Actual crushing tests upon 12 inch cubes of Giant cement concrete, 3
months old, made in the proportions of i cement, 3 sand, and 5 broken
stone, and tested at the government testing laboratory, gave crushing
strengths varying from 3,081 pounds to 4,451 pounds per square inch and
averaging over 4,000 pounds per square inch. Such concrete would sup-
port a column of its own material 3,600 feet high without crushing.
Where thick or heavy masonry work is to be built, the interior mortar
joints are more or less excluded from, contact with air, and consequently
from the carbonic acid contained in the air, hence it is safe to conclude
that it will take long periods of time for the lime mortar to become per-
fectly hardened. During this time the structure is liable to settlement
and deformation, therefore the necessity of some more permanent and
quickly hardening material for mortar.
Baker says in a note on lime mortar, "Lime mortar taken from the
walls of ancient buildings has been found to be only 50 to 80 per cent, sat-
urated with carbonic acid after nearly 2,ooe years of exposure." "Lime
mortar 2,000 years old has been found in subterranean vaults, in exactly
the condition, except for a thin crust on top, of freshly mixed mortar."
For the heavy structures of the present day it is quite apparent that
such defects would be dangerous.
The question of protection from the destructive action of the elements
is often as important a factor of consideration with smaller structures as it
is with the larger ones. The porosity of lime and mortar would allow
moisture and temperature changes to affect the durability of such work.
Cement mortar, upon the other hand, can be made impervious to water.
It will set under water and without contact with air and will continue to
gain strength for an unknown period of time. Cement mortar in the
center of a thick wall will be practically as hard and durable as the ex-
terior surface at the same age.
Lime mortar is of no value in submarine work. Good, durable
masonry can not be laid in water without cement mortar. The fine break-
waters lately constructed by the United States Government at Buffalo and
Cleveland, the concrete jetties at the mouth of the Mississippi river, the
sea-walls around Galveston, Texas, and Havana, Cuba, all illustrate
the value of cement in marine or hydraulic service.
Digitized by
Google
ZO ANNUAL REPORT
NATURAL VERSUS PORTLAND CEMENT.
The question of whether to use natural or Portland cement mortar
depends upon the strength required, upon the time which can be allowed
for the mortar to gain sufficient strength for the immediate requirements
of the work and upon the relative price of the two cements.
It may be that a rapid setting mortar is needed without great ultimate
strength ; if so, a natural cement will fulfill the requirements at a less cost
than Portland cements. The time of setting of both natural and Portland
cements varies so widely with different brands that only general limits may
be stated. Natural cement usually begins to set in from five to forty min-
utes, and attains its permanent set in from twenty minutes to two and one-
half hours, while Portland cement begins to set in from three- fourths of an
hour to three hours and attains its final set in from two and one-half to
eight hours. Some experimental Portlands, however, have been known
to begin setting within three minutes and to have attained hard set in
fifteen minutes.
To illustrate the question of relative cost and strength of the two
cements, suppose the specifications for a certain structure call for a cement
which shall develop an ultimate tensile strength in the work, of 200
pounds per square inch. Most of the natural cements in mortars of i
cement to 2 sand will develop that strength in three months, and greater
strength in six months or a year ; while. Portland cements will give similar
strengths if mixed in the proportion of i cement to 4 or 5 sand. Assum-
ing sand at $1.25 per cubic yard, natural cement at 90 cents per barrel
of 265 pounds, and Portland cement at $2.50 per barrel of 380 pounds,
the material for a cubic yard of mortar will cost :
TABLE 2.
For Portland Cement
Proportions 1 to 6.
0.92 cu. yds. of sand at $1.25
1.12 bbls. cement at 2.50
$1.15
2.80
18.95
For Natural Cement
Proportions 1 to 2.
0.81 CU. yds. sand at $1.25
2.49 bbls cement at .90
fl.Ol
2.24
$3.25
From this comparison it is seen that a natural cement mortar of i
cement, 2 sand, is 70 cents cheaper per cubic yard than a Portland cement
mortar of i cement to 5 sand, both mixtures having about the same ten-
sile strength. These comparisons, however, can only be made under
known conditions, the proportions being dependent upon the weight of
the cement and sand per cubic foot, the voids in the sand and the amount
of water used in the mixing, while the amount saved is greatly influ-
enced by the relative cost of the two cements and the cost of the sand.
Digitized by
Google
STATE GEOLOGIST.
27
When great ultimate strength is required, or when variable strains
occur in a structure, Portland cement should be used.
MIXING THE MORTAR.
In preparing mortar the sand and cement should ^e thoroughly
mixed dry and then water added and the mass carefully mixed again, until
the mortar has a proper and uniform consistency to work easily and
smoothly under the trowel. The strongest mortar for any given sand
and cement is produced when sufficient cement is used to just fill all the
voids in the sand with a thin coating of cement over each grain of sand.
Sands vary greatly in their coefficient of uniformity. Some sands are all
fine, some all coarse, while many are graded from very fine to very coarse.
It follows, therefore, that a graded sand requires less cement to make
a mortar having a given strength, than a sand which has uniform size.
The voids in the latter sand amount to nearly 50 per cent, of the mass.
EFFECT OF VARIOUS SANDS UPON THE STRENGTH OF MORTAR.
Standard quartz sand has about 48 per cent, of voids in it when
measured dry. Lake sand from Sandusky, Ohio, has about 35 to 37 per
cent, of voids. Bank sand from Mock's sand bank northeast of Colum-
bus, containing considerable clay or loam, has about 33 to 34 per cent,
of voids.
A class in civil engineering at the Ohio State University carried on
extensive tests during the winter term of 1902-03, with several brands of
cement and the three kinds of sand named above. The characteristics
of the sands were as follows :
TABLE 3.— Oharaoteristics of Sands.
Fineness— per cent, of
sand remaining on a
Passing a
No. 50
Sieve.
Voids
Per Cent.
Weight
per
cu. ft.
Remarks.
Kind of Sand.
No. 20
Sieve.
No. 90
Sieve.
No. 50
Sieve.
Crushed Quartz
7.7
100.0
47.5 to
49.0
83
Clean.
Lake Sand.
16.7
26.6
49.1
35 to
37
103
Clean.
1
Ba,nk Sand. 11.6
37.0
33.4
18.0
33 to
34
102
Contains
about 7%
of loam.
Among the cements tested were the Atlas, Giant and Dyckerhoff
brands of Portland cement. Seven and twenty-eight day tests were made.
The results of the 28 day tests are given in table number 4. While the
table shows quite a variation in the results obtained by the different
testers, it certainly shows a remarkable uniformity in results when the
Digitized by
Google
28
ANNUAL REPORT
inexperience of the class is taken into consideration. One thing is quite
markedly shown in this table and is corroborated by a large number of
other tests^ i, e,, the increase in tensile strength shown in the mortars made
with the bank sand over those made with the other sands, especially in i
to 2 or I to 3 mortars. One of the important features of the table is its
comparative value for illustrating the personal factor in cement testing.
TABLE 4.
Twenty-eight Day Teste of Cement Mortar Briquettes — 1903.
Tester.
Neat
Standard Sand.
Lake Sand.
Bank Sand.
c
1-1
1-2 ! 1-3
1-1
1-2
1-3
1-1
1-2
1-3
E
<5
0. L. BuRhey
L. Eysenbach
Edw. Thomas . . .
0. A. MeUck
E. R. Brashear...
W. J.Barry
799
800
770
769
778
813
747
767
524
586
701
700
466
400
359
460
240
442
235
231
210
229
246
206
531
685
474
524
595
560
363
322
221
297
380
312
216
182
222
248
204
203
584
602
*393
627
656
632
427
418
♦802
509
556
492
291
367
*265
318
390
342
Atlas
Averages. . .
788
671
360
226
543
316
212
574
451
329
C. W. Schubert..
J. L. Murphy. . . .
O.L. Hill
J. H. Ohubb
Averages. .
908
783
505
252
630
458
278
640
660
364
970
633
456
245
600
436
226
571
431
318
1015
702
557
347
627
468
314
566
463
364
1231
868
600
381
733
476
366
687
649
472
1031
746
529
306
647
457
296
613
521
377
Giant.
J. H. Ohubb
J. L. Murphy
O.L.Hill
O. W. Schubert. .
W. J. Barry
O. A. Melick
E. R. Brashear..
Edw. Thomas. .
L. Eysenbach
O. L. Bushey
Averages
681
490
439
800
614
406
316
570
504
418
567
510
415
280
862
840
288
480
416
856
480
460
434
302
426
362
288
520
444
342
620
486
346
235
409
283
242
650
483
420
531
412
312
174
387
837
242
667
442
882
442
463
374
260
829
264
283
606
510
321
618
467
390
236
898
856
282
680
496
266
487
860
808
284
312
215
212
390
212
282
532
385
378
180
396
317
206
414
433
363
640
425
870
189
854
300
249
418
425
346
560
446
377
288
387
818
255
505
436
339
Dyck
* Oomposed of a mixed sand.
There may be several reasons for this increased strength of the bank
sand mortar over the clean lake and quartz sands ; but it is the writer's
opinion that the principal reason is found in the smaller percentage of
voids and the consequent surplus amount of cement which can go towards
coating every particle in the aggregate. An examination of the table of
characteristics of these sands shows that nearly one-half of the lake sand
passes through a sieve of 2,500 meshes to the square inch. This accounts
at once for the greater percentage of voids which must be thoroughly
filled with cement in order to give greater strength to the mortar. This
fine portion of the lake sand is composed of small, rounded and very
smooth particles of quartz which naturally have no interlocking qualities,
and must therefore depend entirely upon the adhesive powers of the ce-
Digitized by
Google
STATE GEOLOGIST.
29
ment upon their hard smooth surfaces to give strength to the mortar.
The increased strength of the standard quartz mortar is largely due to
the irregular rough grain of the sand which allows the grains to interlock
and which also gives rough adhesion surfaces for the mortar contact.
THE EFFECT OF FINE SAND UPON THE STBIENGTH OF MORTAR.
During the last winter the writer carried on a few tests to determine
the effect of fine sand upon cement mortar. Similar and more extensive
tests have been carried on by others, and in the main all agree that fine
sand weakens cement mortar. The tests made by the writer are here
presented in tabular form :
TABLE 5.
Effect of Fine Sand Upon Cement Mortar.
Per cent.
of
Water.
11
18
1 to 1 Mortar.
Per cent.
of
Water.
1 to 2 Mortar.
Lake Sand.
7
Days.
28
Days.
3
Mos.
7
Days.
28
Days.
Passing a
No. 50
Sieve
315
426
394
395
578
533
668
617
14 '
815
367
Remaining on
No. 60 Sieve
12>^
517
514
629
Remaining on
No. 30 Sieve
12>^
546
625
650
Remaining on
No. 20 Sieve
12K
495
599
613
Remaining on
No. 16 Sieve
12>^
460
521
557
Standard
Quartz
12K
535
632
740
Lake Sand
Unsieved
14
500
581
12K
367
428
The number of the sieve designates the number of meshes to the
linear inch. The tests show a marked increase in strength in mortars from
fine to coarse sand, up to the size of the standard sand grain; from that
size up the strength seems to decrease. It may be noted that the lake
sand between the 20-30 sieves gives results comparable with the standard
quartz mortars. The general results of these tests agree quite well with
the results of other testers.
This series of tests was limited, only about 150 briquettes being
broken. But the tests were made with great care in all details and will
give a general idea of the action of fine; sand upon cement mortar.
The unsieved lake sand occupies a place intermediate between the
50 and the 30 sieve. The results with different proportions of sand would
probably differ slightly, but the two sets of briquettes broken from the i
to 2 mortar indicate relativelv similar results.
Digitized by
Google
30
ANNUAL REPORT
The following table taken from the "Directory of American Cement
Industries," being the result of a series of tests made at the Holyoke
Dam, Massachusetts, and supplemented by some of the writer's tests,
shows the effect of adding sand to cement in reducing the strength of the
mortar. All tests were tensile tests upon mortar briquettes 28 days old,
made from high grade American Portland cements.
The results of the Holyoke test are given in column number 3 and
the writer's tests in column number 4.
TABLE 6.
Strength of Various Grades of Mortar.
Cement.
Sand.
Tensile Strength.
Pounds per square inch.
1
2
3
4
neat
889
878
1
805
681
2
589
428
3
843
275
4
204
211
5
133
161
6
121
7.
71
8
53
9
44
EFFECT OF WATER UPON THE STRENGTH OF MORTAR.
Another feature plainly noticeable in mortars is the effect of the
proportion of water used upon the strength of the mortar. Too much
or too little water greatly reduces the strength of the mortar. Each
sand and each cement influences the amount of water necessary to make
the strongest mortars. In general, fine sands and loamy sands require
more water than coarser and cleaner sands. Natural cements require
more water than Portland cements. Mortars of i sand to i cement re-
quire more water than mortars having greater proportions of sand. Too
little water making a stiff mortar, will increase the cost of working with
it during construction and will decrease the perfect crystallization,
thus decreasing the strength. Too much water acts as a dilutant, leaving
the mortar porous when hardened and consequently not so strong as
dense mortar. When mortar is placed where it will get very little ad-
ditional moisture over that used in mixing it, sufficient water should be
used to thoroughly hydrate the cement.
The writer in making laboratory tests found that a set of briquettes
made from standard sand and Portland cement with 123/^ per cent, of
water, lost one-sixth of the water in the 24 hours that the briquettes
remained in the air, although they were covered with a dampened cloth.
Digitized by
Google
STATE GEOLOGIST. 31
They regained one-half of the lost water after five days under water. If
so much water is lost under such conditions, it is readily seen what a
serious loss must occur in. actual practice unless unusual care is taken to
protect the mortar surface. There is no question but that this loss of
water seriously impairs the strength of the mortar.
The three sets of tests with fine sand in table number 5 illustrate
the effect obtained by using varying percentages of water. The table
shows that from 7 to 10 per cent, of tensile strength is lost by using too
much water, and from 30 to 40 per cent, by using too little. This illu-
strates the necessity of properly proportioning the water to the cement
and sand used. The water used was proportioned by weight to the com-
bined weight of sand and cement.
PERMEABIUTY OF CEMENT MORTAR.
Work of Melick and Shepherd.— In 1899 and 1900, * Messrs. N. A.
Melick and C. W. Shepherd, students in the Ohio State University, carried
on, under the direction of the late Prof. C. N. Brown, a series of tests to
determine the permeability of cement mortar. The question they sought
to answer in reference to permeability, propounded by Mr. Julian Griggs.
M. Am. Soc. C. E., Chief Engineer of the City of Columbus, Ohio, was
this : "Is it necessary to use soap and alum and Silica Portland cement, or
can a cheaper mixture be made impermeable."
The experiments were carried on with three kinds of sand ; standard
sand or crushed quartz passing a No. 20 sieve and being held upon a No.
30 sieve, lake sand, and a mixture of equal parts of lake sand and quartz.
Medusa, Dyckerhoff and Silica Portland cement were used. The water
pressure was applied to the cement mortar by means of a ^ inch pipe
threaded into the back of a metal disc, 15 inches in diameter. Bolted to
the face of this disc was a flanged metal collar, 10 inches in diameter.
Within the collar was rammed from 2 to 3^ inches of the mortar, the
permeability of which was to be tested. A neat cement filler % inch
to 34 inch thick was troweled around the ring to prevent water leaking
between the collar and the concrete. Figure 2 shows a section of collar
filled with mortar ready to bolt on to the disc. The mortar was mixed
so dry that water would just flush to the surface upon thorough ram-
ming. Great care was taken to get uniform ramming over each portion
of the disc ; the rammer being constructed in the form of a double sector
just fitting within the metal collar.
After investigation of actual work being done in Columbus, O., the
young men estimated that in ramming, 8 foot-pounds of work were done
upon each square inch of surface of the concrete which the city was
laying. Hence they designed the rammer so that they could easily do a
similar amount of work upon the ramming of the mortar into these collars.
•Taken from the thesis of MesRrs. N A. Melick and 0. W. Shepherd, civil engineering grad-
uates from the Ohio State University in 1900. The work was done at the suggestion and expense
of the Bngineers' Club, of Columbus, Ohio.
Digitized by
Google
32
ANNUAL REPORT
The mortar collars were kept in moist air for twenty-four hours and
then stored in water for twenty-seven days before being tested. The
water pressure in testing varied from 25 to 65 pounds per square inch.
It was found that discs which did not leak in twenty- four hours would
never leak. To be considered permeable, it was decided that drops of
water must collect upon the outer surface of the mortar disc. Where
discs became damp, but sufficient moisture did not come through to collect
into drops, they were considered to be porous, but not permeable.
Gavnftnt Filler
mimmr^
1^. _.o- ^
Section of Metol Collar witk Concrete Di5c
Figure 2.
The conclusions which they reached were :
First — In plain mortars, permeability depends upon the voids in the
sand. A mortar not poorer than i cement to 2 sand will not leak, no
matter what kind of sand is used. Mortars of i cement to 4 sand will be
impermeable if made of a normal mixture of sand, that is, a sand having
the normal variation in size of grains. Any mortar will become imperme-
able if the water acting against its face contains suspended matter.
Second. — Cement Coatings. — One quarter inch coating of neat ce-
ment will make mortar impermeable.
Third. — Soap and Alum. — Applications of soap and alum on very
permeable mortars do not justify the expense. It is better to procure sand
with less voids, or use a richer mortar. Soap and alum used in the mortar
do not make it impermeable, at least from the beginning.
Work of Kettlcr and Sherman^— In 1901, Mr. F. C. Kettler and J. K.
Sherman, students of the Ohio State University, continued the investiga-
tions upon the permeability of cement mortars, using varying percentages
of water in mixing the mortars, and using loam in the sand. They also
used grout washes on the surface to prevent permeability. The thickness
of the mortar tested was i^ inches. Their method of ramming and test-
ing was the same as in the previous tests. Their experiments with loam
did not show any definite results effected. Loam with standard sand was
very permeable, less so with lake sand.
In the use of a varying quantity of water used in mixing the mortar,
it was found that mortar of standard sand was quite permeable, the
permeability decreasing as the percentage of water in the mortar increased
from 8 to 14 per cent. The permeability of the lake sand mortar was;
Digitized by
Google
STATE GEOLOGIST. 33
much less than with the quartz, but showed no regularity with variations
in the use of water in the mortar. The lowest permeability with lake sand
mortar was obtained when 14 to 16 per cent, of water was used in the
mortar. The wetter mortar required less tamping and more mortar to
fill the collar, thus indicating a denser mixture.
In testing grout washings, the clear mortar disc was first tested under
pressure, then washed with a grout of one part cement and one part water,
allowed to set 24 hours and placed under pressure. Another washing was
applied, the disc allowed to set 24 hours and tested again. According to
the tests the first wash was most beneficial.
The final conclusions were :
The permeability of mortar depends upon —
1st. The ratio of sand to cement.
2nd. The voids in the sand.
3rd. The percentage of water used in making the mortar.
4th. The thickness of mortar..
5th. The head of water pressure.
6th. The amount of tamping.
The permeability can not be materially reduced by the application of
soap and alum solutions, or by finely powdered loam used in the sand.
Permeability can be reduced —
1st. By the application of i to 5 coats of cement grout, the reduc-
tion amounting to from 70 to 98 per cent, of the initial leakage.
2nd. By a coating of neat cement mortar J4 i^ch thick.
3rd. By the mortar surface standing under a head of water con-
taining suspended matter.
LOAMY SAND.
The majority of engineers specify that the sand shall be clean and
sharp. A series of tests carried on at the Ohio State University under the
direction of Prof. C. E. Sherman, also the results obtained by a class in
cement testing under the writer, seems to prove that clay or loam up to
15 per cent, of the weight of the sand adds strength to the mortar. If
additional tests under the varied conditions arising in practice continue
to prove satisfactory, this will mean a great economy in many pieces of
work where bank sand can be substituted for lake or washed sand.
The series of tests referred to were carried on during the winters of
1901, 1902 and 1903, by eight students of the Ohio State University,
grouped two' and two in investigations upon the effect of clay and loam in
sand upon cement mortar. Each thesis embraced the results of such tests
with Dyckerhpflf cement and some standard American Portland. Th£ mor-
tar was made of i part cement and 3 parts sand, a definitepercentageof
3— s. G.
Digitized by
Google
34 ANNUAL REPORT
the sand being clay or loam. Each thesis included three series of tests,
one with standard quartz sand, one with lake sand, and one with equal
proportions of the two mixed. In each series, separate tests were made
with clay and with- loam, used in the following proportions by weight of
the sand, namely, 2, 4, 6, 8, 10, 12 and 15 per cent. A total of about ten
thousand briquettes were broken. The same conclusions were reached
in each of the four theses written, namely, that either clay or loam added
to sand up to the limit of 15 per cent, by weight, did not have any injuri-
ous effect upon Portland cement mortar after the first two weeks and up
to the limit of time which the tests covered, namely, 12 weeks.
On the contrary the clay and loam added considerably to the strength
of the mortar. The results differed with different cements, and with the
different sands. Their conclusions were that the chemical composition of
the cement influenced the results to some degree. In some cases 10 per
cent, seemed to give maximum results, but the majority of the tests