ILL INO I S

UNIVERSITY OF ILLINOIS AT URBANA-CHAMPAIGN

PRODUCTION NOTE
University of Illinois at
Urbana-Champaign Library
Large-scale Digitization Project, 2007.
 ¾

, TESTS- OF COLUMNS:
AN IINVESTIGATION OF TH181 VALU&
CONCRtTE AS REIN FORCEMENT FOR;
SFSTRUCTURAL A1TEEL COLUMNS;
 UNIVERSITY OF ILLINOIS
ENGINEERING EXPERIMENT STATION
BULLETIN No. 56 MARCH 1912

TESTS OF COLUMNS: AN INVESTIGATION OF THE

VALUE OF CONCRETE AS REINFORCEMENT FOR
STRUCTURAL STEEL COLUMNS
BY ARTHUR N. TALBOT, PROFESSOR OF MUNICIPAL AND SANITARY
ENGINEERING AND IN CHARGE OF THEORETICAL AND APPLIED
MECHANICS, AND ARTHUR R. LORD, RESEARCH FELLOW IN
THEORETICAL AND APPLIED MECHANICS

CONTENTS
I. INTRODUCTION
PAGE
1. P relim inary ............... . ........... ............. 3
2. Scope of Bulletin......... .... .. ....... . 4
3. Acknowledgment................................... 5
II. MATERIALS, TEST PIECES, AND METHODS OF TESTING
4. The Steel Columns.......... . . ................. 5
5. Sections of Columns.................... ............. 6
6. Cement and Aggregates .......... ................ 7
7. Concrete......................... ... .. ............. 7
8. M aking of Columns.................................. 8
9. Auxiliary Test Specimens............... ............ . 9
10. Storage and Handling......... ..... ............... 10
11. Method of Measuring Deformations.................... 11
12. M ethod of Testing................. ....... .. ...... . 12
III. EXPERIMENTAL DATA AND DISCUSSION
13 . D ata ................ ............ .. ............... 17
A. PLAIN STEEL COLUMNS-EFFECT OF LENGTH

14. Phenomena of the Tests ...................... . .. 17

15. Stress-deformation Relations.............. ........ . 18
16. Relation of Strength to Length...... .................. 19
B. CORE TYPE-EFFECT OF LENGTH
17. Phenomena of the Tests ........................ . . 23
18. Stress-deformation Relations.................. ....... 25
19. Effect of Test Conditions........................ . . 25
20. Effect of Length.................................. 25
21. Comparison with Plain Steel Columns.................. 27
 C. CORE TYPE-VALUE OF CONCRETE

22. Test Phenomena and Stress-deformation Relations for

Various Mixtures..... ................. .......... 28
23. Basis for Determining Load Taken by Concrete....... 28
24. Development and Amount of Concrete Stress .......... 29
25. Comparison of Cube and Cylinder Strength with Column
Strength........................ ..... ........... 31
26. Values of E and n .... ............................. 32
D. FIREPROOFED COLUMNS

27. Phenomena of the Tests ............................. 34

28. Stress-deformation Relations.................. ...... 35
29. Comparison of Concrete Stresses on Gross Section and
on Core Section ................................... 35
30. Permissible Deformation as Governing Design......... 37
31. Need of a Tie to Prevent Stripping of Shell............ 37
E. SPIRALED COLUMNS

32. Phenomena of Tests and Stress-deformation Relations.. 37

33. Effect of Spiral............................ ......... 40
34. Availability of Spiral Strength for Design. ............ 42
F. SUMMARY
35. General Comments...................... ........... 42
LIST OF TABLES Page
1. Tensile Strength of Cem ent ............................. .................. . ............ 6
2. Proportionsof Concrete Ingredients...................................................... 8
3. Compression Tests of Cubes and Cylinders.................................. .............. 10
4. Gauge Lengths and Precision of Measurements............................................ 11
5. D ata of Test Colum ns......................................................... .............. 14
6. Maximum Loads Carried by Columns and by Steel and Concrete............. ............ 15
7. Loads Carried at Various Unit-deformations.............................................. 16
8. M oduli of E lasticity. ............ ....... ... .......... ........................ ....... ... 22
9. Average Strength of Specimens and Values of n.................... .................. 33
10. Stress in Concrete in Shell and in Core..................................................... 35
LIST OF FIGURES
1. Details of Test Colum ns...................... ....... .. ... ... ............................. 4
2. View of Bearing Block and Attachment of Instruments................................... 16
3. View Showing Core Type of Column beyond the Maximum Load......................... 16
4. View Showing Fireproofed Column beyond the Maximum Load.......................... 17
5. View Showing Spiraled Column after Test....................... ................ ....... 17
6. Load-deformation Diagrams for Plain Steel Columns ..................................... 18
7. Average Deformation Diagrams for Plain Steel Columns of Various Lengths ........... 19
,8. Load-length Diagrams for Plain Steel Columns............ ............................... 20
9. Value of Terms in Straight-line Formula for Plain Steel Columns......................... 20
10. Column Formulas for Maximum Load.............................. ...................... 21
11. Load-deformation Diagrams for Core Type of Column ............ ....................... 24
12. Load-length Diagrams for Core Type of Column.......... ........................ 26
13, Value of Terms in Straight-line Formula for Core Type of Column..................... 26
14. Load-length Diagram in Terms of the Steel Section for Core Type of Column .......... 27
15. Load-deformation Diagrams for Columns with Lean and Rich Concrete ................ 28
16. Comparison of Stress-deformations in Column Concrete and in Cylinders................ 30
17. Comparison of Strength of Concrete in Cubes. Cylinders and Columns.. ............... 32
18. Load-deformation Diagrams for Fireproofed Columns..... ...................... ....... 34
19. Load-deformation Diagrams for Spiraled Columns,-First Test.......................... 36
'.20. Load-deformation Diagrams for Spiraled Columns,-Second and Third Tests .. ......... 38
 TESTS OF COLUMNS: AN INVESTIGATION OF THE
VALUE OF CONCRETE AS REINFORCEMENT
FOR STRUCTURAL STEEL COLUMNS.
I. INTRODUCTION.

1. Preliminary.-In reinforced concrete building construc-

tion columns form an important element, and in the case of very
high or very heavily loaded buildings, the size of the columns and
the space they occupy become important considerations. Various
types of reinforced columns are in use. Columns with longitudi-
nal reinforcement and hooped columns are common. During the
past few years designers have used structural steel columns en-
cased in concrete. Sometimes the structural steel shapes form a
relatively small proportion of the column section and are con-
sidered as reinforcement for the concrete. In other designs the
amount of steel is much larger and the structural shapes will
carry a large proportion of the load so that the column instead of be-
ing a reinforced concrete column is really a steel column reinforced
with the concrete in which it is embedded. Such columns may
occupy less space than the reinforced concrete column as usually
designed.
Two points of view seem to exist with reference to columns
having a large percentage of structural steel: (a) that the con-
crete surrounding the steel simply affords protection from fire
and corrosion and that the additional strength afforded by the
concrete is not considerable in amount and is not available for
design; and (b) that if the concrete be present it must act in uni-
son with the steel and that its strengthening effect and its effect
upon the permissible deformation of the column should be taken
into account. The present building codes either directly or through
the relation of stresses allowed virtually occupy the first position
when the steel column forms more than 8% of the column section.
The series of tests described in this bulletin was planned to
throw light upon the action of columns formed of structural steel
shapes by filling the space between the shapes with concrete or
encasing them in concrete as exemplified in a form of column
section which has been used in reinforced concrete building
construction. It is hoped that the results will be helpful in dis-
cussing fundamental principles underlying the design of such
columns.
4 ILLINOIS ENGINEERING EXPERIMENT STATION

2. Scope of Bulletin.-The investigation was planned with

the view of securing information on the following principal points
for the section and type of column tested: (1) the effect of length
and slenderness on the strength of the plain steel columns; (2)
the effect of length upon the strength of similar columns made
up with a core of concrete; (3) the effect of richness of concrete in
the core filling upon the strength of the column; (4) the effect of
adding an exterior coat around the steel section upon the strength
of the column, and the action of this coat under load; (5) the
effect of spiral hooping upon the strength and stress-deformations
of the column.
The tests were planned in an effort to obtain the most in-
formation on these points with the 32 columns available for the

pe

5,r

FIG. 1. DETAILS OF TEST COLUMNS.

 TALBOT-LORD-TESTS OF COLUMNS

tests. The results may not be applicable to sections or types of

columns which differ from the form used.
3. Acknowledgment.-The steel columns used in the tests
were furnished through the courtesy of the Illinois Steel Company,
Mr. E. J. Buffington, President, and were made at the North
Works of the Company in Chicago. The spirals 'for the six spi-
raled columns were furnished by the American System of Con-
crete Reinforcing, of Chicago. The work of concreting the col-
umns was directly supervised by Mr. D. A. Abrams, Associate in
the Engineering Experiment Station of the University of Illinois.
The tests were made by Mr. Lord with the assistance of Mr. R.
K. Steward and members of the staff of the Engineering Experi-
ment Station. Mr. Lord is responsible largely for outlining the
scope of the tests and for working up the results. Mr. W. A.
Slater, First Assistant in the Engineering Experiment Station,
has given helpful assistance in putting the bulletin into final
form. Especial acknowledgment is made to Professor Frank P.
McKibben and the Department of Civil Engineering of Lehigh
University for the work of testing the four spiraled columns sent
to the Fritz Engineering Laboratory and for the great interest
and care so freely given.

II. MATERIALS, TEST PIECES, AND METHODS

OF TESTING.

4. The Steel Columns.-The structural steel columns used

were of the Gray type and were composed of eight angles, 3 x 21
x T-in., tied at intervals of 16 inches by 5 x -in. plates as
shown in Fig. 1. This type of column has been used in the con-
struction of reinforced concrete buildings and has proved very
satisfactory. The size of angles, radii of gyration, dimensions of
tie plates, slenderness ratio of the flanges between tie plates,
radius of gyration of column and slenderness ratio of columns
give relations which are in many ways comparable to column
sections which have been used in reinforced concrete building
construction.
The method of fabricating these columns was in no sense
unusual. The ends, although milled, did not present a true bear-
ing as they were received in the laboratory. The steel was open
hearth structural steel. Tension tests of 11 x Ax 18-in. speci-
mens cut from an untested flange gave an ultimate strength of
62000 lb. per sq. in. and a yield point of 39800 lb. per sq. in.
6 ILLINOIS ENGINEERING EXPERIMENT STATION

The ultimate compressive strength of flanges composed of two

3 x 21 x A-in, angles riveted back to back averaged 39700 lb. per
sq. in. for two specimens 32 inches long, and 33 200 lb. per sq.
in. for two specimens 7 ft. 10 in. long.
5. Sections of Columns.-Table 5, page 14, gives the schedule
of the columns tested. Ten steel columns were tested without
concrete reinforcement; these columns are called "plain steel
columns" in the schedule and in the discussion. For studying
the effect of concrete in connection with the steel, the space in-
closed within the outline of the structural shapes, as shown in
Fig. 1, was filled with concrete. This combination of structural
shapes and concrete will be termed the "core type of column".
It was adopted as the principal form of test piece because it was
believed wisest to obtain the larger part of the data with the
section which is considered to be the effective section in design.
For purposes of comparison three columns were made with 2
inches of concrete outside of the steel (see Fig. 1), and the action
of this outer shell under load was studied. This combination of
steel and concrete is here termed the "fireproofed type." The
effect of richness of concrete upon the strength of the core type
was sought by the testing of columns with 1-1-2 and 1-3-6 mix-
tures in addition to the 1-2-4 mixture. In six columns the steel
was inclosed within a wire spiral and the space filled with con-
crete to the outer face of the spiral. The spirals were 14 in. in
diameter, were of i-in. steel wire with a pitch of 2 in. and 1 in.,
respectively, the percentages of spiral reinforcement used being
0.75 and 1.0. The percentage of the structural steel section in
terms of the whole column area varied, being 10.8 % for the core
type, 6.1 % for the fireproofed type and 8.5 % for the spiraled
columns.
TABLE 1.

TENSILE STRENGTH OF CEMENT.

Age 7 Days Age 28 Days

Ref. 1:3 1:3

No. 1:3 Sand 1:3 Sand
Neat Standard Used in Neat Standard Used in
Sand Columns Columns
Cand

1 589 198 265 674 278 323

2 684 227 709 283

3 653 240 731 319

 TALBOT-LORD-TESTS OF COLUMNS

6. Cement and Aggregates.-The cement used was furnished

by the Universal Portland Cement Company. Tests of samples
taken at times through the season and made by B. L. Bowling,
Assistant in charge of the Cement Laboratory, are given in
Table 1. Sample No. 1 was taken October 14, No. 2 November 22,
and No. 3 January 15. In the fineness test 98.5 per cent passed
No. 50 sieve, 96.5 per cent passed No. 100 sieve, and 82.5 per cent
passed No. 200 sieve.
The sand used was torpedo sand from Attica, Indiana. It
was of good quality, fairly sharp, clean and well-graded. It
combined with the cement used in a very satisfactory manner
and gave a higher briquette test than did the same cement with
standard Ottawa sand. It was from the same locality and of the
same quality as the sand used in making reinforced concrete test
specimens for several years at the University of Illinois.
A good quality of rather hard limestone from Kankakee,
Illinois, specified to pass through a 1-in. screen and over a I-in.
screen, was used. It is representative of the stone most used in
building construction of reinforced concrete in Illinois and it was of
the grade which has been used in the previous experimental work
of the Laboratory. In the columns tested the failure did not ap-
pear to result from the crushing or breaking of the stone in any
way.
7. Concrete.-Table 2 gives the proportions of the materials
used in the different batches of concrete from which the columns
were made. The sand and stone were first measured by loose vol-
ume. A bag of cement (95 lb.) was considered as 1 cubic foot of
cement. The resulting proportions by weight are given in the
table.
Men skilled in mixing concrete and making test pieces were
employed in the work. The foreman and the other workmen are
experienced concrete workmen; they have made the specimens
for the laboratory for six seasons. The mixing was done with
shovels by hand. The sand and cement were first mixed dry;
the stone, which had previously been thoroughly moistened, was
added and the mix then turned until of a uniform appearance.
Usually the first operation included five or six turnings and the
second three or more. Water was added in sufficient quantity to
produce a distinctly wet mixture which would run rapidly from
the shovel. The whole was then turned until thoroughly mixed.
8 ILLINOIS ENGINEERING EXPERIMENT STATION

8. Making of Columns.-The forms for the core type were ex-

tremely simple inasmuch as the concrete was confined to the
octagon determined by the edges of the steel flanges. Four planks
of the correct length were placed in a vertical position direct-
ly against the flanges and were held in position by means of
yokes. For the fireproofed type octagonal forms of wood were
built to give the required 2 inches of clearance over all faces.
For the spiraled type circular metal forms were placed directly
against the wire spiral and held in place by bands as shown in
Fig. 1 of Bulletin No. 20 of the University of Illinois Engineering
Experiment Station.
TABLE 2.
PROPORTIONS OF CONCRETE INGREDIENTS.

By Volume By Weight
Column

Cement Sand Stone Cement Sand Stone

8907 1 2 2.16 3.55

8908 1 2 2.05 3.55
8912 1 2 2.16 3.55
8913 1 2 2.05 3.55
8917 1 2 2.11 3.35
8918 1 2 1.88 3.62
8922 1 2 2.09 3.42
8923 1 2 2.04 3.54
8925 1 1 1.06 1.81
8926 1 1 1.04 1.76
8927 1 3 3.14 5.43
8928 1 3 3.02 5.30
8929 I 2 2.04 3.52
8930 1 2 2.13 3.60
8931 1 2 2.07 3.60
8933 1 2 2.08 3.61
8934 1 2 2.02 3.51
8935 1 2
8936 1 2 2.02 3.56
8937 1 2 2.04 3.52
8938 1 2 2.08 3.67

The concrete was placed by pouring it into the forms at the

top of the column a bucketful at a time. It was worked around the
sides and in the center by means of a pole. As an aid in secur-
ing uniform concrete successive bucketfuls were taken from dif-
ferent portions of the pile. The forms were filled practically
level with the top of the steel.
Previous to pouring the concrete, the steel column was
placed in a vertical position on a 14 x14 x 1-in. cast-iron bearing
plate (upon which it was later tested) and the forms were placed
in position. The day after pouring, after the concrete in the
column had had time to shrink, the top of the column was pre-
 TALBOT-LORD-TESTS OF COLUMNS

pared for testing by setting another bearing plate upon it, using
a neat cement mortar as a bed. Care was taken to have this
plate rest equally upon the four flanges and also to bear on the
mortar. The layer of mortar between the plate and the steel was
kept as thin as possible. A film of cement over the end of the steel
was necessarily present; but in most cases it was extremely thin
and in no case did this thickness reach i inch. In pouring the spi-
raled type it was found that the concrete did not flow into the
small space at the back of the flanges between the wires of the
spiral, and it was necessary to grout into this space with a sand
mortar. In columns of the fireproofed type the concrete outside
the flanges was found to be less dense and in some cases the sur-
face over the flanges was more or less pitted. As has always
been the practice in the preparation of concrete specimens in the
laboratory, the fabrication was done seriatim, one specimen of a
kind well through the series, followed later with the fabrication
of the second specimen seriatim. By this means any accidental
variation of cement or aggregate or of temperature conditions
would be likely to have the same effect on one type of specimen
as on another, whereas if all the specimens of the same kind
were made at the same time the presence of such variations might
be mistakenly considered to be caused by variations in type.
This practice is explained at some length here, because consid-
erable variation was found in the concrete of specimens of the
same type made at different times. The dates of making speci-
mens were as follows: On October 2, 1910, No. 8907 and 8912
were made from one batch of concrete and No. 8917 from another
batch. On October 29, No. 8922 was made from one batch and No.
8925 from another batch. On November 4, No. 8927 was made.
On November 8, No. 8908 and 8913 were made from the same
batch, No. 8918 from a batch, and No. 8923 from another batch.
November 12, No. 8926 was made and November 19, No. 8928 was
made.
9. Auxiliary Test Specimens.-From each batch of concrete
used three 6-in. cubes and one 8 x 16-in. cylinder were made from
which to determine the properties of the concrete in the columns.
These were stored in damp sand until a few days before the cor-
responding column was to be tested when they were removed to
the testing laboratory and two faces prepared for the test by the
addition of a thin coat of plaster of paris. It was originally in-
tended to test all these specimens at the same date as the cor-
responding column, but an unexpectedly low strength was found
10 ILLINOIS ENGINEERING EXPERIMENT STATION

in the cube tests of the first specimens made. One cube from
each of the later batches was then reserved for test at 90 days in
order to know whether the concrete was poor or whether it mere-
ly lacked curing. It was found that these later specimens gave
what may be considered to be normal strength, and hence no
light was thrown on the cause of the low strength of the concrete
from the first batches. The results of the cube and cylinder tests
are given in Table 3. These auxiliary tests show that the
strength of the concrete at the time of testing was far from uni-
form. The possible causes of this are discussed in a later
paragraph.

TABLE 3.
COMPRESSION TESTS OF CUBES AND
CYLINDERS.

Loads are given in pounds per square inch.

6-in. Cubes
Corresponding 8 x 16-in.
Column Cylinders
No. 60 days 90 days 60 days

8907 1790 1350

8908 2430 2650 1490
8912 1790 1350
8913 2430 2650 1490
8917 1420 1140
8918 2150 2600 1260
8922 1320 970
8923 1970 2640 1150
8825 2970 2420
8926 3280 4000 2520
8927 1320 1800 700
8928 1440 1740 660
8929 1760 2580 1370
8930 2020 2740 1330
8931 1670 2160 1280
8933 1440 2000 1160
8934 2070 2350 1340
8935 2270 2720 1330
8936 1780 2180 1110
8937 1620 2900 1460
8938 1980 2580 1120

10. Storage and Handling.-The columns were stored in the

room where they were made. Forms were removed at the end
of a week and from that time on the columns were occasionally
sprinkled. The records show that the room temperature varied
from 60' to 700 Fahrenheit, but it seems probable that a larger
variation may have occurred in different parts of the building.
It is also probable that the later columns dried out much more and
attained a higher percentage of their final concrete strength than
did the cubes at the same age.
 TALBOT-LORD-TESTS OF COLUMNS

Before removing the columns to the testing laboratory the

bearing plates at top and bottom were connected by rods to pre-
vent displacement, and these rods were not removed until the
column was in its final position in the testing machine.
11. Method of Measuring Deformations.-The extensometer
used and the method of attachment are shown in Fig. 2. For the
determination of deformations in the steel, holes were tapped in
the flanges at the lower point to receive the shaft of a wire-wound
dial extensometer and at the upper point to receive a bolt; from
this bolt a wire was suspended, wrapped once around the drum of
the extensometer and weighted with a nut at the end. The def-
ormation occurring in the gauge length between the bolt and the
TABLE 4.
GAUGE LENGTHS AND PRECISION OF
MEASUREMENTS.

Length of Gauge Least Unit-deformation

Column Length Measurable

2 ft.-0
in. 10 in. .00002
4 ft.-8 in, 40 in.* .000005

10 ft.-0 in, 100 in. .000002

15 ft.-4 in. 150 in. .000001

19 ft.-4 in. 200 in. .000001

*Average.

lower shaft was registered by the movement of the pointer of the

dial. Measurements were made on each of the four flanges, four
instruments being used for the purpose. Where measurements were
desired on concrete faces, specially prepared plugs were inserted in
the column during the pouring and these were tapped to receive
the bolts and the shafts of additional extensometers. The
wire suspended upon the upper bolts was in general 1 inch from
the face of the column, and the accuracy of the observations de-
pends upon the conservation of the plain section in the column as
a whole. The results obtained indicated that this condition was
not fully satisfied in the case of plain steel columns. An arrange-
ment of instruments which gave the deformations at the center
of gravity of the flanges direct was ;used in the test of flanges
and on column No. 8914. The result indicated that the error in
the first arrangement of instruments was not great. The gauge
lengths used varied with the lengths of the spetimens. The dials
could be read to an indicated movement of 0.0002 inch. The gauge
12 ILLINOIS ENGINEERING EXPERIMENT STATION

length and the least unit-deformation obtainable for the different

lengths of the columns are given in Table 4.
In the case of the plain steel columns 10 feet or more in
length, in addition to the four measurements of deformation taken
over the gauge lengths noted in Table 4, six other measurements of
deformation were taken over gauge lengths about one-third as
long. These shorter gauge lengths were located at different
portions of the columns in an effort to detect local bending. The
measurements did not show that material bending occurred at
loads below the maximum load, and if such bending occurred, it
was confined to shorter distances.
12. Method of Testing.-The columns were tested in the
Riehle vertical 600 000-lb. screw-power testing machine in the
Laboratory of Applied Mechanics of the University of Illinois.
It was appreciated in advance that the full strength of the
spiraled columns could not be developed in the 600 000-lb. test-
ing machine, but it was thought that there would be a sufficient
indication of their action to determine the critical yield-point.
However, after testing them it was concluded that a determina-
tion of the action of such columns at higher loads would be of
value, and further tests of four of the columns were made in the
Riehle 800 000-lb. testing machine at Lehigh University. In the
case of the plain steel columns a specimen was placed with its
lower end bearing directly on the weighing table of the testing
machine and centered with respect to the screws. The compres-
sion head of the machine was then lowered until the suspended
spherical bearing block rested on top of the column. This block
was then centered on the column. Although the ends of the col-
umns were milled in the shops, in some cases it was necessary to
file the ends to secure a satisfactory bearing and in some cases to
use carefully prepared steel shims. In the case of the con-
creted columns the bearing plates of the columns rested directly
on the weighing table. When the column was in its final position
in the machine, the rods connecting the end bearing plates, used
while transporting the column to the testing room, were removed
and the spherical bearing block was lowered into position as.
noted for the plain steel columns. An initial reading of the ex-
tensometer was taken with no load on the column except the
weight of the spherical bearing block. The compression head of
the testing machipe was then brought to bear on the bearing-
block and was run down at the slowest speed (0.05 inch per minuteý
 TALBOT-LORD-TESTS OF COLUMNS

until a load of 25 000 lb. was registered. The machine was stopped
at this load and after an interval of 30 to 45 seconds the ex-
tensometers were read. In like manner the load was increased
by increments of 25 000 lb. at the speed of 0.05 inch per minute,
with readings of the extensometers between applications of load,
until the maximum strength of the column was passed or the
capacity of the testing machine was reached. With the first five
columns tested (No. 8905, 8906, 8910, 8915, 8920) no effort was
made to restrain any movement of the spherical bearing block.
In the remaining tests, when a load of 50 000 lb. had been applied,
special wedges were inserted and adjusted in the bearing block
to restrain it from a rolling movement. In the tests of many of the
columns the operation of the machine was continued after the
maximum load was passed, to determine the effect of a further
application of load and a further shortening of the column. This
was done in order that the critical section might be definitely
determined and the nature of the final failure observed.
Three conditions mentioned incidentally above merit more
complete discussion:
(a) It is evident that the speed at which the testing machine
is run is of importance and that a column will carry a higher
maximum load if the load be applied rapidly than it will under a
slow application. In actual use the load is not momentarily
applied but it is a dead weight and follows at once any shortening
or other movement of the column. In testing, the load should be
held a sufficient time to ensure that the material has attained its
full deformation. In the tests herein described the slowest speed
of the machine (0.05 in. per min.) was used in all cases. The in-
struments were read starting about 30 seconds after the indicated
load had been momentarily applied.
(b) The condition of the bearing surfaces at the ends of the
column has an important effect upon the load which the column
will carry. In a building a column receives its load from story
to story and especial attention is given to the bearing arrange-
ment at its base, but in a testing machine the load is applied
directly upon the end section. If the load is to be uniformly dis-
tributed, the bearings of the ends must be perfect. In these tests
care was taken to get a fairly good bearing, but it should be noted
that the load was not always uniformly distributed over the
section and that some of the flanges were more highly stressed
than others. The records of observed deformations of the various
gauge lengths in the different flanges show this effect. When
14 ILLINOIS ENGINEERING EXPERIMENT STATION

TABLE 5.

DATA OF TEST COLUMNS.

Area of A
Column Length 1 Gross Concrete Test
Descri
Description
No. ft. in. r Section Mixture days
sq. in.

8902 Plain Steel 2-0

8905 Plain Steel 4-8
8906 Plain Steel 4-8
8907 Core Type 4-8
8908 Core Type 4-8

8910 Plain Steel 10-0

8911 Plain Steel 10-0

8912 Core Type 10-0

8913 Core Type 10-0
8914 Plain Steel 10-0

8915 Plain Steel 15-4

8916 Plain Steel 15-4
8917 Core Type 15-4
8918 Core Type 15-4
8920 Plain Steel 19-4
8921 Plain Steel 19-4

8922 Core Type 19-4

8923 Core Type 19-4

8925 Core Type 10 0

8926 Core Type 10-0
8927 Core Type 10-0
8928 Core Type 10-0

8929 Fireproofed 10-0

8930 Fireproofed 10-0
8931 Fireproofed 10-0
8933 Spiraled* 10-0
8934 Spiraled* 10-0
8935 Spiraled* 10-0
8936 Spiraled* 10-0

8937 Spiraledt 10-0

8938 Spiraledt 10-0

*0.75 %of spiral reinforcement.

t 1.0 * of spiral reinforcement.

the load is unevenly distributed over the section the tendency

towards bending is increased. The effect of poor end conditions is
more serious in short columns than in long ones; in the 19 ft. 4-in.
columns its effect is probably negligible. In studying the effect
of length on column strength the effect of end conditions must be
borne in mind. Columns No. 8902 and 8914 had the best end con-
ditions of any tested. In the 2-ft. column (No. 8902) a test was
found to be impracticable without carefully turning the ends in a
lathe.
 TALBOT-LORD-TESTS OF COLUMNS

TABLE 6.

MAXIMUM LOADS CARRIED BY COLUMNS AND BY STEEL AND CONCRETE.

To Load in Pounds Pounds per square

inch of Section
Column Load Load
No. nsidered Considered Steel Concrete
s Carried as Carried eel Concrete
y Steel by Concrete

89032 487 300

8905 440 200 33 700

8906 449 000 34 700

8907 577 000 444 600 132 400 34 200 1240

8908 602 000 444 600 157 400 34 200 1470

8910 410 400 31 600

8911 425 600 32 700

8912 510 000 418 000 92 000 32 150 860

8913 584 700 418 000 166 700 32 150 1560

8914 424 000 32 600

8915 398 000 28 300

8916 376 000 28 900

8917 468 500 372 000 96 500 28 600 900

8918 532 200 372 000 160 200 28 600 1500

8920 374 200 28 800

8921 345 000 26 500

8922 491 400 359 600 131 800 27 650 1230

8923 495 500 359 600 135 900 27 650 1270

8925 636 000 418 000 218 000 32 150 2040

8926 655 000 418 000 237 000 82 150 2210

8927 516 000 418 000 98 000 32 150 920

892$ 530 500 418 000 112 500 32 150 1050

8929 600 000*

8930 630 700 418 000 212 700 32 150 1060
8931 635 700 418 000 217 700 32 150 1090

8933 600 000 Maximum load applied five times: not broken.
8934 600 000 Not broken.
8935 600 000 Maximum load applied three times; not broken.
8936 625 000 Not broken.

8937 600 000 Maximum load applied three times; not broken.
8938 600 000 Not broken.

8934 856 000 Second test; near ultimate.

8935 830 000 Second test; near ultimate.
8936 600 000 Second test; not broken.

8937 830 000 Second test; not broken.

8938 827 000 Second test; not broken.

8937 714 000 Third test with spiral and outside concrete removed.

*.ot broken.

(c) The conditions of the end restraint will affect the curve
taken by a column. The lower end, bearing on the unyielding
weighing table of the testing machine, is quite firmly restrained un-
til serious bending occurs in the column. The upper end is loaded
through a spherical bearing block in order to secure adjustment
16 ILLINOIS ENGINEERING EXPERIMENT STATION

with the compression head of the testing machine. With this

arrangement, this adjustable bearing block may finally move on
itself as the tendency to bend in the column overcomes the friction
between the surfaces of the spherical bearing block, and thus only
partial restraint will exist. In the first five plain steel columns
tested (No. 8905, 8906, 8910, 8915, 8920) no effort was made to pre-
vent this movement, and the column was in the condition of one
end fixed and one end partly restrained. Neither the fixedness nor
the freedom of the ends can be considered as in any sense abso-
lute; they must be taken as relative terms. In all the column
tests after the five noted above, special wedges and angle blocks
were driven under the upper head (at a load of 50 000 lb.), the
TABLE 7.

LOADS CARRIED AT VARIOUS UNIT-DEFORMATIONS.

Loads are given in thousands of pounds for the unit-deformation

given in the column caption and in the last column for the maximum
load applied.

Column .00020004 .0006 .0008 0010 ,0012 .0014 .0016 0018 .0020 0030 0040 Max.
No. I 00 I .0I0 .1 I Load

8905 83 160 229 290 341 379 405 422 440

8906 83 161 235 300 355 396 421 449
8907 115 206 286 360 419 467 515 557 577
8908 120 216 311 390 465 529 565 592 602
8910 82 158 226 286 338 372 393 410
8911 82 155 221 281 335 377 402 426
8912 121 207 286 355 414 455 510
8913 123 224 319 390 461 516 548 585
8914 80 160 233 300 355 390 410 419 423 424
8915 84 154 216 271 319 352 368
8916 86 159 221 276 321 353 373 376
8617 109 193 268 336 393 432 468
8918 118 208 288 363 429 482 509 532
8920 80 150 211 262 308 349 374
8921 84 156 214 265. 310 345
8922 115 199 269 331 387 491
8923 114 206 285 358 427 477 495
8925 132 240 336 426 505 568 604 624 636
8926 119 226 320 405 486 553 591 619 C35 647 655
8927 107 190 265 334 397 447 480 500 516
8928 105 187 264 335 399 452 486 500 530
8929 141 257 357 452 534 594 6001
8930 117 218 310 398 474 536 581 611 631
8931 144 254 347 424 492 546 585 610 626 636
8933 126 231 322 409 489 550 588 600t
8934 134 235 321 394 459 511 545 568 588 600t
8935 136 242 338 430 516 578 600 600 600t
8936 129 234 324 406 483 541 579 625+
8937 135 242 338 427 508 567 600 600t
8938 124 230 326 416 491 543 581 6001
8934* 112 220 325 430 525 625 695 715 737 755 808 835 856t
8935* 100 200 300 410 525 640 690 720 743 761 815 830,
8936* 110 200 290 375 460 540 592 600:
8937* 96 194 295 404 520 628 680 710 733 753 815 830:
8938* 110 218 322 430 540 653 705 730 750 768 820 827ý
8937+ 111 225 340 453 567 656 700 712 714 714

*Second test: all second tests made at Lehigh University except column No. 8936.
tThird test; made at Lehigh University.
'The column did not fail under maximum load applied.
 TALBOT-LORD-TESTS OF COLUMNS

spherical bearing block having had opportunity to adjust the

bearing satisfactorily by this time and no appreciable tendency to
bend having yet developed in the column. In this manner the
spherical bearing block was restrained from further motion and
became in effect as rigid a loading block as the base. Fig. 2,
facing p. 16, shows the arrangement used in the later tests. That
the apparatus accomplished the desired purpose is evidenced by
the fact that failure in bending occurred almost invariably at or
near the center of the column. It may be noted in this connection
that the conditions of restraint in the testing machine are probably
not as near fixed-end conditions as are those of a reinforced con-
crete building.

III. EXPERIMENTAL DATA AND DiscuSSION.

13. Data.-In Table 6, page 15, are given the maximum loads
carried by the test columns. The loads carried at various unit-
deformations (interpolated between readings) are given in
Table 7. The load deformation diagrams are given in Fig. 6,
7, 11, 15, 18, 19, and 20. Complete data for all tests are on file
in the Laboratory of Applied Mechanics of the University of
Illinois, but as there were two hundred or more readings of
deformations for most of the tests, only the summarized data are
given in this bulletin.
The discussion of the tests will be made under the following
heads: A. Plain Steel Columns-Effect of Length; B. Core
Type--Effect of Length; C. Core Type-Value of Concrete;
D. Fireproofed Columns; E. Spiraled Columns; F. Summary.

A. PLAIN STEEL COLUMNS-EFFECT OF LENGTH.

14. Phenomena of the Tests.-The plain steel columns showed

test phenomena whic.i were consistently uniform. At loads of
from 225 000 to 250 Of0 pounds cracking sounds were heard and
these continued intei mittently throughout the remainder of the
test.* The time reqr red to add the 25 000-lb. increment of load
gradually became 1 nger with the testing machine running at
a uniform speed, and at the maximum load the weighing beam
floated within a ran!;e of 1000 lb. for a period of 10 or 15 minutes.
At maximum load no bending was visible to the eye except in
the longer columns and very little was shown by the deformation
readings. After the maximum load was reached and the machine
head was run down to complete the failure of the column
18 ILLINOIS ENGINEERING EXPERIMENT STATION

(generally at a faster speed), bending developed very gradually.

In general this bending was fairly symmetrical about the middle
of the length.
15. Stress-deformation Relations.--The load-deformation dia-
grams for the plain steel columns are given in Fig. 6. It is
seen that the load-deformation curves bend at low loads. This
may be due partly to imperfect end bearings of the steel shapes
and to the arrangement of instruments which presupposed the
conservation of a plane section in the column during the test.
Column No. 8914 whose ends were dressed quite carefully to a
fairly true surface and for which the extensometer was arranged
to avoid any effect of bending shows a straight line up to 175 000

500000

_^ /
____ /__
59068905
400000

Q)200000
/_
^300000
899/0 39/

I'00000

... 0

)30000

Q20006

DEFOPM 7T/ONV PER UNI T OF LENGTH

FIG. 6. LOAD-DEFORMATION DIAGRAMS FOR PLAIN STEEL COLUMNS.
 TALBOT-LORD-TESTS OF COLUMNS

lb. load and also an increase in stiffness over the other two col-
umns of the same length. Even in this case a marked bending
is present in the diagram well below the maximum load.
In Table 8, page 22, are given the secant moduli of elastic-
ity, calculated from the curves, for unit-deformations of .0004,
.0007, .0010, and for a point near the ultimate load. From these
computations, it would appear that the modulus of elasticity for
low deformations may be considered as lying between 30 000 000
and 31 500 000 lb. per sq. in. In Fig. 7 the average load-
deformation curves for the several lengths of column are also

ý0
>400000

S300000

200000

c• /00000
/OXO
C11oo
L/ °

7"Y PER
DfFORA/MT/o I//Y/T OF L 5YG T/-
FIG. 7. AVERAGE DEFORMATION DIAGRAMS FOR PLAIN STEEL COLUMNS
OF VARIOUS LENGTHS.

shown. It is apparent that the modulus of elasticity of the col-

umn as a whole decreases as the length of the column increases.
It is also seen that the unit-deformation in the column as a whole
at the maximum load becomes less as the length of the column
increases. This is an indication of localized high stresses in the
column.
16. Relation of Strength to Length.-In Table 6, page 15,
are given the maximum loads carried by the plain steel columns
of the several lengths.
Tests which have been made in the past on relatively small
columns indicate that for columns having a ratio of length (1) to
radius of gyration (r) less than, say, 100, the results may be ex-
20 ILLINOIS ENGINEERING EXPERIMENT STATION

pected to fall along a line slightly inclined to the -r axis. In ar-

riving at an expression for the effect of length upon strength it may
add to our clearness of perception to trace the development of the
effect of length upon strength at various stages of the test. In

ý 30000

'-4

L/enqrn In incnes
Fio. 8. LOAD-LENGTH DIAGRAMS FOR PLAIN STEEL COLUMNS.

Fig. 8 the results for the four lengths of column are plotted for
unit-deformations of .0005, .0008, .0010, and for ultimate load. The
four equations give the relation between- r and unit-stress at
these deformations. The equations show that the effect of length
upon stress is small at low deformations and is relatively much
higher at the higher deformations. In a similar manner results
were plotted for other unit-deformations from .0003 to .0012, and
from these data Fig. 9 has been prepared. In this figure values
P
off and k for the column formula .A-- f- k- are given for a
r
range of values of the unit-deformation. The lower curve also
shows the increasing effect of the slenderness ratio on the
strength as the test progresses. If we should produce the
tangent back to the horizontal axis the intersection is at .0004,
and it is not far from the facts to say that up to a unit-deforma-
tion of .0004 the slenderness ratio has no effect and that beyond
 TALBOT-LORD-TESTS OF COLUMNS

40000

4 30000

'40
20000

, /0000
4

Deforma-/ion per- i/n,' of Leng/ 0

FIG. 9. VALUE OF TERMS IN STRAIGH T-LINE FORMULA FOR PLAIN
P l
STEEL COLUMNS. Column formula: - f-
f -

this deformation it increases in a constant ratio to the increase

in deformation. From these diagrams it appears that a straight
line will represent the results very satisfactorily for any given
unit-deformation.
For the ultimate load, the nature of the strength-length re-
lation is somewhat affected by the fact that the ultimate general
unit-deformation of a column at its maximum load is smaller for
the greater lengths of column, as is shown in Fig. 7. While for
deformations less than the ultimate the straight-line equation

L enqth in /nces$
FIG. 10. COLUMN FORMULAS FOR MAXIMUM LOAD.
22 ILLINOIS ENGINEERING EXPERIMENT STATION

-ý A "

ci 0,
0- oc^ o'O oc)O OcO^ C SOCCiOOC f

o
a -- Q*
___
0 t1-»«1 ^ ^ -I1-
TI vv I~ ^ ^ -- -- _
-^-- _
-~NN ~ *|__

0.

~
§ ~ ~ C - - - - C . . 4)----------------------------------
.0,0

- C
03 00 CC,
- -
___________ 0

ca g = -- a~C)
oo 0o
0
C C) OMo, WI
CaC OT c OaMM-cC0»ccCa*-
¢-z 0 5
ca
C 0

0 __________________

- 0
e
Cý4 Cý 0. 0. 00 0. ~ 0!0 ýC ý ýC ý 0l

H0 _______________________________|

- b§
tc
d• 15-

C CC00C C C C CC 0)
o
.2 __a}O_____________
00
CC *g0 C a
C

C
o8 ~0) 0___________ ____________ *
-C
< Ch 2~ 00 4C C '.SC gtC

0^ -< ^ ^ o^ d^'^- o iCo

.0^
_| ___------------------- 0000 - 0 00 0 . . ~ - 0)
C-
00
av

00

CC Q CISC 000 C C C OO .000C

=W SC -%
_CC_ -C CC C I C j0CC

-- IS--C-CC--C----C-------------I

00* 0) OC O O O O O0 ~'5

---- @ a
----. ^ O" COO
C C C^ C C ^

ccC»
gZ~~~CO SS0a8 0gg CSSSCggCgSSSC g
 TALBOT-LORD-TESTS OF COLUMNS

alone appears to apply, at the ultimate it is possible to construct

an equation of the Rankine type which will represent the result
quite closely. In Fig. 10 such an equation and a straight-line
equation are plotted. The equations are as follows:
Straight-line: - = 36 500 - 155 1
A r
P - 35000
00
Rankine type:
A 1 12
1 12000 r 2
It appears that the straight line equation represents the test
results as closely as the Rankine equation. The ultimate load for
- 0 is higher for the straight-line equation than by the Ran-
r
kine equation, and both are lower than the stress of 37 500 lb. per
sq. in. carried by the test column two feet long and the stress of
39 700 lb. per sq. in. carried by the compression test pieces taken
from the flanges. After making a study of the data it is believed
that the straight-line equation given above may be considered
best to represent the effect of length upon ultimate load for the
plain steel columns tested.
B. CORE TYPE-EFFECT OF LENGTH.

17. Phenomena of the Tests.-The columns in which the core

only was filled with concrete acted in much the same way as the
plain steel columns. The concrete was somewhat restrained by
the structural shapes, but the capacity for carrying an increasing
load accompanied by the development of very high deformations
which has been found in hooped concrete columns was not pres-
ent. The columns exhibited much toughness and gave slow
failures.
All the columns were tested with the upper bearing block
restrained from motion, and in all cases the bending was symmet-
rical about the center'or it occurred below the center. Very little
bending was apparent at the maximum load. The bending shown
in the view in Fig. 3, facing p. 16, occurred some time after the
maximum strength of the column had been developed. In the
final failure of the column, at deformations well beyond the maxi-
mum load, the crushing of the concrete was frequently more
marked at the top of the column than elsewhere, due probably to
the smaller density of the concrete at this place. In the columns
4 ft. 8 in. long practically no bending occurred, and the columns
failed by general crushing which was more marked over the up-
per half. The columns 10 feet long, in the continuation of the
24 ILLINOIS ENGINEERING EXPERIMENT STATION

test beyond the maximum load, failed finally in bending, No 8912

bending about the center and No. 8913 bending sharply near the
base. The columns 15 feet 4 in. long and 19 feet 4 in. long, after
passing the maximum load, bent symmetrically about the center.
In the latter part of the test, the concrete crushed on one side of

DEFORMA7T/ON PER UNIT OF LEN6 TH

FIG. 11. LOAD-DEFORMATION DIAGRAMS FOR CORE TYPE OF COLUMN.

the column at the center and on the opposite side at the top and
bottom. The varying density of the concrete caused by the ob-
struction to settlement and shrinkage in setting offered by the
steel was shown in the final crushing. In all cases the crushing
 TALBOT-LORD-TESTS OF COLUMNS

was more marked at points at or immediately above the tie plates

in the steel columns than at points between these plates. Ap-
parently the concrete was less dense at the first-named places.
As noted above, and as was to be expected, the concrete at the
top of the column was weaker and less dense than that lower
down.
18. Stress-deformation Relations.-Fig. 11, page 24, gives
load-deformation diagrams for these columns. It was evi-
dent from the test results that in these columns the conservation
of plane section was well maintained during the test, and this con-
dition has been assumed in the interpretation of the data. The
load-deformation curves are similar to those of the plain steel col-
umns. In the diagrams in Fig. 11 the average load-deformation
curve for the two plain steel columns of the same length has also
been drawn.
19. Effect of Test Conditions.-The thickness of the mortar
cushion between the end of the steel column at the top and the
bearing plate may have exerted an influence on the strength of
the columns, but it is difficult to arrive at any estimate of the
amount of this. Where the joint was extremely thin, as was
generally the case, its effect was undoubtedly negligible, but in
one case where this joint was nearly i inch thick it may have
exerted an appreciable influence on the strength developed by the
column.
20. Effect of Length.-The maximum loads carried by the
columns are given in Table 6, page 15.
In the discussion of the effect of length of column the slen-
derness function may be expressed in terms of - where 1 is the
length of the column and d is the short diameter of the column
section. In a following paragraph the effect of the length will
be expressed in terms of - , where r is the radius of gyration of
the steel section.
In Fig. 12 are plotted the loads for eight columns for unit defor-
mations of .0005, .0008, .0010 and also for ultimate load. Owing
to variations in the concrete strength these points can not be ex-
pected to show as close agreement as did the plain steel columns.
It may be noted that for each length of column the specimen
showing greater strength was made on Nov. 8 and the weaker one
on an earlier date, the variation in the strength agreeing with
variation of strength of cubes and cylinders, as discussed else-
where. The straight lines on this diagram express fairly well
26 ILLINOIS ENGINEERING EXPERIMENT STATION

the relation between - and the average unit-stress over the

d
cross-section of the column for several unit-deformations and for
the maximum loads. P represents the load on the column and A
the area of the cross-section of both steel and concrete. The
P I
equation for the maximum load is seen to be -= 5150-52 -.
A d

5000
--- :£! _

k4000 .
_P. = C _s_ _
6-0008 __ =2C - 1Z714 ~
3000
q----- _ 3--20_-
'.ýL
S2000

lz^ Lenqth in Inches

FIG. 12. LOAD-LENGTH DIAGRAM FOR CORE TYPE OF COLUMN.

500e

IZ400a

3agw
Q.

:200C

/00a

Deformrnabon per Unif of Length

FIG. 13. VALUE OF TERMS IN STRAIGHT-LINE FORMULA FOR CORE
TYPE OF COLUMN. Column formula : P-= f-k k u
.A-
 TALBOT-LORD-TESTS OF COLUMNS

Fig. 13 gives values of f, the first term of the second member

in the straight-line column formula, and of k, the coefficient of
- , for the range of unit-deformations.
21. Comparison with Plain Steel Columns.-In order to obtain
a comparison of the effect of slenderness in columns of the core
type and in plain steel columns, Fig. 14 has been prepared. In
this diagram, for the purpose of comparison, the total load has

.U0000

*40000 =oo/g ____ 3 o

__ o.

*=.000 -4

0
2000
__ _ _ _ IE 202-5n
000

Len 9 m in /nches
FIG. 14. LOAD-LENGTH DIAGRAM IN TERMS OF THE STEEL SECTION FOR
CORE TYPE OF COLUMN.

been considered to be carried by the steel alone. A is the area of

the cross-section of the steel and r is its radius of gyration. By
comparing with Fig. 8 it is seen that the coefficients of - agree
r
very closely with the coefficients of 1 found in the tests of plain
r
steel columns at the same compression deformations. Thus at
ultimate load the coefficient of - is 155 for the plain steel col-
r
umns and 160 for the reinforced steel columns. This would indi-
cate that in the core type of column within the limits of length
tested, the effect of length upon strength of column is a function
of the slenderness ratio of the steel itself and is almost independ-
ent of the slenderness ratio of the concrete. This conclusion is
in accord with results of tests of long concrete columns made
in the Laboratory of Applied Mechanics, plain concrete columns
20 diameters long giving nearly as great strength as columns 10
diameters long. The discussion of the amount of stress taken by
the concrete is given in a later paragraph.
28 ILLINOIS ENGINEERING EXPERIMENT STATION

C. CORE TYPE-VALUE OF CONCRETE.

22. Test Phenomenaand Stress-deformationRelationsfor Various
Mixtures.-All the columns of the core type showed substantially
the same test phenomena for the different mixtures of concrete,
a fact which may ba accounted for by the presence of sufficient
steel to make the concrete effect the smaller element. As in none
of these columns did the total load taken by the concrete exceed
one-half of that taken by the steel, it would be expected that the
steel would govern the general behavior of the column under test.
The 1-1-2 columns were found to sustain a greater ultimate unit-
deformation than the columns of leaner mixtures. The load-
deformation diagrams for the columns of 1-1-2 mix and 1-3-6
mix are given in Fig. 15. The close similarity of the load defor-
mation diagrams and those of the 1-2-4 mix of the same length,
shown in Fig. 11 and 15, may be noted.

DEFORAIA T/OA/ PEA V7OF LENGTHt

IW/T
FIG. 15. LOAD-DEFORMATION DIAGRAMS FOR COLUMNS WITH LEAN AND
RICH CONCRETE.

23. Basisfor Determining Load Taken by Concrete.-In study-

ing the strengthening effect of the concrete on the steel column
it is necessary to decide upon some basis of division of the total
load into the part considered to be carried by the steel and the
part carried by the concrete of the columns. The load-deforma-
 TALBOT-LORD-TESTS OF COLUMNS

tion diagrams given for the tests of plain steel columns (Fig. 6,
page 18) show that the compression deformations of these col-
umns vary considerably from a straight line, which would be the
form if a constant modulus of elasticity were assumed for the col-
umn. A straight-line relation does not take into account the ad-
justments in the bearing of the rivets and in the loading of the
tie plates or lacing bars, the local flexure of the flanges, and the
later changes in the stiffness of the material itself. Of course, it
can not be told to what extent the presence of the core concrete
in columns will overcome the agencies which cause curvature in
a plain steel column before parts of the steel reach the yield
point. It seems probable that there will be some such action and
that the concreting of the column will add stiffness to the steel
section itself. However, for the purpose of the discussion it seems
best to consider that the steel section in any concreted column
carried the same amount of load as was carried by a plain steel
column of the same length. For this purpose the average load-
deformation diagram for the plain steel columns of the same length
are plotted on the load- deformation diagrams of the tests of columns
of core type. (See Fig. 11 and 15). For any given unit-deforma-
tion we may then obtain the amount of load considered to be car-
ried by the concrete by subtracting the load carried by the plain
steel columns at this deformation from the total load carried by
the concreted column. The results have been plotted on the line
marked "Concrete". It seems possible that the values given by
this line will be somewhat in excess of the part actually taken by
the concrete, though of course it may be of little consequence
whether this small part of the load is taken by the concrete or is
carried by the steel by reason of the greater stiffness given it by
the concrete. A similar method was used for determining the
part of the load taken by the concrete at the maximum strength
of the column. The values thus found are given in Table 6.
These methods of dividing load between steel and concrete will
be used in the later discussion of the ratio of steel stresses to
concrete stresses.
24. Development and Amount of Concrete Stress.-A considera-
tion of these derived curves for the concrete (Fig. 11 and 15) en-
ables the unit-stress on the concrete to be determined (a com-
parative and not an absolute figure) and also the manner of de-
velopment of this stress throughout the test. For all mixtures
it was found that at the beginning of the test the concrete took
the load rapidly, with constant speed in the testing machine,
30 ILLINOIS ENGINEERING EXPERIMENT STATION

much the same as does the concrete in the cube or the cylinder.
In Fig. 16 the stress-deformation diagrams are shown for the
concrete of several columns and for the cylinder made from the
same batch of concrete. In the earlier stages of the test the
curves are not dissimilar. As the deformation increases, how-
ever, the concrete in the columns takes proportionately less and
less stress. In general, too, the concrete in the column may be
said to be less stiff than the same concrete placed in the cylinder.
400000

300000

20000

"300000

S200000

/00000

I.

DEFORMAT/ON PER UN/T/ OF LENG7TH

FIG. 16. COMPARISON OF STRESS-DEFORMATION IN COLUMN CONCRETE
AND IN CYLINDERS.
For the 1-1-2 concrete there are no exceptions to this statement.
For the 1-2-4 mixture 6 out of 8 columns follow the rule and the
other two follow the rule up to a medium load. For the 1-3-6
mixture both columns show less stiffness up to a medium load and
greater stiffness at the latter stages of the test than do the cor-
responding concrete cylinders. None of the concrete curves of
Fig. 11 and 15 reaches its maximum value before the column as a
whole reaches its ultimate load, but the curve becomes very flat at
the higher deformations and shows a tendency for the concrete
 TALBOT-LORD-TESTS OF COLUMNS

stress to become nearly constant in value over a considerable range

of shortening. This tendency is one to be considered in deciding
upon permissible stresses for use in designing columns of the
core type.
The amount of stress considered to be taken by the concrete,
upon the assumptions previously discussed, is given in Table 6,
page 15. It is to be noted that the columns with 1-2-4 concrete
made on October 28 and 29 give concrete stresses much lower
than those made on November 8, so much so that the concrete of
the columns falls into two distinct groups. The cube and cylinder
tests fall into two similar groups. The materials were taken
from the same lot and it is reasonably certain that the measure-
ments and weights of materials are correct. The tests indicate
that the cement used early in the season acquired its strength
more slowly than that used later, (the later specimens acquired
a normal amount of additional strength between the ages of 60 and
90 days) and the room temperature the last of October and the first
of November was lower than later in the year. It is thought then
that part of the difference in strength is due to differences in rate of
hardening. Under the circumstances it will be best to treat the
1-2-4 columns as made of two grades of concrete, dividing them
into two groups, those made October 28 and 29 in group (a) (No.
8907, 8912, 8917 and 8922) and those made November 8 in group
(b) (No. 8908, 8913', 8918 and 8923). The columns with 1-1-2 con-
crete gave a similar but smaller variation. The distinctions
named above have been indicated in Fig. 12 and 14, the solid
symbols representing group (a) and the open symbols group (b).
25. Comparison of Cube and Cylinder Strength with Column
Strength.-In Fig. 17 the values of the ultimate stresses taken by
the concrete determined as described in a preceding paragraph
are plotted as abscissas and the strengths of the corresponding
cubes and cylinders made from the same batch of concrete as
ordinates. The results of the two groups of 1-2-4 concrete are
shown by separate symbols. The relation between the cube and
the column strengths seems to be well expressed by a straight
line; it indicates that the column concrete developed about two-
thirds the strength of the same concrete tested in 6-in. cubes.
The relation between the cylinder and the column strengths seems
more uncertain. . For the 1-2-4 mixtures, the ratio seems to .be
about 1. For the 1-1-2 mixture the cylinder shows higher
strength and for the 1-3-6 mixture the core concrete shows higher
32 ILLINOIS ENGINEERING EXPERIMENT STATION

strength. The average of all the ratios of cylinder-column

strength is about 1.
26. Values of E and n.-An accurate determination of the
modulus of elasticity of the column concrete oan not be made, but
for the purposes of comparison it may be proper to use the
stresses obtained by the method given in "23. Basis for Deter-
mining Load Taken by Concrete". Assuming these concrete stress-

k30 -- -,7- /
oup
--/• -- 4 -- --.----
4 6r o I
Z Cubes / * /-~ / o
up
2500 o /---4 Fnc/6*

-- ~~ --
*^2000 ~0 - - 000---/- -

o 0 ° / /

+/
<
+ / Ci lindc rs

±
^ ~ -- Z -- SOO -- --- - - - - ---

s/ren,/h of Co/umn Concree /n [Link]. /n.

Fia. 17. COMPARISON OF STRENGTH OF CONCRETE IN CUBES,
CYLINDERS AND COLUMNS.

es to be correct, the secant moduli of elasticity (E) have been cal-

culated and are given in Table 8, page 22, for unit deformations of
.0004, .0007, .0010 and for a deformation near the ultimate. The
ratio between this modulus of elasticity and the modulus of
elasticity of the plain steel column at the same deformation
(values of n) are also given in Table 8. This ratio expresses the
relation between the unit-stress taken by the steel and the unit-
stress taken by the concrete. It is to be noted that these values
of the ratio n are larger than ordinarily assumed in reinforced
 TALBOT-LORD-TESTS OF COLUMNS

concrete design. This is probably due largely to the fact that

the concrete in the column is less dense than the same concrete
would be if allowed to settle and shrink freely. A striking fact
is that for widely different values of the unit-deformations the
value of n varies but little.
If the columns with 1-2-4 concrete are divided into the two
groups noted in "24. Development and Amount of Concrete
Stress", we shall obtain the values for the stresses taken by the
concrete and the values of the ratio n given in Table 8. In
discussing the value of n to be used in design it must be borne in
mind that the concrete in the columns was not as well seasoned
as would ordinarily be the case in building construction. It should
TABLE 9.
AVERAGE STRENGTH OF SPECIMENS AND VALUES OF n.
Stresses are given in pounds per square inch.

Compressive Strength
Mixture Average Suggested
Cube Cyinder
Column Values of n Values of n
Cube Cyinder Concrete

1-3-6 1400 690 990 36 35

1-2-4 (a) 1500 1150 1050 35
(b) 2200 1300 1450 23 25
1-1-2 3100 2475 2195 16 16

also be said that, as in this type of column the steel will be used
as the basis of design, the value of the ratio n to be accepted in
design should be greater than the average value found, rather
than less, in order to be on the safe side. In Table 9 values of n
for the different mixtures of concrete are given which seem
reasonable for use in design.
Another Basis for Design.-Another basis for design which
seems rational Is to determine the strength of the steel column
for the of the steel column, taking this from the straight-line
equation on page 23, and then to use as the strength of the
concrete of the core section (without reference to the length of
the column for the column slenderness usual in buildings) a value
taken from the strength of plain concrete, say two-thirds of the
cube strength, in this way combining the strength of steel and
concrete. This seems to be in accord with the results of tests
within the limits of the -here used. Of course a suitable fac-
d
tor of safety would then be applied.
34 ILLINOIS ENGINEERING EXPERIMENT STATION

D. FIREPROOFED COLUMNS.

27. Phenomena of the Tests.-Three columns in which a 2-in.

shell of concrete was added to the core section were tested to de-
termine the additional strength afforded by this covering and to
study the behavior of a fireproofed column under load. During
the earlier stages of the test there was no difference between the
behavior for this type of column and one of the core type except
that the outer shell did not take the expected proportion of the,
load. The concrete shell remained intact until the ultimate def-
ormation of the column was practically reached, the unit-defor-

DEFOPRMA 7T/0V PER U1N/T OF L ENGTH

FiG. 18. LOAD-DEFORMX TION D[AORAMS FOR FIREPROOFED COLUMNS.

mation being .0018 or over at first crack. The steel was then
evidently starting to yield and the concrete in the core was very
close to its ultimate deformation and strength. When the shell
cracked, the total load on the column dropped off about 65 000 lb.
and the strength of the core itself was well evidenced by the
length of time during which the load remained at this second ulti-
mate although the machine was in operation in the meantime.
 TALBOT-LORD-TESTS OF COLUMNS

The load rose 2000 to 4000 lb. after its first drop, showing that the
strength of the steel and of the enclosed concrete was not quite
fully developed when the shell cracked. Fig. 4 gives a view
after the maximum strength had been developed and the shell
had cracked.
28. Stress-deformationRelations.-The stress-deformation dia-
grams are given in Fig. 18. The moduli of elasticity of the col-
umns and the values of n are given in Table 8, page 22. Owing
to the presence of the shell, which constituted 47% of the area of
the total concrete section, and which it will be seen did not carry
its full share of the load, the value of the modulus of elasticity of
the column concrete is somewhat less for the fireproofed type than
for the core type. Where n for the latter averaged about 23 for
concrete of the same grade, it becomes more nearly 30 for the for-
mer. Since undoubtedly it is wisest not to figure on any load on
the shell in designing, this value of n is significant only in show-
ing that the use of values of n herein recommended does not
threaten the safety and integrity of the shell.
29. Comparison of Concrete Stresses on Gross Section and on
Core Section.-The maximum load carried by the fireproofed col-
umns is given in Table 6, page 15. The division of load between
steel and concrete, determined on the assumptions used in the dis-
cussion of the core type of column, is also given in this table.
Of the three fireproofed columns tested two were tested to
failure. No. 8930 took a maximum load of 630 700 lb.; the load
fell to 565 000 lb. when the shell failed, indicating that a load of
65 700 lb. was carried by the shell. For No. 8931 the maximum
TABLE 10.
STRESS CARRIED BY CONCRETE IN SHELL AND
IN CORE OF FIREPROOFED COLUMNS.

Stress in pounds per square inch

Column
No. Gross Shell Core
Concrete Concrete Concrete

8930 1 060 710 1 370

8931 1 090 680 1 440

load was 635 700 lb. and the core load was 572 000 lb., the differ-
ence, 63 700-bk, apparently being carried by the shell. Of the core
load 418 000 lb."may be considered to have been carried by the
steel (determined from the average strength of the plain steel col-
B6 ILLINOIS ENGINEERING EXPERIMENT STATION

umns of the same length), leaving on the core concrete 147 000 lb.
for No. 8930 and 154 000 lb. for No. 8931. The area of the shell
was 93 square inches and of the core concrete 107 square inches.

\600000
"500000

J400000

23000000

200000

/00000

DEFORM47T/ON PER UN/I OFLENG TH

FIG. 19. LOAD-DEFORMATION DIAGRAMS FOR SPIRALED
COLUMNS,-FIRST TEST.
 TALBOT-LORD--TESTS OF COLUMNS

The unit-stresses, as found by the method just outlined, are

given in Table 10. These values indicate that the concrete shell
carried roughly one-half as much load in pounds per square inch
as the concrete of the core.
30. Permissible Deformation as Governing Design.-The ques-
tion always arises in design as to the effect of large deformation
in a column under load upon the integrity of the shell. If the
shell goes to pieces at relatively low unit-deformations, it is evi-
dent that deformation rather than stress must govern in the
selection of the working stresses to be used in design. As al-
ready noted, the shell did not crack or the load drop off until the
unit-deformation exceeded .0018, and this is practically the def-
ormation at the maximum load for the core type of column also.
In other words, the action of the shell during the test did not
seem to be such as to impose any restriction on the selection of
working stresses.
31. Need of a Tie to Prevent Stripping of Shell.-Although the
shell maintains itself intact under high unit-deformation, yet it
seems that a tie of some sort (wire mesh, spiral, or other binder)
should be imbedded in the outer shell to make its permanence
certain. The backs of the flanges occupy one-fourth of the
bonding surfaces between the shell and the core, and in prac-
tice this proportion might be even greater. It seems unwise to
trust the fireproofing shell of the column to stand uninjured un-
der collisions and accidents with so large a surface uncertainly
supported. This tie would also prevent the rapid and complete
failure of the shell at maximun load, as it occurred in the test of
the columns, although this advantage is not very great in actual
construction. The prevention of spalling away from the steel in
case of a severe fire is a more important reason for requiring
a metal binder to hold the exterior concrete in place. From the
tests of spiraled columns it is concluded that a spiral is an excel-
lent tie for this purpose.

E. SPIRALED COLUMNS.

32. Phenomena of Tests and Stress-deformation Relations.-For

loads within the capacity of the University of Illinois testing
machine (600 000 lb.) the load-deformation curves of the columns
with I % and with 1 % spiral reinforcement (see Fig. 19) are
practically identical in nature with those for the core type. The
tests were not carried to the point where the thin concrete coat-
38 ILLINOIS ENGINEERING EXPERIMENT STATION

0 0s . 0 0
DEFO//AM r/OW PER U/V/7M OFL EANGTH
FIG. 20. LOAD-DEFORMATION DIAGRAMS FOR SPIRALED COLUMNS,-SEC-
OND AND THIRD TESTS.
 TALBOT-LORD--TESTS OF COLUMNS

ing over the spiral would be expected to spall. In three of the

spiraled columns the maximum load was applied three to five
times, the result being increased deformations under maximum
load and increased set upon the release of load.
Four of the spiraled columns were subsequently tested a sec-
ond time at Lehigh University, where a load of 830 000 lb. was ap-
plied. Even with the heavier testing machine, the full strength
of the columns was not developed, though No. 8934 evidently was
loaded nearly to its maximum strength. The load-deformation
diagrams are shown in Fig. 20. The deformations at the
first tests of the same columns are also given by the broken lines.
It must be borne in mind that the columns were about three
months older at the second test than when first tested, and the
increased strength of the concrete at a given unit-deformation may
be accounted for by the increased age. To secure some conception
of the amount of the added strength due to the greater age of the
concrete, No. 8937 was tested a third time, the spiral first being
removed, the outer concrete stripped off, and the column reduced
to the section of the standard core type. The column in this
condition carried a load of 714 000 lb., as compared with 547 000 lb.
carried by the two corresponding columns of the core type at an
age of 60 days, (see Fig. 20) and it is evident from this that a con-
siderable portion of the added load carried in the second test was
due to the greater age of the concrete. It is well to call attention
to the fact that this increase of strength was gained after the
concrete had been subjected to high stresses at an age of 60 days.
The large lateral deflection of the four columns tested at
Lehigh University is of interest. In No. 8934, at a load of
752 100 lb. the deflection was .09 in. and at this load the concrete
began to spall. At a load of 856 000 lb. the deflection was .26 in.,
a deflection set of .21 in. remaining when the load was released.
At the fourth application of a load of about 850 000 lb. the
deflection became .54 in., a set of .50 in. remaining with the
release of the load. In No. 8935, at a load of 750 000 lb.
the deflection was .084 in. At a load of 825 000 lb. the deflection
became .31 in., and the concrete at the bottom scaled off
between wires. Upon the release of load the set was .31 in.
At a fifth application of the load, the deflection became .48 in.
and the deflection set shown upon the final release of this load
was .43 in. There was no marked bending in No. 8937 until the
third application of a load of 830 000 lb., when it became .12 in.
40 ILLINOIS ENGINEERING EXPERIMENT STATION

In No. 8938, at a load of 776000 lb. the deflection was .08 in., be-
coming .17 in. at a load of 825000 lb., with a set of .16 in. shown
upon release of load. At the fifth application of the load, the
deflection became .25 in. with a resulting set of .22 in. This
marked bending of the spiraled columns at loads above those
which would be carried by an unspiraled column is in keeping with
the large deflections found in hooped concrete columns of the
usual type.
The phenomena of the second test were not essentially dif-
ferent from those described for the first test. At a load of about
750000 lb. spalling of the outer concrete began on the columns
having i % spiral reinforcement, and this spalling continued dur-
ing the remainder of the test. It seems evident from the action
of the columns and the amount of deformation developed that
these columns were very close to their maximum load at the end
of the second test. The view in Fig. 5 shows the spalling of the
concrete and the buckling of the spacing strip. With the columns
having 1 % of spiral reinforcement the spalling was slight even
at the highest load carried, and these columns evidently would
have carried considerably more load. An examination of the
structural steel after the second test showed no crimping of the
flanges and no movement of the parts relatively to one another,
although, of course, the total shortening of the column was not
large. The spiral seems to have acted to hold the steel in align-
ment and to permit a greater shortening than would otherwise
have taken place.
33. Effect of Spiral.-The maximum loads placed on the
spiraled columns are given in Table 6, page 15. It must be
borne in mind that the capacity of the testing machines used did
not permit the maximum strength of these columns to be
developed.
Within the limits of the first test (600 000 lb. load) the load-
deformation diagrams do not show any eff ect which may be at-
tributed to the variation in the percentage of spiral reinforce-
ment or even to the presence of the spiral. A study of the stress-
es in these columns of the core type shows that the spiral has
little apparent effect upon the action of the column within the
load of 600 000 lb.
In the second test the maximum load applied evidently ap-
proached very closely to the maximum strength of the columns
with I % of spiral reinforcement, but it did not stress the col-
 TALBOT-LORD-TESTS OF COLUMNS

umns with 1 % of spiral reinforcement nearly as severely. The

load-deformation curves (Fig. 20, page 38) show that the marked
yielding of the column takes place at about the same load for
both percentages (650 000 to 700 000 lb.) and that at this yield
point the spiral begins to play an important part. During this
later part of the test the difference in amount of spiraling becomes
apparent and the heavier spirals show greater strength. Since
the tests were not carried to destruction, the amount of this added
strength can not be ascertained, but its existence seems well
established.
For most purposes, then, the significant point on the stress-
deformation curves is the point where the curve bends sharply
and becomes flat and which may be considered to be the load
which would be carried by an unspiraled column. Beyond this
point any additional load is carried only by virtue of very greatly
increased deformations. For the four columns tested this point
lies at or below 750 000 lb., and it will be interesting to compare
this with the strength of the column without spiral reinforce-
ment. For the purpose of this comparison the results of the
third test of No. 8937, after it was stripped of its spiral and
reduced to the standard core type, may be used to estimate the
strength of the concrete at the time of the later test. This col-
umn carried 714 000 lb. load as against 580 000 lb. load carried by
columns of the core type at 60 days of age. If we consider all
the concrete within the spiral in the second test to be as effective
as the core concrete in No. 8937, the load carried without aid
from the spiral would be about 800 000 lb. This figure is prob-
ably a little high, as the concrete outside the core, although in-
side the spiral, would probably take somewhat less load per
square inch than the concrete in the core. This tends to confirm
the view that the load at the point when the. load-deformation
curve bends sharply and becomes flat is approximately equal to
the strength of the unspiraled column, and that only the part of
the strength developed after this yield point is passed should be
attributed to the spiral directly. In No. 8934, this added load
amounts to about 100 000 lb. which is 715 lb. per sq. in. of in-
closed concrete. This is at the rate of 950 lb. per sq. in. of col-
umn concrete per 1 % of spiral reinforcement. Applying this
figure to the 1 9o' spiral columns, the computed probable maxi-
mum load becomes 900 000 lb., a value which the deformations of
the second test seem to indicate as a reasonable expectation for
these columns.
42 ILLINOIS ENGINEERING EXPERIMENT STATION

34. Availability of Spiral Strength for Design.-The tests of

this series of columns indicate that up to a unit-deformation of
.0015 no appreciable difference in the action of columns with or
without spiral reinforcement is found. In building construction
the safe unit-deformation may ordinarily be placed at .0007 or
less. It would seem that any attempt to use an imaginary spiral
strength at working loads could result only in very high actual
unit-deformations. The marked tendency of the column to bend
laterally after the yield point was passed and the large amount of
set found are also evidence of the unavailability of the
higher strengths. The spiral does afford protection against
sudden failure, and gives a tougher and safer column, and
these properties may be considered to warrant the use of
higher unit-stresses in spiraled columns. In the columns
of the core type tested the need of a spiral is much less than
in the ordinary reinforced concrete column, since these columns
are found to possess toughness and the flanges of the structural
angles restrain the core concrete to some extent. It thus appears
that the use of a large percentage of spiral reinforcement in col-
umns of the type here considered is hardly justifiable. A light
spiral may serve to tie the shell together securely and protect it
from accident, but this spiral should not be directly considered
in the computations for strength of column.

F. SUMMARY.

35. General Comments.-The columns tested were of a form

now frequently used in building construction. The percentage
of steel used (area of steel section 10.8 % of the area of the octa-
gon inclosing the structural shapes) is within the range used in
building construction. The conclusions given in the discussion
relate to the properties of columns which have the forms and
sections of the columns tested, and variations in proportions of
metal and concrete may give somewhat different results. The
tests, however, may be expected to throw light on the properties
of columns of the same general type within the limits of ordinary
design. The principal conclusions found in the discussion are as
follows:
 TALBOT-LORD-TESTS OF COLUMNS

1. The maximum load carried by the plain steel columns is

P 1
expressed by the straight-line formula, - 36 500 - 155 , where
'A r
Sis the ratio of length of column to radius of gyration of the
r
section of the steel column.
2. Earlier in the test the effect of length upon load carried
at a given unit-deformation was less proportionately than at
maximum load, the coefficient of -in the equation being only 55
for a unit-deformation of .0008, and 27 for a unit-deformation of
.0005, as compared with 155 in the equation for maximum load.
3. The load-deformation diagrams diverge from a straight
line at loads well below the maximum.
4. In the concreted columns of the core type, the effect of
length upon strength of column was almost identical with that
found in the tests of plain steel columns. In other words the stress
taken by the concrete may be considered to be nearly independent
of the slenderness ratio of the column, within the limits of the
lengths tested, and the stress taken by the steel may be considered
to be the same as that taken by a plain steel column of the
same slenderness ratio.
5. In the tests the concreted columns of the core type showed
considerable toughness, though at the maximum load there
was no material lateral deflection. The final failure of the con-
crete generally occurred at or above tie plates. The discussion
shows that the concrete of the columns was less strong than the
concrete of the cubes and less stiff than the concrete of the
cylinders.
6. The stress taken by the concrete within the core or with-
in the spiral is approximately equal to the strength of concrete
of the cylinders tested and to two-thirds of the strength found
in the 6-in. cubes.
7. The values of the ratio of the modulus of elasticity of the
steel column to that of the concrete, n, under the assumptions
used in the analysis, are much larger than are commonly used in
reinforced concrete design. Values of n for use in designing are
suggested in Table 9.
8. A basis for design which seems rational is to determine
the strength of the steel column by the use of the column formula
for the I of the steel column and then to consider the concrete
r
of the core section (without reference to the length of the column
44 ILLINOIS ENGINEERING EXPERIMENT STATION

for any ordinary length ratio, say a length of 15 diameters) to

have a stress value proportional to the strength of the plain con-
crete, say two-thirds of the cube strength. A suitable factor of
safety would of course be somewhere applied.
9. In the test of the fireproofed type of column (which had a
shell of concrete outside the steel) the concrete shell remained
intact until a deformation was reached as great as that developed
at the maximum load in columns of the core type. This integrity
of section at high deformations indicates that the presence of the
shell need not impose any restrictions upon the working stresses
available for the steel and for the core concrete. Of course,
there are good reasons for the use of a metal binder like wire
mesh or spiral for holding the shell securely in place.
10. The discussion indicates that the stress carried by the
concrete of the shell is only about half of that carried by the core
concrete. This lower strength is not objectionable, since the
shell is not considered in designing the column.
11. The action of the spiraled columns indicates that the
spiral has little effect up to a deformation and load corresponding
to the maximum load for an unspiraled column. Beyond this load
the column compresses rapidly and the presence of the spiral
adds materially to the strength of the column. The tests do not
fix the exact amount of this added strength.
12. In view of the large shortening necessary to make the
added strength due to spiraling available and the general tough-
ness of columns of the core type, it would seem that for building
construction the use of a large percentage of spiral reinforcement
in columns made up of structural shapes and concrete is hardly
justifiable. A moderate spiral may warrant the use of somewhat
higher unit-stresses, since it adds to the toughness of the column
and gives a possible higher ultimate strength, and it will also
serve to tie the concrete of the shell together securely and pro-
tect it from accident, but it does not seem best to consider this
spiral directly in the computations for strength of column.
13. The columns tested possess the qualities of a good struc-
tural member and seem well adapted to more general use in build-
ing construction.
These comments are made on the assumption that the con-
crete is placed in as workmanlike a manner as is obtainable in
the construction of high-grade work in columns reinforced with
longitudinal rods or with rods and spirals.
 PUBLICATIONS OF THE ENGINEERING EXPERIMENT STATION

ALL BULLETINS, THE TITLES OF WHICH ARE NOT STARRED, WILL BE SENT FREE
UPON APPLICATION.

*Bulletin No. 1. Tests of Reinforced Concrete Beams, by Arthur N. Talbot. 1904.

None available.
*Circular No. 1. High-Speed Tool Steels, by L. P. Breckenridge. 1905. None available
*Bulletin No. 2. Tests of High-Speed Tool Steels on Cast Iron, by L. P. Breckenridge
and Henry B. Dirks. 1905. None available.
*Circular No. 2. Drainage of Earth Roads, by Ira 0. Baker. 1906. None available.
*Circular No. 3. Fuel Tests with Illinois Coal. (Compiled from tests made by the Tech-
nologic Branch of the U. S. G. S.. at the St. Louis, Mo., Fuel Testing Plant, 1904-1907,) by L. P.
Breckenridge and Paul Diserens. 1909. Thirty cents.
*Bulletin No. 3. The Engineering Experiment Station of the University of Illinois, by
L. P. Breckenridge. 1906. None available.
*Bulletin No. 4. Tests of Reinforced Concrete Beams, Series of 1905, by Arthur N.
Talbot. 1906. Forty-five cents.
*Bulletin No. 5. Resistance of Tubes to Collapse, by Albert P. Carman. 1906. Fifteen
cents.
*Bulletin No. 6. Holding Power of Railroad Spikes, by Roy I. Webber. 1906. Thirty-
five cents.
*Bulletin No. 7. Fuel Tests with Illinois Coals, by L. P. Breckenridge, S. W. Parr and
Henry B. Dirks. 1906. Thirty-five cents.
*Bulletin No. 8. Tests of Concrete: I. Shear; II. Bond, by Arthur N. Talbot. 1906
None available.
*Bulletin No. 9. An Extension of the Dewey Decimal System of Classification Ap-
plied to the Engineering Industries, by L. P. Breckenridge and G. A. Goodenough. 1906. None
available.
*Bulletin No. 10. Tests of Concrete and Reinforced Concrete Columns, Series of 1906, by
Arthur N. Talbot. 1907. None available.
*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. Fifteen cents.
*Bulletin No. 12. Tests of Reinforced Concrete T-beams, Series of 1906, by Arthur N.
Talbot. 1907. None available,
*Bulletin No. 13. An Extension of the Dewey Decimal System of Classification Applied
to Architecture and Building, by N. Clifford Ricker. 1907. Fifty cents.
*Bulletin No. 14. Tests of Reinforced Concrete Beams, Series of 1906, by Arthur N.
Talbot. 1907. None available.
*Bulletin No. 15. How to Burn Illinois Coal without Smoke, by L. P. Breckenridge. 1908
Twenty-five cents.
*Bulletin No. 16. A Study of Roof Trusses, by N. Clifford Ricker. 1908. Fifteen cents.
*Bulletin No. 17. The Weathering of Coal, by S. W. Parr, N. D. Hamilton, and W. F.
Wheeler. 1908. Twenty cents.
* Bulletin No. 18. The Strength of Chain Links, by G. A. Goodenough and L. E. Moore.
1908. Forty cents.
*Bulletin No. 19. Comparative Tests of Carbon. Metallized Carbon and Tantalum Fila-
ment Lamps, by T. H. Amrine. 1908. Twenty-five cents.
*Bulletin No. 20. Tests of Concrete and Reinforced Concrete Columns, Series of 1907, by
Arthur N. Talbot. 1908. None available.
*Bulletin No. 21. Tests of a Liquid Air Plant, by C. S. Hudson and C. M. Garland. 1908.
Fifteen cents.
*Bulletin No. 22. Tests of Cast-Iron and Reinforced Concrete Culvert Pipe, by Arthur N.
Talbot. 1908. Thirty-flve cents.
*Bulletin No. 23. Voids, Settlement and Weight of Crushed Stone, by Ira 0. Baker. 1908.
Fifteen cents.
Bulletin No. 24. The Modification of Illinois Coal by Low Temperature Distillation, by
S. W. Parr and C. K. Francis. 1908. Thirty cents.
Bulletin No. 25. Lighting Country Homes by Private Electric Plants, by T. H.
Amrine. 1908. Twenty cents.
Bulletin No. 26. High Steam-Pressures in Locomotive Service. A Review of a Report
to the Carnegie Institution of Washington, by W. F. M. Goss. 1908. Twenty-five cents.

* Out of print: price attached.

 PUBLICATIONS OF THE ENGINEERING EXPERIMENT STATION--(Continued)

Bulletin No. 27. Tests of Brick Columns and Terra Cotta Block Columns, by Arthur N
Talbot and Duff A. Abrams. 1909. Thirty cents.
Bulletsn No. 28. A Test of Three Large Reinforced Concrete Beams, by Arthur N.
Talbot. 1909. Fifteen cents.
Bulletin No. 29. Tests of Reinforced Concrete Beams: Resistance to Web Stresses,
Series of 1907 and 1908, by Arthur N. Talbot. 1909. Forty-five cents.
Bulletin No. 30. On the Rate of Formation of Carbon Monoxide in Gas Producers, by
J. K. Clement, L. H. Adams, and C. N. Haskins. 1909. Twenty-five cents.
Bulletin No. 31. Fuel Tests with House-heating Boilers, by J. M. Snodgrass. 1909.
Fifty-five cents.
Bulletin No. 32. The Occluded Gases in Coal, by S. W. Parr and Perry Barker. 1909
Fifteen cents.
Bulletin No, 33. Tests of Tungsten Lamps, by T. H. Amrine and A. Guell. 1909.
Twenty cents.
Bulletin No. 34. Tests of Two Types of Tile Roof Furnaces under a Water-tube Boiler
by J. M. Snodgrass. 1909. Fifteen cents.
*Bulletin No. 35. A Study of Base and Bearing Plates for Columns and Beams, by N.
Clifford Ricker. 1909. Twenty cents.
Bulletin No. 36. The Thermal Conductivity of Fire-Clay at High Temperatures, by J.
K. Clement and W. L. Egy. 1909. Twenty cents.
*Bulletin No. 37. Unit CoAl and the Composition of Coal Ash, by S. W. Parr and W. F.
Wheeler. 1909. Thirty-five cents.
Bulletin No. 38. The Weathering of Coal, by S. W. Parr and W. F. Wheeler. 1909.
Twenty-five cents.
Bulletin No. 89. Tests of Washed Grades of Illinois Coal, by C. S. McGovney. 1909.
Seventy-Alve cents.
Bulletin No. 40. A Study in Heat Transmission, by J. K. Clement and C. M. Garland.
1910. Ten cents.
Bulletin No. 41. Tests of Timber Beams, by Arthur N. Talbot. 1910. Twenty cents.
Bulletin No. 42. The Effect of Keyways on the Strength of Shafts, by Herbert F. Moore
1910. Ten cents.
Bulletin No. 43. Freight Train Resistance, by Edward C. Schmidt. 1910. Eighty cents.
Bulletin No. 44. An Investigation of Built-up Columns under Load, by Arthur N.
Talbot and Herbert F. Moore. 1911. Thirty-five cents.
Bulletin No. 45. The Strength of Oxyacetylene Welds in Steel, by Herbert L.
Whittemore. 1911. Thirty- tve cents.
Bulletin No. 46. The Spontaneous Combustion of Coal, by S. W. Parr and F. W. Kress-
mann. 1911. Forty-flve cents.
Bulletin No. 47. Magnetic Properties of Heusler Alloys, by Edward B. Stephenson.
1911. Twenty-five cents.
Bulletin No. 48. Resistance to Flow through Locomotive Water Columns, by Arthur
N. Talbot and Melvin L. Enger. 1911. Forty cents.
Bulletin No. 49. Tests of Nickel-Steel Riveted Joints, by Arthur N. Talbot and Herbert
F. Moore. 1911. Thirty cents.
Bulletin No. 50. Tests of a Suction Gas Producer, by C. M. Garland and A. P. Kratz.
1911. Fifty cents.
Bulletin No. 51. Street Lighting, by J. M. Bryant and H. G. Hake. 1911. Thirty-flve
cents.
Bulletin No. 62. An Investigation of the Strength of Rolled Zinc, by Herbert F. Moore.
1911. Fifteen cents.
Bulletin No. 53. Inductance of Coils, by Morgan Brooks and H. M. Turner. 1912. Forty
cents.
Bulletin No. 54. Mechanical Stresses in Transmission Lines, by A. Guell. 1912. Twenty
cents.
Bulletin No. 55. Starting Currents of Transformers, with Special Reference to Trans-
formers with Silicon Steel Cores, by Trygoe D. Yensen. 1912. Twenty-five cents.
Bulletin No. 56. Tests of Columns: An Investigation of the Value of Concrete as Rein-
forcement for Structural Steel Columns, by Arthur N. Talbot and Arthur R. Lord. 1912.
Twenty-five cents.

* Out of print; price attached.

 ANNOUNCEMENT CONCERNING A MODIFICATION IN
THE RULES GOVERNING THE DISTRIBUTION OF BULLETINS

The Board of Trustees of the University of Illinois voted, December 3,

1910, that a price should be affixed to certain University publications, among
them being the Bulletin of the Engineering Experiment Station.
This action, so far as it concerns the bulletins of the Engineering
Experiment Station, has for its purpose a threefold object:
(1) To provide a greater degree of control in the distribution of bul-
letins.
(2) To make possible the establishment and maintenance of a trade
circulation through the regular publishing houses.
(3) To regulate the distribution of the reserve or "out-of-print" supply.
IT IS NOT INTENDED THAT THIS ACTION SHALL OPERATE
IN ANY WAY TO ABRIDGE THE PRIVILEGES OF THOSE WHO
HAVE HITHERTO RECEIVED THE BULLETINS OF THE STATION
GRATUITOUSLY, OR TO PREVENT REASONABLE EXTENSIONS
OF THE EXISTING MAILING LISTS.
Practice under the new procedure will be as follows:
(1) All bulletins, hereafter issued, will have a price printed upon the
cover, together with Lhe name of the London sales agent.
(2) EACH BULLETIN, HOWEVER, WILL STILL BE SUBJECT
TO A FREE INITIAL DISTRIBUTION, AS HERETOFORE, ON THE
BASIS OF EXISTING MAILING LISTS.
It will also be placed on sale with authorized agencies, both in this
country and abroad.
(3) THERE WILL BE A LIMITED NUMBER OF COPIES AVAIL-
ABLE FOR FREE DISTRIBUTION, UPON REQUEST, AFTER THE
INITIAL DISTRIBUTION.
(4) As the supply of each bulletin approaches exhaustion, it will be
placed upon a reserve or "out-of-print" list. THE EFFECT OF THIS
ACTION WILL BE TO WITHDRAW SUCH BULLETINS FROM FREE
DISTRIBUTION. Bulletins withdrawn from free distribution will be
available to any applicant upon payment of the assigned price.
(5) A discount from the published price will be allowed to news agents
regularly handling the bulletins.
W. F. M. GOSS
Director.
 I

/ f
:~ ~ ~
(C ~

C\ A ~
C-

C' 417 ~g-

'C CC'.
~ C

'~'C' ~.

CV'
'C C'C$ C,
'C

'C ,CC

~ EDU~. ~' C,

C C~ C-:. ~C

C ~' C,

~C~7
CCC
1%i iil|;iHil iigil

I V \i4A' >9

it^' ^Si : ll>9 s

^: *t! ;^ '^'99
N, ,,,?^^a '>9>9ii',»^^^^ .

t4%^ 4,i?^te1
'A ^ -) 4> V 4'"'

:
' t N
9 19 i 's
'l>e

llxp
I^^^^^?^
~^^%