Você está na página 1de 30

Title No.

Research report sponsored by
the Reinforced Concrete
Research Council

Investigation of Multiple-Panel
Reinforced Concrete Floor Slabs
Design M1ethods-Their
Evolution and Comparison

This is the first of a number of reports to be published on an investigation

of multiple-panel concrete fl~or slabs. !he paper serves as an introducf!o:1
to the later reports present1ng the h1stoncal background of the desig:1
methods for floor slabs in ACI 318-56 and the over-all results of tests on five
structures tested during the investigation.
Key words: flat slab; history; reinforced concrete; slab, multiple-pa:1ci;
test; two-way slab; ultimate load.

• REINFORCED CONCRETE FLOOR SLABS spanning in more than one direction

are classified into two different categories by most building codes: the
flat slab; and the two-way system with supports on all sides, usually
called the two-way slab.
An example of a flat slab is shown in Fig. 1. The slab is supported
directly on the column which may have capitals, and the slab may have
drop panels. In designs for relatively light loads, the capital can be
hidden in the columns and the drop panels in the slab by increasing the
cross-sectional dimensions of the column to include the capital and
making the slab thicker to include the drop panels. The resulting struc-
ture is called a flat plate (Fig. 2).
An example of a two-way slab is shown in Fig. 3. The slab may rest
on reinforced concrete or steel beams or on walls.
Often, the two types are combined in a structure. Flat slabs may
have beams. Two-way slabs may lack beams in certain spans. Never-
theless, the structure is designed for significantly different factors of
safety depending on what the designer decides to call it, a flat slab or
a two-way slab.
The purpose of this paper is (1) to describe the developments that led
to the drastically different methods of design for flat slabs and two-way
slabs. (2) to compare the basic features of the current methods of de-


Fig. 1-The flat slab

Fig. 2-The flat plate

Fi:~. 3-The two-way slab


ACI member Mete A. Sozen is associate professor of civil engineering, University

of Illinois, Urbana, Ill. Dr. Sozen's work involves research in reinforced and pre-
stressed concrete. A previous jOURNAL author he is currently secretary of ACI-
ASCE Committee 421 (321 ), Reinforced Concrete Slabs, and Committee 441 (34 1),
Reinforced Concrete Columns, and a member of ACI-ASCE Committee 428 (328),
Limit Design, and ACI Committee 435 (335), Deflection of Concrete Building
Chester P. Siess, professor of civil engineering, University of Illinois, Urbana, Ill.,
has been active in ACI affairs since 1938. A noted researcher and author, Dr. Siess
has won many honors including the ACI Wason Medal for Most Meritorious Paper
(1949), which he coauthored with N. M. Newmark. He is now serving his third
year as an ACI Director. Active in committee work Dr. Siess is chairman of ACI
Committee 115, Research, chairman of ACI-ASCE Committee 421 (321 ), Reinforced
Concrete Slabs, and a member of Committee 318, Standard Bui Iding Code, and ACI-
ASCE Committee 423 (323), Prestressed Concrete. He is also an ACI member of the
ACI-CEB Collaboration Committee.

sign for flat slabs and two-way slabs in the United States, and (3) to
provide an over-all description of an extensive investigation of multiple
panel reinforced concrete floor slabs at the University of Illinois that
included both the flat and the two-way slabs.
The discussion of the design methods is limited to the rules for the
amount and distribution of the bending moments.
C. A. P. Turner
Centuries of construction with stone and timber preceded reinforced
concrete. Consequently, just as the first motor cars were built to look
like horse-drawn carriages, the first reinforced concrete systems were
conceived in the image of traditional types. In a timber structure, the
planks carried the load to the joists, the joists to the girders, and the
girders to the columns; so must they in a reinforced concrete structure.
Hence, the flat slab had to be invented rather than developed as one of
the obvious applications of reinforced concrete.
Because it escaped the imagination of minds trained in the two dimen-
sions of construction with timber and iron, the flat slab was given the
treatment of a miracle: while it was endorsed blindly by some engineers,
it was resisted savagely by others. Between 1906, when C. A. P. Turner*
built the first flat slab,t and 1921, when Westergaard and Slater1 pub-
lished their comprehensive paper on slabs, the flat slab was a subject
of controversy and therefore of intense interest among practicing engi-
neers, college professors, and lawyers. The bone of contention is illus-
trated dramatically by a comparison made by A. B. McMillan 2 • 3 in 1910
which is shown in Fig. 4. The bars indicate the amount of reinforcement
required by various design procedures in a 20 x 20 ft interior panel of
an 8 in. thick flat slab intended to carry 200 lb per sq ft. Evidently, the
material bill for steel could vary by 400 percent depending on the design
method chosen. There was room for argument.
tThe C. A. Bovey Building In Minneapolis, Minn.•

The arguments about theory did not faze the construction industry.
C. A. P. Turner had found entrenched resistance against his invention
in 1906. Athough the C. A. Bovey Building, built at the risk of the
inventor, had performed satisfactorily in a load test, Turner was still
being asked to put up bond for new work 2 years and a dozen buildings
later. 4 But a few years later, the situation changed. By 1913, over 1000
flat slab buildings had been built around the world. During this period,
Turner's patents were revoked by the Eighth Circuit Court of Appeals
in favor of J. L. Drum who had purchased the 0. W. Norcross patent
for a flat arch which resembled the Turner flat slab and preceded it.
This lawsuit evoked a Dickensian hatred of patent lawyers in Turner
who sustained a relentless campaign through books, articles, and dis-
cussions to seek a recognition of his contribution which he was never
to get. In 1914, Edward Godfrey, a disapproving critic of the flat slab
wrote, "The flat slab has repeatedly been brought before the engineer-
ing profession for consideration and adoption." 5 Turner responded bit-
terly,6 "As sponsor for the original successful flat slab construction, the
writer may say, in answer to Mr. Godfrey, that he has never knowingly

lLJ 10

La.. z z ffi
a:: 0 <(
0 a:: ::z: >
lLJ ::z: ....J
d 1- LIJ 0 ~
z en a:: en i=
a:: <(
::E ~ :J z z
:J (,) a:: <( LIJ <(
t- c.!) ::E III :E ::E 0
Fig. 4-Weight of steel required in the interior panel of a flat slab by vaTious
design methods in 1910

submitted it for the adoption of or appropriation by the engineering

profession at large."
The appropriation of the flat slab by the engineering profession had
its problems created mainly by the comparison shown in Fig. 4. All
design methods could not be correct if the variation in results was
400 percent.
The 1912 progress report of the Joint Committee on Concrete andRe-
inforced Concrete7 had only one paragraph on the flat slab: "The con-
tinuous flat slab with multiple-way reinforcement is a type of con-
struction used quite extensively, and has recognized advantages for
special conditions, as in the case of warehouses with large, open floor
space. At present, a considerable difference of opinion exists among
engineers as to the form of constants which should be used but ex-
perience and tests are accumulating data which it is hoped will in the
near future permit the formulation of the principles of design for this
form of construction."
A. R. Lord
The early proof tests had satisfied the building comm1sswners but
they were not adequate to form a basis for design recommendations.
The severe limitations of the ordinary load test was perceived by most
engineers. The fact that a few panels withstood the test load with no
apparent sign of distress could attest only to the fact that the same
panels might be able to carry the same load if loaded again. The un-
assailable observation that the flat slab designed by almost any method
stood up could not be ignored; nor was it, as evidenced by the sudden
boom in flat slab construction. But this fact alone favored no particular
design method.
In 1910, A. R. Lord made testing history when he reported the strain
readings he had made during the load test of a flat slab floor of the
Deere and Webber Building in Minneapolis, Minnesota. 8 With the help
of such measurements, an insight could be obtained into the response of
the flat slab. Unfortunately, interpretations of the data from this test
and a series of others did little toward resolving the mystery of flat slabs.
For the structural engineer, plate action was an entirely new concept.
The "crossing beam analogy," thinking of the slab as two perpendicular
beams each carrying a certain proportion of the load in relation to their
stiffnesses, helped only to foster the still existing illusion that only part
of the load need be carried in a given direction. Grashof's work 9 had al-
ready been used by the mechanical engineers in boiler plate problems.
However, this work was represented in American engineering literature
either simply as formulas without any derivations 10 or as a basis for
arriving at questionable conclusions. 11

Turner was a prolific writer.* Nevertheless, it is difficult to make

definite statements about his interpretation of the response of a flat
slab. In 1909,13 he wrote, "Such a slab will act at first somewhat like a
flat dome and slab combined,_ but as the deflection gradually increases
it will gradually commence to act as a suspension system in which the
concrete will merely hold the rods together and distribute the load over
them. For example, a slab 24 ft square and 6 in. thick can readily be
loaded so that the deflection would be 8 or 10 in. It would crack clear
through the concrete and would still carry the load to ultimately break
the slab down." However, later he resorted to a more conventional
explanation, the "mathematical flat plate," which was "as any mathema-
tician would readily see," a thin plate. 4 His assuming this approach may
have been caused by the efforts ,of the apologists who used Poisson's
ratio in plate theory as a vehicle to justify Turner's design or by his
desire to differentiate his invention clearly from the Norcross patent
which was "as a mason would understand it" a slab thick enough to act
as an arch:l
The admission that the response of the flat slab was similar to that
of a medium-thick plate together with the observations indicating very
small strains in load tests resulted in conflicting conclusions that are
yet to be resolved.
J. R. Nichols
While the majority of the engineering profession classified the flat
slab as being "beyond the range of pure analysis," J. R. Nichols, a young
engineer from Boston, Massachusetts, cut the Gordian knot with a sim-
ple, straightforward analysis.H
Consider one rectangular panel of a plate having an infinite number
of identical uniformly loaded panels. The plate rests on pin supports.
The equilibrium conditions for one-half of the panel are shown in Fig. 5.
Because of symmetry, shears and twisting moments do not exist along
the edges of the portion considered. Thus,
M.t + MB = Mo = -WL
8 - .............................................. (1)

Eq. 1 does not indicate the relative magnitudes of the positive and
negative moments but it does stipulate that the total moment resistance
provided at the two sections should add to WL/8. For the interior panel
of a flat slab with a round capital and with the assumption that the
reaction is distributed uniformly around the capital (ignoring the twist-
*His interests were broad by any definition and he never made a secret of his strong
opinions. He prefaced a book published in 19341.2 with comments on the quirks of the English
language ("Wormwood is unrelated alike to wood and worms."), devoted half a chapter to
disproving Einstein, and welcomed the F. D. Roosevelt Administration's New Deal. The book
was intended for reinforced concrete design.

Column Center Line

Fig. 5 - Free-body diagram ,/
of one-half of a typical in-
terior panel
Note: Total load on one-half
panel= W/2

ing moments between the slab and the capital) Eq. (1) becomes

Mo = ~L [ 1- : ~ + +( ~ ) 3} _(2)

In the closure to his paper, Nichols suggested that Eq. (2) could be
approximated by

Mo = WBL ( 1 - 2£_)
3L 2 .. ....... (3)

While Niehols' paper took 10 pages to present, the discuss[ons occupied

54 pages. Turner thought the paper "to involve the most unique com-
bination of multifarious absurdities imaginable from either a logical,
practical, or theoretical standpoint." H. T, Eddy stated that the funda-
mental erroneous assumption of the paper appeared in the statement
that statics imposed certain lower limits to the applied forces to be
resisted by the reinforcement. A. W. Buel was "unable to find a single
fact in the paper, nor even an explanation of facts," and considered it to
be "a paper of explanations, not only without facts, but contradicted by
facts." Newtonian mechanics stood very much condemned.
The observed facts were clearly on Turner's side. He used considerably
less steel than would be required by Nichols' analysis, yet his structures
performed satisfactorily. Furthermore, the strain readings made by
Lord and others indicated very small stresses under both working loads
and test loads in flat slabs. When the measured steel strains were con-
verted to bending moments through the use of the straight-line formula
M = A,f,jd = A,E,E,jd . ... ( 4)

the sum of the moments across the negative and positive moment sec-
tions (MA and M 11 in Fig. 5) was only a fraction of M 0 as defined by
Eq. (3) Y In his discussion of a paper by H. T. Eddy, G. S. Binckley 10
conceded "The crushing weight of practical experience and empirical
data under which C. A. P. Turner flattened out Mr. Nichols' purely
theoretical paper, tends to induce caution in others."

C. A. P. Turner need never have fought his battle with the conditions
of equilibrium. Nichols' analysis was right but Turner's design was not
wrong. The problem arose out of comparing incomparables. Turner
realized fully that he could never know exactly what the load on the
structure would be. Hence, W in Eq. (3) was only a guess and there-
fore the whole right-hand side of Eq. (3) was approximate. To the de-
signer and evidently to the structure, it did not matter whether W or
the 1/8 factor was modified. Turner insisted on high-strength steel. Al-
though he might be espousing low steel areas, the force provided was
often adequate to satisfy Eq. (3). Furthermore, Turner was well aware
of the effects of forces in the plane of the slab.
The First Joint Committee adopted the form but not the intent of
Eq. (3). In the final report published in 1916, 17 the design static moment
was given as 85 percent of that based on Nichols' analysis, or

M •• =0.107WL ('1- ;~) 2 .................... (5)

The 1917 proposed revision 18 and the 1920 approved version19 of the
ACI Building Code contained an even more flagrant violation of statics
M •• = 0.09 wL2 (L - qc) 2... ............................... (6)
where qc was twice the distance from center line of the capital to the
center of gravity of the periphery of the half capital. The ratio q was
taken as 2/3 for round capitals and 3/4 for square capitals. A distinction
was made between round and square capitals, although not in exactly
the manner implied by Nichols' analysis.
H. M. Westergaard and W. A. Slater
With the publication of the First Joint Committee Report in 1916 and
the ACI Building Code in 1920, design procedures had become more or
less stabilized. However, the flat slab was still a mystery to a good cross
section of the engineering professron. There were too many conflicting
reconciliations of the load test data and diverse "theories."
In 1921, Westergaard and Slater pubHsihed their monumental work on
the analysis and design of slabs.! This paper included a sound exposition
of the theory of plates, ingenious projections of the available theoretical
solutions to solve practical problems, and •a comprehensive study of the
implications of the then available test's on flat slabs ,and rtwo-way slabs.
One of the accomplishments of the paper was to show conclusively
that the moments calculated from measured strains in flat slab tests
were not as small as they were interpreted to be. Previously, the moment
had been calculated from the measured steel strain using the straight-
line formula
M :::: A,E,E,jd . . ......... ( 4)

Fig. 6 - Relationship be- z
tween bending moment and :::iE
steel strain :::iE



which gives a linear relationship between M and E8 for a given section

as shown by Line A in Fig. 6. The actual moment-strain curve in a load
test is given by Curve B in the same figure. The divergence of the
initial portions of the two curves is due to the tensile strength of the
concrete. After cracking, Curve B approaches Curve A which is based
on the assumption of no tension in the concrete.
The misinterpretations of the test data resulted from the following
series of events: (1) as a result of the applied load a moment M 1 had
been generated at a section (Fig. 6), (2) this moment caused the steel
strain E, 1 which was measured, (3) the strain was converted to moment
using Curve A [Eq. (4) ], and (4) the "measured" moment Mm, turned
out to be a fraction of the actual moment. If this had happened for a
beam test, the fallacy would have been obvious, but a lot could be
hidden in the unfamiliar concepts of the third dimension.
Objections had been made to the interpretations of test data on this
basis almost as soon as Lord's test had indicated extremely small stresses
in flat slabs, but no documented effort to reinterpret the data had been
made. Slater made a painstaking study of the true moment-strain rela-
tionships in reinforced concrete beams and used this study to re-evaluate
data from 14 flat slabs, three of which had been tested to failure. He
found that "On the average, the agreement of (his re-evaluated moment
coefficients from test data) for the higher computed stresses with
[Nichols'] analysis is sufficiently close to warrant the belief that if all
the sources of error in the measurement of deformations and in the
interpretation of test results could be removed, the analysis and the
tests would be in substantial agreement." Nevertheless, Nichols' analysis
was not completely endorsed. "There are some indications. however,
that there was greater strength in the flat slabs than appears from the
conclusion that the test results and the analysis of moments are in fair
Although the truth of Eq. (3) was conceded, it was not granted that
the ratio of the ultimate to the design load would be equal to the ratio

of the yield to the design stress. Slater estimated the average factor of
safety of the structures he studied to be 3.23 if the design was based on
Eq. (5) and 2.72 if based on Eq. (6). Even in the latter case, the factor
of safety was "at least as great as that which can be counted upon in
the most elementary flexural unit ... " and therefore satisfactory.
One significant factor was ignored in estimating the factor of safety:
the effect of adjacent unloaded panels. This is brought out clearly in
the ultimate load for one of the slabs that was tested to failure: the
safety factor for Slab J tested at Purdue University was based on a
capacity of 872 lb per sq ft which was obtained when only one of the
four panels was fully loaded.

Design by "empirical method"

The Westergaard-Slater treatise on slabs consummated the compromise
between analysis and design.
The Second Jo·int Committee-2° adopted a slightly modified version of
Eq. (6) by which the rectangular capital was not recognized

M," = 0.09 WL ( 1- 32c)

L 2 .. ............... (7)

with a footnote in fine print, "The sum of the negative and positive
moments provided for by this equation is about 72 percent of the
moment found by rigid analysis based upon the principles of mechanics.
Extensive tests and experience with existing structures have shown that
the requirements here stated will give adequate strength."
The 1928 ACI Building Code~ 1 provided the comfort of Eq. (7) without
the discomfort of the footnote. The end had justified the means and the
hybrid Eq. (7) graced many a building code in stark defiance of statics.
The 1956 ACI Building Code~ 2 introduced a new factor F based on the
fear that the old tests on flat slabs which had c/L ratios on the order
of 0.2 might not apply to the modern flat plate which had rather low
c/L ratios. Thus, Eq. (7) was changed to

M,. = 0.09 WLF ( 1- 2c )

gL 2 ... ........ (8)

F = 1.15 - c/L but not less than 1 (9)

Design by "elastic analysis"

Design by elastic analysis, which made Hs first appearance in the
1941 ACI Building Code, 23 is in essence a two-dimensional approxima-
tion to the plate problem. However, it is modified to yield answers
comparable to those of the "empirical method."
Westergaard had used the concepts of frame analysis, where appli-
cable, to study critical loadings in flat slabs, 1 and the effect of column

stiffness on the distribution of moments had been analyzed with the

use of approximate methods in Europe. 24 • 25 • 26 However, the ACI frame
analysis finds its beginnings in a paper by H. D. Dewell and H. B.
Hammill published in 1938.27 The paper was based on a report made in
1929 to the technical committees of the Uniform Building Code, a Cali-
fornia edition.
The need for a "rational" method was felt primarily because of the
anticipated effects of pattern loadings on the slab and column bending
moments. The limitations of the empirical method as to panel sizes and
combinations represented another incentive. The main features of the
proposed method were quite similar to those of the current method in
ACI 318-56. The bent was one panel wide. The column-slab joints were
assumed to be rigid. Dewell and Hammill assumed in their calculations
that the columns were hinged at the middle of the distance from the
bottom of the capital to the top of the floor below.
Since the conditions of equilibrium were automatically recognized in
the frame analysis, the resulting moments were 100 percent rather than
72 percent of the static moment. To eliminate this discrepancy, the
negative moments were reduced by 40 percent. The method was given
in the 1933 Uniform Building Code, California edition as an alternate
method for the design of flat slabs.
The ACI method was developed under the direction of R. L. Bertin.
The desideratum was a frame analysis method which would give results
comparable to those obtained by the empirical method of design. This
was effected by stipulating the design negative moment to be that at
a distance xL, from the center of the column width, where

X = 0.073 + 0.57 LA' .................. . (10)

where A' was half the capital width but not greater than L/8.
For a uniformly loaded panel with equal end restraints, the sum of the
positive and negative moments can be made equal to the moment given by
Eq. (7) by taking the negative moment at a distance xL from the column
center line such that

0.09 WL ( 1 -
0.125 WL
+· ~) 2 ( 2L -xL )

For this condition

X = 0.076 + 0.565 LA' . ... . (11)

Eq. (10) was based on "studies covering variations in ratio of adjacent

spans, column to slab stiffnesses, and live to total load."*
*Di StasiobJoseph, and van Buren, M. P., "Background of Chapter 10, 1956 ACI Regulations
for Flat Sla s," unpublished report prepared for ACI Committee 318.

On the other hand, the total design moment was required to be

M •• ::=, 110 W., L ( 1 - 4:L·) 2 ............................... (12)

where Wav was the average of the total load on two adjacent spans and
Aav' was the average of the values of A' at the ends of the span con-
sidered. The procedure is illustrated in a paper by Peabody. 28
The ACI frame analysis was modified further in ACI 318-56. Eq. (10)
was dropped in favor of a "physical" definition of the critical section
for negative moment.* This definition showed substantial agreement
with Eq. (10) for c/L = 0.225 but resulted in more conservative negative
moments for small values of c/L.t Eq. (11) was also eliminated. Fur-
thermore, it was permitted to reduce the design moments "in such pro-
portion that the numerical sum of the positive and average negative
bending moments used in design procedure need not exceed M0 " as given
by Eq. (8) ,22 if the structure analyzed was within the range of applica-
tion of the empirical method.


Crossing beams
Because it looked, at least in shape, like the traditional types of floor
systems, the two-way slab in reinforced concrete was assigned no mirac-
ulous powers.t However, two-way ac'ti'on of the slab was approached
cautiously. The 1912 Progress Report of the Joint Committee7 sug-
gested that square spans "may well be" reinforced in both directions.
A footnote gave the proportion of load to be considered in each direction,
provided the aspect ratio of the panel was less than 1.5:
"The exact distribution of load on square and rectangular slabs, sup-
ported on four sides and reinforced in both directions, cannot readily be
determined. The following method of calculation is recognized as faulty,
but it is offered as a tentative method which will give results on the safe
side. The distribution of load is first determined by the formula:
L' (13)
r = L' + b'
r proportion of load carried by the transverse reinforcement (short span)
L = length of slab (long span)
b = breadth of slab (short span)
*Distance in the direction of span from center of support to the intersection of the center-
line ot the slab thickness with the extreme 45 deg diagonal line lying wholly within the
concrete section of slab and column or other support, including drop panel, capital and bracket.
tSee footnote on previous page.
tThis is evidenced by -the fact that the design methods for two-way slabs are given rou-
tinely In the chapter on "flexural computations" in ACI 318-56 while flat slabs are accorded
a special chapter.

The total amount of reinforcement required at a section could be

reduced by 25 percent, by gradually increasing the bar spacing from
the third-point to the edge of the slab. The design of the beams was to
be based on a distribution of load varying "in accordance with the ordi-
nates of a parabola having its vertex at the middle of the span." In the
1908 Progress Report, 29 the load on the beam varied "in accordance with
the ordinates of a triangle."
In the 1916 final report of the First Joint Committee/ 7 Eq. (13) based
on the crossing beam analogy was replaced by
r =b - 0.5
a linear equation approximating Eq. (13). No reduction was permitted
in the moment at a section. Instead, it was recommended that two-
thirds of the moment be assigned to the center half of the slab and
one-third to the outer quarters.
Although it appeared that the slab was being designed for only a
portion of the load, the complete system including the beams was de-
signed to carry the full load, as it should be. Furthermore, both the
beam and the slab moments included the effects of pattern loadings.
Thus, while the flat slab interior panel in the 1916 Joint Committee re-
port was being designed for 85 percent of the static moment, the two-
way slab panel had to be designed for more than the static moment.
Nevertheless, the comparison was not as inequitable as it was to become
The publication in 1921 of the paper "Moments and Stresses in Slabs" 1
showed quite clearly the fallacy of the crossing beam analogy. The
1924 report of the Second Joint Committee was silent on two-way slabs. 20
The resulting vacuum was filled by the adoption of Eq. (14) in the 1928
ACI Building Code 21 although it was ignored in the 1927 draft of that
Cod€. 30 In the 1928 Code, it was assumed that the load on the beam was
uniform and had the magnitude indicated by the pertinent application
of Eq. (14).
J. DiStasio and M. P. van Buren
The changing attitude of the engineering profession in the United States
toward two-way slabs is reflected in two consecutive editions of the book
Concrete, Plain and Reinforced. The 1917 edition 31 used Eq. (14) at
face value. The 1925 edition 32 warned that the load distribution could
be affected also by the end restraints. This approach culminated in
Method 1 of the ACI Building Code. Method 1 for the design of two-
way slabs was developed for the New York City Building Code and
proposed to the ACI in 1935. It was incorporated in the 1936 ACI Build-

ing Code33 and its derivation was explained in a paper by DiStasio and
van Buren.34
Method 1 is kindred to the "elastic analysis" for flat slabs in that it
is a two-dimensional approximation to the three-dimensional problem of
plates on rigid supports. However, while the "elastic analysis" was
modified for purposes of design, Method 1 was not.
Method 1 was based on two simplifications: an assumption about the
division of load to the two slab spans and an assumption about the dis-
tribution of load along a given span. The division of the load was made
in the same vein as Eq. (13) except that the cube rather than the fourth
power was used. To account for the end restraints, the span between the
points of contraflexure, determined from a frame analysis, was used as
the effective span. Naturally, the division of load to the two spans
would not be sufficient to match a plate analysis. Hence, another modi-
fication factor, ascribed to the nonuniform distribution of the load along
the span, was used to make the result conform to available plate analyses.
The 1936 version of Method 1 provided formulas for the determination
of points of contraflexure under various conditions of end restraint.
Later studies35 showed that this was unnecessary and the formulas were
replaced by a set of constants in the 1947 ACI Building Code which also
contained revised symbols for Method 1.3 6
It should be pointed out that although the divishm of the load to the
two spans in Method 1 may give the impression o1 not providing rein-
forcement for all the load in a given direction, thi~ is not so. What the
slab does not carry, the beams do. Furthermore, the moment coefficients
recognize pattern loadings.
H. M. Westergaard
The 1921 paper by Westergaard and Slater1 had served to show the
shortcomings of Eq. (13), but had not been found practical enough for
general use. In 1926, Westergaard published a paper proposing a method
of design for two-way slabs. 37 This paper contained moment coefficients
for slabs and supporting beams. The coefficients were based on analyses
of continuous plates on rigid supports providing no torsional restraint.
However, they were modified in view of the fact that exceeding the
permissible stress locally was not critical and that flat slabs were de-
signed satisfactorily for only 72 percent of the static moment. Only the
slab moment coefficients were reduced by 28 percent, and then, with
respect to the maximum not the static moment.
Westergaard's moment coefficients were not adopted directly in a
design method since they were given for a series of equal adjacent
panels, a condition rarely met in practice. Hence, the Joint Committee 38
modified and extended Westergaard's procedure to give moment coeffi-
cients for isolated panels with various edge conditions. In preparing this

method, it was assumed that the slab was built monolithically with the
supporting beams. Thus, moments were prescribed at the "discontinuous"
edges and if an unbalanced moment occurred at adjacent negative
moment sections, one-third of the difference was assumed to be resisted
by the beam in torsion. Furthermore, rather than listing a set of moment
coefficients for the beams, it was deemed simpler to specify the load
on each beam. The load carried by each beam was assumed to be the
load in that portion of the panel bounded by 45-deg lines emanating
from each end of the beam and the panel centerline, almost a return to
the 1908 Progress Report of the Joint Committee.
The 1940 Joint Committee method for two-way slabs was adopted as
Method 2 in the 1947 ACI Building Code. 36 Unfortunately, the stipula-
tion for slabs monolithic with their supports was left out in Method 2
leaving it vulnerable to misinterpretation.


In view of the genesis of the design methods for flat slabs and two-way
slabs, it was inevitable that they would be inconsistent with each other.
The design methods for flat slabs evolved out of practice. The method
was made to suit the slab. The design methods for two-way slabs were
derived from theoretical considerations. The slab is built to suit the
In the following paragraphs, a comparison is made in terms of design
bending moments of the results obtained by the "empirical method" for
flat slabs and by Methods 1 and 2 given in ACI 318-56 22 for two-way slabs.
Since the object is a direct comparison, it is assumed for the sake of
simplicity that the columns and beams have no width. The comparisons
are made in terms of the total design moment, defined as the absolute
sum of the positive and the average negative design moments for the
slabs and beams, if any. The total design moment can be related directly
to the factor of safety for a typical interior span if the panel fails in
flexure as a complete unit.
The total design moments specified by ACI 318-56 for various panels
of flat slabs and two-way slabs are plotted in Fig. 7 through 10. The
ordinates represent coefficients of the total design moment in terms of
WL and the abscissas represent the ratios of the sides of the rectangular
The total design moment coefficients for the two-way slabs are differ-
ent for the short and the long spans. In determining the moments for
the two-way slab, the specified distribution of slab moment across a
section was recognized. ACI moment coefficients (ACI 318-56, Chapter
7) were used for the beam moments in Method 2 as well as in Method 1.

For the flat slab, the factor Fin Eq. (8) was assumed to be 1. The fact
that the quantities given are for c/L = 0 is irrelevant since only relative
values are desired. There are slight differences between the total design
moment coefficients for slabs with and without drop panels, and for spans
perpendicular to edge beams of various stiffnesses. However, these dif-
ferences are too small to show in Fig. 7 through 10.
A comparison of the curves for two-way slabs shows that, especially
for side ratios greater than 0.6, the two methods compare quite well.
The largest deviation for side ratios greater than 0.6 is about 5 percent
from the mean. Although it would be desirable to have the agreement
in all panels as good as that for the interior panel, Methods 1 and 2 are
as consistent as they could be expected to be in view of their different
For the interior panel (see Fig. 7), the total design moments for the
two-way slab and the flat slab are beyond comparison. The two-way
slab is designed for a moment 67 percent greater than that for the flat
slab. IfF = 1.15, the difference would be reduced to 45 percent, still a
respectable figure. The quantitative relationship between the design

Two -Wav s ab
f'<- Me hod 2 Met ~Od 1\
t-- \
0 0.12
Flo Sl ~b-1\
-Short Span
--Long Span
0.5 0.6 0.7 0.8 0.9 1.0
Ratio of Sides
Fig. 7-Comparison of total design moments interior panel


Two ~wa " Sl ob


\_ Me •hoc

Mei hod
-- ,_,. - L

0 0.12
==c ...
.Ill2'~ _/
G) Flo Sl ab-
Q 0.08

-Short Span
--Long Span
0.5 0.6 0.7 0.8 0.9 1.0
Ratio of Sides
Fig. 8-Comparison of total design moments edge panel; span perpendicular to

moments for the two-way slab and the flat slab is about the same for
the span in the edge panel perpendicular to the edge as shown in Fig. 8.
The design moment for the flat slab approaches that for the two-way
slab if a spandrel beam extends along the span considered, as shown in
·Fig. 9 and 10; the deeper the beam the larger is the total design moment.
This results from the fact that while the slab is designed for almost the
same moment, the beam is assumed to carry a certain portion of the load
and the beam moment coefficients reflect pattern loadings.
The comparisons made in Fig. 7 through 10 are idealized. They do not
refer to slabs with practical proportions. Nevertheless, they do point out
the basic differences between the design methods.
The two-way slab is just as much a slab as the flat slab. This can be
considered to be self evident. Furthermore, the two-way slab possesses
some structural advantages over the flat slab. It can be reinforced more
effectively to resist torsion and shear. Yet, the two-way slab is designed
for higher moments than the flat slab. The designer is not even per-
mitted to vary the proportion of load assigned to the beams and the slab

Two r-Wa s ob
'-.. c ethc d ~ IV ethc d I 1\
0.16 \
~= =~
E ceep Bee m----'
0 0.12
c .... Flo1 Sla p< s hall pw 3ea nJ

0.08 lN lo 8 eorr 1_/

-Short Span
--Long Span
0.5 0.6 0.7 0.8 0.9 1.0
Ratio of Sides
Fig. 9-Comparison of total design moments edge panel; span parallel to edge

as a function of the beam stiffness,* a privilege given to the flat slab

The handicap imposed on the two-way slab is not the only major incon-
sistency of the current design procedures. The flat slab could be de-
signed more consistently. It does not seem reasonable to increase the
total design moment with the beam stiffness.


Introductory remarks
In a penetrating study of tests on flat slabs published in 1916, Talbot
and Slater concluded, "Progress in obtaining experimental knowledge of
flat slab structures may best be made through a series of tests on struc-
tures designed solely for test purposes and planned systematically to
bring out the fundamental differences between different types of design
and the effect of varying certain elements of design." 15 Thirty-five years.
*The design procedures refer to nondeflecting supports.

later, in 1951, ACI-ASCE Committee 421 (321), Reinforced Concrete

Slabs, came to a similar conclusion and recommended to the Reinforced
Concrete Research Council that a research program be initiated to study
under controlled conditions and compare the strength and behavior of
various types of multiple-panel floor slabs, specifically the flat slab and
the two-way slab. A task committee of the RCRC under the chairman-
ship of Leo H. Corning developed the broad outlines of an investigation
which was initiated in September 1956 in the Structural Research Lab-
oratory of the University of Illinois Civil Engineering Department.
The following sections describe briefly the five test structures and com-
pare their measured strengths. Detailed information about the tests and
their interpretations will be published in subsequent papers.
The test structures
The tests had to be planned to satisfy two objectives which were almost
mutually exclusive. If their results were to serve directly in the re-
evaluation of knowledge accumulated from previous load tests, the test
specimens had to be "typical" structures. Opposed to this was the desire

Two Wa, Sic b

. . . . ..r Met hod 2 Me1 hod I t

.. 0.16
== - ..__- -- --
-- - .L

E r-(J eep Be ~m-
0 0.12
:t Flo Sic b s hall pw I ~ear n_/

... (\ o Beon I ___)"
Q 0.08

-Short Span

--Long Span
0.5 0.6 0.7 0.8 0.9 1.0
Ratio of Sides
Fig. I0-Comparison of ·total design moments corner panel


I .
I. I _,,
. I
~ ' I

. I -~
I -~ . 0
. . I

~ '

I '""
~ I

Fig. I !---Structure Fl
Slab thickness: 1.75 in. Interior column section: 6 x 6 in.
Deep beam: 2 x 5.25 in. Edge column section: 4 x 6 in.
Shallow beam: 4.5 x 2.5 in. Corner column section: 4 x 4 in.


T2 I
Two-way I F3
Modified Fl F2
I Structural
Flat slab
slab design Flat plate Flat slab 1 fabric
Design live load, lb per sq ft 70 70 70 200 200
Design dead load, lb per sq ft 75 75 85 85 85
Total slab design load,
lb per sq ft 145 145 155 285 285
Wall load on beams, lb per ft 400 None 600 600 600
Basis for design ACI 318-56 Limit ACI 318-56 ACI 318-56 ACI 318-56
Method 1 Sec. 1004 Sec. 1004 Sec. 1004

to develop basic information to serve in the preparation of advanced de-

sign methods. This required the test specimens to he as srimple as possible.
The necessity for elaborate instrumentation and control dictated that
the tests be conducted within a laboratory, which put limitations on the
size of the test structures. To fulfill the minimum requirements of being
"typical," the test structure was to include three panels in each direction.
Thus, it was decided to have each test structure comprise nine 5 ft square
panels, arranged three by three. The experimental advantages afforded
by symmetry outweighed any desire for an irregular layout of panels.
The size of the panels made design in strict accordance with the re-
quirements of the ACI Building Code impossible because of the Code
limitations on absolute dimensions such as minimum thickness. Conse-
quently, the designs were made for structures four times as large and
then scaled down. Since the test structures were built of reinforced
concrete, the scaling procedure involved simply the multiplication by %
of every dimension in the design drawings. In view of the questions that
might be raised about the size of the test structures, the test of a larger
structure (15 ft square panels) was carried out by the Portland Cement
Association Research and Development Laboratories in Skokie, Illinois.
The five test structures are listed in Table 1. Plan and profile views
of the structures are shown in Fig. 11 through 13. The supporting
columns were pin-ended and their lengths varied for the different struc-
tures to make the stiffness of the interior columns comparable to the
combined stiffness of columns extending both above and below the slab
and fixed at their far ends. These dimensions were arbitrary since the
selection of the story height (12ft for the F- and 9ft for the T-structures)
was arbitrary as were the column sizes.
The F-structures were designed by DiStasio and van Buren, and Struc-
ture T1 by Paul Rogers and Associates. All four of these structures were
designed in accordance with ACI 318-56. Structure T2 was designed by
the project staff on the basis of limit design. All designs were made for
a nominal concrete strength of 3000 psi and a steel working stress of
20,000 psi. In the design of Structure T2, the yield stress of the reinforce-

1-3/4 " 1-3/4"

-,~- 1l
Fig. 12-Structures F2 and F3
Slab thickness: 1.75 in. Interior column section: 3.75 x 3.75 in.
Drop panel thickness: 2.5 in. Edge column section: 5 x 3.5 in.
Deep beam: 2 x 6 in. Corner column section: 3.5 x 3.5 in.
Shallow beam: 4.5 x 2.5 in.


Two-way F3
T1 slab Fl F2 Flat slab
Two-way Modified Flat Flat Structural
slab design plate slab fabric
Ultimate load w .. , lb per sq ft 537 466 360 551 955
Design load Wd, lb per sq ft 145 145 155 285 285
Yield stress fy, kips per sq in 42.0 47.6 36.7 42.0 70
3.7 3.2 2.3 1.9 3.4

1.8 1.4 1.3 0.92 0.96

f, = 20,000 psi
ment was assumed to be 40,000 psi. The design loads are shown in
Table 1. For the actual test structures, the design wall load was one-
fourth of that shown in the table.
The F-structures were provided with two different depths of spandrel
beams as shown in Fig. 11 and 12. The beams of Structure T2 were pur-
posely designed shallow so that this structure represented a condition
between Structures F1 and Tl.
A small-aggregate concrete was used in the test structures. The aver-
age compressive strength at the beginning of tests on the structure, in-
dicated by 2 x 4-in. cylinders, was 2510 psi for Structure F1, 2760 psi for
Structure F2, 3760 psi for Structure F3, 2830 psi for Structure T1, and
3550 psi for Structure T2. Tests on 4 x 8-in. cylinders indicated 3900 psi
for Structure F3, and 3420 for Structure T1 and 4090 psi for Structure T2.
Tests on 4 x 8-in. cylinders were not made for F1 and F2. *
The slab reinforcement was cut from ¥s in. square bars of C-1018 steel
for Structure F1 and B-1113 steel for Structures F1, T1, and T2. The
steel was annealed to have a stress-strain curve similar to that of inter-
mediate grade reinforcing bars. The yield stresses are listed in Table 2.
The slab in Structure F3 was reinforced with welded wire fabric. The
value shown as the yield stress in Table 2 represents the 0.2 percent
offset stress. The ultimate stress was approximately 5 percent higher.
The longitudinal beam reinforcement was cut from plain #2 bars (f11 =
50,000-psi) for Structures F1, F2, Tl, and T2, and from deformed #2 bars
(fu =- 54,000 psi) for Structure F3.
Tc· simulate a uniform load, the load was applied on each panel at
16 points arranged symmetrically about midpanel. The load on each
pa11el was applied by an individual jack. No "wall load" was applied on
the beams. A top view of Structure Tl is shown in Fig. 14.
Each test structure was instrumented with approximately 300 electric
strain gages to measure strains in the reinforcement. The load applied
on each panel was measured as were the three components of each
•The chronological order of testing was as follows: Structures F1 (1958), F2 (1959), TI
(1960), T2 (1961), and F3 (1962).

L 2"
5'~ 0" 5'- o" f:f-0"
II I' 'I' i I'
I 11 111 I
lj Ill Ill II b
I II l1 I
l----------~ L---------)~ ________ _;
+r L+-r
L"tr- -;-+ -

I i l I
1 I
11 I I
I 11 11 I
I il Ill I 0
I 1
L _________J4----------J ~-------
1 I
- T+
I 11
·l II II I ~
I 11 II I in
I 11 II I
I 11 II I
.L__________ _)p----------
II+,__ ____-=------ , I


1 Fig. 13-Structures T I and T2

Tl T2 Interior column section: 6 x 6 in.
h,~~~- ~~~~
1 ft 4% in. 1 ft 1% in. Edge column section: 6 x 4 in.
Interior beam: 3 x 6 in. 3 x 3 in. Corner column section: 4 x 4 in.
Exterior beam: 3 x 4.25 in. 3 x 3 in. Slab thickness: 1.5 in.

reaction at the bases of the columns. Deflections were measured at the

33 intersections of the column and panel center lines at midspan. A bot-
tom view of Structure F3 is shown in Fig. 15.
A series of tests was performed on each structure to study its behavior
under various load patterns and magnitudes. All panels were loaded
equally in the test to failure. Each test took from 4 to 12 hr.
Measured ultimate loads for the test structures
The strength of one mul1Jipanel flat slab or two-way slab structure
does not necessarily lead to a complete vindication or indictment of the
method of design used. The structure and the methods of design involve
many seemingly minor details which make critical differences in
strength. Nevertheless, a comparison of the measured strengths of the
test structures at face value does indicate the inconsistencies of the cur-
rent design methods insofar as considerations of safety are concerned.
As would be expected from the comparisons shown in Fig. 7 through
10 the highest ratio of ultimate load to design load was registered by
structure Tl, the conventional two-way slab, and the lowest by Structure
F2, the flat slab (see Table 2). The flat slab reinforced with structural

Fig. 14-Top view of Structure T I


Fig. IS-Bottom view of Strudure F3

fabric, Stru~ture F3, had a h~gher ratio of ultimate to design load because
it utilized higher strength reinforcement at the same allowable stress.
Although an oversimplification the ratio w,,f,/wafv, with f.= 20,000 psi,
represents a better means of comparison. In terms of this ratio, Struc-
tures F2 and F3 are comparable. Structure Fl, designed on essentially
the same basis, had a higher value because of the factor F [see Eq. (9)]
and because in some sections the minimum prescribed reinforcement
ratios governed. The ratio for Structure Tl was almost twice that for
Structure F2. The ratio for Structure T2 was what it was intended to be.
Structure T2 was designed for a 20,000 psi working stress for a steel with
an assumed yield stress of 40,000 psi. The actual yield stress was 47,600
The information in Table 2 is given simply to provide perspective, not
as a final judgment on the current state of the art. Detailed descriptions
and analyses of the test structures will be provided in subsequent papers
which should answer the many questions raised by Table 2.
One over-all generalization from the test results should be mentioned
here. The slab will carry almost any uniform load that is put on it. The
problem is to transmit the load out of the slab to the supporting members
and to make the supporting members stiff and strong enough to provide
the necessary reactions in the vertical and horizontal p1anes. The key to
Turner's success was in his handl'ing of these vitally important problems,
unwisely considered by some as men: details.

The paper is based on an investigation conducted in the Structural Research
Laboratory of the University of Illinois Civil Engineering Department. Con-
tributions in support of the project were received from the following organiza-
Reinforced Concrete Research Council of the American Society of Civil
Engineering Foundation
Concrete Reinforcing Steel Institute
Portland Cement Association
Wire Reinforcement Institute
Directorate of Civil Engineering, Headquarters, U. S. Air Force
General Services Administration, Public Buildings Service
Office of the Chief of Engineers, Corps of Engineers, U. S. Army
Bureau of Yards and Docks, Engineering Division, U.S. Navy
The project was initiated by ACI-ASCE Committee 421 (321), Design of
Reinforced Concrete Slabs, and was guided by an advisory committee on which
the following persons have served: L. H. Corning (chairman 1956-62), Douglas
McHenry (present chairman), G. B. Begg, Jr., W. J. Bobisch, F. B. Brown,
M. P. van Buren, Joseph Di Stasio, A. S. Neiman, N. M. Newmark, D. H. Pletta,
J. R. Powers, Paul Rogers, E. J. Ruble, W. E. Schaem, and C. A. Willson.
Acknowledgment is due W. L. Gamble, D. S. Hatcher, J. 0. Jirsa, and M. D.
Vanderbilt for their invaluable work in the execution and reporting of the

1. Westergaard, H. M., and Slater, W. A., "Moments and Stresses in Slabs,"
Proceedings, ACI, V. 17, 1921, pp. 415-538.
2. McMillan, Angus B., "A Comparison of Methods of Computing the Strength
of Flat Reinforced Plates," Engineering News, V. 63, No. 13, Mar. 31, 1910, pp.
3. Brayton, L. F., "Methods for the Computation of Reinforced Concrete Flat
Slabs," Engineering News, V. 64, No. 8, Aug. 25, 1910, pp. 210-211.
4. Eddy, H. T., and Turner, C. A. P., Concrete-Steel Construction; Part 1,
Building, 2nd Edition, Minneapolis, 1919.
5. Godfrey, Edward, Discussion of "Steel Stresses in Flat Slabs," Transactions,
ASCE, V. 77, 1914, pp. 1389-1392.
6. Turner, C. A. P., Discussion of "Steel Stresses in Flat Slabs," Transactions,
ASCE, V. 77, 1914, pp. 1416-1422.
7. Joint Committee, "Progress Report of Special Committee on Concrete and
Reinforced Concrete," Proceedings, ASCE, V. 39, Part 1, Feb. 1913, pp. 117-168.
8. Lord, A. R., "A Test of a Flat Slab Floor in a Reinforced Concrete Build-
ing," National Association of Cement Users (ACI), V. 7, 1910, pp. 156-179.
9. Grashof, Franz, Theorie der Elasticitat und Festigkeit, R. Gaertner, Berlin,
10. Green, C. E., Structural Mechanics, John Wiley and Sons, New York, N. Y.,
11. Eddy, H. T., The Theory of the Flexure and Strength of Rectangular Flat
Plates Applied to Reinforced Conc1·ete Floor Slabs, Rogers and Co., Minneapolis,

12. Turner, C. A. P., Elasticity; Structure and Strength of Materials Used in

Engineering Construction, Developed by the Geometry of Strain and the Energy
Function thru Thermodynamic and Chemical Equations in Such Ways that
Strength May be Figured from Chemical Composition of Rolled Metals and
the Modification of this Resistance by Mechanical Work, Minneapolis, 1934.
13. Turner, C. A. P., "Advance in Reinforced Concrete Construction: An Argu-
ment for Multiple-Way Reinforcement in Floor Slabs," Engineering News, V. 61,
No. 7, Feb. 18, 1909, pp. 178-181.
14. Nichols, J. R., "Statical Limitations Upon the Steel Requirement in Re-
inforced Concre,te Flat Slab Floors," Transactions, ASCE, V. 77, 1914, pp. 1670-
15. Talbot, A. N., and Slater, W. A., "Tests of Reinforced Concrete Flat Slab
Structures," Bulletin No. 84, Engineering Experiment Station, University of
Illinois, Urbana, 1916.
16. Binckley, G. S., Discussion of "Steel Stresses in Flat Slabs," Transactions,
ASCE, V. 77, 1914, pp. 1411-1416.
17. Joint Committee on Concrete and Reinforced Concrete, "Final Report,"
Proceedings, ACI, V. 13, 1917, pp. 509-566.
18. ACI Committee on Reinforced Concrete and Building Laws, "Proposed
Standard Building Regulations for the Use of Reinforced Concrete," Proceedings,
ACI, V. 13, 1917, pp. 410-423.
19. ACI Committee on Reinforced Concrete and Building Laws, "Standard
Building Regulations for the Use of Reinforced Concrete," Proceedings, ACI,
V. 16, 1920, pp. 283-302.
20. Joint Committee on Standard Specifications for Concrete and Reinforced
Concrete, "Standard Specifications for Concrete and Reinforced Concrete," Pro-
ceedings, ACI, V. 21, 1925, pp. 329-425. (Report submitted in 1924.)
21. ACI Committee E-1, Reinforced Concrete Building Design and Specifica-
tions, "Tentative Building Regulations for Reinforced Concrete (ACI E-A-28T) ,"
Proceedings, ACI, V. 24, 1928, pp. 791-833.
22. ACI Committee 318, "Building Code Requirements for Reinforced Concrete
(ACI 318-56) ," ACI JouRNAL, Proceedings V. 52, No. 9, May 1956, pp. 913-986.
23. ACI Committee 318, "ACI Standard Building Regulations for Reinforced
Concrete (ACI 318-41) ," ACI JOURNAL, Proceedings V. 41, No. 4, June 1945,
pp. 559-619.
24. Marcus, Henri, "Die Wirksame Stutzflache Der Tragerlosen Pilzdecken,"
Beton und Eisen (Berlin), Nos. 19 and 20, 1926, pp. 352-358 and 370-377.
25. Hunke, E., "Momententabellen Fur Pilzdecken," Beton und Eisen (Berlin),
No. 17, 1931, pp. 308-313.
26. Podgajetz, P., "Die Berechnung der Pilzdecken als stellvertretender Rah-
men unter Bericksichtigung der wirksamen Stutzflache," Beton und Eisen (Ber-
lin), No. 17, 1937, pp. 314-318.
27. Dewell, H. T., and Hammill, H. B., "Flat Slabs and Supporting Columns
in Walls Designed as Indeterminate Structural Frames," ACI JouRNAL, Pro-
ceedings V. 34, No. 3, Jan.-Feb. 1938, pp. 321-343.
28. Peabody, Dean, "Continuous Frame Analysis of Flat Slabs," Journal.
Boston Society of Civil Engineers, V. 26, No. 3, July 1939, pp. 183-207.
29. Joint Committee on Concrete and Reinforced Concrete, "Progress Re-
port," Transactions, ASCE, V. 66, 1909, pp. 431-482. (Submitted in 1908).
30. ACI Committee E-1, Reinforced Concrete Building Design and Specifica-
tions, "Tentative Building Regulations for the Use of Reinforced Concrete,"
Proceedings, ACI, V. 23, 1927, pp. 643-702.

31. Taylor, F. W., and Thompson, S. E., A Treatise on Concrete Plain and
Reinforced, 3rd Edition, John Wiley and Sons, New York, N. Y., 1916, 885 pp.
32. Taylor, F. W.; Thompson, S. E.; and Smulski, Edward, "Concrete, Plain
and Reinforced: V. 1, Theory and Design of Concrete and Reinforced Structures,
John Wiley and Sons, New York, N.Y., 1925.
33. ACI Committee 501, "Building Regulations for Reinforced Concrete (ACI
501-36T) ," ACI JoURNAL, Proceedings V. 32, No. 4, Mar.-Apr. 1936, pp. 407-444.
34. Di Stasio, Joseph, and van Buren, M. P., "Slabs Supported on Four Sides,"
ACI JoURNAL, Proceedings V. 32, No. 3, Jan.-Feb 1936, pp. 350-364.
35. Bertin, R. L.; Di Stasio, Joseph; and van Buren, M. P., "Slabs Supported
on Four Sides," ACI JouRNAL, Proceedings V. 41, No. 6, June 1945, pp. 537-556.
36. ACI Committee 318, "Building Code Requirements for Reinforced Con-
crete (ACI 318-47)," ACI JouRNAL, Proceedings V. 44, No. 1, Sept. 1947, pp. 1-64.
37. Westergaard, H. M., "Formulas for the Design of Rectangular Floor Slabs
and the Supporting Girders," Proceedings, ACI, V. 22, 1926, pp. 26-46.
38. Joint Committee, Recommended Pmctice and Standard Specifications for
Concrete and Reinforced Concrete, American Concrete Institute, Detroit, June
1940, 271 pp.

M =bending moment c width of capital in direction
M.. total negative moment acting considered
on one edge of a panel A' c/2
Ms total positive moment acting at A'ar average of the values of A' on
panel center line
the two ends of a span
M. total static moment
r proportion of load assumed to
M.. total design static moment
W total load on one panel be carried in direction L
Wov average of the total load on A, cross-sectional area of rein-
two adjacent spans forcement
w unit load on one panel f, steel stress
L span in direction considered jd internal lever arm for a rein-
(center-to-center of columns forced concrete section based
for flat slabs) on the straightline formula
Lz span perpendicular to L steel strain
b short span of a rectangular E, modulus of deformation for
panel steel

Received by the Institute May 27, 1963. Title No. 60-50 is a part of copyrighted Journal of the
American Concrete Institute, Proceedings V. 60, No. 8, Aug. 1963. Separate prints are available
at 75 cents each, cash with order.

American Concrete Institute, P. 0. Box 4754, Redford Station 1 Detroit 19, Mich.

Discussion of this paper should reach ACI headquarters in triplicate

by Nov. 1, 1963, for publication in the Part 2, March 1964 JOURNAL.

Sinopsis- Resumes- Zusammenfassung

lnvestigacion Sobre Losas de Piso de Tablero Multiple de Concreto
Metodo de Diseno- Su Evoluci6n y Comparaci6n
Este es el priJllero de un numero de informes a ser publicados sobre una in-
vestigaci6n en losas de piso de tableros multiples, de concreto. Este articulo
sirve como una introducci6n a los informes posteriores. Se presentanse los
antecedentes hist6ricos a los metodos de disefio para losas de piso publicados en
el ACI 318-56, asi como los resultados generales de las cinco estructuras en-
sayadas durante la investigaci6n.

Investigation des Dalles de Plancher Multi-Panneau en Beton Arme

Methodes de Calcul - Leur Evolution et Comparaison
Cet article est le premier dans une serie en train de publication sur !'investi-
gation des dalles de plancher multi-panneau en beton arme. L'article servira
comme introduction a ceux qui suivront, et qui presentront l'arriere-plan his-
torique des methodes de calcul des dalles de plancher dans l'ACI 318-56 et les
resultats generaux des tests sur cinq batiments eprouves pendant !'investigation.

Berechnung von durchlaufenden Stahlbetonplatten

Berechnungsmethoden- Entwicklung und Vergleich
Dies ist der erste aus einer Reihe von Berichten tiber die Berechnung von
durchlanfenden Stahlbetonplatten die zlir Veroffentlichung kommen. Die Arbeit
dientals Einleitung zu den spateren Berichten. Sie gibt die geschichtliche Ent-
wicklung der Berechnungsmeltoden fUr Betondecken in ACI 318-56 und die allge-
meinen Versuchsergebnisse an flinf Bauwerken die wahrend der Untersuchung
geprlift wurden.