t^o, I C> r j 4. 5 IS o Preprint of Copyrighted Proceedings AMERICAN CONCRETE INSTITUTE Subject to Revision Not Released for Publication AMERICAN CONCRETE INSTITUTE The following paper is to be presented at the annual con- vention of the Institute, Chicago, February 14 to 16, 1921. Written discussion is solicited and should be forwarded to Harvey Whipple, Secretary, American Concrete Institute, New Telegraph Building, Detroit, Mich. TEST OF A FLAT-SLAB FLOOR OF THE NEW CHANNON BUILDING By H. Ff^ONNERMAN* AND F. E. RlCHART** II *» It is the object of this paper to present the results of a test made by the writers in July, 1920, upon a floor slab of the new Henry Channon Building in Chicago. The test was made under the general direction of a commission appointed by the Chicago Building Department and consisting of Prof. A. N. Talbot, of the University of Illinois; Prof. D. A. Abrams, of Lewis Institute, and B. E. Winslow, of the Chicago Building Depart- ment. The architect for the building was A. S. Alschuler, the designing engineers Morrison and Beck, and the contractor R. F. Wilson & Co., all of Chicago. The building in which the test was made is a seven-story reinforced- concrete structure located at the southeast corner of Market and Randolph Sts., Chicago, Illinois, and was erected during the summer of 1920. The type of floor construction used throughout the building was a modified form of the Smulski or S-M-I System; the principal variation from the standard S-M-I design being in the arrangement of rectangular and diagonal bands over the column heads, and in the distribution and amount of steel in the different ring units. The location of the test panels in the building is shown in Fig. 1. General information regarding the floor and the test is summarized below for convenient reference. Slab. General Data of Slab and Test. Area loaded 4 interior panels, 4th floor. Panel dimensions 20 ft. y 2 in. square. Nominal thickness of slab 8 in. Nominal thickness at drop panel 12*4 in. Dimensions of drop panel 6 ft. 6 in. square. Diameter of column capitol 4 ft. 6 in. Age of slab at beginning of load test ... 53 days. ♦Los Angeles, California. ♦♦University of Illinois, Urbana, Illinois. : 2 Flat-Slab Floor Test. Loads. Design load Dead load Loading material for test 200 lb. 350 lb. Load increments and 500 lb. length of time 650 lb. applied. 500 lb. 500 to 0 /SO -S' 200 lb. per sq. ft. 100 lb. per sq. ft. Building brick. per sq. ft. for about 24 hours, per sq. ft. for about 24 hours, per sq. ft. for about 24 hours, per sq. ft. for about 18 hours, per sq. ft. for about 1 week, per sq. ft. for about 2 weeks. Afaz-A-ef S/ FIG. 1. LOCATION OF TEST PANELS. Observations. Number of strain gage lines 337 Number of deflection points 22 Number of observations made 4000 Arrangement of Slab Reinforcement. The reinforcement in the test floor consisted of rings, straight bars and “truss bars.” The general arrangement of the reinforcing bars in the test area as designed, is shown in Fig. 2. The size of all bars was checked before the floor was poured and, with a few slight exceptions the rein- forcement was as shown in Fig. 2. The position of the reinforcement was also inspected before pouring and in general the steel was found to be Flat-Slab Floor Test. 3 Schedule of Reinforcement F ourth F/oor S/ab -Live Load -200 /h per sq ft Co/umns Unit Truss Rods •Sfmigh Rods Rings - size and diameter 5tory Column No. 3 '-O' 4-6" 5'-0"\ 6-0' 7-6 ' 9'-0' /0-6' /2 1 0‘ /3'-3" !4'-6‘ D78 £7 £678 06 £ 6,8 A 3-ft ft ft ft ft ft ft 4 Co/ ze '4 Core 22 " 0 tods //-/"<* Spiral ia-Zp 28’ 0 24’ 0 /6-fo 10-/ ZP B 2 -i'° i-r+ 4 t ft ft ft ft ft ft f* ft c 4-} V l* ft ft ft 3 Co/ 28“ 0 Core 24" f Tods //-/}> Spiral ,i"0-2i‘p 30 "a 26" 4 /'8-fa, f*;2"P T II-} 4 FIG. 2. DETAILS OK REINFORCEMENT. 4 Flat-Slab Floor Test. reasonably close to its designed position in plan. In pouring the concrete the steel became displaced vertically in many places. Care was taken when placing the ring units in the test area to avoid lapping of the bars at places where gage lines were to be located, but in one instance a gage line was located where a lap occurred. The bars of the ring units were generally lapped about 40 diameters, and the position of laps is shown in Fig. 2. A view of a part of the test area just before the concrete was poured is shown in Fig. 3. A large number of corks were placed between the reinforcing bars and the form to facilitate the opening of gage lines in the steel, and some small steel plugs were lightly tacked to the forms FIG. 3. VIEW OF REINFORCEMENT IN TEST AREA. to provide for gage points for measuring compressive strains in the con- crete. Corks were also wired to bars near the upper surface of the slab, and, although they were usually covered up during the pouring of the concrete, they were quite useful in aiding the location of gage lines on the reinforcing bars. Havemeyer deformed bars were used for reinforcement with the excep- tion of a few plain round bars in the ring units. There was not much variation in the strength of the various sizes of bar, as found from tension tests on coupons cut from the reinforcement as it was being placed. The average unit stress at the yield point was found to be 50,000 lb. per sq. in., and the average ultimate strength 84,000 lb. per sq. in. Flat-Slab Floor Test. 5 Concrete in Test Floor. The concrete in the test panels was poured May 7, 1920. It was mixed in the proportion of 1 part Universal portland cement, 2 parts torpedo sand and 4 parts broken limestone and was of quite uniform consistency. The concrete was poured rather wet, excess water being taken up by scat- tering a mixture of cement and sand over the surface and then finishing with trowels. A monolithic floor finish was secured by troweling into the surface of the floor a mixture of equal parts of cement and ironite. During the pouring of the floor, 6 x 12-in. test cylinders were made of concrete from various parts of the test area. This concrete gave a slump of 8 to 9 in. in a 6 x 12-in. cylinder. The results of compression tests made on these cylinders at Lewis Institute are given in Table I. Table I. — Compression Tests of Concrete. No. of Cylinders Tested 5 Manner of Storage In air Age at Test Days 7 Compressive Strength lb. per sq. in. 1270 Initial Modulus of Elasticity lb. per sq. in. 2,840,000 5 a cc 28 2280 3,200,000 6 « u 60 2870 3,630,000 5 a cc 66 3070 3,660,000 5 In moist closet 28 2750 4,110,000 5 a (( (( 60 3800 4,720,000 The Test. The test was performed in the ordinary manner by applying a load and taking observations of deformations and deflections at various stages of the loading. There were 163 strain gage lines on the reinforcing steel and 174 on the concrete, making 337 in all. Leadings of the deflection of the test floor were taken at 22 points. The manner of taking strain read- ings is well known, but the scheme for measuring deflections differed from the usual methods. As shown in Fig. 5, an Ames micrometer dial was mounted on a long wooden pole, the extreme ends of the pole and dial plunger being shod with conical steel points. These points engaged small holes, drilled in steel plates attached to the lower side of the test slab and the floor below. The one instrument was carried from point to point and was frequently checked on a standard gage length. It is believed that the instrument produced very reliable results. This means of measuring deflec- tion is much simpler than that of erecting timber standards at each deflec- tion point, and is less liable to accidental disturbance. The loading material was brick which later was used in the construc- tion of walls of the building. The individual brick were approximately 2y 8 x 3% x 8 in. in size, and their average weight was 4 lb. From meas- urements and counts of the brick as they were piled on the test panels the weight per sq, ft. of area for one layer of brick on edge was found to be 6 Flat-Slab Floor Test. about 30 lb., and this value was used in calculating the various increments of load. The bricks were piled on the test panels by brick masons in a workmanlike manner, and it is believed that the load of 30 lb. per sq. ft. closely represents the actual weight per sq. ft. applied to the floor for one layer of brick. In piling the brick, aisles generally not over 4 in. in width were left as indicated in Fig. 1, in order to prevent arching of the loading material. The space covered by the brick amounted to 96 per cent of the total area of the four loaded panels. The load was applied in four increments, one day being required to place each increment of load.- The load at which strain gage readings and FIG. 4. VIEW OF TEST LOAD. deflection readings were taken were as follows: 200 lb. per sq. ft. (6 layers of brick on edge, 1 layer on side ) ; 350 lb. per sq. ft. (11 layers on edge, 1 layer on side) ; 500 lb. per sq. ft. (16 layers on edge, 1 layer on side) ; 650 lb. per sq. ft. (21 layers on edge, 1 layer on side). The maximum applied load was two and one-sixth times the design live and dead load; with the load of the floor included, the total load amounted to two and one-half times the design live and dead load. Since the floor was already stressed by its own weight, only the strains produced by the applied load were measurable. Small tunnels of timbers and planking were built over the gage lines on the upper surface of the test floor, and since the same number of layers Flat-Slab Floor Test. 7 of brick was piled on the top of the tunnels as elsewhere, the intensity of load on these tunnels was the same as that on other parts of the test floor. The load on the tunnels was transmitted to the floor in such a manner as to cause but little variation in either moment or shear from a condition of uniform loading. Fig. 4 gives a view of a portion of panels A and B showing the load of 650 lb. per sq. ft. in place on the floor. Before load was applied to the test panels duplicate sets of strain - gage readings and of deflection readings were taken on all gage lines and deflection points. One set of strain readings was taken on all gage lines at a load of 200 lb. per sq. ft., and at the other loads two sets of strain readings were generally taken on all gage lines, the second set of readings being taken after the load had been in place from 12 to 14 hours. When the second set of strain readings at a load of 650 lb. per sq. ft. had been taken after this load had been in place for approximately 18 hours, the last increment of load was removed, leaving a load of 500 lb. per sq. ft. on the floor. Deflection readings under the latter load were then taken. A load of approximately 500 lb. per sq. ft. remained on the floor for 7 days, when strain readings and deflection readings were taken on all gage lines and deflection points. During the following 20 days the bricks were gradually removed from the test floor as they were needed in the con- struction of the building. When all the load had been removed, a complete set of strain and deflection readings was taken. Readings of the temperature of the air in the building taken from time to time as the strain readings were being taken ranged from 71° to 87° F. over the period of the test. It is probable that the variation in the tem- perature of the floor slab was much less than the variation in the air temperature and strain readings taken on a reinforcing Bar in the fourth floor of the building well away from the loaded area and, therefore, un- stressed by load, were in no case more than one division on the dial of the strain gage away from the average of the readings, a difference which corresponds to 750 lb. per sq. in. of steel stress. As the usual run of differences was much less than this it was thought not necessary to make a correction for temperature and, accordingly, none was made. Deflections. The location of all deflection points and the deflection of the floor under various increments of load are shown in Fig. 5. The diagram shows a remarkable uniformity of action at corresponding points in the slab. The maximum deflection under the load of 650 lb. per sq. ft. is seen to be 0.59 in. at deflection point 4, while the average of the maximum deflections at the centers of the four panels is 0.54 in. The increase in deflection as each load increment remained on the floor over night is evi- dent from the diagram, especially at the higher loads. The amount of recovery of the slab toward its original position after all load had been removed is also shown in the diagram by circles at the line of zero load. 8 Flat-Slab Floor Test, Flat-Slab Floor Test. 9 Appearance of Cracks. A study of the appearance of the cracks in the test slab is useful not only because the cracks indicate regions of high tensile stresses, but also because the size and distribution of the cracks furnish a good index of the relative amount of tension being carried by the concrete. The cracks on the lower surface of the sfUlb were first observed at the load of 350 lb. per sq. ft. at the center of the loaded panels and at points midway between columns. At the load of 500 lb. per sq. ft. generally a single crack extended along the section of maximum positive moment from center to center of adjoining loaded panels. At the load of 650 lb. per sq. ft. several cracks had formed as shown in Fig. 6, running parallel to the cracks first noted and within a narrow zone about 2 ft. wide. The presence Load- 650 Lbpersqff FIG. 6. CRACKS ON UPPER AND LOWER SURFACES OF SLAB. of such cracks constitutes a good criterion of the spread of higher stresses developed by the positive bending moment. It has been observed in other tests that as the load on the slab approaches the maximum which the slab will carry, cracks are formed outside the belt observed here. For a greater load than 650 lb. per sq. ft., then, a wider belt of cracks would probably be visible, and a smaller proportion of the moment would be carried by the tension in the concrete. Careful search failed to show any cracks following the direction of the ring units in the bottom of the slab, such as have been found in other tests at higher stresses. As nearly all of the upper surface of the floor was covered by the loading material, the development of cracks during loading could only be followed near the tension gage lines in the observation tunnels, but it was undoubtedly similar to that on the lower surface. After the load was removed the cracks on the upper surface were mapped, but it is likely that many of the finer ones had closed up, leaving only the larger ones 10 Flat-Slab Floor Test. visible; the visible cracks are shown in Fig. C. This sketch has several significant features: it shows remarkable uniformity of behavior in the four test panels; it indicates high stresses along inner panel edges, as well as in portions of the unloaded slab along the outer panel edges; and it shows several important cracks around the center column. It is seen that one crack formed just outside of the inner ring of Unit C along its entire circumference. This ring is located 3 in. outside of the edge of the column capital. Another crack w^as found just outside the second ring for a part of its circumference. A third series of large cracks was found just above or a few inches outside of the edge of the drop, and radial cracks FIG. 7. TYPICAL LOAD-STRAIN DIAGRAMS. extended outward from just above the four corners of the drop for a considerable distance toward the center of the test panels. It should be noted that many of the cracks were found near and fre- quently across tension gage lines, and this explains some variations in stresses measured on gage lines having similar locations on the slab. Further information regarding the formation of cracks is suggested by the load-stress curves of Fig. 7, which were plotted from the strain meas- urements. Such diagrams for gage lines lying across or near the sections of maximum moment show a bend or change in slope in the curves at a point between the loads of 200 and 350 lb. per sq. in. This is true for the tensile stresses on both upper and lower sides of the slab. The bend occurs at a measured stress of about 3000 to 5000 lb. per sq. in. Cracks were generally visible at these gage lines at the load of 350 lb. per sq. ft., and their development quite evidently produced the increased rate of stressing Flat-Slab Floor Test. 11 of the steel. At points of less stress the cracks developed later. At the maximum load the largest cracks on the lower side of the slab were ob- served at the centers of panels and were estimated to be in the neighbor- hood of 0.01 in. in width. At the same load, the cracks on the upper side, which ran from column to column, were estimated to be 0.015 in. in width. Around the column head and at places just outside of the drop at the center column, the estimated width of the cracks was 0.015 to 0.02 in. The other cracks noted were smaller. It should be noted that all cracks closed up very well upon removal of the load. Effect of Continued Loading. After the maximum load of 650 lb. per sq. ft. had remained in place 18 hours an increment of 150 lb. per sq. ft. was removed, and the remain- ing load of 500 lb. per sq. ft. was left undisturbed over nearly the whole area for a period of 7 days. At the end of this time a complete set of strain and deflection readings was taken. These measurements showed that the stresses in the steel were about 90 per cent, and the strains in the concrete were about 100 per cent of the corresponding stresses and strains under the load of 650 lb. per sq. in. The measured deflections were about 90 per cent as great as those measured under the maximum load. Since the continuously applied load was only 77 per cent of the maximum load, it follows that the measured strains were from 17 to 30 per cent greater than would be expected from the ratio of the two loads. The load of 500 lb. per sq. ft. was gradually removed from the floor over a period of about two weeks, after which another set of strain and deflection readings was taken. The recovery of strain in the reinforcement was found to be from 50 to 60 per cent, leaving 40 to 50 per cent of the maximum strain remaining in the steel after the test load had been re- moved. The residual strains in the concrete were still greater, being about 70 per cent of those measured at the maximum load. The average recovery in deflection for all observation points was about 50 per cent of the max- imum deflection. The recovery in deflection is shown in Fig. 5. While the proportion of recovery seems rather low for this slab, it must be remembered that the test load remained on the floor longer than is usual in building tests and that the concrete was only about two months old at the time of the test. A considerable amount of plastic deformations is to be expected under these conditions. Measured Stresses and Deformations. The measured stresses and deformations have been used to plot load- stress and load-strain diagrams, a few of which are presented in Fig. 7. These curves are useful principally in showing the relative magnitude of stresses and strains measured at the different increments of load. It is thought best to confine the following discussion mainly to stresses w / / >VXo).00053 | h" | ii 1 ' • g 'ill 'vVvW Jit /’IMMCTCT7 A Vw-ft- AX/ / A / V \ V \A /n / y V 16600 - \s] tt/ v\x , AAA XV > L-U; xN II *,««*« .MW / ,7/7 v X/\ \ \ I | /V^^W' ! ! i' l l I \ §44*?°°-r W«»y/teferoni^&XXvV / / 'v V )l \ ' ' inrnr,m4~ Av~k \ \AX>fx>s/V / fS^.9400 FIG. 8. STRESSES AND DEFORMATIONS ON UPPER SIDE OF SLAB AT MAXIMUM APPLIED LOAD. 6900 J^l / "^-+.+§800 '^_~J75d0'4 [ X id/OQ > K, -^rJ3900 /«^ox Xx x;< - . -oxX HBaar ;* x. X c\Vjsa PI|iI^U$W+W+ Si ^ inr o+oTooo^oX /-V / /V\^ v _X'v y\ \ y v v\ / x' i <^.00050^ n v XXfHX -V^ 15000 m$ +r -±17200 -Rk 14600 /IIJXIS ^ 00034 19100 r ■y.00009 19500 'Y 1/^9000 I8400 Y / X-A #200 X ~ \^.d0006<^r*/5^ v S < x X \I560qXX - X .00007- Symbols x\ / — 1 1 Unit Stress in Steel \ A / lb. per sg. /n. "I y Tension un/ess marJh?d(-\ | 1 1 i \ » Unit Deformation V V \ \ 'C S i°y— in Concrete 27\ V W - Compression unless mar/codfhk A . / /\-i/. 00020 ‘ ' 120600 AUs / / .000^5 - A 7 \ / pooozXp v o§\ \ A §Tv / xT> \ \ IW* FIG. 9. STRESSES AND DEFORMATIONS ON LOWER SIDE OF SLAB AT MAXIMUM APPLIED LOAD. 14 Flat-Slab Floor Test. at the maximum applied load, since at the lower loads there is a possibility of greater relative error and since the stress distribution is masked to a greater degree by the tension in the concrete. For this reason, the maxi- mum measured stresses and deformations at the load of 650 lb. per sq. in. are presented in Figs. 8 and 9 for the upper and lower sides of the slab, respectively. It will be noted that at only a few gage lines did the stress in the reinforcement exceed 25,000 lb. per sq. in., the highest individual stress FIG. 10. COMPRESSIVE UNIT IN SLAB AROUND CENTER COLUMN. being 33,500 lb. per sq. in. The average of the stresses measured across the principal moment sections on both sides of the floor was about 18,000 lb. per sq. in. The compressive strains in the concrete are quite uniformly distributed along the sections used except at places in and near the drop. Fig. 10 shows measured unit deformations in the concrete of the lower side of the slab near the column at the center of the test area. Since neither the stress-strain relation for the concrete nor the modulus of elasticity of the material in the slab is known accurately, no attempt has been made to calculate compressive stresses. The maximum compressive stresses devel- Flat-Slab Floor Test. 15 Mid Section | Colunyi-head5ection\ § § Sit § §saSgg 5&S 3 3 /r m Upper S/de of Slab Mid Section I Column-head SectionX Mid Section I Lower S/de of Slab FIG. 11. STRESSES AND DEFORMATIONS ON JOINT COMMITTEE SECTIONS. 16 Flat-Slab Floor Test. oped appeared to be generally less than one-half the ultimate strength of the concrete; in a few cases they were considerably greater. The greatest individual unit deformation measured was 0.00079. It is seen in Fig. 10 that the deformation in the floor just outside the drop is quite high, as it is also in the diagonal directions on the under side of the drop just next to the column capital. A drop of this deptli evidently causes very abrupt changes in compressive stresses, as well as regions of high stress in the reinforcement above. The stresses and deformations along sections of maximum positive and negative moment are of considerable interest. In Fig. 11, the stresses and deformations along the standard sections designated by the Joint Com- mittee of Concrete and Reinforced Concrete* have been plotted. On the upper side of the slab the highest stresses were observed in the straight reinforcing rods of the mid section (Unit T ). The stresses in the columti head section, particularly those within the drop, were smaller. The unit deformations in the sections of positive moment do not show much variation. On the lower side of the slab the stresses in the reinforcement are fairly uniform for both the inner and outer sections of positive moment, except at the ends near the edge of the loaded area. The unit deformations in the mid section of negative moment are quite uniform, while the de- formations in the column-head section show a characteristic variation at all gage lines near the edges of the drop. It may be noted that the narrowness of the bands of diagonal bars (Unit B, truss bars) enables the reinforcement to act to good advantage, since the bands run nearly at right angles to the sections of maximum moment which they cut, while in the usual four-way system the bars at the edge of the diagonal bands would cross the panel boundaries at points where the intensity of bending moment in the direction of the panel edge differs greatly from that at right angles to it. Stresses Along Rings. It was not found feasible in this test to take observation on con- secutive gage lines along the different rings, but in a number of cases several gage lines were located on a quadrant of a ring. From these observations a good idea of the behavior of the rings in various parts of the slab is obtained. Assuming a uniform variation in stress between observation points the stresses in a number of rings at a load of 650 lb. per sq. ft. are shown diagrammatically in Fig. 12. It is seen that the rings of Unit B show maximum tension at points where they cross the section of maximum positive moment and that the tension is small or changed to compression at points where they cross the diagonal of the panel. The compression is greater in the larger rings. The ♦Final Report of the Joint Committee on Concrete and Reinforced Concrete. July 1, 1910. Flat-Slab Floor Test. 17 LOCATION OF RIN65 Assumed: Uniform variation in stress between gage tines Stresses plotted in radial di- rection. Load= 650 Ib /w 5o.Fr FIG. 12 . STRESSES ALONG KINGS. 18 Flat-Slab Floor Test. inner rings of the unit undoubtedly have more nearly uniform stress throughout their length. The stresses in the rings of Unit A are seen to change from tension to compression in a quadrant of the ring. Obviously, the sections of the ring which cut the outer section of positive moment which is at right angles to the outer section, are in compression. In fact, this mid section was rather highly stressed, as is shown by the compressive deformations in the concrete and the high tensile stresses in the top rods (Unit T) in Fig. 11. Hence it seems probable that with so great a variation in stress in thesfc rings the stress is brought into the bar almost wholly by bond stress. The action of the rings of Unit C and of the adjacent straight bars is more difficult to analyze. Fig. 12 shows two of these rings to be in tension throughout their length, but the tension is greatest at the panel edges and generally less at the panel diagonals. It seems clear that changes in the intensity of stress in a circular ring acted upon by a uni- form internal pressure must be produced by bond stress; if the stress is nearly uniform it would seem to be developed by the lateral pressure or bearing exerted against the inside of the ring. The tension in the rings of Unit G seems to be developed mainly by lateral pressure. The distribution of stresses in cantilever flat slabs similar to the column-head section of the floor tested has been analyzed mathematically by a number of writers. A simple statement of the principles involved may help to explain the relations of the different stresses measured. Because of its great effective depth, the portion of the slab over the column capital does not develop appreciable deformations under an ex- ternal moment. Immediately outside the capital, however, a considerable radial deformation occurs, causing a large radial unit deformation. A circular ring a small distance outside the column capital will have its diameter increased by the increment of radial strain outside the column capital, and its unit deformation will be equal to the radial deformation on the two sides of the capital divided by the diameter of the ring, a comparatively small quantity. It is seen that the radial unit deformation is large at the edge of the capital and decreases with the bending moment at points farther away from the capital. The unit circumferential deforma- tion, or unit deformation in a circular ring, however, is small for a ring near the column capital and increases as the diameter of the ring increases until it reaches its maximum at some distance from the edge of the capital. A comparison of the action of the reinforcement in the column-head section is, therefore, difficult, even if it were not complicated by the dif- ference in effective depth of bars, the difference in the available width of drop in different directions, and the difference in the manner in which the stress is produced in the reinforcing bars. It is evident that the stresses are developed in the bars of the rectangular and diagonal bands by bond stress in the usual way. In the rings in which the stresses are nearly uniform the stress appears to be developed by means of pressure trans- mitted through the concrete and applied as bearing pressure against the Flat-Slab Floor Test. 19 inner side of the ring. This action involves a shortening of the width of a ring of concrete at the top of the slab just inside the reinforcing ring (on the tension side of the slab) and the consequent formation of a crack or cracks of sqme size outside the next smaller ring. The variation of stress in the ring and in the straight bars at and near where they cross the ring was found to be that indicated by theoretical consideration. The greatest stress in a direct or diagonal bar was found over the edge of the column capital, and the stress decreased outwardly, except as influenced by the shape and size of the drop. The smallest two rings of Unit G did not develop their share of stress, and it appears that the rings of this unit are most effective beyond a distance from the capital equal to one or more times the thickness at the drop. Distribution of Stresses and Moments. One of the main objects of the test was to see how effectively the reinforcing steel was distributed throughout the slab and to determine the proportion of the total bending moment developed at the various sections. From the stresses observed the distribution of the steel has been shown to be fairly good; in a few instances a slight rearrangement of the reinforcement would eliminate the extremes of very high and very low stresses. Table II gives average values of stresses in the reinforcement at a load of 650 lb. per sq. ft. as measured on the various sections shown in Fig. 11. Regarding the resisting moments developed by the steel, as has always been found in other tests of reinforced concrete floors at stresses which were considerably below the yield point of the steel, the measured stress over the full-gage length does not account for the full analytical value of the bending moment produced by the load. With a crack present in the concrete across a gage line, it is to be expected that the average unit strain over the gage length will be less than the unit strain over some part of the gage length. The concrete in the earlier stages of loading resists a considerable part of the bending moment. Experience in other tests has shown that as the stress in the reinforcing bars approaches the yield point, the reinforcement gradually takes a greater proportion of the full bending moment and finally assumes its full share. It is evident that the tension in the concrete varies with the percentage of reinforcement, as well as with the quality of the concrete. Some quantitative data on this phenomenon from tests of a variety of structures have been published recently by Prof. Hatt.* From an analysis of the tests of a number of buildings, he found that the total resisting moment of the steel at a *“ Moment Coefficients for Flat Slab Design, with Results of a Test,” by Prof. W. K. Hatt, Proc. A. C. I. 1918. Table II. Stress and Moment Distribution at Load of 648 lb. per sq. ft. (Dead load stresses not included.) Flat-Slab Floor Test. i "§ i 00 O O E- % Flat-Slab Floor Test. 21 measured steel stress of 18,000 lb. per sq. in. was about 40 to 50 per cent of the full theoretical moment of 4_cy 2 wherein c « = diameter of column capital, in i nc it e s, feet lo Z — total load per sq. ft. Jl g — panel length, in inches. To obtain data on the effect of the tension in the concrete, two test beams were poured at the time that the concrete in the test area was poured. However, the percentage of reinforcement used, which was .0071, compares only fairly well for the column-head and outer sections and is FIG. 13. RATIO OF MOMENTS AT VARYING STRESSES IN STEEL. much higher than the percentage for the inner and mid sections of the test slab. Fig. 13 shows the ratio of the resisting moment of the steel to the bending moment at different measured stresses in the two beams. The ratio of the two moments at a measured stress of 18,000 lb. per sq. in. is seen to be about 0.72, but this ratio would be much smaller for a per- centage equal to the average percentage used in the slab. From the average stresses, effective depths and steel areas, the resist- ing moments at the various design sections have been calculated and are given in Table II. A study of these moments shows that only about 36 per cent of the analytical moment is accounted for by the steel stresses 22 Flat-Slab Floor Test. at the load considered. The distribution of the moments, however, agrees fairly well with the distribution commonly used in design. It is found that the negative moment is 60.7 per cent, and the positive moment is 39.3 per cent of the total amount, and that the further subdivision of these moments among the different sections also agrees very well with the usual assumptions regarding moment distribution in flat slab floors. General Comments. The measured stresses at points other than the standard design sec- tions did not appear to be of particular importance, except that high stresses were found on both the tension and compression sides near the edge of the dropped panels. The calculation of moments at intermediate sections was difficult since the position of the point of inflection was not fully known. From the data available it appears that the distance from the panel edge to the line of inflection was about three-tenths of the panel length. The slab was not loaded heavily enough to develop much evidence of the action of shear and diagonal tension. It has been questioned whether a reinforcing system of rings provides properly against shearing failure which might occur at a circular section near one of the rings around the column head at some distance outside the column capital. It was found in the test that tracks followed such a section along rings of Unit C. Whilecracks of this soi;t are also found in slabs having two-way or four- way reinforcement, with the ring reinforcement the cracks are likely to be wider and to be concentrated upon a definite path. It is felt that the building withstood the test very well. The stresses are, if anything, lower than might have been expected at the applied load, and are, with a few exceptions, fairly uniform. It is hoped that the results of this test will add considerably to the knowledge of the behavior of this system of flat-slab reinforcement. The writers wish to acknowledge the hearty and efficient co-operation of It. F. Wilson & Co., who had charge of applying the test load; of the Chicago Building Department, and of the Structural Research Laboratory at Lewis Institute. Acknowledgment is also made t