the test comparable to the separation effected in the pug-mill, wet pan, or blunger. For this reason it would seem as though the most useful data concerning texture or fineness of grain can¬ not be obtained by the present method of analysis. MEANS OF EXPRESSING FINENESS OF GRAIN. If all the particles of clay were considered as being spheres or cubes their superficiaTareas would be inversely proportioned to their diameters. The following calculations show this to be true in regard to the sphere: D 3 Volume of a sphere is equal to Pi —; then if D and d are the diameters 6 PiD 3 Pid 3 of two spheres their volumes would be proportional as-:- 6 6 . The number of spheres required to equal in volume a standard unit PiD 3 6 6 volume would be 1 -or - in the one case, and - in the Pi D 3 Pi d 3 other. Since the surface of a sphere of each size is equal respectively to Pi D 2 and Pi d 2 the total surface area of a collection of spheres, having a 6 6 total volume equal to unity, would be in each case-x D 2 and- Pi D 3 Pi d 3 6 6 x d 2 or-and-respectively. The combined areas of each group PiD Pid rf spheres occupying the same volume but having different diameters are, therefore, inversely proportional to their diameters. This proportional relation of the surface of the particles in the several groups is taken as the surface factor of the respective groups, and the sum of these as the surface factor of the clay. Cushman 1 has shown the error involved in thus taking the mean of the extreme diameters in a given group. According to data given by Cush- 1 Air Elutriations of Fine Powders, Jour. Am. Chem. Soc., Vol. XXIX, No. 4, p. 589, April 1907. PURDY] QUALITIES QF CLAYS FOR MAKING PAVING BRICKS. 151 man, a mechanical analysis of the separate groups would show a predom¬ inance (77 to 87 per cent in Cushman data) of the finer particles of that group. That the mean diameter obtained as described above, is not a true mean of the diameters of the particles in a group, is obvious. The error thus involved cannot, however, be obviated without a much more exten¬ sive subdivision of the groups than is possible under ordinary conditions. It needs no mathematical demonstration to make clear that, theoretically, the more extensive the analysis, the more accurate would the results be. It needs but a short experience with the mechanical analysis by any of the hydraulic methods, to learn that, practically, the more extensive the analysis is made, the larger will be the operating errors. In making a mechanical analysis one must choose between the Scylla and Charybdis of these errors and, naturally, will decide in favor of that one which involves the making of the fewest determinations. In this report the mean of the extreme diameters of each group, irre¬ spective of the distribution by number according to their volume,, of the particles within the respective groups is taken as representing the diam¬ eter of the group. The mean diameter of each group and total surface factors for the clays here reported are shown in Table Y. VALUE OF DETERMINATION OF FINENESS OF GRAIN. As before stated, fineness of grain is the probable cause of several of the other properties exhibited by clays. Since fineness of grain is the cause, and the other properties, in a large sense, the effects, the true significance of this determination can be best discussed by dealing sep¬ arately with the properties induced by size of grain. Numerical Results .—In Table YII is given the per centage by weight of calcined materials in each of the several groups according to sizes of particles. TABLE VII. Sample Number. Hygro¬ scopic water. Com¬ bined water. Percentage Amount by Weight of Particles, Grouped According to Diameters. Total. 1 M. M. l-.IM.M. .1-.01. .01-.001. .001-0.0 K 1. 0.47 5.73 6.92 6.19 54.24 22.92 7.87 104.36 K 2. 1.03 3.77 0.96 1.14 63.75 18.04 13.33 102.04 K 3. 0.97 6.90 1.42 1.47 54.38 23.03 12.00 100.18 K 4. 1.68 5.43 1.30 1.66 46 47 27.76 19.34 104.68 K 5. 1.10 5.60 5.91 1.04 58.01 21.04 9.69 102.41 K 6. 0.70 4.76 1.14 1.74 63.17 23.49 7.62 102.65 K 7 . 0.67 5.46 1.14 3.42 58.82 24.45 9.75 103.73 K 8.. 1.20 7.40 8.76 6.55 45.95 22.48 8.26 100.61 K 9. 0.83 3.32 11.03 1.49 63.80 13.79 6.53 100.83 K 10. 1.74 5.52 0.85 2.09 22.96 40.72 23.93 97.84 K 11. 1.78 8.32 3.02 2.69 40.44 33.76 11.13 101.15 K 12. 1.48 8.66 3.64 2.21 38.03 35.83 12.13 102.00 K-13. 0.92 6.13 1.60 0.79 44.70 38.80 11.37 104.33 K 14. 0.66 5.08 13.66 5.76 41.24 24.34 8.14 98.91 R 1. 1.33 7.69 1.76 5.90 36.59 36.23 12.97 102.49 R 3. 1.10 5.29 10.92 5.84 50.67 20.52 9.93 104.30 R 4. 0.82 6.11 9.47 2.54 47.69 26.78 8.98 102.43 G-II. 1.35 4.28 12.67 2.38 50.78 20.19 12.80 104.48 I-II. 2.01 4.35 4.14 3.62 44.37 25.24 17.53 101.30 H 18. 1.69 14.95 12.14 12.14 24.57 20.88 17.60 103.99 H 20. 1.98 11.97 1.00 1.86 39.73 29.73 16.69 102.79 H 21. 1.63 12.26 0.19 0.50 22.50 39.21 26.39 102.71 H 23.. 2.48 7.71 1.60 2.56 28.11 38.20 21.89 102.57 152 PAYING BRICK AND PAYING BRICK CLAYS, [bull. no. 9 In Table VIII will be found the same data with the hygroscopic water eliminated and the chemical water distributed over the various groups proportionally to the amount belonging to each, as determined by their loss on ignition. TABLE VIII. Sample Number. 1 mm mean diam. 1.25. 1mm mean diam. 0.5. 0.1-.01 mean diam. 0.05. .01-.001 mean diam. 0.005. .001-0 mean diam. 0.0005. Total. Surface factor K 1. 7.27 6.53 56.07 24.86 9.76 104.51 256. K 2. 1.07 1.23 66.24 19.63 . 13.90 102.09 331. K 3. 1.50 2.41 57.15 25.14 13.96 100.18 341. K 4. 1.40 1.74 48.87 29 41 22.24 103.68 514. K 5. 6.38 1.46 60.57 22.93 11.43 102.80 287. K 6. 1.24 1.83 65.83 25.98 7.77 102.66 221. K 7. 1.35 3.75 60.87 25.89 11.81 103.69 300. K 8. 9.66 6.90 48.46 25.40 10.05 100.50 262. K 9. 11.39 1.55 65.50 14.72 7.63 100.80 195. K 10. 1.06 2.42 24 62 44.29 25.52 97.91 604. K 11. 5.36 3.76 43.74 35.45 . 12.94 101.26 339. K 12. 4.49 2.88 40.51 38.82 15.32 102.04 403. K 13. 1.82 1.35 46.74 41.73 13.15 104.80 356. K 14. 14.23 6.31 42.75 26.03 9.67 98.99 254. R 1. 2.16 6.51 38.70 39.32 15.53 102.25 397. R 3. 11.69 6.30 52.90 21.60 11.79 104.29 291. R 4. 10.15 2.84 49.32 29.13 10.85 102.31 275. I-II. 4.64 3.81 45.50 25.94 21.40 101.31 489. H 18. 13.05 17.71 27.57 26.58 19.22 104.16 444. H 20. 1.92 2.75 42.01 32.47 23.97 103.13 553. H 21. 0.34 0.80 24.34 42.77 34.62 102.89 783. H 23. 1.80 2.86 29.95 40.82 27.30 102.74 634. G 2. 13.17 2.47 52.57 20.57 15.72 104.52 366. In Table IX is given the calculated loss on ignition of each group as nearly as it could be determined from the results of analysis. While this loss on ignition has been called “Combined Water,” it must be borne in mind that the loss of many substances other than combined water has been included. Carbon, carbonic acid, sulphur, etc., are driven off on ignition and reduce the weight of the sample. The relations referred to are well expressed by the well-known Kennedy curves. (See Fig. 6.) Table IX. Distribution of combined water over the several groups of particles. Sample Number. 1 M. M. 1-.1M. M. .1-.01 .0—.001 .001-0 1 Total. K 1. 0.32 0.31 1.62 1.64 1.86 5.77 K 2. 0.09 0.08 1.85 1.37 0.43 3.84 K 3. 0.06 0.92 2.28 1.88 1.86 7.03 K 4. 0.07 0.05 1.66 1.25 2.69 5.74 K 5. 0.31 0.41 1.98 1.67 1.64 6.03 K 6. 0.09 0.08 2.22 2.33 0.09 4.82 K 7.. 0.22 0.31 L69 1.29 2.00 5.52 K 8. 0.82 0.30 2.01 2.69 1.71 7.55 K 9. 0.27 0.03 1.18 0.82 1.02 3 34 K 10. 0.19 0.28 1.24 2.83 1.15 5.71 K 11. 2.27 1.03 2.58 1.08 1.61 8.58 K 12. 0.80 0.63 1.95 2.49 3.02 8.91 K 13. 0.20 0.55 1.68 2.61 1.68 6.74 K 14. 0.48 0.33 1.26 1.54 1.48 5.11 R 3. 0 66 0.39 1.72 0.87 1.75 5.40 R 4. 0.60 0,27 1.45 2.13 1.80 6.27 R 1. 0.39 0.55 1.74 2.73 2.43 7.87 I -1. 0.40 0.17 0.24 0.19 3.51 4.53 G -2 . 0.33 0.06 1.13 0.11 2.75 4.39 H 18. 0.79 5.45 2.76 5.49 1.44 15.95 H-20. 0.89 0.85 1.52 2.17 6.95 12.41 H-23. 0.16 0.23 1.13 1.67 4.89 8.11 Fig/ 6. Kennedy curves showing the reactive rate of loss on heating calcareous and non-caicereous clays. (After Bleininger, Geol. Surv. Ohio, 4th Ser., Bull. 4, p. 19.) PURDY] QUALITIES OF CLAYS FOR MAKING PAYING BRICKS. 153 154 PAYING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 5> The distribution of “combined water” over the several groups, as given in Table I, was necessarily calculated by proportion, for the total loss on ignition of the finer groups in some cases amounted to two or three times that which occurred on ignition of the whole sample. Satisfactory explanation of this increase or gain in volatile matter during the process of analysis cannot be given. It is supposed, however, that it is due in part to some organic growth developed in the water, or, possibly, oil from the compressed air that was used in the siphoning off of the supernatant liquid. That this last suggestion will not account for all of this increase, if any, in the volatile matter accumulated in process of analysis, was proved by the fact that when precautions w£re taken to clear the air of all possible traces of solid material, there was still nearly the same increase. There is therefore considerable doubt as to the value of re-dis¬ tribution of the loss on ignition by means of proportions, yet the data obtained in this way are considerably more accurate than they would otherwise be. The irregularities in the data are pointed out solely to call attention to a weak point in this most important determination. Mechanical analysis of clays, as has been stated before, bids fair to become a very essential test in determining the full value of a clay, and attention should be given to the elimination of this increase in volatile matter during the process of analysis. Soil physicists are experiencing the same difficulty, and yet they have learned to give considerable, in fact, a large amount of credit to the mechanical analysis of soils as a means of determining its proper¬ ties for their purposes . Shrinkage in Drying. METHODS OF MEASUREMENT. The amount that a clay will shrink in drying is expressed in per cents of the unit length or volume. In the first instance the shrinkage would be designated as linear shrinkage, and in the second, as volume or cubical shrinkage. In Table I, page ? ?, will be found the percentages of both linear and volume shrinkage for several shale clays as determined by direct meas¬ urement. It will be noted that the variation in linear shrinkage in 60 bricks of each clay is far in excess of reasonable limits. When the linear shrinkage varies from 32 to 133 per cent from the average, the data must be wholly unreliable. In presenting this data it is felt that the failure to produce more consistent results lies in part in the shortness of the shrinkage distance, and in part in carelessness of the operator. In marking the freshly made bricks a stencil devised by J. F. Ivrehbiel was used, so that initially the shrinkage lines were marked upon the brick with accuracy. In measuring the decrease in length of the shrinkage line after the bricks were dried, a vernier shrinkage scale was used that read accurately to the third place. The large variations in the results were therefore a surprise to the operator. PURDY] QUALITIES OF CLAYS FOR MAKING PAYING BRICKS. 155 The volume shrinkage varied within fairly reasonable limits, but even here the variations are quite large considering the size of the bricks used. It is felt that if in one case the variation could be only 0.5 per cent there ought not to be any excuse for a variation of 33.8 per cent in another or an average on all samples of 11 per cent. Inasmuch as the volume shrinkage data proved to be the more accur¬ ate of the two they were used as a basis on which to calculate* the linear shrinkages as shown in the following table: Table X. Comparison of the measured with calculated linear shrinkage. Sample No . Average volume shrinkage in per cents. Calculated linear shrinkage in per cents. Average measured linear shrinkage in per cents. Percentage variation on volume shrinkage. Percentage variation in measured linear shrinkage. K 1 . 6.2 2.1 1.5 33.8 133.3 K 2 . 12.2 4 3 3.5 2.4 70.0 K 3 . 10.5 3.6 2 1 16.7 68.0 K 4 . 10.1 3.5 3.3 5.8 73.6 K 5 . 5.2 1.8 1.6 6.7 129.0 K 6 . 10.1 3.5 4.1 10.3 43.9 K 7 . 9.6 3.3 3.9 12.2 51.2 K 8 . 7.5 2.6 2.1 21.5 95.2 K 9 . 3.5 1.2 0.9 34.1 75.0 K 10 . 18.3 6.5 5.8 0.5 37.8 K 11 . 13.5 4.7 3.3 16.3 73.8 K 12 . 12.7 4.4 3.6 5.9 44.4 K 13 . 10.5 3.6 3.3 7.5 48.4 K 14 .. 6.1 2.1 1.5 11.5 93.3 S 1 . 12.9 4.5 2.7 19.0 60.0 S 2 . 13.1 4.6 4.2 10.6 42.8 R 1 . 13.9 4.9 4.5 5.7 53.3 R 2 . 9.1 3.1 3.3 25.3 36.3 R 4 . 6.1 2.1 3.2 6.7 56.2 B II. 11.5 4.0 5.0 7.8 32.0 C II. 7.3 2.5 1.9 12.5 124.0 H II. 14.3 5.0 5.5 7.6 65.4 I II. 13.8 4.8 4.2 2.9 33.3 J II. 14.4 5.1 4.6 4.1 34.7 L II. 9.7 3.4 3.7 6.8 54.0 H 16 . 7.8 2.7 2.8 19.6 78.5 H 17 .. 21.4 7.7 7.0 4.6 34.2 H 20 . 16.5 5.8 6.8 3.0 26.5 H 21 . 18.0 6.4 7.2 1.7 41.6 H 23 . 20.4 7.3 7.4 5.8 43.2 H 24 .. 11.4 4.0 4.0 10.5 45.0 * If a unit cube shrinks so that each edge is decreased by linear length “a” then the new length of the edges become (1-a). If the decrease in volume of this same cube be represented by “x” then the new volume will be (1-x). Since the edges of the cube are now (1-a) its volume can also be represented by (1-a)3 hence (1-a)3 is equal to (1-x), or a=l-Vl-x. It was by this form¬ ula that the transformation from volume to linear shrinkages were made. 156 PAYING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 The linear shrinkage which probably is the more correct for that sample is underscored. In cases where there is not an underscored linear shrinkage, there is no possible way to judge which one is the most correct. In case the calculated practically agrees with the measured linear shrink¬ age, both are underscored. If the volume and linear shrinkages had been correctly measured, there would have been no discrepancy between the calculated and determined linear data. If any importance at all is to be attached to shrinkage data it is evident that extreme care should be exercised in their determination. When possible, the measured linear should be checked by calculation from the volume shrinkage and vice versa. RELATION OF VOLUME SHRINKAGE TO POROSITY. It will be noted from a glance at Fig. 7 that there does not seem to be any relation whatever between volume shrinkage and pore space in the dried bricks. Fig. 8 also represents the same sort of irregular relation between the volume shrinkage and pore space in the dried brick made from the Iowa loess clays. 1 RELATION OF VOLUME SHRINKAGE TO WATER OF PLASTICITY. The chart, (Fig. 9), showing the relation between the percentage of water of plasticity and the volume shrinkage from the green to the dried condition, proves that while there is some indication of a reciprocal relation between these two factors, this relation cannot be affirmed. o H17 O H23 11210 c Kioo U20 Rio ° HI OK12 °H-ll cKll Si c c-t "i V o 0 52 o H24 RJ( OK 3 ) OL-il oB-II A'40 0 H16 oK 13 OK 6 K7 o C-ll oKS oKli < OK5 >K1 5K9 PERCENT POROSITY OF GREEN BRICK Fig, 7. Diagram showing relations between volume shrinkage and 'porosity of dried brick. (From data in Table I.) 1 la. Geol. Surv., Vol. XIV, 1904,p. 109 and 113. PURDY] QUALITIES OF CLAYS FOR MAKING PAYING BRICKS. 157 RELATION OF YOLUME SHRINKAGE TO WATER IN EXCESS OF THAT REQUIRED TO FILL THE PORES. It would seem that if the volume had been determined at regular in¬ tervals as the bricks lost their mechanical water by evaporation, the per centage up to the time that the brick reached its maximum shrinkage, n? r>« om 0 13 19< i 270 170 031 no O 32 02 nlO 07 ©3 O 22 Oi* 012 °26 020 4.0 °15 o?.i < >30 08 018 O 21 029 ( >5 < >1 - 1 '25 280 in PERCENT POROSITY Fig . 8. Diagram showing relation between volume shrinkage and porosity of loess clays from Iowa. (After Beyer and Williams.) would stand in closer relation to the volume shrinkage than does the total mechanical water and volume shrinkage. It is not known what value such a test would have, but it would probably be considerably more than is the determination of total mechanical water alone. In Table XI is shown the percentage by weight of water that would be required to fill the pores of bricks made from Iowa clays and that which is in excess of the “pore water.” These clays were ground until they would pass through a 40-mesh sieve, 1 then wetted with water and thoroughly wedged. Grinding the clay until it would pass a 40-mesh sieve would reduce the size of the larger grains, and to some extent break down bunches of grains by force that would not have been affected by the water used in wedging. The data is of interest on this account in connection with the problem of shrinkage. 1 la. Geol. Surv.. Vol.XIV,1904,p. 76. 158 PAYING BRICK AND PAYING BRICK CLAYS. [bull. no. 9 TABLE XI. Clay. Per cent Water. Porosity. Sp. Gr. Vol. of clay per 100. Per cent} by wt. of w r aterin pores. Per cent by w 7 t. of 1 clay. Excess water for plas¬ ticity. Flint Brick Co., top stratum. 22.5 17.31 2.41 82.69 7.99 92.01 14.51 Flint Brick Co., middle... 25.0 23.00 2.51 77.00 10.64 89.36 14.36 Flint Brick Co., bottom... 25.0 30.04 2.40 - 69.96 15.83 84.87 9.87 Flint Brick Co., green brick. 25.0 23.20 2.52 76.80 10.71 86.29 14.29 Iowa Brick Co., top stratum. . 17.5 17.43 2.53 82.57 7.52 92.48 9.98 Iowa Brick Co., second from top. 25.0 . 25.73 2.46 74.27 12.92 | 87.08 1 12.07 Iowa Brick Co., third from top. 27 5 29.57 2,40 70.43 15.32 84.68 | 12.18 Iowa Brick Co., fourth from top. 25.0 17.96 2.45 82.04 8.20 91.80 17.80 Iowa Brick Co., fifth from top. 25.0 26.04 2.37 73.96 12.93 87-07 12.08 Iowa Brick Co., bottom stratum. 27.5 28.69 2.36 71.31 14.76 85.24 12.74 Cap. C. Brick & P. Co., top stratum..'. 30.0 30.80 2.69 69.20 14.43 85.57 15.57 Cap. C. Brick & P. Co., second from top. 22 5 25.25 2.48 74.75 11.44 88.56 11.06 Cap. C. Brick & P. Co., third from top. 22.5 17.00 2.53 83.00 7.49 92.51 15.01 Cap. C. Brick & P. Co., fourth from top. 22.5 25.13 2.45 74.87 12.04 87.96 10.46 Jester Clay Bank. 20.0 20.35 2.49 79.65 9.30 90.70 10.70 Harris Brick Yard. 22.5 24.33 2.56 75.67 11.15 89.85 11.35 Dale Brick Co., top loess.. 22.5 18.14 2.44 81.86 8.32 91.68 14.18 Dale Brick Co., bottom shale. 22.5 28.98 2.48 71.02 14.13 85.87 8.37 Corey P. B. Co., red burn¬ ing clay. 25.0 30.10 2.54 69.90 14.49 85.51 10.51 Corey P. B. Co., buff burning clay. 25 0 28.10 2.54 71.90 12.86 87.14 12.14 Colesburg . 27.5 28.36 2.62 71.64 13.12 86.88 14.38 Storm Lake B. & T. Co ... 25.0 19.27 2.42 80.72 8 97 91.03 16.03 Besley Brick Yard, top loess. 22.5 29.77 2.34 70.23 15.34 84.66 7.16 Besley Brick Yard, middle loess. 22.5 25.30 2.32 74.70 12.73 87.27 9.77 Besley Brick Yard, bot¬ tom loess. 22.5 24.03 2.40 75.97 11.64 88.36 10.86 Getham Bros., inland loess. 25.0 22.43 2.41 77.57 10.71 89.29 14.29 Cap. City B. & P. Co., bottom stratum. 22.5 24.59 2.40 75.41 11.96 88.04 10.54 Cap. City B. & P. Co., green brick. 22.5 21.83 2.51 88.17 8.97 91.03 13.53 Granite B. Co., top stratum. 22.5 23,06 2.25 76.94 11.75 88.25 10.75 Granite B. Co., lower stratum. 22.5 22.41 2.42 77.59 . 10.51 89.49 11.99 Clermont B. & T. Co. 20.0 22.66 2.58 77.34 10.10 89.90 9.80 Am. B. & T. Co. 30.0 26.71 2.51 73.29 12.67 87.33 17.33 In making Table XI Byers’ and Williams figures for porosity/ 5 ' specific gravityf an d water of plasticity]; were taken, and the data calculated as follows: If porosity, or volume of pore space, is 29.77 per cent in a unit volume there would be 0.2977 parts by volume of pore space, and 1.0000—0.2977 or 0.7023 volumes of clay. On the assumption that the pore space is filled with water and the specific gravity of the clay is 2.34, there would be *Ia. Geol. Surv., Vol. XIV, 1904. 1 Ibid 83. PURDY] QUALITIES OF CLAYS FOR MAKING PAYING BRICKS. 159 0.2977X1-00=0.2977 parts by weight of water, and 0.7023X2*34=1.6434 parts by weight of clay, or ex¬ pressed as per cents—15.3 and 84.6 per cent respectively of water and clay. This 15.3 per cent of water then is the amount of water by weight that would be required to fill the pore spaces in a brick that would weigh 100 at the time when all the particles have become fixed or arranged in the exact position that they will maintain during the remainder of the drying period. This, it is assumed, would give the weight of water that remains in the pores of the bricks at the time the clay has reached its 0 H 17 H230 - H2H > Kio o o H20 If 170 ORl Kl2 o Kll o ° Ill 0 p ^ O H24 o R 2 K K 13 O S2 03 4 °% L-U K1 11 o KH O O OK5 oK8 OKI o Hie oK9 12 U 16 18 PERCENTAGE WATER OF PLASTICITY 22 Pig . 9 . Diagram showing relation between amount of water required to develope plasticity and volume shrinkage. (Data from Table I.) maximum air shrinkage. This amount of water subtracted from the amount required to develop plasticity would give, if the foregoing as¬ sumption is correct, the amount, the amount of water required to lubri¬ cate the particles sufficiently to cause a state of mobility which we have learned to designate as plasticity. Fig. 10 shows that there is some reciprocal relation between the amount of water in excess of that required to fill the pores of a dried brick, (as given in Table XI) and the volume shrinkage. In Table XII are shown the calculations on the Illinois clays, designed to bring out the same facts given in Table XI. In this table, however, the amount of hygroscopic water is given in each case so that it can be reckoned in as part of the mechanical water, if so desired. It must be borne in mind, however, that the amount of water calculated as being in excess of that required for filling the pores does not in any way include the hygroscopic water. The hygroscopic water is not added in with the 160 PAYING BRICK AND PAYING BRICK CLAYS. [BULL. NO. 9 PERCENT BY WEIGHT OF WATER IN EXCESS OF THAT REQUIRED TO FILL POKES Fig. 10. Diagram showing relation of volume shrinkage to water in excess of that required to fill the pores in Iowa clays. TABLE XII. Sample N umber. 1 Plasticity water. Porosity. Sp. Gr. by pyc¬ nometer. Per cent by weight of water required to fill the pores. Excess water re¬ quired for plasticity. Hygro¬ scopic w'ater. Vol. shrinkage. K- 1. 14.9 26.0 2.67 11.6 3.3 2.01 6.2 K- 2. 16.77 25.7 2.56 11.9 4.8 1.62 12.2 K- 3. 16.82 25.6 2.69 11.4 5.4 2.43 10.45 K- 4. 16.27 27 8 2 67 12 6 3.6 10.12 K- 5. 13.06 25 4 2.65 11.3 1.8 0.923 5.17 K- 6. 17.03 28.9 2.66 13.2 3.8 1.23 10.6 K- 7. 17.57 27.9 2.64 12.7 4.9 1.93 9 62 K-'8 . 14.4 25.2 2.69 11.1 3.3 1.70 7.51 K- 9. 13.4 26.1 2.70 11.5 1.9 0.79 3.54 K-10. 19.6 25.4 2.69 11.2 8 4 2.31 18.29 K-12. 13.35 18.3 2.67 7.7 5.7 5.09 12.74 K-13. 16.3 28.3 2.70 12 7 3.6 2.16 10.54 K-14. 13.6 24.5 2.64 10.9 2.5 0 79 6.13 S- 1. 17.2 23.0 2.64 10.1 7.1 4.76 11.97 S- 2. 16.6 26.4 2.72 11.6 5.0 2.42 13.1 R- 1. 13.4 17.8 2.73 7.9 5.5 1 95 13.9 R- 2. 13.0 24.0 2.72 10.4 2.6 1.53 9.1 R- 4. 13.2 21.8 2.72 9.3 3.9 2.28 5.98 B-II. 17.7 26 9 2.67 12.1 5.6 1.67 11.5 G-II. 11.8 22.4 2,70 9.6 2.2 1.14 7.32 H-II. 16.5 20.7 2.68 8.8 7.7 3.07 14.3 I-II . 14 4 18.9 2.67 8.0 6.4 2.85 13.8 J-II . 16.5 24.2 2.70 10.5 6.0 2 70 14.4 L-Il. 16.4 24.5 2.70 10.7 5.7 3.05 9.7 H-16. 16.2 27.8 2.70 12.4 3.8 1.74 7.8 H-17. 16.6 19.0 2.60 8.2 8.4 3.7 21.4 H -20. 18.3 23.9 2.72 10.3 8.0 2.58 16.5 H-21. 18.0 21.6 2.72 9.2 7.8 3.98 18.0 H-23. 21.4 24.9 2.63 11.1 10 3 2.05 20.4 H-24. 12.8 14.5 2,66 5.9 6.9 1.63 11.4 PURDY] QUALITIES OF CLAYS FOR MAKING PAYING BRICK. 161 water of plasticity, because there is some doubt as to just where and how the clay retains that water on drying. It is supposed to be held either in or between the grains, and does not greatly exceed the amount (on ac¬ count of the natural humidity of the air) that the powdered clay would retain as moisture. The porosity data used are those given in Table I. In the above table there is the same indication of a reciprocal relation between the “excess water” and volume shrinkage as noted in the case of the Iowa clays and shown in Fig. 11. We have here then the promise of a means of obtaining analytically a line on the drying behavior of a clay other than volume shrinkage taken alone. is to s* Fig. 11—Diagram showing relation of volume shrinkage to water in excess of that required to fill the pores in Illinois clays. o — KlO 603.6 552.9 O H20 H21 / 7Q2J9 ' / / / 397.2 OJT1 / fi-l/ / "Rl O S2 33 Vh 4m°i- n / /403.4 Ol K12 K4 n O 355.9 K13 O 220.6p 6 °Bdl S4 M / $24 / / O R2 300.3 o K7 / / / oLrll 3 366.4 o Gil 254.4^ / / 275.3 286.8 o K5 / / / / ’ Cr Fl 4 / Z _ -7 - T 5- T4 EXCESS WA TER PLUS HYGROSCOPIC WATER —11 Gr 162 PAYING BRICK AND PAYING BRICK CLAYS. [BULL. NO. 9 RELATION OF VOLUME SHRINKAGE TO FINENESS OF GRAIN. The volume shrinkage of a clay is a reliable index of its drying be¬ havior only within certain limits. Take for instance K—14 and H—17, which lie close to the extremes of minimum and maximum volume shrinkage; both require considerable care in drying. Roughly we can Fig. 12. Diagram showing relation of volume shrinkage lo fineness of grain. say that clays which exhibit an average shrinkage will dry safely, and that if the ware exhibits either a high or low volume shrinkage it can be assumed to be likely to occasion trouble in drying. But knowing this general fact, how can the drying behavior of a particular clay be esti¬ mated ? It has been suggested that clays which have a fair range in size of grain, i. e., not too large a proportion of either the largest or smallest grains, can be dried with greatest safety. This we proved to be true for PURDY] QUALITIES OF CLAYS FOR MAKING PAYING BRICK. 163 the clays plotted near the middle of a diagonally drawn dotted line in Fig. 12 were the easiest to dry and those at the extreme ends the most difficult. It was demonstrated, however, that while there may possibly be a reciprocal relation between porosity and fineness of grain in the naturally soft and loose-grained clays, there is no trace of such a reciprocal rela¬ tion in the harder clays, like shales, because the cement which holds the grains is not broken by the methods of preparation usually employed. It has also been shown that there is no reciprocal or proportional rela¬ tion between the porosity of the dried ware and the volume of shrinkage. This same lack of proportional relations was found between water of plasticity and volume shrinkage, as well as water of plasticity and por¬ osity. The only factors that seem to exhibit any proportional relation with volume shrinkage are “excess water” and fineness of grain. These factors alone are not, however, sufficient evidence on which to base an answer to our query. Tensile Strength. METHODS OF TESTING. One of the vital factors affecting the drying behavior of clays is their cohesion. Many ways have been devised to measure this cohesion, but the tensile strength test seems to be the most popular. Determinations of tensile strength as usually made and reported, have so large a per¬ centage of variation that they are practically worthless. This has been justly attributed to the personal factors entering into the preparation of the test pieces. It is indeed surprising how variable the results can be even when the operator uses all the care possible in wedging the clay and pressing the briquette. The personal factors have been largely eliminated in the tests here reported by following a method for making briquettes devised by H. B. Fox, of the University of Illinois. Fox Method .—The Fox method is, in the main, as follows: The clay is mixed with just sufficient water to make a thick paste. It is allowed to stand in this condition for some time, generally twelve or more hours, and is then poured onto a slightly moistened plaster slab and allowed to harden until it has assumed about the consistency of “stiff mud.” It is then cut into briquettes by a cutter similar to a biscuit cutter. The clay is forced out of this cutter into the briquette mold by a plunger under a given load; in our case about 50 pounds. While the load is still on, the cutter is removed and the briquette struck off with a wire. By this means the briquette is formed and pressed under uniform conditions without the introduction of personal factors,. with the possible exception of the making up of the slip. The briquettes are then room-dried. In this, care is exercised, for the fine-grained clays and the exceptionally weak clays can be dried so fast as to cause them to “dry check.” It is not always possible to see these “dry checks,” but there is no doubt that a considerable proportion 164 PAYING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 Fia. 13. Krehbiel device for grooving briquettes. purdy] QUALITIES OF CLAYS FOR MAKING PAYING BRICK. 165 The briquettes are then grooved to a slight depth by the use of a file operated in a miter/ making a uniform crpss section in all cases. The object of this grooving is not to obtain a uniform cross section primarily, but to insure the breaking of the briquette at the narrowest section. Be¬ ing uniform, the cross section can be considered as a constant factor, thus making easier the calculation of the results. This grooving was not trusted to give us a constant cross section, however, but each briquette was measured with a vernier shrinkage scale that reads to three places. The results of grooving the briquettes may be noted in the table given below. There it will be seen that the strength per square centimeter cross section is not materially different. In fact the only difference in strength between the grooved and the ungrooved can be said to be within the limits of errors that are unavoidable in this test. The usual contrast between, the variation in results in the grooved and in the ungrooved briquettes which ordinarily exists cannot be seen in the result given be¬ low. The results are exceptionally good in all cases, irregular results due to breaking elsewhere than at the neck not being reported. After the briquettes are grooved they are made bone dry in a hot air- bath and cooled in a dessicator so as to eliminate all moisture, and then broken in a Fairbanks Tensil Strength Machine. Wedging Versus Slip Process .—Clay workers, especially the old potters who make large jars by “throwing” on a wheel, recognize a difference in the working properties of clay when prepared by the slip process and when prepared by the “chaser,” wet pan, or the old-time stamping pro¬ cess. In fact the difference in the clay when prepared in slip, or in one of the “plastic” methods, is so marked, that where ware is to be thrown they install special machinery on which to prepare the clay, and in one of the most up-to-date terra-cotta factories in the west, they keep four men tramping the wet clay with their bare feet, in preference to using the cheaper slip method. In the manufacture of glass pots, tramping with bare feet is the method most generally used in preparing the clay. For this reason the fairness to all clays in casting the slab from which the briquettes were cut was questioned, and the following tests were made to throw light on this point. All the clays for both the “slip” and “wedge” process were made to pass through a 10-mesh sieve. The clay for slip process was cast as in the Fox method. The clay for the wedge process was thoroughly wedged by hand while at its state of maximum plasticity, and then worked into a sheet 1%" thick on the plaster slab by pounding it with a flat board. Briquettes were cut and forced into the mold under constant pressure as in the case of the slip clay. The results are shown in Table XIII: 2 See Fig. 13. 166 PAVING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 TABLE XIII. Slip Process Per Cent Variations Using Aver¬ age Strength as Basis. Wedge Process. Per Cent Variations Using Aver¬ age Strength as Basis. Sample. average STRENGTH IN LBS. PER SQ. CM. average STRENGTH IN LBS. PER SQ. CM. Grooved. Not grooved. Grooved. Not grooved. Grooved. Not grooved. Grooved. Not grooved. K— 14 Western Brick Co., Dan¬ ville, ill. 16.20 17.60 9.9 4.5 28.9 18.00 6.9 12.2 K— 10 Terre Haute, Ind. 29.80 33.30 10.3 11.9 36.8 31.6 17.1 25.3 K— 3 Albion, Ill. 17.65 20.40 10.3 8.8 23.9 23.9 16.7 18.4 K— 11 Brazil shale, Ind. 17.85 21.90 13.3 13.0 24.0 24.3 37.5 30.0 K— 9 Crawfordsville, Ind. 8.25 9.25 8.5 8.05 9.3 9.9 3.35 11.1 K— 8 Veedersburg, Ind. 16.85 18.60 5.37 4.0 21.7 22.1 14.3 21.2 Average. 17.60 20.17 9.61 8.37 24.1 21.60 15.97 16.3 The following conclusions were reached as a result of these tests: First—In every case except that of the Terre Haute not grooved, the wedging process gave higher results. Second—The variation is considerably lower in the slip than in the wedge process. Third—The increased strength .due to wedging was not sufficient to warrant the accompanying increase in percentage of variation. Fourth—Grooving the briquettes did not materially better the results in the slip process and actually made the results worse in the wedge process. It must be remembered in this connection, however, that the results of briquettes that did not break at the necks were rejected. All grooved briquettes broke at the neck. Fifth—Grooving increased the variation in coarse non-plastic clays, such as K—14 and K—9, but did not seem to effect the finer grained clays. Effect of fine Grinding —In view of the fact that grooving aids ma¬ terially in reducing the variation in all, except the less plastic, coarse grained clays, it was thought that perhaps the comparison would be more just if all were finely ground. The dry-pan samples of the two plastic clays, K—10 and K—11, and the two coarse and less plastic clays, K—14 and K—9, were ground of the variations are due to them. wet and also dry until they passed through sieves of 10, 20, 40 and 80 mesh, as follows: A quantity of clay sufficient to make six briquettes was taken from the stock by quartering, making ample allowance for waste. This sample was first crushed to pass a 10-mesh sieve. It was then sieved PURDY] QUALITIES OF CLAYS FOE MAKING PAYING BEICK. 167 through the desired mesh and the residue placed in a small Bonnot mill with 100 Iceland pebbles. Both the wet and the dry samples were taken from the mill every five minutes, and the particles fine enough to pass through the desired mesh were sieved out. The residue left on the sieve was then placed in the mill and ground for another five minutes. This grinding was continued until all the clay passed through the desired mesh. In this manner there was prepared, by both wet and dry grinding, stocks that would just pass the 10, 20, 40 and 80 mesh sieves. The clays that were ground were kept at casting consistency, i. e., quite thick slush, so that when completely ground they were cast into slabs as quickly as convenient. The clays prepared by the dry method were allowed to stand in water until they assumed the thick slip state and then cast on plaster of Paris slabs after standing from 10 to 24 hours. Briquettes were cut and pressed by the Fox method. In table XIY will be found the results of this experiment. The work was done by a man not accustomed to it who could not at first be made to realize the importance of taking .the greatest pains to in¬ sure constant conditions and accurate results. This may account for the higher variations. From these results the following conclusions may be drawn: First—The variations with the grooved briquettes are on the whole lower than those with the ungrooved. Second—The average strength of the grooved is practically equal to that of the ungrooved. Second-—The average strength of the grooved is practically equal to that of the ungrooved. Third—Finer grinding either wet or dry does not materially better the constancy of the results. The fact is, in this experiment, the varia¬ tions in the finer ground samples were higher in many cases than in the coarsely ground samples. Fourth—The average strength of the clay was not materially altered by finer grinding. Fifth—The results by wet grinding differed but little, if any, from those by dry grinding. TABLE XIV. PAVING BRICK AND PAVING BRICK CLAYS, [BULL. NO. 9 Wet Grinding. Per cent variation ungrooved . 22.28 17.97 2.94 27.33 4.14 25.29 22 22 15^92 20.00 15.83 10.00 17.62 11.11 13.5 10.56 10.8 Per cent variation grooved . 21.16 13.21 27.82 32.63 9.89 12 92 25.00 8.81 11.42 1.2 5 13.25 0.684 6.29 18.9 21.79 STRENGTH IN LBS. PER SQ. CM. Ungrooved. Min. 26.5 28.3 33.5 24.4 19.89 17.84 19.85 21.31 6.0 7.0 8.1 8.6 14.3 14.0 11.0 11.5 Max. IO OO lO CO ▼HiOCOCD c^oococo wwoh nncooj ^ co co o’ co in t-oo’cio coco cm cm* CO CO 00 CO CM CM CM CM HrtHH Grooved. Min. C^COCDlO COlOlOlO CM -C~0O ICOWCM lO^^CM tr- lO C— CM CO • CO tr- CM CM CM CM HHHCM HHHH Max. NOONI> O • t- C— CO 05 ** O cgoo^co os c- co c~ • r- oo Tinotoin MNWM HHNW i-( iH t- 4 iH Grinding duration in minimum. ■OiOO • lO o o • o o o • o o o •Hr-ICM • rH CM CM |H HCM -HHN Dry Grinding. Per cent variation ungrooved . 14.99 26.34 7.42 25.74 21.16 12.09 22.27 15.60 15.11 9.19 9.78 12.66 7.26 17.05 18.00 41.50 Per cent variation grooved . 5.41 32.59 11.32 22.5 14.35 7.94 36.79 11.15 1.33 4.93 17.07 17.7 12.00 5.76 14.42 12.42 STRENGTH IN LBS. PER SQ. CM. Ungrooved. Min. CM Oi tCOOCOH Cl Ci CM ^ HidCOO CO O CO o coco^u- cdcocoir- ocococo* c-t-ooco co^co*-* CO CM CO CM rH CM CM CM HHHH Max. 39.08 32.57 39.10 36.9 27.29 25.72 30.40 30.81 8.6 8.7 9.2 7.9 17.9 17.6 20.0 20.0 Grooved. Min. 31.7 27.3 31.4 34.1 17.3 19.7 14.6 26.6 7.4 7.7 6.8 7.0 15.4 14.7 14.8 14.1 * Bj S 33.4 40.5 35.4 44.0 20.2 21.4 23.1 26.9 7.5 8.1 8.2 8.5 17.5 15.6 17.3 16.1 Grinding duration in minimum. • moo -moo -moo • o m o ■MCOO tHNM 'NNM . # • • . . • • Mesh. oooo oooo oooo oooo thSj^OO SlM^OO TH OJ ^ 00 Sample. K10 Teire Haute.. Kll Brazil. K9 [Crawfordsville. K14 Western Brick Co., Danville_ PURDY] QUALITIES OF CLAYS FOE MAKING PAYING BEICK. 169 RESULTS OF TESTS. In the light of the foregoing tests it was decided to dry-grind the clays in a jaw crnsher to pass a 20-mesh sieve. In this the whole sam¬ ple, including the fine and coarse particles, was passed through the jaw crusher. Six briquettes were made by the slip method, as designed by Fox, and grooved to insure breakage at the neck. In this manner the following data was obtained: TABLE XV. Tensile Strength of Claves. Sample. K- 1 Alton, Ill. K— 2 Hydraulic, St. Louis, Mo. K- 3 Albion, Ill. K— 4 Springfield, Ill. K— 5 Edwardsville, Ill. K- 6 Galesburg, Ill. K— 7 Streator Paving Brick Co. K— 8 Veedersburg, Ind. K— 9 Crawfordsville, Ind. K—10 Terre Haute, Ind. K—11 Brazil, Ind . K—12 Brazil Fire Clay. K—13 Clinton, Ind. K—14 Western Brick Co., Danville, 111 K—15 Barr Clay Co.. Streator, Ill. H— 16 Carter, Peoria. H— 18 Sterling, Ill. H—20 Savanna, Ill. H—21 Galena, Ill. H—23 Carbon Cliff, shale. H— 24 Carbon Cliff ‘ fire clay. R— 1 Nelsonville, Ohio. R— 2 Portsmouth, Ohio. R— 3 Canton* Ohio, Imperial plant_ R— 4 CantonOhio, Royal plant. S — 1 Moberly, Mo. S— 2 Kansas City, Mo. B—II Atchison, Kan. G—II Coffeyville, Kan . H—II Topeka, Kan. I —II Caney, Kan. J—II Pittsburg, Kan. L—II Lawrence. Kan. F— 1 Danville Brick Co. Strength in kilograms per sq. cm. Percent of Var¬ iation. Maxi¬ mum. Mini¬ mum. As tested. By elimi¬ nation of irregular¬ ities. 7.356 7.168 7.58 12.292 8.437 31.52 9.934 8.074 18.72 6.84 10.806 8.664 18.8 11.1 5.715 5.216 8.73 8.164 7.516 7.202* 6.985 5.896 15.62 5.359 4 717 11.9 4.373 3.773 13.7 13.245 11.762 5.508 9.525 8.664 9 03 12.971 12.201 6.78 8.346 6.713 19.5 6.713 5.359 20.1 14.9 6.124 5.629 8.08 6.033 5.307 12.03 7.212 8 942 12 5 8.210 5.806 29.2 9.163 7.666 16.3 23.406 15.558 37.8 8.664 7.503 13.4 11.114 8.936 19.6 6.93 9.662 7.393 23.4 5.216 4.717 9.56 5.359 4.717 7.97 10.251 7.892 23.00 10.513 9.252 11.9 9 753 8 664 11.1 9.254 8.014 13.4 14.469 13.880 4.4. 13.742 12.383 7.04 10.069 9.662 8.09 8.925 8.028 10.0 12.111 10.523 13.1 ' CAUSE FOR VARIATION OF MORE THAN 15 PER CENT. K-2. There were two briquettes that broke with high strength and two with low strength. K-3. There was one briquette that broke with low strength. By throw¬ ing out that briquette the variation would be reduced to 6.84. K-4. There was one briquette that broke with high strength. By throw¬ ing out that briquette the variation was reduced to 11.1. K-7. There were two briquettes that broke with low strength. K-13. There were two briquettes that broke with low strength. K-14. There was one briquette that broke with low strength. Elimination of this briquette would make the variation 14.9. 170 PAVING BEICK AND PAVING BRICK CLAYS. [BULL. NO. £ H-20. There was one briquette that broke with high and another with low strength. H123. There was one briquette that broke with high and another w T ith low strength. R-l. There was one briquette that broke with low strength. Elimination of this briquette would make the percentage only 6.93. R-2. Three of these briquettes broke with high and three with low strength. S-l. There was one briquette that broke with high and one with low strength. The results here reported are exceptionally good. The variation in the strength of dry clay, as made by other methods, usually runs from 25 to 50 per cent in nearly every reported instance. In fact, it is seldom, if ever, that a report on tensile strength will show a lower variation than 25 per cent. The placing of 15 per cent as the maxi¬ mum variation to be allowed would be very severe standard ordinarily, but the general character of the work as here reported justifies the limit. RELATION OF TENSILE STRENGTH TO FINENESS OF GRAIN. Curves were plotted from data given by Bies 1 , and also by Beyer and Williams 2 , showing the relation between fineness of grain, as delineated by the surface factor, and tensile strength. There did not appear to be any consistent relation between these two factors, shown by the curves. KlO ,-v 7 ^ / r OHIO 0 / /£4 o Hia / / / O HI O K12 — / /o 'Kilt kN zs "HI oGZ ' / / oKl oKc / yka TENSILE STRENGTH IN KG. PER SQ. CM Fig. 14. Diagram showing relation between tensile strength and fineness of grain. Notwithstanding the apparent contradiction in the case of the New Jersey and Iowa clays, it is believed that fineness of grain in a given clay does bear a relation to the tensile strength. Orton 1 has shown the influence of different sized grains upon a very close grained and tough 1 New Jersey Geological Survey, Vol. 6, p. 89. 2 Iowa Geological Survey, Vol. XtV, p. 84. 1 Trans.Amer.Cer.Soc., Vol. Ill,p.117. PURDY] QUALITIES OF CLAYS FOR MAKING PAVING BRICK. 171 ball clay. This clay is so fine by itself that it is extremely difficult to dry without air checking, but with increasing adulteration of sa$d up to 30 per cent by weight, the tensile strength increased up to a maxi¬ mum in the sample where the sand was of extreme fineness, and here again the tensile strength decreased rapidly. This drop in the curve is credited to the inability of the extremely fine mixture to part with its mechanical water without checking, thus causing flaws in the briquette and very materially weakening it. In this experiment we have at both extremes very fine grained materials; one a pure ball clay and the other the same ball clay adulterated by fifty per cent by weight of a very fine sand, both having a low tensile strength. The intermediate members of this series show increasing strength with decrease of size of grain. So far at least as this one case is concerned, increase in size of grain increases tensile strength. Fineness of grain and tensile strength are, therefore, functions of one another. We know that a fine-grained shale is, in a majority of cases, im¬ proved by adulteration with sandstone, even in the fact of the fact that the sandstone is very co.arse. At Streator, Ill:, there are two strata of shale in one bank, the one, being very gritty, is easily manufactured into a good paver; the other, a close grained plastic shale, gives trouble in every stage of manufacture, and makes a poor paver. Yet these two shales are said to be of very similar chemical composition. The writer believes that the cause of this difference does not lie in their chemical composition, shrinkage, or ability to slake easily, but in their drying behavior. Judging from the results of Prof. Orton’s experi¬ ment on the tough ball clay, it is believed that if many of the plastic, fine grained clays were by addition of coarse material opened suffi¬ ciently to permit ready egress of the mechanical water, they would be excellent paving brick material 1 , while without such a treatment they would be worthless for anything other than building brick, simply be¬ cause the bond of the*clay would be weakened in drying by the expand¬ ing steam inside of the brick which could not readily escape. In Fig. 14 data are plotted showing the relation between fineness of grain and tensile strength. This is indicated by the dotted line. It will be noted that there is a general relation between fineness of grain and tensile strength. This is indicated by the dotted line. There is a remarkable coincidence in the relative positions of the several clays in Fig. 14 and Fig. 11. The same relative positions of the several clays is to be seen also in Fig. 12 which shows the relation be¬ tween volume shrinkage and surface factor. This same relative posi¬ tion of the clays, one with another, was developed also when the rela¬ tion between the sum of the excess and hygroscopic water and the surface factor, and also the relation between the sum of the. excess and hygro¬ scopic water and the tensile strength, were plotted. In the last two in¬ stances, however, the order in which the clays occurred was the reverse of that in Fig. 14 and 11. 1 It is not desired that the reader should infer that this is suggested as a panacea for all clays or that all clays can be “doctored” so as even theroetically to make them fit tor paving brick manufacture. 172 PAVING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 RELATION OF TENSILE STRENGTH TO VOLUME SHRINKAGE. We have seen that there is a greater shrinkage of the mass when dried from stiff mnd to bone dryness, as the grains of the clay decrease in size. If these fine particles are composed largely of clay substance they will possess a degree of cohesion that will cause the dried mass to become quite hard, the hardness increasing directly with increase of cohesion possessed by the individual particles. With increase of exposed surface. $ Fig. 15. Diagram showing relation between volume shrinkage and tensile strength. ought to be an increase in tensile strength, for the closer the particles are to one another the greater will be the bond between them. De¬ crease in size of grain, increase in volume shrinkage and increase in ten¬ sile strength should, therefore, follow one another in this order as — o Kio . / /*H23 / - °jnr OH20 / / ( / >J-I1 / / o Bril < KJlO S \ O K2 / ' 'iuy / 01-11 OK 12 K130 < Sl° / hff O K7 ^ / OR4 < o £8 Kiln i )H16 / C / OKI i CHI • / OKI 4 K90 / . 4 . o 70 73 u 16 TENSILE STRENGTH IN KC PER SQ. CM. PURDY] QUALITIES OF CLAYS FOR MAKING PAYING BRICK. 173 causes and effects. Such a relation is shown in Fig. 15 where the vol¬ ume shrinkage and tensile strength are plotted coordinately. The drying behavier of a given clay then can be said to be a function of four factors acting simultaneously, viz: Volume shrinkage, excess water, fineness of grain and tensile strength. The greater the volume shrinkage and the larger amount of excess water present, the more dan¬ ger will there be in drying. The greater the fineness of grain and the larger the tensile strength, the safer ought the clay to dry, all other things being equal. If a drying modulus were to be formulated it would have in the numerator the surface factor (S), representing the fineness of grain, and tensile strength (T) ; in the denominator there would be the percentage of volume shrinkage (1) and excess water (E), that is, S T - This simple relation is not, however, expressive of the true rela- V E. tive value of the involved factors. It is believed that the formula. -S s T V 3 E approximates th^ truth more closely. 100 Plasticity. THEORIES OF PLASTICITY. There is probably no property of unburned clay which has been more widely discussed than plasticity. To plasticity the clay owes its re¬ sponsiveness to every touch of the potter’s hand and its adaptability to the preservation of every line of the artist’s tool; it is this quality that permits of its being drawn out into sheets and cylinders of the most astonishing thinness. Of the many theories advanced as to the cause of plasticity the fol¬ lowing are the most tenable: Molecular Attraction Theory —To properly appreciate this conception of the cause for plasticity, suppose clay, to be blunged into the form of a slip, as is the practice of the potter before casting a vase. In this slip or fluid condition each grain is surrounded or enveloped by a film of water. If the volume of water is large compared with the total vol¬ ume of clay particles, the mass will behave in every respect like a fluid; indeed, as will the turbid water of the Mississippi. Suppose that, by evaporation, or adsorption by a plaster mold, the volume of the water be decreased. The clay particles will be brought closer and closer to one another, causing the mass to pass from a fluid state through var¬ ious stages of consistency until it assumes a stiff plastic condition; a process to be observed in mud roads after every rain. When in this stiff condition the particles still have an envelope of water or, in other 174 PAYING BRICK AND PAYING BRICK CLAYS. [BULL. NO. 9 words, they are still suspended in water just as truly as they were when the mass was more of a fluid. But, owing to their proximity, it is assumed by those advancing this theory of plasticity, that they are held in position by the molecular attraction which each particle of clay sub¬ stance exerts on the other. Molecular attraction is a known force, and there has been no adequate proof advanced upon which positive claims can be made against such a force operating between clay particles when brought into close proxi¬ mity. The popular conception of a bar of iron is that it is a rigid homogeneous mass, but, as is shown in magnetization experiments, it is made up of individual particles which can be turned about or set up endwise, thus acting independently of one another except in the matter of the molecular attraction that each exerts upon its neighbor, binding or holding the whole together. Aside from composition the degree of molecular attraction determines the hardness of the iron. Iron, then, is a solid fluid, that is, it will flow. The force of gravity is not sufficient to overcome this molecular attraction and cause flowage, but when a force that exceeds that of the molecular attraction is applied, flowage follows in the direction of the greater force. It is in this respect that iron is a fluid. If similar flowage is attempted when the grains in a clay mass are practically dry, or, in other words, not surrounded by water, except per¬ haps that held by absorption, pressure sufficient to overcome the force binding or holding the particles together will disrupt the ware. That is, instead of flowage of the particles in this comparatively dry state, rup¬ ture is a possibility. Further, maximum plasticity or ability to flow i§ not attained until the maximum number of particles is enveloped with the least amount of the suspending medium. This same phenomenon is to be noted with almost all fine insoluble powders. Wheeler 1 has shown, for instance, that the non-plastic slates, Iceland spar, propyllite, gypsum and halloysite can be made to develope a much smaller but still a fair degree of apparent plasticity with water as a floating medium. When dried, the force required to disrupt these masses, while small, is yet comparatively great. The difference, however, between the behavior of clay and these finely pulverized minerals is that the latter can be molded by pressure alone into a shape- that will have a comparatively higher tensile strength than if they were caused to acquire that shape by flowage due only to assumed plasticity. But we know that maximum density and consequent strength can be best developed in plastic clay by the combined influence of pressure and plasticity. Now is it mole¬ cular attraction in the case of clay, as in that of iron which can be bent, stretched, rolled, etc., in the cold without rupture, or is it merely that clay grains may be pressed so close together that fiow T age is permitted so long as water is present in excess, but is resisted by fractional force when dry? ^o. Geol. Surv.,Vol. XI. p. 106. 2 Carhart, H.S.,Univ.Physics,Pt. I. PURDY] QUALITIES OF CLAYS FOE MAKING PAYING BRICK. 175 Text books on physics give as an “expression” 2 for the force of mole¬ cular attraction between two molecules, M and M', MM'f (r). “All that is known about this funtion of r is that it is very large for insensible distances, that it diminishes very rapidly as r increases and that it vanishes while r is still very small. The maximum value of r at which molecular action ceases is estimated by Quincke to the 0.00005 mm. ;# If the particles then were 0.00005 mm. or 0.00002 inches apart, they would be at the extreme distance through which molecular attraction can possibly operate. Grout 1 says, however, “Now a simple calculation, based on the mechanical analysis of the clays, will show that the amount of water needed to place a film 0.00005 mm. thick around each grain is often nearly equal to the amount added in tempering, so that in ordinary plastic clay, it is necessary to consider practically all the water as being under this influence.” Grout 2 bases his reasoning on the following calculations: He found that his “mechanical analyses frequently show a large percentage of grains below 0.001 mm. in diameter, also from 0.001 to 0.005 mm. The average diameter of grains below 0.001 mm. is 0.0005 mm. If these are considered spherical and of specific gravity 2.5, it would re¬ quire 25.5 per cent by weight of water to place around each grain a film 0.00005 mm. thick.” On making these same calculations the following was obtained: PiD3 Vol. of sphere^- 6 Given diameter of sphere 0.00005 Pi — Log — = 1.71899 6 s - Log 0.0005 = 10.09691 11.81590 = Log 6545 X 10-14 volume of clay- sphere. Diameter of sphere plus water film = .0005 -+- .0001 or .0006 Pi — Log — =3 1.71899 6 3 10.33445 Log .0006 =- 1 — 10.05344 = Log 1131 X{ 10-is vol¬ ume in cu. mm. of sphere of clay plus water. Reducing these figures for the sake of convenience to 0.6545 = volume of clay sphere, 1.131 = volume of clay plus water sphere. 0.6545 -^- 1.131 = 0.5775, part of unit volume of clay plus water sphere occupied by the clay. 1.00 — .5776 = 0.4224, part occupied by water film. 1 Jour. Am. Chem.Soc., Vol. XXVII, No.9, Sept. 1905. 2 Loc.cit.,p.1046. 176 PAYING BRICK AND PAYING BRICK CLAYS. [BULL. NO. 9 Given specific gravity of clay = 2.5 Since in the metric system Vol. X Sp. Gr.=Weight 0.5775 X 2.5=1.4540 parts by weight of clay 0.4224 X 1.0= .4224 parts by weight of water 1.8764 total weight. 0.42246 -P 1.8764 = 0.2251, parts by weight of water in a unit vol¬ ume of clay plus water film, or 22.5 per cent. This calculation, so far as the validity of Grout’s argument is con¬ cerned, checks his results. Grout further calculated that if this same volume of clay were con¬ sidered as a square plate one-fifth as thick as wide, instead of a sphere, over 54 per cent of water would be held to the clay particles by this 1 molecular attraction. Supposing it to be fair, inasmuch as the kaolin- ite crystal is “plate-like,”' to consider that in a clay half of the grains are approximately spherical and the remainder plate-like. Grout fig¬ ures that a clay having all its particles the size here assumed would take by virtue of the molecular attraction of the clay particles, 40 per cent of water. In a personal interview the writer suggested to him that he was taking the maximum limit of the distance through which this molecular attraction can be said to operate. His defense was that when the spheres were devoid of a water film they touched one another, but as they gathered to themselves this water film, they need not necessarily be separated .00005 mm., for the film crowded from the points of closest proximity could be considered as filling up the space that would other¬ wise have to be considered as void. It must be admitted by the supporters of Grout’s molecular attraction theory for plasticity, that he used but a portion of a very fine-grained clay on which to calculate his demonstrating example. If he had taken into consideration the data for the sample of clay as published by him instead of only those for the finer portions, quite different results would have been obtained as is shown in Table XYI. The calculations by which the data in the following table were ob¬ tained are, (a) Volume of clay sphere PiD3 where D is the mean diameter of the -- range in each group of the mechanical 6 analysis. Pi (D —0.0001)3 —PiD3 (b) Volume and weight of water film - - 6 6 (c) Weight of dry clay particles as given in the mechanical analysis. id) Total or collective volume of spheres in each group: Weight given -f- Sp. Gr. of the clay. (e) Number of spheres in each group per unit volume: Total volume of d each group -p volume of clay sphere or — a PURDY] QUALITIES OF CLAYS FOR MAKING PAYING BRICK. 177 ( f ) Weight of water film surrounding the sphere in each group of the sample: Weight of water film times the number of spheres or e X b. ( g ) Sum of water required to give each particle in the sample a water film of prescribed thickness. TABLE XVI. Sample. Sp. Gr. Total dry weight of clav par¬ Total weight of water films Analyst. Reference to data used and explanatory ticles by analysis. by calculation notes. S. C. Besley, top clay (1) 2.34 0.9844 S. C. Besley, middle clay 2.32 0.9834 S. C. Besley, bottom clay 2.40 0.9819 Dale Brick Co. 2.44 0.9857 Gethmann Bros. 2.41 0.9739 Clarksburg fire clay (2)... 2.52 . 0.9880 Bridgeport stoneware clay (2). 2.35 0.9870 Charleston river clay (3). 2.66 0.987 Parkersburg pottery clay 2.58 0.984 K-l Alton, Ill. (4). 2.66 1.045 K-2 St. Louis Mo. 2.56 1.021 K-3 Albion, Ill. 2.686 1.002 K-4 Springfield, Ill. 2.67 1.037 K-5 Edwardsville, Ill... 2.65 1.028 K-6 Galesburg, Ill. 2.66 1.027 K-7 Streator. B. B. Co.. 2.636 1.037 K-8 Veedersburg, Ind. 2.689 1.005 K-9 Crawfordsville.Ind. 2.702 1.008 K-10 Terre Haute,Ind... 2.69 0.979 K-ll Brazil shale. 2.659 1.013 K-12 Brazil fire clay. 2.669 1.021 K-13 Clinton, Ind. 2.71 1.048 K-U Western P. B. C., Danville. 2.72 0.9899 R-l Nelsonville, O. 2.73 1.023 R-3 Canton, O. (Imper¬ ial) . 2.66 1.043 R-4 Canton,O.(Royal). 2.72 1.023 —12 G 0.0165 V/ iams. pp. 116 and 123, la. Geol. Surv., Vol. XIV. 0.0219 . .do. Williams’ first group was termed“aboveO,l MM.” 0.0263 ..do. In this group the mean diameter of the particles was assumed to be O. 175 M. M. 0.0320 . .do. 0.00254 . .do. 0.0393 Grout. pp.65and 251 W.Va. Geol. Surv., Vol. III. 0.1750 . .do. pp. 65 and 162 W.Va. Geol. Surv., Vol. III. Attract¬ ed water= 15 -(- per cent. 0.1009 . .do. pp.65 and 200W.Va. Geol. Sury., Vol. III. 0.1043 . .do. pp.65 and 160W.Va. Geol. Surv., Vol. III. 0.0356 Krehbiel and Moore 0.0491 0.0475 0.0703 0.0390 0.0320 0.045 Krehbiel and Moore 0.036 0.029 Calculated by Merry.. 0.0804 0.0467 0.0533 No. 2. fire clay. 0.0479 0.0348 0.0519 No. 2 fire Clay. 0.041 0.03902 178 PAYING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 Table XVI —Concluded. Sample . Sp . Gr. Total dry weight of clay partic¬ les by analysis. Total weight of water films by calculation Analyst. Reference to data used and explanatory notes. I-ll Caney, Kan. 2.67 2.71 2.67 2.72 2.72 2.6,3 1.013 1.045 1.1)42 1.031 1.029 1.027 0.0670 0.0463 0.0593 0.0739 0.10301 0.0867 G-ll Coffeyville, Kan... H-18 Sterling, Ill. H-20 Savanna, Ill . H-21 Galena, 111. H-23 Carbon Cliff, Ill. (shale) .. (1) The Iowa clays are loess. (2) Stoneware or No. 2 fire clays. (3) Alluvial clay. (4) The Illinois clays are shales in every instance except K-12 and R-l. In Table XVI there is but one instance that of the West Virginia stoneware clay, in which the amount of water molecularly attracted even approached that required to develop plasticity. In many instances it does not greatly exceed the hygroscopic water that the clay would re¬ tain when dried in open rack dryers. In fact the maximum amount of water which Grout admits could be so molecularly attracted, agrees quite closely with the water which in Table XI is shown to be in excess of that required to fill the pores. While Grout’s statement of the facts in this case has been proved incorrect, further investigation may find a relation between the molecularly attracted water and “excess water.” As yet, however, such a relation cannot be established. That a clay particle does possess a molecular attraction peculiar to itself is not denied. That this molecular attraction alone is sufficient to cause a plasticity that is peculiar and belongs to no other substance must be discredited until evidence is brought forward that will bear an analysis such as is given in Table XIV. It would be most difficult for supporters of the molecular attraction theory to prove that the kaolin grains in primary clays do not possess every physical property that is attributed to the grains of the clay sub¬ stance in the secondary clays, save that of plasticity. Chemically alike, and differing physically only in this one respect, yet to the one, accord¬ ing to this theory, must be accredited no, or very little, molecular at¬ traction for water, and to the other a strong molecular attraction. Grout 1 may be quoted as follows: “The attraction of two grains may vary with the nature of the grains. The greater the attraction the farther they can be separated without losing coher¬ ence. -- —. Another way in which the films become viscous is the result of molecular attraction, which binds a film over the surface of the grain and renders it viscous. The friction between this film and the solid grain of clay is said to be infinite, compared with water outside of the film. But when forced to move, the resistance would depend on the strength of the attraction of clay and liquid.-. The change in viscosity or 1 Jour. Am.Chem.Soc., Vol .XXVII, No.9,Sept. 1905,p.1016. PURDY] QUALITIES OF CLAYS FOE MAKING PAYING BRICK. 179 in thicknes of the film, seems to be beyond the region of experiment. The quantity is too small to admit the determination of slight changes, but such are constantly assumed in physical problems. W. J. A. Bliss speaks of clay particles and the surrounding adherent liquid as follows: ‘The thickness of this shell depends on the intensity of the attraction between the solid and the liquid.’ J. E. Mills says: ‘Molecular attraction depends primarily on the chemical constitution of the molecule. — —Certain rare organic colloids increase the plasticity by rendering' the water viscous. -. The tendency for tensile strength to vary with plasticity is also easily explained in this way. Molecular attraction between two kaolin grains may be high. If the attraction for water is high, some water will be drawn between the grains and rendered viscous by the attraction; this makes plasticity high. But when the water dries out from such a mass, the kaolin grains still attract each other, and the chances are for greater strength • than when wet, because the water has acted as a lubricant, allow¬ ing a readjustment of grains to fill the space left as the water moved out. The result is a high degree of consolidation.” Mr. Grout’s arguments may be summed up as follows: 1. Attraction varies with the nature of grain, i. e., their chemical con¬ stitution, or in other words, molecular structure. 2. Films become viscous as a result of molecular attraction, the more strongly attracted film being .the more viscous. 3. Organic colloids increase plasticity by rendering the water film viscous. 4. The tendency for tensile strength to vary with plasticity is explained by molecular attraction between grains. 5. Change in viscosity or in thickness of film is beyond the region of ex¬ periment. Granting that these arguments may be valid and may be substan¬ tiated by facts, it will be shown later that they may be considered as establishing the existence of an effect rather than the existence of a cause. Size of Grain Theory of Plasticity —It has been shown earlier in this discussion that the size of the grains as determined in the mechanical analysis does not agTee with the normal fineness of grain in the clay as it issues from the pug-mill; there are bundles of grains that success¬ fully withstand the disintegrating effect of water in the pugging pro¬ cess, but which are to a large extent disintegrated in the process of mechanical analysis. It is obvious therefore that conclusions based wholly on the results obtained in the mechanical analysis cannot be considered as necessarily agreeing with the facts observed in the actual behavior of the clay under factory conditions. In many clays, however, these bundles are broken down to such an extent that the analytical results indicate quite accurately their actual working properties. Because in the mechanical analysis the coarser grains have been re¬ ported as sand and the finer particles as silt and clay, not a few have been led to conclude that clay particles, or at least particles in which clay substances constitute a large proportion, cannot be present in a clay as large grains after thorough disintegration in water. Grout has shown, however, that this conception is entirely erroneous. 1 In Table XYII is given the amount of clay substance that he obtained first from the analytical analysis, second, calculated from ultimate analysis, and third, obtained from mechanical analysis. 1 W.Va.Geol.Surv., Vol.Ill, 1905, p. 26. 180 PAYING BRICK AND PAYING BRICK CLAYS. [BULL. NO. 9 Table XVII. Showing the discrepancy in the reported “ clay substance” in clay, by the three methods for its determination now in vogue. Specimen Number. (Rational Analysis 1 Calculated Kaolin Mechanical Analysis. 4. 67.23 52.30 11.8 17. 36.80 26.39 36.85 41. 72.26 41.65 63.70 62. 70.48 41.14 59.70 76.. 42.41 31.50 33.35 Mr. G-rout has also given 1 results of the chemical analysis of a com¬ plete mixture of the several grades of fineness obtained from 16 samples of clay as follows: Table XVIII. Constituents. .00 to .001 .001 to .005 .005 to .02 .02 to 0.15 0.15 up. Si0 2 . 44.08 54.54 70.30 81.16 73.63 ai 2 o 3 . 28.16 23.00 16.04 9.76 13.01 Fe 2 0 3 . 7.94 5 91 3.21 2.13 4.71 FeO. 0.99 0.99 0.63 0.40 0.18 MgO. 1.36 1.02 0 80 0.39 0.48 CaO. 0.76 0.82 0.72 0.31 0.47 Na 2 0. 0.00 0.29 0.45 0.56 0.00 KoO. 3,05 3 31 2.14 1.78 0.93 h 2 o. 2.80 1.10 •0.56 0.35 0.87 Ignition. 10.86 7.79 4.33 2.59 4.40 Ti0 2 . 0.84 1.12 1.08 0.78 0.60 In this he has proved conclusively that the “clay substance” is pres¬ ent in every grade of fineness. His own conclusions from these analyses are, however, rather startling. He says: “The silica percentage is higher in the coarser portions, where it probably is present in the form of sand or quartz. Alumina is higher in the finer material, but total fluxes are also higher, so that the finest particles are not the purest’ kaolin.” In order better to show the validity of his conclusions his data has been calculated into molecular equivalents as given in the following table: 1W. Va. Genl. Surv., Vol. Ill, p. 61. PURDY] QUALITIES OF CLAYS FOR MAKING PAYING BRICK. 181 Table XIX. Grades of Fineness. Si0 2 ai 2 o 3 Fe 2 0 3 FeO MgO CaO Na 2 0 k 2 o Ti0 2 0 00 to 0 001. 2.66 1.00 0.18 0.05 0.12 0.05 0.12 0.04 0.001 to 0.005. 4.03 1.00 0.16 0.06 0.11 0.06 0.02 0.16 0.06 0.005 to 0.02. 7.45 1.00 0.13 0.06 0.13 0.08 0.05 0.14 0.09 0.02 to 0.15. 14.14 1.00 0.14 0 59 1.02 0.58 0.94 0 19 1.02 0.15 up. 9.62 1.00 0.02 0.002 0.01 0.007 0.008 0.006 A review of Grout’s mechanical analysis of the West Virginia clays discloses the fact that he made 26 determinations: 6 plastic fire clays, pp. 160, 162, 163, 233 and 251, 1 flint fire clay, p. 218, 7 shales, pp. 249, 251, 242 and 262, 10 river clays, pp. 263, 265, 270, 272, 274, and 276. 1 glacial clay, p. 265, 1 residual surface clay, p. 200. It is assumed, therefore, that the samples, the analyses of which are given in Table XVI, are composites of the several grades of grains from the above clays. Being in most cases very impure clays, it is con¬ sidered that although a study of the possible mineral, make-up of each grade is at the best largely based on hypothetical assumptions, such a study would aid in our attempt to understand the constitutional make¬ up of our clays. On the assumption that all the alkali is present as a RO in orthoclase feldspar, the molecular ratio and ratio by weight of kaolin, feldspar and quartz present in each grade would be as follows: Table XX. Showing 1 possible mineral constitution of the several grades of grains in impure clays. Grade. Molecular Ratio. Weight Ratio. Quartz. Kaolin. Feldspar. Quartz. Kaolin. Feldspar. 0.00 -0.001. 0.88 0.12 .18 10 2.9 0.08 0.001-0.005. .82 0.18 1.25 10 4.7 3.5 0.005-0.02. .81 0.19 4.69 10 5.1 13.5 0 02 -0 15. 1.00 8.14 10.0 8. 0.1S up. 0 992 0.008 3.62 10 0.18 8.5 This data checks the fact developed in Table XV, i. e., that clay sub¬ stance is to be found in all of the grades of fineness, in the coarsest as well as the finest. It also shows that more than 50 per cent of the 182 PAVING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 coarsest group, or as it is customarily called, "coarse sand,” may be kaolin, or is at least kaolinitic in composition. As a further analysis of the probable mineral make-up of clays. Grout’s data will be discussed by groups. In this only the most com¬ mon and abundant minerals known to occur in clays are considered. Coarsest grade (0.15 mm.) : This grade of grain, even if all the alkali is considered as being present as a constituent part of feldspar grains, would be assumed to be composed almost entirely of non-dis- integrated kaolin and quartz grains. Only in one case, however, does Mr. Grout 1 speak of the physical character of the grains of this grade. In this particular case the clay examined is a shale. "The 12.9 per cent (referring to coarse sand grade) on 3 mm. screen was mostly flat scales of shale, about 5 mm. in size, of red and greenish color.” The total absence of similar description of this grade in the other 25 samples justifies the conclusion that the grains of this grade were flat or scale like only in this one sample. If this conclusion is true, then it is fair to assume that either the kaolin scales are present in undissolvable bundles, or these grains are not composed of kaolin but some other aluminum compound like gibbsite, etc. On the other hand, it is hardly possible that grains of feldspar of this size could remain unaltered in these old river clays that have been elutriated, mixed and moyed by fresh waters possibly for ages. There is justification for the assumption, therefore, that these coarse grains are bundles of kaolinitic grains cemented together so tightly by some salt that they resist disintegration by water. If the alkalies had been present as constituent parts of feldspar grains of this size, the feldspar crystals could have been easily recognized under the microscope as cubical grains and not flat scales. H. B. Fox, in the Ceramic laboratories of the University of Illinois, separated the grains. of a shale and a glacial clay into the several grades of fineness, and found that all the grades possessed a plasticity that varied directly with the fineness of grain, and that the coarse grains which could not be disintegrated by 20 hours of constant shaking in water, when broken down in a mortar, developed plasticity that in¬ creased as the size of the grains decreased, until when the coarse grains had been reduced to an impalpable powder they developed a plasticity nearly equal to that exhibited in the finest grains that had been separ¬ ated from the original sample, showing, it is believed, that the coarser grains were comprised of materials similar in every respect to those in the fine grains, but cemented in such a way that they withstood success¬ fully the disintegration treatment. (0.02 to 0.15) grade: It is highly improbable that this grade con¬ tained no kaolin or clay substance, but such would have to be the case if all the alkali was present as a constituent part of the orthoclase feld¬ spar grains. The alkalies cannot be present in this case, as easily soluble iw. Va. Genl. Surv., Vol. Ill, p. 249. purdy] QUALITIES OF CLAYS FOR MAKING PAYING BRICK. 18B salts, for the alkaline salts would have been dissolved, carried in solu¬ tion, and would affect only the finest grades. If the feldspar was oligo- clase and not orthoclase, then the 0.5 equivalents of the alumina could be considered as a constituent of kaolin grains. Although there is no statement made as to the presence of mica in the clays from which these grades of grains were obtained, Mr. Grim- sley * 1 ‘states that it is a very common constitutent of the West Virginia clays. Stull 2 gives as the chemical formula of common muscovite mica the following: 0.1243 CaO.-..1 0.1103 MgO.! 1.000 A1 2 0 3 .) 6.399 Si0 2 —0 1)74 H 2 0 0.3280 K 2 0. [0.1857 Fe 2 0 3 .(Comb. Wt. 582.167 0.0929 Na 2 0.J On the assumption that the alkali in this grade is derived wholly from muscovite mica of the composition given by Stull,-the mineral constitu- tents of this grade of grain might be proportioned as shown by the following calculations: .57 Eqv. Mica Si0 2 14.14 3.65 10.49 .86 A1 2 0 3 1.00 0.57 0.43 0.43 Fe 2 0 3 0.14 0.11 0.03 FeO 0.59 MgO 1.02 0.06 0.96 CaO 0.58 0.07 0.51 Na 2 0 0 94 0.05 0.89 K 2 0 0.19 0.19 Ti0 2 1.02 43 Eqv. Kaolin.... 0.59 1.02 .63 0.89 0.57 Eqv. Mica x 582.167 = 331.835 or by proportion 30.0 0.43 Eqv. Kaolin x 258 = 110.940 or by proportion 10.0 0.63 Eqv. Silica x 60 = 37.800 or by proportion 3.4 In this case, the formula most favorable to the supposition that all the IGO is present in the form of mica has been taken. If the theoret¬ ical formula IGO, 3 AbCh, 6 Si Cb, 2 ILO had been taken, there would have been either considerable IGO to account for in some other way, or return to the original hypothesis that this group contained no kaolin. Either supposition leaves considerable alkali unaccounted for, which as has been pointed out, could not possibly be present in an early soluble form. The supposition therefore that this grade is composed in part of kaolinitic grains cemented together by some alkaline salts, finds sup¬ port in any plausible assumption that may be made. (0.005—0.02 and (0.001—0.005) groups: If the kaolin grains in these groups were in their natural condition, i. e., flat plate-like crystals> they should, theoretically, be visible through the microscope. This evi¬ dently was not the case. Beyer and Williams 1 say: “While it is next to impossibleto make out much concerning the crystalline character of the minerals, it is also difficult, because of their minute size in most secondary clays, to' say anything regarding their shape/’—In Other 1W. Va. Geol. Surv., Vol. Ill, p. 12. 2 A. C. S. Trans. Vol. IV, p. 258. 1 la. Geol. Surv., Vol. XIV, p. 94. 184 PAYING BRICK AND PAYING BRICK CLAYS. [BULL. NO. 9 words, the shape of the grains is irregular and non-conformable one with another. If in these grades there is as much kaolin as is shown in Table XVIII, its grains must be, to a very great extent, in bundles. If feldspar or mica is present in such amount and size of grain, as the cal¬ culated data suggest, their grains ought to be detectable with the aid of the microscope. Such, however, is not the case. The supposition, therefore, that all the akalies of these two grades are there as constitu¬ ent parts of feldspar and mica is certainly untenable. (0.000-0.0001) grade: The molecular composition of this group is certainly very instructive. That in such heterogeneous mixtures as shales and river clays the finest particles are found to be composed in the main of kaolinitic grains is certainly astonishing. If the bases present, as shown in Table XIX, page 181, are considered as being present as soluble salts that were either originally present in the clays or in part introduced during the process of analysis, (a most plausible assumption) then there would remain but one conclusion, that is, that the finest insoluble grains are almost entirely kaolinitic in composition. Taking data given by Grout, it was calculated that if all the soluble salts originally in the clays were in the finest group, they would amount to 2.7 per cent of the weight of that group. The 2.7 per cent, together with the soluble salt introduced during the process of analysis from glassware, water, atmospheric dust, etc., would account for nearly all of the alkali in the finest grade. It is not mere assumption therefore, that the finest particles in clay, contrary to Grout’s statement, are the purest kaolin grains. In the course of the research on paving brick clays by the survey there was much speculation as to the number of these submicroscopic kaolin grains in the various shales. This was readily ascertained as follows: By dividing the percentage amount of the group (0.001 to 0) by 100, and considering that as being a part of 1 milligram of the sample, (for the size of the particles is in millimeters) then dividing this amount by the specific gravity of the clays, a figure is obtained that represents the sum or total volume in cubic millimeters of the particles comprising the group. Considering 0.0005 as the mean diameter of the particles, 6 by the formula -the volume of each particle is found to be 654xl0' 13 PiD 3 cubic millimeters. Then for each day, bv dividing the total volume of the particles, by the volume of one particle, the number of grains per milligram of the sample will be obtained. By multiplying the num¬ ber of grains in one milligram by 1,000 there would be obtained the number of grains in 1 gram; or by multiplying by 352,740 there would be obtained the number of grains of this size in 1 oz. of the whole sample. In this way Table XXI was calculated. 1 This mean diameter is twice as large as that given by Whitney for the finest group; U. S. Dept, of Agr. Weather Bureau, Bull., No. 4, p. 35. PURDY] QUALITIES OF CLAYS FOR MAKING PAYING BRICK. 185 Table XXI. Number of grains of group (0.001 to 0) in— Sample No. 1 gram of the clay. 1 oz. of the clay. K 1. 560 trillions. 829.0 trillions. 794.0 trillions. 1,588.7 trillions. 1,940.0 trillions. 440.0 trillions. 197,534 trillions. 292,421 trillions. 280,075 trillion4. 560,398 trillions. 684,315 trillions. 166,205 trillions. K 2. K 3 . H 23. H 21. K 6 . These figures, although beyond the limits of perception of the human mind, are not larger than the figures representing the countless germs that bacteriologists claim can exist in a single drop of a fluid. Startling as this data appears to be, it cannot be other than true if the analytical results of the mechanical separation are correct. If these minute particles were not kaolin grains, would they add to the real plasticity of the clay? Potter’s flint (dry ground) is finer grained than most clays, and particularly more so than the shales, yet it does not exhibit the faintest sign of plasticity. Orton 1 found that glass particles which were so fine that they remained in suspension for hours without settling, when collected exhibited no plasticity. Wheeler 2 found that while quartz crystals ground to 200 mesh, seemed to be ap¬ preciably plastic, on drying the coherence was so slight that it required the gentlest .handling to prevent the molded sample from falling to pieces. Fine quartz dust. and impalpable geyserite or finely precipi¬ tated opal, dried to a very tender mass. The same was true of tripoli. Wheeler 3 found that some plasticity could be developed in powdered slate, prophylite, talc, gypsum, halloysite, etc., but that the plasticity developed was only apparent plasticity, except perhaps in the case of slate. The powdered gypsum when molded and dried formed a rela¬ tively hard mass, but this hardness would be expected on account of the solubility of gypsum in water. The plasticity of the slate, which is a dehydrated shale has caused considerable surprise, and has strengthened the fineness of grain theory of plasticity. That powdered slate should develop plasticity need not be so great a source of wonder, for in the course of the Survey work a shale, after having been held at heat rang¬ ing from 500° to 800° C. for 17 hours, slaked down in water to a red plastic mass in the same manner as the unburned shale at the bank. True, the plasticity of this partially burned shale was not equal to the plasticity shown by the clay before dehydration, but its plasticity was considerably more than that of some of the harder shales before being burned. Fineness of grain in itself then does not seem to be the cause of plasticity. It may said, however, to be a required condition in the operation of the real cause. 1 Brick, Vol. XIV, No. 4, p. 216. 2 Mo. Geol. Surv., Vol. XI, p. 102. 3 Loc. cit., p. 106. 186 PAVING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 In seeming contradiction to this statement regarding the fineness of grain as a sense for plasticity, is the fact that finer grinding of given clay increases its plasticity; but quoting Wheeler 1 : “While it is true that fine clays are usually very plastic and coarse clays much less so, there are very many exceptions.” And again, Grout 2 says that while the majority of clays improve on fine grinding, some are unchanged. Wheeler 3 * 5 reports the physical structure of a few clays as follows: Moberly shale (400 diarn.): Mainly clusters of thick plates with minor portions split off; moderately plastic; suggests fine grinding to develop plasticity. Aldrich shale (325 diam.): One-third dolomite crj^stals; bulk in coarse thick crystals or plates; rest in fine state of division; moderately plastic. Unweathered Leasburg flint fire clay (950 diam.): Almost all fine particles; no plates or scales; devoid of plasticity. Weathered Leasburg flint fire clay (950 diam.): Numerous coarse plates present and occasionally, apparently a few thin plates. Came from same bank the same day a few feet from the unweathered sample. Hartwell loess clay (400 diam.): Large angular fragments which were undoubtedly sand, and apparently some clusters of plate crystals, with only a minor portion of small plates; very plastic. There is sufficient evidence in the above citations to show that any theory so far discussed other than that of molecular attraction, is in¬ sufficient to account for the presence or absence of plasticity. Plate Structure Theory of Plasticity —Grout 1 has recorded the fact that in the case of the Thornton Brick Company’s plastic clay the amount by weight of the particles below 0.005 mm. in diameter rose from 7.7 per cent to 17.8 per cent by weight in one wetting and drying. Fox, in our laboratories, found that the plates, although not disintegrat¬ ed by twenty-four hours of shaking in water, would break down by mechanical crushing or by disintegration in acids and caustic alkali, and that when so broken down the mass became considerably more plastic. Wheeler 1 not only advises fine grinding in the case of the Moberly shales, but relates a most remarkable instance of a clay in which the grains on weathering formed themselves into clusters re¬ sembling plates. It seems highly probable, therefore, that these plates or coarse grains are bunches or bundles of minute grains cemented to¬ gether by salts that are to a greater or less extent soluble in water, and that, depending upon the solubility of the cementing salt in a particular case, or the peculiar compactness of the grains in another, it requires a greater or less amount of time to cause a breaking down of these bundles. It can be readily conceived that the adsorptive power of the particles when combined with their axes in a certain general direc¬ tion, for instance, has greater power in holding certain of these cement¬ ing salts than the solvent action exerted by the water can overcome.' The solvent power of water, in other words, is not sufficient to overcome the adsorptive power of the kaolinitic grains. l Loc. cit., p. 109. 2W. Va. Geol. Surv., Vol. Ill, p. 46. 3 Loc. cit., pp. 104 to 109. 4W. Va. Geol. Surv., Vol. Ill, p. 46. 5 Loc. cit., p. 105. PURDY] QUALITIES OF CLAYS FOR MAKING PAYING BRICK. 187 These coarse grains add to the plasticity of the clay as a whole in a -ratio to the surface exposed. Every exposed kaolin particle is as effec¬ tive in enhancing plasticity as the very small independent particles. The extent to which the larger grains would affect plasticity would, there¬ fore, be in proportion to the exposed surface of the particles of which the bundle or cluster is composed. 1 Further, it is fair to challenge the plate theorist to demonstrate that these small grains when cemented together in a bundle or cluster have not a tendency to line up one with another so that their longest axes 1 will lie in the same relative plane, just as they are in the natural kaolin crystals, i. e., in plate forms. The plate theorist must admit that when these bundles are thus formed they are well-nigh indistinguishable from mica crystals, and that the very large majority of so-called plates or scales of kaolin in a clay are most likely to he mica. It is certainly strange that on one page of a report there will be a statement to the effect that “the clays of this state are quite micacious,” and another page will report the scales -that appear on the stage of the micro¬ scope as “kaolin scales or plates.” ' If the idea that has been put forward in the foregoing is correct, then we must agree with Dr. Ladd 1 when he says: “The question of fineness of grain and shape of the particle becomes, then, largely but modifying factors, affecting degree, and being, within large limits at least, modifiers, rather than determinants of plasticity.” 1 It is quite evident that the peculiar physical make-up of a kaolin grain, so far as the eye by the aid of the microscope can discern, is not fundamentally responsible for their individuality, as expressed in their power to develop plasticity. If the structure of the grains which en¬ ables a mass of them to develop plasticity is not detectable by the micro¬ scope, direct observation and measurement are obviously inadequate in finding the true cause. * Pectoidal Theory of Plasticity —Turning to indirect or circumstantial evidence, there are many facts observed by a good many careful scientists that seem to point to one thing that is more characteristic of kaolin grains than of any other of the inorganic substances or minerals of which a clay is composed, i. e., adsorptive power. Some investigators have even gone so far as to attribute the plasticity of kaolin grains to an adsorptive power or actual taking into the grains themselves of foreign salts from solution. They advance the theory that these minute grains have a micellian structure. To such substance they apply the name “Pectoid,” and to the theory the name, “Pectoid theory.” To many, the absorptive and adsorptive properties of a clay are one and the same thing, and so far as can be judged, the most radical believe in either the adsorptive or the pectoidal theory, and oscillate from one to the other in a manner that induces skepticism. The fact remains, how¬ ever, that both use the same arguments, the only difference being in the conception. It is safe to warrant, that when the pectoid theorist real- l Geol. Surv., of Ga., Bull. 6-A, p. 32. 188 PAVING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 izes than in one gram of clay, the disintegration of which has been effected only by shaking in distilled water, there can exist from 400 to 1500 trillion free and independent sub-microscopic particles, to say nothing about the larger particles, they will find interstices between these grains sufficient to satisfy even the most exaggerated conception of a micellian structure. 1 Dr. Cushman 1 in a brief review of the observations that point toward a colloidal substance as being the prime cause of plasticity, has given the following citations: “Daubree found that wet ground feldspar as¬ sumed a plastic condition, whereas dry ground feldspar did not.” Ost- wald, the eminent German physical chemist, Arons, Bischol, Seger, Rokland, and Van der Bellen, accepted and advanced in substance the colloid theory. T. Way 2 stated that while particles of sand and chalk absorbed water, owing to surface attraction and capillarity, clays and soils with a clay base behaved in a quite extraordinary manner. The more clayey the soil the more water it seemed capable of absorbing. But this was not all; besides water this clay substance exhibited a greater facility for absorbing the bases contained in certain salts which Were dissolved in the water.” 1 E. Bourry 2 says that if clay is mixed with a solution of calcium car¬ bonate, the clay will retain some of the carbonate. “Kaolins do not retain more than 2 per cent of carbonate of lime in solution, while plastic clays can absorb from 10 to 20 per cent of it.” i It is common knowledge among chemists that clay can extract solu¬ ble salts from solution and retain them very persistently against all (attacks by dissolving mediums. Mr. Ackison 3 found that catechu and extract of sumac leaves, spruce bark, tea leaves, oak bark or straw would be absorbed by clays from solutions. * Further, Ries 4 5 advanced the theory that the action of hydro-carbons in solution was to deflocculate the particles. This he claims was proved by the fact that in the untreated Clays the grains were bunched to¬ gether while in the treated clays the particles were separated. 1 In the foregoing citations there is evidence sufficient to formulate a Conception of what changes have taken place in a clay from the time it was first formed in situ by decomposition of the parent rock and left practically devoid of plasticity, until it was deposited elsewhere as a plastic clay. Organic matter would have deflocculated the particles, (and the soluble salts, which are very naturally attracted to the kaolin grains, would soften when wetted, but the water could not extract them owing to the greater force exerted by the kaolin particles. Defloccul^- tion by organic matter, recementation by salts of various kinds, may have formed a cycle of events that in the end would cause a condition Of affairs that makes possible the property described as plasticity. Molecular attraction for foreign substances which is peculiar to kaolin particles, and not to a very large degree to other common constituents of clay, may and does have its influence on plasticity affected by fine- 1 Vol. VI, A. C. S., p. 66. 2 Royal Ag - . Soc., Jour. XI, 1850, cited by Cushman. 3 Emile Bourry , Treatise on Ceramic Industries, p. 54. 4 Amer. Cer. Soc. Trans., Vol. VI, p. 33. 5 A. C. S. Trans., Vol. VI, p. 43. PURDY] QUALITIES OF CLAYS FOR MAKING PAYING BRICK. 189 ness of grain. Plate structure, a natural form in which kaolin grains arrange themselves, is a possibility under the right conditions, but there is also a possibility that organic matter and adsorbed salts may operate in the destruction and formation of the plate grains. A clay that is of secondary origin, like our common clays, could not have passed through the many geological changes with which they are credited without being more or less defloeculated and saturated with these foreign substances. Micellian structure is not a necessary condition. Minuteness of grain and consequently large surface or adsorbing area is sufficient. Adsorption theory of plasticity —Existing data, accumulated for years by scientists, all point to the fact—which is almost beyond the theoretical state, lying wholly within the realm of experimentation— that the plasticity, tensile strength and general working properties of the clay can be traced back to the adsorptive property of kaolin. Fur¬ ther, all the facts that have been cited in support of any and all of the theories are identifiable as conditions that allow of the fullest exhibition Qf the plasticity that seems to follow as a direct consequence of the adsorption of soluble organic and inorganic substances by the kaolin grains. DEVELOPMENT OF PLASTICITY IN THE PRESENCE OF WATER. Whatever may be the fundamental cause of this phenomenon we call plasticity, it is certain that it is manifested only when water is present. It has been shown that mere molecular attraction between the clay grains and the water molecules is not sufficient to account for plasticity. There must, therefore, be factors other than molecular attraction which becomes operative in developing this property, which, when water is' not present, may be said to be latent. Since it is the presence of water that makes the development or expression of plasticity possible, it is important that we consider some of the fundamental and well-known hydrostatic forces. There are at least four forces operating on the water in a wet, un¬ burned brick: First, gravity, or the weight of the water itself; second, surface tension, which is due to attraction (cohesive) between the molecules of water themselves; third, molecular attraction (adhesive) between the water molecules and the mineral particles in the clay; and fourth, surface pressure, which is the opposite of surface tension. Gravity —Surface tension, or the contracting power of any exposed water surface, may act with gravity or against gravity, depending upon circumstances. Molecular attraction between the mineral and water molecules always acts in opposition to gravity. Since, as can be shown, the conditions of capillarity in a mass of clay compressed into the form of a brick is such as to make surface tension the very much greater force, and operating in opposition to that of gravity, gravity will not be considered as one of the component forces in our problem. If we were dealing with “slips” or even soft mud fixtures, the force of gravity would have to be considered. Molecular Attraction —Milton Whitney 1 says: “The potential of a single water particle is the work which would be required to pull it 1 U. S. Dept, of Agr. Weather Bureau, Bull. 4, p. 19. 190 PAVING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 away from the surrounding water particles and remove it beyond their sphere of attraction. It is the total attraction between a single particle and all other particles which surround it.” It is called by some “mole¬ cular attraction.” Surface Tension —Because- it has particles adjoining it only on one side, i. e., molecular attraction is affecting it only from one side, the potentiality of a water particle on the surface is, according to Whitney’s definition, only one-half that of a particle in the center of a drop. That things tend to move from points of low to points of high potential is a well-known law of physics. The particles on the surface, will, there¬ fore, strive to get to the interior of the drop. The results will be sur¬ face tension. Looking at this proposition from the mechanical point of view, the force of molecular attraction operating on the'surface particles, is effec¬ tive along lines that extend from the center of each particle, to the center of the surrounding particles. Since the particle on the surface of a. drop of water is under the influence of other particles only from one side, the several lines of force would extend radically from its center to the center of adjacent particles, having as a resultant a line of force extending from the center of the surface particle to the center of the mass. Surface Pressure —Suppose that instead of a drop we have the same mass of water surrounding a solid particle as a film, say, 0.0005 m m. thick. We should have in this system two combating forces, first, mole¬ cular attraction of water molecules for each other, causing a pull on all water particles toward the center of the film, creating a tension on the outside surface as well as on the surface contiguous to the solid par¬ ticles; second, attraction between the molecules of the solid particles and those of the liquid, tending to create a tension only on the outer surface of the glm. Fig. 16. Diagram showing operation of forces causing surface pressure. Consider the water between four solid particles as shown in the following figure as having a potentiality less than that of the solid particles. All water particles will press outward the solid particles along the resultant lines of force as shown in Fig. 16. In this case instead of tension we would have a pressure. This pressure is known as surface pressure. If, on the other hand, the water had a potentiality that was greater than that of the solid par¬ ticles, the resultant forces of at¬ traction would be toward the cen- purdy] QUALITIES OF CLAYS FOR MAKING PAVING BRICK. 191 ter of the liquid mass as shown in Fig. 17. This would result in surface tension. The practical conclusion from the above discussion of greatest interest in connection with plas¬ ticity, is that when the surround¬ ing fluid has the greater poten¬ tiality, flocculation, or drawing together of the solid particles will result. When the solid 'par¬ ticles have the greatest poten¬ tiality, deflocculation or separa¬ tion of the solid particles will re¬ sult. Citations by the score could be presented showing that clays can be fioccula{ed, or de- fiocculated, depending upon the FIG ' 17 ' D cfurgSSS£ 0n ^ f ° rces material carried in solution by the water used in tempering. It is of interest, therefore, to consider the various solutions and their ef¬ fect on clays. Solutions causing deflocculation —Johnson in “How crops feed” cites a great many instances where solutions of organic compounds have caused deflocculation of soils. Ackison 1 has shown that tannin will de- flocculate clay so thoroughly that when a thin slip of clay suspended in a solution of tamin is poured onto a filter paper the water passing through will be very turbid. Ammonia is used in the water when a clay is being disintegrated preparatory to mechanical analyis. Pe¬ troleum is greedily absorbed by clay because of its low surface tension or potentiality, being held between the minute grains of clay by virtue of the higher potentiality of the clay grains. Whitney 1 has shown that cotton seed, meal, tankage, etc., have similar effects. It will be important to note that the surface tension of solutions which cause defiocculation of grains of pure clay substance is without exception lower than the surface tension of water. It will also be impor¬ tant to note that physical differences in conditions such as degree of concentration of the solution, temperature, etc., that tend to decrease surface tension affect defiocculation. For example, it is a common ex¬ perience of chemists that boiling for the purpose of extracting soluble salts often so thoroughly deflocculates the clay that even filtering through a Gooch crucible will not clearify the filtrate. In the following tables the surface tension of water and of various solutions is given. 1 A. c. S., Vol. VI, p. 44. 2 Bull. 4, Weather Bureau, Dept, of Agr., p. 17. 192 PAYING BRICK AND PAVING BRICK CLAYS, [BULL. NO. 9 Table XXII—(1.) The surface tension of water and alcohol in contact with air. Temperature C° Surface tension in dynes per centimeter. Water. Ethyl alcohol. 0°. 75.6 23.5 74.9 23.1 10.. 74.2 22.6 15. 73.5 22.2 20. 72.8 21.7 25. 72.1 21.3 30. 71.4 20.8 35. 70.7 20.4 40. 70.0 20.0 • 45. 69.3 19.5 50. 68.6 19 1 55. 67.8 18.6 60. 67.1 18.2 65. 66.4 17.8 70. 65.7 17.3 75.. 65.0 16.9 80. 64.3 85. 63.6 90. 62.9 95. 62.2 100. 61.5 Table XXIII (2). Miscellaneous Liquids in Contact with Air. Liquid. Temp. C 0 Surface tension in dynes per centimeter. Authority. Acetone. 14.0 25.6 Average of various. Acetic acid. 17.0 30.2 . .do. Amvl. alcohol. 15.0 24.8 . .do. Benzine. 15.0 28.8 . .do. Butvric acid. 15.0 28.7 . .do. Carbon disulphide. 20.0 30.5 Quincke. Chloroform. 20.0 28.3 Average of various. Ether . 20.0 18.4 . .do. Glycerine. 17.0 63.14 Hall. Hexane . 0 0 21 2 Schiff. Hexane. 68.0 14.2 . .do. Mercury. 20.0 470.0 Average of various. Methyl alcohol. 15.0 24.7 . .do. Olive oil. 20.0 34.7 . .do. Petroleum. 20.0 25.9 Magie. Propyl alcohol. 5.8 25.9 Schiff.'. Propyl alcohol. 97.1 18.0 . .do. Tolnol. 15.0 29.1 . .do. Tolnol. 109.8 18.9 . .do. Turpentine. 21.0 28.5 Average of various. 1 Smithsonian physical tables. Third revised edition, p. 128. 2Smithsonian tables, Ibid. PURDY] QUALITIES OF CLAYS FOR MAKING PAYING BRICK. 193 Table XXIV (1). Salts in Solution. Density. Temp. C° Surface tension in dynes per Cm. BaCl 2 . 1.2820 15-16 81.8 . .do. 1.0497 15-16 77.5 CaCl 2 . 1.3511 19 95.0 1.2773 19 90.2 HC1. 1.1190 20 73.6 do . 1.0887 20 74.5 . .do . 1.0242 20 75.3 KC1. .. 1.1699 15-16 82.8 . .do. 1.1011 15-16 80.1 . .do. 1.0463 15-16 78.2 MgCl 2 . 1.2338 15-16 90.1 . .do. 1.1694 15-16 85.2 . .do. 1.0362 15-16 78.0 NaCl . 1.1932 20 85.8 . .do. 1.1074 20 80.5 . .do. 1.0360 20 77.6 NH 4 C1 . 1.0758 16 84.3 . .do. 1.0535 16 81.7 . .do.. 1.0281 16 78.8 SrCl 2 . 1.3114 15-16 85.6 . .do. 1.1204 15-16 79.4 . .do.'... 1.0567 15-16 77.8 k 2 co 3 . 1.3575 15-16 90.9 . .do. 1.1576 15-16 81.8 . .do.,.. 1.0400 15-16 77.5 Na 2 CU 3 . 1.1329 14-15 7.9.3 . .do... 1.0605 14-15 77.8 . .do. -1.0283 14-15 77.2 K No,. 1.1263 14 78.9 . .do..;. 1.0466 14 77.6 CuS0 4 . 1.1775 15-16 78.6 . .do. 1.0276 15-16 77.0 h 2 so 4 . 1.8278 15 63.0 . .do. 1.4453 15 79.7 . .do ..*.. 1.2636 15 79.7 k 2 so 4 . ... . 1.0744 15-16 78.0 . .do..... 1.0360 15-16 77.4 MgS0 4 . 1.2714 15-16 83.2 . .do. 1.0680 15-16 77.8 Mn 2 SQ 4 . 1.1119 15-16 79.1 . .do... 1.0329 15-16 77 3 ZnS0 4 . 1.3981 15-16 83.3 . .do. 1.2830 15-16 80.7 . .do. 1.1039 15-16 77.8 Smithsonian Tables Ibid. Points to be noted in table—Notes : 1. Solution of organic compounds have a much lower surface tension than water. 2. Surface tension decreases as the tem¬ perature increases. 3. As the density of the solution increases surface tension in¬ creases. 1 From the above the following conclusions regarding deflocculation ap¬ pears : 1. It has been shown that solutions of organic compounds cause defloc¬ culation. It is needless to go into further discussion of this point, for the facts that have been stated are well understood by practical potters and agri¬ cultural chemists. 2. Increased temperature assists in producing deflocculation. Potters who- use hot water in their Plungers and brick manufacturers who use hot water in their pug-mills have learned that clays slake and develop plasticity more easily with hot than with cold water. These cases find their parallel in the laboratory when clay slip is boiled in the process of soluble salt de¬ termination. Deflocculation is increased by the use of hot water. 3. Increased density of a deflocculating solution does not increase its efficiency. “Ammonia (1) has a very marked action in breaking up soils containing particles less than 0.005 mm. in diameter. . . . One drop of l Bull. No. 24, Bureau of Soils, U. S. Dept, of Agr., p. 22-24. —13 G 194 PAVING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 ammonia (added to 5 grams or sample in 50 cc. of water) does not seem to be sufficient to break up the flocculations completely, but no great change is produced by the addition of more than 5 drops to 50 cc. of water” • The preceding facts given by our foremost agronomists, when consid- fered in the light of the fact that increased concentration of a solution increases its surface tension, are proof of the deduction that when the potential of the solid particles is greater than that of the surrounding fluid, deflocculation ensues. In the case of the ammonia solution, in¬ creased concentration by the addition of more than 5 drops of ammonia would so increase the surface tension and consequently the potentiality of the solution as the equalize the potentials of the soil particles and the solution. Solutions causing flocculation —Having discussed the deflocculating solutions in detail, it will not be necessary to dwell at length on the flocculating solutions, for the effect on clay of each class of solution is the converse of the other. It is important to note that solutions which have surface tensions higher than that of water tend to cause floccula¬ tion. The nature of solids affects flocculation in several ways. First, if the clay or soil under examination contains a large quantity of calcium or magnesium carbonate 1 , it has been found that solutions having a surface tension as low as that of ammonia will cause flocculation. Data are not available concerning clay mixtures high in other minerals, but as is about to be shown, clays that have comparatively low content of clay substance probably have as an average for the several mineral grains a low poten¬ tial in comparison with the potential of water. Clays high in products of decomposition of organic matter may be flocculated by ammonia. In fact the “potential” of the impure clays may be so low as to permit am¬ monia solutions to flocculate their grains. Pure clays, i. e., kaolins, re¬ quire for their flocculation solutions having a “potential” that is higher than that of pure water. Second, near 1 the surface of any soil there is a concentration of solu¬ tions. This is adsorption. If 2 the solid is exceedingly porous, this ten¬ dency to concentration near the surface is heightened. It is well known that salts, which are concentrated near the surface of solids are precipi¬ tated or at least are left clinging to the solids when the water is with¬ drawn. Soils 3 4 , even sand, possess the property of attracting and fully absorbing salts which cannot be wholly washed out by new quantities of water. Solutions of many of the salts are materially weakened when brought in contact with solids, because of the adsorption of the salts, but if the surface of the solid be relatively small no weakening of the solu¬ tion may be perceptible. Summary —The well-known facts concerning a plastic clay when wetted with water are, first, that its finer portions are composed of a countless number of minute grains, the composition of which has been shown to agree closely with that of pure clay substance; second, that even the coarser grains are composed largely of kaolinite and other minerals 1 U. S. Dept, of Agr., Bureau of Soils, Bull. 24, p. 24. 2 U. S. Dept. Agr., Rept. No. 64, p. 142. 3 Comp. Johnson, How Crops Feed. p. 173. 4 Comp. Johnson, How Crops Feed, p. 334. PURDY] QUALITIES OF CLAYS FOR MAKING PAYING BRICK. 195 cemented into clusters or bundles; third, that clays having a high con¬ tent of minerals other than kaolin, are flocculated by solutions having a surface tension lower than that of water, while the clays which are practically pure kaolinite in composition require for their flocculation solutions that have a surface tension higher than water; fourth, that clay particles extract salts from solutions and hold them near and on their surface at a high degree of concentration; fifth, that clay substance ex¬ hibits this property of adsorbing salts to a much higher degree than any of the common anhydrous minerals, a fact that makes the extreme fine¬ ness of the "clay substance in clays” of considerable significance. The known facts concerning solutions are: first, that all solutions have a surface tension which is increased with increased concentration; sec¬ ond, that those solutions which have a surface tension higher than that of pure water, tend to cause flocculation of kaolin grains. Oh putting together the known facts concerning clay and water, it is evident that the film of water surrounding the grains of clay , (when the mass is in a plastic condition) has a very high potential, owing to the high degree Of concentration of the salts that are held to the kaolin grains by adsorption. SUPPOSED HISTORY OF THE DEVELOPMENT OF PLASTICITY OF CLAYS IN NATURE. Daubree 1 , Cushman 2 and Mellor 3 have disintegrated feldspar in water either by grinding or by boiling. In all cases, the liquid in which the feldspar was ground contained alkali in solution. Mellor found that not only did the solution give alkaline reactions, but the <( outlines (of the solid particles) could be more readily stained ivith saffranine or with malachite green than before the action” Since the larger part of the clay substance is derived from the dis¬ integration of feldspar, it can be considered that there was formed at the time of "kaolinization” insoluble hydrous silicate, of alumina, soluble potash salt and soluble silicic acid. If feldspar has been disintegrated by atmospheric agencies, water and carbonic acid, and the residual mass is so situated as not to allow the soluble salts to be washed away, they will be retained in part by adsorption and in part by recombination, forming zeolitic masses. Data are not available to warrant the state¬ ment that plastic kaolins formed in situ owe their plasticity to these ad¬ sorbed salts, or that they even contain adsorbed salts. We do know, however^ that nearly all kaolins contain alkalies that can not be shown to be present as constituents of such minerals as feldspar or mica. Fur¬ ther we know that "the less 4 free alkali a clay contains the more will it adsorb.” We know also that clays which have been formed from feldspar under the disintegrating influence of fluorine are not plastic, and con^ tain fluorspar 5 and other fluorine compounds. 1 Am. Rep. Smithsonian Inst., 1862, 228. 2 U. S. Dept. Agr., Bull. No. 92. 3 Eng. Cer. Soc., Vol. pt. 1, p. 72. 4C. F. Binns. A. C. S., Vol. VIII, p. 206. 5 Jackson and Richardson, Eng. Cer. Soc., Vol. II, p. 59. (1903-4.) 196 PAVING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 Cushman 1 reports that the residue left after disintegrating the feldspar and washing out the portions which have been rendered soluble, is com¬ posed of very minute particles. If this insoluble portion, kaolin, had been formed by nature, in whose laboratory reactions, precipitations, etc., extend over an almost infinitely longer time than is given to similar physical-chemical phenomena in the laboratory., there is no doubt but that these minute particles of kaolin would arrange themselves in the thin plate-like clusters that are characteristic of this substance, just as did the Leasburg clay cited by Wheeler. Conditions will control the size of the plate-like crystals so developed. In many kaolins these plate forms are discernable, ranging from those of sub-microscopic dimensions up to those that can be readily measured in the microscope. In a clay examined by the writer not only Avere there crystals of measureable size, but they appeared to be compounded, i. e., made up of several crystals, which could not be separated to an appreciable extent by vigorous shak¬ ing in distilled water. Under natural conditions, therefore, where the disintegrating water readily runs or seeps away, carrying the soluble por¬ tions and leaving the insoluble “residual clay” in situ , we can expect to find a deposit that is more or less crystalline, depending upon the attend¬ ing conditions. These deposits, we know, are practically non-plastic. We know fur¬ ther from Ackison's experiments and the testimony of many agricultural chemists, that these grains can be deflocculated by organic solutions. Since surface waters generally contain organic substances in solution, and since proximity of vegetable growth can give rise to a deposit of decaying vegetable matter on kaolin beds, it is easy to see how such a de¬ flocculation can take place in situ , and especially so if the clay be moved by running water and deposited in the lower lands. By virtue of this deflocculation the clay has a smoother feel, i. e., texture, and thereby assumes a pseudo-plasticity. This fact has. given rise to the fineness of grain theory of* plasticity. These deflocculated particles of kaolin have, as has been shown, a high adsorptive power. Whatever salts may be in solution in the passing waters, or may be carried upward from lower strata by rising waters, will be adsorbed by the kaolin particles. Now, depending upon the de¬ gree of deflocculation, amount of adsorption, and the kind of salt so ad¬ sorbed, plasticity will be developed. When non-plastic kaolins are wetted with water, they are compressed into shapes only with difficulty, and when dried they either fall to pieces, as would so much fine sand, or have so weak a bond that they are easily crumbled. The finer the particles, as with the case of sand, the more readily can they be shaped into pieces that will retain their form, but no matter how finely sub-divided the grains may be, the mass is still very friable. In this fine condition the kaolin no doubt possesses every chem¬ ical and physical property possessed by the plastic kaolin (ball clay) save that of plasticity. It has water chemically combined, molecular attraction, and adsorbing properties. It 'becomes plastic only when it 1U. S. Dept. Agr, Bull. No. 92. PURDY] QUALITIES OF CLAYS FOR MAKING PAYING BRICK. 197 has adsorbed salts, the solution of which exhibits a high surface tension, or as Whitney would express it, which have a relatively high potentiality . Clays having adsorbed salts and consequently, plasticity are no longer friable when molded but, on the contrary, they are exceedingly hard. It is because of this adsorption property which in kaoline grains seems to be manifested to a higher degree than in any other mineral substance, with perhaps the exception of zeolites, that many find reason to believe that plasticity is due to a pectoidal structure of the kaolin grains. Since, however, they cannot show that those substances which are known to have a pectoidal or colloidal structure can be made to show or develop plasticity, and since colloids cannot be extracted from plastic clays, rendering them non-plastic, nor added to non-plastic kaolins rendering them plastic, we must conclude that this theory is hardly tenable. To what this great adsorptive power of clays is due has not as yet been determined. We, however, must accept the existence of this property as a proved fact. We must also concede that when water is added to a clay, that portion which envelops the very minute solid particles having a relatively large surface area in proportion to their volume, and hold¬ ing salts by absorption, will be highly concentrated, that the potential of this film will he very much more than those portions of the water farthest away from the solid particles; and finally, as shown by the flocculation of the clay particles, the potential of this saturated film of water is higher than the potential of the kaolin particles. The writer bases his assumptions as to the cause of plasticity on known facts: Adsorption of salts by the kaolin grains and the con¬ sequent high potential of the water film which surrounds the grains when a clay mass is in a plastic condition. On these assumptions, the cause of the latent plasticity when clay is dry and the developed plasticity when it is wet, are obvious. Fineness of grain, molecular attraction, adsorptive property, are conditions that permit of the adsorption of salts. In other words, they are necessary conditions. METHODS OF MEASURING PLASTICITY. General —There have been many methods devised for measuring placticity. The methods suggested by Zischokke 1 2 and Grout 3 seem to be the most rational of any, for in them the resistance to deformation and amount of flow before rupture, two characteristic properties of plastic bodies, are measured. These methods are based on the same principle as the well-known but crude method of testing plasticity by squeezing a ball of plastic clay between the tips of the forefinger and thumb, and making a mental note of the amount of pressure required to affect a given degree of deformation. The test developed by this Survey involves the tensile strength of the plastic mass rather than its resistance to compression, as in the 1 For description of these methods see “Clays; Occurrence, Properties and Uses” by H. Ries. Wiley .and Sons, 1906. 2 Thon-Industrie-Zeitung, No. 120, p. 1658. (1905.) 3W. Va. Geol. Surv., III. p. 40. 198 PAVING BEICK AND PAVING BEICK CLAYS. [BULL. NO. 9 Zschokke and Grout methods. It is believed that a tensile test gives a more accurate rating of the tenacity with which the several grains cling to one another, for in this, friction between the non-plastic grains and interference to flowage by the larger ones crowding into one another is entirely eliminated. Shape of the test piece —The special features of the shape and size of the briquette employed in this test are, first, narrow neck, (%"), wide ends (—") and straight sides to fit the jaws, as explained later. The smallest portion is cubical in shape, being %" x %" x The clips —In previous experiments it was learned that the Standard Fairbanks clips, using the standard shape and size of briquette, would permit the stretching of the briquettes until they would slip out of the jaws. Special clips were therefore made to fit the briquette. These clips were designed after Orton 1 , differing from his only in dimensions, angle of nip between the jaws, and manner of adjustment. Manufacture of briquettes —Clay was mixed to a thick slip, cast and cut into briquettes by the Fox method, as described under tensile strength. When the cast slab had, in the opinion of the operator, as¬ sumed its maximum plasticity, the briquettes were cut an.d forced into steel molds under a constant pressure of fifty pounds. This weight was applied slowly but the briquette was not allowed to remain under pres¬ sure after it had received its full load. Adjustment and calibration of the machines —Before the Fairbanks machine could be used, the balance beam had to be poised to allow for the difference between the standard and our special clips. For measuring the stretch which the briquettes suffered, the small adjusting wheel was calibrated so that the peripheral distance through which the wheel was turned would represent the distance the under clip had been lowered. The amount of stretch which the briquettes suf¬ fered at any time during the test was measured by the fractional num¬ ber of turns of the adjusting wheel required to lower the under jaw sufficiently to keep the beam in a predetermined position. Method of procedure —The plastic briquette was carefully placed in the clips and the jaws adjusted to it, care being taken to see that the jaws on either side were at the same angle. The lower clips, suspended by counterpoise, were kept in a vertical line by hand guidance. Very small shot was allowed to run into the pail slowly until a rupture oc- cured at the neck of the briquette. As soon as a rupture occurred, the beam dropped with a suddenness that shut off the flow of shot. At- the moment of rupture the amount of initial stretch was noted by the fractional number of revolutions through which the adjusting wheel had been turned. Before removing the “load” the adjusting wheel was slowly turned until the briquettes was completely torn apart. The sec¬ ond peripheral distance. through which the wheel had been turned was noted as “final stretch.” The weight of the shot required to'cause rupture was obtained on a balance that is accurate to one centigram. While the weight thus ob¬ tained is not the force that was required to cause rupture, it does bear a lAmer. Cer. Soc., Vol. VI, p. PURDY] QUALITIES OF CLAYS FOR MAKING PAYING BRICK, 199 constant ratio to that force. The shot was not weighed on the Fair¬ banks machine because it was not sufficiently sensitive. Plasticity modulus —It is obvious that since all three factors, initial stretch, final stretch, and load required to cause rupture, must be consid¬ ered as being affected by the degree of plasticity of the clay, a modulus must be devised that includes all three factors. The one used in our tests was constructed as follows: The central portion of the briquette is a perfect cube %"x 3 / 4"x%". On the assumption that the volume of this portion of the briquette re¬ mains constant throughout the test 1 , and that its cross section decreases proportionally as the length increases, the decrease in cross section in f ' s \ 1.9 3 ^ ^ centimeters due to the initial stretch would be( 1.9 2 x-.lor - V 1.9+a/ 1.9+a where a is the initial stretch. The decrease in cross section after the* final stretch (here it is figured as though there has been no rupture) x 1.9 3 1.9 3 \ would be equal to 1.9+a (~) 1.9+a+b or, by reduction, . 1.9 3 b where b is the final stretch. 1.9 2 +3.8a+1.9b+a 2 +ab Now a measure of the tension that is holding the grains together would be directly proportional to the load and inversely proportional to the decrease in cross section of the briquette due to stretching. The modulus must, therefore,- represent a value that is directly as the load and inversely as the product of the decreases in cross section due to the initial and final stretch. Performing this calculation and collecting; terms the following plasticity modulus is obtained: 1 (6.859 + 10.83 a + 3.61 b + 5.7 a2 + 3.8 ab + as + a2b)=M 24776b in which L = Ioad in centigrams, a the initial stand, b the final stretch. While the modulus is very formidable looking it was found that the test could be made and the plasticity factor calculated quite readily. In fact the entire test required less time than did the Grout test as carried out in our laboratories. With the heavy and far from delicate Fairbanks machine and the clumsy clips, plasticity factors were obtained that varied for any one clay not more than 50 per cent and in some cases only 13 per cent on six: briquettes. This percentage of variation was considered too high to at¬ tach any value to the obtained data, and they are, therefore, not reported. It is believed that with a more delicate apparatus this method of measur¬ ing plasticity would give very close results and .that the data obtained would be a true measure of plasticity. 1 This is no doubt an incorrect assumption. In iron they figure that the length increases four times as much as the cross section decreases, in other words that in the stretch the volume of the test piece actually increases. 2 The decrease in cross section of the briquette is calculated instead of taking the observed increase in length because it and the bond or strength of the mass- are directly proportional while the length is not. 200 PAVING BRICK AND PAVING BRICK CLAYS. [bull. no. 9 CHEMICAL PROPERTIES. Value of Chemical Analyses. Because, in private correspondence and on every public and semi-public oc¬ casion that has afforded opportunity, the writer has taken a stand against the large absolute value and importance of chemical analysis of clays that is contrary to views popularly held by practical clay workers and encouraged by manv of the State Survey reports, the discussion of the chemical properties of clays is intro¬ duced by liberal quotations from the most eminent of ceramic chemists. Dr. Her¬ mann Seger. “Thei demands which the cement and the ceramic industries make on the qualities of clay are as different as the purposes which these industries pursue. “In the manufacture of Portland cement we have in mind the obtaining of a product of a definite chemical composition, and, since the character of clay as such must completely vanish in this, the mutual relation of the indi¬ vidual constituents is to he considered above all things, and the physical condition in which these are found be considered only as far as it opposes greater or less difficulties to the destruction of the clayey character. “The clay industries, on the other hand, pursue a quite different purpose in the treatment of their raw material. The limits within which the chem¬ ical constitution of clay may vary are very wide, and, since the clayey char¬ acter of the material is to be preserved, its physical qualities and those of its essential and accessory constituents are to be placed in the foreground While for such a purpose, the chemical composition of clay, as a whole, ap¬ pears more indifferent and accidental, inasmuch as it depends on the mutual relation of clay, rock flour, sand, accidental admixtures and their chemical constitution, the physical properties of the same, the grain and its form, cap¬ illarity, plasticity, fusibility, etc., are of greater importance, and the chemi¬ cal constitution of each of these constituents is to be considered only as far as it permits us to infer the physical properties of the whole. II is surely a serious mistake to treat material so heterogeneous chemically and me¬ chanically, as the clays and earths used in the ceramic industries, like sub¬ stances chemically and physically homogeneous, as for example glass, and to base conclusions with regard to their properties on their chemical composi¬ tion. “The chemical changes which the materials of the ceramic industries suffer in the course of manufacture, step into the background, with the exception of the loss of chemically bound water, which has as a consequence the loss of plasticity, and must not be produced in the same degree as in the manufacture of cement and glass, or the material will lose its earthy character. In fact, it seems advisable to drop the investigation of the chem¬ ical composition of clay as a whole, and put in its place a deeper study of the composition of the essential and accidental constituents, in order to to infer the properties of the whole from the properties of the compounds thus obtained. For example, we need not ask how much pure clay and silicic acid we have in clay, but what part of the clay and silicic acid be¬ longs to the sandy constituents, what part to the silty, or the clayey con¬ stituent, and what physical properties must we, according to these data, ascribe to the sand, rock dust, clay, etc., individually. “It cannot be denied that in the examinations of clays scrupulously accur¬ ate analyses of the material have heretofore been made, but that little has been learned concerning structure, condition of plasticity, power of absorbing water, shrinkage on drying and burning, form and size of grains of sand, and rock dust, concerning the pecularities of the concretions, and concern¬ ing efflorescences and incrustations. In the consideration of the properties of clay for the purposes of the clay industries we must put ourselves more upon a physical than a chemical standpoint. l Dr. Hermann Segar, the Collected Writings of, Trans, by A. C. S. p. 8-11. PURDY] QUALITIES OF CLAYS FOE MAKING PAYING BRICK. 201 “Ifi chemical analysis has discovered a fixed relation between alumina, silicic acid and flux, we know that these constituents belong essentially to a single well-characterized combination, so that we can take the degree of refractioriness from the laws established by Bischofand Ritchers with a greater or less assurance, .according as this substance is present in a greater or less degree of purity. However, if we should proceed in a similar way in the investigation of brick clays, we would get a theoretical result so very different from the practical results that it would have no value what¬ ever in regard to the knowledge of the material. The chemical analysis gives us only an average of the composition of the components forming the clay, differing very widely in their chemical composition and their physi¬ cal properties. Since the clay, after burning, preserves its earthy charac¬ ter, and the various constituents act only superficially on each other, the chemical analysis gives us absolutely no clue for the deduction of definite properties of the whole. “Two brick clays may have exactly the same composition and still differ in every respect, because the complete analysis, for example, gives us no idea whatever as to how much silicic acid, alumina and flux belong to the clay substance, to the rock dust, and to the sand individually; for instance, in the one case all or the greater part of the flux may belong to the clay substance, in another case to the constituents which make clays lean, and accordingly the properties of the compound be subject to the greatest varia¬ tion with the same percentage composition. In the one case it may be the clay, in another the rock dust or sand, which, with the same percentage composition of the whole, is the most fusible constituent, as admixed iron oxide or carbonate of lime, which according to the manner of distribution are inclined to have the strongest effect sometimes on the clay, sometimes on the rock dust and sand, and thereby produce a number of variations which find not the slighest explanation by a simple chemical analysis. “If we conclude from this that chemical analysis can claim only a limited value for the- discovery of the properties of brick clay, such a judgment would be highly one-sided and inaccurate. . . . For our purpose it is especially the physical properties of clay that are of greatest importance in judging the same, and the chemical properties only as far as they supple¬ ment the former. Here, therefore, to express it in a few words, it will be the task of the chemical analyst to determine the composition of the con¬ stituents that are physically alike, that of the clay, rock dust, sand, and accessory constituents, separately and singly, and to make possible a com¬ parison of these with each other. In this way we are able to get a good idea of the properties of the components,! whereas an analysis of the whole mass would be of little use. We are thus convinced of the necessity of physical analysis of clay simultaneously with, or rather before the chemical, as far as the investigation is made for the purposes of pottery ware, and especially for the manufacture of bricks. Even though scientific men have repeatedly referred to the importance of the mechanical and physical inves¬ tigations, this direction of the investigation has not been pursued with such vigor that the results obtained from it show any real use for the brick industry.” In the foregoing statements Dr. Seger has very forcibly set forth the same doctrines that the writer has come to thoroughly believe as a re¬ sult of observations in the factory and laboratory. In subsequent writ¬ ings Dr. Seger set forth the value of what is known as the “Rational ..analysis” in which “clay substance,” feldspar and free silica are differ¬ entiated. He cited many cases in which, with the aid of the Rational analysis, he was able to obtain more clearly an idea of the constitution 1 Seger, loc. cit., p. 36. 2 Italics not in the original. 3 Italics not in the original. 4 Italics not in the original. 202 PAYING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 and properties of clay than he conld from any other method of analysis. In this, however, he was no doubt over zealous, for later studies by other chemists proved that not only does the “Rational analysis” fail to sharply differentiate between the “clay substance,” feldspar and free silica, but that the analysis is of value only in the purest clays, viz.: China and ball clays. The writer has made ‘rational” t analysis of many types of clays, and, barring those used in the vitreous pottery wares, he is compelled to state that not once has he been able to obtain a clearer insight into the actual constitution of the clays than he could from the gross or ultimate analyses. Predictions concerning a- cla}^ based on a rational analysis in the great majority of cases, go very far wrong. After considerable pains and labor in the execution of the analysis the operator is compelled to make guesses that are much less scientific and accurate than he would if he had merely burned a piece of the same clay in a small muffle furnace, and noted the rate of change in color and dem- sity with increasing intensity of heat treatment, a test that ought not to require more than three hours time, and can be made by any one who has access to a kiln. It was shown in a preliminary report on the fire clays of Illinois 1 that fire clays having the same ultimate chemical composition behave very differently in burning. Indeed the chemical composition and ulti¬ mate fusion period were very often found to coincide in clays which, on the one hand, would remain open and porous through sufficiently long and severe heat treatments to make them fit for use in fire brick, or, on the other, would be nearly vitrified under the heat treatment used in burning stoneware and sewer pipe. Such phenomena are discussed and illustrated in this report under the title of pyro-physical and chem¬ ical properties of clays. In the manufacture of vitreous floor tile the writer learned by prac¬ tical experience that particular effects either in color, vitreousness, ulti¬ mate fusibility, or any other physical property requisite in the produc¬ tion of floor tile, could not be duplicated on the basis of chemical com¬ position. This was also found true in the manufacture of porous white ware bodies, such as are used in jardinieres and art wares, and no doubt is en¬ tertained but that the same would be found to hold true to a large ex¬ tent in the manufacture of vitreous china. In these cases, however, rational analysis, i. e., the determination of the proportional quantity of clay substance, feldspar, and free silica finds value in that these sev¬ eral minerals have decided effect on the expansion and contraction of the blended pottery body, and, consequently, upon the proper fitting on it of a glassy coating (the glaze.) The only instance in which chemical analysis is of positive aid, aside from the explaining of some observed phenomenon, is in the execution and study of a systematically planned series of experiments. Seger’s classical studies that resulted in the invention of the pyrometric cones would probably never have been carried out had he not followed closely 1 Purdy and De Wolf, Preliminary Report on the Fire Clays of Illinois, State Geol. Surv. of Ill. Year-book 1906, pp. 137. PURDY] QUALITIES OF CLAYS FOE MAKING PAYING BRICK. 203 the chemical analyses of the raw materials and planned his series on chemical formulae. Following him, there has been much of exceedingly great value resulting from researches in pottery mixtures that would have been impossible on any other than a strictly chemical basis. In the study of paving brick clays here reported the fact has been discovered that the best paving clays contain a relatively high content of magnesia. Such a discovery has been and would have been impossible from an analysis of an isolated sample. Further, this fact would not have been noted had no systematic researches on a chemical basis been made with pure clays, minerals, and magnesia compounds, showing that mixtures containing a comparatively high content of magnesia have a long fusion range, for, as will be seen later, the value of clays for paving brick man¬ ufacture, or even their fusion range, do pot always .correlate with high magnesia content. The suggestion made by Dr. Seger in the paragraphs quoted in the introduction to this chapter, that a chemical analysis of the several sub¬ divisions of the particles according to size would be of value, is possibly true. In fact it is obvious why such should be the case. The time and trouble involved in making a thorough mechanical analysis of a clay into several groups having different ranges in size of particles in quantities sufficient to make accurate and especially duplicated analyses of each group, places such a determination out of consideration as a commercial test on clays. But for a scientific purpose it is believed that the re¬ sults obtained would justify the expense and trouble involved. Such a study was made by Grout on a composite sample of West Virginia clays. His results are cited and discussed on pages 179 to 180 of this report. So far as the writer is aware, Grout was the first to make such an analy¬ sis, and it is hoped that the deductions drawn from his results show justification for the making of similar studies on single clays. Notwithstanding the fact that up to date it would be but a matter of chance that an interpretation of the results of a chemical analysis would agree with the observed working properties of a clay, it should not be concluded that our chemists may not in the near future devise a method of analysis that will meet the requirements of the case. In fact, so strongly do many believe that this will come to pass, that they see justification in making and reporting chemical analyses of clays by geological surveys, as has been their wont in the past. The writer firmly believes, however, that as a forerunner to such an event, many carefully executed and systematic physical and chemical researches on each type of clay must be made with parallel observations on synthetical mixtures of pure minerals. A few such observations will be made in subsequent paragraphs. Mineralogical Composition of Clays. Clay is a heterogeneous aggregation of minerals in which kaolin is present' in sufficient quantities to give to the mass its characteristic physical properties. If kaolin is not present in sufficient quantities to 204 PAYING BRICK AND PAYING BRICK CLAYS. [BULL. NO. 9 do this the mass should not be called clay. If limestone contains some kaolin entrapped mechanically, the mass is a limestone, notwithstanding the fact that it contains a considerable quantity of kaolin, for it looks and behaves in every way like limestone. But if the lime should be dissolved, as we know has often been the case, until the material is lar¬ gely composed of kaolin, this residual mass is clay. There is a commercial modification of this definition that involves its economic use or value. For instance, if the mass should contain iron in sufficient quantities to render it a commercial source of iron, the mass is more properly called an iron ore, just as a limestone impreg¬ nated with zinc or lead is termed a zinc or lead ore. Aside from cera¬ mic consideration, a clay containing iron or any other substance of com¬ mercial value in sufficient quantities to allow of its being considered a source of that substance from a commercial standpoint would be con¬ sidered an ore. COMPLEXITY r OP MINERALOGICAL COMPOSITION OF CLAY. In the section, the “geology of clays,” Professor Eolfe has set forth in detail the most accepted theory of the origin of clays, the effective agencies of rock decomposition, and the manner in wdiich these agencies operate. It ha^, been shown clearly that the residual mass resulting from rock decomposition may be comprised of a variety of silicates, the . kind and number of silicates formed being dependent upon the con¬ ditions attending the rock decomposition. Professor Cook, 1 after giving the analyses of several of the New Jersey kaolins that differ widely in chemical composition, remarks: “The examples above stated prove conclusively that clays are not altogether uniform in composition, even after throwing out all the ac¬ cidental or foreign constitutents. Either the essential kaolinite is not constant, or our clays consist of this mineral mixed in varying propor¬ tions with other hydrous silicates of alumina. Inasmuch as the greater number of the rich fire and ware clays of the State, and also others which have been here examined, do correspond very closely to the com¬ position and formula assigned to this mineral, the latter explanation is more plausible " After nearly thirty years of constant research Dr. Cook’s problem is no nearer solution; for Dr. Clark in Bull. 125 of the TJ. S. Geological Survey, suggests that there are seven possible combinations of alumina, silica and water of combination, which might form crystalized kaolin. Professor Eolfe has shown that a pure kaolin can be formed only by the decomposition of rocks that consist almost altogether of feldspathic or other highly aluminous minerals, together with comparatively unde- composed minerals like quartz and mica, that are in large part separ¬ able from the kaolinite grains by elutriation. If the parent rock con¬ tains iron or other metallic oxide bearing minerals the residual kaolin will be contaminated with these coloring oxides in such a manner as to render its purification by elutriation impossible. If it is impossible to determine the mineralogical constitution or makeup in the former case, l Report on the Clay Deposits of New Jersey, Geol. Surv. of N. J. 1878, pp. 269-272. PURDY] QUALITIES OF CLAYS FOE MAKING PAYING BRICK. 205 where the residual deposit- is largely 1 pure kaolin, contaminated only with substances that are separable in running water, it is obvious that in the latter case, that of the impure deposit, the problem is far more com¬ plicated. The difficulty of determining the mineralogical composition of a clay .is increased many fold in the case of those that have been transported from the place of origin and contaminated with a heterogeneous as¬ sortment of inorganic materials encountered en route. Shales may vary so widely in their mineralogical constitution that in one case the mass may be highly ferruginous, in another highly calcareous, and so on, depending upon the amount and kind of contaminating substance. Because the shale is highly calcareous it does not follow that it is a simple mixture of calcium carbonate and kaolin, but rather that the predomin¬ ant adulterant is calcium carbonate. Silica, iron compounds, etc., may be and usually are present in considerable quantities in the calcareous shales. The nearest that geologists or ceramists have come to deter¬ mining what inorganic substances are present in a given shale, is sim¬ ply to say that it is calcareous or silicious, etc. It has not been found possible to determine the mineralogical composition of any of the com¬ plex secondary clays either by the microscope or by chemical analysis Approximation to the mineralogical composition of the purer secondary clays like the ball clays, is made possible by “rational analysis,” in which the differentiation of the minerals depends upon their relative solubility in sulphuric acid, yet by this method it is incorrectly assumed that only three mineral components are present, i. e., clay substance, felds|)ar, and quartz, and the results are forced to tally with this as¬ sumption. It is obvious, therefore, that an attempt to distinguish the minerals that occur commonly in clays would be useless in discussing the min¬ eralogical composition of clays in general, and much more so in the case of any particular clay. Granted that it would be possible to make a fairly accurate mineral¬ ogical analysis of a clay, it is doubtful if our present knowledge would enable us to predict its working qualtities or even its fusibility with accuracy. When it is considered that a mixture of 40 per cent quartz and 60 per cent feldspar has approximately the same pyrometric value as feldspar taken alone, and that both have like effect on the green prop¬ erties of clay, some idea of the complexity of the problem is apparent. What is true of a mixture of feldspar and flint is true of a large number of pairs of other minerals. What is true of minerals when considered in pairs, is true to a larger degree when taken in a multiple combin¬ ation. It does not require much imagination to see where one would be led if it should be required to predict the fusion behavior or a hetero¬ geneous mixture of a large number of minerals. This sort of a study is of value and in fact is now looked upon as a necessity in the compounding of artificial mixtures of clays and min¬ erals for pottery purposes, but in these cases the operator is dealing 206 PAYING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 with substances the mineral character of which is to a large degree known, and he is mixing these minerals in predetermined proportions. He has in this case a synthetical mixture of known mineralogical con¬ stitution adn of comparative simplicity (containing at the most not more than four or five different minerals, and, therefore, his practical experience ought to enable him to predict its physical and pyro-chem- ical behavior. In the case of nature’s mixtures, however, man has at present no way of determining their mineralogical constitution, and must depend upon an actual test for obtaining a knowledge of the working properties of the mixture. To illustrate these difficulties reference might be made to the fire clays, which are comparatively pure clay substance or at least rela¬ tively simple mixtures of mineral ingredients. It has been shown 1 that in plotting the position that indicates the relative fusibility of the clays on the basis of their alumina-silica ratio in reference to the position oc¬ cupied by a synthetical mixture having a similar alumina-silica ratio, no concordant relation existed between them. Further the difference between the No. 1 and the No. 2 fire clays of the usual clay workers’ classification having practically the same ultimate fusion point, but differing from one another in the manner of vitrification is no doubt explainable either on account of difference in mineralogical composi¬ tion or character of grains. What is true in the case of simple mineral mixtures like the fire clays would be greatly exaggerated in the case of the exceedingly complex mixtures, such as most of the shales and sur¬ face clays. Ultimate Chemical Composition. By ultimate chemical composition is meant the percentage by weight of the several oxides of the elements that occur in clay irrespective of their original state or combination. Ordinary chemical analyses are re¬ ported in terms of so much silican oxide, aluminium oxide, calcium oxide, etc. All are more or less familiar with such analyses, and not a few brick manufacturers have had repeated analyses of their clays made by chemists. The reports they received are what is known as the “Ultimate Chemical Analysis,” in contradistinction to the rational analysis, that gives the supposed approximate percentage of clay sub¬ stance, feldspar and quartz in the. clay, instead of the individual ox¬ ides of which these substances are composed. The persistent belief in the value of an ultimate chemical analysis on the part of layman and scientist alike is a not wholly unwarranted compliment to the science of chemistry. It cannot be denied that there is some slight foundation for this unflinching confidence in the value of an ultimate chemical analysis, but it is equally true that even after these many years of constant research by scientists the world over, very little advance has been made in ability to interpret the facts that ought 1 Preliminary Report on Fire Clays, State Geol. Surv., Bull. 4, p. 138. purdy] QUALITIES OF CLAYS FOR MAKING PAYING BRICK. 207 to be disclosed by such analyses. Because so many chemists, as well as laymen, do not seem to understand the difficulties that attend the interpretation of such an analysis, a brief review of a few of the recorded facts will be given. In 1868, Richters 1 in his classic work entitled “Refractoriness of Clays,” promulgated laws in regard to the fluxing effect of the various elements in simple mixtures at high heats that are now known as the Richters laws. In 1895, E. Cramer published in the “T'hon-Industrie Zeitung” a review of Richter’s work confirming his laws in every re¬ spect. The fluxing behavior of the various bases according to Richter’s laws are shown in the following curves. Figs. 18 to 21. K 2 0 Na^O CaO MgO FeO Fig. 18. Diagram showing operations of fluxes in kaolin, using equal parts of each. (Ex- ariple. kaolin = 98^, K 2 Q = 2$.) lFrom lecture notes by Prof. Edward Orton, Jr. 208 PAYING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 K 2 0 Na 2 Q CaO MgO FeO Fig. 19. Diagram showing operation of fluxes in kaolin using fractions of their weights;. (Example; Al a 0 3 2Si0,=222.8 at. wt. K a O 94.22 mol. wt. K a O mixture 2 0 1st vertical line. atomic =222.8+ PUR1>Y] QUALITIES OF CLAYS FOR MAKING PAYING BRICK. 209 Fig. 20. Diagram showing operation of fluxes on Al 2 0 3 +2Si0 2 +J4Si0 2 mixtures using equal weights of each. —14 G 210 PAYING BRICK AND PAYING BRICK CLAYS. [BULL. NO. 9 F. 21. Diagram showing the result of Richter’s investigation of various oxides. From Richters 5 and Cramer’s investigations it is learned that the order of fusibility of the different oxides in simple clay mixtures is as shown in the first column of the following table. In the presence of free silica the order is changed somewhat;, as is shown by contrasting the order given in the third column with that given in the first. Go¬ ing to the other extreme of silicate mixture, that of glazes, the order of the fluxing effect of the various oxides is, according to Seger 1 as given in the fifth column. Seger further says: “The law established by Richter and Bischof, concerning the fusibility' of clays, ‘that equivalent proportions of fluxes exert an equal influence on the fusibility, 5 and which appears to be ap- l Seger’s cpllected works. Vol. 2, A. C. S. Translatun, p. 568. PURDY] QUALITIES OF fcLAYS FOR MAKING PAYING BRICK. 211 Table XXV. Showing fluxing behavior of the various oxides in— Pure Kaolin. Kaolin+^j mol. flint. Glazes. Oxide. Molecular weight. Oxides. Molecular weight. Oxides. Molecular weight. Magnesia. 40 Magnesia.. 49 Lead. 222 Calcium. 56 Iron. 72 Barium.... 153 Iron. 72 Calcium... 56 Potash .... 94 Soda . 62 Soda. 62 Soda. 62 Potash. 94 Potash .... 94 Zinc. 81 Lime. 56 Magesia... 40 Alumina... 102 proximately correct for the very high temperatures employed in clay testing, and for the very small' quantities of the fluxes coming into ac¬ tion in the clays, has no bearing on the glasses and glazes, far richer in fluxes and melting at far lower temperatures.” Ludwig 1 having made similar studies with more complex mixtures summarizes his results as follows: “ First —Richters’ law is a special case of the general law of dilute solu¬ tions. Second —This law is restricted by the following correlations: a. It applies only to very dilute solutions, that is, clays with a small amount of fluxes and not to brick clays or glazes. b. Is assumes intimate mixtures. c. Iron shows a different effect, due to its two stages of oxidation, since one molecule of ferric oxide corresponds to two molecules of ferrous oxide. A given percentage of iron contains fewer molecules of ferric oxide than of ferrous oxide, since the former has a higher molecular weight. On changing to the ferrous oxide the number of molecules is doubled, and hence the fluxing action is doubled. Third. The analysis of a fire clay is of great importance in estimating its refractioriness. Fourth. The estimation of refractioriness by means of the percentage of alumina and fluxes leads to erroneous results.” From the above citation it must be concluded that the fluxing power of a given oxide is affected very materially by the kind and number of oxides present, as well as their chemical combination, degree of hydra' tion, oxidation, etc. The facts gleaned from a study of the fluxing effect of a single oxide in a simple mixture do not necessarily hold true in the same degree in complex mixtures. It is well known in glaze manufacture, for instance, that a mixture of several fluxes pro¬ vokes greater fusibility than a mixture of any two or them. What is true of glazes is likewise true of clays. The difficulty of interpreting the results of chemical analyses is more largely due to a lack of experimental evidence on the fluxing behavior of known complex mixtures. Interpretation of the facts concerning 1 Thon-Industrie Zeitung, Vol. XXVIII, p. 773, 190*4. Abstracted by Bleininger, A. C. S., p. 275. Trans. VII, p. 275, 1905. MELTING POINTS EXPRESSED IN CONES 212 PAVING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 a given mixture is impossible until there is more known about mixtures of the same component substances in different proportional combina¬ tions. For example, Seger 1 has shown that the fusibility of mixtures of pure AM). and silica as determined by Bischof can be represented graphically as in Fig. 22. Fig. 22. Seger’s Si0 2 —Al 2 0 3 curve. 1 Seger’s collected writings, Vol. I, p. 545, A. C. S. Trans. PURDY] QUALITIES OF CLAYS FOE MAKING PAYING BRICK. 213 Two important facts are shown in these curves. First. That the kaolin-silica mixtures are more fusible than the alumin¬ ium and silicon oxide mixture of an equivalent chemical composition. Second. That kaolin containing 58.2 per cent flint practically the same fusibility as one containing 88 per cent, while the kaolin-silica mixtures containing percentage of silica between these two limits are more fusible than either. SEGER COXES Fig. 23. Melting points of mixtures of magnesite and Zettlitz kaolin. (After Rieke.) Dr. Rieke 2 has shown that magnesite will flux kaolin, as is shown in Fig. 23. From Dr. Rieke’s results it would seem that a mixture of 85 per cent kaolin and 15 per cent magnesium carbonate has approx¬ imately the same fusion point as 43 per cent kaolin and 57 per cent magnesite. Dr. Mellor 1 has shown a similar fusion phenomenon with mixtures of feldspar and quartz, as exhibited in Fig. 24. Surprising as are the facts shown in these three curves, there has been but very little effort to determine similar relations between the several pairs of oxides and compounds, and practically none to demon¬ strate the fusion behavior of the several oxides and compounds in triple and quadruple combinations, and yet this is the very data that must be worked out before much can be accomplished in the interpre¬ tation of a chemical analysis. When ceramic technology reaches this 1 “Brick,” p. 170, Oct. 1906. 2 Trans. Eng. Cer. Soc., p. 51, 1904-5. 214 PAYING BRICK AND PAVING BRICK CLAYS. [BULL. NO. P Fig. 24. Mellor’s fusion curve for flint—feldspar mixtures. state of progress an explanation can perhaps be made regarding the fact that in some cases the admixture of the refractory kaolin will cause a lowering of the fusion point, while the admixture of a flux such as feldspar to the same mixture raises the fusion point. 2 In the following Tables XXVI and XXVII will be found the per¬ centages of the various oxides into which the clays considered in this re¬ port have been resolved. In Tables XXVIII and ^vXIX will be found the molecular composition of the clays as calculated from the analysis. In Table XXX will be found the results of a rational analysis of clays now used for paving brick manufacture in the State. Xo attempt to interpret this data can be made at this time. After a discussion of the pyro-chemical properties of clays this data will be re¬ ferred to with the endeavor to show all the possible relation there may be developed between the physical, chemical and pyro-chemical properties of clays. 1 See A. C. S. Trans.', Vol. V, p. 158. PURDY] QUALITIES OF CLAYS FOR MAKING PAYING BRICK. 215 Table XXVI. Si0 2 Fe 2 0 3 A1 2 0 3 CaC MgO Na 2 0 K 2 0 Moisture. Ignition. FeO Ti0 2 S K 1 .; 63.36 1.80 15.43 .93 1.58 .56 3.28 .48 1 6.99 4.02 1.00 27 K 3 .. 59.34 3.26 15.36 .76 1.82 .80 3.82 .29 7.89 3.84 1.31 .16 K 4 .. 60.31 5.04 17.74 .41 1.96 1.07 2.88 .81 6.71 1.96 .84 .14 K 5 .. 63.43 1.52 16.89 1.00 2.11 .20 2.03 .46 5.97 4.24 1.07 .11 K 6 .. 63.62 3.02 16.28 .63 1.44 1.50 2.60 .38 5.88 2.90 .96 .11 K 7 .. 59.86 1.42 17! 43 1.05 2.32 .18 2.80 .20 6.35 5.10 1.61 .13 K 14 .. 64.09 2.65 14.16 1.69 1.64 .77 2.90 .51 6.47 3.16 .89 .24 K 15 .. 58.03 2.91 17.72 1.42 1.43 1.40 2.66 .97 6.47 5.77 1.02 .25 F 1 . .. 58.52 4.99 15.67 1 1.05 1 1.45 1.48 2.94 2.02 7.72 1 3.37 .96 .32 Table XXVII. Si0 2 ai 2 o 3 Loss on Ignition. Fe 2 0 3 CaO MgO Na 2 0K 2 0 Moisture. K 2. 63.35 16.27 4.75 7.56 1.01 1.33 3.80 .31 K 8. 60.89 16.40 8.18 8.20 .55 1.61 4.15 .27 K 9. 68.50 16.98 3.54 5.77 .99 1.71 2.97 .50 K 10. 58.35 18.09 7.02 6.14 1.20 2.03 4.58 .81 K11. 55.18 19.22 10.45 8.19 .56 1.67 2.85 1.02 K 12. 54.37 23.61 10.09 6.14 1.58 1 61 2.78 .60 K13. 57.09 19.07 7.97 7.92 .80 1.91 4.69 .43 S 1. 55.02 20.35 9.40 6.26 .87 1.70 3 64 .83 S 2. 56.29 20.32 4.39 7.90 .48 2.01 4.46 .•79 R 1. 58.42 25.05 8.08 3.04 .46 1.52 2.30 1.29 R 2. 63.41 18.61 4.86 5.82 41 1.16 3.60 .68 R 3. 58.57 20.40 5.95 7.40 .63 1.37 3.27 1.06 R 4. 55.51 21.81 8.00 7.66 .56 1.63 3.56 .02 H 20. 47.29 15.51 13.11 4.80 7.33 6.19 3.71 1.31 H 23. 55.37 21.40 8.75 6.72 1.76 .65 2.41 3.39 H-II. 56.25 18.79 7.01 8.02 2.39 1.33 4.61 1.49 H-16. 60.93 17.93 5.73 8.12 1.33 .91 5.01 .55 H-17. 56.56 12.64 6.02 13.56 2 22 2.75 4.82 3.70 H-18. 39.91 16.43 21.20 4,80 7.57 5.08 3.71 .86 H-21. 48.41 18.31 12.79 6.06 5.73 3.13 5.65 .79 G-II. 63.42 16.24 5.14 6.62 1.64 1.87 4.83 .86 B-II. 60.31 19.11 6,70 6.14 2.73 1.73 1.44 3.05 Ill. 68.15 12.89 5.08 7.52 1.02 .59 2.93 1.57 J-II. 62.70 16.95 6.76 8.98 1.17 1.47 3.03 .98 L-II. 58.62 17.74 6.66 8.48 1.26 .98 ‘ 3.92 2.55 i Table XXVIII. Sample No. Location. CaO MgO k 2 o Na 2 0 FeO Fe 2 O s ai 2 o 3 Si 2 0 Ti0 2 K 1. Alton, Ill.. 0.110 0.261 0.231 0.059 0.369 0.074 1.00 6.98 0.083 K 3. Albion, Ill . 0 090 0.302 0.270 0.086 0.354 0.135 1.00 6.57 0.108 K 4. Springfield, Ill. 0.012 0.282 0 178 0.099 0.156 0.181 1.00 5.78 0.060 K 5. Rdwardsville, Ill. 0.108 0.319 0.130 0.020 0.356 0.057 1.00 6.38 0.081 K 6. Galesburg, 111 . 0.070 0.225 0.173 0.152 0.252 0.118 1.00 6.64 0.075 K 7. Streator P. B. Co. 0.109 b. 339 0.174 0.017 0.414 0.052 1.00 5.84 0.118 K 14. Western Brick Co. 0.217 0.295 0.222 0.089 0.309 0.119 1.00 7.69 0.080 K 15. Barr, Streator Ill. 0.146 0.206 0.163 0.130 0.461 0.103 1.00 5.57 0.073 F 1 Danville P. B. Co. 0.122 0.236 0.204 0.155 0.305 0.203 1,00 6.35 0.078 21 PAYING BRICK AND PAVING BRICK CLAYS. [bull. no. 9 Table XXIX. Sample No. Location. CaO r I MgO KNaO Fe 2 0 3 | A1 2 0 3 SiO* K 2. St. Louis, Mo. 0.113 0.208 0.305 0.296 1.00 6.62 K 8.. .. 0.056 0.250 0.331 0.319 1*00 6*31 K 9. Crawfordsville, Ind. 0.098 0 257 0 229 0 217 1 00 6 86 K 10 .... Terre Haute, Ind. 0.121 0.286 0.331 0.216 1.00 5T8 K 11 .... Brazil, Ind. 0.053 0 222 0.194 0.272 1.00 4.88 K 13 .... Clinton, Ind. 0.076 0! 255 0.216 0.265 1.00 5.09 SI. Moberly, Mo.... 0.078 0 213 0.234 0.196 1.00 4.60 S2.... Kansas City, Mo. 0.043 0.252 0.287 0.248 1.00 4 71 R 1. Nelsonville, O. 0.033 0.156 0Y20 0 077 1 00 3*96 R 2. Portsmouth, O. 0.040 0.159 0 253 0 199 L00 5.79 R 3. Canton, O. (Imp.). 0.056 0.171 0 209 0 231 1.00 4'88 R 1. Canton, O. (Metro). 0.047 0.196 0 213 0 224 1 00 4 33 K 12 .... | Brazil Fire Clay. 0.121 0.174 0.154 0.166 1.00 3.91 H-ll. Topeka, Kan. 0.232 0.181 0 321 0 272 1 00 5.09 H 16. Peoria, Ill. 0.135 0.129 0 365 0 289 1.00 5.78 H 17. LaSalle, Ill. 0.320 0 556 0 499 0.684 1.00 7.96 H 18.. . Sterling, 111. . 0 839 0.788 0 295 0 186 1 00 4 13 G-II .... Coffeyville, Kan. 0.184 0.294 0.390 0.260 1.00 6.64 I-II. Caney, Kan. 0.144 0.117 0.297 0.372 1.00 8.99 J- II. Pittsburg, Kan .. .. . 0 122 0.221 0 234 0.378 1.00 6.29 L-II. Lawrence, Kan 0 129 0 141 0 289 0.305 1.00 5.62 B-1I. Atchison,- Kan. 0 260 0.231 0.099 0.205 1.00 5.49 H 20.... Savanna, Ill. 0.861 1.017 0.313 0.197 1.00 5 18 H 21. Galena, Ill. .. 0.570 0.436 0.404 0 211 1 00 4 50 H 23. Carbon Cliff Shale. 0.149 0.074 0.147 0.200 1.00 4.39 Table XXX. Rational Analysis. Sample No. Clay Substance. Quartz. Feldspar. Phos. Carbon. Soluble Salts. K 1. 35.90 46.60 17.50 .093 1.44 .13 K 3. 48.00 26.74 25.26 .078 1.50 .14 K 4. 43.32 43.66 13.02 .024 .72 .04 K 5. 38.92 46.54 14.54 .090 1.26 .08 K 6. 41.02 39.98 19.00 .067 .63 .38 K 7 . 33.14 49.36 17.50 .079 .71 Trace K 14. 25.28 48.54 16 18 .069 1.01 .04 K 15 . 53.36 22.82 23.82 .125 .90 .27 F 1. 51.12 29.38 19.50 .077 .92 .14 PURDY] QUALITIES OF CLAYS FOR MAKING PAYING BRICK. 217 PRO-PHYSICAL AND CHEMICAL PROPERTIES OF PAV¬ ING BRICK CLAYS. [By Ross C. Purdy.] INTRODUCTION. Relations —In the discussion of the physical properties of clays it was shown that there is a possibility of making some correlations between the several physical factors. It was also demonstrated that the physical properties affect the adaptation of clays to processes of manufacture. No relation was found to exist between the chemical composition and working properties, so no attempt was made to correlate the chemical and physical properties. We are now to consider those properties of clays which are* manifested in the process of burning, and it is here that we should be able to trace the combined effect of the physical and chemical properties. In burn¬ ing, the physical and chemical properties of raw clays surely operate as causes having as effects the pyro-physical and pyro-chemical proper¬ ties. If, knowing the physical and chemical composition of the raw clays and the pyro-physical and chemical effects produced in burning, we are not able to trace a logical and invariable sequence between the causes and effects, we will be forced to admit: either (a) That accord¬ ing to the data at hand, clays having similar physical and chemical properties in the raw state, may behave differently in burning, or, (b) That it is at present impossible to determine exactly the physical and chemical condition of raw clays; or, (c) It is impossible to trace the effect of individual physical and chemical properties where so large a number of changes occur simultaneously; or, (d) That, reasoning from analytically determined causes to observed effects is an absurdity if the evidence does not permit of a reverse reasoning, i. e., from effect to cause. The first case, that of clays of similar character behaving differently in burning, is forcibly illustrated in the case of fire clays. Fire clays, having similar ultimate chemical composition and size of grain, may have radically different pyro-physical behavior. The one may burn to an open porous mass at cone 11, being fit for fire brick purposes; the 218 PAYING BRICK AND PAVING BRICK CLAYS. [BULL. NO 9 other may burn quite dense at cone 8, being fit for stoneware, sewer pipe, etc. This fact was noted in the preliminary report on fire clays, 1 and will be illustrated in this report under the topic “Changes That Take Place During Fusion.” The second case, the impossibility of determining exactly the physical and chemical condition of raw clays, is illustrated by the fact that in the more'exact of the two analyses, the chemical, chemists do not claim to be able with ordinary care and attention to details, to determine all of the elements that may be in a clay, nor do they claim to be able to determine the combinations in which these elements exist. The third case, or the impossibility of tracing the effect of several changes in physical and chemical conditions which take place simultan¬ eously, is a well recognized fact. On a rectangular coordinate diagram, two changes; on a triangular coordinate diagram, three changes in properties can be traced with accuracy. No simple 2 method has yet been devised by which the effect of changes in four factors can be traced, and it is certainly beyond the capacity of the human mind to follow the effects of four or more changes, if they cannot he plotted diagrammatically. In the case of several clays, no two of which agree exactly in their several properties, and in all of which there are a great many properties peculiar to the individual clays, it is manifestly beyond our ability* to satisfactorily folloy even all the known details. Broad generalizations with numerous and well-known exceptions are the best that experimenters have been able to make from synthetical mixtures of fairly pure clays. It is obvious, therefore, that with a heterogeneous assortment of impure clays, conclusions concerning the relation between the causes (the physical and chemical properties of raw clay) and ef¬ fects (the pyro-physical and chemical properties) cannot be other than broad generalizations. The last case, that of the absurdity of claiming validity for deduc¬ tions drawn by reasoning from cause to effect in cases where data do not permit of a reverse reasoning, i. e., from effect to cause, is very nicely illustrated in the work of Hoffman and-Desmond 3 where an attempt was made to devise an indirect method of determining the refractori¬ ness-of clays. With a given furnace operating on a predetermined time- temperature schedule, they thought they were successful in determining the relative refractoriness of clays by toning up low grade clavs with the addition of refractory material, and toning down high grade clays by the addition of known amounts fluxes, until the clays had the same refractoriness under the same heat treatment. This scheme worked 1 State Gaol. Surv. of Ill.. Year Book 1906. p. 138. 2 Solid figures are used by physical chemists in depicting the combined effect of more than three factors, but the drawing of such figures to scale according to given data presents difficulties which the writer, at least, has been unable to sur¬ mount. Judging from the fact that ceramists and other technical scientists have not as yet used solid figures it must be inferred that others have also found the involved difficulties insurmountable. 3 Trans. Am. Inst. Min. Eng., Vol. XXIV, p. 32. PURDY] QUALITIES OF CLAYS FOR MAKING PAYING BRICK. 219 nicely until they assumed definite temperatures and attempted to pre¬ pare mixtures that would fuse at these temperatures. In the first in¬ stance they adopted a certain combination of “causes” and measured the “effects.” Tn the second instance they adopted an “effect” and at¬ tempted to determine the combined “causes” that produced this effect. In this they failed so utterly that they abandoned this indirect method of estimating refractoriness. If their careful researches demonstrated no other fact than the futility of attempting to draw conclusions con¬ cerning the relation between cause and effects, when the data show this relation operating only in one direction, i. e., only from cause to effect or from effect to cause, their work was worth while and their report a valuable addition to ceramic knowledge. Relative Importance of Raw and “ Bw'ning” Properties —It is plain that the physical properties of a raw clay influence its behavior mainly in the machines and dryers. True, the physical properties have their influence on the burning behavior of clays, and, as in case of size of grain, if the causes of the physical properties were determinable, their findings would be of value in predicting and explaining the properties developed in burning. Size of grain, as will be shown, is an important factor in the case of pure minerals, but when the grains do not have a homogeneous mineral composition, but are, in the main, clots of minute particles of several minerals, or particles of the same mineral substance cemented together, any data concerning the influence of fineness of grain on the properties developed in burning are apt to be very mis¬ leading. Grout’s analysis of the grains of clays, given-on pages - shows that the grains are not individual particles but are aggregates, and Fox’s results, cited on pages--— confirm the conclusions drawn from Grout’s data. The writer has ground impure clays until they passed sieves of different meshes ranging from 10 to 200, molded the clays -into cones and noted the effect of fine grinding on the refractori¬ ness of the resulting masses. The difference between the ultimate fusion failure to stand erect under high heat treatment of the cones prepared from the same clay but differing in size of grain, was hardly observable. True there was a difference in that the finely-ground sam¬ ples vitrified earlier and did not lag as much in bending over so that they could be said to be a trifle less refractory. In no case, however, was the difference in refractoriness between the 10 and 200 mesh sample of the same clay more than 20 to 40 degrees centigrade, as measured by LeChatelier electric resistance pyrometer. Indirectly, fineness of grain affects the burned product in that in¬ ternal fractures produced in drying and lamination in the machine dies caused by extreme fineness of grain weaken the finished product. These and similar considerations are not properly considered under the topic of Pyro-physical and Chemical Products. The main consideration, in an analysis of the influence of the sev¬ eral properties of clays, is their influence on the character of the pro¬ duct manufactured from the clays in question. In the case of paving 220 PAVING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 brick the desired character of product is toughness or resistance to im¬ pact and abrasion. If coarse as well as fine grained clays, plastic as well as non-plastic clays, and tough clays or clays that show but little tensile strength, can be burned so as to make tough bricks, it is obvious that it will be impossible from such physical data to predict the char¬ acter of ware which a given clay will make. Inability to trace the in¬ fluence of so many factors may be largely responsible for this seeming lack of relation between the physical properties of the raw clay and the properties of the burned ware, but the fact remains that such is the case. On the other hand it can be shown that there is a possible or seem¬ ing relation between pyro-physical and chemical properties and the properties of the burned ware. Such a relation has been shown to exist in the case of fire clays. In the case of paving brick clays there is not quite so distinct a relation between these factors, but still it is observable. The study of the pyro-physical and chemical changes pro¬ duced in clays by heat is, therefore, of considerable more importance in the 'study of paving brick clays than the study of the physical prop¬ erties of the raw clay. DEHYDRATION. Nature of process —Pure kaolin, the basic c^y substance, contains in round numbers 14 per cent of water, chemically combined. At ordin¬ ary drying heats the amount of this chemically combined water in the kaolin is supposed to be unaltered. In fact, there is experimental evi¬ dence to support the belief that there is some water mechanically re¬ tained by the clay even at the highest heat ordinarily attained in any dryer, but this has no relation to the chemically combined water. Since, however, in the ordinary clay or shale but a fractional part of the whole is kaolin, ranging from a possible 100 per cent in the purest varieties down to 25 per cent or less in the more impure clays, it is not surprising that the amount of chemically combined water varies greatly in the different clays. Even in the purest it varies to some extent, amounting in some cases to more than 14 per cent. In these not rare cases some other hydrous minerals are supposed to be present that carry a higher percentage of combined water. It is aside from our purpose to dwell upon the kind and nature of the hydrous minerals that may occur in clay except to note that, if they occur in the purest types of clays, and especially those which have not been moved from their place of formation, it is reasonable to suppose that in a heterogeneous mixture of minerals such as shales seem to be, these highly hydrous minerals may in some cases be present in considerable quantities. Since each hydrous mineral substance retains its chemically combined water with a tenacity peculiar to itself, it follows that the period of dehydration of clays will vary with each variation in quan¬ tity and kind of hydrous minerals present. Likewise the physical alteration in the mass at this period will vary with each variation in purdy] QUALITIES OF CLAYS FOR MAKING PAVING BRICK. 221 kind and quantity of hydrous minerals present. Since, however, it is impossible to gather reliable data as to the mineralogical constitution of the impure clays, the quantities of these hydrous minerals present must be mere speculation. The varying effects produced during the period of dehydration, which probably originate in variable mineralog¬ ical composition, are the only known or determined facts in the case. From the foregoing considerations it is not surprising that the tem¬ perature of dehydration has been considered as ranging from 550/ to 650° centigrade (990 to 1170° Fahrenheit), and that there are Clays which can withstand a heat treatment of 16 hours duration~~alra tem¬ perature which will average during this period at least 650° C. without entire loss of plasticity. Six clays (K 5, H 16, K 8, K 13, K 14, K 15) after subjection to a heat treatment supposedly sufficient to affect complete dehydration, slaked down in water to a red plastic mass similar to that produced from hard shale on weathering. If it is true that on dehydration clay loses the properties that cause the mass to exhibit plasticity then these clays were not dehydrated. If clays that have been subjected to just sufficient heat treatment to cause their complete dehydration still retain consid¬ erable plasticity, then many will have to change their conception as to the cause of plasticity, for surely nearly, if not all, of the physical properties of the. kaolin particles must be altered by dehydration. These six clays tested continued to lose weight after this period. This loss may possibly be accounted for by the loss of volatile matter other than chemically combined water. In the absence of analytical data, however, it was fair to assume that this additional loss was in part at least due to the further expulsion of the chemically combined water. If this, as¬ sumption is correct, these cases would indicate that the usually allotted range in temperature for this period is altogether too limited. If a clay can withstand heat treatment for 16 hours at a temperature that ranges from 500° to 740° C., with an average equal to 650, without complete loss of its combined water, it is fair to conclude that the max¬ imum temperature limit for the dehydration period is above 700° C. Loss due to Constituents other than Combined Water —The actual loss in weight of a clay, aside from loss of the chemical water, up to this temperature may in part, according to Prof. Orton 1 be accounted for as follows: Vegetable tissues, such as roots, leaves, etc., ignite and burn at about 300°C. Bituminous matter, common to shales, ignites and burns between 300 and 400°C. Graphitic carbon, does not ignite much before 500°€. Sulphur distils from iron pyrites between 400 and 600°C. Calcium carbonate decarbonizes between 600 and 1000°C. Ferrous carbonate decarbonizes between 350 and 430°C. lAmer. C. Soc., Vol. V, p. 222 PAVING BRICK AND PAVING BRICK CLAYS. [bull. no. 9 The loss of any or all these constituents would not materially affect the plasticity of clay, and in the main these reactions would be com¬ pleted before or at the same time as dehydration. In caes of K 14 be¬ fore referred to, they had all evidently been completed before the com¬ pletion of dehydration, except perhaps the decarbonation of the small amount of contained calcium carbonate. The bricks were thoroughly oxidized and normal salmon-colored throughout. In this case the only possible conclusion seems to be that dehydration of clay requires more heat than heretofore supposed. It has been demonstrated by Hopewood 1 that, aside from the loss of combined water, solid carbon, carbonic acid gas, sulphur, etc., quite a large variety of acids and bases 2 are expelled from the clay by vola¬ tilization at temperatures below the maximum required for complete dehydration. The evidence given to Hopewood’s experiments, together with the vast accumulation of data by agricultural chemists, makes if appear as though the absorbed as well as the absorbed salts are seriously affected during this period. Direct evidence is not at hand that would throw light on this question, but the value of such evidence is con¬ sidered by the writer to be of such importance that an extensive re¬ search dealing with this subject has been outlined. It is anticipated that the manner in which this period of burning (dehydration) is conducted will be found to play a very significant role in the character of the ware developed in “the finishing heats.” OXIDATION. General Conditions. Definition of terms —“Oxidation” and “Reduction” are chemical terms referring respectively to taking on and giving off of oxygen. When a piece of iron is rusting it is becoming oxidized, i. e., the metal (Fe) is being converted to an oxide of iron (Fe20s) which is red in color. Iron rust can be reconverted to the unoxidized metallic state again by application of heat under reducing conditions, i. e., condi¬ tions that favor separation of the metallic iron and oxygen. When the quantity of oxygen in combination is reduced, then it is said that re¬ duction has taken place. When the quantity of oxygen in combination has been increased then.it is said that oxidation has taken place. Evidence of reduced condition in raw clay— “Blue” clay and dark gray shale owe their characteristic blue color, in the main, to two classes of substances, (1) the ferrous compounds, principally ferrous car¬ bonate and (2) carbon. Partially metamorphosed carbon adds to-a clay mass its characteristic black, just as does lamp black when added to what would otherwise be a white mass. Lamp black, an amorphous form of carbon, is the product of decomposition of carbon compounds 1 Trans. Eng. Cer. Soc., 1904-5, p. 37. 2 R. K. Meade has shown analytical data in support of his claims that the alkalies in cement mixtures are expelled during the burning. “Portland Cement” p. 124. J. W. Mellor also shows with data that the loss of alkalies from fire clays fired at 1400°C., amounts to 20 per cent of the total alkalies present in the unburned clay. Trans. Eng. Cer. Soc. Vol. VI, p. 130. PURDY] QUALITIES OF CLAYS FOE MAKING PAYING BEICK. 223 under the influence of heat, resulting from conditions that prevent its complete oxidation. The carbon in shales, at one time a part of the fibrous tissue of living plants, was buried in deposits of sea mud, and is found today in this same mud hardened into shale. Therefore, the dark iron compounds and the metamorphosed remains of carbon compounds combine to give the characteristic blue color to shales and many fire clays. Evidence of oxidation in raw clay —Where the shale is covered with only a very thin “stripping,” the color of the upper three or four feet of the bank will be red. In the lower portion of these red strata the color shades off gradually into the blue of the more solid strata below. In this red portion near the top of the bank the ferrous compounds have been oxidized to ferric compounds by the action of the oxygen from the atmosphere. Below the belt of weathering, the clay retains its blue color owing to the fact that either air cannot penetrate to those depths or that its oxygen is largely spent before it can reach the lower limit of the belt of weathering. It is observed that oxidation starts at the surface and proceeds downward. The depth to which evidence of oxidation can be seen depends upon the nature and amount of the oxidizable mineral present, the solidity of the rock mass, the prevailing atmospheric conditions and the length of time of exposure. Oxidation of Clay in Buening. The very same processes that are effective in oxidizing the blue shale to “red outcrop” are operative in burning when the blue clay brick is converted into “salmon brick.” In nature, at ordinary temperatures and under varying conditions, this oxidizing process is very slow, but in the kiln at temperatures ranging from 500 to 800° centigrade, with the high draft that is usually maintained at this early stage of the burning, conditions under which oxidizing processes operate are very much intensified and consequently comparatively rapid in their action. In the case of surface clay, and red clays generally, oxidation is so rapid that the lag in time incident to raising heat in a large kiln of relatively cold ware is sufficient to complete the oxidizing processes. In the case of many of the shales, the time required to completely oxidize the clay is so much longer that either the burner must “hold the kiln at red heat” for a time, or, especially in the case of bricks which have been set wet, evidence of incomplete oxidation will be very evident when the bricks are drawn. The change in color from blue in the “green” ware, to red in the salmon is the result of oxidation. Red surface and black centers are results of incomplete oxidation. These changes in color are the same indicators of oxidation and lack of oxida¬ tion noted in the case of shale in the bank. SUBSTANCES IN CLAY" THAT AKE AFFECTED BY OXIDATION. In general terms, the oxidizable substances in clays are carbon com¬ pounds, carbonates, nitrates, sulphites, etc. The most noteworthy ox¬ idizable substances in clays are : Carbon and the carbon compounds, ferrous carbonate and ferrous sulphide. 224 PAYING BRICK AND PAYING BRICK CLAYS. [BULL. NO. 9 Carbon and the Carbon Compounds —Carbon is present in practically all of the secondary clays in forms ranging from unaltered vegetable matter, humus and its compounds, to the Jiighly metamorphic carbon- graphite in graphitic shales. The least altered carbon ignites and ox¬ idizes most easily and the highly metamorphosed carbon most difficulty. To the decomposing carbon compounds and their by-products, the or¬ ganic acids, are due many of the physical properties of clays. It has been shown in earlier pages that organic acids are the main agents that cause deflocculation, a condition that must exist before plasticity can be developed. It could be readily shown that humic acid (C 20 H 2 O 9 ) with its peculiar properties of absorbing and holding heat, moisture and soluble salts, is a very active agent in promoting chemical changes in the mineral ingredients of clay, thus altering the physical condition of the mass. Unaltered carbon compounds and their by-products are, therefore, not only easily oxidized in burning, but have been highly ben¬ eficial in that they have promoted the development of those physical properties which, if the carbon is not in excess, permit of easy manufac¬ ture into wares. The more metamorphosed the carbon compounds the less active they are in promoting physical and chemical alterations in the clay mass and the more difficult are they to oxidize in the kiln. For these reasons fire clays and clay shales in which the carbons compounds have been com¬ pletely converted to graphite are—within small areas—more constant in their properties, thus being more constant in their working and burning behavior, and at the same time, more difficult to burn. Ferrous Carbonate —Ferrous carbonate occurs in clay in various phy¬ sical conditons and sizes of grain. Large concretions—“nigger heads’*— which are often composed mainly of ferrous carbonates, are to be seen in most shale banks. Ranging in size from 12 to 18 inches in diameter, down to minute, almost microscopic particles, these concretionary and globular forms of ferrous carbonate play a role in burning clay wares which, while most peculiar, is but little understood. The ferrous car¬ bonates that exist as finely precipitated powder surrounding the other mineral grains must, in burning, pass through the same chemical alter¬ ations as the ferrous carbonate in lump form, but under such-different conditions that distinction must be made between its behavior when in these two conditions of aggregation. One of these large ferrous carbonate concretions pulverized, pressed into brick form and burned under the same heat' treatment required to produce pavers from the shale in which the concretion was found, produced a brigh red brick which possessed a toughness that was equal to that of the brick made from shale. This experiment proved that the clay mass which is bound together by ferrous carbonate in a mass so hard as to wreck ordinary crushing machines like dry pans, and contain¬ ing a comparatively large quantity of ferrous carbonate, can be burned as safely and into just as good brick as the softer shale containing but a small quantity of ferrous carbonate (3 per cent of total ferrous iron.) PURDY] QUALITIES OF CLAYS FOR MAKING PAYING BRICK. 225' Ill this brick made from the crushed concretion there was practically no carbon, while the- shale contained three quarters of one per cent. While it is true that the carbon content of the shale is so small that no difficulty is experienced in thoroughly oxidizing the mass under the. time temperature schedule required to raise heat uniformly in a large kiln, yet it is a significant fact that occasionally unoxidized brick are drawn from the kilns,, and that the mass containing a large amount of ferrous carbonate was perfectly oxidized under similar kiln treatment. Singer 1 has shown that the acid radical (CO 2 ) is expelled from fer- ous carbonate at temperature below 430 C. The basic radical (FeO) would thus be given ample time to become thoroughly oxidized to FesO« or FeaO before the temperature could be raised sufficiently to cause fusion between the ferrous iron and the silicates. Under normal kiln treatment complete oxidation of the iron would be effected, provided the clay mass contained but a small amount of carbon. In the almost total absence of carbon, our experiment with the concretionary mass proved that the iron could be quite readily oxidized. As the carbon content increased, the difficulty in oxidizing a given amount of ferrous iron would increase, for between carbon and oxygen there is a stronger affinity than between iron and oxygen. In case there is a high content of both carbon and ferrous carbonate, time would have to be allowed in burning to com¬ pletely burn out the carbon before the heat is raised. If this should not be done the ferrous oxide would flux with the silicates causing an early fushion in the unoxidized portion of the brick. In case the carbon is easily ignited and burns freely it has been found that the fires in the furnaces have to be drawn, all air supply shut off and the carbon allowed to smolder until completely burned out. If these precautions are not taken in such cases, the heat from the burning carbon will raise the temperature in the kiln to the point where the fer¬ rous iron will be slagged with the silicates. In fact, the iron that was originally in an oxidized condition would be reduced, and the whole iron content thus be brought to its most active fluxing condition. Where the carbon is less inflammable, a longer time would have to be allowed for its complete combustion, but such stringent precautions would not have to be taken as in the case where the clay contained more inflammable carbon. The chemical explanation of these cases is that although the CO 2 radical is expelled from ferrous carbonate at an early stage in burning, the basic radical (FeO) cannot receive the oxygen required to con¬ vert it to its less active fluxing form, i. e., to Fe 20 s as long as there is carbon left in the clay mass. Carbon having a greater affinity for oxy¬ gen than the ferrous iron will withhold it from the iron. If a clay con¬ tains insufficient carbon of an easily inflammable variety, or, if the car¬ bon, even though present in quantity, is difficultly inflammable, time must be allowed to permit the oxygen to penetrate the brick, for oxida¬ tion proceeds from the exterior towards the interior in a manner similar to the oxidation of shale in the bank from the outcrop downward. l Class exercise under Orton, Ohio State Univ. —15 G 226 PAYING BRICK AND PAYING BRICK CLAYS. [bull. no. 9 High content of ferrous carbonate does not in itself mean that trouble will be experienced in oxidation, nor does a high content of thoroughly oxidized iron considered alone indicate immunity from oxidation troubles. The substance that controls the manner in which the oxida¬ tion .period of burning clay wares must be conducted is carbon. Burn¬ ing carbon not only will prevent oxidation of the ferrous iron, but will reduce the iron that may have originally been in a thoroughly oxidized condition. It depends, therefore, upon the amount and form of carbon present in a given case, as to whether in burning there must be allowed a short or long oxidizing period. Ferrous Sulphide —This very frequently occurs in clays as bright 3 ’ellow or white crystals. The first of these forms is often mistaken for gold because of its similarity in color. It is commonly known as “fooTs gold.” Mineralogically it is known as iron pyrites or marcasite, depending upon its crystalline form. If clay containing pyrites is loosened and allowed to weather, the pyrites will be desulphurized. The iron will, in the dry, oxidize to the hematite (FesCb), or, if moisture i§ present, to limonite (2 FesOs- 3 H 2 O). The sulphur will at the same time oxidize to sulphurous or sulphuric acid. By weathering, therefore, iron pyrites can be thoroughly oxidized and the sulphurous and sulphuric acid removed in solution by percolating waters. These reactions require time, especially under dry conditions. Brick manufacturers cannot, under the existing trade con¬ ditions, weather their clay. The face brick manufacturer, therefore, must allow as little time as possible to elapse from the time > that his clay is mined' until it is under fire in the kiln if he wishes to avoid that bane of the face brick manufacturer, scumming, which results from the formation of soluble salts by the sulphurous and sulphuric acid from iron pyrites. To the front brick manufacturer, the presence of iron pyrites is not, aside from the question of scumming, a serious disadvantage, for the black-slagged specks resulting from ferrous iron from the pyrites fluxing with the silicates is not objectionable to architects. If, however, a clean buff brick is resired or if, for any reason, the smoother and more uni¬ formly distributed black specking by the use of pyrolusite (MnO) is needed, then a clay practically free from iron pyrites must be used. In face brick, soundness and color are the prime requisites. In pav¬ ing brick, toughness alone is the prime requisite. If a clay contains sul¬ phide of iron (pyrites) scattered throughout the mass, local slagged spots scattered all through the brick will be formed in burning. These slagged spots will be spongy or vesicular, i. e., full of cavities, just as is the black warty mass that appears on the face of a brick made from a pyritiferous clay. The local fused spots are detrimental to the tough¬ ness of the brick, not only because they are spongy but also because they tution to Behavior in Fusion( CHEMICAL COMPOSITION. Historical —Search in ceramic literature disclosed the fact that prac¬ tically no data have yet been published that have a direct bearing on the relation of chemical and physical constitution, behavior of clays in fusion, and toughness of the burned ware. Ogden 1 did some prelimin¬ ary work on the relation of composition to toughness in porcelains and found the remarkable fact that increase of clay content from 30 to 60 per cent caused a decrease in the toughness of porcelain. Inasmuch as he employed the “rattler test” in determining relative toughness of l Trans. Am. Cer. Soc., Vol. VII. 240 PAYING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 his bodies, his studies are directly applicable to the. study of paving brick clays. While the development of toughness has not been shown to have a direct relation to the rate and manner of vitrification except in our own results, yet that such a relation exists can be assumed until other evidence proves the contrary. If this assumption is correct, Og¬ den’s results would show that the evidence developed by metallurgists to the effect that addition of either aluminum oxide or silicon oxide not only raises in degrees centigrade the period at which fusion is com¬ pleted, but also increases the viscosity of the molten mass, and the rate at which verification takes place, is not applicable to certain mixtures It must be admitted that before Ogden published his results, ceramists entertained the belief that the greater the content of AhOs and SiO* in clays, the greater would be the toughness. The findings in the case, of fire clays here reported confirm Ogden’s ideas. In the following paragraphs will be given such evidence as seems to bear on this point. Effect of AW* in Ceramic Mixtures —It has been known for some time that the addition of AhOs to clays and clay mixtures increases their refractoriness 1 . Fire clays, high in AhOs, are, as a rule, the most re¬ fractory. AhOs not only raises the actual period at which fusion is completed but also causes the wares made from aluminous clays to soften and deform very slowly. The slower softening and deformation of ware made from aluminous clays has been attributed to increase of viscosity 1 of the mass caused by alumina. The writer has shown 2 that the addition of AhOs as a constituent of clay to stoneware glazes until the proportion of alkali and alkaline earth to alumina was 2.5 to 1, not only rendered the glaze more fusible but also less viscous. Additions of AhOs above this proportional amount in¬ creased the refractoriness of the glaze, if not its viscosity. Addition of AhOs as a constituent of feldspar did not have as great effect on the fusibility of the glaze as did the same equivalent of AhOs from clay, notwithstanding the additional alkali that would be introduced by the feldspar. From these stoneware glaze studies it was concluded that it was not so much a question of quantity of AhOs, but of the manner in which it was added. If added as a constitutent of clay it is already combined with silica and water. Whether it is this mutual solution of calcium carbonate and clay that caused greater ultimate fusibility in the stone¬ ware glazes, when clay was increased to a definite amount, or whether it was a complex case of an eutectic mixture of several substances, is not yet determined. The fact remains that additon of clay did cause greater fusibility and less viscosity, notwithstanding the fact that with each addition of clay the AhOs was being increased. Bleininger 3 has shown experimentally that calcium carbonate reacts with finely pulverized feldspar as readily as with washed kaolin. From his results it would seem as though fusion is initiated between calcium 1 Molasses is more viscous than water, i. e., it flows more sluggishly. Its mole¬ cules are less free to move. Slow-flowing fluids are said to be viscous. 2 Trans. Am. Cer. Soc., Vol. V. 3 Geol. Surv. of Ohio, Bull. No. 3, 4th series, p. 128. PURIFY] PYRO-PHYSICAL AND CHEMICAL PROPERTIES. 241 carbonate and feldspar as early as between calcium carbonate and kaolin (pure clay). This being the case it would seem as though the addition of • clay to stoneware glaze mixtures was merely the formation of a eutectic mixture of minerals. 1 Evidence thus far developed in the case of simple mixtures is sum¬ marized in the following table: TABLE XXXI. Showing the proportions by weight, which cause maximum fusibility between the two mineral substances stated in each case. ( 1 ) ( 2 ) (1) Magnesium carbonate (1) and kaolin (2) .2 3 Calcium carbonate (1) and kaolin (2) .2 3 Finely pulverized flint (1) and kaolin (2) . 2 7 Finely pulverized flint (1) and feldspar (2) .1 3 (1) With quick fire. Any increase or decrease of AkCk outside of the limits given in the above table results in increase of refractoriness of the mixtures as shown in the several curves to which reference has been made. Similar points of greatest fusibility hawe been noted in the case of glazes, but data have not been obtained that permit showing the facts in tabular or curve form. AhO then increases the fusibility of mineral mixtures when added in amounts not exceeding a given proportional limit, the limit being different for different mixtures. Second, in slags, glazes and glasses addition of ALOs above a given amount increases their viscosity, but no limiting points have, as yet, been determined except in the case of slags. Since slags are comparatively simple in composition and usually relatively high in lime, we can learn very little by reviewing in detail the researches that have been made on the vicosity. Third —Increase of AbO in small amount in glasses increases their toughness. So far as data have been obtained increase of AhCh in por¬ celain bodies does not increase their toughness. From these conclusions a query is at once presented concerning the relation between fusibility, viscosity and toughness. At present any dis¬ cussion of this query would be based wholly on assumption, for there are no experimental data bearing on the point. Effect of Silica in Ceramic Mixtures —Anhydrous silica is practically inert at ordinary temperatures, but at the temperature usually attained in brick kilns it becomes very active, forming compounds having very varied oxygen ratios, i. e., amount of oxygen in the basic to the oxygen in the acid oxides. On heating, silica expands considerably, indicating peculiar molecu¬ lar changes. LeChatelier 2 has shown that at 500°C. this molecular 1 The mixtures that gave the greatest fusibility, as shown in each of the figures 19, 20 and 21, are said to be eutectic mixtures. 2 See Bleininger, Ohio Geol. Surv., Bull. 3, p. 28. —16 Q 242 PAVING BRICK AND PAVING BRICK CLAYS. [bull. no. 9 change takes place to a very pronounced degree in all forms of silica, the least in amorphous and the most in highly calcined flint. Per¬ manent expansion in highly silicious bricks and the “punkness” of bricks made from a mixture of clay and sand are evidence of the effect of this peculiar property of silica. No matter how fine the free silica is, it does not seem to be as active in forming new silicate compounds under the influence of heat as is the silica that is previously combined, as for illustration, in clay or feldspar. In other words, silicate combination with free basic elements is affected more readily when the silica is added to the mixture as a constituent of a pre-existing silicate. This was shown very prettily in an experi¬ ment reported by Bleininger. 1 He prepared a mixture of 20 per cent finely ground flint and 80 per cent precipitated calcium carbonate and two other mixtures each containing respectively 20 per cent finely ground feldspar and 20 per cent of kaolin with 80 per cent calcium car¬ bonate. These mixtures were maintained at a temperature of 1100 C. for 75 minutes. At this temperature calcium silicate compounds are formed which are soluble in hot hydrochloric acid and sodium carbonate solutions. The residue left after this acid and alkali treatment is the material which is unattacked or unlocked by the fluxing action of the- lime. In the following table are Bleininger’s results. Table XXXII. Ground Flint. Ground Feldspar. Ground Kaolin. Per cent residue. 28.83 3.75 3.07 Per cent taken into solution. 71.17 96.25 96.93 * Bleininger’s results strongly support the doctrine that has, for the sake of .emphasis, been repeatedly stated in this report, to-wit: That very little can be told concerning the fusing behavior of silicate mix¬ tures from an ultimate analysis, for if this were not the case, feldspar should have reacted far more vigorously with calcium carbonate than did clay. Since cement investigators have found that the hydrous am¬ orphous silica reacts with lime in a manner similar to finely pulverized crystalline quartz, it can be readily seen that misleading data would be obtained even in the rational analysis, in which the hydrous amorphous silica is taken into solution by the sulphuric acid and thus considered as a part of the clay substance. Addition of silica to pure clays like shales increases their refractori¬ ness and, (reasoning from data on slags) possibly, their viscosity. There is no evidence showing that the addition of flint to a clay increases its toughness, but quite the contrary, empirical experiments by several prac¬ tical brick manufacturers have proved that the additon of ordinary bank sand makes the bricks less tough or even very. “punky.” On the other hand an investigation by Worcester 2 proved that Bedford shale, which lLoc. cit.. p. 128. 2 Trans. Am. Cer. Soc. Vol. II, pp. 295. PURDY] PYRO-PHYSICAL AND CHEMICAL PROPERTIES. 243 outcrops near Columbus, 0., is materially benefited by an addition of crushed Berea sandstone from the same locality. Instances are recorded lof addition of certain sands in Europe having proved beneficial, but in neither Worcester’s experiments nor in the European cases was there ‘reported a determination of the effect of sand on the toughness of the - burned mixtures. In the manufacture of floor tile the writer found that a porcelain . body consisting of 40 per cent clay, 45 per cent feldspar and 15 per cent flint was much tougher than a body containing 35 per cent clay and 65 per cent feldspar. It is impossible to say why the body containing flint should be tougher but certainly some credit must be given to the influ¬ ence of the flint. Reviewing the known facts about the effect of silica on either the fusion of clays or development of toughness in clay wares, it must be admitted that we have not at present much positive evidence. Effect of Magnesium Oxide in Ceramic Mixtures —In figure 20 on page 209 is shown graphically in fluxing effect of magnesium oxide with kaolin. Metallurgists report that magnesium oxide is a much “harder” flux than calcium oxide and produces a much more viscous slag. Ceramic investigators have reported conflicting results in their attempts to use MgO as a flux, some claiming that it is more active than CaO and some that it is less active. Claims have been made by some that in glazes it gives greater fusibility and slower fusion, while others claim opposite results. From this accumulation of apparently conflict¬ ing data it has been shown that in short quick burns, as in experimental kilns, MgO is an active flux causing more rapid fusion, but in longer burns its fluxing action begins as early as in the shorter burns but pro¬ gresses less rapidly and requires more intense heat treatment to effect complete fusion. • The lag in the fusion of mixtures containing magnesium oxide is at¬ tributed to either the viscosity of the resulting magnesium silicate, if it enters into combination with the glassy matrix that fills and seals the pores of vitrifying wares, or. to the formation of non-fluid magnesium compounds. 1 Cement investigators claim that the alkaline earth sili¬ cates formed by heating mixtures of clay and calcium or magnesium car¬ bonate at temperatures below that required to cause sintering of the mass into a hard cake or brick are simple silicates of calcium or mag¬ nesium oxide which are not necessarily fluid. At any rate, the effect of magnesium in ceramic mixtures differs from that of calcium in that the magnesium mixtures fuse very slowly over a long heat range, while the calcium mixtures, especially when present in amounts equal to or more than 10 per cent, remain porous up to the time that fusion begins, and then fluxing ensues very rapidly causing the.ware to pass from por¬ ous into the overburned condition within a very short range of- heat treatment. l Eckel states in “Cements, Limes and Plasters,” p. 154, that when magnesia is burned in a quick fire its density (specific gravity) is 3.0 to 3.07, while if burned in a slow, long-continued fire its specific gravity will range from 3.6 to 3.8. 244 PAVING BRICK AND PAVING BRICK CLAYS. [bull. no. 9 The only known facts concerning the influence of magnesium oxide in ceramic mixtures are : (1) magnesium oxide increases viscosity; (2) mag¬ nesium oxide causes slower rate of fusion, at least when it is the pre¬ dominating flux; (3) as has been stated earlier in this report, clays which make good paving bricks contain a larger amount of magnesium than calcium oxide; (4) the Italians are now making low-fired porce¬ lain of which toughness is a special feature, and in which magnesium is the only Ro or fluxing base present. Effect of Calcium Oxide in Ceramic Mixtures —Watts 1 has shown that the presence of a small amount of calcium oxide in porcelain mixtures results in increased toughness of the ware. His investigations are not, however, sufficiently exhaustive to warrant more definite statement. It is known that lime causes a breaking down of the silicates with comparatively little heat treatment, and also that the new silicates formed are probably very simple in composition until higher tempera¬ tures are attained, in which event these simple silicates suddenly fuse, causing the whole to pass rapidly into a fluid mass. Dr. Rieke 2 has shown that in mixtures of from 1 to 10 per cent of calcium carbonate with kaolin very close tight bodies are obtained which have quite a large range of vitrification and in the end fuse quite gradually. Mixtures containing more than 10 per cent of calcium car¬ bonate remain quite open until final fusion begins, at which time the whole mass fuses very rapidly. In comparison with Rieke’s work it is of interest to study results obtained by Nauss, 3 who worked with a mixture similar to Rieke’s high calcium body. The two bodies were as follows: Table XXXIII. Nauss. Rieke. Calcium carbonate. 70 70 Kaolin. . 11.05 30 Flint :. 18.66 In the following table are Nauss* results and in a separate column are placed the data obtained by Rieke. Rieke measured his heat by cones and hence the temperatures obtained in these two studies cannot be com¬ pared closely. Since, however, Rieke used a Seger trial kiln and very short firing periods, his cone readings can be. approximated in terms of degrees centigrade within the accuracy of and discrepancy between the method by which each research was executed. 1 Trans. Am. C'er. Soc., Vol. V, pp. 175. 2 Sprechsaal No. 38, 1906. 3 Reported by Bleining’er, Ohio Geol. Surv. Bull. No. 3, p. 175. PURDY] PYRO-PHYSICAL AND CHEMICAL PROPERTIES. 245 Table XXXIY. Reaction of Calcium Oxide upon Kaolin and Quartz. o 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Temp. °C. Loss on burning.... Loss on completed ignition. Carbon dioxide remaining. Fusible silica—per cent. Remarks. Rieke’s Results. *T3 o o CO •< O ero CD D P 0! S'2, crq 550 585 610 655 700 725 750 775 800 850 900 950 1000 1050 1100 1150 1150 1200 1200 1250 1250 1300 1300 1325 1.79 1.88 1.80 1.91 4.81 6.18 6.99 8.35 12.51 20.34 24.42 27.56 30.00 29.75 29.65 29.74 30.10 30.10 30.10 30.10 30.10 30.10 30.10 30.10 27.79 27.96 28.67 28.83 25.17 24.30 23.02 21.90 17.94 10.64 5.89 2.88 0.60 28.48 28.44 28.84 27.83 25.89 24.58 23.88 21.07 18.72 7.32 5.29 2.17 0.40 . 19.70 19.00 18.90 18.83 18.75 18.75 18.26 16.19 16.10 15.37 13.13 13.34 11.97 10.68 30.3 05 Held at this temperature for 12 hours.. 42.5 ' 2 45.9 5 Dusted, i. e., fell to pieces. Dusted. Dusted. 42.0 8 Began to fuse, “melt”. 10.2 Fused. 10 23-24 Professor BleiningePs conclusions from Nanss* work are: “First —In regard to the decomposition of calcium carbonate, it is clearly shown that it begins to break up between 610° and 650°C., and before 700° is reached the evolution of carbon deoxide is going on quite rapidly. At 1000° the evolution is practically at an end.” “Second —On examining the amounts of insoluble* residue and comparing the percentage with the known amount of quartz in the mixture, 18.66 per cent, and making allowance for the small amount of quartz in the kaolin itself, it is seen that the kaolin is decomposed completely at 850°C., and al¬ most completely at 800°C.” * Third —Free quartz seems to be attacked by the calcium oxide soon after the completion of the decomposition of kaolin, probably at about 950°C., which reaction continues, at an increasing rate up to the highest tempera¬ ture employed in these experiments. It is quite evident, also, that the length of time of burning influences the amount of quartz attacked somewhat, so that by longer burning, at least with temperature over 1100°, more quartz may be rendered soluble than in a short period of ignition.” Prof. Bleininger, continuing, says: “A very interesting fact was brought out by the tendency to dust observed with the mixture at temperatures above 1200°C. While at 1200° the bri¬ quettes were hard, at 1250° they dusted very rapidly, and at 1300° almost instantaneously.” l Insoluble in hydrochloric acid and sodium carbonate solutions. 246 PAVING BRICK AND PAVING BRICK CLAYS. [bull. no. 9 “On calculating the formula of this mixture from the composition we find it to be 1.77 CaO, 0.108 A1 2 0 3 , Si0 2 , that is not quite a singulo calcium sili¬ cate, and hence must properly be classed within the group of natural ce¬ ments. It is not difficult to understand that the dusting must be coincident with a significant molecular change from the condition of the. loose, friable mixture to a hard body breaking down at once to a powder. Might not this fact indicate that up to 1200° these calcareous mixtures are but pozzuolane- like, simple silicates, consisting of silica and base which on further applicar tion of heat become chemically more complex and non-or but slightly hy¬ draulic? This view is strengthened „by the results of another investigation which have shown that on increasing the free silica, with but sufficient base to convert the quartz into the active state, the hydraulicity is practically as great as with a greater amount of base.”i Rieke’s data is evidence that Bleininger’s query can be answered in the affirmative, for it was at this same temperature, 1200°C., that his body ceased to increase in porosity and began to vitrify. From 1200° C. on, Rieke’s body vitrified quite rapidly showing that “a significant molecular change” is taking place. From Nauss’ results it must be conceded that the clay has suffered a very significant change. NcJ doubt it has passed completely into solution with lime and silica. In fact Bleininger’s results given in Table XXXYI page 248 proves this to be so. Rieke’s porosity data show also that prior to this critical tempera¬ ture, 1200°, (rough approximate) the grains must be changing form and size, for the mass is getting more porous with each increase in heat treatment, yet, according to his shrinkage data (1.2 per cent at cone 05 and 3.7 per cent at cone 5) the mass as a whole is decreasing in vol¬ ume. Similar simultaneous increase in porosity and decrease in volume was noted in several instances in our own researches, so this phenom¬ enon is not alone peculiar to simple mixtures high in lime. Important as are these observations, and especially that of complete solution, and possibly the formation of entirely new compounds before the mass begins to decrease in porosity, i. e., vitrify, the more important item to note at this time is, in the writer’s opinion, the difference in the ultimate’ fusion behavior of the two bodies, the one containing fred silica and the other supposedly none. It was shown by Bleininger’s result 1 2 that quartz is not nearly as readily attacked by CaO as is kaolin or feldspar, and hence it could be inferred that the higher the content of quartz in a mixture, the later and slower would the mass fuse. In decided contradiction to such an inference we find that in Xauss’ body, containing 18.7 per cent quartz, the original minerals have been com¬ pletely broken down and the whole began to “melt” at the same tem¬ perature at which Rieke’s body containing no quartz exhibits a porosity of 10 per cent, but complete fusion does not take place until a tem¬ perature of about 1600° C. has been reached. We are learning not to wonder at such apparent discrepancies in experimental work where simple mixtures of two minerals are compared in their fusing behavior with more complicated mixtures of minerals. 1 Italics not in the original. 2 See Table XXXYI p. 248. PURDY] PYE0-PHYS1CAL AND CHEMICAL PROPERTIES. 247 Summarizing these observations the following facts appear: First, Watts-has shown that a small quantity of lime toughens a porcelain mixture. Second, Rieke has shown that in a. simple mixture of kaolin and 1 to 10 per cent calcium carbonate there is quite a large vitrification range and slow fusion, while in mixtures with kaolin containing more than 10 per cent of calcium carbonate the body does not vitrify until late and then rather suddenly fuses. These findings by Rieke and Watts agree with ours in support of the assumption that long vitrifica¬ tion range and slow fusion generally result in the production of tough ware. Third, the results of Bleininger, Nauss and Rieke studied to¬ gether show very forcibly that chemical alterations and reactions, may take place long before vitrification and fusion begin. Also, that each mixture has its own peculiar pyro-chemical and physical behavior, and, as the mixtures become complicated in composition, the deductions drawn from simple mixtures are found to hold true only in very small part. Beyond these studies in simple mixtures by Bleininger, Nauss, and Rieke, and the observation in complicated porcelain mixtures, we have no data that have a bearing on the effect of smaller or larger quantities of lime on toughness of burned wares made from shales. Contrasting the work of Rieke and Nauss, the difficulties that are encountered when attempt is made to trace the effect of lime in such severely complicated mixtures as shales are clearly shown. Effect of Other Oxides in Ceramic Mixtures —Practically nothing is known concerning the influence of oxides other than those considered above, except that in slags titanium causes increased viscosity; that potash silicates are more fluid than soda silicates, and yet, as a rule, less fusible; that phosphoric acid is expelled from ceramic mixtures only at high temperatures, and that, before expulsion it is combined with the bases forming phosphates that are analogous to the silicates. A, detailed study of the influence of the several oxides, alone and to¬ gether, on the fusion of silicate mixtures and the toughness of the burned mixtures, offers a very fruitful and interesting field for re¬ search. INFLUENCE OF SIZE OF GRAIN ON THE FUSION OF CLAYS. Direct evidence —According to Wegemann’s microscopic studies given on later pages, coarse quartz does not enter into the fluxing reactions even at cone 5. With a heat treatment sufficient to fuse cone 5 feldspar is completely fused especially if mixed with free silica, and yet at this cone Wegemann reports that the quartz grains are apparently unaf¬ fected to any noticeable extent until cone 9 is fused down. He affirms that if any reaction has taken place between the free silica and feldspar, the silica must have been supplied from what he terms the ground mass, i. e., the mass that consists of particles too fine to be distinguished through the microscope. According then to Wegemamf’s studies, the melting feldspar in shales affects the coarse flint to but a slight extent. Bleininger 1 experimentally determined the effect of size of flint and feldspar grains on the rate at which lime would decompose them at 1100° C., forming silicates that could be dissolved in hydrochloric acid 248 PAVING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 Table XXXV. Effect of Size of Grain on Extent and Rate of Combination of Silicia and Lime. Sizes. Ground Flint. 150-120 mesh. 1120-100 100-80 80-60 mesh. mesh. mesh. 1 1 1 60-40 mesh. 40-20 mesh. Per cent residue. 1 28.831 1 63.8 78.53; 86.52 86.27 93.78 96.83 Ppr rent taken into solution. 71.17 1 36.2 1 21.47 13.48 13.73 l 1 1 1 6.27 3.17 Table XXXVI. Effect of Size of Grain on Extent and Rate of Combination of Feldspar and Lime. Sizes. Ground Feldspar. 150-120 mesh. I 120-100 mesh. 100-80 mesh, j 80-60 mesh. 60-40 40-20 mesh. mesh. Per cent residue. 3.75 15.45 31.00 64.29 79.63 95.72 1 . Per cent taken into solution. 96.25 84.55 69.00 35.71 20.37 4.28. 1 and sodium carbonate. The data he obtained are given in the follow¬ ing tables: This data, together with Wegemann’s microscopic observations, proves conclusively that a variation of this physical factor—fineness of grain —has an influence on the fusing behavior of clays that is as positive, if not as potent, as a variation in the quantity of the oxides of any of the elements. GENERAL ANALYSIS OF RESULTS. The foregoing detailed discussion of the various elements affecting the manner in which silicate mixtures fuse, has been given in addition to the more general statements on pages 217 and 232 so as to make more plain the deductions that are to be drawn from our own data. This detailed citation, it is hoped, has clearly demonstrated that our present knowledge of the influence of the several factors even in simple mixtures is very fragmentary and that in the more complex mixtures the evidence is, in the main, either conflicting or entirely lacking. In the following analysis of the chemical data obtained by this Survey, and at¬ tempts to show a relation between the chemical and ph} r sical constitution of the clays, their pyro-physical behavior, and toughness of the burned bricks, liberal assumptions must be made and only general conclusions, if any, drawn. These assumptions are: First, Those elements which are supposed to increase the viscosity of the mass when fused lengthen the vitrifying range of the clay and increase the toughness of vitrified wares. Sec¬ ond, Those chemical or physical factors which tend to make the mass more fusible or to hasten the pyro-chemical reactions which result in vitrification are detrimental to development of toughness. Third, That 1 Loc. cit. p. 127. PURDY] PYR0-PHYS1CAL AND CHEMICAL PROPERTIES. 249 .lime is detrimental both to slow fusion and toughness, while magnesia is beneficial. Fourth, That the higher the acid content, or its equivalent, the oxygen ratio, the more viscous will be the fused ingredients and the tougher the burned ware. Fifth, The higher the proportion of Ab(K to other basic oxides the slower will be the fusion, the more viscous the fused ingredients and the tougher the mass. Sixth, The finer the ma¬ terial of which clay is composed, the more rapidly will it fuse and the more brittle will be the burned mass. In the following table will be found the ratio mentioned in the fore¬ going assumptions, as calculated from the chemical data given on pages 215 and 216. In the first column is the ratio of CaO to MgO. In this ratio, CaO is taken as unity. In the second column is given the total oxygen in the basic oxides where AhOa is unity. In summing up the oxygen atoms, the iron oxides were considered as reported, i. e., where FeO is given, only one atom of oxygen to one atom of Fe, and where FeaOa is given, three atoms of oxygen to two atoms of Fe were taken. The difference between the value given in the second column and 3 (oxygen in AhOa) gives the factors for the ' Table XXXVII. w I i i 1 2 3 4 5 6 7 Remarks. ( 1 ( fT 0 (A) Ratio of CaO to MgO.. { 1 1 1 Total oxygen in bases . Oxygen in acid. (B) Oxygen Ratio.. (D) Surface factor, Purdy’s Method... 5 c c c c 0 D j > L 9 > a o f g Rattler loss N. B. M. A. Standard. K — 1 2.39 4.25 13.96 3.27 257 7.3 15.82 K— 2. 1.84 4.51 13.24 2.92 331 3.24 17.48 Good red when vitrified. K— : J.. . 3.36 4.50 13.14 2.90 341 5.70 24 89 K— i t. 6.70 4.3 11.56 2.67 514 8.00 19.11 Not screened when used at factory K— 5. 2.95 4.1 12.76 3.1 287 8.65 19.36 . .do. K— 6. 3.21 4.2 13.28 3.17 221 11 .5 13.25 K— ' 7 3.10 4.2 11.68 2.8 300 7.2 13.89 K"— f 4.47 4.59 12.62 2.75 262 8.85 20.'23 K— 9. 2.62 4.23 13.72 3.22 195 10.6 14.84 Very hard coarse clay. K—10. 2.36 4.39 10.96 2.5 604 2.1 39.36 K—11. 4.20 4.28 9.71 ' 2.27 339 6.6 28.13 K—12. 1.44 3.95 7.82 1.98 403 2.22 K—13. 3.35 4.34 10.18 2.35 356 4.95 31.50 K—14. 1.36 4.5 15.38 3.4 254 3.65 21.24 Very hard coarse clay. K—15. 1.42 4.4 11.14 2.52 18.44 F - 1. 1.94 4.63 12 70 2.75 20.84 S — 1. 2.74 4.11 9.20 2.23 26.25 s — : 2_ 5.89 4.33 9.42 2.17 27.94 R— 1. 4.72 3.54 7.92 2.50 397 16.5 16.92 No. 2 Fire clay. R- ! 3.99 4.05 11.58 2.85 17.80 R— : 3. 3.05 4.13 9.76 2.36 291 6.55 14.80 R— i 1 _ 4.17 4.13 8.66 2.10 275 8.45 15.33 H—16. 0.95 4.5 11.56 2.55 H—17. 1.74 6.43 15.92 2.47 H—18. 0.95 5.48 8.26 1.51 444 0.39 H—20. 1.17 5.78 10.36 1.86 553 0.42 H—21. 0.765 5.04 9.5 1.88 783 0.27 H—23.. 0.497 3.97 8.71 2.2 634 O 53 B—11. 0.895 4.20 10.98 2.61 28.03 G—II. 1.62 4.65 13.28 i 2.85 366 2.3 14.98 Good dark red when vitrified. H—11. 0.78 4.55 10.18 1 2.27 32.97 I—II. 0.82 4.67 17.98 i 3.85 489 | 1.16 25.65 Bright red when vitrified. J— II. 1.77 4.71 12.58 i 2.65 17.14 L— II. 1.09 4.47 11.24 2.51 18.58 250 PAVING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 proportion of oxygen in AbOs to oxygen in the other bases, i. e., (a—3) : 3: :0 in finxes: 0 in AbCk In the third column is given the number of atoms of oxygen in total SiCb. In the fourth column is given the ratio between oxygen in SiCb to oxygen in total bases. This ratio is known as the oxygen ratio and is customarily taken as the ratio of the acids to the bases. In the fifth column is given the surface factor rep¬ resenting fineness of grain by the writer’s method. In the sixth col- A. B, C. umn is a modulus calculated on the formula-= M where D. “A” is the lime-magnesia ratio, “B” the total oxygen ratio, “C” the ratio of oxygen in the Ro bases to oxygen in AbO, and “D” the sur¬ face factor divided by 100. In the seventh column is given the rattler loss determined on commercially manufactured blocks made from each clay. Deductions Drawn from Table XLI —Without going into details con¬ cerning the probable reason for the lack of correlation between the chem¬ ical and physical constitution and the toughness of the burned ware as shown in the above data, it is sufficient to state that it be granted that these data corroborate those of Ogden, proving that our notions about the relation of the chemical and physical constitution of clays to the toughness that is developed in burning are in the main, if not wholly, erroneous. Data on mineralogical composition as obtained by the Rational Analysis, gave results that were still less easily correlated with data on toughness of the burned ware than are those in the above table. Before such data can possibly be of value there must be consider¬ ably more learned concerning the fusing behavior and the physical prop¬ erties of sintered masses of simple mixtures of minerals. There is not much of any hope of learning much concerning these relations from data obtained by any process of chemical analysis now used. THERMO-CHEMICAL AND PHYSICAL CHANGES DURING FUSION. It is indeed very difficult, if not impossible, to determine what the actual thermo-chemical reactions really are, which take place in the fusion of the clay particles, first between themselves, and, secondly, when the whole mass becomes a more or less homogeneous solution. 1 By the aid of the microscope, as will be seen later, more can be told con¬ cerning these changes in an unknown mixture of minerals than bv any other means; inferences from artificial and known mixtures being of no avail. The effect of thermo-chemical reactions, however, can be detected by the changes in porosity and specific gravity. Because of our present inability to ascertain in full the reactions that take place, it seems best to refer to the chemical phases of fusion as “changes” instead of “reactions.” lProf. G. Tamman, Sprechsaal No. 35, 1904, summarizing - his studies on sili¬ cates says, “The volume of the glass is, at the lowest temperatures, larger than that of crystals.” Mellor, Vol. V, p. 78, discusses the volume changes in silicates and cites A. Laurent (Ann. Chim. Phys. (2) 66,96,1837; A. Brongniart, Traite des Arts Ceramiques, 1, 283, 720. 1877) and G. Rose (Pogg, 111,‘123, 1890; A. R. Day and E. S. Shephard, Am. Jour. Science, (4) 22, 262, 1906. Dr. E. Berdel (cited Vol. VII, p. 148 A. C. S. Trans.) described similar physical changes in the heat¬ ing of ceramic materials and bodies. PURDY] PYRO-PHYSICAL AND CHEMICAL PROPERTIES. 251 The greater portion of the constituents of onr clays being mineral substances, many of which do not entirely lose their identity in the burning of clay wares, it is most natural that these should exhibit in nature the same changes when treated separately that they do when heated together in clays. Roth 1 gives the following description of the physical changes in minerals on melting: Table XXXVIII. Mineral. Specific Gravity of the Crystal. Specific Grav¬ ity when melted to glass 1 Per cent Re¬ duction in Spec. Gravity. Quartz. 2.663 2.228 16.3 Quartz .. 2.65 2.19 17.3 Olivine. 3.3813 2.8571 15.6 Mica. 3.0719 2.2405 27.0 Adular. 2.561 2.3512 8.1 Adular. 2.5522 2.33551 8.5 Sanidine. 2.58 2.381 7.6 Orthoclase. 2.574 2.328 9.6 Orthoclase. 2.5883 2.3073 10.9 Microcline. 2.5393 2.3069 9.1 Albite. 2.604 2.041 21.9 Oligoclase. 2.66 2.258 15.1 Oligoclase. 2.6051 2.3621 9.1 Oligoclase. 2.6141 2.1765 16.7 Labradorite. 2.7333 2.5673 6.1 Hornblende. 3.2159 2.8256 12.2 Augite. 3.2667 2.8035 14.2 Epidote. 3.409 2.984 12.5 Red brown garnet.... 3.90 3.05 20.5 Lime-iron garnet_ 3.838 3.340 25.6 Granite. 2.680 2.427 12.9 Granite. 2.751 2.496 9.3 Hornblende granite.. 2.643 2.478 6.2 Felsite porphyry. 2.576 2.301 10.7 Syenite. 2.710 2.43 10.3 Quartz diorite. 2.667 2.403 9.8 Diorite, quartz free.. 2.779 2.608 6.3 Gabbro. 3.100 - 2 664 14.2 1 ‘Not in original table. Remarks. Average. Glass compact. Glass full of fine bubbles. Glass full of fine bubbles. Glass full of fine bubbles, and dark-colored. Glass full of fine bubbles. Glass colorless. Glass colorless. Full of fine bubbles; white glass. Glass full of fine bubbles. White glass; bubbly. Glass full of bubbles. Glass slightly bubbly, with black and white portions Glass compact. Glass compact. Green glass. Green glass; transparent; strongly blebbed. Black glass; opaque; strongly blebbed. Black glass; opaque; strongly blebbed. Transparent; veryblebby; difficult of fusion. Glass homogeneous; dark colored. Glass homogeneous; dark colored. Black glass; opaque; com¬ pact; somewhat difficult to fuse. Black opaque glass; easily fusible. The alterations in the minerals and rocks above cited are those in¬ duced when they are changed by melting, from a crystalline to an amorphous condition. Such complete changes as this cannot be per¬ mitted to take place in the whole mass of clay ware during burning, and yet, as will be shown, the percentage of decrease in specific gravity of many'of our clays from the unburned to the vitreous stage is greater than that given in the above data. This being true, it is evident that there are factors other than the alteration of minerals from the crystal¬ line to the amorphous condition that affect decrease in the specific grav¬ ity of clays. In the following table are given data which show the effect of heat on physical structure of briquettes made from various clays: l Allegemeine und Chemisch3 Geologie, Vol. 11, p. 52. Table XXXIX. 252 PAYING BRICK AND PAYING BRICK CLAYS. [BULL. NO. 9 C3 g . « a co cu g E o B pp co 03 XJ «- G^ 51 ■§&■; .& q,cJ<®d “ 2 aj u 2 G— a 03.G ppc m i u G ® u - « o a « ’o'.o.tp g/3 S'® § olo ti m PI’S®’* nX3/2 0X3 03 r* .G PP 03 Pad ca‘8 pc q. 8 & W 0.0 co 03 o 3 a- 52 o o o G .- O X3 - 2x3 O O • K G os a PC 03 - CJ _ lM 03 g O a 2s U3 “ g M a —■ • - ■ t® O a ^ 4 ) rt-p G «« ^pclg, Z2, o Q, ^ Q-> c § IS §^1 03 y G flrt y ” 03 333 PQ 03 /P CQ CQ MGrOO co bre isPpu.. Q CQ u Q SG3 _ 03 PC 3 E 2^ 72 CCQC X3 03 8*1 cO'u hfi OQ S O it o PP 03 O ±j 03 CX3 3S|C «j&£ tH N H >dco^< o N i~J CM 00 © 00 H t- CM CM C^l y-i Oi frl SO 05 *H CD CO P- CM ^ CO ^ CO 00 CM 00 CO IO ID CO 1 —< §ss| u «u> n 2 2 2 03 o O-fK CL| 03^^ CO CD C- rH CM Cl ID 1T0 ^ ^ ID ^ <008 os ec ca 03 . a o °Z O 03 O -h >rs i/s ogs^ 03 rt o.lP 03 03 03 > & OP^rK rl ITS in CC y-* ec 00 03 . c o r8£ in in eo .03 "Sq §Z CVJ rH Hi m ec 1-1 w WWW ZZZ X uuooo ^ ^ 1 >. >» n rt CO CO cO c0 cl PP P3 pp OO ouo 0 O 0 C/3 C/3 C/3 PC PC PC 03 ©. 03 03 03 2 03 03 O O ££ % X 5 >H QQ « he he he CM CM co coco CO co CO a G _G 0 0 OOO o' d 0 ’> cO ’> C0 ’> cO zzz Z Z z Hi Pu H C 0 C /3 73 COCQ P > > Table XXXIX—Concluded, PURDY ] PYRO-PHYSICAL AND CHEMICAL PROPERTIES. a cn G P oB X cS *0 H G^ O-M O g cffi o co 2°£g cn G T3 asg£ O e O k" » u wXX OtSU> H OJ £ ctf £gtnO T3 . a o a* CJ -m . M & g «8MA Oh^c^ 4j Gx £9 «ts g •■ - „ bees. 05 Q X D Cw ,—i G ->> •g-s SS ho£ a> ti-P cs u A P O ax ’S .. 6 CL) CO C 0) Din D o GX ■gi jg m O . - cj ,* oo X! h-> CJ CS 8? + J tofi 3 ►> b fe fin iJ jO 4 > 0 ) CD CL) cS cS CS cS cS G CS cS cS CS >> w Is -G ’cs'cS X 3 J 3 Is XI Is XJ XI Is X Is X xxxxxx cncncncncn r n X cn X cn X cn X cn U CO C /5 C /5 Cfi Cfi cn C /5 cn xxxxxx M X M Sd J4 M S 4 M _CJ _o ;~j o o _o o o ’C ‘C u '!5 o CJ CJ o »- pqmwwmm 03 « m 05 , H be a be be .5 S ’>’> bo G ’> be G ’> be a '> bo G '> ' he G "> .H .5 .S .S . 5.5 222222 G 2 a 2 G 2 G is cS CS cS CS , cS cS cS cS ’3 '3 '3 " 3 * 3*3 *3 *3 "3 Oh CuO^ Pli Oh Oj Oh Oh 05 05 03 M 05 55 05 05 05 03 253 254 PAYING BRICK AND PAYING BRICK CLAYS. [BULL. NO. 9 It was a surprise to learn that bricks will decrease in volume with¬ out loss of weight, and at the same time'decrease in specific gravity. Had the clay been carried to complete fusion, i. e., to a glass, the de¬ crease in specific gravity would have been credited to the phenomenon as in the case of minerals, i. e., the changing of its constituents from crystalline to amorphous forms. But in the case of a clay briquette, a small portion of which enters into, the fusion, decreasing in specific gravity before the minerals have been rendered amorphous, i. e., fused to a glass or even before vitrification has been completed, .cannot be explained wholly on this basis. Mr. C. H. Wegemann, of the geological department, was, therefore, requested to make a microscopic study of briquettes of two different clays burned at different temperatures. His report follows: Notes on the Microscopic Structure of Certain Paving Brick Clays, at Various Stages of Fusion. [By C. H. Wegemann.] In the hope of explaining some of the phenomena of simultaneous decrease in volume, porosity and specific gravity without loss in weight and to obtain some idea of the manner in which fusion takes place in a vitrifying brick, microscopic sections were prepared from briquettes of two paving brick clays. GENERAL STRUCTURE. Thin sections of the briquettes burned at a low temperature exhibit under the microscope a very fine-grained fragmental ground mass, or matrix, in which are imbedded crystalline and other fragments which were present in the original clay. From these materials are developed, at high temperature, amorphous glasses and crystals. The cavities between the particles of a brick may be divided into two classes: (1) Pores, which are present in pieces fired at low temperatures, due to the incomplete consolidation of the clay. These are the original interstitial spaces of the unburnt clay. (2) Blebs or bubbles, which are formed in the glass at higher temperature by the liberation and expansion of gases. Pores of the first sort are of small size and irregular outline. As the temperature increases, and the material of the matrix gradually .fuses into glass, these interstitial spaces tend to disappear. Cavities of the second sort, which we may for convenience designate as blebs, are simply gas bubbles in glass. They are circular in outline and vary greatly in size. They are not present in the bricks burned at lower temperatures, but appear only after the formation of considerable glass. DESCRIPTION OF SLIDES. R3-14—This briquette was drawn at cone 3 or about 1190°C. The color is red. Under the microscope, the earthy matrix or ground mass is dark brown, the color being due to the presence of iron oxides. The mineral fragments are quartz, feldspar and mica, named in the order of their abundance. They are angular in outline, the thin edges being sharply defined. Glass has formed to some extent throughout the ground mass and in a few instances it has separated out into clear transparent masses, in several of which blebs appear. The blebs, however, are so few and so small that the cavities may be considered as made up almost entirely of pores of the.first class. As estimated under the microscope, the porosity is 1.9 per cent. PURDY] PYRO-PHYSICAL AND CHEMICAL PROPERTIES. 255 R3-16—Drawn at cone 5, or approximately 1230°C; color dark brown. Under the microscope the ground mass appears somewhat denser and darker than in R 3-14. The quartz fragments are apparently unchanged. The feldspar fragments, however, have disappeared.! Mica is present, but in very small quantity. Glass has been formed in considerable amount. It appears in clear trans¬ parent areas, often 0.1mm. in diameter. In some of the glass, needle-like crystals have begun to form, but where free from these the glass is color¬ less. . This fact would seem to indicate that but little iron has entered into its composition. As stated above, fine needle-like crystals are often present, imbedded in the glass. They do not appear to have any definite arrangement with re¬ spect to each other, but occur singly or in dense masses. When viewed singly they are colorless, but when seen in masses, they possess a greenish yellow tint, which they impart to the glass in which they are imbedded. What the crystals are was not determined. The iron oxides present in the matrix have become segregated into dense masses, which, where they transmit light at all, show the red of hematite, but no definite crystals are to be seen. Pores of the first class have dis¬ appeared, and blebs in the glass have become numerous and large, their average diameter being 0.066 mm. The estimated pore space has increased to 4.2 per cent. R 3-18 —Drawn at cone 7, or 1270°C. The fragments of quartz appear un¬ changed. The earthy ground mass is rapidly fusing into glass, which has increased greatly in amount over that in the preceding slide. The fine needle¬ like crystals are also present in greater number. Minute crystals of iron oxide are seen, apparently in the form of rhom- bohedrons, having slightly concave faces. They do not exceed 0.0014 mm. in diameter. The blebs have an average diameter of 0.1 mm. and the pore space has increased to 12.05. R 3-20 —Drawn at cone 9, or approximately 1310°C. Quartz fragments are present as before, but occasionally one is observed the edge of which has fused into a glass. The needle-like crystals are everywhere present in the glass, giving to it the yellowish-green tint before mentioned. The iron ox¬ ides appear much the same as in the last specimen. The blebs are but little changed. R 3-22 —Drawn at cone 11, or approximately 1350°C. The earthy matrix has given place entirely to glass. Quartz particles-are still present, but thin; •their edges have been rounded by fusion. The fine needle-like crystals in the glass have increased greatly in length, being in some cases 0.03 mm. long. They exhibit for the first time a marked tendency to collect in radiating clusters. Often they appear to be attached to the corners of the crystals of 4ron oxide. These latter have increased in number and size, being 0.005 mm. in diameter. In some cases the individ¬ uals unite, forming long serrated columns. ,Blebs have increased in size, their average diameter being 0.128 mm. The pore space as estimated from them is 19 per cent. G 11-10 —Drawn at cone 02, or approximately 1110°C. Color, brick red. As in the R 3 series already described, the mineral fragments consist of quartz, feldspar and mica. Very little glass seems to have developed at this temperature, and no blebs are present. . The pore space is made up entirely of pores of the first class, or those due to the imperfect consolidation of the bricks. The average diameter of these pores is 0.065 mm., and the pore space as calculated is 2.6 per cent. l Hintze gives the fusion points of the feldspar as ranging from 1140C., in sanidine to 1230 C., in labradorite. In the briquette under consideration it is evident that the feldspar has fused into glass. It is to be supposed that in this fusing, it would flux some of the quartz. If it did so, however, the quartz must have been furnished by the ground mass, for the coarser fragments are apparently not changed in outline nor diminished in amount. PAVING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 256 G 11-12 — Drawn at cone 1, or approximately 1150°C. Color red. A little glass appears, but no blebs are seen. The average size of pores is lower than in the last slide, being 0.045, but the pore space as estimated runs a little higher, or 3.6 per cent. It may be remarked that in the slides there is no marked increase in the pore space, as temperature increases, up to the point where blebs appear. From that point on, pore space increases rapidly. G 11-15 —Drawn at cone 3, or approximately 1190°C. Color, reddish brown. Fine needle-like crystals have formed in the glass. A few blebs appear, but are not in sufficient number to affect the pore space materially. As estimated, is is 3.2 per cent, while the average size of the pores of both classes is 0.06 mm. G 11-15 —Drawn at cone 5, or approximately 1230°C. Color, dark brown. Quartz fragments are still present, but the feldspar and mica have dis¬ appeared. Glass has formed in great quantity, being colorless, or when acicular crystals are present, greenish yellow. These crystals are present in great numbers and resemble those described in the former series. Microlites of iron oxide are also present, but have not yet grouped themselves in den¬ dritic forms. Pores other than blebs have disappeared, but the blebs have increased greatly in size, the average diameter being 0.175 mm., while the pore space amounts to 12 per cent. Summary of Changes Observed at Different Heat Treatments. Cone 12—Quartz and feldspar fragments are unchanged. But little glass is developed. No blebs have yet formed. Cone 1—No marked change has taken place over cone 12. Cone 3—A small amount of glass is developed from the ground mass. A few blebs appear. Needle-like crystals are developed in the glass. Cone 5—Feldspar fragments are fused into glass. Quartz fragments are fused into glass. Blebs increase in number and size. Minute crystals of iron oxide develop. Cone 7—Glass increases in amount. Blebs increase in number and size. Quartz fragments are unchanged. Cone 9—Quartz fragments begin to fuse into glass along their edges. Cone 11—Ground mass is completely fused into glass. Some rounded quartz fragments still remain. Blebs have increased remarkably in size and number. Microlites are more numerous. It should be borne in mind that this is but a preliminary study. The num¬ ber of slides examined is too limited to warrant broad generalizations. Specific Gravity, Volume and Porosity Changes of Clays Studied. (by r. c. PURDY.) Owing to the absence of similar data on other clay samples and the incompleteness of the present researches, the writer has no definite con¬ clusions to present concerning the surprising facts presented by Mr Wegemann. This data does, however, establish the facts that neither a mineralogical analysis nor an ultimate or rational analysis of clay will indicate the nature of its pyro-chemical and physical ' behavior. Indeed, the above data would seem to throw doubt on the value of pjrro- chemical and physical study of a synthetic mixture of minerals as a basis on which to interpret the thermal changes in an “unknown” clay mixture. In the following figures 26 and 27 are shown the specific gravity, volume and changes in porosity in the two clays of which microscopic studies were made by Mr. Wegemann. It will be seen that all three - PURDY] PYRO-PHYSICAL AND CHEMICAL PROPERTIES. 257 factors decrease simultaneously, showing that the increases in molecular volume and in bleb structure is not sufficient to counteract the shrink¬ age of the mass as a whole, and is not to be accounted for by the sealing up of the original pores. -—17 G- 258 PAYING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 PERCENTAGE OF ACTUAL POROSITY Fig. 27. Curves showing physical changes in clay at various stages of burning. PURDY] PYRO-PHYSICAL AND CHEMICAL PROPERTIES. 259 Differentiation Between Clays on Basis of Difference in Rate and Manner of Decrease in Porosity and Specific Gravity. INTRODUCTION. Importance of Slow Vitrification —It is the concensus of opinion among those who have given serious thought to the vitrifying proper¬ ties of ceramic mixtures, whether natural, as ordinary clay, or artificial, as pottery bodies, that those mixtures which vitrify most slowly and at a uniform rate, all other things being usual, will produce the strongest and toughest ware. Chemical analysis and synthetical mixtures have failed to reveal the happy combination of minerals or chemical ele¬ ments that will produce this slow, uniform rate of vitrification. A few general rules can be stated as to combinations of ingredients required to produce tough bodies, but none of them can be applied with absolute assurance that they will operate in a given case. With our present in¬ formation empirical trials have to be resorted to find the proper com¬ bination in each case. It is commercially impractical to alter the composition of clays used for paving brick manufacture except in so far as different strata permit of the use or rejection of materials that effect the character of the ware. This the paving brick manufacture has learned by experience, so that the composite “dry pan” sample, before described, is supposed to repre¬ sent the best “mix” that is commericially possible in a given case. On the supposition that, according as its rate of vitrification is slower, one clay is more suited for vitrified paving brick than another, and that there is no means of obtaining information that bears on this problem other than determining this very pyro-physical property in paving brick elays, clays were molded into cones having the same shape and dimen¬ sions of Seger pyrometric cones manufactured by Prof. Edward Orton, Jr. PRELIMINARY TRIALS. Manufacture of Test Cones —The clays in this experiment were dry ground in a mortar to pass a 40 mesh screen, wetted with water from the University mains, wedged thoroughly and molded into cones with a spatula in a regular cone die as used by Orton. On the upper face of each cone was scratched its sample and serial number. After removal from the die the cones were placed in a cool place protected from drafts to dry. Setting of Test Pieces After Drying —One cone each of four different clays was set in a row in the center of a fire clay slab. On either side of the row of test cones was placed a row of three standard Seger cones arranged in opposite order from one another. There were eight groups of such slabs for each set of four test cones, thus allowing eight heat treatments of different intensities on each clay. 260 PAYING BRICK AND PAYING BRICK CLAYS. [BULL. NO. 9 The eight groups with the standard cones were as follows: First group 010-09-08. Second group 07-06-05. Third group 04-03-02-01. Fourth group 01-1-2. Fifth group 2-3-4. • Sixth group 4-5-6. Seventh group 6-7-8. Eighth group 8-9-10. Special saggars were prepared, being 3% inches deep and 8 by 8 inches in area, and having only three sides. These saggars were placed in four bungs in the side down-draft kiln designed by the writer for the ceramic department of the University of Illinois, and shown in Fig. 28. Four of these special saggars were placed in each bung, making 16 sag¬ gars in all. Burning —Four separate burns were made, one of the first four groups, one of the last four groups, and a duplicate or check burn on each. , The kiln was fired with coke, in a manner that maintained oxidizing conditions throughout the entire burn. In all four burns the fire clay slabs were burned to a clean huff color showing no evidence of having been subjected at any time to reducing influences. Inasmuch as the buff color of a fire clay is very sensitive to reducing action, and if once re¬ duced the huff tint is irrevocably bleached, confidence is felt that in these burns we were successful in maintaining oxidizing conditions. When a temperature had been reached sufficient to cause cone 09 to bend, the wicket was opened and the top saggers from each of the four bungs were drawn and placed in the ash pit of the kiln where they cooled slowly. After placing a cover over the exposed cones left in the kiln, the wicket was resealed and the heat raised until cone 06 was bend¬ ing, and so on until the center standard cones of the last set of four sag¬ gers were bending. By this scheme of setting twenty-four clays could be tested in one series of four burns, there being in each draw two slabs of the same group in each of the four saggars. This scheme of burning was made possible by the fact that the openings in the flash wall leading into the firing chamber, and openings in the opposite side of the firing chamber leading into the draft flue caused, with the down draft, an equal lateral distribution of heat. In no instance was there a- failure to have the center test cone bent, although in some cases in the same draw it was bent more than in others. Testing of the Trial Pieces —The cones were detached from the slabs, marked with lead pencil, weighed one at a time on a jolly balance and then placed in clear hydrant water. After twenty-four hours of satura¬ tion, the wet and immersed weights of each cone were made and from the data so obtained, their porosity and apparently specific gravity cal¬ culated. Difficulties Encountered —First, when the cones were detached from the slabs many broke into two or more pieces; second, a few of the cones were bloated at the base, due to a lack of oxidation; third, the cones were PURDY] PYRO-PHYSICAL AND CHEMICAL PROPERTIES. 261 invariably vitrified more at the top than at the base, thus causing ir¬ regularity of results in those that were broken; fourth, in those cones which had softened sufficiently to cause them to bend over, the pore sys¬ tem was not normal, owing to the strain set up on the upper side and compression on the under side of the bent cone; fifth, we were not sue- 262 PAVING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 cessful in making a jolly balance spring that was heavy enough to pre¬ vent the weight of a cone stretching it beyond its elastic limit, give suffi¬ ciently delicate reading. Data Obtained —Although the test as a whole was unsatisfactory, it is believed that the data obtained has a value. The porosity data were ploted on a diagram as shown in Figure 29. In this the linear distance between points on the abscissa, indicating difference in melting periods of the standard cones, is equal to the linear distance assigned to repre¬ sent a difference of two per cent in porosity. The solid line is drawn Fig. 29. Decrease in porosity with burning in terms of cones. through points representing the average data obtained on the two dup¬ licate cones, and the points indicated but not on the heavy black line represents in each case the data obtained from each of the two cones. In case the data for one of the duplicate cones were missing, as in K 2 for instance, the heavy black line traces the points representing the de¬ termined data. PURDY] PYRO-PHYSICAL AND CHEMICAL PROPERTIES. 263 The dotted line was drawn through all possible combinations of three points that were found to lie in line with each other. In some cases there was only one light line and in others more than one, as shown by the data given in Table XL. The lines drawn through three points lying in a straight line are taken as representing the slope of the curve describing the change in porosity with regularly increasing intensity of heat treatment. Where there is a possibility of more than one slope, as indicated by the light lines, each is recorded and their average cal¬ culated. The data in Table XXX is, therefore, the slope or tangent of the angle that the light lines drawn through three points makes with the abscissa. To obtain this data a protector was so placed on the line that the angle could be read. The natural tangent of the angle was then ob¬ tained from a logarithmic table of natural functions. Since the tangent of the angle which a. line makes with a given base line, is the slope or inclination of that line, this tangency can be taken as representing the rate at which the porosity decreases with increasing heat treatment. In the following Table will be found the values; first, of each of the tangents of the angles made by the dotted line and abscissa; second, the average of the tangents; and third, the rattler loss as determined on bricks obtained from factories using the several' clays. Table XL. Rate of Vitrification. Sample. 1 K- 1 K- 2 K- 3 K- 4 K- 5 1.1708 1.9007 2.0323 1.9626 2.0732 K- 6 K- 7 K- 8 K- 9 K—10 K-ll K—12 K—13 2.0503 1.5900 1.4733 2.0965 1.5301 1.1041 1.0538 1.0265 K—14 S - 1 S - 2 R- 1 R- 2 R- 3 . R— 4 . H—16 . H—18 . H—20 . H—21 . H- 1 . H—24 . H—23 . B- 1 . 1.0355 1.5697 1.3270 0.6208 0.9601 1.0538 .9657 1.9347 1.2799 2.5386 2.4142 1.1041 1.1106 2.7475 2.5826 2 3 4 1.9840 1.7675 1.1504 1.3151 1.1508 1.6107 0.9708 1.2685 1.9486 i.8103 1.2167 0.7673 1.7321 1.5061 2.1283 2.1445 0.6128 0.8541 1.1237 1.1882 1.4641 2.1609 0.7954 0.6208 Average. 1.1708 1.9007 2.0323 1.9626 2.0732 2.0503 1.7870 1.6204 1.5208 1.5301 1.3079 1.3322 0.9986 1,3714 1.5783 1.3270 0.6669 1.1677 1.2799 1.5470 1.7558 1.2799 2.5386 1.9391 1.6325 1.2344 2.7475 2.5826 Rattler Loss. 15.82 17.48 24.89 19.77 19.36 13.25 13.89 20.03 14.84 35.74 28.13 31.50 21.24 26.25 27.94 16.92 17.80 14.80 15.33 29.61 28.03 Summary —Owing to the unavoidable inaccuracies of the work and the erroneous assumption that a porosity graph would trace a straight line, the data given in the above Table has but little value. Its principal value lies in the developed fact that as a rule the slower the clay fuses the tougher appears to be the mass. 264 PAYING BRICK AND PAYING BRICK CLAYS. [BULL. NO. 9 FINAL TRIALS. Failing to solve the problem at hand in the above test, another and more thorough investigation was at once started, using not only a large number, but also a larger variety of clays. The manner in which the test pieces for this latter study were prepared was as follows: Wedging —Approximately one pound of dry clay was placed on a dampened plaster-covered table and sufficient water from the University mains added to develop the plasticity required to permit batting the clay into loaves. This was accomplished by adding water in small quantities, and thoroughly working it into the clay each time, until the mass had the desired plasticity. It was then thoroughly wedged by kneading and batting until, on cutting the mass open, it appeared to be compact, i. e. without air blebs. Molding —The loaf was then subdivided into smaller portions, each just sufficient to fill a mold % inch by 2% by 4^4 inches. The slabs were made to fill the mold by pressure applied in a screw press. They were then placed in a miter-box and cut into briquettes %. inches by 1% inch by 2% inches. Marking —The laboratory sample number and a serial number was stamped on each briquette. Drying —The briquettes were dried in an open room at summer heat. It had been found possible to dry even the most tender of clays in this manner, so it was assumed that all clays used in this test could, with¬ out detriment, be subjected to this treatment. Firing —Twenty-four briquettes of each clay were prepared. The ones on which the serial numbers 1 and 2 had been stamped were placed in a saggar to be drawn at cone 010, those on wdiich the serial numbers 3 and 4 were stamped were placed in a saggar to be drawn at cone 08 and so on—each successive pair of briquettes of each clay being placed in a saggar to be drawn at a predetermined heat treatment as follows: Series No. on briquette. Heat at which drawn. Hours interven¬ ing between draws. 1,2. 010 Oxidized at 800° 3,4. 08 for 2 hours. From 800°C to cone 010 6 hours. 2 hours 5,6. 06 2 hours 7,8. 04 2 hours 9,10. 02 2 hours 11,12. 1 2 hours 13,14. 3 2 hours 17,18. 5 7 2 hours 2 hours 19,20.. 9 2 hours 21,22. 11 2 hours PURDY] PYRO-PHYSICAL AND CHEMICAL PROPERTIES. 265 The briquettes in the saggars to be fired from cones 3 to 11 were packed loosely in coarse white placing-sand, as to prevent their stick¬ ing one to another. Only those clays known to be fire clays, or at least sufficiently refractory to withstand severe heat treatment were placed in the saggars to be drawn at the higher cones. The eleven saggars were placed in a coke-fired, side down-draft kiln in a manner convenient for drawing. The “spy” cones were centrally located in the kiln in a shield that protected them at all times from direct contact with the flame. When cone 010 was bent over sufficiently to touch the plaque, the wicket was opened enough to draw the cone •010 saggar, the wicket replaced, and the heat slowly raised as shown :in the above table. Cooling —The saggars in which the briquettes were placed were “tile •setters” 2 inches deep and 8 inches by 8 inches in area. Before plac¬ ing, another saggar was inverted over the one containing the briquettes, so that on drawing, the briquettes were at no time exposed to the rela¬ tively cold temperature of the room, except in one case of accident. The saggars were placed, uncovered, in the ash pit of the kiln, where they were exposed to the direct radiation from the hot grate bars above. In this manenr, the briquettes were cooled rapidly at first, thus pre¬ venting the fused portions in the briquettes from crystallizing very much, but from dull redness down to blackness the cooling extended over a. considerable period. The method of cooling pursued in this investigation was not ideal. The briquettes should have been cooled slowly for the first 200° C. which, as above stated, was not the case. Inasmuch as there is danger nf checking the vitrified briquettes by cooling down to room temper¬ ature too rapidly, some attention should be given to the last as well as to the first stage of the cooling period, but more particularly to the first. It was not possible to cool the briquettes under these ideal con-' ditions, for the services of the kiln were in demand for other purposes, and circumstances did not permit of delaying the burning until such time as the kiln would not be in use. Preparation of Briquettes for Testing —When cooled, sand grains were found to be fused to many of the briquettes, requiring that they be ground off on an emery wheel. Care was taken not to unduly heat the bricks while grinding off the sand, and yet as little water as possible was used. The bricks that were thus ground were washed in distilled water to remove all traces of dirt and adhering particles. From the unground briquettes all adhering particles were removed by a. dry stiff brush. Each briquette was carefully examined for flaws induced during manufacture or cooling, and also in order to remove all adhering por¬ tions, such as broken corners that might have been detached later in the test. Up to this point, all briquettes were handled together, without re¬ gard to sample or series number, except as before indicated. PAYING BRICK AND PAVING BRICK CLAYS. [bull. no. 9 266 In all, 60 clays were prepared for testing, as above described, nsing 16 to 22 briquettes for each. The briquettes were not sorted, those of each clay being treated as a unit, so as to insure like conditions at all times for all briquettes of the same clay. Drying of Briquettes —Briquettes belonging to two or three clays were placed in a drying oven and dried at 240° C. At the expiration of four hours at this temperature, they were cooled in dessicators pre¬ paratory to obtaining the dry weight of each briquette. The dry weight of each briquette was found to the third decimal place on a chemical balance. Saturation of Briquettes —After the dry weights had been obtained, the briquettes were placed in aluminum pans, keeping them arranged in the pans in their regular serial order. Distilled water was added until only the upper surface of each test piece was above the level of the water. This exposure of one face of the briquette was to permit easy escape of the air from the interior of the brick, as it was being displaced by the distilled water. After standing thus in water for 18 to 24 hours, they were completely immersed. After a total of 48 hours in water, the briquettes were placed in water under a bell jar, and -the air exhausted. In nearly every case, when a partial vacuum had been created, the air escaped from the briquettes at such a rate and in such volumes as to cause the water to ap¬ pear to be boiling. From a previous experiment, the data of which are given in the following table, it was thought that in the average case, fairly complete saturation could be attained with 15 minutes treat¬ ment in a partial vacuum. Table XLI. Showing efficiency of vacum treatment in effecting" saturation. Sample. Porosity as de¬ termined after 48 hours’ saturation without air ex¬ haustion. Percentage of Gain in Poro¬ sity at Conclusion of Vacuum Treatment extending oyer PERIOD OF 5 min. 10 min. 15 min. 20 min. S — 2. 3.22 48.1 51.8 57.9 65.0 G—11. 3.3 38.7 42.1 48.4 50.6 K— 4b . . 3.93 27.3 35.6 37.5 K—15d. 4.22 13.48 14.48 18.7 20.8 K—13c. 4.27 44.60 46.60 46.6 46.6 K—15c. 4.51 33.40 37.50 36.8 38.2 R— 4. 5.12 58.2 59.4 61.7 63.7 H—11. 5.29 31.2 35.4 37". 6 38.9 R— 2. 6.1 27.5 32.2 35.6 36.0 K— 6d. 6.46 29.9 31.6 35.3 39.3 K— 2. 6.55 18.6 20.1 21.6 24.3 R— 1. 6.7 10.2 11.0 11.0 11.0 B —11. 6.91 28.0 30.4 31.4 32.0 J — 11. 7.53 11.8 13.7 15.7 16.0 I — 11. 8.64 11.8 12.8 14.1 14.8 K— 8d. 9 06 22.0 23.5 24.0 24.9 B — 1 . 9.39 13.11 20.3 23.4 K—15b. 19.8 6.05 6.22 6.84 7.34 PURDY] PYRO-PHYSICAL AND CHEMICAL PROPERTIES. 267 Each saturated briquette was in turn suspended by a silk thread from the beam of a chemical balance, and its saturated weight taken, allowing for the weight of the thread. Without removal from the balance, a. glass of water was placed on a bridge spanning the scale pan in such a man¬ ner as to cause the briquette to swing absolutely free but completely immersed in the water. The suspended weight of the briquette was thus taken. Calculations —The percentage of porosity of each briquette was cal¬ culated by the formula: Wet Weight — Dry Weight Percentage of Porosity =-•--X 100 Wet Weight — Suspended Weight Plotting of Results —In the previous study, that with clays molded into cones, the writer had arbitrarily established the following propor¬ tion: Linear length on ordinate, equal to 2 per cent porosity; linear length on abscissa equal to difference of heat treatment of one cone: that is, 2:1. This( as before explained, was maintained between the coordin¬ ate factors of the porosity-graphs, so that the rate of decrease in porosity could be expressed numerically in terms of the tangency or slope of the curves, and that the factors so obtained would be comparable one with another at all times. The divisions on the abscissas of the specific gravity curves are the same as those of the porosity curves. The divisions on the ordinate are proportionally; 0.1 Sp. Gr. :2 .cone heat ::1:2. Data obtained —In the following table are the data obtained in the above study. Data for a few more clays were obtained, but owing to their incompleteness they are not recorded at this place: Table XLII. 268 PAVING BRICK AND PAVING BRICK CLAYS. [bull. no. 9 Commercial Possibility as Judged by this test. Paving brick . . .do . . .do . . .do . . .do . . .do . . .do . . .do. . .do. ..do. Building brick. Paving brick . . .do . . . .do . . .do . Building brick . Paving brick . Fire brick . Doubtful brick . Paving brick . ..do .. Decrease in Porosity. Cone. t-H3C-C-t-OSt-»aSt'*U5U5 00l-IOSt-»HeOOSt-l'it- Per cent. 85.0 90.0 90.0 92.0 92.0 84.0 86.0 87.0 90.0 93.8 81.1 71.00 83.90 93.50 87.50 68.2 79.0 28.1 88.7 95.0 94.8 Porosity and Specific Gravity of Briquettes Burned at Cones. tH 4.3 2.32 1.41 1.92 o w • oo cm © n • • t* h . ^cot^too • • • ■ :© CO - - CM C— • •T^Git>GOOt>'^ , C5 • 05 W.^ W050HX • © ^ —- t— CM i-H • •NHCOHCOHlOH • -CON.CM t—* CM CM r— • • • • O CM ^ t-h .... r- tn ■ -inoioowo .... co ih oo w ■ • ■ • co eo ■ eo — CD tH ■ ■lOHOiOtDO-^NOOD'tCOeiJrtDO • • • • CD CO ID Cd 00 M ■ • • ' CM ID O Ol • ■ CD CM >0 CM • • CO CM CM CM CM CM CM CM CM CO CM CM CM • ■ • C-CMCMCM-^CM • ■ • • ^ CM — — CM 40 05 CO 05 to rH CO GO FH CM *-h © ^* © 40 OO • • HtOOO . 1.0 CM CM ^ CM © © ©*** © CO CM CO ^ CM CM CM 40 -ON05H^H • • CM ^ © CO 05 . 40 • C- O ^ — CO CO t~CMCOCM©CMCMCMCMCMCOCM©CM© • D— CM rH CM © CM • • © CM ^ CM © • • • • -40 • CO CM —i CM if? CM i"H # • • i-H . CM CO O-^HOO C— C— CM CO t— OlOCO CO lO C 3 c- C— t-i CM 40 © 40 CO C— OO O —1/3 10 c— ©iO'^C04040^'*ti''^lt-40©^C040©40^l^CM'^©:0^40©^00^©©OT-C-©^CO’— COt-’t OCM*J’4O4OiOCO^4O'<*C-CM30©C-4O©4Oa0 4O 00 CM CM © CM © CM 40 • 40 CM C— CM tH CM CO CM* tr-* CM 00* CM t— CM ©* CM* © CM* OO* CM* C- CM lO CM 00 CM CO CM tO CM © CM CMi-HCMCMtH -CMtHCMCMtHi-HtHCMtHt-I CM t— i—i — © 40; W W C5 ^ CO N - CO ©©© tH OO C— 40 CO 40 CO CO © — CM C~©r-4O4O©©©t~CO00©4O •©©CM©C040©©©^'^iO©40rH40 40'ri40©40C040 ^CMiOCM'^CMtHCMC'-CM©CM 00 CMCM^CM©CM©CMC©CM-HCMt-CM00CM©CM©CM00CM00CM00CMOCM CMtHCMCMt-HCMt-H CMCMCMCMCMCMtHt-. CM CM T- — CM * o 40 -»* © © ▼— i CM 0O CO HlOlO ^ ^ ©40© T-. T— oo © © OC CO 40©CM40C~C^'r*C-C040©C-©^©©CM©©'**©©©'«^^40©40©40C040'«^CM©©40 40 40iOiOiO 05 CM © CM ©* CM OO* CM © CM* CM © CM ^ CM* © CM* OO CM* CM rH* CM CO CM OO CM © CM rH CM © CM t- CM t- CM CM CM — CM CM i—i CO CM CM CO CM CM CM t—< CM CM CM r— CM t— CM CM — CM CM © O GO 40 -h* OO © 40 CO © © CO 40 CO 00 © © © oo CM -TJ* — 40 r-( io *-H©©©40©hJ'©© 40©C— CO lO 40 C— ^ © CO 113 113 *13 1^- 40 ^ iO ^ © C— 40©©©00©©©©CM©C©40 CO CM* © CM 1-H CM* CO CM © CM CO CM 00 CM oo CM © CM t- CM l"- CM* t—I CM* 40* CM © CM © CM 40 CM CM* CM © CM — 03* t- C^ 00 CM rOCMCOCOCMCOCMCMCOCMCMCO CMCMCM — CMCMCOOJ CM 00 o C- 40 © © ©©© 40) 00 ^ 40 00 4 O © 00 i-H © © oo © oo rH © © © © C— ^ © 40 © © CM 4/3 CO © © © t— 40 © © © ^ 00 4/3 CO © 40 40 00 © ^ 40 CM © — © CM © © © © CM CM* CM © CM* © CM i-H CM ^ CM* © CM rH* CM vH* CM* © CM ■rH* CM* CO* CM ©* CM OO CM CO* CM CO CM ©* CM* © CM* © CM 00 CM CM CM CO CO CO CO CO CO CO CO CO CO CO CO CM CM CO CM CO CM CO CM CO 010 ©NOD-t'OO^lO©^ © © CO CM © © © ^ © 40 C— T—1 © 40 40 © C— ©©00©40©CM4040©CM^40 40CO©©©C— © CM 40 t— 40©©©© — © ©CMi-HCM^CM©CMCMCM^CMCOCM©CMCMCMCMCM'^CM^CM©CM©CM^C^1COCMCOCM©CM40CMOOCMOOCM COCOCOCOCOCOCOCOCOCOCOCOCMCMCOCMCOCMCOCMCO Sample Number. H N ^ 413 • . . . © t— ^ 1-H • i-HCMCO^©00i-Hr-irHTH-H t-H CM ^ w CM ^ i-h i i i i i i i i i i i V i i i i i 1 i 7 PURDY] PYRO-PHYSICAL AND CHEMICAL PROPERTIES. 269 co ! « : • c o o : : rt • M M : 24 is 2* ! 00 ^ • • : g ; : o U 4H B : tS « : 24 24 ,a : + : oo u .o he : o ,0*0 03 1-. 23 23 be ‘C bx: 5 • be 'e be be .2 be be B 43 c ^ ' • c 1 iJ a c 3 B B > 8 > s o c > > S c i > > ■*; £ o ! ’> > cS •“ rt B ! « ! -rH t C$ C$ 0) Tj 03 « Pm fe On cq i : : : : Ph Oh Q M Ph Ph rH r-i rH © CO 00 OO kD © CO co • • 00 CO ID r*< ID CM • • H CM CO* CM © CM 14.0 2.51 17.9 2.57 3.7 2.38 . • rH CM rH • 't-HCOiClOH rH CM* © CM 00 CM 1.6 1.72 17.1 2.52 C~ © rH co r- oo ©©©rH~Hrr©00 th^hldcmoocmcmcm 4.8 2.43 17.0 2.51 21.2 2.60 3.6 2.38 11.7 2,41 Ic- OO CO © ©©iDCM©kD00rH rHrHrHCMrHCMt-CM 3.0 1.89 3.4 1.68 2.3 1.78 20.0 2.56 00 CO CM © © 5DH^^^CCHH HNHiJIgNMN ©CO t— r- oo CM iDrH©©lDlDrHCMCMrH CM © CM © CM ID CM CM CM rH ID 00 OO CMOOlDCMCMiDt-r* 00 rH ID CM rH CM © CM 6.5 1 87 2.1 1.73 1.4 2.35 19 2 2.56 1.74 2.29 16.8 2.47 24.3 2.59 5.0 2.41 OO ■H" 7— CO C-rHlD©©©©-fi00rH © CM ©* CM © CM © CM'oO CM © rH rH 1C © © rH rH CM ID 00 • CM O' CM* CM © CM CM CM 00 ID r- ^ * 7.7 2.29 1.8 2.00 1.3 2.31 21.7 2.57i © ID CM r- r-rjD CM »D CM © CM © CM 00 CM rH CM tH CM rH CM IT- CM tH CMOO CM rH CM 00 CM OC CM rH CM CM -r- t-H rH rH rH CM OOCMrHOOOOaOCOlDlDr : H©COOC©rHeOOOOOrHCMCMrHr*i C-^t-^Das^DOlCOOasWOXOlOCOOJ^OCDtOCDCOiOt-LOH^lrt^OliCOOOOHOS^t^iCHift^CD iOCMOOCMr*CM©CMCMrHC-CMOOrHCMCMCMCM©CMCMCMCOCMOOCMrHCMiOCMOOCMCOCM©CM©CM-^CMCMCMrHCM rH CM CM rH CM CM CM CM CM r-i — rH CM rH CM CM CM )COCOM»OCOHOTHOOCOO^C^OOCOI>?DC^liOCO»Ol>'^Wi0^iCHCD' I* CM CM* ID* CM rH CM O CM rfi tH CM CM CO CM CO* CM* H CM OC CM* CM CM rH* CM ID CM 00 CM* CO CM < CM CM CM CM CM © CM C- ID LD©©aD^4D©© CMCM©CMCOCM>DCM CM rH CM CM ©©OOC^OOOO©OOu-tfOCM©OOCO©©rH©©~H©© C-'H'HCOWC£)l-iCt-H^H0500C^^DHiOPCu:XCDCO'X)Ci C- CM »D* CM t- CM >n CM CO CM* O CM "H* CM* C~ CM ©* CM* QO CM 00 CM* t- CM* CO CM rH CM rH CM 00 CM* 00* CM i-i CM CM CM CM rH CM CMCMCMCMCMCMtHCMCMCMCM OOCDOiiCOOCOHCO M »o CM O CM r+n CM CO CM O'! 03 OO rH 00 tH rH ID CD ^ 40 rH U— !iDO0©CO©©©O0rH0O©lDlDO0©rH©kD©0O©OO©©lD' iCM©CMtr-CMOOCM©CMrHCMtr-CM©CM©CM©CMrHCMoOCMiOCM' D CM © CM O CM »D CMID CM - CM 00 CM . OM CM CM CM CM CM CM CO CO CM iiDOO©©©CM©lDkDrH©OOlD©C^rH©0O©aO©©©C-©CMiDCMlDCM©< < CM ID CM t- CMOOCMrHCM©CM©CMOOCM©CM©CM*©CMOOCM©CM©CM-*-CM©CM< l CM CM CM CM CM CM CO CM CM CO CM CM CM CM CM < >©CMlDOOiD©iDOOkDiD© > CM CM* CM* »D CM ID CM 00 CM CO* CM I I I Ih I I I I I I I I I I I I I I > > > > > Upper figure in each case is percentage of pore space. Lower figure, specific gravity. *Porosity data marked with a star were calculated by interpolation. 270 PAVING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 On plotting the data obtained in this experiment they were found in most cases to be consistent, i. e., clay used for particular industries such as paving brick, fire brick, etc., exhibited porosity changes that were so concordant that the possible commercial use for each was pre¬ dicted from the curves and in no case where the clays are now being employed did the predicted use differ from their present use as re¬ ported by those who collected the samples. The curves in every instance were not straight, but curved so that their tangent or rate of declination could not be ascertained without the use of calculus. Inasmuch as the curves did not describe grad¬ ually sloping curves, but in most cases exhibited well defined lags in decrease of porosity, it was found that a simple tangent factor would not describe in full the fusion behavior of the clays. A complicated modulus was devised which was not only a function of the tangents of the sections of the curves between points of lag, but also the length of each section. Considering the fact, however, that this scheme of study¬ ing the fusion phenomenon is here first presented, thus not finding confirmation by other experimenters, and since the modulus does not show more clearly the rate fusion than does the curve, no attempt was made to apply the modulus on the different types of clays. SUMMARY OF RESULTS OF TESTS. In subsequent curves are given the limits of the areas traversed by the porosity and specific gravity curves of the different types 1 of clays. In Fig. 30 are shown the limits of area traversed by porosity-graphs of the fire clays. The fire clays are grouped into three classes according to their rate of decrease in porosity. Number One Fire Clays —The writers of Clay Reports have heretofore failed to recognize that of two clays having similar ultimate chemical compositions and similar ultimate fusion periods, one can be used in No. 1 fire brick, while the other would fail as a first-class fire brick material, 2 and the one failing as a fire brick material would be the only one that could with success be used in the stoneware industry. Sev¬ eral examples of the foregoing were noted in the examination of the Illinois fire clays. In fact, the case is not an uncommon one. In fire brick, maintenance of an open structure through the entire heat range used in the various ceramic industries is essential. On the other hand, in stoneware, closeness of structure at comparatively low temperatures, or early vitrification followed by a long fusion range is absolutely required. It is evident, therefore, that a classification of re¬ fractory fire clays (so called because they withstand heat equivalent to cone 27 or more without failure) should take account of this dif- 1. “Types,” as here used, does not refer to geological origin or age, but rather to the possible commercial use of the clays. 2 By fire brick material is meant what is known in trade as No. 1 fire brick The so-called No. 2 fire bricks are, as a rule, not worthy of the distinctive title “fire brick.” Used in places exposed to fire does not necessarily make a brick a fire brick, for, if this were so, the comparatively fusible Chicago brick placed in the arches of their scove kilns would have to be called “fire brick.” PURDY] PYRO-PHYSICAL AND CHEMICAL PROPERTIES. 271 Fig. 30. Differentiation of fire clays on basis of porosity changes. ference in their manner of fusion. This essential difference in the be¬ havior of fire clays is recognized in a tentative scheme of classifica¬ tion presented by the present writer and Mr. Moore. 1 It will be noted from Fig. 30 that these clays show comparatively little decrease in porosity from cone 010 to cone 11. This decrease averages from 7 to 15 per cent of the initial porosity and in no case does it exceed 17 per cent. The specific gravity, 2 as shown in Fig. 31 remains fairly constant from cone 010 to cone 3 and then, even in the purest clays, it begins to decrease slightly. This decrease in specific gravity in the No. 1 fire clays, even when the porosity remains very high, is considered as evidence of the influence of the adsorbed or cementing salts which, while constituting but a very small part by weight of the whole, are nevertheless potent factors in causing fusion. 1 Trans. Am. Soc., Vol. IX, pp. 239. 2 The specific gravity here referred to is the specific gravity of that portion of a saturated brick not occupied by water. Inasmuch as this water impermeable mass very often, in fact, in the case of impure clays generally does contain in¬ closed or sealed pores known as blebs, the specific gravity so obtained cannot be the actual specific gravity of the material of which this water impermeable portion consists. ''The true specific gravity of the material can be obtained by crushing the brick to fine powder, thus eliminating the sealed pores, and then determining the specific gravity of the powder in a specific gravity bottle as before described. For this reason, the writer has classified specific gravities under three heads: First, false specific gravity, or the weight per unit volume of the whole brick; Second, apparent specific gravity, or the weight per unit volume of the water-impermeable mass; Third, true specific gravity, or weight per unit volume of solid material in the water-impermeable mass. 272 PAVING BRICK AND PAVING BRICK CLAYS. [BULL. NO. & Fig. 31. Curves showing changes in specific gravity of fire clays with pro¬ gressive intensity of heat treatment. dumber Two Fire Clays —It will be noted that while the decrease in specific gravity of this group of clays is about the same as that shown in the No. 1 fire clays, the porosity shows a much larger decrease. The ■earthy vitrification and slow fusion is quite pronounced in this group, permitting their use in the paving brick, sewer pipe, stoneware and ■terra-cotta industries, but not in the manufacture of No. 1 fire brick. Number Three Fire Clays —In Figs. 27.and 28 are shown the limiting area of porosity and specific gravity curves of a class of clays which, in the judgment of the writer, ought to be put in a different category from the preceding group, or number two fire clays. Heretofore, both have been classed together indiscriminately in ceramic and geological PURDY] PYEO-PHYSICAL AND CHEMICAL PROPERTIES. 273 ■literature, as number two fire clays, but they are not the same. Clays of this class differ from the No. 1 and No. 2 fire clays, in that they •seldom have a fusion point exceeding cone 16 or 17, fuse in a very irreg¬ ular manner, and exhibit a much larger decrease in specific gravity owing probably to the presence of iron in nodular form as sulphides or carbonates. Fire Clays in General —These conclusions may be summarized as fol¬ lows: First, While all the types of fire clays here tested maintained •the same range in porosity up to cone 010, there is a marked differ¬ entiation of each at cone 08. Second, From cone 08 ta about cone 1 •the No. 2 and No. 3 fire clays traverse a common area, but at cone 1 •the No. 3 type begins to fuse more rapidly, until when cone 7 is •reached, the No. 3 fire clays have fused sufficiently to be wholly differ¬ entiated from the No. 2. Third, Since the porosity curves in Fig. 27 are composite curves showing the limits of variation in the few clays tested, it is possible that broader limits will be determined when more and a larger variety of clays are tested, yet the data here presented are sufficient to demonstrate that where chemical analysis and fusion period determinations have failed, this method of differentiation has proved successful. Fourth, Differentiation of firer clays on the basis of specific changes will hardly be possible on account of the limited differences between the areas traversed by the specific gravity curves of each type of •fire clay, yet as is shown in Fig. 28 the specific gravity curves parallel and diverge from one another at about the same temperatures as do the porosity curves in Fig. 27. Chemical analysis and ^points of fusion of a few of the fire clays from which curves were drawn are as follows: Table XLIII. No. 1—Fire Clays. Sample Number. Moisture. Volatile Matter. Si0 2 . ai 2 o 3 Fe 2 0 3 TiO z Total. Fusion point. H— 24 .... 0.6 4.63 76.10 15.31 1.10 1.31 99.06 30 V-ll.... 1.74 10.28 56.28 26.68 3.24 1.29 99.50 Not reached... F—18.... 0.84 6.66 66.88 21.87 2.23 1.18 99.86 29 F—19.... 1.19 6.31 68.12 20.08 1.76 1.16 98.62 31 Table XLIV. No. 2—Fire Clays. Sample Number. Moisture. V olatile Matter. Si0 2 ai 2 o 3 Fe 2 O a Ti0 2 Total. [Fusion point. V—4 .... 2.37 8.84 54.80 29.44 1.70 0.82 97.97 Not reached... K-12.,.. 0.60 10.09 54.37 23.61 6.14 *5.97 100.78 ..do. l Total fluxes TiO, was not determined in K-12. —18 G 274 PAYING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 Chemical analyses were not made of all the clays of the Xo. 1 and Xo. 2 type and none of the Xo. 3. From the few that were made, how¬ ever, it is evident that refractoriness and slow fusion are not always de¬ pendent upon the proportional content of alumina and silica, for the two Xo. 2 fire clays have on the average higher AhOs and lower SiO* content that the Xo. 1 fire clays. This is directly contrary to our past teachings and contrary to what might be expected from Segar’s AbCh- SiCb curve, as shown in figure 19, page 208. Paving and Building Brick j Clays —The standardization of tests for first-class paving brick clays has been and perhaps will be for some time the subject of consideration by ceramic investigators. The pyro-physi- cal and chemical tests here reported can be said to give negative rather than positive information, in that they very effectively differentiate the clays they cannot from those that may, be utilized in paving brick man¬ ufacture. Judging from the results so far obtained, they fail, how¬ ever, to differentiate the paving brick clays one from another in re¬ gard to their comparative quality. For example, we have not been able to distinguish by these tests between the clays of 14 per cent and the 24 per cent type, measured in per cents of loss in the rattler test, nor between the clays that preserve their maximum strength through a wide heat range and those which attain and preserve their maximum strength only within a very narrow heat range. The cause of failure of the pyro-chemical studies in this respect is, no doubt, to be found in the fact that inherent strength is not wholly a function of rate of vitrification or development of vesicular structure. As shown in earlier pages, physical tests on the raw clays failed to dif¬ ferentiate paving from building brick clays. The pyro-chemical studies here reported are the only ones that give any clue to cause of toughness or strength of the burned ware. Pyro-chemical studies similar to those here outlined, together with a determination of the maximum strength and the range of temperature in which this maximum strength is developed, would enable the observer to properly classify and differentiate paving brick clays. This, however, amounts to a sub-classification of the paving clays on a basis different from that of the main sub-division. The striking differences between the building and paving brick clays are apparent from figures 32 and 33. Earlier vitrification, irregularity in decrease of porosity and specific gravity, apparently larger quantity of vessicular glass formed within the.mass, or at least a more notable bloating, due, to volatilization of certain constituents, probably the sol¬ uble and adsorbed salts, are the distinguishing features of the strictly building brick class. Sufficient evidence is at hand to warrant the statement that any clay which vitrifies to a porosity of 2 or 3 per cent before cone 5 is reached, in the heat treatment prescribed in this method of burning test pieces, PURDY] PYEO-PHYSICAL AND CHEMICAL PROPERTIES. 275 will be too brittle for use as paving brick material, no matter how little vesicular structure is developed. The fact is, however, that it will be a rare case in which vesticular structure is not strongly developed if the clay shows an early 1 and rapid rate of vitrification. In figures 32 and 33 are shown the upper and lower limits of areas that were traversed respectively by the porosity and specific gravity curves of. clays that either are being or can be used for the purposes in¬ dicated in the figures. Fig. 32. Curves showing- changes in porosity of paving and building brick clays with progressive intensity of heat treatment. 1 The use of the comparative terms “early” and “rapid” in reference to this type of clays, in contrast to their relative use in regard to fire clays, is best illus¬ trated by reference to the curves. 276 PAYING BRICK AND PAVING BRICK CLAYS. [BULL. NO. 9 Fig. 33. Curve showing changes in specific gravity of paving and building brick clays with progressive intensity of heat treatment. The boundary limits shown in these figures are those obtained in these tests, and, therefore, may not show exactly the true limits of the several areas. They indicate, however, approximately the relative man¬ ner in which the clays used for the several industries behave in fusiilg. All clays used for paving and sewer brick may be used for building brick, but what are here defined as strictly building brick clays cannot be used for paving or sewer brick. All paving brick clays can be used in the manufacture of sewer brick, but the sewer brick clays cannot be used for paving brick. The points of differentiation are; first, the paving brick clay fuses more slowly and decreases less in specific gravity; PURDY] PYRO-PHYSICAL AND CHEMICAL PROPERTIES. 277 second, the sewer (and side walk) brick clay fuse more rapidly bnt maintain their shape through a considerable range of heat treatment before failing; third, those clays which are fit only for building brick vitrify rapidly and fail as soon as, or before they are completely vitri¬ fied. The sewer brick clays can be brought with safety to complete vitrification without much danger of loss except perhaps from “kiln marking” while those clays which are fit only for building brick bloat and become spongy as well as soft almost as soon as vitrification takes place. Since the tracing of the porosity curves through the upper or paving. brick clay area does not necessarily signify that they are good for.paving brick manufacture, the lower limit may appear to be superfluous. It remains a fact, however, that, according to the tests here. reported, a clay must have its porosity curve confined within the limiting bound¬ aries shown in order to develop the required toughness. So far as ex¬ perience with the Illinois clays is concerned, the curves for porosity and specific gravity in figure 29 and 30 respectively, denote quite rig¬ idly the allowable variation in rate of decrease in porosity and specific gravity. GENERAL CONCLUSION. In the preceding discussions of physical, chemical, and pyro-physical and chemical properties of clays all of the relations between these prop¬ erties that were known or observed have been shown. A review of these discussions reveals the following as being the most important. 1. Measurement of some of the properties failed to give results that show the factors which affect them or which are involved with them. This was made plain in case of the “individual grains” as obtained by mechanical analysis. We have seen that, the methods universally em¬ ployed to effect the physical disintegration of clay are not sufficiently intensive to produce complete disintegration. It has also been demon¬ strated that the grains or particles so obtained do not usually consist of one mineral substance. As a consequence of this cementation of smaller particles of different substances into bundles or groups, any in¬ ference or conclusion based on fineness of grain cannot be very general in application. 2. Ultimate analysis or gross rational analysis of clay cannot reveal qualities that affect either the “working” or “burning” properties. 3. Either ultimate or rational analysis of the several groups of grains may reveal some important relation of constitution to manifested prop¬ erties. This, however, remains to be demonstrated. It can be said however, that such determinations will not likely become “commercial” tests of clays. On the other hand, however, it seems certain that they will be valuable for research purposes. 4. Vitrification behavior, rate of fusion or toughness of bricks, do not seem to depend within any but very wide limits or in any traceable manner upon chemical or mineralogical constitution of clay. 5. No combinations of physical and chemical properties can be said to be essential to clays from which first-class paving brick may be manufactured. 278 PAYING BRICK AND PAVING BRICK CLAYS. [bull. no. 9 6. The most satisfactory tests tried or developed during the course of these researches for distinguishing between clays on the basis of their commercial availability are rate of decrease in porosity and specific gravity. While even these tests, so far as can be judged by our results, do not make an absolute discrimination, the discussions and curves here given make plain the fact that such tests are the most serviceable of any so Tar developed.. The other tests have special uses and are not to be entirely condemned. 7. Toughness of brick does not bear a consistent relation to degree or range of vitrification. Each clay has its own peculiar range and de¬ gree of vitrification at which its maximum toughness is developed. In . some clays this range is very small and in some quite large. In some clays maximum toughness is attained when the brick still shows an ab¬ sorption of 8 or 12 per cent and in others not until the absorption has been decreased to 2 or 4 per cent. No tests other than the “rattler test” on full size brick which have been burned with different intensity of heat treatment have brought out data which bear on this peculiarity of clays. 8. The pyro-physical studies which have been described suggest a series of determinations which should be more valuable in that they ought to reveal the cause for this want of correlation of toughness and vitrification behavior. The series of determinations referred to is that of the volume changes which take place with increasing intensity of heat treatment. 1 The volume changes which are important are: (a) exterior volume of brick. (b) volume of skeleton of brick. (c) volume of open pores. (d) volume of sealed pores. 1 See Trans. Am. Cer. Soc., Vol. X. }