IL L I N 0 I S

UNIVERSITY OF ILLINOIS AT URBANA-CHAMPAIGN

PRODUCTION NOTE
University of Illinois at
Urbana-Champaign Library
Large-scale Digitization Project, 2007.
T$ OFP

)F 1906
 d... ~

^-. 1
I A.*.;
'
U ^""i

8 1
!£ '
1 /
*y '^;'
i'*''1<
. . '"' * <~

:, -*
- I III
I

HIS BULLETIN is the twelftheof a series of bulletins pub-

A lishied by the Engineering Experiment Station of the
.4~.

UniTversity- of Illinois. The Engineering Experiment

* I~ j
Station was established by actoi of the Board of Trus- C-
1
~I

tees Deicember 8, 1908. It is the pirpose of the Station to carry

U
on investigations along various lines of engineering, and to study ~
U
/
problemsi pf importanice to professional engineers and to the man-
ufacturing, railwayi mining, construcional and industrial inter-
ests of the state.
The control of the Engineering Experiment Station is vested
in the leads of the several departmentsa of the College of En-
gineering. These constitute the Station Staff, and with the Di-
rector determine the character of the investigations to be under-
taken. The work is carried oi under the supervision of the Staff;
sometimes by a Fellow tas L A
graduate work, sometimes by a mem-
ber of the instructional force of the College of Engineering, but
I *
more frequently by a l investigator belonging to the Station Corpe.
 UNIVERSITY OF ILLINOIS

ENGINEERING EXPERIMENT STATION

BULLETIN No. 12. FEBRUARY 1907

TESTS OF REINFORCED CONCRETE T-BEAMS

SERIES OF 1906.

BY ARTHUR N. TALBOT, PROFESSOR OF MUNICIPAL AND SANITARY

ENGINEERING AND IN CHARGE OF THEORETICAL
AND APPLIED MECHANICS.

CONTENTS.

I. INTRODUCTION.
Page.
1. Preliminary ................. ..... ............. ... 3
2. Scope of Tests.......... ....... .................. 3
3. A cknowledgm ent ........... ........................... 4

II. RESISTANCE OF T-BEAMS TO FLEXURE.

4. General ........................... ....... . 4

5. N otation ......... .. .... ....... ... .. . . . ......... 5
6. Longitudinal,Tensile and Compressive Stresses,and
Location of Neutral Axis.-Approximate Solution 6
7. W eb S tresses....... ........ ............................ 9
8. Integrity of Flanges ............... ...... ...... ... 11
9. M ethod of Tjeatm ent............... . ............... 12
10. Other Formulas for T-beams .......................... 12

III. MATERIALS, TEST PIECES, AND METHODS OF TESTING.

11. M aterials ..... ........... . ............ .....1 .

12. Test Specimens.............. . .... ......
...... 14
13. Forms ................ ..... . .....................
* .17
14. Fabrication and Storage .............................. 17
15. M ethod of Testing .................... ..... .... ............ 18
2 ILLINOIS ENGINEERING EXPERIMENT STATION

IV. EXPERIMENTAL DATA AND DISCUSSION.

Page
16. Cube and Cylinder Test Data ... .................... 19
17. Deflection Diagrams......... . ............ 19
18. Phenomena of Tests .............. ...... .... 20
19. Tension in Steel ......... ...... ................. 24
20. Compression in Concrete............... ..... 25
21. Web Stresses ......................... . 26
22. Beam Deflection ......... ..... ................ ... 28
23. Position of Neutral A xis......... ........ ........ 28
24. Applicability of Results ......... ..... .... ..... 29
25. Sum m ary .......... ... ............... .............. 30
Deflection Diagram s ......... ....... ................. 32
 TALBOT-TESTS OF REINFORCED CONCRETE T-BEAMS

I. INTRODUCTION.

1. Preliminary.-The series of tests on T-beams herein

described was undertaken with two objects in view,-to de-
termine whether the width of slab within the limit used in the
experiments is a controlling element in the strength of the
beam, and to test the efficacy of vertical reinforcing stirrups in
resisting web stresses. °
When a reinforced concrete floor and its supporting beams
are built as one piece, the resulting composite structure forms a
system of T-shaped beams. There are differences of opinion on
the action of the T-beams so formed, and also differences as to
the width of flange or floor which may be considered to contribute
to the strength and stiffness of the beam. T-shapes may also be
used in the design of bridge girders and other structures for
special conditions. The large amount of reinforcement which
may be put into a T-beam without encroaching on the com-
pressive strength of the concrete too far and the resulting high
web stresses developed in the stem of the T-beam, make the
T-beam an advantageous form of test piece for determining the
efficacy of various forms and amounts of web reinforcement. It
is felt that this feature of T-beam testing is, in itself, sufficient
reason for the conduct of tests on T-beams.
2. Scope of Tests.-Three top widths of beam were used,
equal, respectively, to two, three, and four times the width of the
stem of the beam. The beams were reinforced vertically
with U-shaped stirrups. One size and spacing of stirrups
was used. To insure that failure woull not occur by slip-
ping of the vertical reinforcement in the concrete, the stir-
rups were made of deformed bars. The amount of longitudinal
reinforcement was made proportional to the width of flange of
the beam. Both mild steel plain round rods and high-carbon
Johnson corrugated bars were used for the longitudinal reinforce-
ment, as it was not known whether the point of elastic limit of
the metal would control the amount of load or method of failure.
As no data on the amount of the web stresses which may be
resisted by vertical stirrups were available and as little seemed
to be known on the effect of the width of flange or slab, the
beams were designed to give considerable latitude in the results
The series was considered as preliminary and leading to a set of
 ILLINOIS ENGINEERING EXPERIMENT STATION

tests which should include various forms and amounts of web

reinforcement.
3. Acknowledgment.-The investigation was made in the
Laboratory of Applied Mechanics of the University of Illinois as
a part of the work of the University of Illinois Engineering
Experiment Station. Assistance in the tests and in the calcula-
tions was given by F. S. Hewes and C. A. Hewes, senior students
in civil engineering, class of 1906, who used the results in their
thesis. Immediate supervision of the work of making the beams
and conducting the tests was given by D. A. Abrams, Assistant in
the Engineering Experiment Station. Acknowledgment is also
made to W. R. Robinson, Assistant in the Engineering Experi-
ment Station, for aid in the preparation of this bulletin. The
stone, sand and cement used in making the beams were furnished
by the Joint Committee on Concrete and Reinforced Concrete.
An analysis of some of the elements of flexure of T-beams
will be given. This will be followed with a description of the
test pieces and method of testing, the experimental data, and a
discussion of the results.

II. RESISTANCE OF T-BEAMS TO FLEXURE

4. General.-The analysis of the resistance of T-beams to

flexure -may be made to follow the general lines of analysis for
rectangular beams. If the tensile strength'of the concrete be not
considered at sections having a maximum bending moment and
if the flange or slab extends down to the neutral axis of the beam,
the resisting moment may be expected to be the same as that for
a rectangular beam of width equal to the width of the flange, pro-
vided, of course, that the integrity of the plane section is conserved.
If the flange does not extend down to the neutral axis. an
examination of the effect of the omission of apart of the com-
pression area must be made to find whether the formula for
rectangular beams is still applicable. With relatively shallow
flanges some modification of the formulas used in rectangular
beams for determining neutral axis, tensile and compressive
stresses, and resisting moment may be required. In the
determination of web stresses, a slightly different treatment will
be needed on account of the narrowed width of beam through
the stem. The stiffness and integrity of the flange next to its
junction with the stem will require investigation.
 TALBOT-TESTS OF REINFORCED CONCRETE T-BEAMS

In the treatment here given, the usual assumptions of beam

action noted in Bulletin No. 4 will be made. These include the
assumptions that a plane section before bending will remain a
plane section after bending and that tensile stresses in the con-
crete at the section of greatest bending moment may be neglected.
The stresses developed in the flange' to conserve the plane section
will be considered separately. The general treatment will follow
that given in Bulletin No. 4 of the University of Illinois Engi-
neering Experiment Station, under I. RESISTANCE OF BEAMS TO
FLEXURE, and equations from this source will be quoted without
demonstration. The term "inclosing rectangle" will be used to
denote the rectangle inclosing the flange of the T-beam and the
stem down to the centroid of the reinforcing bars.
5. Notation.-The following notation will be used, Fig. 1:
h = thickness of flange or slab of T-beam.
b = breadth of flange or slab of T-beam.
b'= breadth of stem of T-beam.
d = distance from the compressive face to the centroid of the
longitudinal reinforcement.
d'-- distance from the center of the longitudinal reinforcement
to center of gravity of compressive stresses.
A = area of cross section of longitudinal reinforcement.
bd= area of inclosing rectangle (rectangle inclosing flange and
stem of T-beam down to center of the longitudinal
reinforcement).
p =dratio of area of longitudinal reinforcement to area of
inclosing rectangle.
o = circumference or periphery of one reinforcing bar.
m = number of reinforcing bars.
E. = modulus of elasticity of steel.
E- = initial modulus of elasticity of concrete in compression, a
term defined in Bulletin No. 4.
n= = ratio of two moduli.

f-= tensile stress per unit of area in longitudinal reinforcement.

c = compressive stress per unit of area in most remote fiber
of concrete.
c'= compressive stress per unit of area which causes failure by
crushing.
 ILLINOIS ENGINEERING EXPERIMENT STATION

e = deformation per unit of length in the longitudinal reinforce-

ment.
e-= deformation per unit of length in most remote fiber of the
concrete.
e'c = deformation per unit of length when crushing failure occurs;
i. e., ultimate or crushing deformation.
S-- = ratio of deformation existing in most remote fiber to
$C ultimate or crushing deformation.
k = ratio of distance between compression face and neutral axis
to distance d.
z = distance from compression face to center of gravity of com-
pressive stresses.
M= resisting moment at the given section.
8 = horizontal tensile stress per unit of area in the concrete.
t = diagonal tensile stress per unit of area in the concrete.
u = bond stress per unit of area on the surface of the reinforcing
bar.
v = vertical shearing stress and horizontal shearing stress per
unit of area in the concrete.

FIG. 1. T-BEAM AND INCLOSING RECTANGLE.

6. Longitudinal, Tensile and Compressive Stresses, and

Location of Neutral Axis.--Approximate Solution.-Under the
assumption that horizontal tensile stresses in the concrete are
not to be considered and that a plane section before bending
remains a plane section after bending, the formulas for
rectangular beams may be made applicable to T-beams with
a close degree of approximation by the substitution of the
rectangle inclosing the flange and stem down to the centroid of
the reinforcing bars (called here the inclosing rectangle) shown
in Fig. 1. If the flange of the T-beam were thick enough to
extend down to or below the neutral axis, the part of the con-
crete not included in the inclosing rectangle would, by the
assumptions quoted above, not affect the resistance of the beam
 TALBOT-TESTS OF REINFORCED CONCRETE T-BEAMS

to flexure. If the flange does not extend to the neutral axis, a

part of the compression area assumed to exist for the inclosing
rectangle is not available in the T-beam, but the portion so cut
off has, by reason of its proximity to the neutral axis, a compar-
atively small influence upon the flexural action of the beam. The
effect is so little that we may, within certain limits of propor-
tions, substitute the inclosing rectangle for the T-section in
determining the position of the neutral axis, the tensile stress in
the reinforcement, the compressive stress in the most remote
fiber of the concrete, and the resisting moment of the section.
For this approximation, the formulas given in Bulletin No. 4
for the proportional depth of the neutral axis may be repeated :
For a constant modulus of elasticity in the concrete (straight-
line stress-deformation relation),

k = 2 pn + p'n -pn ........................... (10)

When the deformation in the most remote fiber is equal to

one fourth of the ultimate or crushing deformation (q-=-4), a
condition approaching that found in beams of the usual per-
centage of reinforcement, the general equation for neutral axis
reduces to,

k= pn~~ ip* 2n- pn....................(11)

Equation (8) of Bulletin No. 4 may be used for the general case
and (9) when the concrete is at the limit of the compressive
strength.
The diagram given on page 16 of Bulletin No. 4 will be useful
in determining the position of the neutral axis if equation (11)
is to be used. With 1% reinforcement, for a ratio n equal to 12,
k will be .40; for a ratio A equal to 15, k will be .43.
For a beam in which the compressive stress developed is less
than the ultimate strength of the concrete (and this condition
covers all the usual cases of T-beams), the formula for the resist-
ing moment of the beam may best be expressed in terms of the
tensile stress in the reinforcing bars, as given in the two equa-
tions:
M = Af (d- z)..................................... (12)
= A fd' ............ ...... ....... .............. (13)
 ILLINOIS ENGINEERING EXPERIMENT STATION

If k = .40, or .43, the latter equation may be written M = .86

A / d, and this equation for the resisting moment of the T-beam
may be used with sufficient accuracy for many purposes, even for
quite a range of reinforcement. If the bending moment is
known, the stress in the steel may be calculated by substituting
the value of the bending moment for X in equation (13) or in the
reduced form M = .86 A f d. The effect of substituting the
inclosing rectangle for the exact T-shape in determining the
position of the neutral axis and the center of the compressive
stresses is so slight that the error may be neglected and the above
formulas used when the flange extends at least two thirds of the
distance to the neutral axis or, say, when the thickness of the
flange is at least one fourth of the depth of the beam. Even for
thinner flanges the error in the use of these formulas will not be
large.
Values for compressive stresses in T-beams are not of very
general usefulness since the percentage of reinforcement (based
on the inclosing rectangle) will generally not be large and the
co upressive stresses iay therefore not need to be considered and
may be sufficiently guarded by limiting the percentage of re-
inforcement. To illustrate, if in a T-beam made up of a beam
and its connecting portion of the floor a width equal to four
times the width of the beam itself be considered to be tributary
to the T-beam and if the area of the reinforcing bars is 4% of the
area inclosing the stem of the T-beam (b'd, of Fig. 1), this re-
inf6rcement will be only 1% of the inclosing rectangle of the
assumed T-beam. Under conditions of good construction, the
compressive stress developed in the concrete, would, as has
been shown for rectangular beams, be well below any danger of a
compression failure. This amount of steel is larger than would
ordinarily be used in such construction.,
If, however, it is desired to compute the compressive stress,
an approximate solution may be made by equation (15) of
Bullet'in No 4,
. 2Af 1 - q_ 2pf 1 - sq
c k T1 - q *
kbd 1-
i q
where b and p refer to the inclosing rectangle. For the con-

ditions of T-beams the fraction 1 will likely range between

1 -
 TALBOT-TESTS OF REINFORCED CONCRETE T-BEAMS

.92 and .97, so that its use or non-use does not affect the results
materially.
There is a greater proportional error in the use of equation
(15) for beams in which the flange does not extend down to the
neutral axis than there is in the use of equation (13), but as the
purpose is only to find whether a limiting value is exceeded, the
limits for depth of flange used with equation (13) may be con-
sidered allowable. This is further borne out by the fact that in
floor systems a width of floor even greater than here used will
generally be tributary to the T-beam, and hence the compressive
stress will be lower than for an assumed ratio of 4.
7. Web Stresses.-Bond, Shear, and Diagonal Tension.-In
T-beams the bond stresses developed are practically the same as
would be found with the same steel in a rectangular beam. The
shearing stresses developed and the corresponding diagonal ten-
sile stresses are higher, since the width of stem is relatively
small; and even with a moderate amount of reinforcement the
resistance to web stresses may constitute the weakest element of
the beam.
The. bond stress developed in a T-beam when the longitudinal
reinforcing rods are laid horizontally throughout the length of
the beam may be determined by equation (17) of Bulletin No. 4,
(p. 19),
V
u - mod' ............. .. .......... . ... (17)

where u is the bond unit-stress developed, V is the total external

vertical shear, m is the number of reinforcing bars, o is the
effective circumference or periphery of one bar, and d' is the
distance from the center of the longitudinal reinforcement to the
center of gravity of the compressive stresses. The only approxi-
mation to be made is in getting the value of d'. As in the pre-
ceding article, d' may be taken to be d (1 - .34 k), and k may be
obtained as before suggested by the use of the method of rec-
tangular beams with the width of the beam taken as that of the
inclosing rectangle. For most conditions d' may be called .86d.
The reasoning for the determination of vertical and horizontal
shearing stresses given on page 20 of Bulletin No. 4 is directly
applicable to T-beams, and formula (18) will give the horizontal
shearing unit-stress (and therefore its equal, the vertical shear-
 ILLINOIS ENGINEERING EXPERIMENT STATION

ing unit-stress), if we use in it the width Y', the width of the

stem of the T-beam.

v= ............ . .................. ............ (18)

In equation (18) v is either the horizontal or the vertical shear-.
ing unit-stress, the two shearing unit-stresses being always equal.
For most conditions d' may be taken as .86d, as in the last para-
graph. The amount of the vertical shearing unit-stress developed
in T-beams may be much higher than in ordinary rectangular
beams, but it will be well below the ultimate shearing strength
of the concrete.
The diagonal tensile stress in the stem of the T-beam is a
function of the shearing stress and the horizontal tensile stress
in the concrete. As the horizontal tensile stresses may not be
well determined, it seems best for our purpose to use the hori-
zontal and vertical shearing stresses as a means of comparison of
the diagonal tensile stress developed. If part or all of the reinforc-
ing bars are bent up at the ends, the problem is further compli-
cated. The beams described in this bulletin- were reinforced
with vertical stirrups, and as the concrete itself failed to resist
the diagonal tension at points well below the maximum load put
on the beam, the resistance of the concrete to diagonal tension
will be disregarded for these tests and the effect of the vertical
stirrups studied.
Two forms of metallic web reinforcement are used to coun-
teract the diagonal tension: (1) bending the reinforcing bars or
strips sheared from them into a diagonal position, and (2) verti-
cal stirrups carried under and around the longitudinal reinforce-
ment and extended upward to the top of the beam or to some
anchorage. The T-beams tested were reinforced with vertical
stirrups, and hence this form of web reinforcement will be con-
sidered further.
The diagonal tension existing in the web may be resolved into
horizontal and vertical components. Considering the longitudi-
nal bars to be all horizontal, we may expect, when the bond re-
sistance is sufficient, that the horizontal component will be taken
by the longitudinal reinforcement. Considering that the test has
passed the point where the concrete of the web resists the diagonal
tension, we may count that the whole vertical component of the
 TALBOT-TESTS OF REINFORCED CONCRETE T-BEAMS

web stresses is taken by the stirrups. The amount of this vertical

component per unit of length of beam is for the T-beams equal to
the horizontal shear for the width of the stem as given by

V
equation (18), and vb' = will give the rate of vertical stress

per unit of length of beam which will go to the stirrups. If the

stirrups are 6 inches apart, the stress to be taken by the two
prongs of the stirrup will be 6 vb'. This calculation is on the
basis of a test loading which gives a constant shear from support
to load point, as, for example, a loading at the one-third points.
For a uniformly distributed load, the value vb' will be the rate of
stress at a given section. To illustrate further, for a T-beam
having a width of stem of 8 in., and d' = 8.6 in., and a load of
60000 lb., V= 30000 lb., and v = 437 lb. per sq. in. With stir-
rups 6 inches apart the total stress on one prong of the stirrup
will be 10 500 lb. These calculations do not take into account
the resistance to bending of the reinforcing bars themselves.
In the T-beams having also a part of the bars bent up, in order
to avoid complicated calculations and to give a general com-
parison, the same formulas will be used herein, although the
results are not an accurate measure of the stresses produced.
8. Integrity of Flanges.-In discussions on T-beams it is fre-
.quently stated that the thickness of the flange must be at least one
half as greatas the width of the stem or rib, in order that. the shear-
ing stresses at the junction of the flange and stem may not exceed
those in the lower part of the stem. As the actual shearing
stresses in both sections are well within the shearing strength of
the concrete, there is no danger of failure by shear in either
section, using the term shear in its strict sense. In the stem we
have used the shearing unit-stress as a measure or method of
comparison for the diagonal tensile stresses. In the flange these
stresses do not need consideration. Longitudinal shearing stresses
in the flange, then, will not require consideration, and the limit
of depth given above is unnecessary.
Another statement sometimes made is that the compressive
stress in the flange varies across the width of the flange from a
maxifnum amount next to the stem to zero at the edge of the
flange. A little consideration will show that this cannot be true.
The flange must transmit stress laterally to its edges, acting in a
 ILLINOIS ENGINEERING EXPERIMENT STATION

way as a beam with a load applied longitudinally and horizon-

tally. The width of the overhang of the flange may be considered
to be the span of this cantilever beam, and the distance from the
load to the support (one-third the span 'length in the beams
tested) may be thought of as the depth. Evidently this will be a
very stiff beam, and the result will be a close approach to uni-
formity of compressive stress at points across the width of the
flange. Little variation from a plane section may therefore be
expected, the change that may exist being produced more largely
by other causes. Of course, there must be a limit to this
assumed integrity of the cross section, but for T-beams with the
ordinary amount of reinforcement the compressive stress devel-
oped is small and the variation may generally be neglected.
As usually constructed, the floor, and hence the flange of
the T-beam, will have reinforcement at right angles to the stem.
This will resist a breaking of the flange next to the stem and
assist in giving the whole flange the curved shape which the
beam takes when the load is applied.
9. iMethod of Treatment.-For the investigation herein
recorded, the preceding method of analysis will be used. The
T-beam will be treated as a rectangular beam of the dimensions
given by the inclosing rectangle, when tensile, compressive, and
bond stresses are under consideration. The width of the stem
will be considered to enter into calculations for shear and for.
diagonal tension. The integrity of the cross section will be
investigated somewhat, and the bearing of the approximation
involved in the above assumptions will be looked into.
10. Other Formulas for T-beams.-If it is desired to take
into account the exact section above the neutral axis the following
formula for the proportionate depth of the neutral axis may be
used, if we consider q = 0, or if we use a straight-line modulus of
elasticity. Other values of q may be used.

k 2pn ( - (1- [pn ( 1

-p Y ) - 1d ................ (10')
For a beam with 1% reinforcement and having a ratio of
depth of flange to effective depth of steel of , using for
L iC) of
 TALBOT-TESTS OF REINFORCED CONCRETE T-BEAMS 18

comparison n = 15, this formula gives k = .44, and the approxi-

mate method previously described gives k = .42. For =
-,

k - .46. It is easily seen that the effect of this difference upon

the calculated resisting moment of the beam (when based on
tension in the steel) is small.

III. MATERIALS, TEST PIECES, AND METHOD OF TESTING.

11. Materials.-The stone was a good quality of rather hard
limestone from Kankakee, Illinois, ordered screened through a
1-in. screen and over a ¼-in. screen. It contained 45% to 50%
voids and weighed 85 pounds per cubic foot. In the determina-
tion of the voids of both stone and sand, the material was poured
slowly into water so that the voids became filled with water and
no air was caught.
The sand was of good quality from near the Wabash river at
Attica, Indiana. It was fairly clean, sharp and well graded, con-
tained 28% voids, and weighed 115 lb. per cu. ft. Table 1 gives
the results of a mechanical analysis of this sand.

TABLE 1.
MECHANICAL ANALYSIS OF SAND.

Per cent
Sieve No. Passing

4 100
10 73
20 36
30 12
74 5
100 2

The cement used was furnished by the Joint Committee on

Concrete and Reinforced Concrete. It was made up of a mixture
of five standard American portland cements,, selected and mixed
by the manufacturers and was of excellent quality.
The concrete was made of the proportions of 1 of cement, 2 of
sand, and 4 of stone, measured by loose volume. The mixing was
done and the beams made by men skilled in concrete work. Care
was taken in measuring, mixing, and tamping, to secure as
uniform a concrete as possible. The mixing was done with
 ILLINOIS ENGINEERING EXPERIMENT STATION

shovels by hand; the stone having been wetted a day or so before

being used. The sand and cement were first mixed dry, the
stone being then added and the mass mixed until uniform in
appearance. Water was added in such proportion that the
tamping of the central portion of the beam really amounted to a
churning action.
The steel used for the horizontal reinforcement in some of
the beams was 1-in. mild steel plain round rods, and in the rest
was 1-in. high steel Johnson corrugated bars. The stirrups were
in all cases J-in. high steel Johnson corrugated bars. Table 2
gives physical tests of these rods, the samples tested being from

TABLE 2.
TENSION TESTS OF STEEL USED IN T-BEAMS.

Average Values.

T-beam Size of Bar I Per cent Yield Maximum Yield Maximum

Elongation Point Load Point Load
No. inches I in 8 in. pounds pounds lb. persq. in. lb. per sq. in.
1 i-in. Johnson 13.5 30 700 51300 54900 91 600
2 bar (new section) 15.0 30 250 48 000 53 750 86 200
3 Area=.56 sq. in.I 14.5 29700 50000 52700 89 000
5 15.0 30100 48900 53380 86980
Av. 14.5 30 200 49 500 53 800 88400

4 .750 plain round 34.5 16600 24600 35 100 55 500

8 .752 plain round 30.8 18100 26600 40700 59900
9 .750 plain round 32.0 17 300 26100 39 100 59 100

Av. 32.4 17 300 25 800 38 300 58 200

the ends left after the bars were cut for the beam. The
average ultimate strength of the plain round rods was 58 200 lb.
per sq. in. and the yield point 38 300 lb. per sq. in. The 4-in.
Johnson bars developed an average ultimate strength of 88400 lb.
per sq. in., and a yield point of 53800 lb. per sq. in.
12. Test Specimens.-Table 3 shows the specimens which
were made and tested. The nine T-beams were 11 ft. long and
the span usel was 10 ft. The following dimensions were con-
stant for all the beams: thickness of flange,-34 in.; width of
web, 8 in.; depth over all, 12 in. and depth from top fiber to
center of steel, 10 in. The width of flange varied, three speci-
 TALBOT-TESTS OF REINFORCED CONCRETE T-BEAMS

mens being made of each of the widths; 16 in., 24 in., and 32 in.
The proportion of reinforcement was from .92% to 1.10%,
based on the area of the inclosing rectangle, as shown in Fig. 2;
that is, it was obtained by dividing the area of the steel by the
product of the breadth of the flange and the effective depth of
the beam. This percentage was made as near 1.00 as the size of
the reinforcing bars would permit. It will be noted that if the
percentage were based on the area of the rectangle inclosing
the stem, it would be very high, the amount for the beams 32 in.
'wide amounting to about 4%. The reinforcing rods were symmetri-
cally arranged with respect to the axis of the beam. In some, a part
TABLE 3.
LIST OF TEST SPECIMENS.
All Beams Have Stirrups of 1 in. Johnson Bars Placed as Shown
in Fig. 2.
• lr -"__ _
T-beams 1Minor Specimens
Reinforcement
T-beam Width of Reinforcement
Area
Flange per Cubes Cylinders
No. inches Kind
sq. in. cent

1 16 3 a-in. Johnson bars 1.68 1.05 1

2 32 6 f-in. Johnson bars 3.36 1.05 2

3 24 4 f-in. Johnson bars 2.24 0.93 3

4 16 4 f-in. plain round 1.76 1.10 4

5 32 J6 ½-in. Johnson bars 3.36 1.05 5
2 bars bent up
6 24 5 f-in. plain round 2.20 0.92
(2 bars bent up
7 16 4 f-in. plain round 1.10 7

8 24 i-in. plain round 0.92 8

bars bent up
4-in. plain round 0.97
9 32
bars bent up

of the rods was turned up beyond the one third points to within 4 in.
of the top of the beam. Ten U-shaped stirrups of 4-in. high steel
Johnson corrugated bars were used in each beam. These stir-
rups were spaced 6 in. apart on the portions of the beam between
the load points and supports. They passed under the corrugated
 ILLINOIS ENGINEERING EXPERIMENT STATION

7
3t/ 114-'ohnsa/ w v./fMJ~'~f/l /', ^
I .i . .II. . I II'U _

N-e 1-6- Zlh,4n,, B ~/x3s,~

I----------...._...- .---- ,
"
VN 2 - S-
-" o ,
a,., 6-/;
A' W 8< 4"-/-4.Y3"/rm/r . ""
3- _

Ii i il i1
iI

i " g ------
u,) , ,
i1.,,, .-- ,y -n nI -.

f -
N- S iIII
~ 0-~.2" I II I I

= L-Li/^^ L-

I
uf ;
NO 7 - 4-'~¼~tnBs
i-
,
i '~
i m-x-j
,
Ti»
......
I
.
i. -
I'3
'
." 11,,1 '
---- , ..

, 1i i i t i i | ' .
l
• i", i 6 _ •

Nl< '-¢^-Sk/'s a/7

0o fs,•rl'.'/^i:4"l','
4-/"8t/P* ,Tt'frt',

40
N 8 - 2 8f

- 41
I i - .I-
11i I•I I- I I

SIr-v, V'ru
A/ a g - /- rounda G ,U

N-2 - 2 i---

' ,2.. . OF
ARRANGEMENT1 ,,IFOCEMN

FIG. 2. ARRANGEMENT OF REINFORCEMENT.

 TALBOT-TESTS OF REINFORCED CONCRETE T-BEAMS

reinforcing bars, alternate ones inclosing only the middle bars.

In two of the 32-in. beams three '-in. Johnson corrugated bars
were placed transversely in the upper side of the flange at and
near the load points. Fig. 2 shows the arrangement of the
reinforcement in all of the beams. In some of the beams in
which the reinforcing bars were placed in two layers, the upper
bars sank down in the mortar until they were in contact with
the lower layer of bars.
Fourteen 6-in. cubes were tested. The numbering is the
same as that for the beams for which the batch was made. Seven
8-in. cylinders were made in a similar manner.
13. Forms.-Fig. 3 shows the plan of the forms used in
making the T-beams. 2-in. planks were used for the sides and
ends. Clamps and struts were of 2-in. x 4-in. pieces. The planks

7Too I/ew n/a'd V'w

FIG. 3. PLAN OF FORMS.
were soaked in water for some time before using, in order to
prevent the absorption of water from the beams.
14. Fabrication and Storage.-The beams were made on
the concrete floor of the laboratory, strips of building paper
being laid on the floor to prevent the concrete from adhering to
it. As already stated, the concrete wis proportioned by loose
volume and mixed by hand. Generally, two batches were mixed
in making a beam.
After the form was set up, a layer of concrete 1-in. to 1I in.
thick was put in, the reinforcement placed, and the rest of the
concrete filled in in layers about 3 in. thick. The sides and
ends were spaded, the concrete tamped, then spaded again and
tamped. This gave a very good surface to the test pieces.
 ILLINOIS ENGINEERING EXPERIMENT STATION

The forms were kept on until a day or so before testing the

beams, the beams being sprinkled twice daily in the meantime.
The temperature of the room ranged from 600 to 70° F. during the
period of storage.
15. Method of Testing.-Fig.4 shows the method of loading
used. The beams were tested in the 600 000-lb. Riehle' testing
machine, the load being applied at the one-third points through
turned steel rollers, which were 2 in. or 4j in. in diameter,
according to the size of the beam. Each of these rested on a
cast iron plate, 2 in. x 3 in. x 2 ft. 8 in., faced on both sides,
which distributed the load laterally over the beam. The supports
had curved bottoms to permit rocking, and plates 8 in. x 11
in. x 8 in. were used as bearing plates. These plates, together
with the bearing plates on the top of the beam, were bedded in

FIG. 4. METHOD OF TESTING.

plaster of paris which was allowed to harden under the weight of

the beam and the aparatus used in loading, before the load was
applied. The movement of the head, in testing, was -y inch per
minute. Center deflections were read on all T-beams and de-
formations were observed on a few. The deflections were
obtained by means of a thread stretched between two points over
the supports and about seven inches from the top of the beam.
The frame used for holding the extensometer dials was improvised
and its lack of rigidity seriously impaired its usefulness, the
readings being sometimes inconsistent and therefore untrust-
worthy. Since at the present time tests are being made here in
which reliable deformation readings are expected, diagrams from
the readings taken in this series will not be reproduced.
The cubes, cylinders, and steel were tested on the 100000-lb.
 TALBOT-TESTS OF REINFORCED CONCRETE T-BEAMS

Riehle' machine, the speed of the head being .1 inch per minute.
The cubes and cylinders were bedded in plaster of paris before
testing.
IV. EXPERIMENTAL DATA AND DISCUSSION.

16. Cube and Cylinder Test Data.-Table 4 gives the re-

sults of the tests of the cubes and cylinders. The breaking
strength of the cylinders is somewhat below that of the cubes.
The concrete showed very good quality.
17. Dedleetion Diagrams.-Fig.7 to 15 at the end of the text
give the load-deflection diagrams for the T-beams. The ordinates
(vertical distances) represent the applied loads, and the abscissas
TABLE 4.
RESULTS OF CUBE AND CYLINDER TESTS.

Cubes Cylinders
Age Maxir aum Load Age Maximum Load
No. at Test Total No. at Test Total
Days Pounds lb. per sq. in. Days pounds
lb. per sq. in.
11 59 62 120 1720
12 59 84 690 2350 1 63 95 000 1890
18 59 66 390 1840 2 63 94 200 1870
3 61 88 200 1760
Av. 1970 4 59 67 000 1330
32 57 42 120 1170 5 58 60 000 1190
57 49 620 1380 7 56 73 000 1450
3s 57 46 200 1280 8 54 88 000 1750
Av. 1280 Av. 1610
51 58 69 800 1940
52 58 81500 2260
Av. 2100
71 56 57 600 1600
72 56 I 71 100 1980
78 56 68 300 1900
Av. 1830
81 54 68 100 1890
8a 54 55 600 1540
83 54 94 500 2620
Av. 2020

Average of all cubes, 1820

(horizontal distances) the corresponding deflection at the center

of the beam. It will be seen that at or near the maximum load
the curve changes direction abruptly, and that for nearly all the
 ILLINOIS ENGINEERING EXPERIMENT STATION

beams the load does not fall off materially until a considerable
deflection has been obtained. The stress-deformation curves
found (not reproduced) show an abrupt change on the tension
side of the beam.
18. Phenomenaof Tests.- In Table 5 (p. 25) are given the maxi-
mum loads which the beams sustained. Fig. 5 and 6 give sketches
showing position of cracks as they appeared after the maximum
load and also the shape of the flange in Beam No. 2 after it broke
off. The general phenomena of the tests of the T-beams were
quite similar to those attending the tests of rectangular beams

Beamw /Vye

If 4
,'i'i 4
I^,

I,

Seam7 ,- 7
4 4
t ,, ,
Be,7iw N-e.,,n'-•'
-
Beam /V-" 3

\ 4

4 4

E
BealM /VA.
FIG. 5. SKETCH OF BEAMS AFTER FAILURE.
 TALBOT-TESTS OF REINFORCED CONCRETE T-BEAMS

having a similar percentage of reinforcement. When the tension

in the steel reached 13000 to 20000 lb. per sq. in., minute verti-
cal cracks became visible in the concrete on the lower part of the
faces of the stem of the T-beam at points in the middle third of
the span length. These grew more distinct as the load was
increased. When the load became sufficient to strain the
steel to its yield point, these cracks opened up and finally became
quite large and extended up into and along the flange of the

Beaw N0 2

f
ae ,m ,a5"
^^=-^==^B N-9^
_T -

I .1I •., Ii,,H

/ //1 1^", , I
-^ (-~13`41

, earn N- 6

FiG. 6. SKETCH OF BEAMS AFTER FAILURE.

 ILLINOIS ENGINEERING EXPERIMENT STATION

beam. The evidences of failure by tension in the steel were

apparent in every beam, as much so in those reinforced with
corrugated bars as in those having plain rods.
At loads somewhat above the amounts at which it would be
expected that failures by diagonal tension in the concrete would
occur in beams not having metallic web reinforcement, minute
diagonal cracks appeared on the face of the stem in the outer
thirds of the span length. These cracks had the direction and
appearance of cracks attending failure by diagonal tension.
Their consideration will be taken up under "21. Web Stresses,"
(p. 26).
In general, the flanges gave no sign of failure. In one (Beam
No. 2) the flange split off after the maximum load was reached,
as shown in Fig. 6. In this beam there was no lateral reinforce-
ment in the flange. It seems evident that the tearing off
occurred by reason of the warping of the flange as bending took
place, and as this effect was not found in the other two beams of
32-in. width, the presence of the lateral rods in the beams was
probably advantageous. Such rods would, of course, be present
in ordinary floor construction, and uneven bending would be
resisted by this lateral reinforcement.
In the representation of diagonal and vertical cracks after
the maximum load had been reached, shown in Fig. 5 and 6, the
tension crack which caused failure is indicated by a heavy line.
The diagonal crack first to appear is numbered 1. These diagonal
cracks generally intersected the plane of the reinforcing bars at
their intersection with the stirrups. The description of the
action of the beams which follows is given only in sufficient detail
to indicate their general behavior, though individual peculiarities
of beams are also noted.
Beam No. 1.-Two tension cracks at points at the level of
the reinforcing bars between the load points appeared at an
applied load of 16 000 lb. Before 20 000 lb. was reached eight
more tension cracks appeared in the middle third of the span
length. At 20000 lb. the first diagonal crack appeared. At
26 000 lb. it had extended to within 2 inches of the flange. With
further loading a large number of diagonal cracks developed.
After reaching a load of 42000 lb., (deflection 0.5 in.), the load
fell off to 41000 lb., as the machine speed did not keep up
with the deflection of the beam under the given load. The load
 TALBOT-TESTS OF REINFORCED CONCRETE T-BEAMS

then-slowly rose to 44500 lb., the deflection increasing to 1.34 in.

During this period the tension cracks opened up quite rapidly,
while the diagonal cracks remained about the same, one diagonal
crack becoming .y in. wide. Finally, at a deflection of 1.6 in.,
and an applied load of 43800 lb., the concrete crushed at one of
the load points.
Beam No. 2.-The failure of this beam was peculiar in that
after it had passed the maximum load and had shown all the
evidences of failure by tension of the steel, instead of holding
about the same load and deflecting considerably more as the
machine was run down, the final breaking was by the flanges
splitting off as shown in Fig. 6. This was the only one of the
three beams with 32-in. flanges which did not have lateral rein-
forcing rods in the flange at and near the load points. The first
tension crack appeared at a point 10 in. west of the center at a
load of 26000 lb. and others appeared at 28000 lb.
The first diagonal crack was observed 19 in. west of the west
load point at 36000 lb., and reached half-way to the flange. A
number of other diagonal cracks appeared as the load was
increased. At 56000 lb. one diagonal crack had reached the
flange and small horizontal cracks formed at the junction of the
flange and stem. The beam carried the load well and reached a
maximum of 78300 lb. applied load. The load gradually fell off
as the deflection was increased, and finally the flange on both
sides from the east end to the west load point broke off suddenly,
as shown in Fig. 6.
Beam No. 3.-In beam No. 3 it was noticed that each diagonal
crack passed through the intersection of the stirrups and the
longitudinal reinforcement. At the maximum load one of the
tension cracks extended well into the flange.
Beam No. 4.-In Beam No. 4, some time after the maximum
load of 29 900 lb. had been reached, the principal tension crack
extended to the flange and then extended along the junction of
the stem and flange, but-no consequences were traceable to this.
As in'the other cases, the failure was by tension in the steel,
followed finally by compression in the concrete.
Beam No. 5.-At an applied load of 25 000 lb., two tension
cracks and one diagonal crack appeared, others following later.
At 74000 lb. the deflection curve makes an abrupt change of
direction, the deflection at this load being .54 in. The tension
 ILLINOIS ENGINEERING EXPERIMENT STATION

cracks increased in size and failure by tension in steel- was

evident. With increased deflection the load ran up slowly,
reaching a maximum of 80000 lb. at a deflection of 1.4 in., when
one roller rolled off its bearing and one pedestal fell forward,
allowing the beam to drop and break. At the maximum load-
there were evidences of crushing in the concrete.
Beam No., 6.-In Beam No. 6, after the maximum load of
37000 lb. was reached the principal tension crack extended into
the flange and finally extended horizontally along the junction of
flange and stem. The test was then discontinued.
Beam No. 7.-In Beam No. 7, the maximum load was main-
tained from a deflection of .3 in. to one of 1.9 in. before final
failure.
Beam No. 8.--In Beam No. 8, the maximum load was main-
tained from a deflection of .3 in. to one of 2.6 in. before final
failure. The stress-deformation diagram gave a compressive
unit deformation of .0005 at the point where the steel yielded.
Beam No. 9.-The stress-deformation curve and the deflection
curve show failure by tension in the steel. The position of cracks
is shown in Fig. 5.
19. Tension in Steel.-As already noted, the abrupt turn in
the deflection diagrams (Fig. 7 to 15, following the text), taken
in connection with the action of the concrete on the compressive
side, indicates that the beams failed in every chse through the
steel becoming stressed beyond the yield point. The stress-
deformation diagrams of the four beams on which observations of
longitudinal deformations were made, also give marked evidence
of failure by tension in the steel. In Table 5 are given the cal-
culated stresses in the, reinforcement at the maximum load, based
on the formula M=-.86 A f d. The weight of the beam and of
the loading apparatus was included in the calculations. The
average stress in the reinforcement in the beams reinforced with
plain mild steel rods at the ma'imum load is 39800 lb. per sq. in.
The average stress in the corrugated bars at the maximum load
is 58700 lb. per sq. in. It will be recalled that the average yield
point for the coupons taken from the ends of the reinforcing
bars was 38300 lb. per sq. in. for the mild steel plain rods and
583800 lb. per sq. in. for the corrugated bars. The calculated
stresses at the maximum load are therefore somewhat above the
average yield point of the metal. It may also be noted that
 TALBOT-TESTS OF REINFORCED CONCRETE T-BEAMS

there is no marked difference in the results obtained for beams

of different width of flange. The tests and calculaticns go to
show that equation (13) may be used for calculating the resisting
moment, and it is believed that this formula may properly be
used at least up to a reinforcement equal to 1% of the inclosing
rectangle.
TABLE 5.
RESULTS OF TESTS OF T-BEAMS.
-M=.86A fd.
All failed by tension in steel.
Longitudinal Rein forcement Maximum Load
Stress
pounds Bending in
Per Kind SWith Per 8 Moment Steel
cent Ap- Beam inches M3 f
14A d AA
ant p- of lb. -in. lb. per
paratus Width sq. in.

1.05 3 f-in. Johnson bars 44 500 46 700 23 350 923 000 64 300
1.10 4 f-in. plain round 29 900 32 410 16 200 631 000 41 500
1.10 4 f-in. plain round 27 300 30 100 15 050 579 000 38 100
0.93 4 f-in. Johnson bars 5O 5CA
i.%••71Jll 55 700 18 570 1107 500 57 500

i-in. bent
0.92 1952 bars plain round 36 800 39 300 13 100 773 500 40 700
up

0.92
S9 a-in. plain round
52 bars 37 300 40 100 13 370 783 500 41 200
bent up.

1.05 6 1-in. Johnson bars 78 300 80 500 20 120 608 000 55 700
f6 f-in. Johnson bars 2
1.05 bars bent up. (6trans- 80 800 83 300 20 820 1 658 000 57 400
verse bars in flange )
(7 f-in. plain round 3
0.971 bars bent up. (6trans- 48 100 50 900 12 720 1 004 000 37 600
( verse bars in flange)
I

20. Compression in Concrete.-The stress-deformation dia-

grams obtained in four beams show clearly that the full com-
pressive strength of the concrete was not developed when the
yield point of the beam was reached, the amount of shortening
being about .0005. The fact that the final deflection of the beam
generally was several times as much as that at the yield point of
 ILLINOIS ENGINEERING EXPERIMENT STATION

the beam shows that with the stretch of the steel beyond the
yield point the neutral axis must have risen very much before
final or ultimate failure by crushing of the concrete occurred,
and this corroborates the preceding statement. The results bear
out the assertion that 'beams with 1% reinforcement will have
only a part of the compressive strength of the concrete devel-
oped, even when steel of 54000 lb. per sq. in. yield point is used.
As in T-beams having a width of flange equal to four times the
thickness of the stem a 1% reinforcement (4% of the stem area)
is very large, it does not seem that the compressive strength of
the concrete in the upper fiber may be expected to be a con-
trolling element in floor construction of this kind.
21. Web Stresses.-Rectangular beams which do not have
metallic web reinforcement and in which the longitudinal bars
are laid horizontally may be expected (for concrete of this qual-
ity) to fail by diagonal tension when the calculated vertical
shearing unit stresses reach, say, 120 lb. per eq. in., provided, of
cohrse, the amount of reinforcement and relation of depth to
length of span are not such that failure by tension or compression
will occur before this amount of stress is developed. This is on
the assumption that the method of calculation given on page 10
is applicable to T-beams so reinforced. For the T-beams tested,
no matter what the width of flange, the load which would
produce failure by diagonal tension in the concrete calculated in
this way would be 17000 lb. As all beams failed by tension in
the steel and as the loads ran as high as 80000 lb. it is apparent
that the metallic web reinforcement was efficient and adequate.
The conditions of testing were not favorable for determining
the time of appearance of the first diagonal cracks (the minute
cracks which appear in the outer third of the span length and
which in beams without metallic web reinforcement presage
early and sudden failure), and in some cases these may have
appeared before their presence was noted. However, the load
at which their existence was observed is given in Table 6. In
Beam No. 7 the load for the first visible crack, 18000 lb., gives by
equation (18) a value of the vertical shearing unit stress of 147 lb.
per sq. in., not much above that at which a beam without
metallic web reinforcement would fail. The general average
time of observation of the first visible diagonal crack corresponded
to a value of about 180.1b. per sq. in. This higher value of v may
 TALBOT-TESTS OF REINFORCED CONCRETE T-BEAMS

be an indication that the stiffness of the stirrups holds back the

growth of cracks of the size at which they become visible or
even that the stirrups act with the concrete in taking these
secondary stresses up to the load where the cracks become
visible, or else that the method of calculation gives too high
results.
After the diagonal cracks have appeared, it must be con-
sidered that the secondary tensile stresses are taken mainly or
wholly by the metallic reinforcement. In Table 6 are given the
values of the vertical shearing unit-stresses for the maximum
loads, calculated by equation (18). As in no case was there a

TABLE 6.
WEB STRESSES.

At First Diagonal Crack At Maximum Load

Beam Flange Load in rounds
Width V Bond, u
in cluding lb. per Shear, v
No. inches Applied Beam and lb. per sq. in. lb. per sq. in.
sq. in.
Apparatus

1 16 20 000 22 200 161 302 340

4 16 29 900 32 100 234 200 236
7 16 18 000 20 200 147 186 219

3 24 28 000 30 500 222 270 405

6 24 20 000 22500 164 194 286
8 24 27 000 29500 214 198 292

2 32 36 000 38800 282 260 585

5 32 35 000 37 800 275 269 605
9 32 32 000 34 800 253 180 370

failure by diagonal tension, we are most interested in the beams

which carried the highest loads and hence developed the greatest
shearing unit-stresses. Beam No. 2 developed 585 lb. per sq. in.
and beam No. 5, 605 lb. per sq. in. These results are several
times as great as the values obtained with rectangular beams
without metallic reinforcement which failed by diagonal tension
of the concrete.
The stresses in the stirrups may be calculated tentatively by
the method given under "7. Web Stresses", (p.'9). For a load of
80 000 l'b., V= 40000 lb.; v = 580. For the width of stem of 8 in.
and stirrups 6 in. apart the tension in each prong of the stirrups,
 ILLINOIS ENGINEERING EXPERIMENT STATION

calculated in this way, will be 13900 lb., equivalent to 55500 lb.

per sq. in. The bond developed is also very high.
In the foregoing discussion the horizontal and vertical shear-
ing stresses have been used as a means of comparing resistances
to diagonal stresses. Viewed by themselves, they are also
interesting. 600 lb. per sq. in. is evidently well below the actual
shearing resistance of concrete, but it is considerably above the
values given as the ultimate shearing strength of concrete by
those who hold that shearing strength is but little more than the
tensile strength of concrete.
22. Beam Deflection.-Considering the beams all as rec-
tangular beams having the full section of the inclosing rectangle
and as having reinforcement equal to 1% of this rectangle, the
deflection in beams of different width of flange should be the
same at loads proportional to the flange width. That is, a load of
10 000 lb. on a rectangular beam 8 in. wide, one of 20 000 lb. on a
T-beam with flange 16 in. wide, one of 30000 lb. on a T-beam
with 24-in. flange, and one of 40000 lb. on a T-beam with 32-in.
flange, all having 1% reinforcement, should, on this assumption,
give the same deflection. However, the T-beams do not have the
full compressive area just above the neutral axis. As the con-
crete on the tension side of the neutral axis also adds to the
stiffness of the beam, especially in the part in the outer thirds of
the span length, the narrow width of the stem will detract from the
stiffness of the beam. It may be expected, then, that the
deflection of the T-beams will increase somewhat as the width of
flange increases, when compared on the basis of loads proportion-
al to the width of flange or on the bEsis of a load which gives the
same stress in the steel. In Table 7 are given the deflections of
the beams for loads which give a stress of 35000 lb. per sq. in. in
the steel, as calculated by the formula, M= .86 A f d. The
results average .27 in. for the 16-in. flange, .27 in. for the .24-in.
flange, and .32 in. for the 32-in. flange. The decrease in stiffness
for the wide flange, whlile apparent, is not very great.
23. Position of Neutral Axis.-Little can be said here con-
cerning the position of the neutral axis except that the data and
analysis go to confirm the statement that for ratios of width of
flange to width of stem not exceeding those used in the experi-
ments, the position given by considering the T-beam as a rec-
tangular beam of the full section of the inclosing rectangle is
 TALBOT-TESTS OF REINFORCED CONCRETE T-BEAMS

sufficiently close for purposes of ordinary calculation. The stress-

deformation diagrams for the beams on which observations were
made, so far as these results may be used, also corroborate this
view. As the compressive stresses in T-beams will generally be
low, the error in any use of the assumed inclosing rectangle likely
to be made will be small enough to be neglected, especially if we
TABLE 7.
DEFLECTIONS FOR A STRESS OF 35 000 LB. PER SQ. IN. IN THE STEEL.
M =.86 A f d.

Beam Flange Per cent Applied

Width Reinforce- Load Deflection
No.
No. inches ment pounds inches
1 16 1.05 25200 0.24
4 16 1.10 26500 0.25
7 16 1.10 26500 0.31
Av. 0.27
3 24 0.93 33600 0.25
6 24 0.92 33200 0.27
8 24 0.92 33200 0.28
Av. 0.27
2 32 1.05 50500 0.31
5 32 1.05 50500 0.31
9 32 0.97 46700 0.34
Av. I 0.32

limit the use to the condition that the flange shall extend two-
thirds of the distance to the neutral axis.
24. Applicability of Results.-The tests are not numerous
enough or sufficiently diversified to show that the results are
generally applicable to beam construction. It seems clear, how-
ever, from the general behavior of these beams that for calcula-
tions on strength the T-section may be considered to be the
equivalent of rectangular beams of the size of the inclosing rec-
tangle for widths of flange equal at least to four times the width
of stem. It seems probable that this relation may be applicable
to even greater widths of flange. However, the actual value
of this limit cannot be of great practical importance, since a
greater width would not materially change the value of the cal-
culated resisting moment (considering it based upon tensile
strength of steel and moment arm measured to center of com-
pressive stress), and since the amount of steel will at any rate
 ILLINOIS ENGINEERING EXPERIMENT STATION

be limited by the space in the stem and by practical considera-

tions in placing and bedding it to an amount which will keep the
maximum compressive stress within a reasonable limit.
The method of calculating the tension in the stirrups is used
tentatively and is not regarded as a final method for use. It is
probable that experiments will show that a different method of
calculation should be adopted. The efficacy of stirrups of this
kind is well brought out, but at the same time, it should be noted
that the size, spacing, and bonding of the web reinforcement are
of larger capacity than is usually given. The fact that all the
beams, and especially the wide beams, escaped failure by diagonal
tension is of especial importance.
Nothing in the observed phenomena of the tests indicates
that there was an appreciable distortion from a plane cross
section even in the flanges of the beam. The tearing of the flange
in one beam was evidently due to unequal bending and seems not
to be traceable to shear or to variations from.a plane cross section.
25. Summary.-The following summary of the discussion is
given:
1. Beams of flange width of two, three, and four times the
width of stem or web and reinforced in each case with steel equal
to 1% of the inclosing rectangle exhibited in a common way the
characteristics of rectangular beams, and the critical failure in
every case came through the longitudinal reinforcement becom-
ing stressed beyond its yield point.
2. The full compressive strength of the concrete at the most
remote fiber was not developed at the yield point of the beam,
even in the beams which were reinforced with steel of 54000 lb.
per sq. in. yield point.
3. The beams with the wide flanges were deflected some-
what more than the narrower beams, as may be expected from
the lack of full width of concrete on both the compression and
the tension sides of the neutral axis, the deficiency affecting the
stiffness of the beam but having little effect upon its strength at
uoints of maximum bending moment.
4. The vertical stirrups used proved to be very effective
web reinforcement. The diagonal tension cracks appeared at or
above the loads at which failure by diagonal tension may be
expected in beams without web reinforcement. A high resistance
 TALBOT-TESTS OF REINFORCED CONCRETE T-BEAMS

to diagonal tensile stresses was developed, as measured by the

calculated maximum vertical shearing unit-stress, which in one
beam was 605 lb. per sq. in. Since no beam failed by diagonal
tension, the limit of strength of the web reinforcement was not
determined.
5. The observed phenomena of the tests give no indication
of distortion from a plane cross section and there was no indica-
.tion that the thin flange was an element of weakness. The tear-
ing of the flange of one beam after the -maximum load was
reached was due to causes which would not exist in a floor system
as usually constructed. It seems clear that the limit of useful
width of flange was not reached in the beams tested.
6. The maximum strength of T-beams to resist horizontal
tension and compression (flange stresses) may well be calculated
by using the ordinary methods and formulas in use for rectangulaz
beams and considering the inclosing rectangle of the T-beam to
be the equivalent rectangular beam. This approximation is at
least applicable for reinforcement not exceeding 1% of the inclos-
ing rectangle. It gives little error when the thickness of flange
is at least one quarter of the depth of beam and the width of
flange not more than four times the width of stem, and may be
used-for an even greater range without great error. The in-
clusion of a greater width of flange than four times the width of
stem would not materially change the calculated strength of the
beam, since the amount of steel which may be put into the stem
is usually limited by considerations which of themselves will hold
the compressive stresses within proper limits, and since the
moment arm of the horizontal couple will not change much with
an increase in width of flange. The web stresses, which here are
very important, will differ from those found for rectangular
beams, and for T-heams the actual width of stem must be used in
the calculations for vertical shear and diagonal tension.
 ILLINOIS ENGINEERING EXPERIMENT STATION

Jv i l . . . . . . . . .I' l . .I1l l1

S I

-- WeIj
1
1 1 3 - 1 1 11 1 1
Q
ili iiIM ill")
9
t 1 1I 1 1 It 11 11 I1 1 1 1 1 1 r
I I' I I II El I I I I I I I I

IQ
2)
Z^
n-r-tSt I I I i - l I I I I . . I I I. 1 i

I
d 1 I 1I 1I 1 1
a~ Mo tu
L
F 6 7
Si i l i i i I
7-BP•7/f A• J

i i 'I'lll'
SF
m--

ii i ii I II I I I I I III
%

N
31J660
I
I I ý TI I T
II I
I -
*M -
..........
- -
1
ý1
I- IT I- -

ezI
It-i
tF -:LLLLLIHHIV1
ii lull
liii
N 111111 I
"I
8 F/8- . 9
Q. T-B~AAf A'."4 7-£P4A'A'/ 7
/6-,,v Pr ~4'VG~
/ff-/w. ^t.di#vee
I
IIIlllll

DEIPL C7T/0/V /A/V

IAI/CHS
 TALBOT-TESTS OF REINFORCED CONCRETE T-BEAMS

1 1 1 1 1 1 1 1 dmA-d6Q44o4 " i I 1 11

'. My""
. .. . . . . .... . .. . . .
I I I. I - . i i i ilI i i ii "1 11
i i :

,.J0Wr
'4

F/6./O
lea.
7-d•BL N,• -, 6'
2-.-/,. Fzdwc/vG

•l,•lr•
| •,€ •'/4¥ ,4"15-V/¥M

T-T Fl~ffi ,_ýl l

_+--
. ... ..
•. . . . ... .. . . . .. .

I II I i IA IJ. I 1

-,-
2' I
- I
•J

'2.I iii I11I11I 1 11 1 i l l i1

Ii'
•T
N
.. .. I .. I . . .. . . . . . . . . . . . .;

* /1.// 1D
1.12
7-Be," IVP 7- 16-141-f/V2 I
ý14-I rILAWEs R4-/At rkxA'Ge
lomm" oowada
PI cap""rm'r 9a'

7
L)FL C 70/bO /11 /INCH/AJ
 ILLINOIS ENGINEERING EXPERIMENT STATION

T TTr"rY1
^ M
-

1 1 1
1
'

I I
J
f
tl
m
//6. /3

TpBf4/AA /V0 9f
I1 - II II I I I I
JA/d/AAid''0VSE,2

t2ii Vrn
•LlJJl I d
l m - I mi
l l s i
l l i
I

111i Ul''07 i t

Il l .
I I I T~r~~r1r1
I Ill I I I I I I I I
m III
I
11111
I I I I I

III"
III 111111111 IIII*I I
Illi l

4(/ffff :11.1
I l l i II l i a l i l I I
S I I I I I I. I I I I . .r .
I

Z. %996f6 *444--4
n ':
I I?! I I I

i l I..
I I I I IIIII..i,,.,,,,

. . . . . . .
jill liii j I
I I I
s.J

I I 1 1I J1il l I 11111 I I
-IiI I I I I I I I I *IIIIIII~, I I I
1111111
I I

I I
ri II I I IIIIIIIi,.,..,,
1-A -1
11 1-I
y /000i f~,'$. /5
ld ra8/Y /vZ r-8t,4A',v2 5
4"/A-3-//. FZ dI6£ JZV# A2 dA'c~
_~- AF
Cweiu~re 8A 2'~'.pt's#rz'o &~'j
m I I I 1 5 I I

I IIII1I11

PC LCTION IV
DOFL //VCH- S
 TALBOT-TESTS OF REINFORCED CONCRETE T-BEAMS

PUBLICATIONS OF THE ENGINEERING EXPERIMENT STATION

Bulletin No. 1. Tests of Reinforced Concrete Beams, by

Arthur N. Talbot. 1904.
Circular No. 1. High-Speed Tool Steels, by L. P. Brecken-
ridge. 1905.
Bulletin No. 2. Tests of High-Speed Tool Steels on Cast
Iron, by L. P. Breckenridge and Henry B. Dirks. 1905.
Circular No. 2. Drainage of Earth Roads, by Ira 0. Baker.
1906.
Bulletin No. 3. The Engineering Experiment Station of the
University of Illinois, by L. P. Breckenridge. 1906.
Bulletin No. 4. Tests of Reinforced Concrete Beams, Series
of 1905, by Arthur N. Talbot. 1906.
Bulletin No. 5. Resistance of Tubes to Collapse, by Albert
P. Carman. 1906.
Bulletin No. 6. Holding Power of Railroad Spikes, by Roy
I. Webber. 1906.
Bulletin No. 7. Fuel Tests with Illinois Coals, by L. P.
Breckenridge, S. W. Parr and Henry B. Dirks. 1906.
Bulletin No. 8. Tests of Concrete: I. Shear; II. Bond, by
Arthur N. Talbot. 1906.
Bulletin No. 9. An Extension of the Dewey Decimal System
of Classification Applied to Engineering Industries, by L. P.
Breckenridge and G. A. Goodenough. 1906.
Bulletin No. 10. Tests of Plain and Reinforced Concrete
Columns, Series of 1906, by Arthur N. Talbot. 1907.
Bulletin No. 11. The Effect of Scale on the Transmission of
Heat through Locomotive Boiler Tubes, by Edward C. Schmidt
and John M. Snodgrass. 1907.
Bulletin No. 12. Tests of Reinforced Concrete T-beams,
Series of 1906, by Arthur N. Talbot. 1907.
 4'~ ~
~A ~ ~

~
~?w 22
4, 4,
1
" A ~

VA
I' ~ -i
4
y~ 2~-
A7

A YUNIVERSITY op ILLINOIS

THE UNIVERSITY INCLUDES THE

4,
'A
COLLEGE OF LITERAI URE AND ARTS (Ancient and Modern
'Languags and Literatures, Philosophical and Political Sci-
hA
ence Groups of Studies, kcononics, Cpmmerce and Industry)
A COLLEGE OF ENGINEERING (Unexcelled library; spacious *4
buildings; well-equiipped laboratories and shops. Graduate 4"
4/
~ '~,
and Undergraduate courses in Architecture; Architectural
Engineering; Architectural Decoration; Civil Engineering;
Mun'icipal and Sanitary Engineering; Electrical Engineering;

<I
Mechanical Engineeing; Railway Engineering).
COLLEGE OF SCIENCE (Astronomy, Botany, Chewistry, Geology,
Mathematics, Physics, Physiology, Zo6logy).
A
COLLEGE OF AGRICULTURE (Animial Husbandry, Agronomy, 'A
Dairy Husbandry, Horticulture, Veterinary, Science, House-
old Science),.
COLLEGE- OF LAW (Three years' course).
COLiLEGE OF MEDICINE (College of Physicians and Surgeons,
44 v
Chicago). (Four years' course).
COLLEGE OF DENTISTRY (Chicago). (Three years' course). '4,
4
VV V
V ~VV

SCHOOLS-MfUSIC (Voice, Piano. Violin). LIBRARY SCINCE.

 '-,7- )'',,,.

^.4 •"\ ^''4

)' •
4% V 7
. '4^
^^ "lf
4 4 '^

44.4

.7<
'.4.4 .4 -
14,4 ' '<4

4 4'

.4.

* A,
.4 4, ii•,••!•i!•ii
•¸
S , ".4%
t ..
4
4 4
'.4 4,^
,44 ~< 44'

I
.4 7,, ~.44 4~4 '4 .44 4
'44 &
•
••i•ii•ii,•b•
'44

44 .4

-+
A /