(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Cement and concrete"

CEMENT AND CONCRETE 



BY 



LOUIS CARLTON SABIN, B. S., C. E. 

i' 

ASSISTANT ENGINEER, ENGINEER DEPARTMENT, U. S. ARMY ; MEMBER OF THE 
AMERICAN SOCIETY OF CIVIL ENGINEERS 



SECOND EDITION, REVISED AND ENLARGED 



NEW YORK 

McGRAW PUBLISHING COMPANY 
1907 






V 




COPYBIGHTED, 1904, 1907 
BY 

L. C. SABIN 



Stanbopc rPrccs 

H. OIL80N COMPANY 
BOSTON. U.S.A. 



C. F. 



PREFACE 



THAT the use of cement has outstripped the literature on 
the subject is evidenced by the number and character of the 
inquiries addressed to technical journals concerning it. This 
volume is not designed to fill the proverbial "long felt want/' 
for until within a few years the number of engineers using 
cement in large quantities was quite limited. These American 
pioneers in cement engineering, under one of whom the author 
received his first practical training in this line, needed no formal 
introduction to the use and properties of cement; their knowl- 
edge was born and nurtured through intimate association and 
careful observation. 

To-day the young engineer frequently finds a good working 
knowledge of cement one of the essentials of success, and the 
gaining of this knowledge by experience alone is likely to be 
too slow and expensive, judged by twentieth century standards. 
In fact, the variety and extent of the uses to which cement is 
applied, and the knowledge concerning its properties, have of 
late increased so rapidly that even the older engineer, whose 
practice may have directed his special attention along other 
channels for a few years, finds it difficult to follow its progress. 

One who wishes only a catechetical reply to any question 
that may arise concerning cement and its use will be somewhat 
disappointed in these pages; on the other hand, he who would 
devote special attention to the subject must, of course, go far 
beyond them. The author has attempted to take a middle 
course, avoiding on the one hand a dogmatic statement of facts, 
and on the other too detailed and extended series of tests, but 
giving, where practicable, sufficient tests to support the state- 
ments made, and endeavoring to show the connection between 
theory and practice, the laboratory and the field. 

The original investigations forming the basis of the work 
were made in connection with the construction of the Poe 
Lock at St. Marys Falls Canal, Michigan, under the direction 



359470 



iv PREFACE 

of the Corps of Engineers, U. S. Army. To the late General 
0. M. Poe, the Engineer officer in charge of the district at that 
time, and to Mr. E. S. Wheeler, his chief assistant engineer, 
may be credited a very large share of the value of the results 
obtained, since the accomplishment of a series of experiments 
of so comprehensive a character was made possible only through 
the broad views held by them as to the value of thorough tests 
of cement. 

The author wishes to express his appreciation of the courtesy 
of General G. L. Gillespie, Chief of Engineers, U. S. A., in grant- 
ing permission to use the data collected, and of the kindness 
of Major W. H. Bixby in presenting a request for this per- 
mission. 

When not otherwise stated, the tables in the work are con- 
densed from the results of the above mentioned investigations. 
In supplementing this original matter, much use has been made 
of the experiments of others as published in society transac- 
tions, technical journals, etc., to all of whom credit has been 
given in the body of the work. 

If this attempt to place in one volume a connected story of 
the properties and use of cement serves to make the road to 
this knowledge a little less devious than that followed by the 
writer, the latter will be rewarded. 

L. C. S. 

SAULT STE. MARIE, MICH. 
January 3, 1905. 



CONTENTS 



PART L CEMENT: CLASSIFICATION AND 
MANUFACTURE 

CHAPTER I. DEFINITIONS AND CONSTITUENTS 

PAGE 

ART. 1. GENERAL CLASSIFICATION OF HYDRAULIC PRODUCTS .... 1 

ART. 2. LIME: COMMON AND HYDRAULIC 3 

ART. 3. PORTLAND CEMENT 4 

ART. 4. SLAG CEMENT 7 

ART. 5. NATURAL CEMENT 10 

CHAPTER II. MANUFACTURE 

ART. 6. MANUFACTURE OF PORTLAND CEMENT 12 

Materials. Wet Process. Dry Process. Semi-dry Process. 
Details of the Manufacture : Burning, Grinding. Sand-Cement. 

ART. 7. OTHER METHODS OF MANUFACTURE OF PORTLAND .... 35 

ART. 8. MANUFACTURE OF SLAG CEMENT 36 

ART. 9, MANUFACTURE OF NATURAL CEMENT 37 



PART II. PROPERTIES OF CEMENT AND 
METHODS OF TESTING 

CHAPTER III. INTRODUCTORY 
Desirable Qualities. Uniform Methods of Testing 42 

CHAPTER IV. CHEMICAL TESTS 
ART. 10. COMPOSITION AND CHEMICAL ANALYSIS 45 

CHAPTER V. THE SIMPLER PHYSICAL TESTS 

ART. 11. MICROSCOPICAL TESTS. COLOR 50 

ART. 12. WEIGHT PER CUBIC FOOT, OR. APPARENT DENSITY .... 51 
ART. 13. SPECIFIC GRAVITY, OR TRUE DENSITY 53 



vi CONTENTS 

CHAPTER VI. SIFTING AND FINE GRINDING 

PAGE 

ART. 14. FINENESS . 59 

Importance of Fineness. Sieves. Methods. Specifications. 

ART. 15. COARSE PARTICLES IN CEMENT . . . ' 66 

Effect on Weight, Time of Setting and Tensile Strength. 

ART. 16. FINE GRINDING 72 

Effect on Weight, Time of Setting and Tensile Strength. 

CHAPTER VII. TIME OF SETTING AND SOUNDNESS 

ART. 17. SETTING OF CEMENT 79 

Process of Setting. Rate. Variations in Rate. 
ART. 18. CONSTANCY OF VOLUME 90 

Causes of Unsoundness. Tests. Discussion of Methods. Hot 

Tests for Natural Cements. Conclusions. 

CHAPTER VIII. TESTS OF THE STRENGTH OF CEMENT 
IN COMPRESSION, ADHESION, ETC. 

ART. 19. TESTS IN COMPRESSION AND SHEAR 103 

ART. 20. TESTS OF TRANSVERSE STRENGTH 104 

ART. 21. TESTS OF ADHESION AND ABRASION 106 

CHAPTER IX. TENSILE TESTS OF COHESION 

ART. 22. SAND FOR TESTS 109 

Value of Tests of Sand Mortars. Uniformity in Sand. Com- 
parison of Different Kinds. Tests with Natural Sand. Fineness. 

ART. 23. MAKING BRIQUETS Ill 

Proportions. Consistency. Temperature. Gaging: Hand and 
Machine. Methods. Amount of Gaging, Form of Briquets. 
Molds. Molding. Briquet Machines. Approved Methods 
of Hand Molding. Marking the Briquets. 

ART. 24. STORING BRIQUETS 131 

Time in Air before Immersion. Moist Closet. Water of Im- 
mersion. Storing in Air; in Damp Sand. 

ART. 25. BREAKING THE BRIQUETS 137 

Testing Machines. Clips. Clip-breaks. Comparative Tests of 
Clips. Requirements for a Perfect Clip. Form Recommended. 
Rate of Applying Tensile Stress. Treatment of Results. 

ART. 26. INTERPRETATION OF TENSILE TESTS OF COHESION .... 151 

CHAPTER X. RECEPTION OF CEMENT AND RECORDS 
OF TESTS 

ART. 27. STORING AND SAMPLING 158 

Storage Houses. Percentage of Barrels to Sample. Method of 
Taking and Storing the Sample. 



CONTENTS vii 

PAGE 

ART. 28. RECORDS OF TESTS 160 

Value of Records. Marking Specimens. Records at St. Marys 
Falls Canal. 



PART III. THE PREPARATION AND PROP- 
ERTIES OF MORTAR AND CONCRETE 

CHAPTER XI. SAND FOR MORTAR 

ART. 29. CHARACTER OF THE SAND 168 

Shape and Hardness of the Grains. Siliceous vs. Calcareous 
Sands. Slag Sand. Sand for Use in Sea Water. 

ART. 30. FINENESS OF SAND 173 

Relation Volume and Superficial Area. Effect of Fineness. 

ART. 31. VOIDS IN SAND 176 

Conditions Affecting Voids: Shape of Grains; Granulometric Com- 
position. Effect on Tensile Strength of Mortar. Moist Sand. 

ART. 32. IMPURITIES IN SAND 182 

ART. 33. CONCLUSIONS. WEIGHT AND COST OF SAND 184 

CHAPTER XII. MORTAR: MAKING AND COST 

ART. 34. PROPORTIONS OF THE INGREDIENTS 186 

Capacity of Cement Barrels. Equivalent Proportions by Weight 
and Volume. Richness of Mortars. Effect of Pebbles. Con- 
sistency. 

ART. 35. MIXING THE MORTAR 191 

Hand Mixing. Machine Mixing. 

ART. 36. COST OF MORTARS 193 

Ingredients Required. Tables of Quantities. Estimates of 
Cost. Tables of Cost of Portland and Natural Cement Mortars. 

CHAPTER XIII. CONCRETE: AGGREGATES 

ART. 37. CHARACTER OF AGGREGATES 200 

Proper Materials. Screenings in Broken Stone. Foreign 

Ingredients. 
ART, 38. SIZE AND SHAPE OF FRAGMENTS AND VOLUME OF VOIDS . 202 

Conditions Affecting Voids. Effect on Strength of Concrete. 

Gravel vs. Broken Stone. 
ART. 39. STONE CRUSHING AND COST OF AGGREGATE 208 

Breaking Stone by Hand. Stone Crushers. Cost of Aggregate. 

Examples. 



viii CONTENTS 

CHAPTER XIV. CONCRETE MAKING: METHODS 

AND COST 

PAGE 

ART. 40. PROPORTIONS OF THE INGREDIENTS 214 

Theory of Proportions. Determination of Amount of Mortar 

Required. Aggregates Containing Sand. Required Strength. 
ART. 41. MIXING CONCRETE BY HAND 217 

Hand vs. Machine Mixing. Method of Hand Mixing ; Number of 

Men and Output ; Examples. 
ART. 42. CONCRETE MIXING MACHINES 221 

General Classification. Description of Machines. Basis of 

Comparison. 

ART. 43. CONCRETE MIXING PLANTS AND COST OF MACHINE MIXING 226 
ART. 44. COST OF CONCRETE 232 

Ingredients Required for a Cubic Yard. Examples of Actual Cost. 

CHAPTER XV. THE TENSILE AND ADHESIVE STRENGTH 
OF CEMENT MORTARS AND THE EFFECT OF VARIA- 
TIONS IN TREATMENT 

ART. 45. TENSILE STRENGTH OF MORTARS OF VARIOUS COMPOSITIONS 

AND AGES 241 

ART. 46. CONSISTENCY OF MORTAR AND AERATION OF CEMENT . . 246 

ART. 47. REGAGING OF CEMENT MORTAR 250 

ART. 48. MIXTURES OF CEMENT WITH LIME, PLASTER PARIS, ETC. . 257 
Mixtures of Portland and Natural. "Improved" Cement. 
Ground Quicklime with Cement; Slaked Lime; Plaster of Paris. 
Conclusions. 

ART. 49. MIXTURES OF CLAY AND OTHER MATERIALS WITH CEMENT 267 
Effect of Powdered Limestone, Brick, etc. ; Sawdust ; Terra Cotta. 

ART. 50. USE OF CEMENT MORTARS IN FREEZING WEATHER .... 274 
Effect of Frost on Set Mortars. Effect of Salt; Heating Materials; 
Consistency; Fineness of Sand. Conclusions. 

ART. 51. THE ADHESION OF CEMENT 284 

Adhesion between Portland and Natural. Adhesion to Stone and 
Other Materials. Effect of Consistency; Regaging; Character of 
Surface of Stone. Effect of Plaster of Paris. Adhesion to Brick; 
Effect of Lime Paste. Adhesion to Rods of Iron and Steel. 

CHAPTER XVI. COMPRESSIVE STRENGTH AND MODULUS 
OF ELASTICITY OF MORTAR AND CONCRETE 

ART. 52. COMPRESSIVE STRENGTH OF MORTARS ...:...... 302 

Ratio of Compressive to Tensile Strength. 
ART. 53. CONCRETES WITH VARIOUS PROPORTIONS OF INGREDIENTS . 305 

Effect of Consistency; Amount and Richness of Mortar; Methods 

of Storage. 



CONTENTS ix 

PAGE 
ART. 54. CONCRETES WITH VARIOUS KINDS AND SIZES OF AGGREGATES 312 

ART. 55. CINDER CONCRETE AND EFFECT OF CLAY 316 

ART. 56. MODULUS OF ELASTICITY OF CEMENT MORTAR AND CONCRETE 320 



CHAPTER XVII. THE TRANSVERSE STRENGTH AND 
OTHER PROPERTIES OF MORTAR AND CONCRETE 

ART. 57. TRANSVERSE STRENGTH 327 

Transverse Strength of Mortars Compared to Tensile and Com- 
pressive Strength. Richness of Mortar; Consistency. Transverse 
Tests of Concrete Bars: Variations in Mortar Used; Consistency; 
Mixing ; Aggregate ; Screenings. Deposition in Running Water. 
Use in Freezing Weather. 

ART. 58. RESISTANCE TO SHEAR AND ABRASION 342 

ART. 59. EXPANSION AND CONTRACTION OF CEMENT MORTAR, AND 

THE RESISTANCE OF CONCRETE TO FIRE 345 

Change in Volume during Setting. Coefficient of Expansion of 
Mortar and Concrete. Fire-Resisting Qualities of Concrete. 
Aggregate for Fireproof Work. 

ART. 60. PRESERVATION OF IRON AND STEEL BY MORTAR AND CONCRETE 350 
Action of Corrosion. Tests of Effect of Concrete. 

ART. 61. POROSITY, PERMEABILITY, ETC 354 

Porosity. Permeability. Waterproof Mortars and Concretes. 
Washes for Exteriors of Walls. Efflorescence. Pointing Mortar. 
Cements in Sea Water. 



PART IV. USE OF MORTAR AND CONCRETE 

CHAPTER XVIII. CONCRETE: DEPOSITION 

ART. 62. TIMBER FORMS OR MOLDS 365 

Sheathing. Lining. Posts and Braces. 

ART. 63. DEPOSITION OF CONCRETE IN AIR 372 

Transporting, Depositing, Ramming. Rubble Concrete. Fin- 
ish; Plastering; Facing; Bushhammering ; Colors for Concrete Finish. 

ART. 64. PLACING CONCRETE UNDER WATER 383 

Laitance. Tremie, Skip, etc. Depositing in Bags ; Cost. 
Block System: Molds; Cost. 

CHAPTER XIX. CONCRETE-STEEL 

ART. 65. MONIER SYSTEM . . . 395 

ART. 66. WUNSCH, MELAN, AND THACHER SYSTEMS 397 

ART. 67. OTHER SYSTEMS OF CONCRETE-STEEL 399 

Hennebique, Kahn, Ransome, Roebling, Expanded Metal. 

ART. 68. THE STRENGTH OF COMBINATIONS OF CONCRETE AND STEEL 401 



x CONTENTS 

PAGE 
ART. 69. BEAMS WITH SINGLE REINFORCEMENT 404 

Formulas for Constant Modulus Elasticity; for Varying Modulus. 

Excessive Reinforcement. Tables of Strength. 

ART. 70, BEAMS WITH DOUBLE REINFORCEMENT 417 

ART. 71. SHEAR IN CONCRETE-STEEL BEAMS 419 

CHAPTER XX. SPECIAL USES OF CONCRETE: BUILD- 
INGS, WALKS, FLOORS, AND PAVEMENTS 

ART. 72. BUILDINGS 424 

Roof ; Floor System ; Columns. Building Forms. N. Y. Build- 
ing Regulations. 

ART. 73. WALKS 434 

Foundation; Base; Wearing Surface; Construction; Cost. 
ART. 74. FLOORS OF BASEMENTS, STABLES, AND FACTORIES .... 440 

ART. 75. PAVEMENTS AND DRIVEWAYS 442 

Pavement Foundations. Concrete Wearing Surface. Construc- 
tion. Example. 

ART. 76. CURBS AND GUTTERS 445 

ART. 77. STREET RAILWAY FOUNDATIONS 447 

CHAPTER XXI. SPECIAL USES OF CONCRETE (CONTINUED) : 
SEWERS, SUBWAYS, AND RESERVOIRS 

ART. 78. SEWERS 450 

Methods and Cost. Forms. 
ART. 79. SUBWAYS AND TUNNEL LINING 457 

Waterproofing. Subways. Tunnels in Firm Earth; in Soft 

Ground ; in Rock. Examples ; Methods ; Cost. 
ART. 80. RESERVOIRS: LININGS AND Roo^s 467 

Details of Construction. Groined Arch. Forms. Examples; 

Cost. 

CHAPTER XXII. SPECIAL USES OF CONCRETE (CONTINUED): 
BRIDGES, DAMS, LOCKS, AND BREAKWATERS 

ART. 81. BRIDGE PIERS AND ABUTMENTS AND RETAINING WALLS . 478 

Bridge Piers ; Steel Shells. Repair of Stone Piers. Retaining 

Walls and Abutments: Coping; Rules for Use of Concrete. 
ART. 82. CONCRETE PILES 485 

Building in Place. Concrete-Steel Piles: Molding; Driving. 
ART. 83. ARCHES 488 

Design; Centers; Construction; Finish and Drainage. Examples 

and Cost, 
ART. 84. DAMS 498 

Concrete vs. Rubble. Quality of Concrete. - Construction. - 

Examples. 
ART. 85. LOCKS 502 

Me'thods of Building. Examples. 
ART. 86. BREAKWATERS . 50? 



CONTENTS xi 

CHAPTER XXIII. CONCRETE BUILDING BLOCKS: THEIR 

MANUFACTURE AND USE 

PAGE 

ART. 87. GENERAL METHODS OF MANUFACTURE 512 

Tamped Blocks; Machine; Curing. Pressed Blocks. Poured 
Blocks. 

ART. 88 MATERIALS AND FINISH 520 

Proportions. Special Forms. Freezing. Waterproofing. 

ART. 89. COST AND LAYING 526 

Volume. Weight. Cost. Laying. Comparison of Concrete 
Block and Brick. 

ART. 90. STRENGTH OF CONCRETE BLOCKS AND BUILDING REGULA- 
TIONS 533 

APPENDIX I. 
Uniform Methods of Testing suggested by the Am. Soc. C. E. Committee. 

APPENDIX II. 

Standard Specifications for Cement adopted by the American Society for 
Testing Materials. 

APPENDIX III. 

Methods of Analysis suggested by the New York Section of the American 
Chemical Society, 



PART I 
CEMENT 

CLASSIFICATION AND MANUFACTURE 



CHAPTER I 

DEFINITIONS AND CONSTITUENTS 
ART. 1. GENERAL CLASSIFICATION OF HYDRAULIC PRODUCTS 

1. The use of a cementitious substance for binding together 
fragments of stone is older than history, and it is known that the 
ancient Romans prepared a mortar which would set under 
water. So far as our present knowledge of cement manufac- 
ture is concerned, however, the credit of demonstrating that a 
limestone containing clay possessed, when burned and ground, 
the property of hardening under water, is due to Mr. John 
Sm'eaton, who announced this as the result of his experiments 
made in 1756 in seeking a material with which to build the 
Eddystone Lighthouse. After this discovery by Smeaton nearly 
sixty years elapsed before M. Vicat gave the true explanation 
of this action, namely, that the lime during burning combined 
with the silica to form silicate of lime, the essential ingredient 
of hydraulic limes and cements. 

In 1796, Parker, of London, obtained a patent for the manu- 
facture of a cement from septaria nodules, and aptly named his 
product " Roman Cement." In 1824, Joseph Aspdin of Leeds, 
England, patented a process of manufacture of ''Portland 
Cement." 

2. The cements in general use in the United States to-day 
are of two kinds, Portland cements and natural cements, and in 
what follows our attention will be directed almost entirely to 
these two products. 

Common limes were formerly used largely in engineering 
construction, but have of late been almost entirely superseded, 



CEMENT AND CONCRETE 

for this purpose, by cements. Since the hardening of lime 
mortar depends on the absorption of carbonic acid from the 
atmosphere, these limes are sometimes called "air limes," while 
the hydraulic products which set under water are, for a similar 
reason, styled " water limes/' Hydraulic limes, though playing 
an important role in foreign countries, are not manufactured or 
used to any extent in the United States. The European prod- 
uct known as " Roman" or "Vassy" cement, somewhat re- 
sembles our natural cement, but is usually inferior to the Ameri- 
can article. Our chief interest in these products, which are used 
only abroad, is to know what relation they bear to the cements 
with which we are familiar. The following classifications are 
selected as being authoritative: 

3. The conferences of Dresden (1886) and Munich (1884) on 
Uniform Methods of Testing for Materials of Construction, clas- 
sified the hydraulic products as follows : - 

(1) Hydraulic limes: made by roasting either argillaceous or 
siliceous limestones. They slake partially or wholly on the ad- 
dition of water. 

(2) Roman cements: made from argillaceous limestones hav- 
ing a large proportion of clay. They do not slake by the addi- 
tion of water and hence must be mechanically ground to powder. 

(3) Portland cements: obtained by burning to the point of 
insipient vitrificatio'n either hydraulic limestones or mixtures of 
argillaceous materials and limestones, and afterward grinding 
the product to fine powder. 

(4) Hydraulic gangues: natural or artificial materials which 
do not harden alone, but which furnish hydraulic mortars when 
mixed with quicklime. 

(5) Pozzolana cements produced by an intimate mixture 
of powdered hydrate of lime and finely pulverized hydraulic 
gangues. 

(6) Mixed cements: the products of intimate mixtures of 
manufactured cement with certain materials proper for such a 
purpose. Mixed cements should always be designated as such 
and the materials entering into the composition should be stated, 
but it may be added parenthetically that these things are 
seldom done. 

4. MM. Durand-Claye and Debray divide cements into six 
classes, namely, (1), Grappier cements obtained by grinding. 



LIME 3 

the pieces of hydraulic lime which do not slake; (2), quick-set- 
ting (Vassy) cements formed by burning very argillaceous 
limestones at a low temperature; (3), natural Portland cements, 
or those cements made from natural rock which correspond to 
artificial Portland in character; (4), mixed cements; (5), arti- 
ficial Portlands; and (6), slag cements. 

M. H. LeChatelier, an eminent French authority, divides 
hydraulic products into four classes, namely: 1 Portland ce- 
ments, hydraulic limes, natural cements, and mixed cements. He 
subdivides the third class, natural cements, into quick-setting, 
slow-setting and grappier cements, and includes natural Port- 
lands among the slow-setting natural cements. Slag cements, 
which are put in a separate class by MM. Durand-Claye and 
Debray, are included in " mixed cements" by M. LeChatelier. 

5. Prof. I. 0. Baker gives a classification that is better 
adapted for use in this country than any of the above. 2 He 
divides the products obtained by burning limestone, either pure 
or impure, into lime, hydraulic lime and hydraulic cements. He 
then sub-divides cement into Portland, Rosendale (preferably 
called natural) and Pozzolana. 

ART. 2. LIME: COMMON AND HYDRAULIC 

6. Common lime is the product obtained by burning a pure, 
or nearly pure, carbonate of lime. On being treated with water 
it slakes rapidly, evolving much heat and increasing greatly in 
volume. It is now seldom used in engineering construction and 
will not be considered further. 

7. Prof. M. Tetmajer has thus defined hydraulic limes: Hy- 
draulic limes are the products obtained by the burning of argil- 
laceous or siliceous limestones, which, when showered with water, 
slake completely or partially without sensibly increasing in 
volume. According to local circumstances, hydraulic limes may 
be placed on the market either in lumps, or hydrated and pul- 
verized. The following table gives a classification of hydraulic 
limes according to M. E. Candlot * who states that the first 



1 "Tests of Hydraulic Materials," by H. LeChatelier. Trans. Am. Inst. 
Mining Engrs., 1893. 

2 "Masonry Construction," p. 48. 

3 "Ciments et Chaux Hydrauliques," par E. Candlot. 



CEMENT AND CONCRETE 



class is seldom used for important work and that the fourth 
class is quite rare. 

TABLE 1 
Classification of Hydraulic Limes. E. Candlot 



Class. 


Per Cent, 
of Clay in 
Limestone. 


Per Cent, 
of Silica 
and Alumi- 
na in Fin- 
ished Prod- 
uct. 


Hydraulic 
Index, or 
Ratio of 
Silica and 
Alumina to 
Lime. 


Approx. 
Time to 
Set, 
Days. 


Feebly Hydraulic Lime 
Ordinary " u 
Real " " 


5 to 8 
8 to 15 
15 to 19 


9 to 14 
14 to 24 
24 to 30 


.10 to .16 
.16 to .31 
.31 to .42 


16 to 30 
10 to 15 
5 to 9 


Eminently " u 


19 to 22 


30 to 33 


.42 to .60 


2 to 4 



Hydraulic limes should be burned slowly, and at such a tem- 
perature that sintering does not take place. The best hydraulic 
limes have a composition very similar to that of Portland cement. 
The comparatively low temperature at which they are burned 
permits them to slake on the addition of water. They gain 
strength much more slowly than cements. 

Having considered the classification of hydraulic products as 
a whole, we may proceed to the discussion of Portland and nat- 
ural cements, the hydraulic products which have by far the 
greatest importance here, and the only varieties which will be 
taken up in detail in the present work. 

AKT. 3. PORTLAND CEMENT 

8. DEFINITION. Although questions may still arise as 
to whether a given product is entitled to the name Portland 
cement, yet ideas are now pretty well crystallized as to what 
shall be included under that title. All authorities agree that 
the proportion of the constituents must be confined within 
comparatively narrow limits, that certain ingredients must be 
present only in small quantities, if at all, that the calcination 
must be carried to a point just short of vitrification, and that 
the resulting product must be ground to a fine powder. 

Two definitions recently adopted are worthy of special men- 
tion: That of the Association of German Portland Cement 
Manufacturers specifies that the raw materials shall be inti- 
mately ground, then calcined at a clinkering temperature and 
reduced to proper fineness; the product is to contain not less 



PORTLAND CEMENT 5 

than 1.7 parts lime by weight to each part silica + alumina + 
iron oxide, and its specific gravity is to be not less than 3.10. 

The definition adopted by the Board of Engineer Officers of 
the United States Army in 1901 is as follows : "By a Portland 
cement is meant the product obtained from the heating or cal- 
cining up to incipient fusion of intimate mixtures, either natural 
or artificial, of argillaceous with calcareous substances, the 
calcined product to contain at least 1.7 times as much lime, 
by weight, as of the materials which give the lime its hydraulic 
properties, and to be finely pulverized after said calcination, 
and thereafter additions or substitutions for the purpose only 
of regulating certain properties of technical importance to be 
allowable to not exceeding 2 per cent, of the calcined product." 

It is noticed that neither definition attempts to specify the 
kind of raw materials which shall be used in the manufacture. 
The first definition requires the raw materials to be ground 
before calcination, while the second does not. The question is 
sometimes raised as to whether a product prepared from a 
rock which, as it occurs in nature, contains the proper propor- 
tions of silica and lime shall be designated Portland cement. 
In the United States, cements are made from rocks having so 
nearly the correct proportions of ingredients that one part of a 
ledge may require addition of lime and another part may need 
more clay to correct the mix. If some of the rock is found to 
be correct in composition without addition of either lime or 
silica it is preeminently suited to the manufacture of Portland 
cement. To insure uniform composition and burning, however, 
preliminary grinding is a wise precaution if not absolutely essen- 
tial. A definite understanding of what is meant by the term 
Portland cement will prove a very great convenience, and may 
save law suits, but the engineer must not neglect to ascertain 
whether a given product has the properties he desires under 
whatever name it may appear on the market, and if imposed 
upon he should not blame the definition. 

9. COMPOSITION. As a result of an extended series of 
experiments M. Le Chatelier first published in 1887 1 his conclu- 
sion that the essential composition of Portland cement consists 
of (1) tricalcium silicate, 3CaO, Si0 2 ; (2) tricalcium aluminate, 

1 Annales des Mines, 1887 and 1893. Jour. Soc. Chem. Ind., Mar. 31, 
1891, p. 256. 



6 CEMENT AND CONCRETE 

3CaO, A1 2 3 ; and (3) "a fusible calcium silico-aluminate whose 
chief function is that of a flux during burning to promote the 
necessary chemical reactions." It is generally agreed that the 
first component named is the real cementitious ingredient in 
Portland cement, while Le Chatelier states that the tricalcium 
aluminate is " mainly active during setting, contributing some- 
what to the subsequent hardening." In accordance with this 
theory he considered that the basic constituents should not 
exceed three equivalents of the acid constituents, and proposed 
the following formulas: 

CaO + MgO < 
Si0 2 + A1 2 3 ~ ' 
and 

CaO+MgO > 
Si0 2 - (A1 2 3 + Fe 2 3 ) 

in each case the quantities in the formulas being the chemical 
equivalents of the substances, not weights. 

If weights be substituted for equivalents in the first equation 
it is found that the silica and alumina should not be less than 
.56 the lime and magnesia. This ratio of the acid and basic 
constituents is known as the hydraulic index. In the light of 
recent investigations, however, this index is of limited value. 
One reason for this is that in the above formulas magnesia is 
considered of equal basic value with lime, while Dr. Erdmenger 
and other authorities have shown that magnesia does not 
replace lime in Portland cement mixtures. Later researches, and 
notably those of Messrs. Newberry, l have also shown that the 
lime and alumina probably combine to form the dicalcite alumi- 
nate, 2CaO,Al 2 O 3 , their conclusions being stated as follows: 

" First Lime may be combined with silica in the propor- 
tion of 3 molecules to 1, and still give a product of practically 
constant volume and good hardening properties, though harden- 
ing very slowly. With 3J molecules of lime to 1 of silica the 
product is not sound, and cracks in water. 

"Second. Lime may be combined with alumina in the 
proportion of 2 molecules to 1, giving a product which sets 
quickly, but shows constant volume and good hardening prop- 
erties. With 2J molecules of lime to 1 of alumina the product 
is not sound." 



Jour. Soc. Chem. Ind., Nov., 1897. 



SLAG CEMENT 7 

Since to produce a sound cement all the lime must be com- 
bined, it follows from the above that the total lime must be 
sufficient only to furnish 3 molecules to each molecule of silica 
and 2 molecules to each molecule of alumina. As the mole- 
cular weights of lime, silica, and alumina are 56.1, 60.4, and 
102.2 respectively, the maximum permissible amount of lime 
by weight would be 2.78 times the silica, plus 1.1 times the 
alumina. This does not take account of the ferric oxide, Fe 2 O 8 , 
present. The action of the latter has not been determined 
with certainty, but Professor Newberry considers it should be 
omitted from the calculation. 

Expressing the above statement as a formula we shall have 

per cent lime = 2.78 (% silica) + 1.1 (% alumina), 
or, 

per cent lime 

the value of the lime-acid ratio = . . m rrr -= ^= = : r > 

2.78 (% silica) + 1.1 (% alumina) , > 

must not exceed one. 

The mean value of this ratio for twenty representative Ameri- 
can brands is found to be .915. Values of .76 and .99 respec- 
tively were given by two well-known German brands. 

The above discussion is based upon the theory evolved by sev- 
eral authorities from the results of experiment. Much doubt has 
been thrown upon this theory, however, by the recent investiga- 
tions of Messrs. Day and Shepard in the Geo-physical Laboratory 
of the Carnegie Institution. In the Journal of the American 
Chemical Society for September, 1906, Messrs. Day and Shepard 
presented their results, showing that the only definite compounds 
of lime and silica are the metasilicate CaOSiO 2 and the orthosili- 
cate Ca 2 SiO 4 . This discovery will necessitate the modification of 
present accepted theories, but as yet the investigations have not 
been carried far enough to lead to definite conclusions. 

10. The analyses of thirty-two brands of Portland cement, 
most of them of American manufacture, given in Table 2, are 
selected from various sources and illustrate the proportions 
obtaining in commercial products. 



CEMENT AND CONCRETE 



TABLE 2 
Analyses of Portland Cement 






RAW MATERIALS. 


PRO- 
CESS. 


BRAND. 


LOCATION. 


1 


Limestone and Shale . . . 


s 


Alpena 


Alpena, Mich. 


f 


Cement Rock and Limestone 


D 


Alpha 


Alpha, N. J. 





Chalk and Clay ...... 


W 


Alsens 


Itzehoe, Germany. 


-1 


Cement Rock and Limestone 


D 


Atlas 


Northampton and Cop- 










lay, Pa. 


i 


Marl and Clay 


W 


Brunson 


Bronson, Mich. 


!, 


Marl and Clay 


S 


Buckeye 


Harper, Ohio. 


7 


Cement Rock 


D 


Buckhorn 


Manheim, W. Va. 


8 


Limestone and Clay .... 


D 


Catskill 


Smiths Landing, N. Y. 


9 


Limestone and Clay .... 


D 


Diamond 


Middle Branch, Ohio. 


10 


( Limestone and Clay . . . 
) Limestone and Marl . . 


D ) 

Wf 


Dyckerhoff 


Amoeneburg, Ger. 


11 


Cement Rock and Limestone 


D 


Edison 


New Village, N. J. 


12 


Marl and Clay 


S 


Empire 


Warners, N. Y. 


13 


Marl and Clay 


W 


Germania 


Lehrte, Germany 


14 


Cement Rock and Limestone 


D 


Gtiant 


Egypt, Pa. 


lo 


Limestone and Clay .... 


D 


Hudson 


Hudson, N. Y. 


1(> 


Limestone and Clay .... 


D 


Iron Clad 


Glens Falls, N. Y. 


17 


Chalk and Clay 


W 


K. B. & S. 


England. 


18 


Cement Rock and Limestone 


D 


Lehigh 


Allentown, Pa. 


19 


Marl and Clay 


W 


Medusa 


Bay Bridge, Ohio. 


20 


Marl and Clay 


W 


Medusa 


Syracuse, Ind. 


21 


Dement Rock and Limestone 


D 


Nazareth 


Nazareth, Pa. 


22 


Dement Rock and Limestone 


D 


Northampton 


Stockertown, Pa. 


23 


Limestone and Clay .... 


D 


Did Dominion 


^raigsville, Va. 


24 


Marl and Clay 


W 


Dmega 


Jonesville, Mich. 


25 


Marl and Clay 


W 


Peerless 


Union City, Mich. 


26 


Dement Rock 


D 


R,ed Diamond 


Oemopolis, Ala. 


27 


Dement Rock and Limestone 


D 


Saylors 


^oplay, Pa. 


2S 


Slag and Limestone . . 


D 


Universal 


Chicago, 111. 


29 


Dement Rock and Limestone 


D 


Vulcanite 


Vulcanite, N. J. 


W 


Marl and Clay 


W 


Wolverine 


^oldwater, Mich. 


31 


Clay and Waste Limestone 


W 


^yandotte 


Wyandotte, Mich. 


32 


Chalk and Clay 


W 


Yankton 


Yankton, S. Dak. 



NOTE. Wet process is indicated by W; dry, by D; and semi-wet, 
byS. 

ART. 4. SLAG CEMENT 

11. Slag cement is manufactured to a considerable extent 
in Europe and is beginning to assume some importance in the 
United States. It is a pozzolana cement in which the silica 
ingredient is supplied by blast furnace slag. Pozzolana 
cements have been denned as ''products obtained by intimately 
and mechanically mixing, without subsequent calcination, pow- 
dered hydrates of lime with natural or artificial materials which 
generally do not harden under water when alone, but do so 



SLAG CEMENT 



TABLE 2 (continued) 
Analyses of Portland Cement 



No. 


SiO 2 . 


A1 2 3 . 


Fe 2 3 . 


CaO. 


MgO. 


NaaO, 
K 2 O. 


S0 3 . 


H 2 O& 
Loss. 


1 


20.26 


8.62 


2.71 


63.22 


2.34 


1.34 


0.76 




2 


22.62 


8.76 


2.66 


61.46 


2.92 




1.52 




3 


24.90 


8.00 


3.22 


59.38 


0.38 


o.'so 


0.98 


2.16 


4 


21.30 


7.65 


2.85 


60.95 


2.95 


1.15 


1.81 


1.41 


5 


22.90 


6.80 


3.60 


63.90 


0.70 


1.10 


0.40 


0.60 


6 


21.30 


6.95 


2.00 


62.30 


1.20 




0.98 


4.62 


7 


23.72 


6.85 


3.41 


61.54 


1.44 




1.72 


0.90 


8 


23.44 


6.35 


3.99 


63.21 


1.15 




1.22 




9 


21.24 


7.85 


4.14 


63.22 


0.28 


o.'es 


1.11 


1.32 


10 


19.35 


7.00 


4.50 


63.75 








5.40 


11 


20.14 


7.51 


3.33 


62.71 


2.34 




1.64 




12 


22.04 


6.45 


3.41 


60.92 


3.53 




2.25 




13 


21.14 


6.30 


2.50 


66.04 


1.11 






2.91 


14 


23.36 


8.07 


4.83 


58.93 


1.00 


0.50 


0.50 


2.46 


15 


21.60 


9.27 


3.80 


61.78 


1.27 




1.53 




16 


21.50 


10. 


50 


63.50 


1.80 


0.40 


1.50 


0.80 


17 


19.75 


7.48 


5.01 


60.71 


1.28 


0.75 


0.97 


3.38 


18 


22.26 


9. 


49 


62.46 


2.65 




1.56 




19 


23.20 


7.03 


2.41 


64.19 


0.97 






2.20 


20 


22.06 


4.80 


1.66 


65.44 


3.82 




0.90 




21 


19.06 


7.47 


2.29 


61.23 


2.83 


1.41 


1.34 


4.52 


22 


21 18 


7.03 


2 41 


63 06 


2 05 








23 


21.20 


7.90 


2.82 


63.14 


2.40 




1.37 




24 


22.24 


7.26 


2.54 


64.96 


2.26 




0.41 


0'33 


25 


20.30 


7.60 


3.40 


62.60 


1.70 


1.70 


1.30 


1.10 


26 


20.54 


8.55 


3.84 


63.85 


0.66 






1.34 


27 


21.25 


8.25 


4.21 


61.25 


1.50 


2.00 


i.'ss 




28 


23.62 


8.21 


2.71 


61.92 


1.78 




1.32 


0.'52 


29 


21.08 


7.86 


2.48 


63.68 


2.62 




1.25 




30 


21.22 


7.51 


3.83 


63.75 


0.82 




1.58 


1.02 


31 


23.20 


8.00 


2.40 


62.10 


2.00 




0.90 


0.80 


32 


22.00 


7.74 


4.61 


59.50 


0.90 


1.20 


0.80 





when mixed with hydrates of lime (such materials being pozzo- 
lana, Santorin earth, trass obtained from volcanic tufa, furnace 
slag, burnt clay, etc.), the mixed product being ground to ex- 
treme fineness." l 

Slag cement somewhat resembles Portland in its properties, 
but is more like some of the natural cements in its constituents, 
while in the manner of occurrence of these constituents and in 
the method of manufacture it is quite different from either of 
these classes. 



1 "Report of Board of Engineers on Steel Portland Cement," Washing- 
ton, 1900. 



10 CEMENT AND CONCRETE 

12. As this cement is a mixture of lime and pozzolanic 
materials, its value depends largely upon its extreme fineness 
and the intimate mixture of the ingredients. Its specific gravity 
is low, about 2.7 to 2.8, and it sets very slowly, although the 
setting may be hastened by the addition of certain substances 
such as caustic soda. On account of the sulphide present, 
most slag cements are not suited to use in air, as they crack 
and soften in this medium; neither are they suitable for use in 
sea water, nor in freezing weather, but when mixed with two 
or three parts of sand and kept constantly wet with fresh water, 
they give quite satisfactory results. 

Slag cement has an approximate composition of silica, 20 to 
30 per cent., alumina, 10 to 20 per cent., and lime, 40 to 50 per 
cent. It usually contains calcium sulphide, the amount some- 
times reaching three or four per cent. The characteristic green- 
ish tint which slag cements exhibit when they harden in water 
is due to this ingredient, as is the odor of hydrogen sulphide 
sometimes given off by a briquet when broken, especially if it 
has hardened in sea water. Some slag cements have also quite 
a percentage of magnesia. 1 

ART. 5. NATURAL CEMENT 

13. Natural cement, as its name implies, is made from rock 
as it occurs in nature. Argillaceous limestones or argillo- 
magnesian limestones, having the proper proportion of clay, 
magnesia and lime, may be used for the production of natural 
cement. The burning is not carried so far as in the manufac- 
ture of Portland cement, and the resulting product is of lighter 
weight and usually quicker setting, though some natural ce- 
ments are quite slow setting. The properties of these cements, 
coming from different localities, vary greatly. In fact, it is 
difficult to distinguish some natural cements from Portland, 
and they may be considered to grade into the natural Portlands. 
Light burning in manufacture, light weight per cubic foot, and 
slower rate of acquiring strength, may be considered the dis- 
tinguishing characteristics froip a physical point of view. 

1 For an excellent resume of the qualities and distinguishing character- 
sties of slag cements, the reader is referred to "Report of Board of Engi- 
neers on Steel Portland Cement as used in United States Lock at Plaque- 
mine, La." Washington, 1900. 



NATURAL CEMENT 



11 



14. Analyses. Table 3 gives the results of a number of 
analyses of natural cement, compiled from various sources. 

TABLE 3 
Analyses of Natural Cements 



Brand. 


Locality. 


| 

i 


Alumina. 


Iron 
Oxide. 


Lime. 


Magnesia. 


Potash and 
Soda. 


Carbonic AcidJ 
Water, and I 
Loss. 1 


Bonneville Improved . 


A 


30.40 


10.36 


2.60 


52.12 


0.21 




3.07 


Round Top 


B 


28.02 


10.20 


8.80 


44.48 


1.00 


o.'so 


7.00 


Cumberland 


C 


28.30 


10.12 


4.42 


49.60 


3.76 






Double Star 


D 


23.32 


6.99 


5.97 


53.96 


7.76 




2.00 


Black Diamond . . . 


E 


22.54 


8.24 


2.14 


42.31 


5.39 


2.' 82 


. 


Fern Leaf 


F 


26.40 


6.28 


1.00 


45.22 


9.00 


4.24 


7.86 


Hulme 


G 


25.28 


7.85 


1.43 


44.65 


9.50 


4.25 


7.04 


Star 


H 


23.29 


5.96 


2.16 


41.28 


15.39 


1.98 




Norton 


I 


27.98 


7.28 


1.70 


37 59 


15.00 


7.96 


2.49 


Rosendale 


J 


26.54 


5. 


89 


45.30 


17.06 






Brooklyn Bridge . . . 


K 


30.84 


7.75 


2.11 


34.49 


17.77 


4.'00 


3.04 


Hoffman 


L 


27.30 


7.14 


1.80 


35.98 


18.00 


6.80 


2.98 


Lawrence 


M 


22.77 


10. 


43 


34.54 


21.85 


3.63 


4.43 


Newark & Rosendale . 


N 


28.71 


5.88 


3.60 


27.00 


30.00 




3.52 


Clarks Utica .... 


O 


34.66 


5.10 


1.00 


30.24 


18.00 


6.16 


4.84 


Milwaukee 


P 


23.16 


6.33 


1.71 


36 . 08 


20 . 38 


5.27 


7.07 


Mankato 





27 70 


7.06 


1 86 


37.00 


22.63 




2 46 


James River .... 


^6 

R 


49.53 


11. 


29 


25.15 


13^77 




0^26 


Howard 


S 


22.58 


7.23 


3.35 


48.18 


15.00 




3.66 


Union Akron .... 


T 


17.14 


7.61 


2.00 


36.83 


25.09 


3.64 




Akron Star 


u 


20.40 


6.22 


2.56 


40.64 


25.80 




1.47 


Buffalo 


V 


16.48 


4.40 


2 00 


39.20 


26.52 


1.85 


6.80 





















The localities indicated by letters are as follows: 

A, Siegfried, Pa. ; B, Hancock, Md. ; C, Cumberland, Md. ; D, Fort Scott, 
Kan. ; E, F, and G, Louisville, Ky. ; H, Speeds, Ind. ; I, K, L, M, and N, Ulster 
Co., N. Y.; J, Howes Cove, N. Y.; O, Utica, 111.; P, Milwaukee, Wis.; Q, Man- 
kato, Minn.; R, Locker, Virginia; S, Cement, Ga.; T and U, Akron, N. Y.; 
V, Buffalo, N. Y. 

Comparing these analyses with those given for Portland 
cement in Table 2, it is seen that natural cements have a higher 
percentage of silica, about the same amount of alumina, and a 
much smaller content of lime, than have Portlands. Many natu- 
ral cements have a large percentage of magnesia, but the mag- 
nesia and lime together of natural cements usually do not equal 
the percentage of lime in Portlands. In other words, the 
hydraulic index is usually higher than in Portland cements. 



CHAPTER II 
MANUFACTURE 

ART. 6. THE MANUFACTURE OF PORTLAND CEMENT 

15. HISTORICAL. It is said that as early as 1810 a patent 
wa,s obtained in England for the manufacture of an artificial 
product by calcining a mixture of carbonate of lime and clay. 
This, however, was not called cement, and it was not until 1824 
that Joseph Aspdin, of Leeds, England, in obtaining a patent 
for the manufacture of a similar material, called his product 
" Portland Cement." This name was probably suggested by 
the fact that the color of the hardened product resembled that 
of a limestone quarried on the Island of Portland. The industry 
was introduced into Germany about thirty years later, and has 
since grown to very substantial proportions in both of these 
countries, as well as in France, Austria, and Russia. 

David O. Saylor was the first to manufacture Portland 
cement in the United States, at Coplay, Pa., about 1872, and 
works were established at that point in 1875. These were 
soon followed by other factories in Pennsylvania and Indiana, 
and at present cement is successfully manufactured in nearly 
half of the states of the Union, the production having steadily 
increased. 

16. MATERIALS REQUIRED. The materials requisite for the 
manufacture of Portland cement are carbonate of lime and 
silica. The former may be in the form of limestone, chalk or 
calcareous marl, while the latter may be in the form of clay or 
shale, or it may occur as an impurity in the limestone, as in 
the so-called " cement rock" of the Lehigh Valley region. The 
proportions required for ordinary materials are 25 per cent, 
clay or shale to 75 per cent, carbonate of lime. The cement 
rock worked in the Lehigh Valley region usually requires the 
addition of 10 to 20 per cent, of purer limestone, though in 
some deposits a small percentage of clay must be added. Some 
of the chief requisites for all Portland cement materials are 

12 



PORTLAND CEMENT 13 

that they shall be of uniform composition, and that they shall 
not contain such proportions of sulphur and magnesia as to 
give injurious percentages of these ingredients in the finished 
product. ( 50 and 52.) Formerly authorities did not agree 
as to whether the alumina in the clay or shale was an unwel- 
come constituent for Portland cement manufacture, but it is 
now generally agreed that the alumina, as well as the iron 
oxide, is useful as a flux in burning and that the dicalcic alumi- 
nate formed plays a role in the setting of the cement, and 
probably also in the subsequent hardening. If the ratio of 
silica to alumina and iron oxide falls below 2.5, however, the 
cement is likely to be too quick setting, while if this ratio 
exceeds 3 the mix is likely to be too refractory. 

In regard to physical characteristics neither material should 
contain flinty nodules, difficult to crush and of a different com- 
position from the main body of the rock. Marl should be as 
free as possible from hard shells, since their presence necessi- 
tates more careful preliminary grinding. Clay should not 
contain any considerable amount of coarse sand, for although 
silica is the most useful constituent of the clay it must not be 
in this insoluble form. In the highly argillaceous limestone 
known as cement rock, the mixing of the proper materials has 
been partially accomplished in nature, and with this material 
there is somewhat less danger from poorly proportioned mix- 
tures than in dealing with practically pure carbonate of lime 
and clay or shale. The soft materials, marl and clay, have an 
advantage because they are usually of fine grain and easily 
reduced and mixed, but since they are so hygroscopic the ex- 
pense of drying them may easily offset this advantage. 

A few analyses of materials suitable for Portland cement 
manufacture are given in Table 4. 

17. The materials for Portland cement manufacture, lime- 
stone, marl, clay, shale, etc., are widely disseminated, but the 
suitability of a certain locality for successful commercial manu- 
facture depends upon the manner of occurrence of these requi- 
sites. In England the clay is dug from the old beds of the 
Thames and Medway Rivers, and chalk, which occurs in abun- 
dance, furnishes the carbonate of lime in most cases, though 
limestone is sometimes used. In Germany both chalk and 
marl are used; the chalk being a soft white marl similar to the 



14 



CEMENT AND CONCRETE 



deposits in this country, and the marl a "more or less hard 
limestone rock containing clay." In the United States both 
limestones and marls are used. The most important cement 
producing region in the United States is in the Lehigh Valley, 
where an argillaceous limestone is employed. The factories 
using marl are situated in New York, Ohio, Indiana, Michigan, 
etc., where the marl is found overlying or in the vicinity of beds 
of clay or shale suitable for cement making. In the Lehigh 
Valley region many advantages are combined. The cement 
rock of that locality has nearly the correct composition for 
Portland cement manufacture, requiring usually a small admix- 
ture of pure limestone. The supply of this rock is almost 
inexhaustible, the managers of the works have had long 
experience in the production of cement from these materials, 
and a market for the product is near at hand. 

TABLE 4 
Analyses of Portland Cement Materials 



Material. 


Portland 
Cement 
Company. 


State. 


SiO,. 




< fe 


CaO. 


1 


i 


C0 2 . 


H 2 O 
and 
Loss 


Limestone . . 
Limestone . . 
Limestone . . 
Limestone . . 
Limestone . . 
White Marl . 
White Marl . 
White Marl . 
Chalk .... 
Cement Rock 


Atlas 
Hudson 
Alpena 
Northampton 
Diamond 
Newaygo 
Peninsular 
Sandusky 
Western 


Mo. 
N. Y. 

Mich. 
Penn. 
Ohio 
Mich. 
Mich. 
Ind. 
So. Dak. 
Penn. 


0.4 
1.0 
3.1 

5.6 
9.5 
0.8 
0.2 
1.8 
4.1 
Q 5 


0.4 
1.0 
1.2 
2.4 

2.5|2.7 
0.1 
0.8 
1.2 
4.5 
4.7 


54.9 
54.3 
52.1 
50.5 
45.7 
53.6 
51.6 
49.6 
51.0 
45.2 


0.2 
Tr 
.1 
.0 
.0 
.9 
.3 
.3 

? 3 


\A 

0.'9 
0.5 


43.3 
42.7 
42.1 
39.6 
37.0 
43.1 
40.5 
40.4 
40.0 
38.1 


5 7 
4 2 


Cement Rock . 
Cement Rock 


Northampton 


Penn. 
Penn 


15.7 

m ?. 


7.9 
7.2(0.9 


39.6 
35.5 


1.7 

?, ? 




33.2 
30.3 




Clay .... 
Clay . . . 


Sandusky 
Peninsular 


Ind. 
Mich. 


55.3 
61 4 


10.23.4 
25.1 


9.1 
1.4 


5.7 




7.1 
6. 


6 


Shale .... 
Shale .... 
Shale .... 
Shale .... 


Alpena 
Western 
Lehigh 
Hudson 


Mich. 
So. Dak. 
Ind. 
N. Y. 


57.4 
58.0 
59.6 
62.4 


23.3 

22.8 
19.117.6 
27.6 


4.3 
1.8 
0.3 
0.3 


3.2 
1.8 
2.3 
1.5 


0.4 
1.3 


8. 
12. 
4. 
8. 


2 
1 

7 




Information collected from following sources: 

Cement Materials and Industry, Bulletin 243 U. S. GeoL Survey, by E. C. 
Eckel; Annual Reports, State Geological Surveys; and direct from com- 
panies. 

Deposits of cement materials are of value only when the 
limestone or marl, and clay or shale, are found in large quanti- 



PORTLAND CEMENT 15 

ties and near together, when the physical character of the 
materials is such as to render them easy of comminution and 
mixture, when coal or other suitable fuel may be had at low 
prices, and when the market is not too far removed. 

The following estimate of the relative quantities of cement 
made in the United States in 1903 from the several classes of 
materials has been made by Mr. E. C. Eckel: 1 

Argillaceous limestone and pure limestone 56 per cent. 

Marl and clay 14 

Soft limestone and clay 3 

Hard limestone and clay 28 

18. GENERAL DESCRIPTION OF PROCESSES. The essen- 
tials of any method of Portland cement manufacture are that the 
materials shall be correctly proportioned, very finely comminuted 
and thoroughly mixed; that the mixture shall be carefully 
burned to just the proper degree of calcination, and the result- 
ing clinker ground to extreme fineness. How these essentials 
can be best accomplished depends upon the character of the 
raw materials and the cost of fuel and labor, so that the de- 
tails of the method vary with the materials used and with the 
local conditions. 

In order that the proportions may be accurately determined, 
it is usually necessary to dry one or both of the raw materials. 
The ingredients may be ground separately and afterward mixed, 
though with certain materials the grinding and mixing may be 
done at the same time. In this mixing, a large amount of 
water may be used, giving a thin slurry, as in the wet process; 
a moderate amount of water may be employed, as in the semi- 
'dry process, giving a mix of the consistency of puddled clay; or 
the dry process may be employed, in which after the materials 
are once dried no water is added before the burning of the 
clinker. The burning may be accomplished in any one of 
several styles of kilns, although the advantages of the rotary 
kiln are so great that it has nearly superseded all other styles 
in this country. The grinding is a simple mechanical problem, 
to secure the required degree of fineness with least cost. 

19. THE DRY PROCESS. This method of manufacture is 
especially adapted to such materials as cement rock, lime- 



' Cements, Limes and Plasters," Wiley and Sons, New York, 1905. 



16 CEMENT AND CONCRETE 

stone, and shale, that must be ground before they can be mixed, 
although it is occasionally employed for marl and clay. In 
order that the proper chemical combinations shall take place 
in the kiln it is essential that the material shall be finely ground 
as well as thoroughly mixed. The rock is obtained by the 
ordinary quarry methods and is usually conveyed to the plant 
in cars which dump into the hoppers of large rock crushers. 
From these the crushed rock is passed to driers, usually of the 
rotary type. The drying of the raw materials serves two pur- 
poses it prepares the rock for further reduction in mills in which 
wet material would tend to clog, and it facilitates the proper 
weighing of the two ingredients to form the mixture. 

After drying, the two raw materials may be combined, and 
the continuation of the process is on the mix so formed. This, 
however, is not the universal practice; some manufacturers com- 
bine the two ingredients even before crushing, others mix them 
before they pass to the drier, while others keep them separate 
until after the second or third stage of the preliminary grind- 
ing. The second stage of the grinding is frequently accom- 
plished in ball mills or in kominuters, although rolls and grind- 
ers of the impact type are also employed. The final mixing and 
grinding preliminary to burning is usually done in tube mills, 
this type being considered most efficient for fine materials. 

From the tube mills the mix passes directly to the rotary 
kilns for burning. As the clinker falls from the lower end of 
the rotary, a small stream of water is allowed to trickle over it. 
This partially cools the clinker, renders it somewhat easier of 
reduction, and probably makes the cement somewhat slower 
setting. After further cooling, either in coolers or in sheds, 
the clinker is ground, usually by two or more stages, ball mills 
or rolls being frequently used for the first reduction, and tube 
mills or Griffin mills for the final grinding. Before the grind- 
ing is completed a proportion of calcium sulphate is added to 
regulate the rate of setting of the cement. The ground cement 
is stored in large bins and packed as desired in cloth or paper 
sacks, or in barrels,* for shipment. 

20. THE WET PROCESS. Although water may be used to 
facilitate the mixing of materials that require previous grind- 
ing, the wet process is particularly adapted to such raw mate- 
rials as are easily acted upon by water. The present types of 



PORTLAND CEMENT 17 

grinding machinery are such that it is no longer necessary to 
add water to assist in the mixing and grinding, and the use 
of the wet process is now confined almost entirely to the treat- 
ment of the softer raw materials, marl and clay, which, as 
they occur in nature, contain a considerable percentage of 
water. This contained water constitutes, from an economic 
standpoint, one of the serious problems in treating this class of 
materials, because of the expense involved in evaporating it 
at some stage of the process; it is, however, not without its 
advantages. Aside from the fact that the mixing is more 
readily accomplished in the wet state, the grinding is compar- 
atively easy, the slurry may be pumped from point to point 
instead of being carried and elevated, and the correction of the 
mix, by adding the desired amount of over-clayed or over- 
limed slurry, is more readily accomplished than is the case with 
the dry process. 

The marl and clay are commonly loaded in cars by dredges 
or steam shovels and drawn to the works by locomotive, cable 
or horses. The stones and sticks are removed from the marl 
by separators if necessary, and as a first step the clay is usually 
dried and ground to facilitate the calculation of the mix. The 
marl is sometimes passed through a pug mill before mixing with 
the clay, though more frequently the weighed amount of dried 
clay is added during the first pugging. Here sufficient water 
is added so that the slurry may be pumped to tube mills in 
which the final grinding of the raw materials is usually accom- 
plished. Either just before or just after passing the tube mill 
the slurry is caught in vats where it is kept agitated either by 
compressed air or mechanical agitators. Here the mix is 
analyzed and corrected if need be by adding the proper amount 
of an over-clayed or over-limed slurry kept in separate tanks 
for this purpose. As the slurry is pumped to the rotaries it 
usually contains as much as fifty per cent, water, which must 
be evaporated in the upper end of the kiln by the gases from 
the burning which is taking place in the lower half. 

It may be remarked here that in the older form of the wet 
process, developed in England and formerly used here to some 
extent, the thin slurry was run into "backs" or shallow reser- 
voirs to settle, the clear liquid was run off, and the slurry, after 
further drying, was cut into bricks to be artificially dried and 



18 CEMENT AND CONCRETE 

then burned in dome kilns. The adaptation of the rotary kiln 
to the burning of wet materials has put the original form of 
the wet process out of use in this country. 

The fuel consumption jn the wet process is appreciably 
greater than in the case of a dry mixture. The cooling and 
grinding of the clinker do not necessarily differ from the corre- 
sponding operations in the dry process. 

21. THE SEMI-DRY PROCESS. During the first thirty 
years of Portland cement production in the United States this 
process of manufacture was prominent, but the successful 
development of the rotary kiln has rendered it almost obsolete, 
and the modern cement mill is designed to use one of the pro- 
cesses already outlined with such minor modifications as the 
designer may consider advantageous. 

With wet materials the semi-dry process consists in forming 
the mix into bricks after it has passed a pug mill, drying the 
bricks artificially either in drying ovens or in a portion of the 
burning kilns, and then burning the bricks in a stationary kiln. 
With dry materials, sufficient water is added after the prelimi- 
nary grinding to permit pugging the mixture and making the 
bricks. The method is not only expensive in labor account, 
but considerable loss results from imperfect burning, and the 
uniform results attainable by the methods already described 
are not usually realized with this process. 

22. DETAILS OF THE MANUFACTURE: Preparation and Mix- 
ing of the Raw Materials. The main points in the prepara- 
tion of the raw material for burning are: First, the proper 
amount of each ingredient must enter the mixture; second, the 
materials must be reduced to an extremely fine state of divi- 
sion; and third, the mechanical mixing must be as perfect as 
possible. Unless the ingredients are dried, the first require- 
ment is not easy to fulfill, especially with marl and clay, as the 
absorptive power of the materials renders it difficult to properly 
apportion them. The wet process, which thus offers the greatest 
difficulty in the first proportioning, furnishes a ready oppor- 
tunity for correction of the mix before burning. With the dry 
process the variation in the analysis of the materials is the 
most frequent cause of variations in the mix, and should be 
carefully watched. A slight error in proportions is apparently- 
less dangerous with argillaceous limestone, such as cement rock, 



PORTLAND CEMENT 19 

than with pure limestone and clay, and for this reason some of 
the works producing cement from this material are prone to 
become careless in proportioning and correcting the mix. 

Referring to 9 we have seen that the per cent, of lime 
must not exceed 2.78% silica + 1.1% alumina. In practice, 
however, the percentage of lime in the mix seldom exceeds 2.7 X 
percent, silica + per cent, alumina; and the percentage of lime 
in the finished cement is still less, because of the loss of some 
volatile products and the addition of a small amount of ash 
and kiln lining in the process of burning. 

To illustrate the method of calculating a cement mixture 
let us consider that we have a Lehigh cement-rock of composi- 
tion lime 37.5; silica 17.0; alumina 7.0, and purer limestone 
of composition lime 51.5; silica 4.5; alumina 0.5, and desire to 
determine the proper proportions in which to mix them. 

Using the above formula as fixing the proper proportion of 
lime we have, 

For 100 parts Cement Rock: 

PARTS 

Lime required by Silica 17.0 X 2.7 = 45.9 
Lime required by Alumina = 7.0 

Total Lime required for 100 parts rock 52. 9 
Lime in rock 37.5 



Lime to be added 15.4 

In one part of Limestone: 

PARTS 

Lime required by Silica . 045 X 1 . 7 = . 0765 
Lime required by Alumina = . 0050 

Total Lirne required for 1 part limestone . 0815 
Lime contained in 1 part limestone . 515 

Excess of Lime in 1 part limestone . 4335 

Then to 100 parts cement rock we must add ' == 35 . 5 parts limestone. 

. 4ooo 

More than three-fourths of the Portland cement manufac- 
tured in the United States is made from limestones, either pure 
or argillaceous. These must be ground before they can receive 
the required addition of clay, or of purer limestone, as the case 
may be, and they are usually dried to facilitate the grinding 
as well as to permit of determining the correct proportions of 
the ingredients for the mix. The first step in the process with 



20 



CEMENT AND CONCRETE 



hard materials is to pass them through ordinary rock crushers. 
These are commonly of the gyratory type and of large size. 
Rolls and "dry pans" are also sometimes employed for the 
softer limestones and shales. Rolls and ordinary rock crushers 
are too familiar to need description here. A "dry pan" con- 
sists of a cylinder 6 to 12 feet in diameter in which revolve two 
heavy rollers. The rollers, being on opposite ends of a hori- 
zontal shaft attached to a vertical axis, turn about the shaft 
as well as revolve about the axis, the crushing being done 
between the edges of the rollers and the floor of the pan. This 
style of mill is also called an edge runner. The rock works to 
the outside and is discharged through perforated plates. Edge 
runners are not suited to reducing coarse lumps of very hard 
material, but work well on clay and soft rock. A similar grinder 
used on wet materials is called a "wet pan." 

In the dry process the material passes from the rock crushers 
to the driers. The rotary drier, the style in most common use, 
resembles the rotary kiln; it is a cylinder about four or five 
feet in diameter and about forty feet long, lined with brick, 
and is set slightly inclined to the horizontal and revolves slowly 
on rollers. The drier may be heated by the gases from the 
burning kilns or by a separate furnace, the hot gases entering 
at one end and escaping into the stack. The material is intro- 
duced at the higher or stack end and works slowly to the lower 
or furnace end. One style of drier, the Ruggles-Coles, has an 




RUGGLES-COLES DRIER 



inner cylinder, and the materials pass through the annular space 
between the two, the products of combustion being forced first 
through the inner cylinder and then back through the raw 
material. The stack drier has also been used, in which the 



CEMENT KILNS 21 

material is fed at the top and is deflected by baffle plates as it 
falls through the hot gases. 

The logical time for mixing the ingredients would seem to be 
immediately after they pass the driers, and this is usually done, 
though in some cases the materials are not mixed until after 
further reduction, while in a few mills the mixing precedes the 
drying. The further grinding of the raw materials is usually 
accomplished in two stages, a typical combination being ball 
mills and tube mills, though Kominuters, Griffin, Kent, Wil- 
liams, and Huntington mills are also employed. The thorough 
grinding of the raw materials is of the greatest importance, not 
only because of its effect on the quality of the product, but 
because an extremely fine mix may be burned at a smaller 
expense for coal. As the same styles of grinding machinery 
are used in pulverizing the clinker they will be briefly described 
under that head. 

In the wet process, which is seldom used for materials other 
than marl and clay, only the clay is dried. The grinding of 
the clay may be accomplished in one of the forms of disinte- 
grators, such as the Williams or Stedman mill, or in a dry pan 
or edge runner, described elsewhere. After the addition of the 
clay to the marl the further grinding or mixing is done in ordi- 
nary pug mills, in wet pans, or in wet Griffin, ball and tube 
mills. These mills are described under the head of grinding 
( 28 and 29). After thorough grinding the material is run 
into tanks for correcting the proportions and here the thorough 
mixing is completed by revolving paddles working in the tanks, 
or by agitation caused by jets of compressed air allowed to 
escape into the slurry. 

23. STYLES OF KILNS. The various styles of kilns in use 
may be divided into four classes, namely: (1) Common dome 
kilns, (2) Continuous kilns, (3) Chamber and ring kilns, and (4) 
Rotary kilns. The Dome Kiln is the simplest type. The cham- 
ber is usually egg-shaped. Cement-brick and coke are piled in 
alternate layers, the use of the proper amount of the latter 
requiring much skill, as it is a matter of experience. As the 
draft in the kiln varies with the weather, this method of burn- 
ing is more or less at the mercy of the winds. When the burn- 
ing is complete, the kiln is allowed to cool before removing the 
clinker, and thus much heat is lost, and the lining of the kiln 



22 CEMENT AND CONCRETE 

is destroyed by alternate heating and cooling. The amount of 
underburned and overburned clinker is likely to be large. The 
output is small, and fuel expense high. This style of kiln was 
formerly used for Portland cement, but is now confined almost 
entirely to burning natural cement and is then used as a con- 
tinuous kiln, material being charged at the top as the clinker is 
withdrawn at the bottom. 

The Dietsch Kiln is one of the best examples of the second 
type, or continuous kiln. The slurry, in the form of bricks, is 
introduced at the base of the stack, into what may be called 
the heating chamber. Below this there is a right angle with a 
short horizontal section, over which the hot slurry is raked, 
to fall into the burning chamber. The clinker in the lower 
part of the latter is cooled by the air entering through the grates,, 
while the slurry in the upper chamber is heated by the gases 
from the burning zone. At intervals a portion of the clinker, 
partially cooled, is removed at the bottom; this causes a general 
settlement in the kiln and leaves a space at the top of the burn- 
ing chamber, into which the dried clinker from above is raked, 
and more fuel added. This kiln uses small coal for fuel and is 
more economical than the dome type. 

The distinguishing feature of the Schbfer kiln is the con- 
traction of the dome at the point where combustion takes 
place, concentrating the draft at this point. The air entering 
the shaft at the bottom cools the clinker already burned, while 
the gases from the clinker burning in the central section serve 
to dry the raw bricks above. Several kilns of this type are in 
successful operation in this country. 

24. Chamber Kilns are used largely in England with coke as 
fuel. The gases from the kiln are made to pass over the slurry 
spread on brick floors, the kiln proper being at one end of this 
chamber and the stack at the other. These kilns are inter- 
mittent, have a comparatively small output, and require con- 
siderable labor. 

The Hoffman Ring Kiln consists of a series of compartments 
built around a large central stack. The chambers communicate 
by means of flues in such a way that the smoke and hot gases 
from one may be passed through other chambers before reach- 
ing the chimney. The kiln may be either "up draft" or "down 
draft," according to the direction in which the heat is drawn 



ROTARY KILNS 23 

through the chamber. The compartments are charged from 
the sides, and when the moisture has been driven off from the 
material in the chamber first fired, the gases from this chamber 
are passed through the adjacent chambers, which have in the 
meantime been filled with raw materials. Although this kiln 
is economical of fuel if run continuously, much labor is re- 
quired to charge and empty it. This type is not used in the 
United States, though it has been employed to some extent 
in Germany. 

25. Rotary Kilns. Although rotary kilns for other purposes 
had been in use for some time, the first patent for a process of 
manufacture of cement by their use was issued in 1877 to Mr. 
T. R. Crampton. The method, apparently, did not pass beyond 
the stage of laboratory experiment until 1885, when Frederick 




THE BONNOT COMPANY ROTARY KILN 

Ransome, of England, patented a rotary kiln, which, how- 
ever, required many important modifications to make it a 
success. 

About 1888 Mr. J. G. Sanderson and Dr. Geo. Duryee made 
some successful experiments with the rotary kiln for wet mix- 
tures, and in the following year experiments were begun at the 
works of the Atlas Portland Cement Co. under Mr. P. Giron, 
which resulted in the construction of a practical kiln for burning 
dry mixtures. Prof. Spencer B. Newberry, at about the same 
time, perfected the rotary process for wet materials at Warners, 
N. Y., and Sandusky, Ohio. 

A rotary kiln as used for the burning of cement consists of 
a steel cylinder from five to seven feet in diameter and sixty 
feet or more in length. This cylinder rests on two sets of rol- 
lers with its axis inclined to the horizontal about one foot in 
twenty, and is revolved slowly by gearing. (See cut.) The 
mixture to be burned is introduced at the upper or stack end 



24 CEMENT AND CONCRETE 

of the cylinder, while a jet of powdered coal, or sometimes of 
fuel gas or oil, is injected at the lower end through a special 
burner. As the cylinder revolves, the material works slowly 
towards the lower end, the clinkering temperature being main- 
tained, roughly speaking, throughout the lower twenty feet of 
the kiln. This portion of the kiln must be lined with the most 
fire-resisting brick; as the clinker acts upon fire brick lining 
to form a fusible compound at the high temperatures required 
in the burning, a coating of fused cement clinker is sometimes 
beaten down upon the lining while it is still plastic; this adheres 
to the brick and protects them somewhat from further injury. 
The clinker as discharged from the lower end is in rough, green- 
ish black, balls varying in size from a quarter inch to an inch 
in diameter. 

The dried and pulverized coal is fed through a special burner 
supplied with a blast, the greater part of the air necessary for 
combustion entering through openings in and around the hood 
of the kiln. The burning may be controlled by the rate of 
feeding the mix, the rate of revolution of the kiln, or the amount 
of coal and air supply, the temperature being determined by 
the appearance of the molten mass. The crowding of a kiln 
to increase the output increases considerably the fuel cost per 
barrel on account of the great quantity of air introduced which 
must be heated. 

While until recently the ordinary length of a rotary kiln 
was 60 feet, it was realized that, especially for wet mixtures, a 
greater length was economical, and recently a much advertised 
plant has installed, among several other innovations, kilns 
150 feet long, and these have shown such results that just now 
a period of evolution to kilns of greater length seems to be 
fairly inaugurated. While the results at different mills vary 
greatly, and it is not easy to give a correct estimate of output 
and fuel consumption, since these vary with the character of 
the raw materials and fuel as well as with the details of the 
kiln, it may be said that with the dry process using cement 
rock and limestone a 60-foot kiln will give an output of 
about 200 barrels per day with a fuel consumption of 100 pounds 
of coal per barrel, while a kiln 150 feet long may give 300 barrels 
per day with only about 75 pounds of coal per barrel. With 
wet mixtures an output of 100 barrels per day may require as 



ROTARY KILNS 



25 



much as 200 pounds of coal per barrel with a 60-foot kiln, more 
than half of the coal being required to evaporate the water 
in the mix. The advantage of the longer kiln in this case 
would probably be still greater. Dr. Richards found from 
tests of a 60-foot kiln in a dry process plant that of the 110 
pounds of coal required to burn a barrel of cement the heat 
from about 80 pounds was carried in the waste gases, half of 
which was due to excess of air introduced, while the heat from 
12 pounds remained in the hot clinker issuing from the kiln, 
and the equivalent of 14 pounds of fuel was lost in radiation. 
The keen competition in cement making, and the occasional 
low prices prevailing, have led to efforts to utilize the heat 
contained in the waste gases. With a longer kiln the waste 
gases are expelled at a much lower temperature. 

26. ADVANTAGES OF THE ROTARY KILN. A comparison 
of the average output of the several styles of kilns described 
above, and the approximate fuel consumption, are given in 
the following table. Where it is necessary to dry the materials 
before introducing them into the burning kiln, the fuel required 
in drying is not included. 

FUEL REQUIRED FOR VARIOUS KILNS 



Style of Kiln. 


Barrels per Day. 


Pounds of Coal 
per Barrel of 
Cement. 


Int6rniitt6irt dome 


30 


80 to 120 


Hoffman. (per chamber) 


25 


60 to 80 


Dietsch and Schofer 


50 to 75 


60 to 80 


Chamber 


30 


160 to 200 


Rotary wet process, 60-foot kiln 


80 to 120 


140 to 200 


Rotary wet process, 110-foot kiln 


130 to 200 


100 to 140 


Rotary, dry process, 60-foot kiln 


175 to 250 


95 to 120 


Rotary dry process, 150-foot kiln 


300 to 500 


75 to 100 









Although, as seen from the above table, the burning of 
cement in a rotary kiln of the ordinary length requires a larger 
fuel consumption than with some other types, the ability to 
use a cheaper form of fuel and the saving in the amount of 
labor required, much more than offset this disadvantage. Since 
either wet or dry materials may be fed into the kiln, the neces- 
sity of forming the slurry into bricks, drying and stacking 
them in the kiln, is eliminated, and it is possible to so arrange 



26 , CEMENT AND CONCRETE 

a plant that the material is handled entirely by machinery 
from the raw state to the finished product. In fact, the con- 
veying and elevating devices in a cement plant form a most 
important part of its equipment, and the mechanical problems 
in cement manufacture are given, in one sense, more careful 
attention than the chemical questions. The control possible 
in burning with the rotary is much better than with fixed 
kilns, since the clinker may be watched while burning and the 
temperature may be gaged according to its appearance. The 
fact that the pieces of clinker are much smaller with the rotary 
process also contributes to uniform burning as well as to greater 
ease in grinding. The remarkable development of the Port- 
land cement industry in the United States is due in no small 
measure to the adoption and perfection of the rotary kiln, for 
the labor expense in manufacture is so reduced thereby that 
we are able to successfully compete with cements made abroad 
where lower wages prevail. 

27. GRINDING. In grinding it is not sufficient that the 
cement be so reduced that a certain percentage of it will pass 
a sieve having, say, 10,000 holes per square inch; but it is de- 
sired that as large a proportion as possible shall be of the finest 
floury nature. To accomplish this result it has been claimed 
that French buhr millstones are the best, but their great con- 
sumption of power has led to the introduction of other forms 
of grinding machinery, so that at present millstones find their 
chief use in natural cement manufacture. 

It is usually considered that the greatest economy results 
from a gradual reduction of the clinker as it passes from one 
form of grinder to another, each machine being supplied with 
the size of pieces it is best adapted to handle. Large pieces of 
clinker are first passed through an ordinary rock crusher, such 
as the Gates or Blake. Where rotary kilns are in use, this step 
in the process may be omitted, as the clinker comes from the 
kiln in small, nut-like pieces, and the grinding is completed in 
two stages. The first reduction may be accomplished in ball 
mills, Kominuters. Kent mills, Disintegrators or rolls, while 
the final grinding may be done in tube mills, Griffin mills, 
rolls or Huntington mills. 

28, BALL AND TUBE MILLS. The ball mill is perhaps in 
most common use for the first reduction. It consists of a short 






GRINDING 



27 



cylinder of large diameter partially filled with flint or steel 
balls. (See cut.) The grinding plates arranged about the 
inner circumference overlap one another and form a series of 
steps. The material is fed to the mill through one axis, and 
as the mill revolves the balls drop down the steps formed by 
the plates and roll over one another, pulverizing the clinker 
between them. The grinding plates are perforated, and as the 
clinker is pulverized it passes through the perforations and 
drops on a screen, and finally on a second finer screen. The 
particles too coarse to pass the screens are returned to the 





KRUPP BALL MILL 



grinding chamber through the openings between the steps 
formed by the grinding plates, while the material passing the 
screens is discharged to the conveyor. A ball mill working 
on ordinary rotary kiln clinker will grind from 15 to 24 barrels 
per hour to pass a 20-mesh screen and will require from 30 to 
50 H.P. 

The Kominuter is a modification of the ball mill, the prin- 
cipal point of difference being that the grinding plates are 
not perforated, but the material must travel the full length 
of the drum, escaping to the screens outside the plates through 
openings at the end opposite the inlet. It then travels back 



28 



CEMENT AND CONCRETE 



across the screens surrounding the plates, the material re- 
jected being returned to the grinding chamber when it reaches 
the inlet end. (See cut.) The machine is said to grind 40 
to 50 barrels per hour, giving a product passing 20-mesh 
screen. 

The Tube Mill is another style of pulverizer depending upon 
the action of balls for the grinding. It is usually employed for 
the final reduction either of clinker or raw material, and from 




THE LINDHARD KOMINUTER, F. L. SMIDTH & CO. 

its very general adoption for this duty must be considered a 
very satisfactory machine. It consists of a steel cylinder 15 
to 30 feet long and about 5 feet in diameter, lined with chilled 
iron or with stone, and mounted on trunnion bearings about 
which it is slowly revolved. (See cut.) 

The tube is at least half full of flint or steel balls, the former 
being a natural product imported from Europe, while the 
latter are made in this country. As the tube revolves the balls 
are carried up the rising side, but before reaching the highest 
point of revolution they drop to the bottom of the tube, im- 
pinging upon each other and crushing the product between 
them. The material is fed in at one end and gradually works 
toward the other; the fineness of grinding is regulated by the 
rate of feed, and no sieving is necessary. The capacity of a 
tube mill 5 feet by 22 feet is said to be from 8 to 20 barrels per 
hour when working on the product of a ball mill, and the power 



GRINDING 



29 



consumed is variously stated at 30 to 70 H.P. The renewing 
of the flint pebbles forms one of the heaviest repair items, and 




THE DAVIDSEN TUBE MILL, F. L. SMIDTH & CO. 

the cost of these has been estimated as high as one-half cent 
per barrel of cement. 

29. The Griffin Mill is an American invention that has 
found much favor for grinding clinker. A heavy steel crush- 
ing roller is rigidly attached to the lower end of a steel shaft 
which is suspended by ball and socket joint at its upper end. 
The shaft is thus free to swing in any direction within a case 
which incloses a ring or die surrounding the crushing roller. 
When the shaft is revolved rapidly by means of a horizontal 
pulley at the top, the roller presses by centrifugal force against 
the inner surface of the ring where the grinding takes place, and 
thus runs around the ring in a direction opposite to that 
in which the pulley is revolved. The material which drops 
below the roller is thrown up again by steel blades attached 
to the bottom of the shaft, and above the roller are fans which 
create a draft, forcing the finished product through the screens 
above the ring whence it falls into a hopper. Although the 
repair expenses of a Griffin mill are high, it has proved a very 
satisfactory machine and is extensively used, especially for the 
final reduction of clinker. The size of mill is designated by 
the diameter of the ring or die. A 36-inch mill, with roller 24 
inches in diameter, 6-inch face, and weighing 175 pounds, 
consumes 25 to 35 H.f*. and grinds 8 to 10 barrels of cement 
per hour when working on clinker not coarser than one-half 
inch. The manufacturers of the Griffin mill now make a 3-roll 
mill of somewhat similar construction. 



so 



CEMENT AND CONCRETE 




30-INCH GRIFFIN MILL ARRANGED FOR DRY PULVERIZING 



GRINDING 



31 



The Huntington Mill is very similar to the Griffin mill, the 
single roller of the latter being replaced by three rolls suspended 
from a revolving head. As the head is revolved, the rolls press 
against the steel ring surrounding them, and by reason pf this 
contact the rolls revolve upon their shafts. The output and 
power consumption are said to be about the same as in the 
Griffin mill. 

The Fuller-Lehigh Pulverizer Mill is a more recent style of 
grinder, said to be suitable for Portland cement clinker as well 
as raw material and coal. The machine consists of an up- 
right horizontal ring or die with a circular groove or track in 
which travel four balls. The balls rest upon a horizontal 
annular plate the revolution of which causes the balls to revolve 
and press against the circular track. The die, plate and balls 
are of Swedish iron. The balls are 9J inches in diameter and 
weigh 112 pounds each, and as they revolve at the rate of 
about 210 revolutions per minute the pressure against the 
surrounding track is estimated to be about 1,600 pounds. The 
material entering the center of the mill falls into a pan below 
the balls and is thrown 
up between the balls and 
track. The ground mate- 
rial is driven upward by 
fans and through a screen, 
the rejected portion return- 
ing for further grinding. 1 

The Kent Mill resem- 
bles the Griffin mill in that 
the crushing is done be- 
tween rolls and die, but the 
arrangement is quite dif- 
erent. There are three rolls, 
one of which is driven 
while the others revolve 
only because of their con- 
tact with the exterior ring which is revolved rapidly by the 
contact of the driven roll upon its inner face. (See cut.) The 
material is held upon the revolving ring by ce'ntrifugal force 



.ADJUS 
SCREW 




OH FLOOR 

INTERIOR OF KENT MILL 



For cut and detailed description see Engineering News, June 21, 1906. 



32 CEMENT AND CONCRETE 

and passes under the three rolls in succession. The rolls are 
held against the concave surface of the ring by springs which 
permit some yielding, to pass lumps. 

This mill was at first used only as a preliminary grinder or as a 
feeder for a tube mill, but it has been found that by screening 
the material, either with an inclined screen or with an air 
separator, it may be used economically as a fine grinder. When 
working on rotary kiln clinker this mill has given 40 to 50 
barrels per hour of 20-mesh material with a power consumption 
of 24 H.P.; in grinding the same clinker to pass a 100-mesh 
screen, with 88 to 90 per cent, passing 200-mesh, it has given 
10 to 12 barrels per hour. 

30. Rolls. Although rolls have not as yet been exten- 
sively employed for cement grinding, one mill has recently 
been equipped entirely with rolls for both coarse and fine grind- 
ing, and in other mills rolls are employed for a first reduction 
of the clinker. The rolls are 24 to 30 inches in diameter with 
16 to 18 inch face. The shells may be of forged steel, chrome 
steel, or chilled cast iron, with beveled edges to prevent break- 
ing the corners. The distance between the rolls varies with 
the fineness of the clinker and of the finished product; the 
feed is regulated by an automatic device. The rolls are some- 
times mounted in sets of three for double passage of the material, 
only one roll being actuated, the others being driven by fric- 
tion, a peripheric speed of about six feet per second being 
maintained. The three-roll fine grinders in the Edison Port- 
land Cement Mill at New Village, N. J., have rolls 28 inches 
diameter and 8-inch face. The rolls are held together with a 
pressure of about 9 tons per square inch, this pressure being 
obtained by an endless rope passing over loose s-heaves on the 
ends of the outer rolls at each side of the machine; a sheave in 
the bight of this endless rope is attached to a piston working 
in a cylinder to which compressed air may be admitted. This 
relieves the bearings of all pressure except the weight of the 
rolls. These machines have a large power consumption, but 
give a correspondingly large output. 

31. Disintegrators. This class, including all the machines 
which crush the material by impact, may be briefly described 
here, although seldom employed in the manufacture of Por.t- 
land cement. Under this head may be mentioned the Williams 



GRINDING 33 

Crusher, the Raymond Pulverizers, and the Stedman Disinte- 
grator. The Williams Mill consists of a series of hammers 
hinged to a horizontal rotating shaft, the whole being inclosed 
in a case. The grinding is accomplished by the rapid blows of 
the hammers upon the material fed into the case. 

In the Raymond Automatic Pulverizer large square blades 
are attached to discs keyed to a horizontal shaft that is re- 
volved rapidly. The blades are set at such an angle with the 
shaft that the material is thrown alternately to right and left, 
the pulverizing being the result of the impact of the particles 
against each other and against the walls of the inclosing 
chamber. 

The Stedman Disintegrator consists of a series of circular 
cages composed of steel bars, and fitting one within another. 
The cages are mounted on a horizontal shaft, and are arranged 
to revolve, two right-handed and two left-handed. The mate- 
rial being fed to the inside of the central cage is thrown out to 
the bars of the second cage moving in the opposite direction, 
and so on through the series. This mill is especially adapted 
to materials not so hard as Portland cement clinker. 

32. The grinding of the clinker, being an important and 
expensive part of the process of cement making, has been the 
subject of much experimentation on the part of the manufac- 
turers. That they have not all arrived at the same conclusion 
is partly due to variations in the character of the material. A 
high-limed hard-burned clinker is harder to grind than one of 
higher silica and alumina content which is burned at a some- 
what lower temperature. The fineness requirements in cement 
specifications have been constantly increasing, and manufac- 
turers are vying with each other to produce a finely ground 
product without incurring prohibitory expenses in the grinding. 
While the reduction from clinker to finished cement was form- 
erly sometimes accomplished in one machine, the one-stage 
reduction process, nearly all plants now pass the material 
through two mills, usually of different styles. Ball mills for 
first reduction and tube mills for final reduction probably now 
form the most common combination. The user of cement is 
interested chiefly in having a large percentage of the finest 
material present, and it has been found that in this regard 
there is little to choose between several of the best types. 



34 CEMENT AND CONCRETE 

33. SAND-CEMENT. This product, which is also called silica 
cement, is composed of Portland cement and silicious sand mixed 
in any desired proportion and then ground to extreme fineness. 
This product is placed on the market by dealers, but rights to 
use the process may be purchased. In the construction of Lock 
and Dam No. 2, Mississippi River, between Minneapolis and 
St. Paul, Major F. V. Abbot l used the process ; grinding with 
a tube mill one part of Portland cement with one part fine 
sand. The cost, exclusive of plant, is estimated as follows: 

\ barrel of Portland cement at $2.85 $1.42 

\ " " sand at .05 03 

Cost of grinding .50 

Cost of royalty .05 

Cost of one barrel Silica cement. $2.00 

This cement has given remarkably high tests considering the 
adulteration with sand, and is claimed to be specially useful in 
making impervious mortar and concrete. 



ART. 7. OTHER METHODS OF MANUFACTURE. OF PORTLAND 

CEMENT 

34. Portland Cement from Blast Furnace Slag. The prep- 
aration of a true Portland cement from blast furnace slag has 
been followed in Germany and elsewhere in Europe for several 
years, and recently has been introduced in the United States. 
As this process utilizes a waste product, its popularity is likely 
to increase. Whereas, for the manufacture of slag cement only 
the slag from gray pig iron is available, it is found that in most 
cases the slag from white pig iron may be used for the produc- 
tion of Portland cement from slag. 

The method of manufacture is briefly as follows: The slag 
as it comes from the blast furnace is subjected to the action of 
a stream of water, which granulates it and changes it chemi- 
cally, the water combining with the calcium sulphide, which is 
injurious to cement, to form lime and sulphuretted hydrogen. 



1 Report of Mr. A. O. Powell, Asst. Engineer, Report Chief of Engineers, 
U.S. A., 1900, p. 2779. 



SLAG CEMENT 35 

The granulated slag is then dried, mixed with the correct pro- 
portion of dried limestone, and ground to extreme fineness. 
The mixture is next burned in rotary kilns, the remainder of the 
process being the same as that employed when ordinary raw 
materials are used. While a cement made from slag by this 
method may have some peculiarities due to the nature of the 
raw materials used, and should be very carefully tested before 
it is used in important work, it should not be confounded with 
slag cement, which is a mixture of granulated slag and hydrated 
lime subsequently ground, but not burned together. 

35. Portland Cement from By-Products of Soda Manufacture. 
The Michigan Alkali Company has installed at Wyandotte, 
Mich., a cement plant to utilize the large amount of limestone 
which they have as waste in the manufacture of soda products. 
The limestone which has served its purpose in the soda manu- 
facture is in a finely divided and semi-fluid state; to this is 
added the proper percentage of clay, which has been dried and 
pulverized. The two are then very thoroughly mixed and 
ground by pug mills and tube mills, the mix corrected by proper 
combination of over-clayed and over-limed slurry, and finally 
burned in rotary kilns. 



ART. 8. THE MANUFACTURE OF SLAG CEMENT 

36. Slag Cement is made by adding calcium hydrate to a 
granulated basic slag resulting from the manufacture of gray 
pig iron. The slag must be carefully selected as to its chemical 
composition ; Professor Tetmajer having found by extended 
experiments that slags containing silica, alumina, and lime in 
the ratio 30 to 16 to 40 are best adapted to the purpose. As the 
molten slag runs from the blast furnace it is suddenly chilled 
by being run into water, or is partially disintegrated by being 
treated with a strong current of water, air, or steam. It is thus 
reduced to coarse particles resembling sand, or to a spongy or 
fibrous mass which, after drying, is readily ground to a fine 
powder. The process of chilling results in a certain chemico- 
physical change that renders the powder capable of combining 
more readily with the slaked lime which is subsequently added. 
Slag which has been allowed to cool slowly will not form an 
hydraulic product when mixed with the lime, although the 



36 CEMENT AND CONCRETE 

chemical composition of the slag may be identical in the two 
cases. The lime is dipped into water, or treated with steam, 
until slaked to a fine dry powder, and is then added to the 
powdered slag in proportions of about one part of the former 
to three parts of the latter, this proportion depending upon the 
composition of the slag used. The powdered slag and lime are 
sifted, then mixed and reground together to an extreme fine- 
ness, thus insuring an intimate incorporation of the ingredients. 
Since there is no burning in the process, it is evident that the 
finished product is merely a mixture, not a chemical compound 
as is the case with Portland cement. 

37. One of the largest mills for the manufacture of slag 
cement in the United States is conducted by the Illinois Steel 
Company, and the following description of the process is con- 
densed from a statement of Mr. Jasper Whiting, 1 manager of 
the cement department, and patentee of the process: Slag of 
the proper composition is chilled as it comes from the furnace 
by the action of a large stream of -cold water under high pres- 
sure. The slag is thereby broken up, about one-third of its 
sulphur is eliminated, and it is otherwise changed chemically. 
A sample of the slag thus granulated is mixed with a proportion 
of prepared lime, and ground in a small mill whereby actual 
slag cement is produced. If the tests upon this trial cement 
are satisfactory, the slag is dried and then ground, first in a 
Griffin mill and then in a tube mill, where it is mixed with the 
proper amount of prepared lime and the two materials ground 
and intimately mixed together. The resulting product is said 
to be so fine that but 4 per cent, is retained on a sieve having 
200 meshes per linear inch. The lime is burned from a very 
pure limestone and stored in bins, beneath which are two 
screens of different mesh, the coarser at the top. A quantity 
of lime being drawn on the upper screen is slaked by the addi- 
tion of water containing a small percentage of caustic soda. 
The lime passes through the two screens as it slakes and is 
then heated in a drier; the slaking being thus completed, the 
lime may be incorporated with the slag. The purpose of the 
caustic soda added in the above process is to render the cement 
quicker setting. 



" Report of Board of Engineers on Steel Portland Cement," Appendix I. 



NATURAL CEMENT 37 

ART 9. THE MANUFACTURE OF NATURAL CEMENT 

38. History. The American product called natural cement 
was first manufactured at Fayetteville, Onondaga County, 
N. Y., in 1818, and used in the construction of the Erie Canal. 
Other early dates of manufacture are given as 1823, near Rosen- 
dale, N. Y., and 1824 at Williamsville, Erie County, N. Y., the 
products being used in the construction of the Erie and the 
Delaware & Hudson Canals. Factories were soon started in 
other states, and at present nearly every State in the Union has 
one or more natural cement factories, the total annual produc- 
tion being now about nine million barrels. 

39. Materials Required. The composition of rock from 
which natural cement may be made varies within wide limits. 
Natural cement is essentially a combination of acid constituents, 
silica, alumina, and iron oxide, with the basic constituents, lime 
and magnesia, but the proportions in which the several mate- 
rials shall occur in the natural cement rock is not confined to 
the narrow limits we have seen were necessary in Portland 
cement manufacture. When argillaceous limestone is used 
the product is sometimes referred to as aluminous natural 
cement, its essential ingredient being a bisilicate of alumina 
and lime, while the product made from argillo-magnesian lime- 
stone is called magnesian cement, and is probably composed 
of a triple silicate of lime, magnesia, and alumina. The Mary- 
land cements are typical of the former or aluminous variety, 
containing only one to five per cent, of magnesia, while the 
Rosendale and Milwaukee are magnesian cements containing 
15 to 25 per cent, magnesia. (See Table 3, page 11.) 

Rocks suitable for the manufacture of natural cement are 
widely distributed throughout the United States. If this 
were not so the use of natural cement would be very much 
curtailed, for if it must be transported long distances it cannot 
successfully compete with Portland cement, since the cost of 
transportation makes a much greater proportionate increase 
in the price of the product than is the case with Portland. 

In the following table are given a number of analyses of 
raw cement rock quoted by Mr. Eckel in his report on Cement 
Materials and Industry of the United States. 1 



Bulletin No. 243, U. S. Geological Survey. 



38 



CEMENT AND CONCRETE 



ANALYSES OF NATURAL CEMENT ROCK 



Authority. 


; 
Locality. 


SiO 2 . 


A1 2 3 


Fe 2 0, 


CaO. 


MgO. 


CO 2 . 


Mineral Industry 
C. Richardson 
C. Richardson 
W. A. Noyers . 
W. A. Noyers . 
L. C. Beck . . 
C. Richardson 
F. W. Clarke . 
Mineral Industry 
C. Richardson 
G. Steiger . . 


Coplay, Pa. 
Hancock, Md. 
Fort Scott, Kan. 
Louisville, Ky. 
Louisville, Ky. 
Howes Cave, N. Y. 
Lawrenceville, N.Y. 
Utica, Ills. 
Milwaukee, Wis. 
Maukato, Minn. 
Akron, N. Y. 


18.34 
19.81 
21.80 
18.33 
9.80 
11. 
21.41 
12.22 
17.00 
12.14 
9.03 


7. 
7.35 
3.70 
4.98 
2.03 
50 
10. 
9.39 
4.25 
4.62 
2.25 


49 
2.41 
3.10 
1.67 
1.40 
1.50 
09 
3.90 
1.25 
1.84 
0.85 


37.60 
35.76 
35.00 
30.41 
29.40 
31.75 
25.80 
24.40 
24.64 
22.66 
26.84 


1.38 
2.18 
3.50 
8.04 
16.70 
14.91 
10.09 
10.43 
11.90 
16.84 
18.37 


31. 0( 

33 .'OC 
32.76 
41. 4t 
40.34 


32! 46 
39.07 
40.33 



The materials found at any locality may vary considerably 
as to chemical composition, especially among the several strata. 
In some' cases the different strata are utilized to make two or 
more brands, which differ somewhat in their characteristics as 
to time of setting, etc. It is common also to mix two or 
more layers together in the manufacture, with the idea that 
the ingredients lacking in one stratum will be supplied by the 
others. 

40. DESCRIPTION OF PROCESS. - As the ingredients have 
been incorporated by nature in such proportions as to yield a 
natural cement, much of that portion of the process of Portland 
cement manufacture preliminary to the burning is unnecessary. 
The rock occurs in strata and is quarried in open cut where 
near the surface, or is mined in case the stripping is heavy. 
In open cut a face of 20 feet or more is sometimes worked. As 
has already been stated, the strata vary in chemical composi- 
tion, and while two or more brands are sometimes made at the 
same mill, the rock from several strata are frequently mixed 
in the production of one brand, the idea being that if one layer 
has an excess of acid constituents it may be corrected by 
another containing too much lime or magnesia. As the rock 
is not finely pulverized before it enters the kiln, each lump 
burns by itself and makes a certain cement; the piece of cement 
next it must make as distinct a product as though burned in a 
separate kiln. What is obtained, then, by this method is a 
mixture of several cements, and it would seem that the mere 



NATURAL CEMENT 39 

mechanical mixing of an over-limed and an over-clayed cement 
can not produce a well-balanced product, but that whatever 
advantage may be gained by this process is due to the pozzo- 
lanic reaction which plays a part in the hydration of natural 
cements. Moreover, the fact that wide variations in com- 
position must call for similar variations in the temperature of 
burning precludes the probability that the best results are 
obtained in this way. This practice undoubtedly accounts, 
in a great degree, for the large variations that occur in the 
cement from a single mill, variations which are often, however, 
more noticeable in short-time tests than in longer ones. 

41. Burning. As the rock is quarried it is broken into 
pieces varying in size up to six inches. To secure uniformity 
in burning, the pieces should be approximately of the same 
size, and to this end the rock should be passed through the 
ordinary rock crusher. It is then conveyed to the kilns, usually 
by tramway. 

The style of kiln almost universally used in burning natural 
cement is the cylindrical dome kiln, briefly described under 
Portland cement manufacture ( 23). These kilns are from 
20 to 50 feet high and 8 to 16 feet in diameter, built of masonry 
or steel lined with brick, and are operated as continuous kilns. 
The coal used for fuel is about pea size, either bituminous or 
anthracite being used, according to the locality. The rock and 
fuel are spread on the top of the kiln in alternate layers, the pro- 
portion of fuel being usually regulated by the man in charge, 
though a machine is sometimes employed to govern automat- 
ically the amount of coal used. 

The temperature reached in the burning varies quite widely 
according to the character of the rock. Materials high in 
alumina, iron oxide, and silica, that is, with a low lime-acid 
ratio ( 9), are most readily broken up. The carbon dioxide 
is first driven off, then takes place the combination of the lime 
and magnesia with alumina and iron oxide, and finally with 
the silica. Such rocks may be burned at a temperature of 
about 1000 to 1200 C., while those having a high lime-acid 
ratio may require 1200 to 1400. 

There is little question that the cement rock used in most 
natural cement plants is capable of producing a much better 
and more uniform quality of product than is actually obtained. 



40 CEMENT AND CONCRETE 

Many plants depend largely upon a local market, and there 
has not been the incentive to improvement that comes from 
sharp competition in a broader field, and consequently the rule 
of thumb methods, handed down from the early operation of 
the plant, are still employed. To select the rock with great 
care and burn each stratum at the temperature required to 
develop the best results, as determined by experiments on a 
practical scale, would undoubtedly improve the product. As 
this would call for the installation of rotary kilns and a higher 
cost of production throughout, it is usually considered that it 
would not be profitable, but this has not been established. 

42. Grinding. The calcined rock is conveyed first to some 
sort of stone crusher; a common form is known as a "pot- 
cracker," and consists of a corrugated conical shell in which 
works a cast iron core, also corrugated. The elements of the 
shell and core have a smaller angle to the vertical in the lower 
part of the mill, providing for gradual reduction and greater 
power on the smaller particles. The Berthelet and McEntre 
crackers are mills of this general type. A small cracker, about 
15 inches diameter and 18 inches high, working as a coarse 
grinder only, will pass 250 to 300 barrels per day. 

The ordinary buhr millstones have been used from the 
inception of the natural cement industry and are still very 
generally employed. The " rock-emery" mill, a stone in which a 
layer of harder composition is set about the exterior face, has 
replaced the older buhr mills in some plants, as it requires less 
frequent dressing. While ball mills, Williams crushers, Sted- 
man disintegrators, and dry pans have been introduced to 
some extent in natural cement plants, tube mills have, perhaps, 
proved the most efficient for this work of any of the modern 
types of reducers. These are described in Article 6 under 
Portland cement manufacture. When, as is now common 
practice, the reduction is accomplished in several stages, the 
fine material or finished product is separated after each stage, 
the power consumption for a given output being reduced 
thereby, although the proportion of extremely fine dust is 
probably also diminished. Mr. Berthelet of the Milwaukee 
Cement Company has perfected a screening system for this 
purpose that has proved very successful, and has been installed 
in other cement plants. 



NATURAL CEMENT 41 

After passing the grinding mills the product is sometimes 
conveyed to " mixers" by means of which the cement is thor- 
oughly mixed to promote uniformity. It is now ready for 
packing, and may be placed in bins or sent direct to the chute 
from which the bags or barrels are filled. Most natural cement 
is packed in paper or cloth sacks, as barrels add too large a 
proportionate cost to the product. 

It is seen that the manufacture of natural cement is very 
similar to that portion of Portland cement manufacture suc- 
ceeding the preparation of the raw material for burning. In 
general, less care is requisite with natural cement, the burning 
is carried on at a lower temperature, and the calcined rock is 
softer, so that less expense is incurred in grinding. 



PART II 

THE PROPERTIES OF CEMENT AND 
METHODS OF TESTING 



CHAPTER III 

INTRODUCTORY 

43. In the tests of such structural materials as wood and 
steel it will not usually be difficult to determine the suitability 
of the material for the intended purpose, provided the test 
pieces truthfully represent the members to be used. It is known 
that so long as these members are protected from oxidation and 
over-loading they will retain their qualities, and there is always 
a reasonably clear understanding of what these qualities should 
be. On the other hand, in the testing of cement, one may be 
perfectly sure that from the moment the cement is manufac- 
tured until long after it has been in service in the structure its 
properties will be ever changing; and, further, the qualities 
which it is desirable the cement should possess are not always 
clearly in mind. 

44. Desirable Qualities in Cement. The desirable elements 
in a cement may be stated as follows: 1st, That when treated in 
the proposed manner it shall develop a certain strength at the 
end of a given period. 2d, That it shall contain no compounds 
within itself which may, at any future time, cause it to change 
its form or volume, or lose any of its previously acquired strength. 
3d, That it shall be able to withstand the action of any exterior 
agency to which it may be subjected that would tend to decrease 
its strength or change its form or volume. When it is deter- 
mined that a cement has these three qualities, it is certain that 
it is safe to use it, but it is further desirable to know that the 

42 



UNIFORM METHODS 43 

cement in question will accomplish the given object as cheaply 
as any other cement. 

The cohesive and adhesive strengths of cement are not usu- 
ally considered in the design of the structure into which cement 
enters. The design of a masonry arch does not comprehend any 
adhesive strength in the cement, except as it may be recognized 
as an additional factor of safety, and a masonry dam is so de- 
signed that there shall be no tension at the heel. These facts 
are due in a large measure to the very imperfect knowledge we 
have of the behavior of cements in various contingencies. With 
the increasing use of concrete, as in arches, locks, floors, roofs, 
etc., the tensile and transverse strengths of cement are coming 
to be relied on to a certain extent; and as its properties become 
better known, and as means of recognizing these properties 
become more certain and widespread in their application, ce- 
ment will be more extensively employed in a scientific and eco- 
nomical manner. 

Cement may be compared in one sense to timber and cast 
iron. A large factor of safety is employed when dealing with 
these materials because of hidden defects that may exist. The 
defects which lie hidden in cement may be even greater than 
these in proportion to its possible strength, and defects in ce- 
ment are often more treacherous because their development 
may be deferred for some time. The importance of knowing 
whether the cement fulfills the second and third requirements 
noted above is therefore evident. 

45. Having considered the qualities a cement should have, 
we may proceed to the detailed consideration of the various 
tests employed to disclose the presence or absence of these qual- 
ities. The strength a given cement will develop is investigated 
by chemical analysis, by obtaining the specific gravity and fine- 
ness, and by actual rupture tests, whether they be tensile, com- 
pressive, transverse, or shearing. By tests for change of volume 
and by chemical analysis, it is sought to determine whether a 
cement has within itself elements of destruction. For the power 
to withstand external agencies there are no adequate tests, 
though chemical analysis is considered an aid. The methods 
of use, the proportions of the materials, their incorporation and 
deposition are of great importance in insuring against external 
causes of injury. 



44 CEMENT AND CONCRETE 

46. Uniform Methods of Cement Testing. In order that 
uniformity should prevail in the methods employed in testing 
cements, various societies have discussed the subject in detail, 
usually through committees, and much valuable work has been 
done along this line. The engineers of public works in many 
European countries have adopted specifications and laid down 
more or less detailed rules for testing. The Corps of Engineers, 
U. S. A., has recently adopted a similar code of rules. 

The International Society for Testing Materials, with which 
the American Society for Testing Materials is affiliated, has con- 
sidered the subject and still has committees at work upon it. 
The New York section of the Society of Chemical Industry has 
recently formulated a method for analysis of materials for the 
Portland cement industry. The American Society of Civil En- 
gineers received a report in 1885 from a committee appointed 
to consider methods of cement testing, and in order to keep the 
subject abreast of the latest developments in the manufacture 
and use of cement, a second committee was appointed several 
years ago, which has been making a thorough discussion of the 
subject, and has submitted a preliminary or progress report. 

47. Notwithstanding that so much has been done toward 
unification of methods, it may never be possible to determine 
accurately the value of one cement as compared with another 
tested in a different laboratory; though in tests of iron and 
steel no such difficulty is experienced. Certainly, as at present 
carried out, strength tests of cement are purely relative tests 
and do not show the absolute strength which may be developed 
in the structures; nor can the results be compared with the re- 
sults obtained in other laboratories and any fine distinctions of 
quality drawn. To attempt to carry out acceptance tests in 
such a way as to show directly the strength which will be de- 
veloped in actual construction, is only to introduce causes of 
irregularity in the tests. 



CHAPTER IV 

CHEMICAL TESTS 
ART. 10. COMPOSITION AND CHEMICAL ANALYSIS 

48. Value of Chemical Tests. The definite aid which chem- 
ical analysis may render in determining the quality of a cement 
is limited by the following considerations. It is not definitely 
known just what part is played by each of the compounds that 
go to make up commercial cement, and chemical analysis does 
not tell the manner of the occurrence of these compounds. A 
cement may have a chemical composition that is thought to be 
perfect, but if the burning has not been properly accomplished, 
it may be a dangerous product and analysis would show no de- 
fect. Some of the best authorities say that chemical analysis 
is useful principally in tracing the cause of defects which, by 
other tests, have been found to exist. However, there are some 
constituents which it is fairly well known a cement should not 
contain in any considerable quantities. An analysis may be of 
value in estimating quantitatively such constituents, while it 
may also be of service in detecting adulterations. It is not im- 
possible, then, that chemical tests may yet play a more impor- 
tant role in cement testing, especially if the method of analysis 
can be made more simple and rapid, without too great a sacri- 
fice of accuracy. 

49. Lime. The proportion of lime in Portland cement may 
vary from 59 to 67 per cent. A much greater range than this 
is allowable in natural cement, the percentage usually being 
from 30 to 45, according to the amount and character of the 
other active constituents. An analysis of Portland cement which 
shows a percentage of lime far outside of the limits mentioned 
above, should be regarded with suspicion and submitted to very 
thorough tests before acceptance. As already stated, the ratio 
of the silica and alumina to the lime in a cement is called the 
hydraulic index. The value of this ratio is usually between .42 
and .48 for Portland cement. 

45 



46 CEMENT AND CONCRETE 

Cement mixtures containing a large percentage of lime re- 
quire a high temperature for calcination, are difficult to grind, 
and yield a slow-setting product. The danger in highly limed 
cements is that they will not be properly calcined and a por- 
tion of the lime will be left in a free state. The demand for 
high strength in short-time tests has led manufacturers to 
make a heavily limed product, and in some cases the limits of 
safety have probably been overstepped. The introduction of the 
rotary kiln, however, has so improved the facilities for burning 
cement that a higher percentage of lime is now possible. 

There is no method known at present for determining quanti- 
tatively the amount of free lime in a cement, and it seems doubt- 
ful whether its presence can be detected with certainty by chemi- 
cal analysis. The method usually employed for this purpose 
depends on the hydration of the lime and subsequent absorption 
of carbonic acid. 

50. Magnesia. The detection of magnesia in several con- 
crete structures that had failed, led to the conclusion that mag- 
nesia, in quantities exceeding two or three per cent., was a 
dangerous element in Portland cement. In 1886-87 Mr. Har- 
rison Hayter 1 mentioned several failures of masonry and con- 
crete which he considered were due to magnesia, and concluded 
that cement should not contain more than one per cent. Later 
investigations, however, indicated that such failures could be 
explained in other ways, and that the magnesia found in the 
failing structure had come from the sea water and replaced the 
lime in the cement. Mr. A. E. Carey 2 has considered that "an 
excess of caustic lime or magnesia causes first, disintegration 
by expansion due to hydration, and second, being soluble, when 
conditions permit of their washing out, leave the concrete in a 
honeycombed state." Notice that this refers to caustic mag- 
nesia, and Prof. S. B. Newberry 8 has stated that "it is doubtful 
if magnesia is ever combined in Portland cement. Our own 
experiments tend to confirm the opinion of many German 
authorities that magnesia remains free in cement and does 
not combine with the constituents of clay after the manner 
of lime." 



1 Proc. Inst. C. E., Part 1, Session of 1886-87. 

2 Ibid., 1891-92. 

8 Municipal Engineering. October, 1896. 



COMPOSITION AND ANALYSIS 47 

On the other hand, M. H. LeChatelier * says that the " acci- 
dents occasioned by certain magnesian elements, and the similar 
results obtained in laboratory experiments, have been due to 
the employment of badly proportioned cements, containing free 
uncombined magnesia and too small a quantity of clay. Cor- 
responding mixtures containing lime instead of magnesia would 
have caused still more serious accidents, yet it would not be con- 
cluded that there must be no lime in cement." Again, Dr. 
Erdmenger characterizes magnesia as an adulterant only, and 
considers that its effect is nil if a greater percentage of lime is 
.added in the manufacture. 

Some authoritative information on the amount of magnesia 
allowable in Portland cement is contained in the report of the 
magnesia commission of the Association of German Cement 
Makers, 1895: Three members of this committee, Messrs. Schott, 
Meyer and Arendt concluded that "the presence of magnesia 
up to ten per cent, causes no harmful expansion or cracking of 
the cement, even after several years." Mr. Dyckerhoff, how- 
ever, presented a minority report, in which he pointed out that 
while a large amount of magnesia, not sintered, may not have 
an injurious effect, yet a content of more than four per cent, of 
sintered magnesia, whether added or substituted for part of the 
lime, has an injurious effect after long periods. The committee 
continued the ruling of 1893 that "a magnesia content of five 
per cent, in burnt cement is harmless," but held the question 
open for further investigation, indicating that this limit might 
be raised. 

In view of the disagreement among such eminent authori- 
ties it is impossible to arrive at a satisfactory conclusion, but if 
the effect of magnesia depends upon the manner of its occur- 
rence, whether free or combined, sintered or unsintered, then 
chemical analysis can be of but limited value as a test of quality 
in this regard. Natural cements frequently contain large pro- 
portions of magnesia replacing lime, and in this case an analysis 
is of the same value as an analysis for lime. 

51. Alumina and Iron .Oxide. The amount of alumina 
which a cement should contain is not well established. Its 
presence tends to facilitate the burning, and it renders the prod- 



1 Trans. Amer. Inst. Mining Engrs., 1893. 



48 CEMENT AND CONCRETE 

uct quicker setting. Cements containing large percentages of 
alumina are inferior for use in air or sea water, and it is probable 
that the percentage of alumina should not exceed eight or ten 
to obtain the best results in all media. A slag cement may be 
detected by its large content of alumina. Oxide of iron acts 
as a flux in burning, but in the finished product is little more 
than an adulterant. 

52. Sulphuric Acid. French specifications say that Port- 
land cements shall not contain more than one per cent, of sul- 
phuric *acid or sulphides in determinable proportions. This is 
doubtless intended for cement to be used in sea water. Adul- 
terations with blast-furnace slag may sometimes be detected 
by the amount of sulphides present, but small quantities of sul- 
phuric acid in the cement may be derived from the coke used 
in burning and have no injurious effect for use in fresh water. 
A content of 1.75 per cent, of sulphuric anhydride, SO 8 , is now con- 
sidered the maximum permissible. Sulphates mixed with the 
raw materials and burned with the cement may be harmless, 
while the same amount added after burning would not be per- 
missible. [For tests on the effect of adding sulphate of lime to 
cement, see Art. 48.] 

53. Water and Carbonic Acid. The determination of these 
may give some idea of the deterioration of a product by storage, 
and they may also indicate defective burning. M. Candlot con- 
siders that in the case of Portland cement, a loss on ignition 
(water and carbon dioxide) exceeding three per cent, "indicates 
that the cement has undergone sufficient alteration to appre- 
ciably diminish its strength." Natural cements may, however, 
contain considerable proportions of these ingredients and still 
give good results. 

54. Conclusions. Finally, then, the determination of silica, 
alumina, magnesia and lime may be of value, first, in classify- 
ing a product, and second, as indicating whether the proportions 
contained in it are such that if properly manufactured it is 
capable of giving good results. What these proportions should 
be for Portland cement has already been stated, 9. The de- 
termination of certain injurious ingredients is also of some 
value, but it must be remembered that the dangerous elements 
most commonly occurring, namely, free lime and magnesia, are 
not determinable by chemical analysis. It has been stated by 



COMPOSITION AND ANALYSIS 49 

M. LeChatelier that " neither complete nor partial chemical 
analysis of the constituents of hydraulic materials can be ranked 
among normal tests. But chemical analysis may render real 
service in controlling the classification of a product concerning 
which there is reason to doubt the declaration of the manufac- 
turer. Thus, a slag cement can be distinguished from a Port- 
land by its tenor in alumina and water; certain natural cements, 
by their contents of sulphuric acid, etc." 1 

The methods of analysis for Portland cement are given in 
considerable detail in a little book, " The Chemical and Physical 
Examination of Portland Cement," by Richard K. Meade. The 
method of analysis suggested by the New York Section of the 
Society of Chemical Industry is published in the Engineering 
Record of July 11, 1903, and in Engineering News of July 16, 
1903. 



'Tests of Hydraulic Materials," H. LeChatelier. 



CHAPTER V 

THE SIMPLER PHYSICAL TESTS 
ABT. 11. MICROSCOPICAL TESTS. COLOR 

55. Microscopical examinations are of some interest and 
value to those who are thoroughly versed in the chemistry of 
the burning and hardening of cements, as an aid in determining 
the part played by each compound in the hardening. 

Examinations may be made either of the dry powder, or of 
thin sections of hardened cement, or clinker. Dry powder of 
Portland cement appears to be made up of scaly particles, many 
of which are clearly denned and semi-transparent, while natural 
cement particles are more nearly opaque and less angular. Thin 
sections of Portland cement clinker have been found to exhibit 
colorless crystals somewhat cubical in structure, which are 
thought to form the essential hardening constituent; thin sec- 
tions of hardened Portland cement show a clear crystalline 
structure. Prof. Hayter Lewis found that the particles in good 
Portland cement were angular in form, consisting of scales and 
splinters, while the particles of cement of poor quality were 
rounded or nodular. 

Microscopic examinations have no place at present in ordi- 
nary tests of quality. 

56. Significance of Color. The color of cement is chiefly 
derived from its impurities, such as oxides of iron and manga- 
nese, rather than from its essential ingredients, and the color is 
therefore of minor importance. Other things being equal, a 
hard burned Portland cement will be darker in color than an 
underburned product. An excess of lime may be indicated by 
a bluish cast, and excess of clay or underburning may give a 
brownish shade. Gray or greenish gray is usually considered 
to be indicative of a good Portland. 

57. The colors of natural cements have a wide range, vary- 
ing from a light yellow to a very dark brown, without reference 
to quality. Owing to a popular idea that dark color indicated 

50 



WEIGHT PER CUBIC FOOT 51 

strength, some manufacturers have been said to add coloring 
matter to their product, but although this may have been true 
at one time, the correction of this false idea has doubtless ren- 
dered such a practice quite unnecessary now. Variations in 
shade in different samples of .the same brand of natural cement 
may indicate differences in burning or in the composition of the 
rock; but the interpretation of color for any given brand must 
be the result of close study, for some cements become lighter 
on burning and others become darker, while in some cases no 
variation in shade can be detected for different degrees of 
burning. 

ART. 12. WEIGHT PER CUBIC FOOT OR APPARENT DENSITY 

58. Significance. Since a hard burned Portland cement 
will usually be heavier than a light burned one, a test of the 
weight per cubic foot was once thought to be of great value in 
judging of the degree of burning. But it has been shown re- 
peatedly that the weight per cubic foot depends quite as much 
on the fineness as on the burning. It also depends on the age 
of the cement, and its chemical composition. As a test for 
quality, the determination of the apparent density has therefore 
been discarded. However, it is an aid in classifying a product, 
since Portland cements weigh from 70 to 90 pounds per cubic 
foot when loosely filled in a measure, while natural cements 
weigh from 45 to 65 pounds. A knowledge of the weight per 
cubic foot is also useful in reducing proportions given by weight 
to equivalent volumetric proportions, and vice versa. 

59. Method. This test may be made with a very simple 
apparatus, and the results obtained, though not strictly accu- 
rate, are sufficient for all practical purposes. A metal tube, 

2 feet 4 inches long, about 6 inches in diameter at the top, and 

3 or 4 inches at the bottom, is supported by a frame resting on 
four legs. A metal cylinder, 6 inches in diameter and 6^V 
inches deep, holding one-tenth cubic foot, is placed on the floor 
below the tube. A coarse sieve, through which all of the ce- 
ment will pass, is placed on top of the tube and three feet above 
the bottom of the measure. The cement passes through the 
sieve, falling freely to the cylinder below, which is struck off 
level when full. The cement must not be heaped too much, 
and great care must be taken that the measure is not jarred 



52 CEMENT AND CONCRETE 

while it is being filled or struck off. The cement is in such a 
light condition that a very slight jar is sufficient to cause it to 
settle. 

The above apparatus is on the same plan as that used by 
Mr. E. C. Clarke on the Boston Main Drainage Works, and is 
described here for general use when it is desired to compare 
the results obtained by operators at different points. Should 
one wish simply to obtain a series of results on different cements 
which are to be compared among themselves, it is quite suf- 
ficient to sift each sample through a coarse sieve, and then with 
an ordinary scoop carefully fill a measure of any known capac- 
ity, without other apparatus. 

Mr. Henry Faija has described an apparatus consisting of a 
funnel* with a screw at the mouth which carries the cement 
horizontally to the point where it falls freely into the measure. 
Various other devices have been employed, but none seems to 
have met with universal favor. 

60. To determine the relative accuracy obtainable with the 
simple form of apparatus first described, the author made a 
series of tests which may be summarized as follows: 

1st Method. Cement passed a wire mesh sieve, holes .033 
inch square and fell freely two feet through a 6-inch tube into 
a measure holding J cu. ft. Five trials with a sample of Dycker- 
hoff Portland, highest weight per cubic foot, 81 Ibs. 4 oz., 
lowest, 79 Ibs. 2 oz., difference, 2 Ibs. 2 oz. Three trials with 
Alsen's Portland, highest weight, 73 Ibs., lowest, 72 Ibs., dif- 
ference, 1 Ib. 

2d Method. Measure same size filled with scoop without 
other apparatus, and cement not shaken or jarred in measure. 
Five trials with Alsen's Portland, highest result, 73 Ibs. 8 oz. 
per cu. ft., lowest result, 72 Ibs. 12 oz., difference, 12 oz. Five 
trials with different sample of same cement, highest, 72 Ibs. 
4 oz., lowest, 72 Ibs., difference, 4 oz. 

3d Method. Measure filled with scoop, and cement well 
shaken down as filling proceeded. Five trials with Alsen's 
Portland, highest result, 100 Ibs. 8 oz., lowest, 97 Ibs. 14 oz., 
difference, 2 Ibs. 10 oz. 

It appears from these tests that when the measure is filled 
with the scoop, the results are about as uniform as when the 
apparatus is used, provided the filling is always done by -the 



SPECIFIC GRAVITY 53 

same person. But the results obtained by different operators 
with the same sample of cement would probably vary less, one 
from the other, when the apparatus is employed. In other 
words, the personal factor is more nearly eliminated when the 
cement is passed through a sieve and allowed to fall freely 
from a given height. 

61. As to the effect of age on the weight per cubic foot, it 
was found in one case that cement which weighed 93J pounds 
per cubic foot when freshly ground, weighed but 88 pounds 
when a few days old, and 78 and 74 pounds after six months 
and one year, respectively. 1 

Many experiments have been made to show the effect of 
fineness on the weight per cubic foot, but as this subject will 
be taken up again under " fineness," it will suffice to quote one 
series of tests made by Mr. E. C. Clarke, 2 giving the " weight 
per cubic foot of the same sample of German Portland cement 
containing different percentages of coarse particles as deter- 
mined by sifting through the No. 120 sieve." 

Samples containing 0, 10, 20, 30, and 40 per cent, of coarse 
particles retained on No. 120 sieve gave the following weights 
per cubic foot: 75, 79, 82, 86 and 00 pounds, respectively. 

It may be repeated that the weight per cubic foot is no 
longer considered an indication of quality, but should it be 
desired to specify a given weight, the method by which the 
test is to be made should also be stated. 

ART. 13. SPECIFIC GRAVITY OR TRUE DENSITY 

62. The apparent density or weight per cubic foot is in- 
fluenced to such an extent by the degree of fineness of the 
cement that this test has been almost superseded by the test 
for specific gravity. Although the true density, or specific 
gravity, is not affected by the fineness, it is influenced by the 
composition, the degree of burning, and the age, or amount of 
aeration of the sample. 

The method commonly employed in this test consists in de- 
termining the absolute volume of a given weight of the cement 



1 "Cement for Users," by H. Faija, p. 54. 

2 " Record of Tests of Cements for Boston Main Drainage Works," Trans. 
A. S. C. E., Vol. xiv, p. 144. 



54 CEMENT AND CONCRETE 

powder by measuring the amount of liquid which it will dis- 
place. A simple form of apparatus may be constructed in 
any laboratory as follows: In a wide mouth bottle, having 
straight sides and holding 200 c.c. or more, fit a perforated 
cork. Through the cork slip a burette graduated in cubic 
centimeters from to 50, placing the zero end down. Fill the 
bottle and the tube up to the zero mark, with some liquid such 
as turpentine, benzine or kerosene oil, but preferably benzine 
(62 Baume naptha). By means of a funnel in the top of the 
burette, add slowly 100 grams of cement; then jar the bottle to 
remove air bubbles and read the burette. This reading, x, 
represents the volume of 100 grams of cement; and 100, the 
volume of 100 grams of water, divided by x gives the specific 
gravity of the sample. 

63. Among other forms of apparatus which are also of sim- 
ple construction and tend to facilitate the test, may be men- 
tioned the following: - 

M. Candlot 1 devised an apparatus consisting of a graduated 
tube terminating in a bulb at the upper end, the lower end of 
the tube being ground to fit the neck of a flask. The tube and 
flask being disconnected, sufficient liquid is placed in the bulb 
so that when connected with the flask and placed upright, the 
level of the liquid will be at or near the zero mark on the tube. 
The actual level of the liquid is read after standing a few minutes; 
the apparatus is again inverted and the flask disconnected to 
allow of the introduction of 100 grams of cement. The flask is 
then replaced and the contents of the apparatus well shaken to 
expel air-bubbles. When the latter have been completely ex- 
pelled, the flask is placed upright, and after standing a short 
time the level of the liquid is again read, the difference between 
the two readings indicating the absolute volume of 100 grams 
of the cement powder. 

The apparatus devised by M. H. LeChatelier 2 consists of a 
flask of a capacity of about 120 c.c., and having a neck some 
20 c. in length, halfway up which is a bulb having a capacity 



1 "Ciments et Chaux Hydrauliques" par. E. Candlot. 

2 ''Report of Commission des Methods d'Essai des Materiaux de Con- 
struction," The Engineer (London); Illustrated also in Meade's "Examina- 
tion of Portland Cement," Spaulding's "Hydraulic Cement," and Engineer- 
ing News, January 29, 1903. 



SPECIFIC GRAVITY 



55 



-31- 



-30- 



of 20 c.c. Near the bottom of the tube, or flask, is the zero 
mark, and above the bulb the tube is graduated for a length 
corresponding to a capacity of 3 c.c., each graduation repre- 
senting .1 c.c. The diameter of the tube is about 9 mm. The 
zero mark on the tube is below the bulb. The method of opera- 
tion is similar to that described _ 

*~_ * 

above. 

64. The following st} T le of 
apparatus (see Fig. 1) is sug- 
gested as a very convenient 
form, and one which may be 
used for another test soon to 
be described. In this form, the 
flask, of a capacity of about 
200 c.c., has straight sides and 
a flat bottom. The lower part 
of the burette is of large diame- 
ter, about 15 mm., to allow the 
cement to pass readily, while 
the UDper portion is made 
smaller, about 8 mm., to per- 
mit more accurate reading, and 
is graduated from 30 c.c. to 40 
c.c., the divisions being 0.1 c.c. 
Half divisions may be esti- 
mated. The zero mark is in the 
larger part of the burette, but it 
is less difficult to make an ac- 
curate reading at the zero mark, 
since at the time of taking this 
reading the liquid is clear; this 
mark should entirely surround 
the burette. The mouth of the 
bottle and the lower end of the 
burette should be ground to fit, 

and a ground glass stopper should form a part of the apparatus, 
A long pipette will be found convenient for adjusting the level 
of the liquid to the zero mark. 

65. Turpentine is frequently employed for this test, but it 
is somewhat inconvenient to use, since its volume is so sensi- 



o 



/ns/cfe eft*, 



'Botile ZOO Cu.Cent. 



FIG. 1. SPECIFIC GRAVITY APPA- 
RATUS 



56 CEMENT AND CONCRETE 

tive to changes in temperature. This sensitiveness renders it 
imperative that the temperature at the time of taking the final 
reading be the same as when the initial reading is taken, or 
that a correction be applied. To assure this condition the ap- 
paratus should be immersed in a water bath, and the tempera- 
ture of the cement should be the same as that of the turpentine. 
The use of water in the apparatus does not offer this inconven- 
ience, but it is possible that the hydration of the cement during 
the experiment might be sufficient to so affect the volume as 
to change the result, especially with quick-setting cements. 
Light oils, such as benzine and kerosene, are rather volatile, 
but the former (62 Baume naptha) is recommended in the 
preliminary report of the Committee of the American Society 
of Civil Engineers. With the precautions mentioned above, 
turpentine may be used with good results; that which has 
been dried by standing over cement or quicklime is to be 
preferred. 

66. This test may be extended to give interesting and valu- 
able results, in the following manner: When the cement has 
settled in the bottle, leaving the liquid clear, pour off a portion 
of the latter and replace the burette by a glass stopper. Thor- 
oughly agitate the remaining liquid and cement until the latter 
is in suspension; allow the cement to settle again without dis- 
turbance, and it will be found that it is graded in the bottle 
according to its fineness, the coarsest particles being at the 
bottom. With Portland cement, if a portion of the sample is 
underburned it will appear as the top layer, and be indicated 
by its yellow color. It will also be interesting to note what 
proportion of the cement is so fine that the separate grains 
are indistinguishable. That the bottle should have straight 
sides and a flat bottom is to accommodate this part of the 
test, which also dictates the use of some other liquid than 
water. 

67. Effect of Composition, Aeration, Etc. It has been said 
above that the composition of a cement affects its specific 
gravity, a highly limed cement having a higher density. On 
this account an analysis for lime is valuable in connection with 
this test, in order to determine whether a high specific gravity 
is due to a high percentage of lime or to hard burning. 

The age, or aeration of a sample affects its specific gravity 



SPECIFIC GRAVITY 57 

because of the absorption of water from the atmosphere* The 
absorption of two per cent, of water is sufficient to lower the 
specific gravity from 3.125 to 3.000. The following may be 
given as illustrating this point: a certain sample of natural 
cement when taken from the barrel had a specific gravity of 
3.106; after it had been spread out in the air for two months its 
specific gravity was 3.000. A quantity of this aerated cement 
weighing 120 grams was placed in an iron vessel and heated 
over an oil stove for about one hour; at the end of this time 
the cement had lost two grams in weight. The specific gravity 
of the fresh cement being 3.106, 118 grams would have an ab- 
solute volume of 38 c.c.; two grams of water would occupy 2 
c.c., hence 120 grams of the aerated cement would occupy 40 
c.c., and 120 -?- 40 = 3.00, the specific gravity of the aerated 
cement as found above. It is not always possible to thus 
drive off all of the water absorbed, since a portion of it may 
enter into combination with the cement; but a sample should 
always be heated for at least thirty minutes at a temperature 
of 100 C. before making the test for specific gravity, and 
should any appreciable loss of weight occur, it is an indication 
of aeration. 

68. A determination of the specific gravity is primarily a 
test for burning, but it may also be of much value in detecting 
adulterations, as with blast furnace slag or ground limestone. 
An admixture of 10 per cent, of either of these substances would 
suffice to lower the specific gravity from 3.15 to about 3.10. 
The specific gravity of Portland cement ranges from 2.90 to 
3.25, but a first-class product should not show a lower specific 
gravity than 3.05. If fresh Portland gives a result below this 
it is probably either underburned or underlimed, or, perhaps, 
has been adulterated. 

The specific gravity of natural cements has been found to 
vary from 2.82 to 3.25. The specific gravity of one sample of 
underburned natural cement was found to be lower than a 
sample of the same brand which was overburned, but it seems 
very doubtful whether this is true of other brands made from 
rock of a different character. It was also found that the spe- 
cific gravity of the coarse particles of some natural cements is 
lower than that of the fine particles (see Table 10, Art. 15), 
while the opposite is true in the case of Portland cements. 



58 CEMENT AND CONCRETE 

No general rules can be given at present for the interpreta- 
tion of this test that are applicable to all natural cements; it is 
thought that the test will be of value in comparing samples 
of the same brand, though it seems doubtful whether it will 
prove of value in comparing one brand of natural cement with 
another, since it is quite probable that the interpretation may 
vary with the variety of rock used in the manufacture. The 
value of the test for Portland cements is, however, well 
established. 



CHAPTER VI 

SIFTING AND FINE GRINDING 

ART. 14. FINENESS 

69. Importance of Fineness. The fineness of cement is al- 
ways conceded to be one of its most important qualities, and 
the determination of fineness is omitted in none but the very 
crudest tests. Unfortunately, however, sieves that are so coarse 
as to give delusive results are usually employed. It is very easy 
to show that grains of cement as large as one-fiftieth of an inch 
in diameter are practically valueless, but much more difficult 
to determine the point of fineness at which the particles begin 
to have cementitious value. 

70. A moderately coarse sieve is easier to operate than a 
very fine one, less time being consumed in sifting. The impres- 
sion seems to be quite general also that there is a fixed relation 
between the proportions of the different sized grains in different 
samples. Many specifications require that a certain percentage 
"shall pass a sieve having 2,500 holes per square inch." Now, 
there is little doubt that grains of cement larger than .005 inch 
in one dimension have very little cementitious value, and hence 
a cement, all of which would pass holes .015 inch square, while 
but 50 per cent, of it would pass holes .005 inch square, is little 
better than one which leaves a larger residue on the coarser 
sieve but the same residue on the finer. 

In America and Germany it is the usual practice in the pro- 
cess of manufacture to pass the cement through a screen which 
will reject particles larger than about .015 inch in diameter; the 
futility in attempting to determine, with a sieve no finer than 
this, the proportion of the particles which are fine enough to be 
of value, is therefore apparent. Since the English cement 
makers have not been so progressive in the practice of screen- 
ing, they have obtained the reputation of producing a coarse 
product. In many cases this reputation is probably a just one, 
but when tested with a very fine meshed sieve, some of the 

59 



60 CEMENT AND CONCRETE 

English cements do not compare so unfavorably with those of 
German manufacture. It is a curious fact in this connection 
that the English are the most conservative in holding to the 
use of the coarse sieve in testing, which makes their cement 
appear so very much coarser than the American or German 
product. 

71. SIEVES. Sieves for cement testing may be made either 
of wire or silk gauze, set in metal or wood frames. Sieves of 
perforated metal plate are sometimes employed for sifting sand, 
but seldom for cement. It is with considerable difficulty that 
accurate gauze sieves are obtained. They are usually desig- 
nated by numbers corresponding to the number of meshes per 
linear inch; this is in some respects an unsatisfactory method, 
for the size of the wire, which is quite as important as the 
number of meshes, is frequently not given at all, or stated 
in terms of some wire gage which is capable of various 
interpretations. 

As usually supplied by different manufacturers, sieves pur- 
porting to have the same number of meshes per linear inch may 
vary in this regard as much as 10 or 15 per cent. Likewise the 
size of wire used by different makers, in sieves having the same 
number of meshes per inch, may vary quite as much. Again, 
on account of irregularities in the gauze, the holes in a given 
sieve vary one from another; in some cases an opening may be 
but 60 or 70 per cent, as large in one dimension as an adjacent 
one. 

An ideal sieve should conform to the following requirements: 
(1) holes to be of uniform size and shape throughout, (2) sides 
of the holes to be very smooth, and (3) the spaces between the 
holes to be of such size and shape that particles will not easily 
rest there. 

It is evident that the largest holes determine the character 
of the sieve. For example, a sieve having half its holes 0.01 
inch square and the other half 0.02 inch square, would, if used 
long enough, separate the cement exactly as it would if all 
the holes had been 0.02 inch square. Hence, if 'a very small 
percentage of the holes are larger than the normal, it seriously 
impairs the accuracy of the sieve by introducing an inde ter- 
mination; but holes smaller than the normal have no greater 
objection than that, as the sifting proceeds, they become spaces 



FINENESS 



61 



between the real or larger holes, and as such do not fulfill the 
third requirement mentioned above. The shape of the holes, 
whether round, square or hexagonal, seems of minor importance 
so long as uniformity is maintained. The second requirement 
is necessary, because, should particles adhere to the sides of 
the hole, the size of the latter would be decreased to that ex- 
tent. The third requirement is for convenience, but would 
require consideration if the style of the sieve were changed to a 
punched metal plate. 

72. The Committee of the American Society of Civil Engi- 
neers, in their report on " A Uniform System for Tests of Cement " 
in 1885, recommended three sizes of sieves for cement: No. 50 
(2,500 meshes to the square inch) wire to be of No. 35 Stubbs' 
wire gage; No. 74 (5,476 meshes to the square inch) wire to 
be of No. 37 Stubbs' wire gage; No. 100 (10,000 meshes to the 
square inch) wire to be of No. 40 Stubbs' wire gage. For sand, 
two sieves were recommended, No. 20 and No. 30 (400 and 900 
meshes per square inch) wire to be of No. 28 and No. 31 Stubbs' 

TABLE 5 

Sieves: Number of Meshes per Linear Inch and Sizes of 
Openings, as Found by Measurement 







No. OF 










MESHES 

PER 


DlAMKTER OF WIRE 

IN DECIMALS OF 


MEAN SIZE OF OPENING IN DECIMALS 
OF AN INCH. 




o . 


LINEAR 


AN INCH. 








INCH. 






fe 


S > 








w 




a g 


f> 


<*H 








. 


QO 


. 


Os 






^cc 


$> 





Web. 


Woof. 


o 


09 


g 


8 


r< 









g| 


an 

(>" qj 

m 






1 


SI 


|? 


1 


"o 

1 


Remarks. 










o^ 


^ 


Diam- 


Diam- 


s 


y> & 


"S 5 


SH 










t> 


| 


eter. 


eter. 


A 


W 





P 


s* 






a 


b 


c 


d 


e 


/ 


9 


h 


j 


j 




1 


20 


20 


19 


.0185 


.0169 


.0016 


.0315 


.0337 


.0022 


.93 




2 


20 


20 


19 


.0165 


.0168 


.0003 


.0335 


.0358 


.0023 


.93 




3 


30 


30 


28| 


.0119 


.0119 


.0000 


.0214 


.0229 


.0015 


.93 




4 


30 


30 


30 


.0118 


.0118 


.0000 


.0215 


.0215 


.0000 


1.00 




5 


30 


30 


29.V 


.0116 


.0122 


.0006 


.0217 


.0217 


.0000 


1.00 




6 


40 


40 


36 


.0095 


.0095 


.0000 


.0155 


.0183 


.0028 


.85 




7 


50 


50 


47 


.0082 


.0083 


.0001 


.0118 


.0130 


.0012 


.90 




8 


74 


80 


80 


.0054 


.0054 


.0000 


.0071 


.0071 


.0000 


1.00 




9 


100 


101 


88J 


.0040 


.0040 


.0000 


.0059 


.0073 


.0014 


.80 




10 


120 


120 


120 


.0037 


.0037 


.0000 


.0046 


.0046 


.0000 


1.00 




11 


200 


210 


170 


.0022 


.0022 


.0000 


.0026 


.0037 


.0013 


.70 


Approx. 



62 



CEMENT AND CONCRETE 



wire gage, respectively. It seems to be impracticable to com- 
ply with these sizes of wires, because neither manufacturers nor 
engineers appear to agree as to what diameters of wire corre- 
spond to No. 37 and No. 40 Stubbs' wire gage. 

73. The conferences of Dresden and Munich decided that 
fineness should be determined by sieves of 900 and 4,900 meshes 
per sq. cm., respectively, for Portland cement, and 900 and 
2,500, respectively, for other hydraulic products, the size of the 
wires being as follows: for 4,900, .05 mm.; for 2,500, .07 mm.; 
and for 900, .10 mm. These sieves would have respectively 
31,600 (178 X 178), 16,000 (127 X 127), and 5,800 (76 X 76) 
meshes per square inch, and the sizes of the holes would be 
approximately .0037 inch square, .005 inch square and .009 inch 
square, respectively. It was also decided that for sifting sand, 
punched metal plates were preferable to wire cloth sieves. 

74. In Table 5 are given some of the results obtained by the 
writer which will serve to show what variations may exist in 
sieves which have been selected from a considerable number 
offered for use. 

Table 6 gives the data available concerning certain sieves 
that have been used or recommended in this country and else- 
where. 

TABLE 6 
Sizes of Openings in Sieves Recommended or in Use 





wj 

s 


3rt 

Sg 


SIZE 




REF. 


1.3 S 

^~ 


*K 

Eta S 


HOLE, 
INCH 


REMARKS. 




M*" 1 


N2 


SQUARE. 






g 


oaQ 









a 


b 


c 




1 


178 


.00197 


.00366 


Established by Conferences, Dresden & Munich. 


2 


127 


.00276 


.00512 


tt u tt tt tt 


3 


76 


.00394 


.00920 


u tt tt tt tt 


4 


76 


.00437 


.00875 


Present German Standard. 


5 


76 


.00591 


.00721 


Recommended by H. LeChatelier. 


6 


176 


.00162 


.004 


Silk mesh Vyrnwy Reservoir. 


7 


103 


.0022 


.0075 


tt tt tt tt 


8 


170 


.00279 


.00309 


Cornell University, Marx & Mosscrop, 1887. 


9 


80 


.00651 


.00599 


tt tt tt tt tt 


10 


50 


.00881 


.01119 


tt tt tt tt tt 


11 


30 


.01214 


.02119 


tt tt tt tt tt 


12 


20 


.01899 


.03101 


tt tt tt tt tt 


13 


200 


.0024 


.0026 


Progress Report, A. S. C. E. Committee, 1903. 


14 


100 


.0045 


.0055 


tt tt tt tt tt tt ' 



SIFTING AND GRINDING 



63 



75. The time sifting should be continued will depend on the 
fineness of the meshes, the diameter of the sieve, the amount 
of cement taken, and the manner of sifting ; it will also depend 
upon the fineness of the cement, as well as its nature, and its 
condition as to dry ness. But, although some care is necessary 
concerning these points, very large variations in results due to 
variations in the time the sifting is continued may easily be 
avoided. The diameter of the sieve is usually made greater 
for the finer meshes, but this is not always the case. It is a 
common practice in America to use one-tenth of a pound of 
cement in testing the fineness, using a scale weighing in ten- 
thousandths of a pound. Where the metric system is in use 
(and it may well be adopted in a cement laboratory), 100 grams 
of cement are usually taken. 

76. M. H. LeChatelier recommends a sieve having 900 meshes 
per sq. cm., of wire 0.15 mm. diameter, giving holes 0.18 mm. 
(.0072 inch) square. He prefers machine screening, but says 
that for current tests it might be sufficient to screen by hand 
for ten minutes with a sieve three decimeters (about 12 inches) 
in diameter. 

Table 7 is taken from experiments made by M. Durand- 
Claye and M. Candlot, and shows what differences may arise 
from varying the length of time that a sample is screened. The 
cements used were not the same in the two cases, but the sieves 
had each 5,000 meshes per sq. cm. (about 180 per linear inch), 
and 100 grams of cement were taken in each case. Had a coarse 
sieve been used, the differences would have been much less 
after the same lengths of time. 

TABLE 7 
Fineness: Mechanical and Hand Sifting Compared 



M. 'DURAND-CLAYE. 


M. CANDLOT. 


MECHANICAL SIEVE MAKING 200 


HAND SIEVE, 12 INCHES IN 


REVOLUTIONS PER MINUTE. 


DIAMETER. 


No. 
Revolutions. 


Per Cent. 
Retained. 


Ditf. 


After 
Minutes. 


Per Cent. 
Retained. 


Diff. 


600 


41.2 




5 


2&6 




1,000 


39.4 


1.8 


10 


29.1 


0.5 


1,500 


38.6 


.8 


20 


28.4 


0.7 


2,000 


38.0 


.6 


30 


28.0 


0.4 


2,500 


37.6 


.4 


40 


27.7 


0.3 



64 



CEMENT AND CONCRETE 



77. Table 8 gives the results obtained by the author in 
sifting several samples. The No. 80 sieve was about 6J inches 
in diameter, and Nos. 120 and 200 about 5J inches. One hun- 
dred grams of cement were taken in each case, and the sieve 
was shaken vigorously by hand. It is seen that coarse samples 
require less time for sifting than fine samples, and that natural 
cements require a longer time than Portlands. With the No. 
80 sieve, five minutes usually suffices to obtain the fineness, 

TABLE 8 

Effect of Time of Sifting on the Result Obtained in Testing 

Fineness 



REFERENCE. 


SIEVE. 


CEMENT. 


PEK CENT. BY WEIGHT THAT HAD 
PASSED SIEVE AFTEB SIFTING. 


Approximate 
No. of Meshes 
per Inch. 


Size Holes, 
Inch Square. 


Kind. 


Brand. 


Sam- 
ple. 


1 Minute. 


3 Minutes. 


5 Minutes. 


7 Minutes. 


1 
o 

J 


15 Minutes. 


20 Minutes. 


25 Minutes. 


a 


b 


c 


d 


e 


/ 


ff 


h 

93 

86 
96 


i 
93 


k 


I 


m 


1 

2 
3 
4 

5 
6 

7 
8 

9 

10 
11 
12 
13 
14 
15 
16 

17 

18 
19 
20 
21 
22 
23 
24 


80 j 
120 | 
200 j 


.0071 

by 

.0071 
.0046 

by 

.0046 
.0036 

by 

.0037 


Port. 
Nat. 
Port. 
Nat. 
Port. 
Nat. 


X 

Y 

S 
Z 
Bn 
An 
In 
Gn 

X 

Y 
S 
Z 
Bn 
An 
In 
Gn 

X 

Y 

S 
Z 
Bn 
An 
In 
Gn 


685 

42s 
34s 
43s 
27s 
G 
28s 
108 T 

685 

42s 
34s 
43s 

27s 
G 
28s 
108 T 

685 

42s 

34s 
43s 

27s 
G 
28s 
108 T 


91 

82 
95 
100 
68 
78 
85 
75 

45 

41 
72 
54 
27 
45 
13 
5 


92 

85 
96 




































71 

80 
89 
89 

78 

73 
89 
94 
62 
73 
42 
16 

58 

65 

72 
74 
57 
67 
41 
40 


71 
82 
90 
90 

82 

77 
90 
96 
65 
76 
64 
27 

65 

68 
76 
78 
59 
69 
53 
49 












82 
90 
91 


















91 
84 

78 
91 
97 
66 
78 
82 
64 

68 

71 
78 
81 
60 
70 
69 
64 


84 

78 
91 
98 
66 
78 
83 
82 

70 

72 
80 
82 
60 
72 
75 
76 


























'sV 

71 

72 
81 
83 


86* 


72 
77 
79 


78 



FINENESS 



65 



and with the No. 120 sieve, but little cement usually passed after 
the sifting had continued ten minutes, though with one brand 
of natural, Gn, it appears that the true fineness would not be in- 
dicated by sifting less than 20 minutes. With the No. 200 
sieve 20 minutes is usually required, and in the case of two samples 
of natural cement, a still longer time appears to be necessary. 

78. Conclusions. Until there is a proper standard in the 
United States concerning sieves and methods of sifting, the best 
that can be done is to select, from the sieves that manufacturers 
have to offer, those which appear to be most nearly uniform in 
size of mesh, and then actually determine the size of the holes. 
This may be done by counting, under the magnifying glass, the 
number of meshes per inch each way, and determining the size 
of wire with a micrometer wire gage. 

As to the time sifting should be continued, one can easily 
find by trial the time required in using a given sieve in order to 
confine the error within given limits. A fine natural cement 
should be selected to determine this, as such a cement requires 
the longest sifting. Care should be taken that the cement is 
well dried before making the test for fineness. It will be found 
that for sieves having holes between .003 inch and .004 inch 
square (sieves approximating 170 to 200 meshes per linear 
inch) 20 to 30 minutes are required, while for sieves having 
holes .007 to .009 inch square (approximately 70 to 100 meshes 
per linear inch) from five to ten minutes will usually suffice. 

79. Specifications for Fineness. The following table has 
.been compiled to show what are considered reasonable require- 
ments for fineness. In most specifications there is the usual 
indetermination concerning the sizes of holes in the sieves. 

TABLE 9 
Requirements as to Fineness 



SPECIFICATION. 


DATE. 


PER CENT. REQUIRED TO PASS 
SIEVE HAVING 10,000 
HOLES PER SQUARE INCH. 


Portland. 


Natural. 


U. S. Army Engineers .... 
U. S. Navy Department . . . 
City Pittsburg Pa 


1901 

1900 
1897 
1896 
1895 


92 
95 
90 
90 
95 
85 


80 

'??' 

80 


New East River Bridge . . . 
Topeka, Kan., Bridge . . . 
Master Builders' Exchange, Phila. 



66 



CEMENT AND CONCRETE 



ART. 15. COARSE PARTICLES IN CEMENT 
80. The Effect of Coarse Particles on the Weight of Cement. 
- To remove the coarse particles by sifting will reduce the 
specific gravity of a sample of Portland cement, as the un- 
ground particles are from the harder burned and denser por- 
tion of the clinker, and to remove these denser particles will, 
of course, decrease the average density of the sample. This is 
not always the case with natural cements, as is shown by the 
following tests : 

TABLE 10 

The Relative Specific Gravity of Coarse and Fine Particles of 

Cement 





CEMENT. 






Kind. 


Brand. 


Fineness. 


SPECIFIC GRAVITY. 


Portland 


K 


As received . . . 


3.086 




u 


50-100 


3 145 




X 


Pass 50 . . . 


3.039 




' 


Ret. on 50 


3.125 


Natural 


Gn 


Pass 100 . 


2.874 




u 


Ret. on 50 


2.817 




An 


Pass 60 . 


2.945 




" 


Ret. on 50 


2.817 



The apparent density or weight per cubic foot of Portland 
will be reduced more than the specific gravity by the removal 
of the coarse particles; because not only will the true density 
be decreased, but the packing, which is facilitated by a wide 
range in the sizes of the particles, will be less perfect than when 
the coarse particles are present. In 89 a table is given show- 
ing the changes in specific gravity and weight per bushel oc- 
casioned by removing the coarse particles by sifting. 

81. Effect of Coarse Particles on the Time of Setting. 
Table 11 gives the results of a number of tests on Portland and 
natural cements to determine the relative time of setting of 
samples from which the coarse particles had been removed by 
the No. 200 sieve, while Table 12 gives results obtained with 
a sample of natural cement of varying fineness. 

In Table 11, 30 per cent, of water was used for all Portland 
cements, and 36 per cent, for all naturals, but the consistency 



COARSE PARTICLES 



67 



varied as stated in the table. It is seen that in nearly every 
case the setting was hastened by removing the coarse particles, 
though this may have been due in part to the fact that with the 
same percentage of water the finer cement gave a stiffer paste. 
For the tests in Table 12, the attempt was made to make all 
of the mortars of the same consistency by varying the percent- 
age of water. As would be expected, the coarse particles are very 
slow setting. In fact, what hardness they attained was prob- 
ably due largely to the fine dust that adhered to the grains. 
These coarse particles may be considered as practically inert, 
and their presence in a sample would naturally make it slow 
setting. To show this by actual test, however, is very difficult, 

TABLE 11 
Effect of Coarse Particles on the Time of Setting 





CEMENT PASSING No. 20 


CEMENT PASSING No. 200 


CEMENT. 


SIEVE. 


SIEVE. 






Time to bear 




Time to bear 




Kind. 


Brand. 


i Ib. wire. 
Minutes. 


Consistency. 


J Ib. wire. 
Minutes. 


Consistency. 


Portland 


Y 


39 


Trifle moist 


13 


Trifle dry 


" 


X 


9 


Moist 


4 


0. K. 





Z 


432 


Trifle moist 


354 


Trifle dry 


it 


s 


556 


<t 1 1 


341 


tt it 


Natural 


Gn 


31 




29 




d 


Bn 


143 


Trifle moist 


151 


Trifle dry 


u 


In 


397 


u tt 


256 




it 


Hn 


256 


. . . 


233 





NOTE : 30 per cent, water used for all Portlands. 

36 per cent, water used for all natural cements. 

TABLE 12 
Effect of Coarse Particles on Time of Setting 

Natural Cement, Brand Gn All pastes appeared same consistency. 





WATER USED AS 


TIME TO BEAR 


TIME TO BEAR 


FINENESS. 


PER CENT. OF 


J LB. WIRE. 


1 LB. WIRE. 




CEMENT. 










Minutes. 


Minutes. 


Pass No. 20 sieve 


33 


14 


159 


" 50 " 


36 


29 


219 


" 100 " 


38 


24 


214 


Retained on 50, 








reground to pass 100 


28 


73 


670 


Pass No. 50, retained 








on No. 100 


34 


205 


800 



68 CEMENT AND CONCRETE 

as the amount of water required to bring the mortars to the 
same consistency varies with the amount of coarse particles 
present, and as there is no very satisfactory method of testing 
the consistency, the tests for time of setting have in them this 
indetermination. 

82. Effect of Coarse Particles on the Tensile Strength. A 
cement having a certain quantity of coarse particles will fre- 
quently give a higher tensile strength when tested neat than a 
cement from which the coarse particles have been removed by 
screening. The reason for this may be found in the fact that 
a wide range in the sizes of grain of the powder facilitates pack- 
ing, both when dry and when mixed with water to form a paste. 
Another reason is that the unground particles are stronger 
than the hardened mortar, and, considering the broken section 
of a briquet, the break does not take place through these par- 
ticles, but they are pulled out of their bed; this virtually in- 
creases the area of section. Were the same sample of cement 
reground, so that a certain proportion of the coarse particles 
was rendered active, it might then give a higher strength, neat, 
than at first. If so, the reason would be found in the fact that 
the coarse particles, being the hardest burned, were really from 
the best part of the cement clinker, and rendering these parti- 
cles active by fine grinding increased the cohesive properties of 
the cement so much as to overcome the physical effect of the 
coarse particles, which, when judged by neat tests, appear to be 
beneficial. The above serves to illustrate the difference be- 
tween sifting and fine grinding which are so frequently con- 
fused in treating this subject. 

83. Among the many tests that have been made to show 
the effect of sifting on the cohesive and adhesive strength of 
cements, a few may be given as follows: - 

Mr. Maclay l gives a few experiments to show that the pres- 
ence of coarse particles increases the cohesive strength, neat, 
seven days. 

Lieut. W. Innes 2 gives two tables of results obtained by ex- 
perimenting on very coarse cements. The tables show that 
removing the particles that would not pass through sieves of 



1 Trans. Am. Soc. C. E., Vol. vi. 

2 Minutes Proc. Inst. C. E., Vol. xxv. 



COARSE PARTICLES 69 

1,296 meshes and 2,500 meshes per square inch, decreased the 
strength when tested neat at the ages of three months and six 
months; but increased the strength when sand mortars were 
used. The differences at six months were relatively somewhat 
less than at three months. By separating a sample of cement 
into two parts, that passing a sieve having 2,500 meshes per 
square inch and that retained on the same sieve, and then 
remixing the screenings with the fine portion, he found 
that the highest strength, neat, six months, was given by the 
mixture containing the largest amount tried (70 per cent.) of 
screenings. 

84. In the tests of cement for the Cairo Bridge 1 a series of 
experiments was made to determine the effect of coarse parti- 
cles on the value of both Portland and natural cements. The 
cement was separated into two parts, by a sieve having 10,000 
meshes per square inch. Briquets were made both neat and 
with sand, the cement used being made of 100, 90, 80, 70 and 
60 volumes of sifted cement to 0, 10, 20, 30, and 40 volumes, 
respectively, of cement screenings. The briquets were broken 
when six months old. 

It was found that in the case of Portland cement, neat, the 
highest result was obtained with the largest (40) per cent, of 
screenings, but with one and two parts sand, the strength 
steadily fell as larger amounts of screenings were used. With 
Louisville natural cement the presence of screenings seemed to 
have little effect on neat tests; and with one part of sand to one 
of cement, the use of as much as 30 per cent, of screenings to 
70 per cent, of sifted cement did not appear to decrease the 
strength. With two parts sand to one cement, the results were 
slowly diminished by successive additions of larger percentages 
of screenings. 

85. M. R. Feret is said to have replaced with sand the grains 
of cement retained on sieves having 5,800 and 32,300 meshes 
per square inch, and found that, except in the case of neat ce- 
ment mortars, the substitution of sand for coarse particles -of 
cement did not decrease the strength. In experimenting on 
this subject Mr. Eliot C. Clarke 2 found that the coarse particles 



1 Jour. Assn. Engr. Soc., 1890, and Engineering News, Jan. 31, 1891. 

2 Trans. A. S. C. E., Vol. xiv, pp. 158-162. 



70 



CEMENT AND CONCRETE 



TABLE 13 
Effect of Removing Coarse Particles from Natural Cement 



CEMENT. 


No. PABTS 
SAND TO 
ONE 
CEMENT 

BY 

WEIGHT. 


WATER AS 
PER 
CENT. OF 
WEIGHT 
DRY 
INGREDI- 
ENTS. 


TENSILE STRENGTH, POUNDS PER SQUARE INCH. 


7 da. 


28 da. 


3 mo. 


6 mo. 


2 years. 


A 


None 


33.3 


120 


275 








B 


" 


35.7 


100 


266 








C 





38.5 


83 


253 








D 


( 


28.3 


202 


264 


. 






E 





35.0 


127 


143 


. . . 


. . . 


. . . 


A 


One 


19.0 


121 


253 




330 


380 


B 


ii 


19.6 


104 


251 


. 


344 


398 


C 


it 


20.0 


94 


261 


. . . 


331 


385 


D 


t( 


16.0 


286 


360 


. . . 


396 


385 


A 


Two 


16.1 




168 


215 


223 


210 


B 


K 


16.1 


. 


203 


245 


267 


302 


C 


u 


16.1 


. 


218 


297 


317 


358 


D 


K 


13.9 


. . 


227 


230 


245 


262 


E 


it 


15.8-16.1 


. . . 


71 


40 


57 


50 


A 


Three 


14.5 








127 


128 


B 


u 


14.5 


. 


. 


. 


167 


164 


C 


u 


145 


. 




. 


205 


234 


D 


It 


12.1 






. . . 


125 


118 



Fineness of Cement 





PER CENT. PASSING SIEVE No. 


50 


100 


120 


Cement A 


82 
100 


70 
85 
100 


64 

78 
91 


Cement B . 


Cement C 





NOTE: All cement from same barrel, Brand Bn, Sample 27s. 

Sand, crushed quartz 20-30. 

All briquets made by one molder and stored in one tank. 

All results, mean of 5 briquets, except two which are means of 

ten and two briquets, respectively. 

A Cement passing No. 20 sieve, holes .033 inch square. 
B- " " " 50 " " .012 " 

C- " " 100 " " .0065 

D " retained on No. 50, reground to pass No. 100. 
E " passing No. 50, retained on No. 100. 



COARSE PARTICLES 



71 



of cement were somewhat better, for use in mortar, than fine 
sand, but very little better than coarse sand. 

86. The tests given in Table 13 were made under the aiH 
thorns direction to determine the effect of sifting and the value 
of coarse particles. It is seen that in neat tests the strength is 
slightly diminished by sifting out the coarse particles; in the 
tests of mortars containing equal parts by weight of sand and 
cement, there is little difference in the strength of the three 
samples, though the coarser cement appears to gain its strength 
a little more rapidly. With two parts sand to one of cement,, 
the greater value of the fine particles is very noticeable, and 
with one-to-three mortars the difference is still more marked, 
the sifted cement giving 80 per cent, greater strength than the 
unsifted. 

87. In Table 14 these results are arranged in a different 
way. If we assume that the particles that will not pass the 
No. 120 sieve are not cement at all, but equivalent to sand, 

TABLE 14 

Effect on Tensile Strength of Removing Ccarse Particles from 
Natural Cement 









PARTS SAND AND 






PER CENT. 


PARTS SAND 


COARSE PAR- 


STRENGTH 


CEMENT. 


PASSING SIEVE 


TO 


TICLES TO ONE 


OF MORTAR AFTER 




No. 120. 


ONE CEMENT. 


PART 


Two TEARS. 








FINE PARTICLES. 




A 


64 


3 


5.2 


128 


B 


78 


3 


4.1 


164 


A 


64 


2 


3.7 


210 


C 


91 


3 


3.4 


234 


B 


78 


2 


2.8 


302 


G 


91 


2 


2.3 


358 


A 


64 


1 


2.1 


380 


B 


78 


1 


1.6 


398 


C 


91 


1 


1.2 


385 



and that all particles passing this sieve are cement, we obtain 
a new set of proportions of sand to cement. Thus the sample 
of cement passing No. 20 sieve, sample A, would be composed 
of 64 parts cement and 36 parts sand, and the 1 to 3 mortar 
would have in reality the proportion 64 cement to 336 sand, or 
1 to 5.2. It is seen that the tensile strength bears a closer 
relation to the richness of the mortar when considered in this 
way. There is, of course, no abrupt division in size such that 



72 



CEMENT AND CONCRETE 



TABLE 15 
Value of Coarse Particles of Cement, Natural and Portland 



ocooo-ac50T^cou>i-i | REFERENCE NUMBER. 1 


TENSILE STRENGTH, POUNDS PER SQUARE INCH. 


Neat Cement. 


1 Part Standard 
Sand to 1 Cement. 


3 Parts Standard 
Sand to 1 Cement. 


3 Parts Limestone 
Screenings, 
(g) to 1 Cement. 


3 mos. 


4 mos. 


lyr. 


3 mos. 


lyr. 


3 mos. 


lyr. 


3 mos. 


lyr. 


d 


e 


/ 


g 


h 


i 


j 


ft 


l 


330 

259 
295 
306 
309 


'630 
550 
621 
615 
591 


390 
336 
334 
370 
343 
706 
553 
665 
651 
764 






108 
193 
203 
102 
92 
378 
423 
455 
357 
362 


139 
217 
224 
106 
96 
395 
463 
469 
399 
354 


160 
269 
251 
137 
139 
472 
538 
561 
425 
405 


234 

332 
319 
170 
155 
589 
677 
676 
568 
506 


















786 
755 
745 
746 
765 


812 
838 
891 
841 
837 



Ref. No. 1. Natural cement passing No. 20 sieve. 
" " 2. Natural cement passing No. 80 sieve. 
" " 3. Natural cement reground before sifting, until all passed 

No. 80 sieve. 

" " 4. Natural cement, 64f per cent, of cement passing No. 80 
sieve mixed with 35^ per cent, of limestone screen- 
ings retained between Nos. 20 and 80 sieves. 

" " 5. Natural cement, 64f per cent, of cement passing No. 80 
sieve mixed with 35| per cent, of crushed quartz 
retained between Nos. 20 and 80 sieves. 
" " 6. Portland cement passing No. 40 sieve. 
tt d 7 Portland cement passing No. 80 sieve. 
" " 8. Portland cement reground (before sifting) until all passed 

No. 80 sieve. 

" " 9. 81 per cent, of cement passing No. 80 sieve mixed with 
18 per cent, limestone screenings retained between 
Nos. 40 and 80 sieves. 

" " 10. 81 per cent, of cement passing No. 80 sieve mixed with 
18 per cent, crushed quartz retained between Nos. 
40 and 80 sieves. 

Of the natural cement passing No. 20 sieve, 35 per cent, was retained 
on sieve No. 80, while 64f per cent, passed the No. 80 sieve. In lines 4 and 
5 the coarse particles of cement (20-80) were removed and replaced by an 
equal weight of sand grains, retained between sieves 20 and 80. 

Of the Portland cement passing No. 40 sieve, 18J per cent, was retained 
on sieve No. 80, while 81 per cent, passed sieve No. 80. In lines 9 and 10 
the coarse particles of cement (40-80) were removed and replaced by an equal 
weight of sand grains retained between sieves 40 and 80. 

All briquets made by same molder, each result mean of five specimens. 



FINE GRINDING 73 

coarser particles act only as sand, while finer ones enter into 
combination as cement; part of the coarse particles will have 
some cementitious value, while some of the finer particles will 
have somewhat the effect of sand. 

As to the sample composed of coarse particles reground, it 
must be considered that although this sample was passed through 
the No. 100 sieve, yet it was in reality much coarser than sam- 
ple C, because the particles were harder, and the grinding 
in the mortar less thorough than the original grinding. Since 
this sample of reground cement gives so high a strength neat 
and with one part sand, it appears that the hard particles from 
which it was made are of excellent quality if ground fine enough, 
and the relatively lower results with larger proportions of sand 
must be attributed to imperfect grinding. 

The coarse particles retained between sieves 50 and 100 gave 
a higher strength neat than was expected, but much of this 
strength may be due to the floury portion of the cement that 
doubtless adhered to the coarse particles instead of passing 
through the sieve. 

88, The tests in Table 15 were made to determine whether 
the coarse particles of cement are of greater value in mortar 
than the same quantity of fine sand. The coarse particles of 
the cement were sifted out and replaced with sand grains of 
about the same size. The conclusion drawn from the preced- 
ing tests would indicate that some of the coarse particles of 
cement might be replaced by sand without diminishing the 
tensile strength; but the tests given in this table indicate that 
this is not the case when it is a question of substituting sand 
grains of the same size. Although such a substitution has 
little effect on the strength of rich mortars, it results in a de- 
creased strength with mortars containing as much as three 
parts sand to one of cement by weight. (See 85 in this con- 
nection.) 

ART. 16. FINE GRINDING 

89. Effect of Fine Grinding on the Weight of Cement. Fine 
grinding will decrease the weight per cubic foot, the fine ce- 
ment not packing as closely as the coarser product. In " Ce- 
ment for Users," by Mr. Henry Faija, the following results are 
given, showing the relation between fineness, weight, and spe- 



74 



CEMENT AND CONCRETE 



TABLE 16 
Relation of Fineness to Specific Gravity and Weight per Bushel 

From " Cement for Users " 



SAM- 
PLE. 


SPECIFIC GRAVITY. 


WEIGHT PER BUSHEL. 


a 


b 


c 


a 


b 


c 


d 


e 


1 


3.00 


2.97 


3.07 


116.5 


107.5 


121.0 


112.0 


115 


2 


3.03 


2.94 


3.04 


116.0 


104.0 


130.5 


109.0 


115 


3 


3.02 


2.91 


3.035 


114.0 


100.0 


128.0 


104.5 


109 



cific gravity: (a), cement as delivered; (b), sif tings that passed 
through sieve with 2,500 holes per sq. in.; (c), coarse, retained 
on above sieve; (d), cement all ground to pass above sieve; (e), 
coarse particles reground to pass above sieve. 

90. Effect of Fine Grinding on Time of Setting. Since the 
coarse particles of cement are practically inert, there is every 
reason to believe that finer grinding will increase the activity 
of a sample, since it will render some inert particles active. 
For the reason mentioned in 81, however, it is difficult to show 
this difference in time of setting by actual tests. 

Tests reported by Mr. David B. Butler 1 showed that several 
Portland cements which took an initial set in 20 to 30 minutes 
and hard set in 45 to 120 minutes would, when reground to pass 
a sieve having 180 meshes per linear inch, begin to set in from 
1 to 7 minutes and set hard in 5 to 15 minutes. These may be 
considered extreme results; the rise in temperature of these 
cements during setting was so great as to indicate they were 
not normal cements, and variations in consistency of the pastes 
may have influenced the time of setting. 

91. EFFECT OF FINE GRINDING ON STRENGTH. since the 

best burned clinker of Portland cement is the hardest, it follows 
that the unground particles would, if ground fine enough to 
become active, form the best portion of the cement. This is 
not, a priori, true of natural cements, because burning renders 
some varieties of cement rock softer at first, but when the burn- 
ing is carried beyond a certain point they become harder again. 
The coarse particles in a natural cement may thus be either 



Proceedings Inst. C. E., 1898. 



FINE GRINDING 



75 



from underburned or overburned rock; hence it is possible that 
in some cases it might be better to leave the hardest particles 
in an unground state. Thus, while it has been generally ac- 
cepted that fine grinding improves Portland in a twofold de- 
gree, by bringing into action the best burned clinker, as well 
as by rendering a given weight of cement capable of coating a 
larger number of sand grains, a similar conclusion concern- 
ing natural cement is not well established. 

TABLE 17 

Effect of Fine Grinding of Natural Cement on the Tensile 
.' Strength of Mortar 



REFERENCE. 


TENSILE STRENGTH, POUNDS PER SQUARE INCH. 


Neat 
Cement. 


1 Part Stand- 
ard Sand 
to 1 Cement. 


2 Parts Standard Sand 
to 1 Cement. 


3 Parts 
Standard 
Sand to 1 Ce- 
ment. 


4 Parts 
Standard 
Sand to 1 Ce- 
ment. 


7 da. 


6 mo. 


7 da. 


28 da. 


28 da. 


3 mo. 


6 mo. 


2yr. 


6 mo. 


2yr. 


6 mo. 


2yr. 


a 


b 


c 


d 


e 


/ 


9 


h 


i 


J 


k 


I 


1 
2 
3 
4 
5 


268 
283 
278 
392 


538 
473 

538 
592 


224 

230 
307 
368 


381 
350 
433 
538 


207 
245 
292 
271 
21 


354 
433 
469 
344 


291 
426 
406 
369 
73 


70 
102 
92 
160 
45 


202 

302 
305 

274 


48 
65 
61 
110 


156 
212 
240 
205 


49 

78 
65 
90 




















REFERENCE. 


FINENESS OF CEMENT, 
PER CENT. PASSING 
Sieve Number. 


100 


120 


1. Cement as received passed through No. 20 76.5 
2. Cement as received passed through No. 100 100.0 
3. Reground in mortar, not sifted .... 95.8 


72.4 
94.6 
91.5 



Cement; Natural, Brand Jn. 
No. 1. Passing No. 20 sieve. 
" 2. Passing No. 100 sieve. 
" 3. Reground before sifting. 

" 4. Particles retained on No. 50 sieve, reground to pass No. 100 sieve, 
" 5. Particles retained on No. 50 sieve, reground to pass No. 50 sieve, 

but retained on No. 100 sieve. 

All briquets made by one molder and immersed in one tank. In general, 
each result is mean of five specimens. 



76 CEMENT AND CONCRETE 

92. Some tests bearing upon the value of fine grinding have 
already been given in Table 15. Samples 3 and 8 were reground 
with mortar and pestle before being sifted. If we compare the 
results given by sample 3 with those obtained with samples 1 
and 2, not reground, it appears that the regrinding diminishes 
the strength in neat mortars but increases it in mortars con- 
taining three parts sand to one of cement. Regrinding ap- 
pears to be no better, however, than sifting. Comparing sam- 
ple 8 with samples 6 and 7, it is seen that regrinding Portland 
cement does not diminish the strength in neat mortars to the 
same extent as sifting does, and in sand mortars regrinding 
generally results in a greater increase in strength than sifting. 

93. The results in Table 17 were obtained with another 
sample of natural cement and are of greater practical value as 
indicating the importance of fine grinding, since in these tests 
a sample is included obtained by regrinding the original cement 
without previous sifting. The conclusions concerning the ce- 
ment retained on No. 50 sieve reground to pass No. 100, and 
the coarse particles alone retained between sieves 50 and 100, 
are practically the same as those drawn from Table 13. 

As to the other three samples, the No. 20 sieve removed only 
a very few coarse particles, and that passing this sieve may be 
considered to represent the cement as received The No. 100 
sieve removed about 24 per cent, by weight from the original 
cement, and the cement that was reground contained but about 
4 per cent, of particles which would not have passed the No. 
100 sieve. The third sample, reground cement, may be com- 
pared with the first to indicate the improvement obtained by 
finer grinding, and it may be compared with the second to de- 
termine the difference between removing the coarse particles by 
sifting and reducing them by finer grinding. In considering 
these results it will be best to neglect the two-year tests, since 
all of the samples failed at this age. A comparison of the re- 
sults obtained with these three samples indicates that while the 
advantage of finer grinding is not apparent in neat tests, in 
sand mortars the value of finer grinding is more marked the 
larger proportion of sand used, so that with three or four parts 
sand, the strength with the fine samples is about 50 per cent, 
greater than with the cement as received. It also appears that 
the reground sample gains its strength more rapidly than the 



FINE GRINDING 77 

sifted sample, though at six months it seems to make little dif- 
ference whether the coarse particles are removed by sifting or 
reduced by grinding. 

94. Conclusions as to the Effect of Fine Grinding and Sifting 
on Tensile Strength. The general conclusions to be drawn 
concerning fine grinding and sifting may be summarized as fol- 
lows: According to the tests given, it appears that to remove 
the coarse particles from a sample of natural cement by sifting, 
or to reduce them by finer grinding, generally diminishes the 
strength obtained in tests of neat cement mortars. In one-to- 
one mortars, the strength of the finer samples is not much 
greater than when the coarse particles are present; but in mor- 
tars containing greater proportions of sand, the advantage ob- 
tained by eliminating the coarse particles is very marked 
in the case of natural cement, the strength given by the 
finer samples sometimes exceeding that of the original cement 
by more than 60 per cent. While the advantages of sifting 
and finer grinding are also important for Portland cements, 
there does not result such a large proportionate increase in 
strength. 

Reground samples of natural cement gain strength more 
rapidly than resifted samples, but eventually the strength 
attained is about the same. In Portland cements regrinding 
seems to be of greater value than resifting. A sample of natural 
cement made from coarse particles reground gains strength 
rapidly, and for mortars with small proportions of sand, gives 
good results. The fact that such samples do not give a high 
strength with large proportions of sand is doubtless due to the 
fact that the grinding is "not thorough, and the indications are 
that the material of which such coarse particles are composed 
would form a valuable part of the cement if ground fine 
enough. 

The coarse particles of either natural or Portland cement 
may be replaced by grains of sand of the same size without 
materially affecting the strength attained by neat and one-to- 
one mortars, but for mortars containing larger proportions of 
sand, such a substitution results in a decreased strength. 

95. Finally, it may be said that the process of manufacture 
and the character of the materials from which cement is made 
have such an influence on the relative proportions of fine and 



78 CEMENT AND CONCRETE 

coarse particles that the percentage of finest particles cannot 
be determined by testing with a coarse sieve. While it is not 
known at what point of fineness grains of cement begin to have 
cementitious value, or what proportion of the cement should be 
the finest flocculent matter, it is certain that a cement should 
leavers small a percentage as possible on a sieve having holes 
.004 inch square, in order to have the greatest sand carrying 
capacity. 

There is, however, a reason for using a comparatively coarse 
sieve in connection with the fine one. Overburned lime, which 
is likely to occur in Portland cements, is more dangerous in the 
form of coarse particles than an equal quantity in a fine condi- 
tion, because coarse particles slake more slowly and it is better 
that expansion should occur early in the process of hardening 
if it is to occur at all. For the same reason a cement that would 
be unsound normally may be rendered less dangerous by re- 
grinding. 

As fine grinding is expensive, it is only a question as to when 
the increased strength obtained is offset by the extra expense 
incurred in grinding. There is now little trouble in obtaining 
either natural or Portland cement of which from 70 to 85 per 
cent, will pass holes .004 inch square. (See 77.) 



CHAPTER VII 

TIME OF SETTING AND SOUNDNESS 

ART. 17. SETTING OF CEMENT 

96. Process of Setting. When cement is gaged with suffi- 
cient water to bring it to a paste, and is then left undisturbed, 
it soon begins to lose its plasticity and finally reaches such a 
condition that its form can no longer be changed without pro- 
ducing rupture. This change of condition is known as the 
"setting" of cement and is considered to be, in a measure, dis- 
tinct from " hardening." Setting usually takes place within a 
few hours, or perhaps minutes, while the hardening is continu- 
ous for months or years. 

The precise chemical changes that take place in the setting 
and hardening of cements are not thoroughly understood. The 
chief cementitious ingredient in Portland cement is considered 
to be a tricalcium silicate, 3 CaO, SiO 2 ; in contact with water it 
forms hydrated monocalcic silicate and calcium hydrate. This 
process is believed to contribute more to the final hardening of 
the mortar than to the setting, though the hydration of the 
finer particles of this important compound also contributes to 
the first setting. It is considered that the calcium aluminates 
play an important role in the first setting of cement, as they set 
rapidly in contact with water, and it has been suggested that 
they form the chief active constituents of natural cement. 1 

These chemical changes cause the formation of crystals 
which by their interlocking and adhesion give strength to the 
new compounds. For a scientific and detailed treatment of 
this subject, the reader is referred to the articles of M. H. Le 
Chatelier in Annales des Mines, 11, pp. 413-465, Trans. Am. 
Inst. Mining Engineers, August, 1893; to the conclusions of 
S. B. and W. B. Newberry, Cement and Engineering News, 
1898; and to "The Constitution of Portland Cement from a 



1 S. B. Newberry, " Mineral Resources of the United States," 1892. 

79 



80 CEMENT AND CONCRETE 

Physico-Chemical Standpoint/' a paper by Mr. Clifford Richard- 
son read before the Association of Portland Cement Manufac- 
turers at Atlantic City, June 15, 1904, Engineering Record, 
August 13 and 20, 1904, Engineering News, August 11, 1904. 

97. THE RATE OF SETTING AND ITS DETERMINATION. The 
setting of cement being a gradual and continuous process with- 
out well-defined points of change, it is necessary, in order to com- 
pare the rates of change in condition of different samples, to 
adopt an arbitrary standard. The method usually adopted is 
to determine the resistance of the mortar to the penetration of a 
wire or needle. The wires used by General Totten and rec- 
ommended by General Gilmore for this purpose are now in 
general use in this country. One of the wires is T V inch in diame- 
ter and is loaded to weigh J pound; the other is ^ of an inch 
in diameter and loaded to weigh one pound. The paste is said 
to have reached "initial set" and "end of set" when these two 
wires, respectively, fail to make an impression on the surface. 

98. M. Vicat also suggested a needle test as follows: The 
cement paste is placed in a conical ring, 4 cm. in height and 7 
cm. in diameter at the base. The consistency should be such 
that a rod 1 cm. in diameter and weighing 300 grams does not 
entirely pierce the mass. This consistency having been ob- 
tained by trial, a needle of circular cross-section having an area 
of 1 sq. mm. and loaded to weigh 300 grams, is gently lowered 
on the paste. The moment when this needle no longer pene- 
trates the mass is called the beginning of the set, and the time 
in which it fails to make an impression upon it is called the end 
of setting. It may be mentioned in passing, that, according to 
a few comparative tests made by the author, when a cement 
paste has "set" by Gilmore's "heavy" wire, ^ inch weighing 
one pound, it requires about 1,100 grams weight on the Vicat 
1 sq. mm. needle to make an impression on the paste. Vicat's 
method was indorsed by the Munich Conference and was sug- 
gested in the recent progress report of the Committee of the 
American Society of Civil Engineers. 

99. M. LeChatelier has suggested a modification of this 
method by substituting for the rod 1 cm. in diameter a disc of the 
same diameter carried by a slender rod, the disc being loaded 
to weigh 50 grams, the normal consistency being such that the 
disc will stop midway in the ring, or "vase." The beginning 



SETTING OF CEMENT -81 

and end of setting he would define by the penetration of the 
needle (1 sq. mm. in section) to mid-depth in the ring, tjie 
weights being 50 grams and 3,000 grams, respectively. 

100. An approximate method of determining time of setting 
is also in use as follows: After mixing the cement paste to the 
proper consistency, place enough of it on a glass plate to form 
a thin cake, or "pat," about three inches in diameter and one- 
half inch thick at the center, thinning toward the edges. When 
the pat is sufficiently hard to bear a gentle pressure of the fin- 
ger nail, the cement is considered to have begun to set, and 
when it is not indented by a considerable pressure of the thumb 
nail, it may be said to have set. 

101. Mr. Henry Faija objected to all methods which are 
based upon the rates of acquiring hardness, on the ground that 
there are periods in the early stages of hardening that may be 
more rationally defined. He considers that the time at which 
the water leaves the surface of the pat, depriving it of its glossy 
appearance, is really the beginning of setting, and that this 
time may or may not correspond to the result obtained by the 
use of the needle. 

102. Variations in the Rate of Setting. Some of the quali- 
ties which determine the actual rate of setting of a cement 
are, its composition, degree of burning, age and fineness. Aside 
from these qualities of the cement itself, the addition of certain 
salts subsequent to the manufacture also influences the rate. 
The observed rate of setting will be influenced by the details 
of the test, such as the quantity, temperature and composition 
of the water used in gaging, the amount of gaging, the tem- 
perature of the cement, and the temperature and character of 
the medium in which the pat is placed after molding. 

103. An over-limed or highly limed cement is usually slower 
setting than an over-clayed one. Among natural cements, those 
of the aluminous variety are usually quick setting. Other 
things being equal, a well-burned Portland cement will be slower 
setting than an underburned sample. It is not certain that 
such is the case for all natural cements, though it probably is 
true of most of them. It has been said that underburned ce- 
ments owe their quick setting to their porosity, but the forma- 
tion of different compounds in the higher temperature may also 
account for the difference. 



82 



CEMENT AND CONCRETE 



104. The effect of the age of cement on its time of setting 
is very marked, but varies widely with different samples. The 
idea that cements invariably become slower setting by storage 
is a false one. The origin of this error may be found in the 
fact that by the time cement has reached its destination, it 
has usually passed through the earlier and more rapid changes 
in characteristics. Dr. Erdmenger l has stated that some Port- 
land cements become slower setting, while some set more rapidly 
as a result of storage. Dr. Tomei made experiments on several 
Portland cements 2 which show that they generally become 
quicker setting at first (from one to four months after grind- 
ing), and then become gradually slower setting, until at the 
end of a year they set in about the same length of time as 
when fresh. The writer has seen this trait exhibited very 

TABLE 18 

Time of Setting of Five Samples of Natural Cement as Affected by 

Aeration 











TIME SETTING 


TIME SETTING 








WATER. 


H 


CEMENT 


CEMENT AERATED 




i 






H 

a 


FROM PACKAGE. 


19 DAYS. 






a 


S 


. 


g 


3 


S 


6 




i 









REFER 


M 

GO 


II 


II 
s" 


cJ 

s 





S 


Diff. 

/- 


Hi 


A 


Diff. 
i-h. 


REMARKS. 






ft 




H 


























Min. 


Min. 


Min. 


Min. 


Min. 


Min. 






a 


6 


c 


d 


e 


/ 


9 


h 


i 


J 




1 


84 R 


32.0 


65 


67-73 


52 


110 


58 


54 


173 


119 


Five samples, 


2 
3 


83 R 
82 R 


(i 






60 
44 


100 
100 


50 
56 


51 
48 


164 
166 


113 

118 


same brand. 
TJ 2 and O 2 re- 
quired more 


4 


U2 


347 






60 


280 


220 


100 


326 


226 


and less wa- 


5 
6 


2 
84 R 


29.3 
40.0 






101 

87 


349 
1200 


248 
1110 


147 
130 


306 
1241 


159 
1111 


ter respect- 
ively than 
the others to 


7 


83 R 


ii 






80 


1178 


1098 


122 


1233 


1111 


make same 


8 


82 R 


it 






72 


1202 


1130 


125 


1227 


1102 


consistency. 


9 


Us 


42.7 






109 


1256 


1147 


202 


1221 


1019 




10 


2 


37.3 






192 


1247 


1045 


234 


1216 


982 





plainly by samples of Portland cement of American manufacture, 
but has not noticed it in natural cements. Table 18 gives the 
results of some tests on the effect of aeration on the time of 



1 "Notes on Concrete," by John Newman, p. 11. 

2 Trans. A. S. C. E.. Vol. xxx, p. 12. 



SETTING OF CEMENT 



83 



setting of five samples of natural cement from the same 
factory. 

105. The coarse particles in a cement retard the setting be- 
cause they are inert. Either fine grinding or sifting will doubt- 
less hasten the rate of setting, but, as has been stated above, 
the detection of changes in the rate is difficult. Table 11, 
81, gives the results of a few tests on this subject. 

106. Addition of Salts. The time of setting of a cement is 
sometimes regulated at the factory by addition of sulphate of 
lime to the finished product. Such additions are admitted to 
the extent of two per cent, by the regulations of the Asso- 
ciation of German Portland Cement Makers, and are now quite 
generally made by American Portland cement manufacturers. 
Table 19 gives the results of a few experiments on the effect 
of plaster of Paris on the time of setting of several cements. 

TABLE 19 

Effect of Plaster Paris on Time of Setting 





"^TJ 


Time to Bear i Ib. 


Time to Bear 1 Ib. 




O ^-2 


Wire, Minutes, with Plaster 


Wire, Minutes, with Plaster 


CEMENT. 


M^3 


Paris as Certain Percent- 


Paris as Certain Percent- 




llj 


age of Cement and 
Plaster Paris. 


age of Cement and 
Plaster Paris. 


Kind. 


Brand. 


|~g 


0% 


1% 


2% 


3% 


6% 


0% 


1% 


2% 


3% 


6% 


Portland 


s 


24 


232 


477 


460 


425 


40 


498 


917 


910 


860 


832 


k< 


R 


24 


95 


375 


381 


358 


75 


345 


745 


776 


778 


750 





X 


26 


4 


258 


287 


268 


84 


305 


625 


725 


668 


694 


Natural 


Gn 


34 


38 


106 


107 


86 


42 


543 


414 


527 


671 


632 


" 


An 


34 


93 


179 


302 


295 


93 


193 


439 


592 


725 


698 



It is seen that small percentages retard the initial setting in 
a marked degree, the maximum effect usually being given by 
2 per cent, of the plaster. Larger percentages tend to make 
the cement quicker setting again, so that with 6 to 10 per cent, 
added, the cement may begin to set quicker than without the 
addition of plaster. The final set (time to bear one pound 
wire) does not appear to be thus hastened by large percentages. 
This might be considered to indicate that the hastening of the 
initial set is caused by plaster of Paris taking up the water from 
the cement and obtaining sufficient hardness to bear the light 
wire. 

The probable explanation of the action of a small amount of 



84 CEMENT AND CONCRETE 

sulphate of lime in retarding the setting is that suggested by 
M. Candlot, l namely, that the aluminate of lime, to which is 
due the initial setting, dissolves less readily in a solution of 
sulphate of lime than in pure water. If the aluminate does 
not commence to hydrate until the silicate of lime has set, the 
subsequent combination of the sulphate and aluminate may 
cause the mortar to disintegrate. 

107. Solutions of common salt have been found to retard 
the setting, but when a large percentage of salt is used, it some- 
times forms a crust on the top which may resist a light wire and 
thus make the paste appear to be quicker setting. Sea water 
generally retards the setting somewhat more than solutions of 
common salt, probably on account of the magnesian salts pres- 
ent, but M. Candlot says that cements to which sulphate of 
lime has been added set more rapidly when gaged with sea water 
than when gaged with fresh water. 

The effect of calcium chloride on the setting of cements is 
entered into in detail in M. Candlot's treatise on "Cements and 
Hydraulic Limes," and may be summarized as follows: A weak 
solutioli^of calcium chloride renders Portland cement slower 
setting because the aluminate of lime dissolves more slowly in 
such a solution than in pure water. On the other hand, the 
aluminate dissolves rapidly in a concentrated solution of calcium 
chloride, and therefore such a solution hastens the setting of 
Portland cement. Aluminous cements, i.e., cements containing 
a very high percentage of alumina, are not appreciably affected 
by gaging with a comparatively weak solution of calcium chlo- 
ride on account of the large excess of aluminate of lime present; 
and on the other hand, cements containing no alumina are not 
affected, as in such cements the hardening is due to the silicate 
of lime. A weak solution of the chloride hastens the hydration 
of the free lime, and therefore a cement which contains a dan- 
gerous percentage of the latter may be made sound by gaging 
with such a solution, as the lime may thus be hydrated before 
the cement sets. The chloride of calcium test for soundness is 
based on the supposition that the free lime may be hydrated by 
the action of the chloride soon after the setting of the cement, 
and thus the expansive action be hastened. 



"Ciments et Chaux Hydrauliques," par E. Candlot. 



SETTING OF CEMENT 



85 



The effect of sugar on the time of setting does not seem to 
be well known, but it is said l that the presence of saccharine 
matter may either accelerate or retard the setting of the cement, 
depending on the amount of sugar present, the character of the 
cement and the amount of water used. 

108. The quantity of water used in gaging has a most impor- 
tant influence on the test for time of setting, an increased quan- 
tity of water retarding the setting. This may be seen from 
Table 20. 

TABLE 20 
Effect of Consistency of Mortar on the Time of Setting 





Water as per cent, of 


















cement by weight . . 


26.7 


28.6 


30.8 


33.3 


36.4 


40.0 




w a^ 


Minutes to bear T \ inch 
















a 2 


wire weighing ^ pound. 


20 


23 


30 


42 


46- 


55 




5z O 


Minutes to bear ^ inch 


















wire weighing 1 pound 


28 


41 


57 


76 


78 


85 






f Water as per cent, of 
















Q 
fc H 


cement by weight . . 


24 


26 


28 


30 


32 


34 


36 


Sl< 


Minutes to bear y^inch 
















8* 


\ pound wire . . . 


2 


2 


3 


7 


21 


28 


38 


PH 


Minutes to bear ^ inch 


















1 pound wire . 


160 


188 


279 


289 


371 


403 


583 



As might be supposed, this influence varies with different 
samples, and M. H. LeChatelier 2 has given the following table 
which illustrates this point. 

TABLE 21 
Effect of Consistency of Mortar on Time of Setting 



CEMENT. 


PER CENT. 
WATER. 


TIME SETTING, 
Minutes. 


Portland A < 


24 


20 


Portland B . . . . j 


34 
25 


85 

7 


Quick settin " Vassy . . < 


35 

50 


45 
5 




58 


10 



"Masonry Construction/' I. O. Baker, p. 98. 
2 "Tests of Hydr. Materials," p. 33. 



86 



CEMENT AND CONCRETE 



109. It is necessary, then, in writing specifications and in 
making tests, where the time of setting is at all carefully con- 
sidered, to note the consistency of the paste used in the test. 
Practically, it is preferable to use a paste rather thinner than 
that usually employed for briquets. 

The consistency is sometimes defined by M. Vicat's apparatus 
of a rod 1 cm. in diameter, or by M. LeChatelier's modification 
of the same mentioned above, or by the requirement that it 
shall be at the point of ceasing to adhere to the trowel. Another 
definition is that it shall, when placed on a glass plate, flow 
toward the edges only on repeated jarring of the plate. This 
last is a very fair approximate method, though giving a rather 
thin paste. 

That mortars set more slowly than neat cement paste is 
largely due to the increased amount of water present in the 
former, this excess of water being required to moisten the 
grains of sand. The relation between the time of setting of mor- 
tars and neat cement paste is not definite. M. Candlot found 
the time of setting of one-to-three mortars to be from two to 
twenty times as great as that of the paste of neat cement of 
normal composition. 

110. The temperature of the cement and water also has an 
important bearing on the observed time of setting. As the 
temperature of the materials is increased, the time of setting 
diminishes in about the same proportion. The following table 
gives a few of the results obtained by M. Candlot l with Port- 
land cements. 

TABLE 22 

Effect of Temperature of Materials on Time of Setting 





TEMPERATURE, 
Degrees C. 


TIME OF SETTING, 
Minutes. 




6 
15 


60 
25 


Cement No 2 ] 


25 

7 
20 


4 

350 

295 




30 


190 



Table 23 gives the results of similar tests made under the 
author's direction. The temperatures of cement and water 



"Ciments et Chaux Hydrauliques," par E. Candlot. 



SETTING OF CEMENT 



87 



were varied while the temperature of the room in which the 
tests were made remained nearly constant, or from 63 to 67 

Fahr. 

TABLE 23 

Effect of Temperatures of Cement and Water on the Time of 

Setting of Paste 



Temp, cement and water, / 


















Degrees, Fahr. . ( 


40 


50 


60 


70 


80 


90 


100 


110 


Minutes to bear j'.y ^ 


















inch wire weigh- Portland 


270 


247 


225 


196 


175 


158 


135 




ing * pound. J Natural 


102 


90 


84 


72 


60, 


54 


55 


43 



111. Amount of Gaging. If a cement paste containing a 
moderate amount of water be insufficiently gaged, it will appear 
dry, when a more thorough working might make it plastic. 
Thus an insufficient gaging may make a cement appear quicker 
setting. It is also the case that when a cement is regaged after 
having begun to set, the second setting will take place more 
slowly; this, however, is a somewhat different matter. 

112. The temperature and character of the medium in which 
the pat is kept during the setting process will have a decided 
influence on the rate of setting. 

This is clearly shown by the following table, given by M. 

TABLE 24 

Time of Setting as Affected by Temperature of the Water and 
of the Medium in which Cement Sets 



SAMPLE. 


TEMPERATUKE 


TIME REQUIKED TO 


Of water at time 
of gaging. 


Of air during 
setting. 


Begin to set. 


Set. 


Degrees C. 


Degrees C. 


lir. Min. 


Hr. Min. 


' { 



16 


1 

16 


6 47 
20 


11 
2 23 


2 I 




16 


1 

10 


5 30 
52 


8 8 
5 13 


3 I 



15 


3 
15 


12 
43 


20 
3 3 


* 1 



15 


3.5 

17 


24 
20 


1 3 
45 



88 CEMENT AND CONCRETE 

Paul Alexandre, 1 from which it appears that different samples 
are affected in very different degrees. It is seen that the 
higher the temperature, the more rapid the setting. 

113. At temperatures below 32 F. (0 C.), setting seems 
to be entirely suspended. If a cement paste, which has been 
submitted to such low temperatures since gaging, is brought 
into a warm room, the setting process begins as though the 
mortar had just been gaged. It must not be concluded, 
however, that freezing has no evil effect on mortars. (See 
Art. 50.) 

114. Setting in Air and Water. A cement paste sets much 
quicker in air than in water. This is due to the percolation of 
water to the interior of the pat, when it is immersed as soon as 
made, being analogous to using an excess of water in gaging. 
When a pat sets in dry air, the evaporation of water from the 
surface hastens the hardening of that portion. If immersed 
directly after it has set in air, it re-softens, and this is also 
true of some briquets immersed when twenty-four hours old. 
The time of setting of cements that are to be deposited under 
water may well be tested in that medium, when they should 
be protected by a mold of some form to retain their shape. 
Ordinarily the time of setting should be tested in moist air. 

Cements are said to set more quickly in compressed air than 
in free air; this may be partially due to the higher temperature 
usually existing in the former. 

115. Requirements as to Time of Setting. What is desir- 
able as to time of setting will, of course, depend on the work 
in hand; certain purposes requiring that the cement shall be 
able to retain its shape soon after deposition, while in other 
cases ability to mix large quantities at a time, without fear of 
the cement setting before it is in place, may be very convenient. 
An extremely quick setting cement should be regarded with 
suspicion until it has proved itself of good quality. It is some- 
times stated that where a quick setting mortar is desired, nat- 
ural cement must be used, but this is not true; either Portland 
or natural may be found with almost any rate of setting de- 
sired. As a general rule, however, among cements that have 
been stored several months, the Portlands are slower setting. 



1 "Recherches Experimentales sur les Mortiers Hydrauliques," par. Paul 
Alexandre. 



SETTING OF CEMENT 89 

Portland cement will ordinarily begin to set in from twenty 
minutes to six hours, and natural cement in from ten minutes to 
two hours, though there are many cements the time of setting of 
which is outside of these limits. 

116. Conclusions. The purpose aimed at in the test for 
time of setting will, to a certain extent, regulate the method 
to be employed. The pressure of the ringer nail will be suf- 
ficient to determine (after a little experience) whether a cement 
will answer a certain purpose in this regard. But, if one is 
working to rigid specifications, or pursuing investigations as to 
the effect of different treatment on time of setting, it becomes 
very desirable to have a method of determining and defining 
the consistency of the mortar, and an accurate method of de- 
termining the rate of setting. 

In the author's experience, the Vicat consistency apparatus 
as modified by M. LeChatelier (see 99) has proved unsatisfac- 
tory except for thin pastes of neat cement or mortars contain- 
ing less than two parts of sand. If the paste is not of such a 
consistency as to run freely into the ring, or "vase," an error 
may be introduced in the method of filling the latter. In oper- 
ating with a natural cement it was found that a neat paste, in 
which the water used was 32 per cent, of the dry cement, re- 
quired a gross weight of 640 grams to make the disc (1cm. diam- 
eter) penetrate midway in the vase; with 33 per cent, water, 
a weight of 410 grams was required; 34 per cent., about 250 
grams; 35 per cent., 175 grams; 37 per cent., 155 grams. It 
would seem that some modification of this apparatus might be 
made which would not only indicate when a thin, neat cement 
paste has the assumed " normal" consistency, but which would 
also define the consistency of a given mortar, whether of neat 
cement or of sand mixture. 

General Gilmore's wires are very simple, and will perhaps 
answer the purpose of obtaining the time of setting as well as 
any method in use. They can be used somewhat more accu- 
rately if the wires are made to slide vertically in a frame, than 
when held in the hand. 

The necessity of care in all of the details of this test, tem- 
perature and amount of water, amount of gaging, character 
of medium, etc., has been sufficiently emphasized in the preced- 
ing paragraphs. 



90 CEMENT AND CONCRETE 

ART. 18. CONSTANCY OF VOLUME 

117. That a cement should not contain within itself ele- 
ments which may lead to its destruction, is evidently a most 
important quality. It is probable that nearly all cements un- 
dergo a slight change in volume during induration, contracting 
in air and expanding in water. But it is the detection of those 
larger changes, which result from bad proportions or defective 
manufacture, and which cause deterioration or even complete 
disintegration, that is the object of the tests for soundness. 

118. Causes of Unsoundness. The most frequent cause of 
unsoundness is considered to be the presence of free lime or 
magnesia. (See 49 and 50.) Any one of the following causes 
may account for the presence of free lime in cement: (1) An 
excessive percentage of lime may have been used in proportion- 
ing the raw materials; (2) the raw materials may not have been 
sufficiently mixed to render the mass homogeneous; (3) hard 
particles of lime, such as shells, may not have been ground fine 
enough in making the mix to permit them to enter into com- 
bination with the other ingredients during burning; or (4) the 
cement may have been underburned, so that part of the lime 
did not enter into combination. 

The particles of free lime which occur in cements are nat- 
urally rather difficult to slake on account of their impurity and 
the high temperature at which they have been calcined, and 
the same thing is probably true of magnesia. It may thus 
require weeks or months of exposure to the atmosphere to cor- 
rect tendencies to expand due to the presence of free lime or 
magnesia. Likewise when such defective cements are immersed 
in water of ordinary temperature, the expansion may not occur 
for a considerable period. This fact has led to the use of hot 
tests of various kinds to detect such faults, but before touch- 
ing on these so-called " accelerated tests/ 7 the ordinary cold- 
water test will be described. 

119. TESTS FOR SOUNDNESS. The Committee of the Amer- 
ican Society of Civil Engineers on a " Uniform System for 
Tests of Cement" recommended, in 1885, the following test for 
soundness: "Make two cakes of neat cement two or three inches 
in diameter, about one-half inch thick, with thin edges. One 
of these cakes, when hard enough, should be put in water and 



CONSTANCY OF VOLUME 91 

examined from day to day to see if it becomes contorted, or if 
cracks show themselves at the edges, such contortions or cracks 
indicating that the cement is unfit for use at that time. In 
some cases the tendency to crack, if caused by the presence of 
too much unslaked lime, will disappear with age. The re- 
maining cake should be kept in air and its color observed, which 
for a good cement should be uniform throughout, yellowish 
blotches indicating a poor quality; the Portland cements being 
of a bluish-gray, and the natural cements being light or; dark, 
according to the character of the rock of which they are made." 
For the ordinary cold test this method will probably give as 
valuable results as any of the forms that are suggested. 

120. The German regulations require a very similar test, 
except that in the case of slow setting cements the pat is not 
immersed until twenty-four hours old. While a cement that is 
decidedly bad may show its defects in from one day to one week 
by this cold water test, it may be the case that cracks will ap- 
pear only after several months' immersion. It has therefore 
been proposed to hasten the destructive action of the free 
lime or magnesia by submitting the cakes of cement to steam, 
hot water, or dry heat. 

121. The Kiln Test, recommended by Prof. Tetmajer in 1890, 
consists in placing in an air bath, pats which have been kept 
in moist air for twenty-four hours; and then gradually raising 
the temperature of the air bath to 120C. This temperature 
is maintained for at least one-half hour after the disengage- 
ment of steam has ceased. The pats should show no tendency 
to expand under this treatment, but if cements fail to pass the 
test, the results of the ordinary cold water treatment are to be 
awaited. This test is intended for cements that are to be 
used in air. 

122. The Boiling Test, which was also recommended by Prof. 
Tetmajer, consists in placing the pats, twenty-four hours after 
made, in water of ordinary temperature, and gradually heating 
the water to bring it to the boiling point in about an hour; 
five or six hours in the boiling water should develop no defects. 
This is a severe test, and has been objected to on the ground 
that cements which have been well proportioned, but which 
are a trifle underburned, will fail to pass this test while giving 
good results in mortars to be used in the air. This test, how- 



92 CEMENT AND CONCRETE 

ever, is steadily gaining in favor, and is used in many cement 
works as a test of quality. 

123. The Warm Water Test. Mr. H. Faija was an early 
experimenter in accelerated tests for soundness, and about 1882 
he began the use of a "steamer," using a temperature of about 
110 Fahr. After eleven years' use he still believed this tem- 
perature to be high enough to detect tendencies to expand in 
faulty cements. The apparatus 1 " consists of two vessels, one 
within the other, a water space being thus maintained between 
them, which assists in equalizing the temperature of the inner 
or working vessel." The latter is partially filled with water 
and is provided with a rack or shelf near the top. A ther- 
mometer is inserted through the cover of the inner vessel, and 
the water within is kept constantly at 110 Fahr. As soon as 
the pat is gaged, it is placed on the rack in the vapor, which 
will be at about 100 Fahr. After six or seven hours in this 
moist heat, the pat is immersed in the warm water. "In the 
course of twenty-four hours it is taken out and examined, and if 
then found to be quite hard and firmly attached to the glass, the 
cement may at once be pronounced sound and perfectly safe 
to use; if, however, the pat has come off the glass and shows 
cracks or friability on the edges, or is much curved on the 
under side, it may at once be decided that the cement in its 
present condition is not fit for use." Mr. Faija also recom- 
mended, in case of failure in the first test, that the cement be 
spread out in a thin layer for a few days and a second test 
made. If the cement passes this second test, it is pronounced 
sound and fit for use after being stored a sufficient length of 
time. 

124. The Hot Water Test. The temperature to be used in 
accelerated tests for soundness is a point which has received 
much attention and is still under discussion. In 1890 M. Deval 
described a series of experiments he had made, in which he 
employed a temperature of 80 C. While this is much more 
severe than the temperature used by Mr. Faija, it is still mild 
in comparison to some temperatures that have been advocated. 

125. Mr. W. W. Maclay, who was probably the first engi- 
neer in this country to introduce a hot test requirement in 

1 "Portland Cement Testing," by H. Faija, Trans. A. S. C. E., Vol. xvii, 
p. 222. 



CONSTANCY OF VOLUME 93 

specifications, gave the results of his experiments in a paper 
presented to the American Society of Civil Engineers in 1892. 
The method used " consists in molding six pats of pure cement 
and water, about one-half inch thick and about three inches in 
diameter, on thin glass plates, and of the same consistency as 
for the briquets for tensile strength." The treatment to which 
these pats are submitted is as follows: 

No. 1, in steam (vapor) bath, temperature 195 to 200 F., 
as soon as made. 

No. 2, in same vapor bath when set hard (bear ^ inch wire 
weighing one pound). 

No. 3, ditto, after twice the length of time in air allowed the 
second pat. 

No. 4, ditto, after 24 hours. 

No. 5, in water of temperature about 60 F. when set hard. 

No. 6, kept in moist air at temperature of about 60 F. 

"The first four pats are each kept in the steam bath three 
hours, then immersed in water of a temperature of about 200 
Fahr. for twenty-one hours each, when they are taken out and 
examined. To pass this test perfectly, all four pats, after being 
twenty-one hours in hot water, should, upon examination, show 
no swelling, cracks, nor distortions, and should adhere to the 
glass plates. The latter requirement, while it obtains with 
some cements nearly free from uncombined lime, is not insisted 
upon; the cracking, swelling and distortion of the pats being 
much the more important features of this test. The cracking 
or swelling of No. 1 pat alone can generally be disregarded." 

126. DevaPs Method. Making tests of mortar briquets, 
which have been kept in hot water, seems, to be the most rational 
accelerated test for soundness. This method was used in Ger- 
many several years ago, when it was claimed that a definite 
relation existed between the results thus obtained and the longer 
time cold water tests. This theory being disproved, threw dis- 
credit on the hot test, but M. Deval l has since made many 
experiments showing that it is of much value in detecting 
bad products. 

The method consists in making briquets with three parts 
sand to one of cement, and after twenty-four to seventy-two 



1 "Hot Tests for Hydraulic Cements," M. Deval, Bull. Soc. d' Encourage- 
ment, etc., 1890, pp. 560-583. 



94 



CEMENT AND CONCRETE 



hours in moist air, according to the rate of setting, immersing 
them in water maintained at 80 C., the briquets being broken 
after an immersion of from two to seven days. These hot 
water briquets are to be compared with briquets stored in 
water of the ordinary temperature and broken at seven and 
twenty-eight days after immersion. 

127. Among other tests M. Deval compared the results ob- 
tained with six samples of Portland cement as follows: 
No. 1. Good finely ground cement of modern make. 
No. 2. Coarsely ground cement of good quality, but partially 

aerated. 
No. 3. Quick setting cement with low per cent, lime and 

lighter burn. 

No. 4. Made from clinker having property of disintegrating 
spontaneously while cooling; large proportion of inert 
material. 

No. 5. Under-burnt cement; contains free lime. 
No. 6. Over-limed cement. 
The results of the tests are given in the following table: 



TABLE 25 

Cold and Hot Tests on Six Samples of Portland Cement 
(M. Deval) 





TENSILE STRENGTH IN KILOS PER SQ. CM. 


CEMENT. 


Cold. 


Hot. 




7 days. 


28 days. 


2 days. 


7 days. 


1 


15.0 


23.3 


17.2 


24.3 


2 


6.7 


13.7 


7.6 


11.0 


3 


6.2 


16.5 


7.3 


16.2 


4 


2.9 


3.9 


) 


6 


6.1 


12.2 


Disintegrated. 


6 


7.6 


20.2 


) 



No. 4, when allowed forty-eight hours to set, gave 3.2 kilos 
at two days, and 4.3 kilos at seven days, when tested hot. 
Among the cements which disintegrated in the hot water, the 
only one that gave a high result cold was No. 6, and this sam- 
ple, it is stated, would crack and swell badly even in cold water 



CONSTANCY OF VOLUME 95 

if mixed neat. It is quite possible, however, that a sample 
might be found which, not having quite as flagrant defects as 
No. 6, would pass all the cold tests but be condemned by the 
hot test. 

128. The conclusions drawn from these experiments have 
been stated as follows : 

"(1) Tests made cold do not indicate the quality of the 
cement, inasmuch as cement containing excess of lime, and, in 
consequence, deplorably bad, may give excellent results." 

"(2) Portland cement of good quality, mixed with normal 
sand in the proportion of one to three, resists water at 80 C. 
Its strength at two and seven days after setting is about equal 
to that which it would have at seven and twenty-eight days 
in the cold." 

"(3) Poor cement containing much inert material does not 
resist the action of water at 80 C. unless the setting be allowed 
to proceed for some days before immersion." 

"(4) Cements containing free lime do not withstand the ac- 
tion of water at 80 C. if immersed twenty-four hours after 
setting." Comparison of the strength hot and cold will suffice 
for the detection of even small quantities of free lime. 

129. Before passing to the comparison of the tests for sound- 
ness already outlined, a few other tests which have been sug- 
gested for use may be briefly mentioned. 

The Chloride of Calcium Test depends on the fact that slak- 
ing of free lime is hastened by feeble solution of chloride of 
calcium. (See 107.) Concerning this test, Prof. F. P. Spald- 
ing 1 says he "has found it to give true indications in a number 
of cases, including some unsound magnesian cements. It con- 
sists in mixing the mortar for the cakes with a solution of 40 
grammes chloride of calcium to one liter of water, allowing 
them to set, immersing them in the same solution for twenty- 
four hours, and then examining them for checking and soften- 
ing as in other tests." 

130. M. H. LeChatelier's Method. The method recom- 
mended by M. H. LeChatelier for testing soundness requires 
the use of a cylindrical mold, about 1J inches in diameter and 
of about the same height, which is made of thin metal and 

1 "Notes on the Testing and Use of Hydraulic Cement," by Fred. P. 
Spalding. 



96 CEMENT AND CONCRETE 

slit along a generatrix. The mortar is to be placed in the 
mold as soon as made, and immersed at once in cold water; 
the mold is held firmly by a clamp, and a flat plate at either 
end of the mold retains the mortar in shape until set. When 
setting has taken place, the mold is undamped and the widen- 
ing of the slit indicates the expansion of the mortar. If de- 
sired, the swelling may be increased and hastened by transfer- 
ring the mold and its contents to hot water as soon as the ce- 
ment is set. The same writer has suggested a modification of 
the hot test by placing briquets in cold water and gradually 
heating to near the boiling point, this temperature being main- 
tained for six hours. 

Various other tests have been suggested, such as the effect 
of regaging; withstanding immersion as soon as gaged; allow- 
ing large thin cakes to harden in air and striking them to obtain 
a musical sound. Most of these tests, however, are worthy of 
passing notice only. 

131. Discussion. There are but few experiments to show 
that a cement which will actually fail and disintegrate when 
properly used, may still pass the cold water neat pat test; yet 
there is no doubt that inferior cements may pass this test per- 
fectly, " inferior cements" being those which will not give the 
best results in practice, though they do not disintegrate. 

Cement is at present used in a very crude way, and it is only 
in exceptional cases that a poor quality of material may be 
detected in the completed structure. This is sufficient reason 
why so few failures can be found in cement work which may 
be attributed to a poor quality of cement. But in the more 
economical manner in which this material is, even now, begin- 
ning to be used, it is absolutely essential to know what its fu- 
ture behavior will be. That the cement will never be exposed 
to hot water in actual use, is a weak argument against hot 
water tests. It must be remembered that the chief object of 
testing cement is to arrange the various products in their true 
order of merit, and any system which will effect this result is 
perfectly legitimate. On the other hand, it is due to the man- 
ufacturers that a test which will occasionally reject perfect 
cements should not be adopted when it is possible in any other 
way to detect poor products with certainty. 

132. It is possible that the temperature used and recom- 



CONSTANCY OF VOLUME 97 

mended by Mr. Faija is sufficiently high to detect unsoundness 
or a tendency to "blow/ 7 It has never been clearly proved 
that it is not, but the higher temperature of 70 to 100 C. has 
appeared to meet with greater favor. The writer made a few 
experiments to compare results obtained with mixtures of 
Portland cement and lime when using the temperature of 110 
Fahr. (43 C.) with those obtained in water at 190 Fahr. (88 
C.), and in water at the ordinary temperature of 60 to 65 Fahr. 
(16 to 18 C.). Quicklime, in proportions varying from one 
to ten per cent., was added to the cement, and seven pats were 
made from each mixture of cement and lime. 

These pats were subjected to the following treatment: 

Pat No. 1, placed in vapor of water at 110 F. when made. 
" 2, " " " 110 F. when set. 

" 3, " " " 110 F. after 24 hours. 

4, " " " 190 F. when made. 

" 5, " " " 190 F. when set. 

" 6, " " " 190 F. after 24 hours. 

Above six pats immersed in the hot water after three hours in 
vapor. 

Pat No. 7, placed in cool water when set. 

When no lime was added, pats 1, 2 and 3 revealed no defects; 
pats 4 and 5 showed small cracks in two days, but pat No. 6 
still adhered to the glass after eight days. Pat No. 7 was perfect 
after two months. With 2 per cent, lime added to the cement, 
pat No. 1 was slightly warped and cracked, and Nos. 2 and 3 
were off glass; Nos. 4 and 5 were cracked and warped; No. 6 
was off glass, and No. 7 became detached from glass after two 
months, but was otherwise perfect. With 4 per cent, lime, all 
the pats failed, the one in cool water being off glass, cracked 
and warped after one day. 

It must be remembered that the free lime occurring in cement 
is of a different character from the quicklime added in these 
tests, because the former contains impurities and has been cal- 
cined at a very high temperature, and would therefore slake 
more slowly. It has been said that as small an amount as 1 
per cent of free lirne in cement is dangerous. If this is true, 
and it probably is, the temperature of 110 Fahr. would seem 
to be inadequate to quickly indicate a tendency to "blow." 



98 



CEMENT AND CONCRETE 



133. Some of the results obtained by M. Deval have already 
been given ( 127). Mr. Maclay made similar tests on several 
samples of Portland cement, using a temperature of 200 Fahr., 
but these tests only permit of comparing the strength acquired 
in cold water in seven and twenty-eight days with the strength 
in hot water at ages of from two to seven days. Long time 
tests, showing that the cements which give low results in hot 
water and normal results in cold water on short time tests, 
give in reality a low strength at the end of six months or more, 
have been almost entirely lacking until very recently. 

Table 40, 226, gives some of the results obtained by the 
author in hot tests and long time cold tests on Portland cement. 
It is seen that the hot test at 80 C. indicated, in seven days, 

TABLE 26 
Cold and Hot Tests on Samples of One Brand of Portland Cement 



CEMENT. 


PARTS. 

SAND 


DATE 
MADE. 
1894. 


AGE. 


TENSILE 
STRENGTH. 


BRIQ 


UETS STORED. 






Mo. Da. 






Moist air. 


Water. 


B' 


2 


4 16 


5 da. 


8 


1 da. 


80 C. 4 da. 


A 


2 


7 2 


5 da. 


235 




4 


B' 


2 


4 16 


7 da. 


13 




* 6 


A 


2 


7 2 


7 da. 


229 




6 


B 


8 


7 2 


7 da. 


197 




15 to 18 C. 6 da. 


A 


3 


7 2 


7 da. 


108 




6 


B 


3 


7 2 


28 da. 


298 




27 


A 


3 


7 2 


28 da. 


198 




27 


B' 


2 


4 16 


7 mo. 


411 




' 7 mo. 


A 


2 


7 2 


6 mo. 


465 




6 " 



BEHAVIOR OF PATS MADE JULY 2, 1894 

No. 1 in vapor, when held \i wire. ) Immersed in water 80 C. after three 
No. 2 in vapor, when held 1# wire, f hours in vapor. 
No. 3 in tank, when held 1# wire. 
No. 4 in tank ; two hours after held 1# wire. 

Cement: A, No. 1 off glass in two days; No. 2 warped some in two days. 
" A, No. 3 O.K. after twenty-one days; off glass and warped in 

fifty-two days. 
" A, No. 4 loose on glass in twenty-one days; off glass and warped 

in fifty-two days. 

" B, No. 1 off glass and warped some in two days; No. 2 entirely 
disintegrated in two days; No. 3 loose on glass in twenty- 
one days; off glass and warped in fifty-two days; No. 4 loose 
on glass in twenty-one days; off glass and warped in fifty-two 
days. 



CONSTANCY OF VOLUME 99 

the inferior quality of sample W, although it gave normal re- 
sults in cold water up to twenty-eight days; the two year tests 
with mortars containing two parts or more sand, show it to be 
inferior. If we attempt to carry the analogy too far, however, 
we fall into the error which placed the hot test in disrepute for 
several years, that is, we must not expect that the strength in 
cold water after a long time will be exactly proportional to the 
strength developed in hot water in a few days. 

134. In Table 26 are given the results of tests by the author, 
on samples of a single brand of Portland cement. The por- 
tion marked "A" had been spread out in open air for seventy- 
seven days in a thin layer. The portion marked "B" was 
taken directly from the barrel July 2d, and B ' was taken 
from the same barrel April 16th. Samples B and B' are not 
identical, because the cement had undergone some change, 
though stored in the barrel. Each result is the mean of five 
briquets. 

In the short time cold tests there was nothing to indicate that 
the cement directly from the barrel was not good, except the 
very small evidence in the fact that pat No. 3 was loose on glass 
plate after twenty-one days. In fact, the cold water briquet 
tests at seven and twenty-eight days unmistakably declare in 
favor of the sample B. On the other hand, how sharply did 
the hot tests bring out the defects, two days in hot water being 
sufficient to entirely disintegrate one of the pats. Although 
sample B' showed a considerable tensile strength at seven 
months with two parts sand, yet the pats of neat cement failed, 
even in cold water, after two months, altogether too late a date 
to be of any value in preventing the use of the cement. 

13. In a paper read before the American Society for Testing 
Materials, July, 1903/ Mr. W. P. Taylor of the City Testing 
Laboratory, Philadelphia, gives some very interesting data con- 
cerning the behavior of cements that failed to pass the boiling 
test. The method employed was to make cakes of cement in 
the form of a small egg, keep them in moist air for twenty-four 
hours, then place them in cold water which is gradually raised 
to the boiling point and maintained at that temperature for 
three hours. The results cited show that some unsound ce- 



Proceedings Amer. Soc. for Testing Materials, 1903. 



100 CEMENT AND CONCRETE 

ments may be much improved by sifting out the coarse parti- 
cles, and that a cement failing in the boiling test when fresh 
may pass it satisfactorily after four or five weeks. 

Examination of the results showed that 96 per cent, of a 
large number of specimens which did not pass the hot water 
test failed within three hours, and 99 per cent, in four hours. 
This fixes a practical limit to the time necessary to continue 
the test. Some very valuable tests are cited to show the 
ultimate failure in cold water of samples that failed in the 
hot tests. Ten cements which passed the cold water pat test 
of twenty-eight days 7 duration, but which failed in the boiling 
test above described, gave normal results in one-to-three mor- 
tars at twenty-eight days, showing a tensile strength of 217 to 
252 pounds per square inch, but gave only 47 to 147 Ibs. per 
square inch at four months. 

Another valuable comparison is given by Mr. Taylor: A 
compilation of data, covering over a thousand tests on many 
varieties of cements, showed that "of those samples that failed 
in the boiling test but remained sound at twenty-eight days (in 
cold water), 3 per cent, of the normal pats showed checking or 
abnormal curvature in two months; 7 per cent, in three months; 
10 per cent, in four months; 26 per cent, in six months and 48 
per cent, in one year; and of these same samples, 37 per cent, 
showed a falling off in tensile strength in two months; 39 per 
cent, in three months; 52 per cent, in four months; 63 per cent, 
in six months and 71 per cent, in one year." 

136. It may be of interest to introduce here some of the 
opinions that have been expressed concerning hot tests. M. 
Candlot l says that cements of normal composition, the burning 
of which has not been carried to the point of vitrification, would 
be condemned by the hot test of neat cement, although mor- 
tars made with them show no signs of alteration in sea water, 
and, when preserved in air, give entirely satisfactory results. 
Referring to the tests of one-to-three mortar briquets in water 
at 80 C., he considers that "cements containing free lime give 
in hot water, lower resistances than in cold water; cements of 
good quality give resistances at least equal and nearly always 
greater in hot water than in cold. Cements well proportioned 



1 "Ciments et Chaux Hydrauliques," par M. Candlot, pp. 144-145. 



CONSTANCY OF 



and homogeneous, but not having obtained the maximum burn- 
ing, give satisfactory results with this test." 

In using the slit cylinders mentioned in 130, M. H. Le 
Chatelier found 1 that the addition of 5 per cent, of lime could 
be detected by cold tests in a few hours, while 5 per cent, of 
magnesia could not be detected in twenty-eight days. The 
cement containing 5 per cent, lime disintegrated almost at 
once in hot water, while the sample to which 5 per cent, of mag- 
nesia had been added, swelled considerably in one day. 

Mr. A. Marichal 2 found that "the percentage of water en- 
tered in combination, after ten days in hot water, was the same 
as for six months in cold water, and that the strength of the 
cement was increasing with the amount of water entered in 
combination. It was discovered incidentally, that cement con- 
taining over 5 per cent, of magnesia, or 3 per cent, of uncom- 
bined lime, would not stand the boiling test." 

137. Hot Tests for Natural Cements. All that has pre- 
ceded concerning hot tests refers to their use for testing Port- 
land cements. Very little is known concerning the value of hot 
tests for natural cements. There are comparatively few natural 
cements that are absolutely bad, but to distinguish between the 
first and second quality of this variety of products is much more 
difficult than to make a similar distinction with Portlands. One 
point is certain, natural cements must not be expected to with- 
stand boiling water. Mr. de Smedt experimented with fifteen 
brands of natural cement, and found that thirteen of them 
went to pieces in boiling water in two hours, although none of 
them was thought to contain caustic lime. Prof. Tetmajer 
has stated that for Roman cements, boiling water, or even 75 C., 
is not at all conclusive, and recommends 50 C. for trial, but our 
natural cements are not strictly comparable with Roman ce- 
ments. 

138. The author has experimented with three temperatures, 
namely, 50, 60, and 80 C., and is inclined to consider that 
80 C. is likely to give the most useful information for sand 
mortar briquets but not for neat cement pastes. Table 41, 
227, gives the results of hot briquet tests on six brands of 



" Tests Hydr. Materials," by H. LeChatelier. 
3 Trans. Amer. Soc. C. E., Vol. xxvii, p. 438. 



O O 



102 CEMENT AND CONCRETE 

natural cement. It is seen that, with two parts sand, brands 
Jn, Hn, and Bn, give very low results at 80 C., and these brands 
are really inferior cements as shown by the two-year cold tests. 
Brand Jn is the only one that gave a lower result at seven days 
than at five days when tested at 80 C., and this brand failed 
entirely at two years, though it gives normal results in cold 
water up to six months. Neat cement pats of this brand, after 
being stored in cold water for nearly one year, were found to be 
cracked, although they had been perfect after one month in 
cold water. It was also found that neat cement pats of this 
brand warped and cracked in two days when placed in water of 
60 C. when set. 

139. CONCLUSIONS. It may be said that although the 
limits within which the hot tests are reliable have not been well 
established, and although a strict adherence to them may at 
times reject a usable product, yet it is believed that sufficient 
experiments have been made to indicate that they are of much 
value, and should be made in all cases where the quality of the 
cement is of high importance. 

The present indications seem to be that Portland cements 
may well be tested in the form of neat cement pats and sand 
mortar briquets at a temperature of about 80 C. Natural ce- 
ments in the form of neat paste should not be called upon to 
resist a temperature above 60 C., but 80 C. will probably give 
the most useful information with sand mortars. In either case, 
the mortar should be allowed to set in moist air of ordinary 
temperature, then transferred to the vapor, to remain two or 
three hours before immersion in the hot water. It is not rec- 
ommended that these hot tests should replace the ordinary 
cold tests, but simply that in cases where the extra work in- 
volved is not prohibitive, the hot tests should be made in con- 
nection with the cold tests. 



CHAPTER VIII 

TESTS OF THE STRENGTH OF CEMENT IN COMPRES- 
SION, ADHESION, ETC. 

140. IN testing the strength of cement the object is three- 
fold : 1st, to obtain an idea of the strength that may be ex- 
pected from the cement as used in the structure; 2d, to obtain a 
basis for comparing the value of different cements in this regard; 
and 3d, to determine the ability of the cement to withstand 
destructive agencies, whether these agencies be due to exterior 
causes or emanate from the character of the cement itself. To 
illustrate the last point it is only necessary to mention such de- 
stroying agents as free lime (interior) and frost (exterior). It is 
evident that the stronger the cement the more effectually will 
these agencies be resisted. 

The strength of cement may be tested by compression, 
shearing, bending, adhesion, abrasion and tension. The tensile 
test is the one most frequently used, but the tests will be con- 
sidered in the order named. 

ART. 19. TESTS IN COMPRESSION AND SHEARING 

141. Value of Test. In practically all forms of masonry 
construction, cement is called upon to resist compression. In 
consequence of this fact, the opinion is somewhat general that 
the greatest amount of information would be obtained by com- 
pressive tests. But the compressive strength of cement is so 
much greater than its tensile strength, that when failures occur, 
they are likely to be due to other forms of stress. In short, the 
ratio of the compressive strength to the crushing force it is 
likely to be called upon to resist, is usually much greater than 
the corresponding ratio in tensile strength. 

142. There is no doubt that compressive tests are of much 
interest and value, especially so since the use of concrete and 
steel in combination has become general, but as yet the facili- 
ties for making, the test are not available without considerable 
expense. This is on account of the larger force required (the 

103 



104 CEMENT AND CONCRETE 

compressive strength being six to ten times the tensile) and be- 
cause the uniform distribution of the stress over the surface of 
the specimen, and the accurate recording of the force exerted, 
are even more difficult than the corresponding operations in 
tensile tests. Prof. Sondericker, 1 in a paper read before the 
Boston Society of Civil Engineers, describes an apparatus in 
which he seems to have overcome a part of these difficulties. 

A convenient specimen for compressive tests is a cube meas- 
uring two inches on a side. The specimens are prepared and 
treated in the same way as briquets for tensile tests. Before 
testing, two opposite faces of the cubes are usually ground so as 
to be true planes, parallel to each other, or the opposite sides 
may be faced with plaster of Paris, though this is not recom- 
mended. Grinding two surfaces to true planes increases very 
much the work involved in testing, so that several tensile tests 
may be made in the time required to make one compressive test. 

143. Conclusions. Although tests of compressive strength 
are of interest from a scientific point of view, it is not considered 
that they would give much greater information concerning the 
relative qualities of cements than is given by tensile tests, and 
therefore they need not be included in an ordinary series of 
acceptance tests. 

144. Tests of Shearing Strength. Although cement is fre- 
quently called upon to withstand a shearing stress, tests of this 
kind are very seldom made. Some of the difficulties encoun- 
tered in compressive tests are also present in tests of shearing. 
Prof. Cecil B. Smith made quite an extended series of shearing 
tests by cementing together three bricks, the middle one pro- 
jecting above the other two, and the pressure being so applied 
as to avoid any transverse stress. It is evident that by this 
method the adhesive strength is also brought into play. Shear- 
ing tests need not be included in normal tests of quality. 

ART. 20. TESTS OF TRANSVERSE STRENGTH 

145. It is probable that the earliest rupture ^ests of cement 
were made by submitting rectangular prisms to a bending 
stress; but such tests have long held a place subordinate to 
trials of tensile strength. A mass of masonry, taken as a whole, 
is very apt to be subjected to a bending stress, but it is a ques- 



1 Jour. Assoc. Engr. Soc., Vol. vii, p. 212. 



TRANSVERSE TESTS 105 



tion whether a transverse test on a small specimen gives any 
better idea of the ability of a large beam to carry its load, than 
do simple tensile and compressive tests. 

In Engineering News of December 14, 1893, appeared an 
article giving the comparative results obtained in tensile and 
transverse tests. The tensile specimens had an area of one 
square inch at the smallest place, and the transverse specimens 
also had an area of cross-section of one square inch. It was 
found that the modulus of rupture computed by the common 

formula/ = , , 2 was from 1.1 to 3.8 times the tensile strength 

developed by the briquets. Some comparative tests made at 
St. Mary's Falls Canal are discussed in Art. 56. 

146. The objections to transverse tests are: 1st, if the speci- 
mens are made but one inch in cross-section, it is difficult to 
handle them without injuring them, and if the section is made 
much larger than one inch square, a much greater amount of 
cement is required to make the specimens and more room re- 
quired to store them; 2d, it would seem that the results ob- 
tained might be less trustworthy than those in tensile tests 
because of the greater influence of the outside layers, which are 
subjected to the greatest accidental variations, on the apparent 
strength of the specimen. On the other hand, it may be said 
that, when no testing machine is at hand, the apparatus requi- 
site to make a crude test may easily be improvised. All that is 
required is a rectangular wooden mold, three knife edges, and a 
pail with a quantity of sand or water. 

147. When it is a question of making tests of transverse 
strength accurately and rapidly, the apparatus required is no 
more simple than the apparatus for tensile tests. In the con- 
struction of metal molds in large quantities it makes little dif- 
ference whether the form requires curved or straight lines. As 
far as breaking is concerned, there is a certain force to be applied, 
and a machine that will answer for one test may also be used 
for the other. In the matter of clips, there may be a slight 
advantage as to simplicity in a clip designed for transverse 
breaking. 

In making transverse tests the author has used a form two 
inches square and eight inches long. By placing the end sup- 
ports five and one-third inches apart, the modulus of rupture 



106 CEMENT AND CONCRETE 

3 Wl 
by the formula / = , , 2 becomes equal to W, the center load 

applied. 

148. Finally, it may be said that there is little objection to 
substituting transverse tests for tensile tests, although no evi- 
dent advantage would be gained. It would also seem that 
there is no object in making tests for quality by both trans- 
verse and tensile tests, though from a scientific standpoint 
comparative tests of transverse and tensile strength are of great 
interest. 

ART. 21. TESTS OF ADHESION AND ABRASION 

149. ADHESION. The test for adhesion is also one of long 
standing, being used during that time when engineers were con- 
tent with an approximate idea of what might be expected of an 
hydraulic product. It has been stated above that when failure 
occurs in a mass of masonry, it is more frequently a failure in 
tension than in compression; it may be added, that it is also 
more likely to fail in adhesion than in cohesion. Hence, an 
adhesive test is a very proper one to make, and will give most 
valuable results. In fact, it is perhaps the most rational rup- 
ture test, and were it not for the difficulties involved in its ap- 
plication, it would doubtless come into general use. 

150. One of the greatest difficulties experienced in making 
adhesive tests is the preparation of the specimens of stone or 
other material to which the mortar is to adhere. In early ex- 
periments common brick were used, or pieces of stone were cut 
to the same shape as brick, and two or more pieces cemented 
together. In later methods the flat surfaces of two specimens 
are sometimes joined with their axes at right angles, thus mak- 
ing the cemented surface square. The upper brick being held 
on two supports, a load is applied to the lower brick. 

151. Mr. I. J. Mann, in a paper presented to the Institution 
of Civil Engineers, 1 described a method of testing adhesion in 
which are used test pieces 1 inches long by 1 inch wide by \ to 
f inch thick. These are cemented together in a cruciform shape, 
and a simple spring balance machine with properly arranged 
levers pulls them apart. The upper block is supported at its 
ends and an inverted U-shaped piece bears upon the ends of 

1 Proc. Inst. C. E., Vol. Ixxi, p. 251. 



TRANSVERSE TESTS 107 

the lower block. The stress is applied through a conical shaped 
pivot bearing on the U-shaped saddle. Mr. Mann states that 
test pieces may be made either of plate glass or close grained 
limestone, the latter being sawn into pieces of the right size. 

152. Another method is to make test pieces to fit one end 
of the mold used for tensile tests, and after placing the piece of 
stone in the mold, to fill the other end with the mortar to be 
tested. The objection to this method is the expense of pre- 
paring pieces of this form. It has been suggested to substitute 
artificial stone for the cut stone samples. Thus, suppose it is 
required to test the adhesion of a certain mortar to granite: 
mold half briquets of a mixture of ground granite with cement, 
and after these have well hardened, replace them in the mold 
and fill the other end of the mold with the mortar to be used. 
It is quite certain that the same result would not be obtained 
in this way as though the specimens were cut from a piece of 
solid granite. 

153. One of the simplest methods of applying this test is 
one which the author has used for some time. The test pieces 
are in the form of flat plates one inch square and one-fourth 
inch or less in thickness. These plates being placed in the 
center of a briquet mold, the ends of the mold are filled with 
mortar. The plates may be improved by cutting shallow 
grooves in two opposite sides to make a more perfect fit with 
the sides of the mold. This m.ay easily be done with a round 
file. Besides the simple form of the test pieces and consequent 
ease of making them, this method has the further advantage 
that a test may be made almost as readily and accurately as a 
tensile test of cohesion. Also, since the adhesive area is one 
square inch, the results may be compared with cohesive tests 
on specimens having the same area of cross-section. 

154. The experiments on adhesive strength made by Mr. 
Mann were probably more extensive than any others published. 
His results are useful mainly as showing the lack of cementitious 
properties in the coarser grains of cement, and this point he 
proves very clearly by quite a large number of experiments. 
It was also developed that cement that had been rendered slow 
setting by aeration or " cooling" gave a lower adhesive strength 
than samples directly from the makers, which set more rapidly. 
But the method followed by Mr. Mann, of immersing the speci- 



108 CEMENT AND CONCRETE 

mens as soon as cemented together, may have had something to 
do with this result; the quicker setting samples would earlier 
resist the injurious action which is likely to follow the immer- 
sion of such small quantities of mortar before they have set. 

155. All of the things which influence the results in testing 
the cohesive strength must also be considered as affecting the 
adhesive test. The consistency of the mortar, the method of 
gaging, the pressure applied in cementing the specimens, and 
the conditions of storage until the time of breaking, will all 
have an influence on the result obtained. In addition to these, 
the character of the samples as to the kind of stone used, its 
structure, the physical condition of the surfaces, etc., must all 
be considered. It is therefore clear that many difficulties must 
be met before the test for adhesion can ever be included in 
standard tests. 

156. Special tests directed toward ascertaining the compara- 
tive adhesion of cement to different varieties of stone, the effect 
of the various differences in manipulation, the comparative ad- 
hesion of mortars containing various proportions of sand, etc., 
are of undoubted value. But, before the adhesive test can be 
considered a normal one for cement, much of this experimental 
work will be required. 

The results of a number of adhesive tests made under the 
author's direction are given in Art. 51. 

157. TESTS OF ABRASION. . Abrasion tests of cement are 
not at all common, and for the ordinary uses to which cement 
is put, its resistance to such action is of little interest except as 
it may imply other kinds of strength. Occasionally, however, 
it may be desired to have a mortar which will withstand wear, 
as, for instance, in making concrete walk. In such cases, tests 
for resistance to abrasion have some interest and value. 

The test is usually made on a sample prepared as for tensile 
or compressive tests, by submitting it to the wearing action of 
an emery or grindstone, or a cast iron disc covered with sand. 
The number of revolutions of the stone or disc is recorded, 
automatically if possible, and the loss of weight is determined 
after a given number of revolutions. 

A few tests of this kind made at St. Mary's Falls Canal are 
given in Art. 58. 



CHAPTER IX 

TENSILE TESTS OF COHESION 

158. THE testing of cement by applying tensile stress to a 
previously prepared briquet, containing definite proportions of 
cement and water, or of cement, sand and water, is the strength 
test which is now in most general use. The value of this method 
in comparison with that of other forms of rupture tests has al- 
ready been briefly discussed. 

That cement fails oftener in tension than in compression is 
one reason for preferring the tensile test. Its ready applica- 
bility is a still more important point in its favor. 

ART. 22. SAND FOR TESTS 

159. Whether the tensile test should be applied to neat ce- 
ment briquets or to those prepared from sand mortars has been 
a disputed point, but there are now but few authorities who 
recommend the use of the neat test exclusively. When tests 
for soundness are not carefully made, the behavior of the cement 
in neat briquets gives, perhaps, a better idea as to the reliability 
of the cement than do sand tests, but otherwise the sand test is 
a better index of the value of the cement. The principal ob- 
jection to the sand test is that the use of sand introduces another 
cause of variation in the results obtained by different experi- 
menters. This objection has considerable weight, because it is 
impracticable to find sand in widely separated localities which 
is absolutely the same in composition and physical properties; 
but two cements which appear to be of equal value when tested 
neat may exhibit quite different characteristics when used with 
sand, and it is believed that this fact far outweighs the objec- 
tion noted. As soon as regularity in sieves is established, the 
size of the sand grains may be regulated. The chemical and 
physical properties of the sand and the shape of the grains is a 
more difficult matter. The crushed quartz that is used in the 
manufacture of sandpaper was recommended by the Committee 

109 



110 



CEMENT AND CONCRETE 



of the American Society of Civil Engineers of 1885, and if some 
care is taken to select that which is clean and made from pure 
quartz, there is little difficulty in obtaining a uniform product 
of this kind. 

160. The German Normal Sand is obtained by washing and 
drying a natural quartz sand. In various parts of Germany 
sand answering the purpose may be found. Some tests made 
in this country to compare the " normal" German sand with 
American crushed quartz have shown the sand to give a some- 
what higher strength, while other tests have shown an opposite 
result. 1 A few of these tests are given in Table 27. 



TABLE 27 

Results Obtained with German "Normal" and American 
ard" Sand in Three Laboratories 



Stand- 



SAND. 


AGE. 
Days. 


STRENGTH OF MORTAR, 
1 CEMENT, 3 SAND, OBTAINED AT 
LABORATORY NUMBER 


3 


4 


5 


Normal 


7 
7 
28 
28 


218 
253 
317 
334 


173 
219 
341 
300 


201 
211 
281 
283 


Standard . 


Normal 


Standard 


Normal 


PER CENT. OF WATER USED. 


8 
9 


9 
10 


. . . 


Standard . . 









Mr. Max Gary has stated that "the Russian standard sand 
gives markedly lower, and the Swiss sand considerably higher, 
strength than the German." 

161. Tests with Natural Sand. It is not to be concluded 
from what has preceded that one must make mortar tests with 
a "standard" sand only. On the contrary, one may obtain 
valuable results by using in tests the sand which it is proposed 
to use on the work. The only point to be insisted upon is that 
a cement shall not be rejected on account of the poor quality 
of the sand used in testing. It is thus very desirable that a 
certain proportion of the tests be made with a pure quartz 
sand, .and by making parallel tests with the natural sands, the 



Article by Clifford Richardson, Engineering Record, Aug. 4, 1894. 



MAKING BRIQUETS 111 

coefficient of the latter may be obtained. In any case it is 
necessary, in order to obtain comparable results, to sift the 
sand used for tests. 

162. Fineness of Sand for Tests. The American practice in 
using crushed quartz is to reject the coarser particles by a sieve 
having 20 meshes per linear inch (holes about .03 inch square) 
and to reject the finer particles by a sieve of 30 meshes per linear 
inch (holes about .02 inch square). The size of grain of German 
normal sand is practically the same. In using a natural sand 
it is not necessary to use this size of grain, but it is better to do 
so, or at least to use some definite size or definite combination 
of sizes; as, for instance, one-half of 20 to 30 (passing holes .03 
inch square and not passing holes .02 inch square) and one-half 
30 to 50 (passing holes .02 inch square and failing to pass holes 
.012 inch square). Such a method will permit of duplicating a 
given size of grain at any time, while if the sand is used as it 
occurs in nature, considerable variations will be found. The 
effect of the quality of sand on the strength obtained is dis- 
cussed in Chapter XI. 

ART. 23. MAKING BRIQUETS 

163. Proportions. The proportions of the ingredients should 
always be determined by weight rather than by measure. It 
will be found more convenient to use metric weights for the 
dry ingredients. The water should then be measured in cubic 
centimeters, which is equivalent to weighing it in grams. The 
proportion of sand to be used for mortar briquets will depend 
upon circumstances, but for short time (seven day) tests good 
results are not usually obtained with natural cement if more 
than two parts of sand by weight are added to one part of ce- 
ment. Portland cement may be tested at seven days with 
three parts of sand to one of cement. If too large a proportion 
of sand is used, the briquets are liable to be injured in handling, 
and very low strengths are not as accurately recorded by the 
testing machine. 

164. CONSISTENCY: DETERMINATION. The consistency of 
the mortar has such a marked influence on the strength obtained 
that its importance can hardly be overestimated. The difficul- 
ties attendant upon specifying the consistency of a given mortar 
have already been touched upon in 116. The Committee of 



112 CEMENT AND CONCRETE 

the American Society of Civil Engineers of 1885 recommended 
the use of a " stiff plastic" mortar, but this phrase has had va- 
rious interpretations. 

The present Committee in its progress report 1 recommended 
the use of the Vicat apparatus: "In making the determination, 
500 gr. (17.64 oz.) of cement are kneaded into a paste, and 
quickly formed into a ball with the hands, completing the oper- 
ation by tossing it six times from one hand to the other, main- 
tained six inches apart; the ball is then pressed into the rubber 
ring ( 98) through the larger opening, smoothed off, and placed 
(on its large end) on a glass plate, and the smaller end smoothed 
off with a trowel; the paste, confined in the ring resting on the 
plate, is placed under the rod bearing the cylinder, which is 
brought in contact with the surface and quickly released. The 
paste is of normal consistency when the cylinder ( 1 cm. in di- 
ameter and loaded to weigh 300 grams) penetrates to a point in 
the mass 10 mm. (0.39 in.) below the top of the ring. Great 
care must be taken to fill the ring exactly to the top." 

The following simple test taken from French specifications 
will determine a good consistency of mortar to use for briquets. 
It should be capable of being easily molded into a ball in the 
hands, and when dropped from a height of one and a half feet 
on a hard slab, this ball should retain its rounded form without 
cracking. The mortar should also leave the trowel clean when 
allowed to drop from it. Were a smaller quantity of water 
used, the mortar would be crumbly and the ball would crack 
when dropped on the slab, while a larger amount of water would 
cause the mortar to adhere to the trowel and the ball would be 
flattened by striking the slab. 

165. Another method of determining the proper consistency, 
which the author believes will prove very satisfactory, is to 
make several batches of mortar containing the same weights of 
cement and sand, but having different percentages of water. 
As each batch is mixed, the volume of the resulting mortar is 
measured by pressing it lightly into a metal cylinder (a small 
tin pail will answer the purpose), taking pains to fill the cylinder 
in the same manner each time. That batch of mortar which 



1 Proc. Amer. Soc. C. E., Jan. 1903; also Engineering News, Jan. 29, 
1903, and Engineering Record, Jan. 31, 1903. 



CONSISTENCY OF MORTAR 



113 



occupies the least volume, when thus lightly packed, is the one 
in which the amount of water used is most nearly correct. 
Should either the mortar which contained the least water or 
that which contained the most water chance to have the least 
measured volume, then more trials must be made until such a 
consistency is obtained that either more or less water will in- 
crease the bulk of the mortar. This method will give a con- 
sistency somewhat more moist than that which gives the highest 
results on short time cohesive tests, but it is believed that where 
briquets are made by hand, more uniform results will be ob- 
tained when the mortar is a trifle moist. This method is not 
suited to daily use, as it requires too much time, but is valuable 
as a check on one's ideas of proper consistency. 

166. EFFECT OF CONSISTENCY ON TENSILE STRENGTH. 
Tables 28 and 29 give a few of the results obtained by the author 

TABLE 28 

Variations in Consistency of Mortar. Effect on Tensile Strength, 
Neat Natural Cement 







TENSILE STRENGTH, POUNDS PER SQUARE INCH. 


CEMENT. 




WATER USED EXPRESSED AS 




AGE OF 


PER CENT. OF DRY INGREDIENTS BY WEIGHT. 




BRIQUETS. 




Brand. 


Sample. 




25% 


30% 


35% 


40% 


45% 


Gn 


83 R 


7 days 


1136 


205d 


122/ 


720 


5U 


Gn 


84 R 




926 


72d 


58/ 


540 


39h 


An 


G 




162c 


165e 


108/ 


750 


64A 


An 


N 




1526 


194c 


204e 


134/ 


790 


Hn 


26 S 




226eZ 


. 176/ 


990 


56A 


35? 


Gn 


83 R 


28 days 


1896 


244d 


211/ 


1820 


135fc 


Gn 


84 R 




1496 


l6Sd 


114/ 


1070 


WSh 


An 


G 




210c 


22Se 


165/ 


1020 


SOh 


An 


N 




1736 


286c 


254e 


208/ 


1500 


Hn 


26 S 




333d 


309/ 


2170 


12U 


89i 


Ln 


31 S 


Iday 


162c 


148e 


97/ 


630 


36h 


Ln 


u 


7 days 


178c 


177e 


124/ 


710 


45A 


Ln 


u 


28 days 


207c 


257e 


202/ 


1400 


SSh 


Ln 


it 


3 mos. 


300c 


389e 


333/ 


2640 


197/1 



SIGNIFICANCE OF LETTERS 

a barely damp. 

b very dry ; no moisture shown on surface briquets. 

c dry ; slight moisture shown on surface briquets. 

d trifle dry. 

e about right consistency. 

/ trifle moist. 

moist. 

h very moist ; would just hold shape. 

1 extremely moist ; would not hold shape. 



114 CEMENT AND CONCRETE 

in tests to determine the effect of consistency on the tensile 
strength of natural cement mortars. All of the briquets were 
made in the usual manner and stored in fresh water until time 
of breaking. Each result given is the mean of from two to ten 
briquets. The letters affixed to each result indicate the degree 
of moisture which the mortar appeared to have when mixed, 
varying from "a," barely damp, to "i," so wet that the mortar 
could not hold its shape when laid on a glass slab. 

The results in Table 28 were obtained with neat cement mor- 
tars of several brands of natural cement. The first point to be 
noted is the variation in the amount of water required by dif- 
ferent samples to give the same consistency; thus, Brand An, 
sample N, when mixed with 35 per cent, water, appeared to 
have about the same consistency as did sample G of the same 
brand mixed with 30 per cent. It is also apparent that the 
strength of all samples is not affected alike by given variations 
in the amount of water used in mixing; comparing the results 
obtained when 45 per cent, water is used with that given when 
25 per cent, water is used, it is seen that at seven days the wet 
mortar gives 42 per cent, of the strength obtained with dry mor- 
tar for sample 84 R, Brand Gn, while with the sample of Brand 
Hn the strength of the wet mortar briquet is but 16 per cent, 
of that given by the dry mortar. Of the six samples tested at 
seven days and twenty-eight days, three gave the highest 
strength at seven days when mixed with 25 per cent, water, 
and five gave the highest strength at twenty-eight days when 
30 per cent, water was used. The results on Brand Ln show 
the greater proportionate gain with age of the wet briquets. 

Table 29 shows similar results for mortars made with one, 
two and three parts sand. With one part sand the wet mortar 
made from Gn, 21 R, which gave but 22 pounds per square 
inch at seven days, gave 429 pounds, or nearly the highest 
strength, at six months. A similar result is shown for sample 
15 R of the same brand when mixed with two parts sand, the 
highest strength at one year and two years being given by the 
mortar containing the greatest per cent, of water. That mor- 
tars containing three parts sand to one cement may be more 
easily damaged by an excess of water, is indicated by the re- 
sults on Brand Ln in this table. 

167. The effect on the strength of Portland cement mortars. 



CONSISTENCY OF MORTAR 



115 












^^il 


.3 




g 


S 


!5g| 


| 



jg 




^ 


M 


CO iO (M CO 
S<l <M t- 

r^ CO CO 


n 

CD 




a 
M 




a 

8 


1 


CO CO 00 to 
C.^ t !>. 
f 


If 



4 

1 


S 

be 
S 


S3 


T-H CO "^ 1 


!! 


M 

K 
fc 




** 

Q 


ii 


Itlf 8|ii If E 






w o> 


POUND 


1 


00 


i>- t~- l>- T-H CO 10 lO CO O O C^ GO 




SI 


H 
H 
1 
IB 


S 


S 


T-H CO (M (M T-H <?q 


II 




H 

H 
X 




><n 


TiC<ICOCO T-HCOCOCO r- 1 C<l C<l <?5 


1 
friS 


ENSILE 


X 


I 3 


rO-O-OrO '^'^'^'XS'^j'^'^S 1 ^ - 

*O O> T i CO *O rH *^ CO OO CO CO C^ ^* 


K- 

rt 

rH jj 


H 


5 


<N 


V) 


II 




i 


IS 


CO Ci GO O O O ^ Oi 


.2 S 

o 

a fl 






oo 


(N CO GO O Jj 


- Variatioi 
Natural Ce 




AGE OF 


BRIQUETS. 


c 

"d S 'd S r CP>s>>T3S>>>j r ^ S 

t^QOCOCO t-COCOCO OOCOrH(M OOCOr-HiM t-COCOCO 
<7<J CN (7^ 3^ cq r_ 

- - i. 


oj 

(M 




Q 

Mi 


I S 


c 
2 


H 






yl 


n 












P3 OQ ^ ^ 




| 





^dtuBg 


c^ w S S co 




Q 




putug 


O" >-5 " O""" O""" h^ 2 "" 



116 



CEMENT AND CONCRETE 



of variations in consistency, has been investigated by Mr. Eliot 
C. Clarke/ M. Am. Soc. C. E., and byM. Paul Alexandre, 2 Chief 
Engineer, Fonts et Chaussees. The results of one series of ex- 
periments made by M. Alexandre are given in Table 30. The 
mortars were mixed with fresh water and the samples immersed 
in sea water. 

TABLE 30 

Variations in Consistency of Mortar 

EFFECT ON TENSILE STRENGTH, PORTLAND CEMENT MORTAR. 

25 pounds cement to 1 cu. ft. sand (about 1 to 4 by weight). 



CONSISTENCY. 


WATER 
PER 
CENT. OF 
SAND. 


RESISTANCE, LBS. PER SQ. IN. AT AGE OF 


1 

ft 
eo 


I 


28 Days. 


8 

eo 


jj 


0> 
IN 


CO 


1 


Dry . . 

Ordinary . 
Wet . ., 


14 
22 
30 


31 
25 
16 


56 
46 
35 


73 

74 

55 


77 
116 
89 


69 
153 
126 


67 
170 
136 


88 
162 
180 


Disintegrated 
190 
189 



From "Recherches Experimental sur Les Mortiers Hydrauligues," 
par M. Paul Alexandre, Annales des Fonts et Chaussees, Sept., 1890 

It is seen that the highest strength at three days and seven 
days is given by the dryest mortar, at twenty-eight days to two 
years by that of the ordinary consistency, and at three years by 
that containing the highest per cent, of water. All of the sam- 
ples exhibited white spots in the broken section at three years, 
and at four years the dry mortar briquets had lost their cohe- 
rence on account of their porosity permitting the sea-water to 
permeate them. 

168. Conclusions. It may be concluded, then, that the 
consistency of the mortar has a very marked effect on the ten- 
sile strength obtained; that different samples of cement are not 
affected in the same degree by given variations in consistency ; 
that the effect of consistency is usually shown most plainly in 
short time tests; and that while the dryer or stiffer mortars give 
the highest results on short time tests, the moist mortars attain 
a greater strength after a certain time. 



1 "Records of Tests of Cement made for Boston Main Drainage Works." 
Trans. A. S. C. E., Vol. xiv. 

2 Annales des Fonts et Chaussees, Sept., 1890. 



TEMPERATURE 117 

169. Temperature of the Ingredients and of the Air where the 
Briquets are Made. The temperature of the mortar and of 
the air in which the briquets are prepared is a matter of some 
moment. In 1877, Mr. Maclay * reported a series of experi- 
ments on Portland cements from which conclusions may be 
drawn concerning the effects of the temperature of the mortar. 
These experiments indicate that mortar having a temperature 
of 40 Fahr. when gaged, will attain greater strength in from 
seven days to three weeks than a mortar having an initial 
temperature of 70 Fahr. One is most likely to work some- 
where between these two temperatures, but it may be mentioned 
that according to Mr. Maclay's experiments, it appears that 
mortars gaged at a temperature of 90 or 100 Fahr. also at- 
tain a higher strength than those gaged at 70 Fahr. 

Similar experiments made by M. Candlot 2 indicate that mor- 
tars gaged with cold water give but feeble resistance at first, 
but in from two weeks to one month, such mortars surpass in 
strength those gaged with warm water. M. P. Alexandre 3 im- 
mersed some briquets at a temperature of about 90 C. (194 
Fahr.) for forty-eight hours and then at 15 to 18 C. (60 to 
65 Fahr.) until broken, while other briquets were maintained 
at the latter temperature from the time of molding. The bri- 
quets that were broken at the age of four days showed that 
the highest strength had been obtained by the briquets which 
had been kept hot for forty-eight hours, but at twenty-eight 
days and three months those briquets which had not been sub- 
jected to this high temperature gave the highest strength. 

170. Table 31 gives a few of the many experiments on this 
point made under the author's direction. It appears that the 
briquets made in a low temperature (34 to 37 Fahr.) are usu- 
ally stronger than those made in the ordinary temperature of 
65 to 68 Fahr. In some cases the difference was not very 
great, and in some of the tests the briquets made in the ordi- 
nary temperature gave higher results at one day and seven 
days than those made in the cold; but at twenty-eight days the 
cold-made briquets were nearly always in the lead, and in many 

1 "Notes and Experiments on the Use and Testing of Portland Cement," 
Trans. A. S. C. E., Vol. vi, p. 311. 

2 "Ciments ei Chaux Hydrauliques" 

3 "Les Mortiers Hydrauliques." 



118 



CEMENT AND CONCRETE 



TABLE 31 

Temperature of Materials and of Air where Made. Effect on Tensile Strength, Natural 
and Portland Cements 


TENSILE STRENGTH, POUNDS PER SQUARE INCH. 


Age of Briquets When Broken. 


i 


moo 


. . . co 


NOTES: All briquets made by same molder; each result is mean of five to ten specimens. 
Results in columns headed "warm," temperature of materials used and air where made = 65 to 67^ Fahr. 
Results in columns headed "cold," temperature of materials used and air where made = 34 to 37 Fahr. 
a. In damp closet 36 hours, except 1 day specimens which were 12 hours in air where made and 12 hours 
in damp closet. 
6. In damp closet 24 hours, except 1 day specimens which were 24 hours in air where made. 
c. In damp closet 19 to 21 hours, except 1 day specimens which were 3 to 5 hours in air where made, 
to 2 hours in damp closet and 16 to 21 hours in tank. 


nut? AY 


t- 


. . . <N . . 


6 Months. 


woo 


* ^ co ! S 


UU13AY 


CO cq O 

'. >* co ! T}< 


3 Months. 


PPO 


CO O C5 O f-H cq 
CO (M CO G<l t^ > 


HUB AY 


O CD I>- O t-~ O 

^ 00 "^ *O O ^O 


28 Days. 


P,oo 


^ OOTt^ O O ^H 

10 r-l i-H TtH 00 

(N (7<l C<I i i t^ CO 


* 


o co o 10 ci c^ 

I-H O 1 OO CO 1 H ^ 
(M (N T-H t^ CO 


i 

p 


P,o 


j S ^ '. CO <M 


ui.n?AY 


CD t^ $ 


! 

co 


pio 


:S 


* 


:1 


! 


'PI0 


'ii ; 


uu.* 


(M CO --i 
CD t^ O 

T 1 rH 


fe 

o 
W 

1 


IJ9SOIO dltlBQ 


llfa^ 


9.I8HAY J IV 


<N ^ "? C<J ^ 00 

^ * 6 ^ * 


AHd *x^taj3 
Had 'aasn naxvAY 


rH r-n' CO CO (?J 

CO CO rH l-l (N rH 


'XHOiaAY AH xKiawao 
axo ox og-06 zxavn) 
aansnno ao sxavd 


O O -H CO O CO 


N < " 
IB 

9 




CO - rH - rH - 
rH " (N (M 

C ^ C ^ t_i ^ 

<j- O - K - 

l ss= i" 



MORTAR MIXING 119 

cases this difference held good at three months and six months. 
Some of the results indicated that if the briquets were allowed 
to remain twenty-four hours or more in the cold air, it tended 
to counteract the beneficial effects of cold molding, but this 
point was not satisfactorily established. 

171. From the foregoing the following conclusions may be 
drawn: To make briquets of cold materials and allow them 
to remain some hours in cold air, retards the hardening of the 
briquets; but when briquets so treated are, after a few hours, 
placed in a medium of ordinary temperature, they gain strength 
more rapidly than briquets made of warm materials and kept 
continuously at the ordinary temperature of 60 to 70 Fahr. 
After being placed in a warmer medium, the briquets made 
with cold materials in cold air frequently gain strength at such 
a rate as to surpass in strength the warm-made briquets at seven 
days; the former almost invariably surpass the latter at twenty- 
eight days. In some cases it appears that this superiority of 
cold-made briquets is maintained up to six months, but in other 
cases the difference seems to disappear after three months. 

Although these variations in temperature have not as marked 
an effect on tensile strength as have many other variations in 
manipulation, yet in carefully conducted experiments one should 
always operate in a constant temperature. As a matter of 
convenience, 65 to 70 Fahr. will commend itself, and this 
temperature may well be taken. 

172. GAGING BY HAND. The objects to be attained in 
gaging are to thoroughly incorporate the cement and sand, to 
evenly distribute the water throughout the mass, and, if pos- 
sible, to give the mortar a certain tenacity resembling that of 
putty. This last object is not always possible of attainment 
with mortars containing a large dose of sand. 

The ordinary method of preparing mortars in the laboratory 
is to gage with a trowel on a glass, slate, or marble slab. In 
gaging mortars, the cement and sand are first mixed dry; the 
materials are then drawn away from the center, leaving a crater 
to receive the water, which is all added at one time. The dry 
material is then gradually turned from the edges toward the 
center until all of the water is absorbed, after which the mass 
is thoroughly worked with the trowel in such a way as to rub 
the material between the trowel and plate until the consistency 



120 CEMENT AND CONCRETE 

is uniform throughout. A batch of mortar sufficient for five 
briquets cannot usually be properly gaged by this method in 
less than five minutes. 

The Committee of the American Society of Civil Engineers, 
in their preliminary report on methods of manipulation, sug- 
gested that "as soon as the water has been absorbed, which 
should not require more than one minute/ 7 the mortar should 
be kneaded with the hands for one and one-half minutes, the 
process being similar to that used in kneading dough. 

173. HOE AND BOX METHOD. Mr. Alfred Noble used for 
many years a form of gaging apparatus, consisting of a box 
with sloping bottom, in which the mortar is worked by means 
of a hoe. The author has used an iron box made on this prin- 
ciple (Fig. 2), which has given excellent results. The box is 
2 feet 1\ inches long, 6 inches wide at the bottom, and at the 




fr 



Side Elevation 

FIG. 2. MIXING BOX. 




center is 6 inches deep. The level part of the bottom is 3 inches 
by 6 inches, and from this level part the inclined portions of 
the bottom slope up toward the ends at an inclination of about 
22 degrees. The sides of the box extend below these in- 
clined planes to give a level bearing for the box when in use. 
It is also well to have the sides flare enough to give a width 
of 6 inches at the top to prevent the hoe from becoming 
wedged. A " German clod hoe," which is strong and heavy, 
yet a trifle flexible in the blade, is used in connection with the 
box. 

The weighed quantities of the dry ingredients being put in 
the box and well mixed, the measured volume of water is added. 
Two minutes of hard work, in which the operator may put all 
his strength, is sufficient to bring the mass to plasticity if the 
amount of water added is correct. A return to the trowel and 



MORTAR MIXING 121 

slab method of mixing is not likely after a trial of this simple 
device. 

174. MACHINE FOR MORTAR MIXING. As the mixing by 
hand is a rather slow and tedious method, and the hoe and box 
method are not very generally known, several machines have 
been devised to do the work. None of them, however, has 
given such satisfactory results as to bring it into general use. 

One of the machines is called a "jig," or "milk shake" 
machine, 1 and consists of a cup which moves rapidly up and 
down, this motion being imparted by means of a hand wheel,, 
crank and connecting rod. The dry cement and water being 
placed in the cup and tightly covered, a few rapid turns of the 
wheel are sufficient to reduce the cement to a paste. This, 
form is only applicable to neat cement mortars, and has been 
said to give unsatisfactory results even for these, though in 
some laboratories this machine has been used for all neat 
mortars. 

Other forms have been made in which the mortar is thoroughly 
stirred by means of forks or blades projecting into the mortar 
from a horizontal arm above. The gager devised by Mr. Faija 
is constructed on this principle, and similar machines may be 
obtained from manufacturers of testing apparatus. 

175. Steinbrtich's Mortar Mixer is a German machine oper- 
ating on a different principle. It consists of a circular shell 
having on its upper side and near its outer edge a circular groove, 
or trough, to receive the mortar to be mixed. In this trough 
rests a wheel on a fixed horizontal axis, which is above the pan 
and normal to the axis of the pan. A cross-section of the rim 
of the wheel is a semicircle fitting the groove in the pan. The 
gearing is such that the pan is made to revolve about its vertical 
axis, and the wheel about its horizontal axis, the inner surface 
of the trough and the under side of the periphery of the wheel 
where the two are in contact moving in the same direction at 
a given instant. The mortar is thus rubbed between the two. 
Small blades, or plows, scrape the sides of the trough as the latter 
revolves, thus keeping the mortar in the bottom of the trough. 
The wheel and the plows are mounted on hinged axes, or sup- 
ports, so that they may be raised from the pan when the mortar 



1 S. Bent Russell, Engineering News, Jan. 3, 1891. 



122 



CEMENT AND CONCRETE 



is to be cleaned out. The mixing requires about two and one- 
half minutes. The price of the machine is about $130. 

176. The amount of gaging which a mortar receives has an 
important effect on its consistency and the strength it will 
attain. This was found to be the case in several experiments 
where mortar gaged eight minutes in the box described above, 
gave from 15 to 35 per cent, greater strength at one year than 
that which was gaged but two minutes, the amount of water 
used being the same in the two cases. Experiments on this 
point are given in Table 78, 364. It is therefore important to 
eliminate, if possible, the variations which must follow hand 
mixing, but as yet no apparatus has seemed to meet with gen- 




FlG. 3. FORM OF BRIQUET 
USED ON THE CONTI- 
NENT OF EUROPE. 



FlG. 4. -FORM OF BRIQUET 
USED IN THE UNITED 
STATES. 



eral approval, though among machine mixers those similar to 
that used by Mr. Faija seem to have given the best results. 
The hoe and box method described in 173 partially eliminates 
the personal equation, and for facility of operation and thor- 
oughness of mixing leaves little to be desired. 

177. FORM OF BRIQUET. The shape and size of the briquet 
have been the subject of much discussion and experiment. 
Mr. John Grant, a pioneer in tensile tests, tried many forms 
before finally adopting one quite similar to the form afterward 
recommended by the Committee of the Amer. Soc. C. E. in 
1885. Mr. Alfred Noble also made a series of experiments" on 



FORM OF BRIQUET 123 

different styles of molds and clips, and presented the results 
in a paper read before the American Society of Civil Engineers. 1 

There are two forms of mold that are now in quite general 
use. On the continent of Europe the form most generally used 
is that shown in Fig. 3. It has a cross-sectional area of five 
square centimeters (.775 sq. in.) at the smallest place, and the 
heads of the briquet are elliptical in form, the major axes being 
transverse to the briquet axis. The curve forming the side of 
the briquet in the central portion is of very short radius, giving 
the effect of a semicircular notch on either side of the briquet 
at the smallest section. These notches have the effect of con- 
fining the break to this place. 

The other form of mold is the one mentioned above as recom- 
mended by the Amer. Soc. C. E. Committee, and used in America 
and England. A briquet of this form is shown in Fig. 4. The 
cross-sectional area at the center is one square inch, and the 
increase of section toward the ends is gradual, the radius of the 
curve at the side of the briquet being f inch. 

178. Area of the Breaking Section. Formerly a section of 
2\ square inches was more commonly used here and in England, 
while an area of 16 square centimeters (2.48 sq. in.) was com- 
mon in France and other continental countries. The larger 
the area of the breaking section, the smaller will be the com- 
puted strength per square inch; this point seems fairly well 
established, although the experiments recorded in a very ex- 
cellent paper by Mr. Eliot C. Clarke 2 indicate no apparent dif- 
ference in strength between briquets 1 square inch and 2\ 
square inches in section. 

M. Durand-Claye found that the tensile strength of a briquet 
varied more nearly as the perimeter than as the area of the 
section. The experiments of M. Candlot do not point to this 
conclusion, though they clearly show that the indicated strength 
per square centimeter is very much greater for a briquet hav- 
ing an area of five square centimeters at the small section than 
for a briquet of 16 square centimeters area. 

Mr. D. J. Whittemore 3 experimented with briquets that were 



1 Trans. Amer Soc. C. E., Vol. ix, p. 186. 

2 Trans. Amer. Soc. C. E., Vol. xiv, p. 141. 

3 "Tensile Tests of Cements," etc. Trans. A. S. C. E., Vol. ix, p. 329. 



124 CEMENT AND CONCRETE 

circular in cross-section. He found that while the ultimate 
strength of a briquet was about proportional to the periphery 
of the breaking section for the ordinary solid briquet, yet if a 
core were inserted in the mold, giving the cross-section an annu- 
lar form, this proportion was not maintained. It was con- 
cluded from this that the apparent peripheric strength could 
not be explained by saying that the surface of the briquet had 
gained a greater strength than the interior, but that the expla- 
nation must rather be sought in the method of applying the 
stress in breaking the briquet. The force being communicated 
to the surface of the briquet, the stress is not uniformly dis- 
tributed throughout the breaking section, because of the low 
elasticity of the mortar. 

M. Paul Alexandre showed that the difference in strength 
per unit area decreased with age, although it did not entirely 
disappear at one year. It would therefore seem that the expla- 
nation of this phenomenon may be found in a combination of 
these two causes; more rapid hardening of the smaller speci- 
mens, and greater inequality of stress in breaking the briquets 
of larger section. 

179. Form of Briquet Suggested. As a result of experi- 
ments which will be described under the head of "Clips," 1 
(Art. 25) the following conclusions were drawn as to the desir- 
able features for a briquet: 

1st. The smallest section should not have an area much less 
than one square inch. Probably an area of five square centi- 
meters would represent a minimum. 

2d. The area of the section of the briquet between opposite 
gripping points should be about one and three-fourths times 
the area of the smallest section. 

3d. The distribution of stress over the smalles'i/ section 
should be as nearly uniform as possible. 

4th. The curve of the sides at the breaking section should 
not be very sharp; one-half inch might be taken as a minimum 
radius. 

5th. The area of the vertical section from the gripping point 
to the plane of the end of the briquet the section subjected 



1 These experiments were described by the writer in detail in "Municipal 
Engineering," Dec., 1896, Jan. and Feb. 1897. 



FORM OF BRIQUET 



125 



to shear when the stress is applied should be nearly as great 
as the area of the neck of the briquet. 

6th. The face and back of the briquet should be parallel 
planes, to permit of easy storage. 

7th. The total volume should be kept as small as is consis- 
tent with the other conditions. 

Fig. 5 represents a form of briquet which will, it is thought, 
satisfactorily fulfill the above requirements, and in which it is 
believed the full strength of the smallest section may be more 
nearly developed than with present forms. The curve at the 
central section has a radius of one inch, and the line of the 
side of the briquet is con- 
tinued in a tangent one- i< 

half inch in length, having 
an inclination of nearly 
45 degrees with the axis 
of the briquet. The total 
length of the briquet is 
four inches, the ends be- 
ing formed by straight 
lines tangent to the curves 
forming the corners. If 
the clip is so formed that 
the gripping points bear 
at the centers of the one- 
half inch tangents form- 
ing the sides of the briquet, 
the distance between op- 
posite gripping points will 
be 1J inches. 

180. Comparison with 
other Forms. Compar- 
ing this briquet with the forms in common use, the German and 
the form shown in Fig. 5 both have an area between opposite 
gripping points about If times the area of the smallest section, 
but in the form shown in Fig. 4 this ratio is too small to fulfill 
the second specification. 

The unequal distribution of stress over the breaking section 
of the briquet has already been mentioned as a probable partial 
cause why briquets of small cross-section show a greater strength 




FIG. 5. FORM OF BRIQUET SUGGESTED 

FOR USE 



126 CEMENT AND CONCRETE 

per unit area than those having a larger area of cross-section. 
In Johnson' 3 " Materials of Engineering" is given the theory of 
the distribution of stress over the breaking section of a briquet, 
as developed by M. Durand-Claye, and published in Annales des 
Fonts et Chaussees of June, 1895. Applying the formulas there 
given to three styles of briquet, the A. S. C. E. form of 1885, 
the German standard, and the form shown in Fig. 5, it is found 
that the ratios of the maximum stress to the mean stress are, 
for the three forms respectively, 1.54, 1.52 and 1.22. From a 
theoretical point of view, this means that with a total pull of 
100 pounds on each briquet, the outer fiber of the briquet 
shown in Fig. 4 would be subjected to a stress of 154 pounds 
per square inch, while with the form suggested above, the 
stress on the outer fiber would be but 122 pounds per square 
inch ; briquets of the latter form should, therefore, theoretically, 
show a breaking strength 1.27 times the strength given by 
briquets of the same mortar made in the A. S. C. E. form of 
1885. 

The German form has too sharp a curve at the sides to fulfill 
the fourth requirement given above. All of the forms comply 
with the first, fifth and sixth requirements. 

As to the volume of the briquet, the author's form having a 
total length of four inches, has about 50 per cent, greater volume 
than the A. S. C. E. form of 1885. 

181. MOLDS. In the early tests of cement, wooden molds 
were employed, but they absorb water from the mortar and 
soon warp out of shape. Iron molds have also been used to a 
considerable extent, but these are apt to become rusted if not 
in constant use. Brass, bronze or some similar metal not easily 
corroded should be used, and molds of this character can be 
obtained of dealers in testing apparatus. 

The molds may be made single, or in " nests" or " gangs" 
of three to five. The two halves of the mold may be entirely 
separable, or may be hinged at one end and fastened by a clip 
at the other end. The gang molds are somewhat cheaper than 
the single ones. The hinged molds and those held with patent 
clip are rather difficult to clean, while the gang molds, if made 
heavy enough to prevent spreading, are unwieldy, and briquets 
are removed from them with greater difficulty than from the 
single molds. It is considered, therefore, that the most con- 



MOLDING BRIQUETS 127 

venient form is the single mold, in which the two halves are 
held together by a screw clamp of simple design. 

182. To clean these molds, place ten in a row with clamps 
removed ; scrape the upper faces with a piece of zinc, brush 
with a stiff " horse-brush," and wipe with oily waste. Turn 
them over and repeat the process. Then separate the two 
halves of each mold, place the twenty halves in line with inner 
surfaces up, forming a trough twenty inches long. Wipe this 
trough thoroughly with oily waste, finishing with some that is 
only slightly oiled. 

183. MOLDING. Methods of molding briquets vary widely 
and have a considerable effect on the results obtained by differ- 
ent operators. The mold may be placed on a glass or marble 
slab, or on a porous bed. This difference in treatment will 
affect the results chiefly because a porous bed will extract 
moisture from the briquet, and, unless it is already mixed very 
dry, will make it give a higher result on a short time test. The 
use of a porous bed probably originated with a desire to more 
closely imitate the use of mortar in actual work, but it intro- 
duces another source of variation in results and should not be 
followed. 

184. In hand work the whole mold may be filled at once, 
or small amounts of mortar may be added at a time, and each 
layer packed; the mortar may be tamped into the mold with a 
rod, in which case the pressure used may vary widely; or the 
mortar may be pressed in with the fingers, or with the point of 
a small trowel; and, finally, the pressure applied on the top of 
the whole briquet may be light or heavy. It is evident that it 
is almost impossible to so describe all these details of manipula- 
tion that another operator may follow the same system and 
obtain the same results. The practice of ramming the mortar 
into the mold by means of a metal rod or a stick faced with 
zinc is objectionable, because of the possible wide variation in 
the force thus applied. This method is sometimes used by 
manufacturers, since by making the mortar quite dry and ram- 
ming it into the molds very hard, a high initial strength is 
obtained. But the foremost cement makers are now eschewing 
such methods and are aiming to make fair tests. Some experi- 
ments made under the author's direction indicate that the 
pressure applied to the top of the briquet is the salient point in 



128 



CEMENT AND CONCRETE 



the process of molding, and that the other details are of minor 
importance. 

In Germany a heavy trowel or iron plate weighing about 
250 grams, and provided with a handle, is used in making one- 
to-three mortar briquets. The mortar is made rather dry 
(about 10 per cent, water), and after the mold is filled and 
heaped, the mortar is beaten with the trowel until it becomes 
elastic, and water appears on the surface. The excess of mor- 
tar is then scraped off with an ordinary trowel or spatula. 

185. Several machines have been devised for making bri- 
quets, some of which are said to give good results. Among 
these the most prominent is the Bohme hammer apparatus, 
which is much used in Germany, although not employed to any 
extent in the United States. It consists of a plunger which 
fits the mold and upon which a given number of blows are 
struck by a hammer. The mortar is first gaged as for hand 
molding, and placed in the form. A pinion, turned by a hand 
crank, is geared to a wheel provided with ten cams. These 
cams operating on the wrought iron handle of the hammer 
cause a certain number of blows to be delivered to the plunger. 
The mechanism is automatically shut off after the proper number 
of blows has been delivered. The following results were ob- 
tained by Professor Bohme with his apparatus: 

TABLE 32 

Comparison of Hand Made Briquets with Those Made by Bohme 

Hammer 









MEAN TENSILE 


No. 


METHOD. 


WEIGHT OF 
BRIQUETS. 


STRENGTH AT 
7 DAYS IN KGS. 








PER SQ. CM. 


1 


By hand 


160.0 


16.06 


2 


Hammer, 75 blows 


158.0 


12.75 


3 


100 " 


159.5 


13.25 


4 


" 125 " 


159.5 


14.56 


5 


" 150 " 


159.0 


15.56 



186. Several American engineers have devised machines for 
briquet-making, but none of them has been generally adopted. 

An apparatus designed by Prof. Charles Jameson, of Iowa 
University, is* said to work very rapidly. The mortar is packed 
in the mold by a plunger of the form of the briquet. This 



MOLDING BRIQUETS 129 

plunger works in a chamber of the same shape as the briquet 
mold. The mortar is placed in a hopper at the side of this 
chamber, and is delivered to the mold automatically when the 
plunger is raised. The force is applied to the plunger by hand, 
but it should be so arranged that this be done by a weight, to 
prevent variations in pressure. In this method the briquet is 
removed from the mold as soon as made, and this would appear 
to be an objectionable feature. 

Professor Spalding, of Cornell University, in his excellent 
little book on " Hydraulic Cement," states that he has found 
that "a pressure of about 500 pounds upon the surface of the 
briquet is sufficient to produce a compact and homogeneous 
briquet, and a crude appliance consisting of a lever arranged to 
bring a pressure upon the mortar in the mold by means of a 
weight suspended at the end of the lever, has been found to 
increase both the rapidity and the regularity of the work, and 
especially to diminish the variations in results obtained by dif- 
ferent men." 

A machine which would give more uniform results and work 
more rapidly than hand molding, would commend itself for 
general use. 

187. Method Recommended. In making briquets by hand, 
the mortar may well be packed into the molds by the fingers, 
which should be protected by rubber tips. When the mold is 
filled and slightly heaped, the trowel should be placed on top, 
and the molder put about 60 pounds pressure on the trowel. 
The excess mortar is then cut off by the trowel and the top of 
the briquet is smoothed by drawing the trowel across the face. 
The results obtained by four molders using this method in the 
same laboratory are given in Table 33. 

188. The recent progress report of the Committee of the 
American Society of Civil Engineers on uniform tests of cement 
contains the following, under " Molding": 

" Having worked the paste or mortar to the proper consist- 
ency, it is at once placed in the molds by hand. 

"The Committee has been unable to secure satisfactory re- 
sults with the present molding machines; the operation of 
machine-molding is very slow, and the present types permit of 
molding but one briquet at a time, and are not practicable with 
the pastes or mortars herein recommended. 



130 



CEMENT AND CONCRETE 



TABLE 33 
Results Obtained by Different Molders when Using Similar Mortar 









, o 




HEAN TENSILE 










g 


i I 


1! 




STRENGTH. 


If 






1 

1 


PARTS 
CEMEN 


ljl 


!<3 

- Ec| 
3g 


AGE. 


fc 
M 


* 


H 


|l 

o 


DATE. 




H 
H 

H 
4 


k 



OH 
K K 


sS 1 ^ 


5Gd 

S 




i 


3 


H 

O 


H 




H 


P3 




H b 


s 







o 


o 










1 


^ Q 


HI 




i 


i 


i 


P 








a 


6 


c 


d 


e 


/ 


9 


k 


i 




1 





31.6 


62-65 


7 days 


81 


92 


89 


5 


10-22 


Clear 


2 





" 


" 


28 " 


197 


213 


220 


5 






3 


1 


18.7 


67-62 


7 " 


79 


91 


89 


5 






4 


1 


u 


^ 


28 " 


235 


257 


259 


5 






5 


1 


(( 


63-68 


3 mo. 


515 


541 


519 


5 






6 


1 


It 


it 


1 year 


558 


569 


555 


5 






7 


2 


15.2 


70-65 


28 days 


196 


186 


197 


5 






8 


2 


" 


u 


3 mo. 


423 


383 


406 


5 






9 


3 


13.3 


65-61 


3 mo. 


253 


263 


239 


5 






10 


3 


" 


" 


1 year 


260 


232 


236 


5 






11 


Su 


m of M 


eans 




2797 


2827 


2809 




















Holder 


Holder 




















S. 


T. 








12 





31.6 


62-65 


7 days 


. 


60 


60 


5 


10-28 


Cloudy 


13 





t< 


" 


28 " 


. 


145 


167 


5 


*' 




14 


1 


18.7 


65 


7 " 





67 


71 


5 


i 




15 


1 


^ 


(i 


28 


. 


223 


211 


5 


i 




16 


1 


u 


" 


3 mo. 




435 


449 


5 


i 




17 


1 


(1 


U 


1 year 




504 


491 


5 


' 




18 


2 


15.2 


67 


28 days 


. 


182 


179 1 


5 


t 




19 


Sum of Means 







1616 


1628 









Cement, Brand Gn, Sample 21 R. Sand, Crushed Quartz 20 to 40. 
All briquets in same line received same treatment after made and were 
immersed in same tank until broken. 
1 Mean of ten specimens. 

" Method. The molds should be filled at once, the material 
pressed in firmly with the fingers and smoothed off with a 
trowel without ramming; the material should be heaped up on 
the upper surface of the mold, and, in smoothing off, the trowel 
should be drawn over the mold in such a manner as to exert a 
moderate pressure on the excess material. The mold should be 
turned over and the operation repeated. 

"A check upon the uniformity of the mixing and molding is 
afforded by weighing the briquets just prior to immersion, or 



STORING BRIQUETS 



131 



upon removal from the moist closet. Briquets which vary in 
weight more than 3 per cent, from the average should not be 
tested." 

189. Marking the Briquets. The briquets made in a given 
laboratory should be numbered consecutively, so that no con- 
fusion can arise, and this one number is all that should be placed 
on the briquet. The record of the brand of cement, the pro- 
portions used, etc., should be placed in a book opposite the 
briquet number. The briquets should be numbered on the 
face, near the end. Steel stamps furnish a ready means of 
numbering, and when mortar contains more than two parts of 
sand to one of cement a thin strip of neat cement paste plastered 
across one end of the briquet will aid in making the numbers 
legible. 

ART. 24. STORING BRIQUETS 

190. The Time in Air before Immersion. As soon as the 
briquets are molded they should be covered with a damp cloth 

TABLE 34 

Variations in Length of Time Briquets are Left in Moist Air before 
Immersion Natural Cement 









TENSILE STRENGTH, POUNDS PEH 








SQ. INCH. 


CEMENT. 


PARTS CRUSHED 
QUARTZ, 20-30 

TO 

1 CEMEXT 


AGE 
WHEN 
BROKEN. 




HOURS IN MOIST AIR BEFORE 
IMMERSION. 








8 


12 


24 


48 


72 


168 


Brand. 


Sample. 


Gn 


loR 





7 days 


123 




139 


151 


161 


237 




i 


1 


7 days 


91 


. . . 


106 


114 


114 


182 




16 R 





28 days 


110 


. . . 


106 


109 


89 


113 






1 


28 days 


142 




138 


139 


152 


175 






2 


28 days 


102 




105 


112 


113 


115 


An 


G 





7 days 


. 


168 


181 


194 


185 


238 









28 days 


. 


200 


210 


224 


241 


243 






1 


7 days 




108 


137 


141 


157 


160 






1 


28 days 




278 


283 


297 


297 


301 






3 


28 days 




120 


130 


137 


139 


152 



NOTE : All briquets made by same molder. Each result is mean of ten 
specimens. 

until they are ready to be removed from the molds, when they 
should be transferred to a "damp closet," lined with zinc or 
other non-corroding metal. It was formerly the practice to 
immerse the briquets as soon as they were considered to be 



132 



CEMENT AND CONCRETE 



sufficiently set; but for the sake of uniformity, they are now 
left in moist air for twenty-four hours before immersion, whether 
the cement is quick or slow setting. Briquets which are to be 
broken at twenty-four hours, however, are usually immersed 
as soon as set hard. 

Table 34 gives the results obtained by allowing natural 
cement briquets to remain in moist air different lengths of time 
before immersion. In general, the strength is greater for seven 
and twenty-eight day tests the longer the briquets are allowed 
to remain in the moist air. It appears that, while the time in 
moist air should be made as nearly uniform as possible, a varia- 
tion of a few hours will not cause an important difference in 
strength. 

TABLE 35 

Gain or Loss in Strength of Natural Cement Briquets by Immersion 









TENSILE STRENGTH, POUNDS 








PER SQ. INCH. 


TIME IN 
MOIST AIR. 


TIME IN TANK. 


AGE WHEN 
BROKEN. 






One Part Stand- 








Neat Cement. 


ard Sand to 










One Cement. 


20 hours 




20 hours 


151 


94 


18 hours 


6 days 


7 days 


147 


153 


2 days 




2 days 


192 


126 


2 days 


5 days 


7 days 


160 


158 


3 days 


. 


3 days 


205 


141 


3 days 


4 days 


7 days 


177 


155 


4 days 


. 


4 days 


218 


165 


4 days 


3 days 


7 days 


191 


165 


5 days 




5 days 


230 


175 


5 days 


2 days 


7 days 


192 


169 



NOTE : All briquets made by same molder. Each result is mean of 
five specimens. 

Table 35 shows the early action of the water on the briquets. 
These tests were made in sets of ten; five briquets of a set were 
immersed after twenty hours, forty-eight hours, etc., while the 
other five of the same set were broken at the time the first five 
were immersed. With this sample of natural cement, it appears 
that the briquets lose part of their strength by immersion, and 
that some time is required to regain this lost strength. Thus, 
with neat cement mortar the briquets broken at twenty hours 
without immersion were as strong as those broken at seven 
days which had been immersed the last six and one-fourth days. 
With briquets of one-to-one mortar, it appears that if immersed 



STORING BRIQUETS 133 

at the end of four days, the gain in strength during the last 
three days (in water) is about equal to the loss of strength due 
to immersion. If immersed earlier than this, the gain is greater 
than the loss, but if immersed later, the loss is greater than 
the gain. 

191. For storing briquets the required twenty-four hours 
before immersion a moist closet is very convenient, tends to 
promote uniformity of treatment, and may be very easily 
made. The use of a damp cloth for covering briquets is incon- 
venient, as the cloth may dry out. If it is used, the end of 
the cloth should rest in a pail of water, so it will keep wet by 
capillarity; it should also be kept from touching the briquets by 
a wire screen or by wooden slats. 

A moist closet may be made of slate, glass or soapstone, or 
of wood lined with metal. In the bottom of the box is a pan of 
water, or a sponge kept constantly wet. The shelves may well 
be of glass, and should be so arranged that any shelf may be 
removed without disturbing the others. 

192. Water of Immersion. When the briquets are ready to 
be immersed, i.e., usually, twenty-four hours after made, they are 
placed in a tank, containing water that is kept fresh by frequent 
renewals. The water in the tank should also be maintained at 
a nearly constant temperature. It is sometimes the case that 
briquets are subjected to considerable variations of temperature 
while in storage. It also frequently occurs that the water is 
allowed to become stale. A few of the many experiments 
made at St. Marys Falls Canal to show the effect, on the tensile 
strength of natural cement briquets, of variations in the tem- 
perature of the water of immersion, are given in Table 36. The 
details of these experiments, as well as other tests on the same 
point, may be found in the Annual Report, Chief of Engineers, 
U. S. A., for 1894, page 2314. 

The very marked effect which the temperature of the water 
may have on the rate of hardening of natural cements is clearly 
shown. When broken at the age of one day or seven days, 
the effect on the strength may not be evident, or the briquets 
stored in cold water may develop a greater strength, but the 
more rapid hardening of the briquets stored in warm water is 
usually very evident at twenty-eight days, and increases up to 
two or three months. Some samples of cement are affected 



134 



CEMENT AND CONCRETE 



less than others, and a few experiments indicated that the 
differences in strength due to the temperature of water of im- 
mersion decrease after three months and become almost nil at 
one year. 

193. The conclusion drawn from these tests may be briefly 
stated as follows: Between certain limits the early strength of 
natural c'ement mortars is usually developed faster in cool 

TABLE 36 

Variations in Temperature of "Water in which Briquets are 

Immersed 







o 














( 


TENSILE STRENGTH, POUNDS PER SQUARE 


fc 


NATURAL 


^ S _H 


3 


INCH, WHEN IMMERSED IN 





CEMENT. 


O H~CB 




WATER OF APPROXIMATE TEMPERATURE, 


o 

1 




1||| 


s| 


DEGREES FAHR. 


1 






H? ^ 


W 9 

o 


















a 


Brand. 


Sample. 


9) 


^ 


38 


40 


50 


55 


60 


65 


70 


80 


i 


Gn 


15 R 


o 


7 days 


146 




137 


125 






126 


154 


2 









14 days 


144 




131 


125 


131 


150 


168 


208 


3 









28 days 


166 




178 




184 


. . . 


247 


280 


4 






1 


7 days 


83 


. . . 


88 


'84 


89 


98 


97 


121 


5 






1 


14 days 


84 


. . . 


111 




123 




150 


191 


6 






1 


28 days 


96 


. . . 


156 


187 


. 


221 


243 


288 


7 


Ln 


31 S 





1 day 


. . . 


143 


. . . 


124 


120 


. . 


109 


109 


8 









7 days 


. . 


204 


201 


. . . 


183 


. . . 


193 


186 


9 









14 days 


. . . 


184 


203 


. . . 


204 


. . . 


229 


245 


10 









28 days 


. . . 


221 


245 


. . . 


254 


. . 


281 


303 


11 









2 mos. 




261 


292 




348 


. 


382 


429 


12 


An 


G 


1 


7 days 




134 


140 




150 


. . . 


154 


158 


13 






1 


14 days 


. . . 


149 


162 




189 




182 


216 


14 






1 


28 days 


. . . 


198 


223 


. . . 


250 




281 


296 


15 






1 


2 mos. 


. . . 


251 


286 


. . . 


337 




386 


403 


16 






3 


14 days 


. . 


50 


58 


. . . 


69 


. . . 


73 


100 


17 






3 


28 days 


. . . 


67 


87 


. . . 


100 


. . . 


102 


157 


18 






3 


2 mos. 




104 


127 




147 





194 


231 



water, but after the first seven days, and sometimes after a 
shorter time, the strength is developed more rapidly in warm 
water, and the strength at any time between seven days and 
three months is approximately proportional to the temperature. 
After three months, the effect of the temperature seems to 
diminish, and may entirely disappear in time. 

M. Paul Alexandre * made quite a number of experiments 
on this point with Portland cement. In these experiments the 



"Recherches Experimentales sur les Mortiers Hydrauliques." 



STORING BRIQUETS 135 

gaging was done in about the same temperature as that at which 
the water of immersion was maintained, so that a double cause 
of variation was present. However, it was found that in all 
cases the higher strength was attained at seven days by the 
briquets made and stored in the higher temperature (15 to 
18 C., 60 to 65 Fahr.) while at twenty-eight days the briquets 
of the lower temperature (0 to 5 C., 32 to 40 Fahr.) were 
ahead in the case of neat cement, and nearly as high as the 
warm briquets in the case of mortar. At three months the 
differences seemed to disappear. 

194. Stale Water. Some experiments made to compare the 
strength of briquets which were alike in all other respects, but 
were immersed in different tanks in which the water had not 
been frequently renewed, showed very clearly the possible varia- 
tions from this source. Natural cement briquets, neat, and with 
one and two parts sand, gave, when immersed in one of the 
tanks, only from 40 to 60 per cent, of the strength attained 
in another tank by briquets entirely similar. 

To store briquets in running water is going to the other 
extreme; this appears to be the best method, at least for short- 
time acceptance tests, provided the temperature can be regu- 
lated. However, in some cases where this has been adopted, 
the strength of the briquets is said to have fallen off very much 
after four or five years. Whether this is due to the action of 
running water is a very interesting point, and a valuable one 
from the practical standpoint of the use of cement, but it has 
not yet been thoroughly investigated. 

195. It appears from the foregoing that variations in the 
temperature and freshness of the water in which the briquets 
are immersed is an uncertain contingent, and therefore that all 
such variations should be carefully avoided. As a matter of 
convenience, the tanks may well be maintained at 60 to 70 
Fahr., but if one does not care for a comparison of his results 
with those obtained in other laboratories, then any other con- 
stant temperature between 40 and 75 may be adopted. The 
water in the tanks should be renewed at least once a month, 
and preferably once a week. 

196. Storing Briquets in Sea Water. When the cement 
under test is to be used for constructions in the sea, some of 
the briquets should be stored in sea water to indicate the be- 



136 CEMENT AND CONCRETE 

havior in this medium. Many tests have been made in this 
way by several experimenters, but the varied results obtained 
only indicate the different effects of such treatment on different 
samples of cement. One of the effects of storing in sea water 
has been touched upon under the head of consistency of mortar, 
where it is shown that porous briquets may disintegrate in this 
medium. A small specimen like a briquet will of course be 
more quickly affected than a large mass of concrete, but on the 
other hand, the concrete in work is likely to be more porous 
than the briquet. The effect of sea water upon cement will be 
taken up in another place. 

197. Other Methods of Storing Briquets. It has been 
thought that briquets, made to test cement that is to be used 
in air, should be hardened in the same medium in order that 
the tests should more nearly approach the conditions of use. 
Several points, however, should be borne in mind in interpret- 
ing the results obtained with air-hardened specimens. In actual 
work the mortar is usually in a large mass, or is protected from 
the influence of a warm, dry atmosphere, so that it remains 
moist for a long time, whereas a briquet placed in the open 
air is much more affected by changes in atmospheric condi- 
tions. If the briquets are allowed to harden in a room, such a 
small quantity of mortar may become quite dry in a few days, 
and, unless the amount of moisture in the air is regulated, an- 
other source of variation is introduced in the tests. 

It has been found impossible to obtain uniform results from 
briquets made as nearly alike as possible and stored side by 
side in the air of the laboratory. The regular acceptance tests 
should, therefore, it is thought, be made in the ordinary man- 
ner, but if cement is to be used in locations where it is likely 
to become very dry, a few special tests should be made to assure 
one that the brand of cement in question is one that will yield 
good results in such exposure. It may be found that certain 
kinds or brands should be entirely avoided for use in such lo- 
cations. A few tests of this character are given in Tables 72 
and 73, 359, 360. The results in any given line of the table 
are from briquets made the same way but treated differently 
in the method of storing. It is seen that these brands harden 
well in dry air. The effect of the amount of water used in 
gaging appears to follow somewhat the same law, whether the 
briquets are stored in air or water. 



BREAKING THE BRIQUETS 137 

A method more nearly approaching conditions that fre- 
quently prevail in practice is to bury the briquets in damp 
sand. Table 120, 409, gives the results obtained with a large 
number of briquets stored in this way. While the results are 
somewhat more irregular than those for water-hardened speci- 
mens, since the conditions cannot be made so nearly uniform, 
yet this method gives better results than dry air storage. 

ART. 25. BREAKING THE BRIQUETS 

198. THE TESTING MACHINE. The function of the testing 
machine is simply to furnish a means of applying the tensile 
stress, and of measuring the amount of force required to break 
the briquet. Aside from the clips, which hold the briquet, 
any contrivance which may be conveniently operated, and 
which will accurately measure the force applied, may be used 
for this purpose. 

There are several forms of testing machines on the market, 
all designed on the lever principle, though differing slightly in 
the method of application. The force is applied either by al- 
lowing water or shot to run into or out of a vessel suspended 
at the end of the longer arm of a lever, or a weight is made to 
run along the lever arm, which is graduated so that the force 
applied may be read from the beam. 

199. In machines of the first class the delivery of shot is 
cut off automatically the instant the briquet breaks. The ad- 
vantage of this style is that the flow of shot may be so adjusted 
as to approximately regulate the rate of applying the 'stress; 
but little skill is required to operate it, and, since in its best 
form two levers are used, the shorter arm of one acting on the 
longer arm of the other, the machine occupies but little space. 
This machine does not permit rapid operation, since the shot 
must be weighed each time a briquet is broken. One of the 
main disadvantages of this form has been that in the case of 
strong briquets, a certain initial strain had to be applied in 
order that the stretch of the briquet and the slipping of the 
clips should not allow the shot to be cut off before the briquet 
broke. This objection, however, has recently been met by the 
makers, who have provided means of taking up this slip by a 
hand crank. 

200. Another objection urged against the short-lever shot 



138 CEMENT AND CONCRETE 

machines is the fact that as the stream of shot flowing into the 
scale pan is cut off by the breaking of the briquet, a certain 
amount of shot on its way to the pan falls into the pan after 
the briquet breaks, and is weighed, although not acting on the 
briquet at the time of the break. A form of shot machine is 
now on the market, however, in which this objection has been 
overcome. The load is applied by means of a weight hanging 
from one end of a lever. This weight is at first counterbalanced 
by a pail of shot at the other end of the lever, but as the shot 
is allowed to run out of the vessel, the unbalanced portion of 
the weight acts, through suitable levers, upon the briquet. 
The flow of shot is shut off automatically by the breaking of 
the briquet, and the shot that has escaped is weighed on a 
special scale to determine the load acting on the briquet. 

201. In the other form of machine the weight is made to 
move along the arm by means of a cord and hand-wheel. This 
style may be operated much more rapidly, but some skill is 
required to use it properly, and as now made it occupies too 
much space. These machines are preferable for laboratories, 
while the shot machines may well be used in cement factories 
and small works where a foreman does the testing. 

202. It would seem that a machine could easily be made 
which would combine the desirable features of both of these 
forms, by placing a heavy weight provided with rollers upon 
the upper lever arm of the shot machine, and using it in the 
same way that the hand power machine is now used. This 
would involve placing a hand wheel and cord upon the machine 
to operate the moving weight, the shot attachment being re- 
moved. Such a machine would combine the compactness of 
the shot machine, with the accuracy and speed of the single 
lever machine; the graduations on the beam could represent 
five pounds each, instead of two pounds, the value of the grad- 
uations now on the single lever machines. 

203. FORMS OF CLIP. Since cement has been tested by 
tensile strain, it has ever been a problem to obtain a clip which 
would give a perfectly true axial pull on the briquet. Various 
forms of clips have been used from time to time, but none of 
them has proved satisfactory in all respects. To trace the his- 
tory of the development of the clip is not warranted by its 
interest, but it may be said that in some of the early forms the 



BREAKING THE BRIQUETS 



139 



head of the briquet was held between two plates and clamped 
tight enough to develop sufficient friction to transmit the stress. 
The later forms of briquets are made with a shoulder or with 
wedge-shaped ends to allow the clip to grasp them. Mr. John 
Grant, Mr. Alfred Noble, General Gilmore, Mr. J. Sondericker 
and Mr. D. J. Whittemore have each designed or adapted dif- 
ferent forms, and more recently Mr. S. Bent Russell and Mr. 




I" 3 /f Vi K I 2 

FIG. 6. RIEHLE "ENGINEERS' STANDARD" CLIP 

W. R. Cock have each devised a clip which will be mentioned 
below. 

204. Form in Most General Use. The clip in most general 
use in the United States is of the general style shown in Fig. 6. 
It differs only in detail from the form recommended by the 
Amer. Soc. C. E. Committee of 1885, which has been called 



140 CEMENT AND CONCRETE 

the "Engineers' Standard. 7 ' The general form is pear shaped; 
the briquet is grasped at the points of reverse curve at the side 
of the briquet, giving an area between opposite gripping points 
of about one and a quarter square inches. The gripping points 
are rather too sharp, when new, as they have a tendency to 
crush the briquet locally. The width of the bearing increases 
with the amount of wear the clip sustains. The clip is provided 
with a conical pivot, which rests in a cone-shaped cavity at- 
tached to the machine, so that the two parts of the clip are 
free to swing. In a form which was previously used to a con- 
siderable extent, each bearing surface was designed to be about 
an inch square, the jaw being made to conform to the outline 
of the briquet. This form, however, did not give satisfactory 
results; a particle of sand between the briquet and the bearing 
surface of the clip would give an eccentric pull, and strong 
briquets would sometimes break in the head of the briquet 
transverse to the axis, in several curved layers joining opposite 
gripping surfaces. 

205. CLIP-BREAKS. When a briquet is inserted in the or- 
dinary clip, the gripping points will not, in general, grasp the 
briquet symmetrically. The gripping points have a tendency 
to slide on the surface of the briquet in order to assume a sym- 
metrical position; there is friction to resist this sliding, and 
when this resistance overcomes the tendency to motion, the two 
clips and the briquet become a rigid system, and bending strains 
may be introduced. Again, if the briquet is not too badly ad- 
justed in the clips, it is apt to break in a line joining two op- 
posite gripping points, instead of at the smallest section; this 
is called a " clip-break." The tendency to form clip-breaks is 
greater if the gripping points are very narrow or have sharp 
edges; neat cement briquets exhibit this tendency much more 
than briquets from sand mortars, and some samples of cement 
are much more likely to give clip-breaks than others. 

206. Cause of Clip-breaks. When a briquet breaks in this 
manner, the broken section is usually about normal to the side 
of the briquet at the point where the jaw was in contact. This 
indicates that a clip-break is caused by compression at that 
place: there is evidently compression along the plane joining 
the two opposite gripping points, and tension at right angles 
to that plane, and the briquet fails here as a result of the two 



BREAKING THE BRIQUETS , 141 

stresses. If the briquet is not properly adjusted in the clips, 
but is so placed that its longest axis is at one side of the line 
joining the points of application of the forces (in the " Engi- 
neers' Standard" clip, the line joining the pivot points), then 
the bending strain that is introduced is greatest at the central 
section of the briquet; this may cause the briquet to break at 
the smallest section, when if it were properly adjusted in the 
clips it would develop a clip-break. The bearing surfaces of 
the clip should not be too small, as this increases the intensity 
of pressure, but on the other hand there appears to be no prac- 
tical advantage in making this area more than T 3 F to inch 
wide (the length being limited by the thickness of the briquet, 
one inch). 

207. Prevention of Clip-breaks. The method most fre- 
quently adopted to prevent clip-breaks is to cushion the grip- 
ping points with some compressible material, such as thin rub- 
ber or blotting-paper. This device prevents clip-breaks, but 
the result of about three hundred tests made under the author's 
direction showed clearly that it also lowered the apparent 
strength very materially. 1 Briquets broken with the bare clips 
showed a mean strength of 606 pounds per square inch, while 
the cushioned clips gave an apparent strength of but 521 pounds, 
or 86 per cent, of the strength without the cushion; of the bri- 
quets broken with the bare clips, 33 per cent, were clip-breaks; 
with the cushioned clips no clip-breaks occurred. The rubber 
was applied by slipping two rubber bands over each end of the 
briquet, giving cushions about T V inch thick. 

208. Strength of Briquets that Develop Clip-breaks. It was 
also found in breaking 277 briquets with two styles of clips 
without cushions that 129 of them that gave clip-breaks aver- 
aged 611 pounds per square inch, while 148 which did not de- 
velop clip-breaks had a mean strength of 590 pounds. This 
result is easily accounted for by saying that some of the bri- 
quets that broke in the small section were made to do so by the 
cross-strain introduced by imperfect adjustment in the clips. 

When a briquet breaks at other than the smallest section, it 
is certain that the smallest section has a greater strength per 

1 For a report of these tests in detail, see Annual Report Chief of Engi- 
neers, U. S. A., 1895, p. 2913. Also "Municipal Engineering," Dec., 1896, 
Jan., Feb., 1897. 



142 



CEMENT AND CONCRETE 



square inch than is shown by the result obtained; how much 
greater cannot be told. But it follows that if clip-breaks could 
be eliminated in a proper way, one which would not cause center 
breaks by the introduction of cross-strains or other undesirable 
conditions, the strengths thus obtained would be greater than 
when clip-breaks occur. The fact that the use of a rubber 
cushion gives lower strengths, shows that this is not the proper 
method of preventing clip-breaks. 

209. Mr. W. R. Cock has devised a clip, with rubber-covered 
gripping points, which has attracted some attention. It has 




I 2 

FIG. 7. RUSSELL CLIP 



3 in. 



sometimes been assumed that because this clip eliminated clip- 
breaks it must give a higher apparent strength than the rigid 
form. No extensive series of experiments have been published 
which permit of comparing this clip with other forms, but 
from the results obtained above, in using rubber cushions, it 
would appear that the Cock clip may give lower apparent 
strengths. 



BREAKING THE BRIQUETS 



143 



210. The form of clip designed by Mr. S. Bent Russell is 
constructed on the "evener" principle, each clip having free- 
dom of motion imparted by four pin-connected joints (see Fig. 
7). It is sought to prevent any but an axial pull being ap- 
plied to the briquet. On account of details of construction, 
into which it is not necessary to enter here, the clip must be 
in its normal position when the briquet is inserted, in order 
that the possibility of cross-strain shall be effectually removed. 
As a result of many tests with this form and the ordinary "En- 




r 3 A te * o z 

FIG. 8. SINGLE GIMBAL CLIP 



din. 



gineers' Standard," it was found that they gave very nearly 
the same strength. But that the evener motion itself was of 
some value was shown by a series of experiments in which part 
of the briquets were broken by this form of clip without modi- 
fication, while part were broken by the same clip when it had 
been changed to a rigid form by means of a clamp that elimi- 
nated the evener motion. It is believed that with some modifi- 
cations this clip will give good results, and it may be used al- 
most as rapidly as the ordinary rigid form. 



144 CEMENT AND CONCRETE 

211. Several experiments were made with a clip in which 
the gimbal principle was applied, the stress passing from the 
machine to the gripping points through knife edges placed in 
the line joining opposite gripping points and midway between 
them 1 (Fig. 8). Higher results were obtained with this form, 
the " Single Gimbal," than with any of the styles with which it 
was compared, but it was made only for experimental purposes, 
and unless modified is not convenient enough to be recom- 
mended for general use. 

212. In the course of these experiments it was shown that to 
increase the distance between gripping points, grasping the bri- 
quet nearer the head, increased the apparent strength and 
diminished the number of clip-breaks. With the Russell clip, 
increasing this distance from IfV inches to l T 7 g- inches gave an 
increase of about six per cent, in the apparent strength; and a 
similar increase in the width between jaws of the Gimbal clip, 
from ly\ to 1 T %- inches, gave an increase in apparent strength 
of about five per cent. It was found later that Mr. J. Son- 
dericker had previously arrived at similar results, 2 and as the 
form of briquet used by the latter had permitted extending 
the experiment, he found that when the points were about If 
inches apart (making the area of the briquet about If square 
inches between opposite gripping points), nearly all the frac- 
tures occurred at the smallest section. 

213. Effect of Improper Adjustment. The effect of not 
properly adjusting the briquets in the clip was also investigated. 
In some cases the briquets were placed in the proper position 
as nearly as possible. In the other cases they were in a de- 
cidedly distorted position, much worse than they would be 
placed with the most careless manipulation. It was found that 
if the briquet was so placed in the " Engineers' Standard" 
clip that the gripping points on one side of the briquet were 
farther apart than those on the other side, the decrease in- break- 
ing strength was very marked (about 35 per cent.), while if the 
planes determined by the lines of contact of the gripping points 
of each clip were parallel, there appeared to be no effect. The 



1 This clip was devised by the author at the suggestion of Mr. E. S. 
Wheeler, M. Am. Soc. C. E. 

2 Jour. Assoe. Eng. Soc., Vol. vii, p. 212. 



BREAKING THE BRIQUETS 145 

reason of this is evident: in the former case the line of force, 
joining the two pivot points, does not pass through the center 
of the smallest section of the briquet, and transverse stresses 
are introduced, while in the latter case the line of force does 
pass through the center of the smallest section, though not at 
right angles to its plane. With the Russell and Gimbal clips 
the distortion seemed to have little effect, provided, that in 
the case of the former, the clip was itself in its normal position 
when the briquet was inserted. 

214. Conclusions Derived from Tests of Several Styles of 
Clips. From the tests described above, 1 the following conclu- 
sions may be drawn : - 

1st. When using the ordinary form of clips with metal 
gripping points, the briquets which break at the places of con- 
tact of the jaws give higher apparent strengths than those 
which break at the smaller sections. 

2d. A rubber cushion between the briquet and the jaw of 
the clip prevents clip-breaks, but materially lowers the stress 
required to break the briquet. 

3d. The form of clip designed by Mr. S. Bent Russell gives 
somewhat less irregular results than are obtained with the 
Riehle " Engineers' Standard" rigid clip. Although the results 
given by the Russell clip in its present form are a trifle lower 
than those given by the Riehle, it seems probable that these 
lower results are due to defects in detail which may readily 
be eliminated. 

4th. By the application of the gimbal principle to cement 
testing clips, higher, as well as more nearly uniform, results 
may be obtained. 

5th. In using the rigid form of clip, careless manipulation 
in adjusting the briquet may result in serious error due to the 
introduction of cross-strains, while with either the Single Gim- 
bal or Russell clip slight deviations in adjustment are not im- 
portant. 

6th. With the form of briquet recommended by the commit- 
tee of the American Society of Civil Engineers in 1885, the break- 
ing stress may be somewhat increased, and the number of clip- 



1 These tests were described in greater detail and discussed by the 
writer in "Municipal Engineering," Dec., 1896, Jan. and Feb., 1897. 



146 



CEMENT AND CONCRETE 



breaks may be very materially decreased, by such a modifica- 
tion of the clip as to allow grasping the briquet nearer the head. 

215. Requirements for a Perfect Clip. As a logical result 
of these conclusions, the ideal clip should fulfill the following 
requirements : - 

1st. It should impart a true axial pull to the briquet with- 
out subjecting it either to cross-strains or to compressive forces 




FIG. 9. FORM OF ARTICULATED CLIP SUGGESTED FOR USE 

sufficient to cause it to break at other than the smallest section. 
2d. The bearing surfaces of the gripping points should not 
be more than about one-fourth of an inch wide, since this is 
sufficient to prevent crushing the briquet at these places, and 
too wide a jaw will not usually bear uniformly over its whole 
surface. 



BREAKING THE BRIQUETS 



147 



3d. Its parts should have sufficient strength and stiffness, 
so that they will not bend appreciably when in use. 

4th. It should permit rapid operations, and 

5th. It should be as light as consistent with the above 
requirements. 

216. Form Suggested. Fig. 9 shows a style of clip which 
closely conforms to the above specifications. The evener 
form devised by Mr. Russell has been selected for modification. 
The S. G. clip would more nearly meet some of the requirements, 
and, so far as the principle is concerned, this form is considered 
quite the equal of the evener clip. But no method of applying 
the gimbal principle has commended itself as affording such 
rapid manipulation as does the evener motion, and since it is 
thought that either form will obviate cross-strains in a plane 
parallel to the face of the briquet, the evener form has been 
adopted on account of convenience. 

The defects in detail of the Russell clip which have already 
been mentioned have been obviated in the present form. The 
gripping points are made one-fourth of an inch wide, and a little 
more material has been used between the gripping points and 
the first pin to stiffen the clip. This form is designed for use 
with the briquet shown in Fig. 5 (see 179). 

217. RATE OF APPLYING THE TENSILE STRESS. Table 37 
gives the results of several hundred experiments made by Mr. 

TABLE 37 
Relation of Apparent Tensile Strength to Rate of Applying Stress 



RATE OF APPLYING 


TENSILE STRENGTH 


I 


STRESS, 


OBTAINED, 


POUNDS 


PER MINUTE. 


POUNDS PER SQ. INCH. 




50 


400 




100 


415 




200 


430 




400 


450 




6,000 


493 



Henry Faija l to show the effect on tensile strength of varying, 
the rate of applying the stress. 

A few of the results obtained from nearly 900 tests, made 



"Cement for Users," by Mr. Henry Faija. 



148 



CEMENT AND CONCRETE 



under the author's direction to illustrate this point, are given 
in Table 38. Some of these results accord very well with those 
given in Table 37, but the results in the latter table were doubt- 
less obtained from neat Portland briquets only, while the ex- 
periments given in Table 38 were made with briquets neat 
and with two parts sand, and on natural as well as Portland 
cement mortars. 

TABLE 38 

Relation of Apparent Tensile Strength to the Rate of Applying the 

Stress 









TENSILE STRENGTH, POUNDS PER 


CEMENT. 


PROPORTIONS. 


AGE OF 
BRIQUETS. 


SQUARE INCH, FOR STRESS APPLIED AT 
BATE OF POUNDS PER MINUTE. 








100 


300 


500 


700 


900 


Portland 


Neat cement 


7 and 14 days 


453 


485 


621 


520 


528 


it 


Neat cement 


3 months 


. . . 


590 


617 


622 


640 


t{ 


1-2 


3 months 


445 


467 


487 


507 


510 


Natural 


Neat cement 


7 days 


150 


169 


186 




202 





Neat cement 


3 months 


309 


351 


363 


378 


390 





1-2 


3 months 


255 


299 


327 


329 


354 



218. It appears from all these results that the increase in 
the breaking strength due to increasing the rate of applying 
the stress is considerable in the case of low rates of speed, but 
when a rate of 500 or 600 pounds per minute has been reached, 
a further increase in rapidity does not make a material increase 
in the apparent strength. Since certain variations in rate are 
sure to occur, until some device is used to automatically regu- 
late it, a rate should be adopted which would allow of slight 
variations without materially changing the result of the test. 
A rate of 600 pounds per minute would fulfill this requirement, 
and, with certain machines at least, would be still more con- 
venient than the rate of 400 pounds per minute which has here- 
tofore been quite generally used. 

An analysis of the experiments made to determine the de- 
gree of uniformity obtained by using each of the given rates, 
showed there was but little difference in this regard, but if 
any choice could be made on this basis it seemed to lie with 
the more rapid rate. 

219. With the shot machines it is not difficult to approxi- 



BREAKING THE BRIQUETS 149 

mately regulate the rate at which the stress is applied. In 
operating a machine in which a handwheel moves a weight 
along the graduated beam, it must be remembered that the rate 
at which the weight moves is the controlling factor, and not 
the movement of the lower wheel, which simply serves to take 
up lost motion, the stretch of the briquet under strain, and 
the slipping of the briquet in the jaws of the clip. A mistaken 
idea concerning this matter has sometimes led to the adoption 
of a device to regulate the motion of this lower wheel. Until 
one is accustomed to applying the stress at a given uniform 
rate, he will find it an aid to hang near the machine a pendulum 
of such a length that a certain number of vibrations correspond 
to a complete revolution of the handwheel. 

220. Treatment of the Results. The number of briquets 
which are made to test the strength of a given sample of cement 
will depend on the accuracy which it is desired to attain. If 
but two briquets are made, neither of the results may be re- 
jected; however widely they may differ one from the other, 
the mean of the two must be considered the result of the experi- 
ment when nothing is known as to their comparative value. 
But if several briquets are made from the same sample, and 
they vary one from another, the final result is sometimes ob- 
tained by rejecting certain of the observations. In some cases 
if five or six specimens are made, the highest and the lowest 
ones are omitted, while sometimes the two lowest are rejected, 
and the mean of the three or four highest is taken. 

221. While the absolute mean of all of the observations will 
ordinarily be quite sufficient, and should usually be considered 
the result of the test, yet where tests are very carefully made 
to compare two samples, or two methods of manipulation, it 
may be desired to reject certain observations that appear to 
be abnormal. The beginner in cement testing, unfamiliar with 
observations of this character, may not feel confidence in his 
own judgment as to what observations may be rejected, and the 
criteria sometimes used in more accurate work are entirely too 
complicated for this purpose. To serve as a guide in such 
cases, the writer would suggest the following simple method 
which, though entirely arbitrary, is more justifiable than either 
of the methods mentioned above. As the experimenter be- 
comes more familiar with the work, he will doubtless prefer to 



150 



CEMENT AND CONCRETE 



depend on his own judgment in the rejection of observations, 
taking into account the general accuracy of the work. 

First obtain the absolute mean and the difference between 
this mean and each individual result; let us call this difference 
the " error" for each result. Reject any observations whose 
error is, say, ten per cent, of the absolute mean, and obtain the 
mean of the remaining observations as the true result. 

222. For example, suppose that we have broken ten bri- 
quets obtaining the strengths given below, and wish to deter- 
mine the result of the test. The absolute mean is found to be 
213.9 pounds, or, the nearest whole number, 214 pounds. 

TABLE 39 
Rejection of Observations 



NUMBER 

OF 

BRIQUETS. 


OBSERVED 
STRENGTH. 


ERROR. 


OBSERVED 
STRENGTH. 


NEW 
ERROR. 


1 


209 


5 


209 


1 


2 


226 


12 


226 


16 


3 


227 


13 


227 


17 


4 


184 


30 


. . . 


. . 


5 


217 


3 


217 


7 


6 


252 


38 






7 


200 


14 


200 


10 


8 


195 


19 


195 


15 


9 


193 


21 


193 


17 


10 


236 


22 


. . . 





Sum . 


2,139 


177 


1,467 


83 


Mean 


213.9 


17.7 


209.6 


11.9 



The " errors" are given in the third column, and it is seen 
that three of them are greater than ten per cent, of the mean. 
Omitting the results having these large errors, we obtain a new 
mean of 209.6 pounds, which is to be considered the result of 
the test. An inspection of the first column of errors shows that 
the mean of the errors is 17.7 pounds; if we divide this by the 
mean of the tensile strengths, we obtain 17.7 -* 213.9 = .0827. 
Expressing this as a percentage, we may call 8.27 per cent, 
the "average error." The same result is, of course, obtained 
by dividing the sum of errors by the sum of the strengths. 
Now if we consider column five, we see that the new average 
error will be but 83 *- 1467 = 5.66 per cent. 



INTERPRETATION OF TESTS 151 

223. In giving the results of a series of tests, it is a common 
practice to state only the absolute mean, but it is of considerable 
interest to know the variations that occurred in breaking in 
order that one may judge of the reliability of the results, or, 
in other words, to make a rough approximation as to the prob- 
able error. For this purpose the highest and lowest result may 
be given, but a much better index to reliability would be to 
give the " average error" as explained above. However, in 
reporting a large number of tests, the extra labor involved in 
obtaining this " average error" is usually considered too great 
to be attempted, and in such cases the absolute mean and the 
highest and lowest results must serve the purpose. 

224. Accuracy Obtainable. When an operator has become 
expert and is working under good conditions, he may expect 
to obtain results within the following limits: The extreme varia- 
tions between the results in a set of ten briquets (the difference 
between the highest and lowest) not exceeding 20 per cent, of 
the mean strength of the set, the maximum variation from the 
mean not exceeding 12 per cent, of the mean, and the " aver- 
age error," as explained above, not exceeding 8 per cent. 

ART. 26. THE INTERPRETATION OF TENSILE TESTS OF 
COHESION 

225. One of the problems presented in the inspection of 
cement is to foretell the ultimate relative strengths of two 
samples from the results of short time tests. Formulas have 
been presented purporting to solve this problem, such formulas 
being based on the assumption that the strength gained at the 
end of months or years is a function of that developed in a few 
days. In fact, the raison d'etre of tensile or other short-time 
strength tests for the acceptance of cement, rests, in a sense, 
upon this same assumption. 

The value of strength tests as one of the guides in determin- 
ing in a short time the probable quality of a cement is unques- 
tioned. One is apt, however, to seek too close an agreement 
between the results of such tests and the actual quality of the 
cement. It would be easy to select examples illustrating the 
harmony between short and long time tests; but it will be of 
greater value to show, rather, some of the many exceptions to 
such a rule, and thereby emphasize the fact that it is only by 



152 



CEMENT AND CONCRETE 



a close analysis of all of the information obtainable concerning 
a sample, and a general knowledge of the behavior of the dif- 
ferent grades of cement, that one may hope to arrive at a tol- 
erably accurate opinion. 

226. Comparative Tests of Portland Cements. In Table 
40 are given the results of tests on four brands of Portland 
cement at seven days, twenty-eight days and two years. From 
the tests at two years it appears that T and U are the best 
cements, V is nearly as good, but W gives a much lower result. 
Turning now to the seven and twenty-eight day tests of bri- 
quets maintained at the ordinary temperature, it is seen that 
W gave in every case higher results than T, and nearly as high 
as U or V. Among the short time tests it is only the results 
of briquets maintained at 80 C. that indicate the inferiority 
of Brand W. 

TABLE 40 

Interpretation of Short Time Tests of Portland Cement, Several 

Brands 









TENSILE STRENGTH, POUNDS 


PARTS 
SAND TO 1 
CEMENT 

BY 


TEMPERA- 
TUBE WATER 

OF 


AGE OF 
BRIQUETS. 


PER SQUARE INCH. 


Brand. 


WEIGHT. 


IMMERSION. 


















T 


U 


V 


W 


2 


Hot, 80 C. 


7 days 


339 


278 


284 


222 


3 


I. U 




221 


191 


180 


134 


3 


" 60 C. 




144 


142 


169 


144 





Ordinary 




426 


510 


487 


565 


1 






327 


425 


400 


396 


2 






172 


275 


256 


236 


3 






73 


150 


160 


150 


1 




28 days 


526 


577 


557 


556 


2 






312 


394 


387 


332 


3 






142 


241 


223 


247 


1 




2 years 


719 


753 


763 


654 


2 






554 


553 


513 


407 


3 






386 


373 


340 


287 



227. Comparative Tests of Natural Cements. From the 
nature of natural cements a much greater variation in strength 
among different brands, and even among different samples of 
the same brand, is to be expected. With Portland cements 
made in accordance with ordinary methods, the variations in 
strength among ten or twenty brands will usually be compara- 
tively small. One of them may possibly prove unsound, and 



INTERPRETATION OF TESTS 



153 



one or two others may give inferior strength, but the variations 
in strength among three-fourths of the samples will not gener- 
ally exceed 20 per cent. With the same number of brands of 
natural cements, variations of 50 to 200 per cent, may be 
expected. 

TABLE 41 

Interpretation of Short Time Tests of Natural Cement, Several 

Brands 









TENSILE STRENGTH, POUNDS 








PER SQUARE INCH. 


PABTS 
SAND TO 1 
CEMENT. 


TEMPERA- 
TURE WATER 

OF 

IMMERSION. 


AGE OF 
BRIQUETS. 




Brand. 




















Jn 


Hn 


Bn 


Mn 


Nn 


Kn 


2 


Hot, 50 C. 


7 days 


152 


192 


84 


133 


160 


277 


2 


Hot, 60 C. 




170 


270 


79 


154 


164 


254 


2 


Hot, 80 C. 




58 


136 


128 


179 


166 


221 





Ordinary 




174 


203 


130 


189 


210 


189 


1 






125 


198 


103 


164 


169 


164 







28 days 


208 


344 


293 


203 


316 


289 


1 






237 


342 


247 


247 


252 


385 


2 






132 


223 


148 


158 


184 


217 


3 






64 


113 


85 


93 


104 


101 


1 




2 years 


177 


271 


358 


631 


665 


532 


2 






106 


157 


195 


515 


550 


561 


3 






99 


130 


117 


340 


328 


372 



In Table 41 six brands of natural cement are compared by 
tests at seven days, twenty-eight days and two years. These 
six brands have been arranged in the table according to their 
value as shown by the two year tests, and it is seen that the 
first three, Jn, Hn and Bn, are especially poor, while the last 
three, Mn, Nn and Kn, are exceptionally good. In the short 
time tests of briquets maintained at ordinary temperature, Jn 
and Bn gave low results and Nn and Kn gave fairly high results, 
in harmony with the long time tests; but Hn, which proved to 
be one of the poorest samples, gave in every case the highest, 
or next to the highest, result in seven and twenty-eight day 
cold tests. In this table we find again that the results of the 
briquets maintained at 80 C. for seven days gave, in a general 
way, the best indication of the relative values of the six brands. 

228. Several Samples of One Brand. To show that short 
time tests do not always indicate the relative values of several 
samples of cement, even when all of the samples are of the 



154 



CEMENT AND CONCRETE 



same brand, Tables 42 and 43 are given. All of the results in 
these tables are from samples of the one brand of natural ce- 
ment. 

TABLE 42 

Comparison of Short and Long Time Tests of Samples of One 
Brand of Natural Cement 



SERIES. 


SAND. 


AGE. 




TENSILE STKENGTH, LBS. 
PER SQUARE INCH. 


Kind. 


Parts to 
1' Cement. 


A 


. . . 






28 days 
6-7 months 


Number 
of Samples 
Tested. 


3 


7 


2 


3 


5 





84 
121 


123 
186 


177 
241 


220 
301 


297 
381 


B 


Std. 
(i 


1 
1 


7 days 
6 months 


Number 
Samples. 


17 


20 


17 


16 







62 
462 


74 

468 


86 
442 


146 
367 




C 


P.P. 

u 
II 


1 

1 and 2 
2 


7 days 
6 months 1 
7 days 2 


Number 
Samples. 


50 


50 


19 


19 







49 
473 
273 


54 
426 
249 


73 
381 
265 


128 
321 
283 





D 


p.p. 




2 

2 


7 days 
1 year 
7 days 2 


Number 
Samples, 


13 


48 


38 


18 







66 
473 
257 


80 
422 
234 


95 
377 

277 


147 
325 
215 


. . 


E 


. . . 



2 


7 days 
6 months 


Number 
Samples. 


12 


21 


18 


9 






74 
535 


83 

477 


120 
424 


167 
373 





F 


C Cr.Qtz! ) 

\ 20 to 40 } 



2 


28 days 
( 6 months ) 
( and 1 year f 


Number 
Samples. 


287 


170 


41 









135 
565 


191 
454 


235 
367 









1 Mean one-to-one and one-to-two mortars. 

3 Briquets immersed six days in water maintained at 60 C. 



INTERPRETATION OF TESTS 155 

In Table 42 the results are selected from a large number of 
tests of this brand, and are arranged in groups according to the 
strength shown at a certain age. For instance, in Series A 
the results of twenty samples are given, arranged according to 
the strength at twenty-eight days. Three of the samples gave 
less than 100 pounds per square inch, neat, at twenty-eight 
days; the same three samples gave a mean strength of 121 
pounds per square inch, neat, six to seven months. Seven 
samples, the strength of which fell between 100 and 150 pounds 
at twenty-eight days, gave a mean strength of 186 pounds at 
six to seven months. The results of this series show the 
harmony between short and long time tests when it is a question 
of comparing neat cement mortars. 

In Series D of this table the samples are arranged in order 
according to the strength developed by one-to-two mortars 
one year old. Thirteen samples had a strength at this age of 
between 450 and 500 pounds, average 473 pounds. The same 
samples gave but 66 pounds, neat, seven days. Forty-eight 
samples, giving between 400 and 450 pounds, average 422 
pounds, gave but 80 pounds, neat, seven days, while eighteen 
samples that developed only 300 to 350 pounds mean, 325 
pounds at one year, showed a mean strength of 147 pounds, neat, 
seven days. 

A little study of this table will show that the samples which 
were comparatively weak in seven and twenty-eight day tests, 
either neat or with sand, gave the best results in the long time 
tests of sand mortars. Series A shows that the neat tests at 
seven days and at six months are consistent, but in all cases 
where sand mortars are tested at six months to one year, the 
highest results are given by the samples showing the lowest 
strength in the short time tests in cool water. It is very sel- 
dom that this conclusion has not been indicated by the author's 
tests of this brand. It is not invariably true, however, for 
some samples which were selected as being defective in burn, 
gave low results both in short and long time tests. The con- 
clusion stated above must therefore be understood to have 
limits even for this brand, and may not apply at all to many 
brands. 

As to the results of short time tests of briquets stored in hot 
water, Series C and D indicate that such results are more nearly 



156 



CEMENT AND CONCRETE 



consistent with the long time tests, yet it is evident that even 
with hot tests one could not readily and accurately differen- 
tiate the best from the mediocre samples. 

TABLE 43 

Natural Cement: Rate of Increase in Strength, Hardening in Water 

and Dry Air 







TENSILE STRENGTH PER SQ. IN., OF SAMPLES. 


SAND, PARTS 
TO ONE 
CEMENT. 


AGE OF 
BRIQUETS 
WHEN BROKEN. 


Hardened in Water. 


Hardened in Air of 
Room. 






84 


U' 


0' 


84 


U' 


0' 


1 


7 days. 


74 


53 


103 


107 


68 


187 


1 


28 days. 


228 


189 


228 


188 


95 


256 


1 


3 mos. 


415 


346 


331 


158 


100 


248 


1 


6 inos. 


506 


381 


307 


425 


161 


359 


1 


2 years. 


446 


383 


209 


151 


147 


403 


3 


28 days. 


99 


97 


64 


112 


61 


180 


3 


3 mos. 


244 


241 


129 


153 


81 


194 


3 


6 mos. 


255 


232 


162 


92 


69 


173 


3 


1 year. 


274 


264 


186 


229 


70 


144 


3 


2 years. 


258 


268 


167 


274 


152 


228 



Sample 

Fineness : Per cent, passing Sieve No. 120, Holes 

.0046 inch square 

Time Setting to bear^/' ^lb. Wire, min. . . . 
Specific Gravity 



84 



0' 

9.7 



80.5 87.8 
54 23 97 
3.012 2.950 3.145 



U', underburned, 0', overburned. All samples same brand, Gn. 

229. The results in Table 43 will serve to illustrate the same 
point by showing the very different rates of increase in strength 
of three samples when the briquets are stored in water and in 
dry air. One of these samples, 84, was taken at random from a 
shipment, while U' and O' were supposed to be defective in 
burn. Of the water-hardened specimens, No. 84 gained in 
strength up to six months or one year and then suffered only 
a slight falling off. The underburned sample showed a con- 
tinuous gain, but the overburned cement showed a marked 
decrease in strength after six months or one year. The air- 
hardened specimens were very irregular in strength, but the 
underburned sample gave very low results throughout. 

Table 44 gives similar results obtained with several samples, 
the briquets being hardened in water as usual. 16 R is a fair 



INTERPRETATION OF TESTS 



157 



sample of the best cement of this brand, and its rate of increase 
in strength with one to three parts sand is shown. Samples 
M and L were tested together, as were CC and DD. M and 
CC are of the class giving comparatively high results at seven 
days, while L and DD give high results at seven days, but 
develop only a moderate ultimate strength. 

TABLE 44 

Natural Cement: Difference in Rates of Increase in Strength of 
Several Samples of the Same Brand 



IlEFEKENCE. 


CEMENT. 


SAND. 


TENSILE STRENGTH, POUNDS PER SQ. IN. 

AT A(iE OF 


d 

q 

c3 

m 


^3 
'a 

ED 


Kind. 



3^ 
3 s'Sci 
*2-3 

S 




1 
a 
t- 




>i 

oo 

C-J 

142 
101 
73 

199 
88 
155 
150 
232 
218 
185 
189 


02 

I 

<N 




CO 


5 

Q 

to 


i 

0> 

>> 


1 
>> 
<N 


E 

i 

>> 

00 


1 

2 
3 
4 
5 
6 
7 
8 
9 
10 
11 


Gn 


16 R 

( t 

M 
L 
M 
L 
CC 
DD 
CC 
DD 


Crushed Qtz. 20-30 
it 

1 1 

Point aux Pins 


1 
2 
3 


2 
2 
1 
1 
2 
2 


94 
59 

iis 

40 
63 
30 
123 

77 


334 

289 
204 


399 
341 
243 

256 
148 
216 
296 
276 
327 
268 
326 


430 
335 
252 
248 
146 
241 
415 
269 
337 
242 
303 


500 
386 
268 
300 
167 
252 
369 

279 
373 


445 
354 

262 


248 










317 
474 
279 
359 





230. Conclusions. From the above tables one should not 
draw the conclusion that all strength tests are valueless be- 
cause likely to be misleading. Some lessons, however, seem to 
be plain; conclusions drawn from the results of short time tests 
of strength alone are likely to be far from infallible. This is 
especially true of natural cements. The correctness of one's 
conclusions concerning the value of a sample is likely to de- 
pend very much upon his knowledge of the behavior of that 
particular brand, and the beginner in cement testing should 
not have too great confidence in his early conclusions. Samples 
under inspection should be tested in comparison with other 
samples of known quality, and the results of the strength tests 
studied in connection with all the information obtainable from 
the other tests of quality already outlined. 



CHAPTER X 

THE RECEPTION OF CEMENT AND RECORDS OF TESTS 

ART. 27. STORING AND SAMPLING 

231. STORAGE. The storage houses provided for the ce- 
ment should be such as will effectually preserve it from damp- 
ness, the floor being dry and strongly built. A circulation of 
air under the floor will insure dryness. 

In building houses for storage, due regard should be given 
to the ease of getting the cement in and out, and facilities pro- 
vided for the use of block and tackle in tiering. 

When the cement is received, whether in sacks or barrels, it 
should, if possible, be so tiered in the warehouse that any pack- 
age is accessible for sampling. In the case of barrels this may 
readily be attained by tiering in double rows, the barrels lying 
on the side. It has been found that ordinary cement barrels 
will withstand the pressure if tiered five high with a "binder" 
row on top; and when so piled, a warehouse 32 feet wide and 
100 feet long will readily hold 2,200 barrels, an allowance of 
about one hundred fifty square feet of floor space for one hun- 
dred barrels. 

232. Where storage space is limited, the barrels may be 
numbered and sampled before they are placed in the warehouse, 
and they may then be piled solid, but this should be avoided 
if practicable. Sacks cannot be quite so neatly stored, and 
since a smaller quantity is contained in a sack, they may be 
tiered so that every third or fourth sack is accessible. It is 
desirable where work is executed with the greatest care that 
every package be numbered for future identification, but this 
may sometimes prove impracticable, especially when the ce- 
ment is in sacks, and in such cases the sampled packages only 
may receive numbers. 

233. Percentage of Barrels to Sample. The amount of ce- 
ment which shall be accepted on the test of a single sample 
must be determined by each user of cement according to his 

158 



STORING AND SAMPLING 159 

knowledge as to the uniformity and reliability of the brand in 
use, and according to the character of the work in which the 
cement is to be used. In a few isolated cases every barrel is 
tested, while sometimes several tons of cement are accepted on 
a single test. As the improvements in methods have decreased 
the work involved in making the simpler tests, the tendency 
has been to test a larger percentage of the packages. 1 

The report of the committee of the Amer. Soc. C. E. in 
1885, contains the following concerning sampling: " There is no 
uniformity of practice among engineers as to the sampling 
of the cement to be tested, some testing every tenth barrel, 
others every fifth, and others still every barrel delivered. Usu- 
ally, where cement has a good reputation, and is used in large 
masses, such as concrete in heavy foundations, or in the back- 
ing or hearting of thick walls, the testing of every fifth barrel 
seems to be sufficient; but in very important work, where the 
strength of each barrel may in great measure determine the 
strength of that portion of the work where it is used, or in 
the thin walls of sewers, etc., every barrel should be tested, 
one briquet being made from it." 

234. Taking the Sample. The sample should be taken in 
such a manner as to fairly represent the package, and for this 
purpose a " sugar trier" may be used, by which is obtained a 
core of cement about one inch in diameter and eighteen inches 
long. As any tool used for boring cement barrels soon becomes 
dull, and as a sugar trier is somewhat difficult to sharpen, the . 
author prefers to use an ordinary bit and brace to penetrate 
the barrel head, and then extract the sample with a "trier," 
or a long, slender scoop of similar form provided with a handle. 

For storing the sample until it is tested, it has been found 
convenient to use covered tin cans holding about one pint, 
the cover of the can being labeled with the number of the pack- 
age from which the sample is taken. 



1 In a paper read before the Institution of Civil Engineers in 1865-66, Mr. 
John Grant states that " after using, during the last six years, more than 
70,000 tons of Portland cement, which has been submitted to about 15,000 
tests, it can be confidently asserted that none of an inferior or dangerous 
character has been employed in any part of the work in question. " (The 
Metropolitan Main Drainage, London.) This is an average of one test to 
twenty-five barrels. 



160 CEMENT AND CONCRETE 

ART. 28. RECORDS OF TESTS 

235. Value of Records. In conducting work in which the 
use of cement enters as a prominent factor, it is not only neces- 
sary to know that the cement used is of a good quality, but also 
to be able to show at any future time what tests were made to 
establish its value. This fact, as well as the convenience of 
the work, demands that a record shall be kept of all the tests 
made. These records may be more or less elaborate, according 
to the kind and amount of the work in hand, but in any case, 
enough detail should be given to make them intelligible to other 
engineers. 

236. Marking Specimens. There is sometimes a tempta- 
tion, in making tensile specimens, to stamp upon them many 
details of the test, and for this purpose an elaborate cipher 
system has sometimes been used. But this method is to be 
strongly deprecated. Each briquet should receive its proper 
consecutive number, as mentioned in 189, and the details 
concerning it should be placed in the record book. 

237. RECORDS KEPT AT ST. MARYS FALLS CANAL. In the 

tests of cement at St. Marys Falls Canal, during the construc- 
tion of the Poe Lock, a system of records was used that gave 
entire satisfaction. At the time the largest amount of cement 
was being used three molders were employed, each making 
fifty briquets per day of eight hours. Over one hundred thou- 
sand briquets were made in five and one-half years. Although 
the system of records used at this point may be more elaborate 
than is often necessary, yet the system will be described, and 
certain modifications will be suggested for places requiring less 
complete records. 

238. Barrel Records. The barrels receive consecutive num- 
bers after they are tiered up in the warehouse. The " receiv- 
ing book" is a simple transit book in which are entered the 
date of the receipt of each cargo, the name of the boat (or the 
car number, if shipped by rail), the brand of cement, the num- 
ber of barrels, the first and last barrel number of the cargo and 
the warehouse in which the cement is placed. The next book 
to be used is the " barrel book," in which the numbers of the 
barrels are entered consecutively in a column at the left, each 
barrel being given one line. This book is also of transit size, 
but might well be larger. The headings are given below. 



RECORDS OF TESTS 



161 



SAMPLE PAGE OF "BARREL RECORD" 



No. 
BBL. 


SAMPLED. 


DEFECTS. 


AC- 
CEPTED. 


RE- 
JECTED. 


ISSUED. 


REMARKS. 


88251 

3 

4 

5 
6 
7 
8 
9 
88260 
1 
2 

3 

4 
5 


M. D. 

5 16 


M. D. 

6 5 


S7 = 48% 


M. D. 

6 13 


M. D. 


July 25 
July 25 
July 25 

Sept. 27 

July 25 
July 25 
July 25 
July 25 
July 25 
July 25 
July 25 
July 25 

Sept. 28 

July 25 
July 25 


67 = 65% 


5 19 


6 6 
6 5 




6 13 


6 12 


S7 = 102% 
(Removed by 
| Contractor 


i 87= 32% \ 
\ S7 = 35% f 














5 19 






5 26 


. . . 






















5 19 







5 26 


. . . 






5 19 


6 6 

6 5 




6 13 


6 12 


S7 = 105$ 
( Removed by 
\ Contractor 


j S7 = 40$ { 
I S7 = 32% \ 



























When the barrels are sampled and briquets made, the date 
sampled is entered in the second column of the barrel record 
book. The other columns will be explained later. 

239. Holders' Records. Separate sheets of paper ruled and 
headed as shown on page 148 are used by the molders to record 
the details concerning the making of briquets. 

Separate sheets properly ruled and headed are also given to 
the assistants who test time of setting and fineness. These 
record sheets, when filled in by the assistants, are copied the 
following day by the bookkeeper, into the permanent " record 
book." At the end of the month these separate sheets, con- 
taining original records of work done, are folded and filed for 
future reference. 

240. Briquet Record. The briquets are made in sets of 
ten for convenience. Each set is given a page in the "record 
book," as is indicated on page 149 where the form for this book 
is given. The size of page is 9 by 12 inches. Paper having the 
same ruling and column headings is convenient for reporting 
tests to the chief engineer. 

241. Summary Book. The data for each set of briquets 
are copied from the record book, in a condensed form, into the 
"summary book," one line of the latter containing a page of 
the former. In the summary book each brand is given a few 



162 



CEMENT AND CONCRETE 



I I 

I * 



ox 



>-( 05 



r 



'9 '}<&$ 



RECORDS OF TESTS 



163 



fipnop 



NS IN MORTAR 
WEIGHT. 



6Q 



SETTING 
CEMENT. 



3ASIIS 
GOT 'OR O.MISSVcT 



CO 



OS 



xanorag 
AO -on 



II 



164 



CEMENT AND CONCRETE 



03 



PROPORTIONS IN 
MORTAR BY WT. 






sxsmtxnig; 



sxanorag 



'ON: 



RECORDS OF TESTS 



165 



pages by itself, so that this book corresponds to a ledger in form. 
By this means a large number of tests on the same brand may 
be looked over at once. The summary book might be omitted 
where a smaller number of tests are to be made, or it might 
be slightly modified and take the place of the record book. A 
sample page is given below. 

242. Records of Fineness, Time of Setting and Soundness. 
Although provision is made in the record book for recording 
time of setting and fineness, it has been found that where a 
large amount of cement is being tested it is more convenient 
to have separate books for each test. Especially is this true 
as it has been judged necessary to test but a very small per- 
centage of the barrels for fineness, while a larger percentage of 
the barrels are tested for time of setting and soundness. The 
"fineness book" is as simple as possible and need not be illus- 
trated. A sample page of the "pat book" is given below. 

Sample Page of "Pat Book" or Record of Time of Setting and 

Soundness 

Lagerdorfer Portland Cement Pats, Two from Every Third Barrel 



No. 
BBL. 


SPECIAL MARKS! 
ON PATS. 




fi~ 

B^ 

s w 

^ 
M 


3s 
^ 

SS 

s 

2fe 


9a 

s 




TREATMENT 
OF PATS. 


EXAMINED. 


REMOVED. 


RE- 
MARKS. 


Stmr. 


Tank. 


Date. 


Condi- 
tion. 


Date. 


Condi- 
tion. 


110274 

' '7 

'so 

3 

' 'e 
' '9 
'92 
'95 
'93 

' SOI 


1 

2 
1 
2 
1 

2 

1 

2 
1 
2 
1 
2 
1 
2 
1 
2 
1 
2 
1 
2 


9:23 
9:28 


117 

'37 


337 
232 


IPat. No. 1 in vapor when set hard, im- 
mersed 3 hrs. later, temperature 176 F. 


Pat. No. 2 in tank when set hard. 


Mo.D. 
8 1 


O.K. 


Mo.D. 

7 15 
8 17 
7 15 
8 17 
7 15 
8 17 

7 15 

8 17 
7 15 
8 17 
7 15 
8 17 
7 15 
8 17 
7 15 
8 17 
7 15 
8 17 
7 15 
8 17 


O.K. 

f Surface 
< cracked 
(,& scaled. 
O.K. 


Water 
24%. 


8 1 


O.K. 


9:32 


15 


268 


8 1 


O.K. 


9:34 


143 


383 


8 1 

's' i 

8 ' 1 


O.K. 

'O.K. 

O.K. 


9:41 


24 


259 


9:50 


15 


250 


9:53 
9:57 
10:28 


12 
123 
'97 


247 
363 
367 


8 1 


O.K. 


8 1 


O.K. 


8 1 


O.K. 


10:32 


103 


365 


8 1 


O.K. 













166 



CEMENT AND CONCRETE 



3 03 









-OR 



a 



00 



^^eo ^ $* 



H 8 

31 



00 



>-H GO us eo ^J 



p. 






88 




RECORDS OF TESTS 167 

243. The Diary. When the bookkeeper has copied the 
data contained on the record blanks into the record book and 
summary book, he turns to the proper page in the diary and 
records the briquets to be broken. Thus, if briquets made 
May 17th are to be broken at three months, he enters the num- 
bers of these briquets and the tank in which they have been 
placed under the date Aug. 17th. This leaves no chance of 
allowing briquets to go beyond the proper time of breaking. 

244. Acceptance or Rejection. If all of the tests on a given 
sample are satisfactory, the date of acceptance is placed in the 
proper column of the barrel book. It only remains then to 
mark the barrels "0. K.," and issue them when needed, placing 
the date of issue in the column indicated. If, however, some 
of the tests have given unsatisfactory results, the failure is 
noted in the " defects" column of the barrel book, and the 
barrel is resampled to determine whether the failure was due to 
faulty manipulation. If finally rejected, the barrel is promi- 
nently marked to prevent its being issued for use. 

It is seen from the above that the history of each barrel is 
given in the barrel book, and the record of any brand is given 
in a condensed form in the summary book. 

245. Special Tests. When special tests are made to inves- 
tigate the effects of variations in manipulation, or for any 
other special purpose, such as to test the value of certain kinds 
of sand, it becomes convenient to have still another form which 
may be called a " series book." In this the results are so ar- 
ranged that they may be studied for conclusions, and tables 
for reports may be copied directly from it. A sample form is 
given on preceding page. Should extra rulings be needed, 
they may be placed at the right in the "remarks" column. 



PART III 

PREPARATION AND PROPERTIES OF 
MORTAR AND CONCRETE 



CHAPTER XI 

SAND FOR MORTAR 

246. Mortar. When cement is mixed with sand and water, 
the resulting paste is called mortar. The term "neat cement 
mortar" is sometimes used to designate a cement paste with- 
out sand, but when the term mortar is not qualified, it refers 
to the mixture containing sand. The primary function of mor- 
tar is to bind together pieces of stone of greater or less size, 
though it is sometimes used alone to prevent the percolation of 
water, to make a smooth exterior finish, or in places too confined 
to permit of placing concrete. 

There are comparatively few cases in which it is judicious 
to use cement without the addition of sand, fbr such an ad- 
mixture not only cheapens the mortar, but actually improves 
it for nearly all purposes. The quality of sand used is only 
second in importance to the quality of the cement. Indeed, if 
one does not know how to select either a good cement or a good 
sand, he is in greater danger of going amiss in the selection of 
the latter than the former; for the cement has been placed upon 
the market by a manufacturer who has a reputation to estab- 
lish or maintain. 

ART. 29. CHARACTER OF THE SAND. 

247. Various kinds of rock are capable of producing sand of 
good quality. The natural sands are usually siliceous in char- 
acter, but calcareous sands are also met with and may give 
excellent results in mortar. Good artificial sand may be made 
from almost any kind of rock that is not liable to chemical 

168 



CHARACTER OF SAND 



169 



decay, even though it be only moderately hard. One of the 
most essential features of a good sand is that the grains should 
be perfectly sound. Evidences that chemical decay is going 
on in the grains would indicate that the sand is of very inferior 
quality. 

248. SHAPE AND HARDNESS OF THE GRAINS. It is gener- 
ally believed that the grains of sand should be angular in order 
to give the best results; this is probably true, although in test- 
ing three varieties of calcareous sand, M. Paul Alexandre l ob- 
tained results which seemed to indicate that if rounded grains 
are disadvantageous, the other properties of the sand may 
readily counterbalance this disadvantage. 

M. Alexandre used three sands which were reduced to the 
same fineness by sifting into different sizes and then remixing 
them in fixed proportions (equal parts of five sizes). The three 
sands were, 1st, white marble, very hard with sharp corners; 
2d, moderately hard limestone; and 3d, chalk, very soft with 
rounded grains. The proportions used were 400 kg. of cement 
to one cubic meter of sand, the amount of water varying from 
twenty-five to thirty per cent, of the sand, according to the 
amount required to produce plasticity. The tensile strength of 
the mortars, in pounds per square inch, is given in Table 45. 

TABLE 45 
Results Obtained with Three Varieties of Calcareous Sand 



CHARACTER OF SAND. 


TENSILE STRENGTH, POUNDS 
PER SQUARE INCH AT 


7 da. 


28 da. 


6 mo. 


1J yrs. 


1 Marble 


45 
72 

86 


107 
148 
129 


171 
222 
205 


220 
266 
262 


2 Limestone ... ... 


3 Chalk 





As these sands varied in the structure and hardness as well 
as in the shape of the grains, it cannot be concluded that rounded 
grains are as good as sharp and angular ones for mortar-making. 
There is little question that if two samples of pure quartz sand, 



'Recherches Experimentales sur Les Mortiers Hydrauliques." 



170 



CEMENT AND CONCRETE 



differing in sharpness but alike in all other respects, including 
the percentage of voids, were tested side by side, the rounded 
grains would be found inferior. (See also 253.) 

M. Alexandre also made tests on sands differing both in 
chemical and physical characteristics, but having the same fine- 
ness, namely, twenty per cent, each of five sizes of grain. Some 
of the results are given in Table 46. 

TABLE 46 
Results Obtained with Various Sands 



SAND. 


WATER 
PER 
CENT. OF 
VOLUME 
OF SAND. 


TENSILE STRENGTH, IN LBS. PER 
SQ. IN., OF MORTARS CONTAIN- 
ING 400 KG. OF CEMENT TO 1 Cu. 
METER SAND, AT AGES OF 


7 da. 


lyr. 


3 yrs. 


Sea Sand 


21 
28 
21 
20 
20 
28 


69 
78 
65 
63 
79 
35 


165 
198 
158 
174 
178 
99 


245 

267 
201 
215 
244 
132 


Calcareous (Renville stone) . 
Granitic 
Siliceous (cliff quartz) 


Siliceous (Cherbourg Quartzites) . 
Coke 





249. Siliceous vs. Calcareous Sands. The above tests 
would seem to show that sand to be used in mortar need not be 
siliceous. In. experimenting on different varieties of sand, both 
natural and artificial, the author has obtained results that 
point to a similar conclusion. Some of these tests are given in 
Tables 47 to 50. 

Table 47 gives the results obtained with four varieties of 
siliceous sand. The first was an artificial sand made by crush- 
ing sandstone, the second and third were natural sands con- 
taining a large percentage of quartz grains, and the fourth 
appeared to be almost pure quartz. Only the fine particles of 
the sands were used in the tests given in this table. 'The dif- 
ferences in strength at the end of two years are not great, but 
the two natural sands appear to give somewhat lower results. 

In Table 48 the two natural sands were again compared, 
but this time in connection with a calcareous sand formed by 
crushing limestone. The latter gave the best results. Only 
the finer grains were used in these tests. 

250. Tables 49 and 50 are more valuable in this connection, 



CHARACTER OF SAND 



171 



TABLE 47 

Values of Different Varieties of Fine Siliceous Sand for Use in 
Portland Cement Mortar 

Two PARTS SAND TO ONE CEMENT BY WEIGHT 











TENSILE 


. 






WATER, 


STRENGTH, LBS. 


8 


SAND. 


FINENESS. 


Per 


PER SQ. IN. AT 






3 
p 

B 






Cent. 


6 Mo. 


2Yr. 


9 


a 


b 


c 


d 


e 


i 


( Screenings from ( 


Pass 40 sieve . 


18.5 


388 


470 


2 


crushing Pots- ? 
( dam sandstone ( 


Pass 40 sieve, 
retained on 100 


17.5 


478 


539 


3 


Bank sand, siliceous 


Pass 40 sieve . 


13.3 


433 


445 


4 


River sand, siliceous 


Pass 40 sieve . 


12.1 


382 


437 


5 


Clean quartz 


Pass 40 sieve . 


13.3 


398 


506 



NOTE. Holes in No. 40 sieve 0.015 inch square, holes in No. 100 sieve 
about 0.0065 inch square. 

TABLE 48 
Different Varieties of Fine Sand for Portland Cement Mortar 











TENSILE STRENGTH, POUNDS 










PER SQUARE INCH. 


H 






PER CENT. 
WATER 




3 


SAND. 


FINENESS. 




1 Part Sand to 1 


2 Parts Sand to 











Cement by Wt. 


Cement by Wt. 


h 

8 






Itol 


1 to 2 


6 mo. 


IS 
mo. 


3yr. 


6 mo. 


18 
mo. 


3yr. 




a 


b 


c 


d 


e 


/ 


g 


h 


i 


J 


1 


River sand, sili- 






















ceous . . . 


Pass 40 sieve 


14.0 


12.4 


715 


725 


776 


491 


575 


581 


2 


Bank sand, sili- 






















ceous . . . 


Pass 40 sieve 


14.5 


12.6 


664 


699 


759 


442 


502 


524 


8 


Calcareous sand 






















from crushing 
limestone . . 


Pass 40 sieve 


18.2 


17.7 


721 


770 


788 


531 


632 


680 


4 


Calcareous sand 
from crushing 
limestone . . 


Pass 40, re- 
tained on 100 


17.5 


17.0 


753 


783 


844 


597 


659 


727 



since the coarser particles of the sand were used with the fine. 
The sand was separated into four sizes by sifting, and then 
remixed in equal proportions. Table 49 gives the results ob- 



172 



CEMENT AND CONCRETE 



tained with natural cement, and Table 50 refers to Portland. 
The superiority of the screenings is very clearly shown, the 
limestone giving especially good results. Indeed, the strength 
obtained with three parts limestone screenings to one part of 
either Portland or natural cement is remarkably high. The 
mortar made from such sand is peculiarly plastic when fresh, 
and soon gains a high strength which it appears to maintain. 

TABLE 49 
Values of Different Varieties of Sand for Natural Cement Mortar 









'* 


TENSILE STRENGTH, LBS. 






X 


^_ M 


PER SQ. IN., 3 PARTS SAND 


i 


SAND. 





0^ 


TO 1 CEMENT BY WT. 









Sg 


28 Da. 


6Mos. 


lYr. 


2 Yrs. 




a 


b 


c 


d 


e 


/ 


d 


i 


Clean crushed quartz . . . 


MX. 


15.4 


117 1 


344 


356 


332 


2 


River sand, siliceous . . . 


MX. 


13.3 


93 


297 


339 


308 


3 


Limestone screenings . 


MX. 


16.7 


143 


467 


526 


601 


4 


Potsdam sandstone screenings 


MX. 


18.2 


113 


316 


416 


462 


5 


Clean crushed quartz . 


20-30 


12.6 


118 


330 


342 


324 



1 13.6 per cent, water, trifle dry. 

NOTE. Fineness MX. means 25 per cent, each of 20-30, 30-40, 40-50 

and 50-80. 
Expression 20-30 means passing No. 20 sieve and retained on 

No. 30 sieve. 

TABLE 50 
Values of Different Varieties of Sand for Portland Cement Mortar 







. 


' 


TENSILE STRENGTH, LBS. 






He 


fc H 


PER SQ. IN., 3 PARTS SAND 


fe 




H 


W S" 1 


TO 1 CEMENT BY WT. 


w 


SAND. 


g 


O "^ 




P5 




(z; 


P3^ 
















^ 


28 Da. 


6Mos. 


lYr. 


2 Yrs. 




a 


b 


c 


d 


e 


/ 


9 


1 


Clean crushed quartz . . 


MX. 


12.5 


255 


327 


359 


335 


2 


River sand, siliceous . . . 


MX. 


11.1 


206 


284 


329 


324 


3 


Limestone screenings . . . 


MX. 


12. 5 1 


407 


574 


667 


665 2 


4 


Sandstone screenings . . . 


MX. 


12. 5 1 


321 


438 


495 


492 3 


5 


Clean crushed quartz . . . 


20-30 


11.1 


259 


344 


369 


335 



1 Trifle dry, plastic. 2 13.3 per cent, water. 3 14.3 per cent, water. 
NOTE. Fineness MX. means 25 per cent, each of 20-30, 30-40, 40-50 
and 50-80. 



FINENESS OF SAND 173 

251. Slag Sand. To turn to good account some of the 
immense quantities of blast furnace slag produced yearly, the 
use of granulated slag in place of ordinary sand has been ad- 
vocated. In a paper read before the Engineers' Society of 
Western Pennsylvania, in March, 1904, Mr. Joseph A. Shinn 
described some experiments he had made, in which it was shown 
that "slag sand," with Portland cement, natural cement, or 
common lime, gave a higher strength than the sample of river 
sand used in the comparison. 

The "slag sand" is produced by projecting two flat jets of 
water into the stream of molten slag, the resulting sand being 
heavier, finer and more nearly uniform in size of grain than the 
ordinary slag granulate. 

252. Sand for Use in Sea Water. It has been said that 
granitic sands when used in sea water do not give good results 
on account of the felspar of the granite being attacked by the 
cement when the concrete is impregnated with sea water. M. 
Paul Alexandre would proscribe the use of argillaceous sands 
in sea water, but he found that sands containing calcareous 
marl gave excellent results in the sea, and others have stated 
that the mixture of crushed limestone with concrete has been 
known to hinder the action of sea water upon it. Since porous 
and permeable mortars are most liable to disintegration by 
sea water, it is evident that it is especially desirable to employ 
a sand in which the proportion of voids is small. 

ART. 30. FINENESS OF SAND 

253. The size and shape of the grains are important ele- 
ments in the quality of sand. Considering grains of the same 
shape but differing in size, the larger grain will have a smaller 
surface area in proportion to the volume than the smaller grain, 
since the volume varies approximately as the cube of one di- 
mension while the surface varies as the square. Since, in order 
to obtain the best results in mortar, each grain of sand must be 
coated with cement, it follows that, other things being equal, 
the coarser grained sands will give the best results, because 
they will be more thoroughly coated; this will be especially true 
when the amount of sand in the mortar is relatively large. 

Following the same reasoning given above as to the relative 
volume and superficial area of sand grains, it would appear 



174 



CEMENT AND CONCRETE 



that spherical grains would be better than cubical or angular 
ones (see 248). This, however, is not thought to be the case, 
for the better bond obtained with angular grains seems to coun- 
terbalance the advantage which the small superficial area would 
appear to give to the spherical grains. For this reason a len- 
ticular shaped grain, while having a very large area relative to 
its volume, will give excellent results in mortar if otherwise 
suited to the purpose. 

It is usually desirable to have all of the voids in the sand 
filled by the cement paste, as this renders the mortar less por- 
ous, and makes it more certain that all the grains are coated 
with cement. On this account a mixture of fine and coarse 
particles is excellent. 

TABLE 51 

Effect on Tensile Strength of Varying Fineness of Limestone 
Screenings Used with Portland Cement 



AGE 
BBIQUETS WHEN 
BROKEN. 


TENSILE STRENGTH, POUNDS PER SQUARE INCH 

FINENESS OF SCREENINGS. 


10-20. 


20-30. 


30-40. 


40-50. 


40-80. 


PASS 50. 


6 months . 


718 


657 


633 


516 


.... 


403 


2 years . . . 


812 


754 


656 


. . . 


516 


488 


4 years . 


845 


782 


714 


. . . 


571 


516 



SIGNIFICANCE OF FINENESS 





SIEVE NUMBER. 


APPROXIMATE 










Passing. 


Retained on. 


GRAIN. 








Inch. 


10-20 


10 


20 


.057 


20-30 


20 


30 


.028 


30-40 


30 


40 


.020 


40-50 


40 


50 


.015 


40-80 


40 


80 


.012 


Pass 50 


50 




.008 



NOTES. Three parts screenings to one cement by weight. 

All briquets made by one molder and immersed in one tank. 
Variations in consistency were slight, the largest percentage of 
water being used for the finest particles. 



FINENESS OF SAND 



175 



254. TESTS ON EFFECT OF FINENESS OF SAND. Many of 
the experiments made to show the effect of the fineness of sand 
on the strength of the mortar are defective, because the sand 
used varies in the shape of the grains and in chemical charac- 
teristics as well as in fineness. The experiments given in Table 
51 were made with screenings obtained in crushing limestone* 
and thus all causes of variation aside from the fineness of the 
sand were absent, except the differences in consistency of the 
mortar, the uniformity in consistency depending on the judg- 
ment of the operator. The results show quite clearly the su- 
periority of the coarser sand. 

255. The Relative Effect of Fine Sand on Portland and Nat- 
ural Cement. The tests in Table 52 were made to determine 

TABLE 52 

Coarse and Fine Sand, Relative Effects with Portland and 

Natural Cement 







TENSILE 




a 


TENSILE 






"! . 


STRENGTH, 


5tf s 


fc 


STRENGTH, 


o &T - 


AGE OF 
BRIQUETS 

WHEN 

BROKEN. 


qH 
^ 

sHs 
fc 


POUNDS PER 
SQ. IN. WHEN 
SAND is 


^S 1 

f O H # 
fcfcfc^ 
H H f* O 

sl*? 


BRAND 
PORTLA 

3EMENT 


POUNDS PER 
SQ. IN. WHEN 
SAND is 


2g a i 

im 

- - 












&l 




20-30 


40-80 


2 s g 


h 
O 


20-30 


40-80 


$* 2 


28 days . . j 


Bn 
In 


197 

89 


145 

57 


74 
64 


A 

u 


406 
352 


337 
275 


83 
78 


6 months . . j 


Bn 
In 


216 
364 


188 
267 


87 
73 


A 

u 


520 
499 


446 
415 


86 
83 


2 years . . j 


Bn 
In 


256 
450 


250 
419 


98 
93 


A 
U 


546 
567 


451 
496 


83 
89 



NOTES. Sand, limestone screenings ; three parts to one cement by 

weight. 
20-30 means sand passing sieve with 20 meshes per linear 

inch, and retained on sieve with 30 meshes per linear 

inch. 
Columns 5 and 9 show percentage that strength with finer 

sand is of the strength with coarser sand. 

the relative effects of fine sand on Portland and natural cements. 
Limestone screenings of two sizes of grain were used in con- 
nection with two brands of each kind of cement. At twenty- 
eight days the natural cement shows the decrease in strength 
due to the use of fine sand more than Portland cement does. 



176 CEMENT AND CONCRETE 

At six months the fine sand seems to have about the same 
effect on Portland and natural, but the two-year results in- 
dicate that the ultimate effect is less on the natural cement 
than on the Portland; the mean ratio of the strength obtained 
with fine sand to that given by coarse sand being ninety-six 
in the case of natural, and only eighty-six in the case of Port- 
land. The effect of fine sand appears to decrease with age, 
especially with natural cement. 

The fineness of sand will be treated further in the following 
article relating to voids. 

ART. 31. VOIDS IN SAND 

256. Conditions Affecting Voids. The voids present in a 
given mass of sand will depend upon the shape of the grains, 
the degree of uniformity in size of grains, the amount of moisture 
present, and the amount of compacting to which the mass has 
been subjected. If all of the grains in a given mass of sand are 
of uniform size, the percentage of voids will be independent of 
what that size may be. In other words, the percentage of 
voids in a cubic foot of buckshot will be the same as in a cubic 
foot of bird shot; but if we take a cubic foot of a mixture of 
buck and bird shot we will find that the voids are much 
less. 

257. Effect of Shape of Grain. M. Feret has published in 
France the results of a large number of experiments made by 
him as to the voids in sand and broken stone. 1 Table 53 gives 
the results he obtained concerning the effect of the shape of 
the grains on the percentage of voids present. He first divided 
each sand into three parts by means of three sieves, which we 
will call A, B and C. Sieve A had four meshes per sq. cm. 
(about five meshes per linear inch), sieve B had 36 meshes per 
sq. cm. (about fifteen meshes per linear inch), and sieve C had 
324 meshes per sq. cm. (about forty-five meshes per linear 
inch). The grains that passed A and were retained on B were 
designated G, the grains that passed B and were retained on C 
were designated M, and the grains that passed C were desig- 
nated F. These different sizes were then recombined by tak- 
ing five parts of G, three parts of M and two parts of F, and 



1 Abstracted in Engineering News, Vol. XXVII, p. 310. 



VOIDS IN SAND 



177 



the resulting sand was designated G 5 M 3 F 2 . Thus, all of the 
sands tested had the same " granulometric " composition. 

TABLE 53 

Voids in Sands Having Different Shaped Grains 
FROM M. FERET 



NATURE OF SAND. 


VOLUME OF VOIDS REMAINING IN 
ONE LITER OF SAND. 


Unshaken. 
C.C. 


Shaken to Refusal. 
C.C. 


Natural sand with rounded grains. 
Cherbourg quartzite, angular grains. 
Crushed shells, flat grains. 
Residue of Cherbourg quartzite crushed 
between jaws, laminated grains. 


359 
421 
443 

475 


256 

274 
318 

346 



It is seen that the rounded grains have the smallest percent- 
age of voids, or about thirty-six per cent, unshaken, while the 
laminated grains gave the largest percentage. It may also be 
noticed that the angular grains were compacted more by shak- 
ing than any of the others. 

258. Effect of Granulometric Composition of Sand on the 
Percentage of Voids. To determine the effect of uniformity of 
size of grain upon the percentage of voids and the strength of 
mortars, the author has experimented with an artificial sand 
formed by crushing limestone. That portion of the product 
that passed the coarse screen of the crusher varied in fine- 
ness from particles three-eighths of an inch in one dimension to 
a very fine powder, the particles of which were less than .0065 
inch in one dimension. Such material admits of division into 
parts that differ widely in fineness, but which are essentially 
of the same composition, and it is therefore excellent for an 
experiment of this kind. 

The four sieves used in first separating the material into 
parts had, respectively, 10, 20, 40 and 80 meshes per linear inch, 
the sizes of the holes being, respectively, about as follows: 0.08 
inch, 0.033 inch, 0.017 inch, and 0.007 inch square. The sev- 
eral sizes of grain are designated as follows: 

"C," Coarse, passing No. 10, retained on No. 20. 
"M," Medium, " " 20, " " 40. 

"F," Fine, " " 40, " " 80. 

"V," Very fine, " " 80. 



178 



CEMENT AND CONCRETE 



M. Feret's method of designating the granule me trie compo- 
sition, namely, to represent by exponents the number of parts 
of each size of grain, has been adopted. 

259. The voids were obtained by first weighing a given 
volume of the sand; dividing the weight by the specific gravity 
of the limestone, as previously determined, gives the amount 
of solid material in the measure, and this subtracted from the 
volume of the measure, gives the voids. This method is con- 
sidered more nearly accurate than the usual one of measuring 
the amount of water required to fill the voids in a measure of 
sand, especially so for a sand of uniform character and one 
which absorbs water quite freely. 



TABLE 54 

Voids in Limestone Screenings, Showing Effect of Variations in 
Granulometric Composition 



FINENESS OF 
GRANULOMETRIC 
COMPOSITION. 


WEIGHT OF 
ONE LITER OF 
SAND, DRY, 
GRAMS. 


VOLUME SOLID 
SAND IN 
ONE LITER 
(SP. GR. = 2.667) 
Cur. CENT. 


PER CENT. 
VOIDS 
IN SAND. 




Loose. 


Shaken. 


Loose. 


Shaken. 


Loose. 


Shaken. 


a 


b 


c 


d 


e 


/ 


9 


C = Coarse 10 to 20 


1126 


1358 


422 


509 


57.8 


49.1 


M = Medium 20 to 40 


1140 


1362 


428 


511 


57.2 


48.9 


F = Fine 40 to 80 


1150 


1392 


431 


522 


56.9 


47.8 


V = Very fine, pass 80 


1165 


1609 


437 


603 


56.3 


39.7 


. C 




1395 




523 




47.7 


M 




1439 





540 




46.0 


F 




1459 




547 




45.3 


V 





1656 




621 




37.9 


C 65 , M 25 , F 15 , V* 


, 


1606 




602 




39.8 


O, M 30 , F 20 , V 10 


, 


1732 




649 


t 


35.1 


C 25 , M 2 *, F'25, V26 




1912 


. 


717 




28.3 


C 30 , M 25 , F* 5 , V 80 




1850 




694 


. 


30.6 


C 50 , M, F, V 50 




1991 





746 





25.4 



The results obtained are given in Table 54. Comparing the 
voids in C, M, F and V, it is seen that the first three have nearly 
the same percentage, but V has less voids than the others. 
This is explained by the fact that this sample was made up of 
all sizes smaller than the holes in No. 80 sieve, down to the 
fine powder. Comparing the mixed sands, it is seen that the 
sample made up of equal parts of coarse and very fine had 



VOIDS IN SAND 



179 



the least voids, the percentage being only a little more than 
half of that obtained with coarse particles alone. The next 
lowest percentage was given by the sample having equal parts 
of four sizes. 

It is apparent that the granule metric composition has a 
very important effect on the percentage of voids. When one 
desires to make a compact mortar with as small a quantity of 
cement as possible, similar tests might well be made with the 
materials available for use. 

260. Effect on Strength of Mortars of Varying the Granulo- 
metric Composition of Sand. Table 55 gives the results of 
tensile tests of mortars made with limestone screenings of vari- 
ous granulometric compositions. The differences in strength 
are not very great, but it appears that with one-to-three mor- 
tars the highest strength is developed at six months, with the 
coarse grains alone, but when poorer mortars are in question 
the result is affected by the percentage of voids in the sand. 

TABLE 55 

Limestone Screenings with Portland Cement. Effect on Tensile 
Strength of Variations in Granulometric Composition of Sand 







TENSILE STRENGTH AT 




GRANULOMETRIC COMPOSITION 




6 Mos. POUNDS PER 


WEIGHT OF 


OF SAND. PER 




SQ. IN. WITH PARTS SAND 


BRIQUETS IN 


CENT. OF EACH SIZE GRAIN. 


VOIDS. 


TO ONE CEMENT BY 


GRAMS. 




% 


WEIGHT. 




C 


M 


F 


V 




3 


5 


3 


5 





100 








46 


509 


324 


1465 


1438 


40 


30 


20 


10 


35 


505 


392 


1466 


1480 


25 


25 


25 


25 


31 


470 


356 


1445 


1455 


30 


25 


15 


30 


28 


496 


391 


1448 


1470 


50 








50 


25 


487 


349 


1455 


1460 



CEMENT. Portland, Brand R. For significance of composition of sand, 
see text. 

261. Table 56 gives the results of similar tests of both Port- 
land and natural cement with Point aux Pins sand dredged 
from St. Marys River and containing a very large percentage 
of quartz grains. The sand was divided into but three parts 
by sifting, and was then remixed, the proportion of each size 
being indicated in the table. The results verify the conclusions 



180 



CEMENT AND CONCRETE 



already drawn that the coarser sands give the higher strength. 
It appears that not more than one-half of the grains should be 
very fine if the best results are desired. 

TABLE 56 

Varying the Granulometric Composition of River Sand. Effect on 
Value of, for Use in Cement Mortar 



COMPOSITION OF SAND AS 
TO FINENESS. 


TENSILE STRENGTH, POUNDS PER SQUARE INCH. 


Parts Used 










that Passed 
No. 20 Sieve 
and Re- 
tained on 
No. 30. 


Parts 
Used, 
30-40 


Used that 
Passed 
No. 40 
Sieve. 


Portland Cement with 
Two Parts Sand to One 
Cement by Weight, at 
age of 


Natural Cement with 
Three Parts Sand to One 
Cement by Weight, at 
age of 


M 


F 


V 


28 da. 


6 mo. 


lyr. 


2yr. 


28 da. 


6 mo. 


lyr. 


2yr. 


10 








342 


471 


560 


591 


77 


267 


348 


341 


4 


1 


5 


300 


448 


515 


507 


77 


237 


304 


319 


2 


4 


4 


290 


425 


494 


503 


79 


278 


291 


325 


1 


3 


6 


246 


384 


455 


442 


46 


222 


234 


251 


1 


2 


7 


271 


366 


456 


438 


67 


226 


247 


251 



NOTE. River sand, mostly quartz, obtained at Point aux Pins. Each 
result mean of five briquets, all made by one molder. 

262. Effect of Moisture. * The effect of a small amount of 
moisture on the bulk of a given weight of sand is not usually 
appreciated, but it may easily be shown that it is very marked. 
The results in Table 57 were obtained by adding small amounts 
of water to a given bulk of dry sand. Each time, after the 
water was added, the sand was stirred up and the weight of a 
given volume of the moist sand was obtained. It appears that 
the finer sands are affected more than coarse ones. 

In the case of the limestone screenings 40-80, if we add but 
3.7 per cent, water to a given quantity of dry sand, the bulk 
of the sand is so increased that if we take 1,000 c.c. of the. moist 
sand it will contain but 720 c.c. of dry sand. The voids are, 
of course, correspondingly increased from 54.5 per cent, to 
67.2 per cent. 

The cause of this increase in bulk is that each grain of sand 
is surrounded by a film of water which prevents the grains 
from lying close together after they have been disturbed. A 
large amount of air is also imprisoned in the mass. It may be 
noticed that the difference in bulk between moist and dry sand 
is greater when the measurements are made " loose. " 



VOIDS IN SAND 



181 



TABLE 57 
Volume of Sand and Voids as Affected by the Addition of "Water 



N 

1 


SAND. 


a 
* T 


WEIGHT OF 
PRY SAND IN 
ONE LITER 
OF MOIST 


VOLUME OF 
DRY SAND 
IN ONE 
LITER OF 
MOIST 


PERCENT. 
VOIDS IN 
SAND BY 
VOLUME. 






W 




4 




&l 


SAND. 


SAND. 




m 



Kind. 


9 
P 


a >> 


-oo 


C~33 


-~S 


s"S 




d 


W 












$ S3 


4) S 




0) 






i 


* 


oj 


|| 




J3 3 


| 


J| 












GO 


o 


020 








a 


6 


c 


d 


e 


/ 


9 


h 


i 




Crushed 


















1 


Limestone. 


10-20 


0.0 


1288 


1489 


1000 


1000 






2 


ii 


it 


4.8 


1094 


1367 


849 


919 






3 


" 


" 


7.7 


1023 


1295 


794 


869 






4 


ii 


a 


11.9 


996 


1276 


773 


857 






5 


ii 


40-80 


0.0 


1214 


1481 


1000 


1000 


54.5 


44.6 


6 


H 


ii 


0.85 


1124 


1489 


926 


1005 


57.9 


44.2 


7 


" 


a 


1.5 


1059 


1470 


872 


993 


60.3 


44.9 


8 


a 


n 


2.2 


950 


1383 


782 


934 


64.4 


48.2 


9 





ii 


3.7 


875 


1298 


720 


877 


67.2 


51.4 


10 


it 


t ( 


6.3 


824 


1274 


679 


860 


69.1 


52.3 


11 


ii 


ii 


7.8 


799 


1266 


658 


855 


70.0 


52.6 


12 


" 


a 


12.3 


817 


1280 


672 


864 


69.4 


52.0 


13 


i i 


it 


16.8 


829 


1306 


683 


881 


69.0 


51.1 


14 


i 


ii 


20.2 


836 


1274 


689 


860 


68.6 


52.3 


15 


i 


a 


25.3 


891 


1357 


783 


916 


66.6 


49.1 


16 


' 


ii 


30.3 


1049 


1270* 


864 


858* 






17 


i 


Pass 80 


0.0 


1185 


1500 


1000 


1000 






18 


' 





2.4 


1038 


1394 


873 


929 






19 


' 


" 


5.1 


835 


1281 


704 


854 






20 





" 


12.2f 


806 


1310 


680 


873 






21 


' 


t< 


17.7t 


806 


1260 


680 


840 








Point aux 


















22 


Pins. 


r 


0.0 


1725 




1000 








23 


(i 




2.0 


1405 




815 








24 


ii 




4.0 


1400 




810 








25 


ii 


+ j 


6.0 


1400 




810 








26 


ii 




10.0 


1415 




820 








27 


ii 




11.6 


1425 




825 








28 


ii 


- 


18.4 


1485 




860 















* Not jarred down in measure as much as usual. 
t Sand crumbled like damp earth. 

f Sieves No. 20 

I Approx. size 



Water rose to surface. 



Fineness of Point 
aux Pins Sand 



30 40 50 



80 



holes - .033 .022 .017 .012 .007 
t Per cent, passing 96.0 82.3 46.6 6.7 1.2 

NOTE. 10-20 = passing No. 10 sieve (holes about .08 in. sq.) and retained 
on No. 20 sieve. 



182 CEMENT AND CONCRETE 

263. This subject is of great importance in proportioning 
mortars, because, in construction, the amounts of cement and 
sand are usually measured. Suppose it is desired to use a mix- 
ture of one hundred pounds of cement to four hundred pounds 
of sand, and for convenience we will suppose the packed cement 
and dry sand each weigh one hundred pounds per cubic foot. 
If now we use damp sand, containing about 3.5 per cent, water, 
instead of dry sand, and measure the materials, we would have 
four cubic feet of damp sand to one cubic foot of cement; but 
damp sand would contain only about 4 X 75 = 300 pounds of 
dry sand, and we would really have a one-to-three mixture 
instead of a one-to-four. 

ART. 32. IMPURITIES IN SAND 

264. The usual specification for sand is that it shall be 
" clean, sharp and siliceous." We have shown that it need not 
be siliceous, and we have also noted that one authority con- 
siders that it need not be sharp, though this latter does not 
appear to be proven; let us see what interpretation should be 
given to the word " clean" if it must be retained in all speci- 
fications for sand. 

Mr. E. C. Clarke, in the tests for the Boston Main Drainage 
Works, showed that "clay in moderate amounts" (ten per 
cent, to thirty per cent, of the sand) "does not weaken cement 
mortars." Calcareous marl might be considered an impurity, 
but we have seen that M. Alexandre found that sands contain- 
ing this material gave excellent results. On the other hand, 
there seems to be no doubt that loam, peaty matter or humus 
will very materially decrease the strength of mortars, or even 
destroy them entirely. Likewise, decayed particles of some 
kinds of stone, or grains which readily break up into thin scales, 
should be strenuously avoided. 

265. Detection of Impurities. Clean sand when rubbed in 
the hand will not leave fine particles adhering to it, but should 
the sand not prove to be clean, the character of the impurities 
should be investigated before finally rejecting it. When there 
is not time for making proper tests, it will, of course, be safest 
to use only such sand as has no foreign matter whatever; but 
when strictly pure sand can only be obtained at great cost, tests 
may show that a small percentage of impurities may be tolerated. 



IMPURITIES IN SAND 183 

Another simple test, beside the one of rubbing in the hand, 
is to place a little of the sand in a test tube filled with water; 
if any impurities are present, they may separate from the sand 
on account of their lighter weight, or if in a very fine state of 
division, the water may be rendered murky in appearance. 
This test is not absolute, however, especially for calcareous 
sand, as the fine particles of limestone will give the murky 
appearance to the water, although not objectionable except on 
account of their extreme fineness. 

The use of poor sand will result in a larger proportionate 
decrease in strength for a mortar containing a large amount 
of sand than for one made with a small amount. The effect of 
incorporating various foreign substances in cement mortar is 
treated in Art. 49. As some of these substances may occur 
in sand, the article referred to should be read in connection 
with this subject. 

266. SAND WASHING. When impurities occur, they may 
sometimes be removed by washing, but such work must be 
carefully inspected if the foreign matter be of a really danger- 
ous character. 

In the construction of the Canal at the Cascades, Columbia 
River, Oregon, quite an elaborate concrete plant was estab- 
lished, which had in connection a sand and gravel washer and 
separator. 1 This consisted of a tube about two and one-half 
feet in diameter and seventeen feet long, made of one-quarter- 
inch boiler iron and revolving about an axis slightly inclined to 
the horizontal. Angle irons were riveted on the inside of the 
tube to carry the material up on the side and drop it again, 
while a spray of water issued from a perforated pipe inside the 
tube. The materials were separated by screens near the lower 
end of the tube. The material contained considerable earthy 
matter and is said to have been fairly well washed by this pro- 
cess. 

Another style of sand washer was designed by the contract- 
ors for the construction of Lock No. 3, improvement of Alle- 
ghany River. 2 The sand contained earthy matter and some 
coal, the latter being hard to remove by ordinary processes. A 



1 Report of Lt. Edw. Burr, Report Chief of Engineers, 1891, p. 3334. 

2 W. H. Rober, Engineering News, Feb. 16, 1899. 



184 CEMENT AND CONCRETE 

large barrel or tank, nine feet in diameter and seven feet high, 
was provided with double floor, the upper one being pierced 
with one-inch holes. Paddles were attached to a vertical shaft 
in the axis of the tank and revolved by suitable gearing, while 
water was forced into the space between the two floors. The 
water finding its way through the holes in the upper floor, passed 
up through the sand and overflowed at the top, carrying with 
it the coal and sediment. The cost of washing is said to have 
been about seven cents per cubic yard, but it is evident that 
methods of handling would have to be quite perfect to keep the 
cost at so low a figure. 

ART. 33. CONCLUSIONS. WEIGHT. COST 
267. REQUIREMENTS FOR GOOD SAND. In conclusion, then, 
we may say that good sand may consist of grains of almost 
any moderately hard rock that is not liable to future alteration 
in the work. The grains may be of any shape, but preferably 
should be sharp and angular or lenticular in form, not rounded 
and smooth. The sand should not contain such impurities as 
loam or humus, but for most purposes a small percentage of 
clay or fine rock dust is not objectionable. Clay should not, 
however, be permitted in sand for use in sea water. 

Coarse grained sands are better than fine grained ones, but 
a mixture of fine and coarse is excellent, especially where but 
a small amount of cement is used, because such a mixture con- 
tains less voids and will make a less permeable mortar, while giv- 
ing a good strength. As might be expected, the deleterious effect 
of poor sand is more apparent the larger the dose of sand used. 
268. Weight of Sand. It is evident from what has pre- 
ceded that the weight of sand per cubic foot will vary greatly, 
not only with the character of the rock from which it came, 
but also with its physical condition. Natural sand, as it or- 
dinarily occurs, will weigh about as follows, according to its 
condition: 

Moist and loose 70 to 90 pounds per cu. ft. 

Moist and shaken 75 to 100 " 

Dry and loose 75 to 105 " 

Dry and shaken 90 to 125 " 

When settled in water, weight of wot 

I, voids full 100 to 140 



WEIGHT AND COST OF SAND 185 

If the rock from which the sand is made weighs, say, one 
hundred sixty pounds per cubic foot solid (specific gravity, 
2.56), then the sand will weigh per cubic foot 120, 100, and 
80 pounds, for voids of 25, 37.5 and 50 per cent., respectively. 

269. Cost of Sand. The cost of sand will, of course, vary 
with the locality. In exceptional cases where it is found di- 
rectly at the works, it may not cost more than twenty to thirty 
cents per cubic yard to deliver it on the mixing platform. If 
it has to be pumped from the bed of a river or lake and can be 
conveyed to the work in scows with a tow of not more than 
ten miles, it may be delivered at the work for from forty to 
sixty cents per cubic yard. If it must be hauled in wagons for 
some distance, it may cost from fifty cents to one dollar per 
yard; and again, if sand is so difficult to obtain that it must be 
made by crushing rock, it may cost from one dollar to three 
dollars per yard. Usually from sixty cents to a dollar is a fair 
price for sand. Several examples of cost of sand will be given 
in connection with the subject of cost of concrete. 



CHAPTER XII 

MORTAR: MAKING AND COST 
ART. 34. PROPORTIONS OF THE INGREDIENTS 

270. CAPACITY OF CEMENT BARRELS. Since there is no 
standard size for cement barrels, the capacities vary consider- 
ably, Portland cement barrels ranging from 3.1 to 3.6 cu. ft., while 
natural cement barrels contain from 3.4 to 3.8 cu. ft. In Ger- 
many cement is packed to weigh three hundred ninety-six 
pounds per barrel, gross, the net weight being about three hun- 
dred seventy-five pounds. American Portland usually weighs 
four hundred pounds gross or about three hundred eighty 
pounds net. 

In 1896 the Boston Transit Commission had a number of 
measurements made of the capacity of Portland cement bar- 
rels, and these have been compiled by Mr. Sanford E. Thompson. 1 
Table 58 presents some of the averages obtained from this series 
of tests. It is seen that the capacity of the barrels varied from 
3.12 to 3.50 cu. ft., the mean, volume being 3.29 cu. ft. The 
difference between the capacity of the barrel and the volume 
of the packed cement contained is due to the fact that there 
is usually a small space beneath the head not filled with cement. 
A barrel of packed cement makes about 1.25 barrels, measured 
loose . 

271. Natural cements made in the East are packed to 
weigh three hundred pounds net, while some of the Western 
natural cements weigh but two hundred sixty-five pounds per 
barrel net. Any of the natural cement factories will doubtless 
pack their cement to suit customers on large orders, and there 
seems to be little reason for this variation in weight between 
the West and the East. There would perhaps be some trouble 
in getting three hundred pounds of a very light, finely ground, 
natural cement in the ordinary sized barrel, but two hundred 



1 Engineering News, Oct. 4, 1900. 

186 



PROPORTIONS IN MORTAR 



187 



TABLE 58 
Capacity of Portland Cement Barrels 



. 


HIGHEST. 


LOWEST. 


MEA.N. 


Height of barrel between heads, feet .... 
Capacity between heads, cubic feet 


2.19 
3 50 


2.01 
3 12 


2.09 
3 29 


Volume of packed cement in barrel, cubic feet . 
Volume of loose cement in barrel, cubic feet . 
Net weight of cement in barrel, pounds . 
Weight per cubic foot of cement as packed in 
barrel pounds 


3.48 
4.19 
387.0 

123 16 


3.03 
3.75 
370.7 

113 81 


3.18 
4.07 
377.4 

118 79 


Weight per cubic foot, loose, pounds .... 


100.49 


88.52 


92.63 



NOTE. Results are averages of thirty-one tests with seven brands, four 
of which were American. The above data compiled by Sanford E. Thomp- 
son and published in Engineering News of Oct. 4, 1900. 

eighty pounds may be put in a barrel without difficulty, and it 
would seem that a compromise might be made on this weight. 

272. QUANTITY OF SAND. The amount of sand to be used 
in mortar will depend entirely on the character of the work 
and the quality of the cement and sand. If it is merely a 
matter of strength to be developed, no special care need be 
taken to have the voids in the sand rilled with cement, but if 
an impervious mortar is desired, the mortar must not be too 
poor in cement, even though only a moderate strength is re- 
quired. 

In France the proportions of cement and sand are usually 
given in terms of kilograms of cement to one cubic meter of 
sand. In England and America the proportions are usually 
given by volume, as so many parts of cement to one of sand, 
while in Germany the proportions are given by weight. The 
bulk of cement varies so much according to the degree of pack- 
ing, and the volume of sand is so varied by the amount of mois- 
ture contained, that the German method of stating proportions 
by weight seems to be the most logical one to adopt. 

273. Proportions by Volume. It has been shown that the 
volume of a given quantity of cement may vary twenty-five 
per cent, according as it is measured packed or loose, and that 
likewise the volume of sand may vary twenty per cent, accord- 
ing to the amount of moisture contained. This makes it ne- 
cessary to take great precaution in proportioning mortars by 



188 CEMENT AND CONCRETE 

volume if the desired richness of the mortar is to be assured. 
Nevertheless, mortars for use in actual construction are usually 
proportioned by volume. The usual method is to state the 
proportions as one part of packed cement (as it comes in the 
barrel or bag) to so many parts of loose sand, but proportions 
are sometimes stated as volumes of loose sand to one volume 
of loose cement. 

274. Equivalent Proportions by Weight and Volume. As 
cement is now so frequently sold in sacks of one-fourth barrel 
each, in which the cement is not so compact as in a barrel, we 
have assumed the contents of a barrel to be 3.45 cu. ft. for 
Portland, and 3.75 cu. ft. for natural, which are somewhat 
higher than the mean actual capacities of stave barrels as 
shown by tests. At three hundred eighty pounds and two 
hundred eighty pounds net weight respectively for Portland 
and natural, this is equivalent to one hundred ten pounds per 
cubic foot and seventy-five pounds per cubic foot packed. If 
we also assume that loose cement weighs eighty-five pounds per 
cubic foot for Portland and sixty pounds per cubic foot for 
natural; and that loose, dry sand weighs one hundred pounds 
per cubic foot, while loose, damp sand weighs eighty pounds per 
cubic foot, we may obtain the following comparisons, Table 59. 

275. It is evident that in all specifications and in reports 
of tests, as well as in the use of cement, the method of stating 
proportions should be made clear, and in interpreting the re- 
sults of tests this must be borne in mind. For instance, in 
tests to compare the value of limestone screenings with quartz 
sand, proportions by weight will favor the quartz, while pro- 
portions by volume will favor the screenings, since the latter 
are lighter. 

276. Richness of Mortar. Mortars containing small amounts 
of sand are often stronger than neat cement mortars. Es- 
pecially is this true of most natural cements. Some of these 
will give as high strengths when mixed with two parts sand by 
weight as when neat, and usually the one-to-one mortars are 
stronger than the neat mortars. These remarks refer to tensile 
tests where a good quality of sand is used and the mortars are 
three months old or more. The neat cement mortars gain 
their strength more rapidly, short time tests usually not show- 
ing the results mentioned, Portland cements of good quality 



PROPORTIONS IN MORTAR 



ISO 



TABLE 59 



Comparison of Proportions by Weight and Volume 



PARTS DRY 
SAND 

TO ONE 
CEMENT BY 

WEIGHT. 



EQUIVALENT PARTS SAND, PROPORTIONS STATED BY VOLUME. 



PORTLAND CEMENT. 



II 



1.10 
2.20 
3.30 
4.40 
5.50 
6.60 



1 ol 

0200 



1.38 

2.75 
4 12 
5.50 

6.88 

8.25 



0.85 
1.70 
2.55 
3.40 
4.25 
5.10 



1.06 
2.12 
3.19 
4.25 
5.31 
6.38 



NATURAL CEMENT. 



||| 
3 < s3 



0.75 
1.50 
2.25 
3.00 
3.75 
4.50 



0.94 

1.88 
2.81 
3.75 
4.69 

5.62 



fi 



0.60 
1.20 
1 80 
2.40 
3.00 
3.60 



i 5 ! 



0.75 
1.50 
2.25 
3.00 
3.75 
4.50 



In preparing the above table the following assumptions are made : 



MATERIAL. 


WEIGHT 

IN A 

BARREL. 


VOLUME 

OF A 

BARREL. 


WEIGHT PER CUBIC FOOT. 


Packed. 


Loose 
Dry. 


Loose 
Damp. 


Portland cement . 
Natural cement . 
Sand .... 


380 

280 


3.45cu. ft. 
3.75cu. ft. 


110 

75 


85 
60 
100 


'so' 









usually give about the same tensile strength neat and with 
one part sand by weight. Tests showing the rate of decrease 
of strength with added sand are discussed in 363 to 365. 

Portland cements are usually mixed with from one to three 
parts sand by weight, and natural cements are mixed with 
from one to four parts by weight (or three-fourths part to 
three parts by measure). For certain special purposes poorer 
mortars are sometimes employed. To arrive at the proper 
proportion to use in mortar for a given purpose, the tables of 
strength given in Chapter XV will be of value. 

277. Effect of Pebbles. If the sand contains pebbles, the 
proportions should be considered in a little different way. 
Suppose we make a one-to-three mortar with sand that con- 
tains ten per cent, of pebbles. We have in reality, then, 3 X .90 
= 2.7 parts of sand to one of cement, and .3 part pebbles em- 
bedded in this richer mortar. This point is of special signifi- 
cant in making concrete from gravel containing some sand, or 



190 CEMENT AND CONCRETE 

from broken stone from which the fine particles or screenings 
have not been removed. Such fine particles serve to weaken 
the mortar by increasing the dose of sand, while the pro- 
portion of aggregate is diminished. In using aggregates con- 
taining some fine material, then, or in using sand containing 
pebbles or fine gravel, one should not permit himself to be de- 
ceived as to the actual richness of the resulting mortar or 
concrete. 

278. AMOUNT OF WATER FOR MORTAR. The amount of 
water required for mortar will vary with the proportion of sand 
to cement, the character and condition of the ingredients, the 
weather, and the purpose which the mortar is to serve. If the 
water is stated as such a percentage of the combined weight 
of cement and sand, the amount required for a rich mortar 
will be greater than for a poor one, since the cement requires 
more water than the sand. Fine cement will require more 
water than coarse; the same is true of sand. Sand from ab- 
sorbent rock will require a larger amount of water. On a hot, 
dry day, more water must be used to allow for evaporation; 
and again, if the mortar is to be placed in contact with brick 
or porous stone, the mortar must be more moist than when 
used in connection with metal, or with hard rocks such as 
granite. All of these points must be borne in mind when 
determining the proper consistency for a given purpose. 

279. We may arrive at the approximate amount of water 
required in the following manner: find what proportion of 
water is required for the neat cement. This will vary among 
different samples, and especially between Portland and natural 
cements; the former requiring twenty to twenty-eight per cent, 
of water (by weight), and the latter thirty to forty per cent. 
Then find the amount of water required to bring the sand alone 
to the consistency of mortar. This will vary considerably, fine 
sand requiring much more water than coarse, etc., as men- 
tioned above. Having these two quantities, we may find the 
amount of water required for a mortar having any given pro- 
portions of these samples of cement and sand. Thus, suppose 
we find that the neat cement requires twenty-five per cent, 
water and the sand ten per cent, water to bring them to the 
proper consistency. If we wish to make a one-to-three mortar 
from these ingredients, using one hundred pounds of cement, the 



MIXING MORTAR 191 

required amount of water is (100 X .25) -f (100 X 3 X .10) 
= 25 + 30 = 55 pounds. 

280. However, it will usually be better to experiment di- 
rectly upon the mixture which it is proposed to use, and for 
this purpose the following rule will be found of value. For 
ordinary purposes, that amount of water should be used which 
for given weights of the dry ingredients will give the least 
volume of mortar with a moderate amount of packing. In the 
actual use of mortars it is not practicable to state that a cer- 
tain definite amount of water shall always be used with given 
quantities of the dry materials. It is the resulting consistency 
of the mortar that must be specified and insisted upon, while 
the amount of water required to produce this consistency will 
vary from day to day and must be left to the discretion of the 
inspector or foreman. For a discussion of the relation of con- 
sistency to the tensile strength of the mortar, see Art. 46. 

ART. 35. MIXING THE MORTAR 

281. Having decided upon the proportions of cement, sand 
and water, it remains to incorporate these into a plastic, homo- 
geneous mass. The size of the batch should be so adjusted, if 
possible, that a full barrel of cement shall be used, and for 
careful work the amount of sand should be weighed instead of 
measured. Where this is impracticable, the condition of the 
sand from day to day, as regards the amount of moisture con- 
tained, should be taken into account (see 262 and 263). 

Mortar is usually mixed by hand, but where large amounts 
are to be used, machine mixers may profitably be introduced. 

282. HAND MIXING. For hand mixing, a water tight plat- 
form or shallow box should be used, of such a size that the given 
batch will not cover the bottom more than four inches deep. 

If the sand is measured, a bottomless box, provided with 
two handles at each end, will be found more convenient than 
the bottomless barrel which is often employed for this purpose. 
When the sand is delivered on the mixing platform in barrows, 
the latter may be fitted with rectangular boxes to avoid re- 
measuring. A two-wheel cart, the box of which may be in- 
verted to discharge the contents on the mixing platform, will 
also be found very serviceable when the runway is suited to 
such a vehicle. 



192 CEMENT AND CONCRETE 

The proper amount of sand is evenly spread on the plat- 
form, the cement is then dumped on top of the sand and spread 
out over it to an even thickness. With either hoes or shovels 
the dry materials are then thoroughly mixed, until, when a 
small amount is taken in the hand, it will appear of uniform 
color throughout. From two to five turnings of the materials, 
according to the expertness of the workmen, will be required to 
produce this result. The dry mixture is then drawn to the 
edges of the platform to form a ring, and the requisite amount 
of water is added at one time in the center. The mixture is 
then gradually incorporated with the water, and the mass is 
thoroughly worked until plastic and homogeneous. Should it 
be found that too little water has been used, a small amount 
may be added from a sprinkling pot or rose nozzle, but the 
mass should always be worked over again after such addition. 
Four shovels may be used at one platform, but if the mixing is 
done by hoes, not more than two can be used to advantage 
with a batch of ordinary size. 

Some engineers prefer one method and some the other, but 
in whatever manner done, the mixing should not be stinted. 
From two to four turnings of the mass are usually considered 
sufficient, but as a general rule it will be found that further 
mixing, beyond that required to just give the mass a uniform 
appearance, will be amply repaid in the strength of the result- 
ing mortar. (See Art. 47.) 

283. MACHINE MIXING. Where large quantities of mortar 
are required, machine mixers are sometimes used. A very 
complete plant for mortar-making was used in building the 
Titicus Dam. 1 In this case machinery was used in measuring 
the proportions of, cement and sand as well as in making the 
mortar. The measuring apparatus consisted of two cylindrical 
troughs, one for cement and one for sand. Each trough was 
divided, by means of six radial vanes and four discs, into eigh- 
teen equal compartments. These cylinders revolved in cast 
iron boxes which were so constructed as to serve as hoppers 
for filling the compartments. Three compartments were pre- 
sented to the hoppers at once, and slides were provided by 
which any of the hoppers could be cut off at will. The cylin- 



Engineering Record, August 3, 1895. 



INGREDIENTS REQUIRED 193 

ders being geared to the same pinion, it was possible, by means 
of the slides, to make any desired proportion of cement and sand 
from neat cement, to three parts sand to one cement. 

The mixing machine " consisted essentially of a semi-cylindri- 
cal wrought-iron trough with extended flaring sides, with ele- 
ments slightly inclined to the horizontal, and in its axis a re- 
volving shaft with oblique radial blades set at an incline of 
ninety degrees to each other and of a length to just clear the 
bottom of the trough." 

284. Another form of machine that is sometimes employed 
consists of a semi-cylindrical trough in which rotates an axis 
carrying a blade in the form of a screw. The materials are 
fed to the mixer at one end and the screw mixes them while 
working the mass toward the other end. 

ART. 36. COST OF MoRTAR 1 

285. INGREDIENTS REQUIRED FOR ONE CUBIC YARD OF 
MORTAR. The character of the ingredients used in making cement 
mortar varies so much that it is difficult to accurately deter- 
mine the quantities of materials required for a proposed mortar 
except by experimenting with the materials that are to be 
employed. It has been shown that the weights per cubic foot 
of both cement and sand vary greatly according to the condi- 
tions of packing, the moisture, etc. The percentage of voids 
in the sand is one of the most important variations affecting 
the amount of mortar made with certain materials mixed in 
given proportions. The consistency of the mortar also has a 
marked effect, and different cements show a considerable varia- 
tion in the volume of mortar that a given weight will yield. 
In any general treatment of the question, then, we may expect 
only approximate results, and the tables given in this connec- 
tion must be considered in this light. 

286. Results of Experiments. The tests from which Tables 
60 and 61 were derived, were made with a natural sand weigh- 
ing about one hundred pounds to the cubic foot, dry, and having 
about three-eighths of the bulk voids. The grains varied in 
size from 0.01 in. to 0.1 in. in diameter with a few grains out- 



1 Portions of this article were contributed to " Municipal Engineering," 
and appeared in that magazine, Feb., 1899. 



194 



CEMENT AND CONCRETE 



o 



Proportions by Volume Dry 
Loose Sand to Loose Cement, 
Loose Cement Assumed at 
85 Ibs. per Cu. Ft. 


02 3 



g 


O 5 GO CO S S 


Cement. 


I-I 


~ 


i' co <N I-H i-i i-i ' 


Pounds. 


- 


linn ; ; 


Proportions by Volume Dry 
Loose Sand to Packed Cement, 
Cement Assumed at 114 Ibs. per 
Cu. Ft. or 380 Ibs, per Bbl. of 
3.33 Cu. Ft. 


02 53 

O 


- 


-^ ^ CM t^ Oi > i CO 
O r-- CO 30 CO OS d 


Cement. 


Bbls. 
of 

380# Net. 


c* 


S3I3532S 


Pounds. 


.- 


OOOOtOOOO 
GO O T-I CO CD >O -^ -^ 


Proportions by Volume Dry 
Loose Sand to Packed Cement, 
Packed Cement Assumed 
at 104 Ibs. per Cu. Ft. or 380 Ibs. 
per Bbl. of 3.65 Cu. Ft. 


JS 

o 


- 


8 CD t- -* QO O 
vO t- CO CO Oi ' 


I 

o 


H 


X 


O t^~ "^ CO C^ CO "^ 

* i-H CO O CO CO rH ' 

!>' T}H CNI <M' i-H r-i i-i ' 


Pounds. 

i 


, 


O O O O O 1C O 
I-H 00 CO CO I-H O CO 
00 *Q O t- CD *Q <# ' 
<M i i i i 


PROPORTIONS BY 
WEIGHT, DRY SAND AND 
CEMENT. 


1^ 


e 


i^ co * co I-H ro 
_ \O t~ CO CO OS Oi ' 

. 


Cement. 


H 





O CO CD O CO OS O . 
"* O t- O O <M_ i-t 


Pounds. 


>e 


ssssgis 

CO O O t- O * ^ 

<M l-H i-H 


JLO 1 OI CIKVS SXHV<I 


8 


**- 



INGREDIENTS REQUIRED 



195 



13 

3 

W 



S *; 

<4H 

Is 
3 ^ 

1 

2 : 

1 2 
*O a 

I? 



ons by Volume Dry 
d to Loose Cement, 
ment Assumed at 
r Cubic Foot. 



Proportio 
oose Sand 
se C 

60# 




d 
ed 
ed a 
t. or 
Bbl. 



Proportions by Vol- 
ume Dry Loose 
Sand to Packed Ce- 
ment, Packe 
Cement A ssurn 
75# per Cu. Ft 
0# Net pe 



sponoj 



spunoa; 



P 

ce S 



spunoj 



portions b 
me Dry Lo 
d to Pack 
ent, Cem 
umed at 71 
ubic Foot 
5# Net per 



PROPORTIONS BY 
EIGHT, DRY SAND AND 
CEMENT. 



00 =fc^ 

' 



spunoi 



r i r- 
O CO CO . 
O . 



t O O I 

<* CC5 2<1 . . 
Ir^ CO S<i T-H . . 



SCO '-O CO 
CO CO O 
CO CO >^' -H' 



>O CO C5 t>. ' ' 

* T* t- . 

00 ^n i -H . . 



O CO CO 00 t- 
O J> CO CO OS 



t-- t~ *> O t^ 



t~ *> O t^ O 

i i t- O O CO . 



O O *O O O 
*^H xo ^ O l^- 
(M C<l CO O Tf 



80t-rt<CO 
O t~-- GO CO 



p CD 

cd T* <M" <M' rn 



O O O O 
-^ I-H CO lr^ 
CN 7^ 1>- iO 

<M T i 



*O lO CO r*H O 
rt; rj^ CO 
CO Tji 3^1 <N I-H 



3*0 S^ 
n ^^ 



spunoj 



JO I OX 



SXHYd 



1-1 <M O -<ti t^ OS 
iO t-. CO CO CO CO 



10 CO 
t- Tfl CO (M - rH rH r-J 



SCO CO t^ CO CO 
. OS rtj >0 CD 
CO* *' CO (N <N r-< 



>O 1-H 

CO O' CO" <M~ T<i r-i 



O O O O O O 
""^ CO t^- ^ !> t^- 

(N CO OS t^ iO * 



O <-> <N CO -^ *0 



196 CEMENT AND CONCRETE 

side of these limits. The consistency of the mortar was such 
that when struck with the shovel blade the moisture would 
glisten on the smooth surface thus formed. In the experiments 
the proportions were determined by weight, and the results for 
proportions by volume were deduced from them. The results 
for neat natural cement mortar and for the natural cement 
mortars containing more than four parts sand by weight were 
derived by analogy. 

287. Explanation of Tables. The first section of Table 60 
gives the amount of materials required for Portland cement 
mortar when the proportions are stated by weight; the second 
and third sections refer to proportions by volume of loose sand 
to packed cement when the size of the cement barrel is as- 
sumed at 3.65 cu. ft. and 3.33 cu. ft., respectively. The fourth 
section gives the materials required when the proportions are 
given in terms of loose sand to loose cement. Likewise, the 
first section of Table 61 for natural cement refers to proportions 
by weight; the second, third and fourth sections, to propor- 
tions by volume of loose sand to packed cement when the 
cement weighs 265 pounds, 280 pounds and 300 pounds, net. 
per barrel, respectively; while the fifth section refers to propor- 
tions of loose sand to loose cement. 

As has been shown, the method of stating proportions by 
weight is the most accurate, but when the sand does not ap- 
proximate the weight of 100 pounds per cubic foot when shoveled 
dry into a measure, the sections of the tables referring to weight 
proportions may require a correction, and it may be simpler 
to use the sections giving proportions by volume of loose sand 
to packed cement. The method of stating proportions by vol- 
umes of loose sand to loose cement is to be deprecated, but since 
it is occasionally used, provision is made for it in the tables. 

In using those portions of the tables where the proportions 
are stated by volume, it should be borne in mind that if the 
sand is damp when used it will weigh less per cubic foot, and 
hence more, by measure, will be required to make a cubic yard 
of mortar. 

288. Estimating Cost of Mortar. With the data given in 
Tables 60 and 61 and a knowledge of unit prices of the mate- 
rials used in the mortar, one may estimate the cost of the ma- 
terials in a given quantity of mortar. The cost of the mixing 



COST OF MORTAR 



197 



will, of course, depend upon the cost of labor, the method em- 
ployed, etc., and may vary from fifty cents to a dollar and 
fifty cents per cubic yard. If we assume, for illustration, that 
natural cement can be delivered on the mixing platform for 
$1.10 per barrel of 280 pounds net, that sand costs 60 cents 
per cubic yard, and the mixing costs $1.00 per yard of mortar, 
then we have for the cost of a mortar composed of one part 
cement to two parts sand by weight: 

3.46 bbls. cement at $1.10 $3.80 

0.72 cu. yd. dry sand at .60 43 

Cost of mixing per cu. yd 1.00 

Total cost of one cu. yd. of mortar $5.23 

289. For approximate results, Tables 62 and 63 give the 
cost of the materials used in a cubic yard of mortar for different 
prices of cement. In Table 62 the proportions by weight only 
are indicated, since for Portland the proportions by volume of 
loose sand to packed cement vary so little from proportions 
by weight. 

TABLE 62 
Cost of Portland Cement Mortar 

COST OF CEMENT AND SAND IN ONE CUBIC YARD OF PORTLAND CEMENT 
MORTAR. SAND, 75 CENTS PER CUBIC YARD 



COST OF PORT- 
LAND 
CEMENT PER 
BARREL 
OF 380 POUNDS 
NET. 


COST OF INGREDIENTS IN MORTAR, IN DOLLARS. 
Proportions in Mortar by Weight, Parts Sand to One of Cement. 





1 


2 


3 


4 


5 


6 


11.20 


8.90 


5.33 


3.89 


3.03 


2.56 


2.23 


2.02 


1.30 


9.62 


5.73 


4.17 


3.23 


2.72 


2.36 


2.13 


1.40 


10.36 


6.14 


4.44 


3.43 


2.88 


2.49 


2.24 


1.50 


11.10 


6.55 


4.72 


3.63 


3.03 


2.62 


2.35 


1.60 


11.84 


6.96 


5.00 


3.83 


3.19 


2.75 


2.46 


1.70 


12.58 


7.37 


5.27 


4.03 


3.35 


2.88 


2.57 


1.80 


13.32 


7.77 


5.55 


4.23 


3.51 


3.01 


2.68 


1.90 


14.06 


8.18 


5.82 


4.43 


3.67 


3.13 


2.79 


2.00 


14.80 


8.59 


6.10 


4.63 


3.82 


3.26 


2.90 


2.10 


15.54 


9.00 


6.38 


4.83 


3.98 


3.39 


3.01 


2.20 


16.28 


9.41 


6.65 


5.03 


4.14 


3.52 


3.12 


2.30 


17.02 


9.81 


6.93 


5.23 


4.30 


3.65 


3.23 


2.40 


17.76 


10.22 


7.20 


5.43 


4.46 


3.78 


3.34 


2.50 


18.50 


10.63 


7.48 


5.63 


4.61 


3.91 


3.45 


2.60 


19.24 


11.04 


7.76 


5.83 


4.77 


4.04 


3.56 


2.70 


19.98 


11.45 


8.03 


6.03 


4.93 


4.17 


3.67 


2.80 


20.72 


11.85 


8.31 


6.23 


5.09 


4.30 


3.78 


2.90 


21.46 


12.26 


8.58 


6.43 


5.25 


4.42 


3.89 


3.00 


2220 


12.67 


8.86 


6.63 


5.40 


4.55 


4.00 



198 



CEMENT AND CONCRETE 



TABLE 63 
Cost of Natural Cement Mortar 

COST OP CEMENT AND SAND IN ONE CUBIC YARD OF NATURAL CEMENT 
MORTAR. SAND, 75 CENTS PER CUBIC YARD 





o 




METHOD OF STATING 


gg 


COST OF CEMENT PER BARREL, DOLLARS. 


PROPORTIONS, AND 


<j g 




WEIGHT OF CEMENT IN 
ONE H \ KRKI 


CO H 
.0 






*0 


0.60 


0.70 


0.80 


0.90 


1.00 


1.10 


1.20 


1.30 


1.40 


1.50 







5.07 


5.92 


6.76 


7.60 


8.45 


9.30 


10.14 


10.98 


11.83 


12.68 




1 
2 
3 


3.50 

2.74 
2.23 


4.02 
3.10 
2.50 


4.54 
3.47 

2.78 


5.06 
3.83 
3.05 


5.59 
4.20 
3.32 


6.11 
4.57 
3.59 


6.63 
4.93 
3.86 


7.15 
5.30 
4.14 


7.67 
5.66 
4.41 


8.19 
6.03 
4.68 




4 


1.90 


2.12 


2.33 


2.54 


2.76 


2.98 


3.19 


3.41 


3.62 


3.83 







4.48 


5.23 


5.98 


6.72 


7.47 


8.22 


8.96 


9.71 


10.46 


11.20 




1 

2 


3.14 

2.48 


3.60 

2.80 


4.06 
3.12 


4.52 
3.45 


4.98 
3.77 


5.44 
4.09 


5.90 
4.42 


6.36 
4.74 


6.82 
5.06 


7.28 
5.38 




3 


2.04 


2.28 


2.52 


2.76 


3.00 


3.24 


3.48 


3.72 


3.96 


4.20 




4 


1.75 


1.94 


2.13 


2.32 


2.51 


2.70 


2.89 


3.08 


3.27 


3.46 


By volume ; parts dry j 
loose sand to packed 1 



1 


5.07 
3.13 


5.92 
3.58 


6.76 
4.02 


7.60 
4.46 


8.45 
4.91 


9.30 
5.36 


10.14 
6.M) 


10.98 
6.24 


11.83 
6.69 


12.68 
7.14 


cement. Cement as--! 


2 


2.29 


2.57 


2.85 


3.14 


3.42 


3.70 


3.99 


4.27 


4.55 


4.84 


sumed 265 Ibs. per bbl. 
of 3 75 cu. ft. 


3 


1.83 


2.04 


2.24 


2.45 


2.65 


2.86 


3.06 


3.26 


3.47 


3.67 




4 


1.63 


1.79 


1.95 


2.11 


2.27 


2.43 


2.59 


2.75 


2.91 


3.07 


By volume ; parts dry f 


1 


2.94 


3.36 


3.77 


4.19 


4.61 


5.03 


5.44 


5.86 


6.28 


6.70 


loose sand to packed 
cement. Cement 9*8 I 


2 


2.22 


2.50 


2.77 


3.05 


3.32 


3.60 


3.87 


4.15 


4.42 


4.70 


sumed 300 Ibs. per bbl. | 


3 


1.82 


2.02 


2.22 


2.42 


2.62 


2.82 


3.02 


3.22 


3.42 


3.60 


of 3.75 cu. ft. t 


4 


1.59 


1.75 


1.91 


2.06 


2.22 


2.38 


2.53 


2.69 


2.85 


3.00 



In Table 63 the cost of materials in one cubic yard of natural 
cement mortar is given, 1st, for various parts of sand to one 
of cement by weight when 'the cost of cement refers to a barrel 
of 265 pounds; 2d, when this cost is for a barrel of 300 pounds 
net; 3d, for various parts sand to one cement when the propor- 
tions are expressed as parts by volume, dry loose sand to one 
volume of packed cement weighing 265 pounds per barrel; and 
4th, when the proportions are expressed as parts of dry loose 
sand to one volume of packed cement weighing 300 pounds per 
barrel. The quantities in the table are based upon the as- 
sumption that the sand used is similar to that used in the ex- 
periments from which Tables 60 and 61 were derived, and that 
the cost of sand is seventy-five cents per cubic yard. 



COST OF MORTAR 199 

I _ 

290. Example. To indicate the use of these tables, let us 
determine the cost per cubic yard of natural cement mortar 
composed of one volume of packed cement to three volumes of 
loose dry sand when the cement weighs 300 pounds per barrel, 
net, and costs $1.25 per barrel, while sand costs $1.00 per cubic 
yard. In the fourth section of Table 63, opposite three parts 
sand and under $1.20 and $1.30, we find, respectively, $3.02 and 
$3.22; then with cement costing $1.25 and sand $0.75, we should 
have cost of mortar per cubic yard $3.12. But in our example 
sand is assumed to cost $1.00 per cubic yard, or twenty-five 
cents more than the price for which the tables are computed, 
and from Table 61 we find that for this mortar 0.83 cubic yard 
of sand is required. We must therefore add to $3.12, .83 X 25 
= 21 cents, giving $3.33 as cost of materials in one cubic yard 
of the mortar. The cost of mixing the mortar must be added 
to obtain the total cost per cubic yard. 



CHAPTER XIII 

CONCRETE : AGGREGATE 

291. Cement concrete is composed of a mixture of cement 
mortar and fragments of stone, brick or other moderately hard 
substances to which the mortar may adhere. Put in place while 
plastic, it soon obtains a strength and hardness equal to good 
building stone. This property, combined with its cheapness 
and adaptability to monolithic construction, renders it one of 
the most useful of engineering materials. 

ART. 37. CHARACTER OF AGGREGATE 

292. MATERIAL FOR AGGREGATE. Many of the points men- 
tioned concerning the selection of a good sand are also applicable 
to broken stone. The latter may be produced from almost 
any moderately hard rock, provided it is not subject to decay. 
The best material for broken stone is a rather hard and tough 
rock, which breaks into angular fragments with surfaces that 
are not too smooth. 

Gravel makes a good aggregate, although its surfaces are too 
smooth and rounding to give the best results. Coarse gravel 
may be improved by running it through a rock crusher to render 
some of the fragments angular and rough. A mixture of gravel 
and broken stone gives excellent results (see 454). The gravel 
assists the compacting of the mass, and the fragments of broken 
stone furnish a good bond. A mixture of this kind also leaves 
but a small percentage of voids in the mass, and this decreases 
the amount of mortar required. 

293. Sandstones are sometimes said to be better than lime- 
stones, but this will depend on their relative hardness and 
structure, and the use to which the concrete is to be put; no 
general rule will apply. Some limestones seem to be particu- 
larly adapted to concrete-making, as the cement adheres to the 
surface so firmly. Granite, syanite and trap are excellent for 
the purpose. Fragments of brick and of other burnt clay 

200 



CHARACTER OF AGGREGATE 201 

products give good results up to the limit of the strength of the 
pieces, but this limit is not high. Table 155 gives the results 
of transverse tests of concrete bars made under the author's 
direction, to show the comparative value of different kinds of 
stone. The results of these tests are discussed in 454. 

Mr. E. L. Ransome * has pointed out that "for fireproof 
work, care should be taken to avoid such aggregates as contain 
feldspar," and that limestone should not be used if the con- 
crete is likely to be subjected to a long continued red heat. 
The same writer mentions the fact that finely crushed granite 
may be inferior to finely crushed limestone for use in concrete; 
one reason for this being that, "owing to the brittle quality of 
granite, in crushing it is not only broken into small pieces, but 
many of these pieces are so bruised or contused that upon a 
little pressure being exerted upon them, such, for instance, as 
can be applied by the finger or thumb, they will crumble." 

294. The care required in the selection of a proper quality 
of broken stone or gravel will depend upon the required strength 
of the concrete. If a strong concrete is required, rich mortar will 
not be able to make up a deficiency in the strength of the stone ; 
but if a low strength is sufficient, and consequently a poor mor- 
tar is to be used, but little will be gained by having a very 
strong rock from which to obtain broken stone. In this case 
a rock which presents a good surface to which mortar may 
adhere is the principal requirement, and a very hard rock need 
not be insisted upon. 

295. PRESENCE OF SCREENINGS IN BROKEN STONE. it is 
frequently required that the broken stone shall be freed from 
all fine material, resulting from the crushing of the stone, before 
the mortar is added to form concrete. The wisdom of this 
requirement is not always clear and depends upon the kind of 
stone. It has already been stated that some forms of crusher 
dust or screenings give, if not too fine, most excellent results 
in mortar; this is especially true of limestone screenings. Again, 
to retain in the broken stone all of the screenings, will result in 
diminishing the percentage of voids in the aggregate, and thus 
decrease the amount of mortar necessary. 

On the other hand, if a stone is covered with a layer of 



Engineering Record, Nov. 17, 1894. 



202 CEMENT AND CONCRETE 

moistened, floury dust, it cannot be so readily brought in direct 
contact with the mortar, and if the mortar does reach the 
stone it is made less rich by the dust, which acts as so much 
fine sand. It must be said, however, that so far as our ex- 
periments go, they do not confirm this latter theory when a 
moderate amount of fine material is in question, especially with 
crushed limestone. There is a reason, however, in some cases 
why the very fine material which acts as sand should be screened 
out of broken stone, even if it is again used in the mortar for 
the concrete; the fine material collects in certain parts of the 
bin or pile, making the proportions irregular, so that one batch 
of concrete may have a rich mortar with a comparatively large 
amount of stone, while another may have a poor mortar with 
but little stone. If, therefore, all of that portion of the broken 
stone finer than, say, one-eighth of an inch, be screened out 
and used as so much sand in making the mortar, the resulting 
concrete will be better and more nearly uniform in quality. 

296. Impurities. Material that is really foreign, such as 
vegetable mold or loam, will be detrimental to the strength of 
the concrete. Even clay is not permissible here if it adheres to 
the stone, because if the surface of a piece of stone is smeared 
with clay, the mortar will not be able to adhere as well to that 
surface. Clay in a granulated form and not adhering to the 
stone may be permitted, however, .in small amounts, possibly 
as much as ten per cent., without seriously injuring the concrete 
for many uses. 

When old masonry is torn down, the stones are sometimes 
crushed for use in concrete, but such stones, having particles 
of mortar adhering to the surfaces, will not be of first quality 
for the purpose; their cheapness, however, will frequently out- 
weigh such objections. 

ART. 38. SIZE AND SHAPE OF THE FRAGMENTS AND THE 
VOLUME OF VOIDS 

297. As in the case of sand, the shape of the fragments and 
the degree of uniformity in size have an important effect on 
the proportion of voids in the mass, and all of these elements 
affect the value of broken stone for use in concrete. As in 
mortar each grain of sand should be completely covered with 
cement, so in concrete should each piece of stone be completely 



SIZE OF AGGREGATE 



203 



covered with mortar. As the pieces in a given volume of broken 
stone will have a smaller total superficial area when the frag- 
ments are large than when they are small, we should conclude 
that the larger fragments will require less mortar or be more 
thoroughly coated with a limited amount. From the same point 
of view we should expect that round fragments w^ould require 
less mortar than those of irregular shape. 

It is found however, in practice, that these theoretical con- 
siderations must be modified to correspond with the facts. 

TABLE 64 

Voids in Broken Stone and Gravel Varying in Granulometric 

Composition 







WEIGHT OF 




CHARACTER STONE. 


GRANULOMETRIC COMPOSITION. 


BROKEN 
STONE, LBS. 

PER 


PER CENT. 
VOIDS. 






Cu. FT. 




Limestone . . . 


K 
V 


88 

89 


47 
44 




F 


90 


43 




M 


91 


42 




C 


85 


46 




FSO, M 


91 


42 




F 50 , C 50 


94 


40 




Rao, F,. MSO 


102 


35 




K20 ? v 20 , F 20 , M 2 '}, C 20 


104 


34 


Potsdam sandstone 


C, 120 Ibs., K, 33 Ibs. 


ill! 

86 


29 
45 


u u 


M 


84 


44 


it U 


V33, F 67 


88 


43 


11 11 


V 33 , F 33 , M 33 


92 


40 


it it 


K 25 , V 25 , F 25 , M 25 


97 


36^ 


Gravel .... 


V 


110 


32 ' 


u 


F 


108 


33 


it 


M 


106 


34 


it 


V 33 , F 38 , M 33 


112 


30 


t< 


V 50 , M 50 


114 


29 


Potsdam sandstone I 


P. 6 cu. ft. on in. screen ) 
G. 2 " " in. to Jin. " f 


100 


39 


and gravel . . j 


P. 4 cu. ft. on in. screen j 
G.4 " " in. to Jin. " J 


109 


33 



NOTE. Stone jarred down in measure for all trials. 

K passed holes \ inch square, failed to pass holes T V inch square. 
y ^ M u i (i 



204 



CEMENT AND CONCRETE 



When the pieces of broken stone are too large, they do not 
bed themselves well in the matrix of mortar, but become wedged 
one against another, leaving voids in the concrete. While 
round fragments have a small superficial area in relation to 
their volume, have a small percentage of voids, and pack to- 
gether readily, yet they are lacking in ability to form a good 
bond, and hence do not give the best results. 

298. Relation of Size of Stone to Volume of Voids. As illus- 
trating the effect of the size of fragments and granulometric com- 
position of stone on the volume of voids, Table 64 gives a number 
of results obtained at St. Marys Falls Canal. 

Table 65 gives some of the results obtained by M. Feret in 
similar tests. 1 

TABLE 65 
Size of Stone and Volume of Voids 



COMPOSITION, BY WEIGHT, OF SMALL STONE. 


PER CENT. OF VOIDS BY VOLUME. 


Fragments Passing a Ring of 






90 mm. 


60 mm. 


40mm. 


20 mm. 
















Tl lr ^f 


and Retained on a Ring of 






60 mm. 


40mm. 


20mm. 


10 mm. 






1 











41.4 


52.1 





1 








40.0 


53.4 








1 





38.8 


51.7 











1 


41.7 


52.1 





1 





1 


35.6 


47.1 


1 


4 


1 


1 


33.5 


48.8 


1 


1 


1 


4 


35.6 


46.4 



The percentage of voids in a mass of broken stone of uniform 
size should be independent of what the size may be, and the 
first few lines in Table 65 show this to be nearly the case with 
the four samples tested. It is seen from both tables that the 
more complex mixtures give smaller percentages of voids, and 
that for all sizes the voids are much less in the gravel than in 
the broken stone. 



1 " Sand and Stone Used for Cement Mortar and Concrete," by M. Feret. 
Abstracted in Engineering News, March 26, 1892. 



SIZE OF AGGREGATE 



205 



299. M. Feret's Experiments. To show the effect of the 
variation in sizes of fragments on the strength of the concrete 
made, M. Feret experimented with four mixtures of three sizes. 
The proportions used in the mortar were one part by weight of 
Portland cement to three parts of Boulogne gravel, gaged with 
an amount of water equal to seventeen per cent, of the total 
weight of cement and sand. The volume of mortar used in 
each case was made equal to the volume of the voids in the 
stone. The concrete was thoroughly mixed and then rammed 
into a large cylindrical mold. After four months' exposure to 
the air, twelve cubes were cut from the cylindrical block, four 
cubes being cut from each of three consecutive horizontal 
layers. These cubes were placed in sea water and crushed 
after one month, being then five months old. The results of 
the tests are given in Table 66. 

TABLE 66 
Strength of Concrete. Varying Size Stone 











MEAN RESISTANCE IN KG. 










PER SQ. CM. 






VOL. OF 


WEIGHT 




GRANUL.OMETRIC 
COMPOSITION OF 


VOL. OF 
VOIDS PER 


MORTAR 
PER Cu. 


OF CON- 
CRETE 


OF 4 CUBES FROM 


IjL 




Cu. METER. 


METER OF 
STONE. 


PER Cu. 
METER. 




j 


g 












1 


1 


| 


sl a 


G 


M 


F 


Cu. Meter. 


Cu. Meter. 


Kg. 










4 


1 


1 


0.492 


0.492 


2296 


144 


143 


173 


153 


1 


4 


1 


0.494 


0.494 


2272 


141 


141 


154 


145 


1 


1 


4 


0.486 


0.486 


2276 


106 


121 


133 


120 


2 


2 


2 


0.478 


0.478 


2264 


115 


132 


151 


133 



Size "G" of broken stone passed a ring 60 mm. (2.4 inches) 
in diameter and was held by a ring 40 mm. (1.6 inches) in 
diameter; "M" passed 40 mm. (1.6 inches) ring and was held 
by a ring 20 mm. (0.8 inch) in diameter; while "F" passed 
the 20 mm. (0.8 inch) ring and would not pass a ring 10 mm. 
(0.4 inch) in diameter. 

The following conclusions are drawn from this table: (1) In 
each block the lower layers, which had been submitted to 
longer continued ramming than the upper layers, offered a 



206 , CEMENT AND CONCRETE 

greater resistance. (2) The mean resistance varied according 
to the granulometric composition of stone used, and was greater 
with the increasing proportion of large stone in each block. 
Since the amount of mortar used was in all cases equal to the 
volume of voids in the stone, the effect of voids on the strength 
was not noticeable. 

300. Further Experiments. Tables 153 and 155 give the 
results of some experiments made under the author's direction 
to test the effect of size and character of broken stone. In 
these tests the proportions are generally 35 pounds of cement 
to 105 pounds of sand and 3.75 cubic feet of broken stone, the 
stone being measured after jarring it down in the vessel. The 
amount of mortar made was sufficient to fill the voids in the 
stone when the latter did not exceed about thirty-three per 
cent (452, 454). 

It is seen that, in general, a higher result was given by mix- 
tures of various sizes than by any one size alone, and the fine 
stone gave higher results than the coarse. In these tests the 
effect of voids is shown, since in some cases there was not suf- 
ficient mortar to fill the voids. 

301. GRAVEL VS. BROKEN STONE AS AGGREGATE. The ele- 
ments entering into the analysis of the superiority of one kind 
of aggregate over another are given above, but since the ques- 
tion of the relative merits of gravel and broken stone is so 
frequently discussed, a word may be added here to show the 
special points involved in such a comparison. 

Gravel is composed of hard, rounded pebbles, the surfaces 
of which are usually quite smooth. On account of the manner 
of its formation and occurrence, the sizes of the pebbles are 
usually graded from coarse to fine. Occasional beds of gravel 
are found, however, in which the sizes of the several fragments 
are nearly the same. In broken stone the fragments are angular 
and usually have rough surfaces, though the degree of rough- 
ness depends upon the kind of stone. The sizes of the frag- 
ments as they come from the stone crusher vary from coarse 
to fine, but by regulating the crusher jaws and by screening, 
any desired size may be obtained. 

302. In determining the value of a certain material for 
aggregate, at least six characteristics are to be considered, the 
strength and durability of the stone, the size and shape of the 



GRAVEL VS. BROKEN STONE 207 

fragments, the volume of the voids, and the character of the sur- 
face to which the cement must adhere. As gravel is usually from 
the igneous rocks, its strength and durability are not often open to 
question. This may or may not be so in the case of broken 
stone, but the question of relative value of gravel and broken 
stone, which is so frequently conclusively settled either one way 
or another, seldom hinges on this point. As to the average size 
of the fragments, it is evident that as a general proposition it 
must be allowed that by proper screening either broken stone 
or gravel may be obtained of any desired size. 

Of the three remaining characteristics, the shape of the 
fragments, volume of voids, and character of surface, the first 
is probably the least important and the third of the greatest 
moment. The round pebbles of the gravel slide readily one on 
another, and do not interlock to give a good bond. The angular 
fragments of broken stone give a better bond, but on the other 
hand, if not thoroughly tamped, are likely to bridge, or arch, 
and thus leave holes in the mass. On account of the shapes of 
the fragments and because the sizes are usually more varied in 
gravel, the latter has generally a smaller percentage of voids; 
thirty to thirty-seven per cent, voids in gravel, and forty to 
fifty per cent, in broken stone, may be considered to give, in 
a general way, some comparative figures. Coming now to the 
character of surface, cement will not usually adhere so firmly 
to the smooth surface of the gravel as to the freshly broken 
surface of the fragments of stone, but this cannot be con- 
sidered a universal rule, for the strength in adhesion is not 
simply a matter of smoothness or roughness as it appears to 
the eye or the touch. The adhesion to limestone may be 
very much stronger than to a sandstone which has a rougher 
appearance. 

303. Summing up the relative advantages, we find that the 
gravel is suitable for concrete because, first, it is not likely to 
bridge and leave holes in the concrete; if mixed rather wet, very 
little tamping is required to compact it; and second, the usual 
smaller percentage of voids makes it possible to secure a com- 
pact concrete with a smaller amount of mortar than would be 
required for broken stone. On the other hand, the angular 
fragments of broken stone will knit together, as it were, to form 
a strong concrete if properly tamped, and the very important 



208 CEMENT AND CONCRETE 

question of a suitable surface for adhesion is usually in favor 
of the broken stone. It is evident, then, that this matter must 
resolve itself into a question of relative cost and suitabil- 
ity, and a general statement that either gravel or broken stone 
is superior, is not tenable. One experimenter using a small 
percentage of mortar in the concrete, so that the voids in the 
broken stone are not nearly filled, may conclude that gravel is 
the better, while another experimenter using a larger amount 
of mortar, filling the voids in the broken stone but giving a 
large excess of mortar for the gravel, will conclude that broken 
stone is much to be preferred. 

ART. 39. STONE CRUSHING AND COST OF AGGREGATE 

304. Breaking Stone by Hand. When but a small quantity 
of concrete is to be made, and broken stone cannot be pur- 
chased in the vicinity, the stone for concrete may be broken by 
hand. This is an extremely tedious process, however, and is 
generally avoided, since broken stone prepared in this way will 
cost from two dollars and a half to four dollars per cubic yard. 
In the reconstruction of the breakwater at Buffalo, the cost of 
breaking stone by hand was two dollars and eighty-six cents 
per cubic yard, and loading on boat cost thirty-nine cents, 
making total cost about three dollars and twenty-five cents per 
cubic yard. 1 

305. STONE CRUSHERS. The most common forms of rock 
crushers are the gyratory and the movable jaw types. The 
jaw breaker, or Blake crusher, consists of one fixed plate or 
jaw and one movable one. The latter is hinged at the upper 
end and the lower end is moved backward and forward through 
a short space by means of a toggle joint or other mechanism. 
The jaws are several inches apart at the upper end, depending 
on the size of the machine, and converge toward the bottom. 
The distance between the jaws at the bottom regulates the size 
of fragments delivered, and this distance may be adjusted at 
will. 

The Gates crusher is of the gyratory type and consists of 
a corrugated cone of chilled iron, called the breaking head, 



1 Report of Capt. F. A. Mahan in Report Chief of Engineers, U.S.A., 
1888, p. 2034. . 



STONE CRUSHING 209 

within a larger inverted cone, or shell, which is lined with chilled 
iron pieces. The vertical shaft bearing the breaking head is 
pivoted at the upper end while the lower end travels in a small 
circle ; an eccentric motion is then imparted to the head, so that 
it approaches successively each element of the shell. The size of 
opening can be regulated by raising or lowering the breaking head. 

Stone crushers are made of various sizes having capacities 
up to one hundred tons per hour. The cost of running a stone 
crusher is not great, the principal expense being incurred in 
breaking the stone into pieces of proper size to feed the crusher, 
the delivery of the stone to the crusher, and taking it away 
when broken. 

Crushing plants are usually provided with revolving screens 
into which the broken stone is delivered from the crusher. 
These screens are usually made of perforated steel plate, 
the holes being such as to separate the material into the sizes 
desired. 

Where large amounts of concrete are required, and the stone 
is to be crushed on the work, the arrangement of the crusher 
plant should receive careful study to facilitate the transporta- 
tion of the rock to and from the crusher. The broken stone 
should be discharged from the crusher into bins, from which 
the carts or cars may be filled by gravity, or from which the 
material may be led directly to the mixer through a chute or 
other form of conveyor. In quarries preparing aggregate for 
sale, and on important works, very complete stone crushing 
plants are erected. 1 

306. COST OF AGGREGATE. The cost of aggregate varies 
greatly according to the proximity of the stone to the crusher, 
the character of the stone, and the amount required. In ex- 
ceptional cases gravel suitable for use in concrete is so near 
at hand that it may be delivered on the mixing platform for 
from twenty-five to forty cents per cubic yard. When it must 
be brought from a distance, the cost is correspondingly in- 
creased. Where a considerable quantity of stone is to be broken, 
the cost of crushing, aside fro'm transportation of the materials 



1 The stone crushing and sand and gravel washing plant used in the con- 
struction of the Canal at the Cascades of the Columbia, Ore., is described 
and illustrated in Report of Chief of Engineers, 1891, p. 3332. 



210 CEMENT AND CONCRETE 

to and from the site of the work, would usually be from thirty 
to forty cents per cubic yard. 

In one case where the stone was delivered to the crusher in 
carts after having been sorted from spoil banks containing 
much poor stone that had to be handled over, the cost per 
cubic yard of crushed stone was approximately as follows for 
about six thousand cubic yards crushed in one season: 

Labor, including sorting and delivering to crusher, per cubic 

yard of crushed stone $.67 

Rent of power plant 04 

Fuel 05 

Tools, supplies, breakages, etc 12 

Interest and depreciation of plant 12 

Total cost per cubic yard $1.00 

307. The following data concerning the cost of breaking a 
large amount of stone for road material are given by Messrs. 
Spielman and Brush. 1 "The stone was broken by a ten-inch 
Blake stone crusher at the rate of about twenty cubic yards in 
ten hours. The size of the stones as they came from the crusher 
was: fifty per cent., two inches size; twenty-five per cent., one 
and one-half to one inch size; twenty-five per cent., screenings 
and pea dust. The cost of the crusher, engine, boiler, etc., 
set up complete, was about twenty-five hundred dollars. The 
cost of working per day independent of the original cost of the 
machinery and interest thereon, and also independent of any 
royalty on the stone, was found by the contractor to be as 
follows: 

Repairs, lubricants, wear and tear on crusher and engine, about $6.00 
1 engineman, $2.50; 1 feeder, $1.50; 1 screener, $1.50; 5 laborers 

quarrying and breaking up stone at $1.00 10.50 

1 team hauling stone 5.00 

\ ton coal 2.50 



Cost of preparing and crushing 20 ou. yds. of stone, $24.00 
Cost of one cubic yard, $1.20. 

308. The cost of breaking trap on the Palisades is given as 
follows: 2 "Two crushers deliver thirty-five cubic yards of two- 



1 Trans. Am. Soc. C. E., April, 1879. 

2 "Construction and Maintenance of Roads," by Mr. Edward P. North, 
M. Am. Soc. C. E., Trans. A. S. C. E., April 16, 1879. 



COST OF CRUSHING STONE 211 

inch stone per day. when working well, the stone being sledged 
to go into the jaws readily; fifteen per cent, of the time is lost 
by breakdowns: 

1 engineman and fireman $2.50 

2 laborers feeding, at $1.25 2.50 

2 laborers screening, at $1.25 2.50 

Coal, 1 ton 3.50 

Oil and waste 1.00 

Breakages 5.00 

$17.00 
or about fifty-seven cents per cubic yard. 

"On Snake Island, three crushers were arranged in a row, 
and the broken stone was carried by an endless belt to the 
revolving screen, whence it fell into the bins, so that no screen- 
ers were employed. The engine had one cylinder ; eight inches 
by twenty-four inches, and was running with eighty pounds of 
steam. The product was said to be one hundred eighty cubic 
yards per day when there was no breakdown." The cost was 
as follows : l 

1 engineman and fireman $2.50 

3 laborers feeding, at $1.25 3.75 

2i tons coal, at $3.50 8.75 

Oil, etc 2.00 

Breakages 15.00 

$32.00 

"Allowing for the fifteen per cent, lost by breakdowns, the 
cost would be about twenty-one cents per cubic j^ard." 

At another place on the Hudson, two crushers, set face to 
face, nine-inch by fifteen-inch jaws, could deliver at the rate of 
one hundred twenty cubic yards per day when no trouble 
occurred, but one hundred cubic yards was a fair average. 

COST. 

1 engineman and fireman $2.50 

3 feeders 3.75 

2 screeners 2.50 

li tons coal, at $4.00 6.00 

OD, etc 2.50 

Repairs 10.00 

$27.00 
or twenty-seven cents per cubic yard." 



1 "Construction and Maintenance of Roads," by Mr. Edward P. North, 
M. Am. Soc. C. E. Trans. A. S. C. E., April 16, 1879. 



212 



CEMENT AND CONCRETE 



It is noticeable that in all the above cases the item for 
repairs is very large. The wages paid are lower than at 
present, 

309. The following data concerning the cost of quarrying 
and crushing about five thousand six hundred yards of broken 
stone at Baraboo, Wis., is taken from an article by Mr. W. G. 
Kirchoffer, C. E. 1 

Cost per Cubic Yard of Crushed Rock 



ITEMS. 


1901. 


1902. 




$ .040 


$ .027 


Dynamite, at 24 to 27 cents pound . . 
Tools, repairs, depreciation, supplies and 
improvements 


.056 
.200 


.110 
.218 


Labor, quarrying and tending crusher . . 
Fuel, at $4.60 per ton, and oil .... 
Rent of engine 


.714 

.078 
.085 


.544 
.053 
.066 


Superintendence, including livery . . . 
Hauling stone to city 


.086 
.500 


.165 
.500 








Total cost per cubic yard . . . 


$1.76 


$1.68 



The cost of common labor was fifteen cents an hour, quarry- 
men and drill runners, seventeen and one-half to twenty cents, 
engineers and engine, thirty-five cents, and team and driver, 
thirty cents. 

310. The cost of crushing cobble stone with a rented plant 
at Port Huron, Mich., is given by Mr. Frank F. Rogers, C. E. ; 
from which the following data have been derived. 2 





JULY AND 
AUGUST. 


OCTOBER 

AND 

NOVEMBER. 


Hours run 


171 5 


94 


Stone crushed cubic yards 


1145 


522 


Average cubic yards crushed per hour . . 
Average rental cost per cubic yard . . . 
Average fuel cost per cubic yard .... 
Average labor cost per cubic yard . . . 
Average total cost of crushing per cu. yd. 


6.67 
11. 6 cents 
3.7 " 

22.2 " 
37.5 " 


5.55 
16.1 cents 
7.1 " 
27.9 " 
51.1 " 



1 Engineering News, Jan. 15, 1903. 

2 Michigan Engineers' Annual, 1902, abstracted in Engineering News, 



March 6, 1902. 



COST OF CRUSHING STONE 213 

In the construction of the defenses at Portland, Me., 1 a No. 
5 Champion Crusher was used, driven by a thirty horse-power 
portable engine. Granite was purchased at one dollar per ton 
on the wharf. Hauling to crusher cost thirteen cents per ton. 
Cost of crushing, twenty cents per cubic yard of crushed stone, 
making total cost of crushed stone in bin at crusher one dollar 
eighty-three cents per cu. yd. 



1 Report of Charles P. Williams; Officer in charge, Maj. Solomon W. 
Roessler, Corps of Engineers,U.S. A. ; Report Chief of Engineers, 1900 ; p. 757. 



CHAPTER XIV 

CONCRETE MAKING: METHODS AND COST 

ART. 40. PROPORTIONS OF THE INGREDIENTS * 

311. Concrete is simply a class of masonry in which the 
stones are small and of irregular shape. The strength of the 
concrete largely depends upon the strength of the mortar; in 
fact, this dependence will be much closer than in the case of 
other classes of masonry, since it may be stated as a general 
rule, that the larger and more perfectly cut are the stone, the 
less will the strength of the masonry depend upon the strength 
of the mortar. 

In deciding, then, upon the proportions of ingredients to use 
in a given case, the quality of the mortar should first be con- 
sidered. If the concrete is to be subject to but a moderate 
compressive stress, the mortar may be comparatively poor in 
cement; but if great strength is required, the mortar must be 
of sufficient richness; while if imperviousness is desired, the 
mortar must also possess this quality and be sufficient to thor- 
oughly fill the voids in the stone. 

312. THEORY OF PROPORTIONS. The usual method of stat- 
ing proportions in concrete is to give the number of parts of 
sand and aggregate to one of cement. These parts usually 
refer to volumes of sand and stone, measured loose, to one vol- 
ume of packed cement. However, there is no established prac- 
tice in regard to this and a " 1-2-5 concrete" may mean five 
volumes of loose stone to two volumes loose sand to one volume 
loose cement, or any one of several combinations. 

This method of stating proportions leads to confusion unless 
one is careful to explain what is meant by such an expres- 
sion as " 1-3-6 concrete." The evils of similar methods of 
stating proportions in mortars, and the desirability of fixing 
upon some standard system of weight or volume, have already 



1 Portions of this article were contributed to Municipal Engineering by 
the author, and appeared in that magazine, May, 1899. 

214 



PROPORTIONS OF THE INGREDIENTS 215 

been pointed out. The only circumstances under which such 
expressions as the above may be used with propriety are when 
one wishes to give only an approximate idea of the character 
of concrete used. 

From tests of strength it is known that to obtain the strong- 
est concrete with a given quality of mortar the quantity of the 
latter should be just sufficient to fill the voids in the aggregate. 
The strength is notably diminished if the mortar is deficient, 
and is also impaired by a large excess of mortar. This last 
statement is subject to one exception: if the mortar is stronger 
than the stone, then an excess of mortar does not weaken the 
concrete. This case, however, should never be allowed to occur, 
since it is evident that the strength of the stone should be at 
least equal to the required strength of the concrete. Further, 
the ordinary uses of concrete are generally best served by a 
compact mixture containing as few voids as possible. 

For these reasons, then, one should consider concrete not as 
a mixture of cement, sand and stone, but rather as a volume 
of aggregate bound together by a mortar of the proper strength. 
The volume of voids in the aggregate, the per cent, of this 
volume filled with mortar, and the strength of this mortar be- 
come then the important considerations in proportioning con- 
crete. When thus considered, it is an easy matter to determine 
the required volume of mortar for a given volume of stone, 
and the amount of cement and sand required for a given volume 
of mortar has already been considered. 

313. Determination of Amount of Mortar to Use. The bulk 
of a given quantity of broken stone is not so variable as the 
volume of sand. The volume of the stone, and consequently 
the voids, will vary with the degree of packing, but the packing 
is not influenced appreciably by the amount of moisture present. 

The proportion of voids in the broken stone may be obtained 
as follows: Find the weight per cubic foot of the broken stone 
in the condition in which the volume of voids is sought, being 
careful to use a measure holding not less than two or three cubic 
feet. Also obtain the specific gravity, and hence the weight 
per cubic foot of the solid stone. Then one, less the quotient 
obtained by dividing the weight per cubic foot of the broken 
stone by the weight per cubic foot of the solid stone, will be 
the proportion of voids in the aggregate. , 



216 CEMENT AND CONCRETE 

For example, suppose the weight per cubic foot of the broken 
stone is 102 pounds. The specific gravity of the solid stone 
determined in the ordinary manner is found to be 2.724. Then 
weight per cubic foot of solid stone is 62.4 X 2.724 =170 

102 
pounds and 1 r\ = -40, voids in stone. 



Another method is to fill a vessel of known capacity with 
the stone to be used, and to pour in a measured quantity of 
water until the vessel is entirely filled. The volume of water 
required indicates the necessary amount of mortar to use. The 
stone should be moistened before placing in the vessel, to approxi- 
mate more nearly its condition when used for concrete, and to 
avoid an error from absorption of the water used to measure voids. 

314. As to the degree of jarring or packing to which the 
stone should be subjected in filling the measure, if the stone 
is filled in loose, and it is proposed to ram the concrete in place, 
the amount of mortar indicated will be a little more than the 
required quantity. If the concrete is to be placed without 
ramming (as in submarine construction), the amount of mortar 
indicated will not be too great. On the other hand, if the stone 
is shaken down in the vessel to refusal, the voids obtained will 
be less than the amount of mortar which should be used, be- 
cause it is not possible to obtain a perfect distribution of mor- 
tar in a mass of concrete, and because the concrete will usually 
occupy a greater space than did the stone when shaken down. 
And again, for perfect concrete, pieces of stone should be sepa- 
rated one from another by a thin film of mortar, and hence the 
volume of the concrete will be greater than the volume of the 
stone measured in a compact condition without mortar. A 
deficiency of mortar is usually more detrimental than an excess. 
It is safer, therefore, to measure the voids in the stone loose, 
or when but slightly packed, and make the amount of mortar 
equal to, or a trifle in excess of, the voids so obtained. 

315. Aggregates Containing Sand. If in the case of broken 
stone all of the fine particles are used, or if gravel which con- 
tains a considerable amount of sand is employed, then this 
fine material or sand must be considered as forming a part of 
the mortar. This will not change the method of obtaining the 
amount of mortar required for such broken stone or gravel, 
but it will change the composition of the mortar used. Thiis, 



MIXING BY HAND 217 

suppose we have a gravel ten per cent, of which is sand (grains 
smaller than one-tenth inch in diameter) and we find the voids 
to be thirty-three and one-third per cent. To three cubic yards 
of this gravel we will add one cubic yard of a one-to-three 
mortar. The voids will be filled, but instead of having three 
cubic yards of stone imbedded in one cubic yard of a one-to- 
three mortar, we will in reality have a little less than that 
amount of stone imbedded in a mortar composed of one part 
of cement to about three and three-tenths parts sand. 

316. Required Strength. In the paragraphs just preced- 
ing, an attempt has been made to indicate the general principles 
to be applied in proportioning the materials in concrete. To 
decide on the actual proportions of the ingredients to use for a 
given purpose, one must have clearly in mind the strength that 
will be demanded and any special condition to which the con- 
crete is to be subjected. A reference to Art. 57 concerning the 
strength of concrete, will be of service in deciding on the proper 
proportions to use in a given case. 

ART. 41. MIXING CONCRETE BY HAND 

317. Necessity of Thorough Mixing. Too much stress can 
hardly be laid upon the necessity of thoroughly mixing the 
concrete if the best results are to be attained. It has already 
been shown that thoroughness in mixing mortar is repaid by 
greatly increased strength, and the result is even more marked 
in the case of concrete. Every grain of sand should be coated 
with cement, and every piece of stone should be covered with 
mortar. In general, the cost of mixing is from one-tenth to 
one-fifth of the total cost of the concrete in place. If by doub- 
ling the cost of mixing we can increase its strength more than 
one-tenth or one-fifth in these respective cases, or permit a 
corresponding decrease in the amount of cement necessary for 
a given result, the additional labor in mixing is justified. 

318. Concrete may be mixed by hand or by machine. Opin- 
ions vary as to the comparative merits of the two systems, but 
as a machine properly installed usually furnishes much the 
cheaper method of mixing, it is usually employed. The saving 
by this method, however, will evidently depend upon the cost 
of labor, the total amount of work to be done, and the degree 
of concentration of the work, or facilities for distributing the 



218 CEMENT AND CONCRETE 

concrete. In certain sections where cheap labor is abundant, 
the cost of hand mixing may be as low as machine mixing. 

With proper supervision, hand mixing may be thorough, and 
the chief argument against it, aside from its cost, is that such 
hard work is likely to be slighted. The best forms of mixers 
now on the market, however, give results quite equal to the best 
hand work. 

319. METHOD OF HAND MIXING. We will assume that the 
materials have been brought within easy reach of the mixing 
place. If the concrete is to be mixed near the point where it 
is to be deposited, the mixing platform must be made portable. 
Three platforms, each 8 by 14 feet, built of two-inch plank or 
of two layers of one-inch boards, nailed to four 2x6 inch longi- 
tudinal scantlings laid flat, will be suitable for such a case. 
The platforms should be made without vertical sides, though if 
desired a narrow piece of one-inch board may be laid flat around 
the edges and nailed. A short piece of rope attached to each 
corner of the platforms, or to the ends of the longitudinal scant- 
lings, will be found convenient in moving them. These mixing 
boards are placed side by side. 

The sand, which may be delivered to the mixing platform 
in wheelbarrows, is first dumped on the board and spread 
evenly over the surface. If the sand is measured, the barrows 
should be so arranged as to hold the required amount after 
" striking " with a straight edge. This will make the measure- 
ment independent of the judgment of the shoveler. If the sand 
is delivered in cars, bottomless boxes of two or three barrels 
capacity, according to the proportions used, will be found 
more convenient for measuring than barrels. If the sand is 
determined by weight, which as has been shown is the more 
accurate method, the scales should be set at a weight which is 
a factor of the total weight, and but little time will be required 
to bring the scales to a balance for each barrow. 

If it is possible, the batch should be of such a size as to 
take either one or two full barrels, or a certain number of full 
sacks of cement. This will obviate the necessity of measuring 
or weighing the cement. The sand having been spread over the 
surface of the mixing board, the cement is dumped upon it and 
spread evenly over the sand. These ingredients are then mixed 
dry, the required amount of water is added at one time in the 



MIXING BY HAND 219 

center of a ring formed of the dry materials, and the whole is 
thoroughly mixed as described under the head of mortar-making. 

320. The mortar having been spread evenly over the board, 
the broken stone is dumped upon it and evenly distributed 
over the surface. Four shovelers then mix the concrete. Each 
shoveler starts at a corner of the board and turns each shovel- 
ful completely over, casting toward the end and spreading the 
mortar a little as he draws the shovel toward him. The two 
shovelers at each end work toward each other, and meeting at 
the axis of the platform, return to the side and repeat. When 
the four shovelers meet at the center of the board, they turn the 
mass again by casting toward the center in a similar manner. 
If in mixing the concrete it is found that sufficient water has 
not been used, more may be added from a rose nozzle, or sprink- 
ling pot, previous to the last turning of the mass. The shovel 
should always be used at right angles to either the side or the 
end of the board, never diagonally; and it should always scrape 
the mass clean from the board, never cut it at mid-depth. From 
three to five turnings are required to thoroughly mix the concrete. 

The mode of mixing has been thus minutely described, be- 
cause if a gang of men are started properly they will soon be- 
come expert, working in unison; whereas if each man is allowed 
to mix according to his notion, confusion is sure to result. It 
is sometimes preferred to spread the stone on a separate board 
and cast the mortar upon it, but this necessitates one handling 
of the mortar which does not appear to contribute much to the 
incorporation of the ingredients. 

While the shovelers are engaged in mixing the concrete on 
one platform, the mortar mixers have proceeded to the next 
platform to mix another batch of mortar, and the cement and 
sand are being placed upon the third platform. Thus the work 
proceeds in regular progression without delays. The shoveling 
of concrete is hard work, and it will be found necessary not 
only to pick good men for this duty, but to cull them until the 
evolution results in the proper men for the work. An extra 
compensation for men who perform satisfactory service in the 
mixing of concrete will usually be repaid in the character and 
quantity of the output. 

321. With the method described above, a working gang 
would consist of the following men under ordinary conditions: 



220 CEMENT AND CONCRETE 

Measuring and supplying cement and sand 1 

Mixing mortar * 2 

Delivering stone from bin, one man with horse and cart, or two 

men with barrows 2 

Shovellers to mix concrete and cast or wheel to place .... 4 

Water boy 1 

Spreading and tamping concrete 1 

Total men required 11 

If it is found impracticable to mix the concrete near the 
place of deposition, it may be necessary to put on two or more 
extra men to wheel the concrete to place. This, gang of eleven 
men may be doubled and still work on the same three platforms 
when so desired. 

With a moderate length of wheel for the materials and the 
finished concrete ; a gang of eleven picked men, working ac- 
cording to system, will be able to make from twenty-five to 
thirty cubic yards per day of ten hours, or about two and a 
half yards per man. The double gang of twenty-two men may 
not work to quite as good advantage, and will probably not 
put in more than from forty to fifty cubic yards per day. It 
would therefore be somewhat more economical to work two 
gangs of eleven men each on separate sets of platforms, espe- 
cially as in this way a rivalry is created. Lack of room, however, 
will frequently preclude this arrangement. 

322. COST OF MIXING BY HAND. The amount of concrete 
stated above, two and a half yards per man, may be taken as 
a maximum. With wages at $1.75 per day this would corre- 
spond to a cost of about seventy cents per yard, exclusive of 
the wages of a foreman. Numerous examples might be cited 
where the mixing costs more. Colonel Mendell, in writing of 
the fortifications at Fort Point, California, 1 states that a fore- 
man (at $4 per day) and twenty laborers (at $2 per day) made 
forty-five cubic yards per day of eight hours, the cost of mixing 
being thus about $1 per cubic yard. It is stated that "the 
circumstances were exceptionally favorable." 

As an instance where hand mixing was done at a very low 
cost, the Lonesome Valley Viaduct 2 may be mentioned. At 



1 Jour. Assn. of Engr. Soc., March, 1895. 

2 Construction of Substructure for Lonesome Valley Viaduct, Gustave R. 
Tuska, Trans. A. S. C. E., Vol. xxxiv, p. 247. 



CONCRETE MIXERS 221 

this point colored labor was used at a cost of $1 for eleven hours' 
work. A gang of men, distributed as follows, would mix and 
lay forty cubic yards of concrete per day: 

Filling sand barrows and handling water 1 

Filling rock barrows 2 

Mixing sand and cement 4 

Mixing stone and mortar 4 

Wheeling concrete 2 

Spreading concrete in the molds 1 

Tamping concrete in the molds . . : 1 

Foremen _1 

Total T6 

Fifteen men at $1 per day, and foreman at $2.50 per day, 
makes a cost of $17.50 for forty yards of concrete, or at the 
rate of forty-four cents a yard for mixing. Had the laborers 
received $1.75 per day, however, the cost would have been 72 
cents per yard. 

323. In the construction of the Forbes Hill Reservoir and 
standpipe" at Quincy, Mass., 1 all concrete was mixed and placed 
by hand. "The ordinary concrete gang was made up of a 
sub-foreman, two men gaging materials, two men mixing mor- 
tar, three men turning the concrete, three men wheeling con- 
crete, one man placing, and two men ramming. Two gangs 
were ordinarily employed, placing about twenty cubic yards 
per day each, or about 1.43 cubic yards per man. The con- 
crete was turned at least three times before placing." With 
labor at $1.75 per day, this would give the cost of mixing and 
placing $1.22 per cubic yard. The actual cost of mixing and 
placing varied from $0.97 to $1.53, according to the character 
of the work. 

ART. 42. CONCRETE MIXING MACHINES 

324. General Classification. Concrete mixing machines may 
be divided into two general classes, batch mixers and continu- 
ous mixers. In the former, sufficient materials are proportioned 
(10 make a convenient sized batch for the mixer. They are then 
charged into the machine at once, given a certain amount of 
mixing, and then discharged at once. In the continuous mixers 

1 Described by C. M. Saville, M. Am. Soc. C. E., Engineering News, Mar. 
13, 1902. 



222 CEMENT AND CONCRETE 

the materials are dumped on a platform, and after being prop- 
erly proportioned, are delivered gradually to the mixer, and if 
fed uniformly, the concrete is discharged continuously by the 
machine. In the latter method care must be taken to feed 
the cement, sand and stone together and at a uniform rate. 
If one man shovels cement, two men shovel sand and four men 
handle the stone, and the cement man stops to fill his pipe, 
there is likely to be a poor streak of concrete. It is therefore 
desirable in feeding a continuous mixer to spread the measured 
quantity of stone on the platform, and on top of this place the 
weighed quantities of sand and cement. Then if each shoveler 
gets his shovel blade under the whole mass, he will have some 
of each ingredient. 

325. There are many styles of concrete mixers of both classes 
on the market. One of the oldest, as well as one of the best, 
is the cubical box mixer which consists of a box four or five 
feet on a side, supported by trunnions at opposite corners, and 
made to revolve about this axis. A hinged door is provided 
near one corner of the box by which the latter is charged and 
emptied. The dry materials may be first charged and mixed 
and the water added later, either through the door or through 
a perforated pipe in the axis, or the water may be added with 
the dry materials; after from ten to thirty revolutions of the 
box, the mixed concrete is discharged into a skip or on a car, 
to be conveyed to the place of deposition. 

The great merit of this mixer is that the materials are thrown 
back and forth from one side of the cube to another and a 
thorough commingling results. The chief disadvantage is the 
difference in elevation between the receiving hopper and point 
of delivery, making it necessary to elevate the materials; one 
other defect is that the batch is not in view while being mixed, 
so that the amount of water cannot be regulated according to 
slight variations that may occur in the moisture of the sand and 
stone when -charged. 

326. To obviate this latter difficulty as well as to facilitate 
to some extent the charging and dumping of the batch, a form 
of box mixer is made in which the corners of the box in the axis 
of revolution are truncated, and the trunnions are replaced by 
collars which support the box, and through which the materials 
may be fed and discharged. The collars are supported in a 



CONCRETE MIXERS 223 

tilting cradle which permits the delivery end to be depressed 
after the batch is mixed. This style of mixer with truncated 
corners and tilting cradle is known as the Chicago Improved 
Cube. 1 

Mixers working on the same principle are sometimes made 
in other forms than the cube. One of these is the cylindrical 
mixer, which is made of boiler plate and may be four or five feet 
in diameter and five or six feet long. This is rotated about a 
diagonal axis. It is said to be more easily and cheaply made 
than the cubical mixer, and dumps more quickly and cleanly, 
while the cost of operation is about the same, and the mixing 
is as satisfactorily done as in the cubical form. 

327. The so-called " Dromedary Mixer " 2 is a batch mixer 
specially designed for use on street work. The mixing chamber 
is a cylindrical steel drum with closed ends, mounted between 
two wheels. It is hinged along an element of the cylinder so 
that it opens into two halves like a clam shell bucket, to dis- 
charge. A trap door is provided for filling. The cart is drawn 
by a horse, 'and the chamber may be thrown in or out of gear 
with the cart wheels. The cement and sand being first added 
and the trap door closed, the horse draws the cart to the stone 
pile. The stone and water are here added and the cart is drawn 
to the work; the concrete, mixed on the way, is dumped by the 
driver, who merely raises a lever which not only separates the 
two halves of the mixer, but throws it out of gear so that it 
stops revolving. The chamber may be thrown out of gear at 
any time without dumping if desired. 

328. The Ransome Concrete Mixer 3 " consists of a hollow 
rotary dome, having upon the inner surface of its periphery 
directing guides or flanges, and hinged shelves, by means of 
which the materials are thrown together and perfectly com- 
mingled. A discharge chute, or spout, is arranged to deliver the 
material into the barrow or cart when properly mixed." The 
mixer is also provided with an automatic device for proportion- 
ing the materials, and a conveyor to carry them to the mixer. 
Water is supplied to the mixer through a pipe with facilities 
for regulating the supply. 



1 Municipal Engineering and Contracting Co., Railway Exch., Chicago, 111. 

2 Manufacture now discontinued. 

3 Ransome Concrete Machinery Co., 11 Broadway, N. Y. 



224 CEMENT AND CONCRETE 

329. The Smith Mixer l is a batch machine made of two 
truncated cones placed base to base, and provided on the in- 
terior with deflecting plates designed to throw the materials 
from one end of the mixer to the other as the machine is re- 
volved. At the junction of the two cones, on the outer cir- 
cumference, is a spur gear by which the chamber is actuated. 
The latter rests upon rollers in a swinging frame, so arranged 
that the machine may be tilted for dumping while the drum is 
revolving. In operating this mixer it has been found advanta- 
geous to charge the broken stone or gravel. first, and give one or 
two revolutions before adding the cement and sand, as this cleans 
the mortar from the corners. This form seems to be particularly 
adapted for a portable machine. They may be had mounted 
on trucks, with or without an engine, as desired. 

330. The McKelvey Mixers 2 are made in two styles, con- 
tinuous and batch. Both styles are cylinders revolving on 
friction rollers, and having, on the interior, deflecting blades 
and a patent "gravity shovel" which lies against the rising 
side of the drum and casts the materials downward when the 
cylinder has revolved far enough to overturn the blade. The 
batch mixer has a shorter cylinder and can be discharged at 
will. These mixers may be fed by shovels, or they may be 
provided with a hopper into which the materials may be dumped 
from carts or 1 barrows. They discharge directly into wheel- 
barrows. The mixer, and an engine and boiler to run it, are 
mounted compactly on a truck, or the mixer is furnished on a 
steel frame without an engine. 

331. The pan mixer 3 consists of a large shallow pan into 
which may be lowered a framework carrying a series of plows. 
The materials are spread in the pan in layers, the plows are 
lowered into it, and the pan is revolved about its vertical axis, 
the plows remaining stationary. The plows are so arranged as 
to move the materials radially toward and away from the cen- 
ter of the pan. The water may be added from a rose nozzle. 
For dumping, an opening is made in the bottom of the pan by 
withdrawing a slide. Were the plows made to revolve in a. 



1 Contractors' Supply Co., 232 Fifth Ave., Chicago. 

2 McKelvey Concrete Machinery Co., N. Y. Life Bldg., Chicago. 

3 Clyde Iron Works, Duluth, Minn. 



CONCRETE MIXERS 225 

stationary pan, the concrete would be more conveniently 
dumped in a pile, or in a car, instead of being scattered about 
under the pan. 

332. The Cockburn, 1 a continuous mixer, is in the form of a 
long box square in cross-section, surrounded at either end by 
circular rings supported on friction rollers. By suitable gear- 
ing the mixer is revolved about its longest axis, which has a 
slight inclination toward the discharge end. The materials are 
added through a hopper at one end, and fall from one side of 
the box to the adjacent side as the machine revolves, working 
gradually toward the delivery end, which is open. The water 
is added through a pipe at about one-third of the length of the 
box from the feed end. While this machine has no complicated 
system of blades to become clogged, the mortar has a tendency 
to stick in the corners of the mixer, making the interior cylin- 
drical, and thus much less effective in mixing. Striking the 
sides of the box with a heavy hammer will detach the mortar, 
and this requires occasional attention. 

333. A common form of continuous mixer consists of a screw 
working in a cylinder. The materials are fed to the cylinder 
near one end and are mixed while being gradually worked toward 
the other end by the screw. The water is added through a 
fixed perforated pipe at a point about one-third of the distance 
from the feed end of the cylinder, and the mixed concrete falls 
from the outlet at the other end. This style is frequently 
made in a light form and mounted on wheels, and is then con- 
venient in the laying of concrete for pavements. 

A modification of the screw mixer consists of a semi-cylin- 
drical trough, in which revolves a shaft carrying blades set at 
right angles to the shaft and to each other. The trough is 
sometimes given a slight inclination to the horizontal, and the 
blades are so shaped as to assist in working the materials toward 
the delivery end. 

334. The Drake Mixer 2 is of the general form just described. 
One of the machines made by this company is a semi-cylindrical 
trough in which revolve in opposite directions two shafts, each 
carrying some thirty blades. Most of the blades are straight, 



1 Cockburn Barrow and Machine Co., Jersey City, N. J. 

2 Drake Standard Machine Works, 298-302 W. Jackson Boul., Chicago. 



226 CEMENT AND CONCRETE 

but some of them are curved to work the material toward the 
delivery end. 

335. Gravity Mixer. An appliance recently devised, which 
is called a concrete mixer, consists of a steel trough provided 
with staggered pins and deflecting plates. The trough is sup- 
ported in an inclined position and has a hopper at its upper 
end. Water is supplied through spray pipes at the side of the 
trough. The materials, stone, sand and cement, are spread in 
layers on the mixing platform, with the stone at the bottom. 
The materials are then thrown into the hopper; they are mixed 
as they descend through the pins, and the product is caught in 
barrows or carts at the bottom. 

336. In a very able article on concrete mixers, 1 Mr. Clarence 
Coleman, M. Am. Soc. C. E., makes an analytical discussion of 
the relative efficiencies of the several forms. In this analysis 
he gives the following weights to the several requirements for 
a perfect mixer. That the entire mass of concrete shall be so 
commingled that the cement shall be uniformly distributed 
throughout the batch is given a weight of forty; that the amount 
of water shall be subject to control is given a weight of twenty- 
five; perfect dry mixing and relative time of mixing, each ten; 
and receiving materials, discharging concrete and self-cleaning, 
are each given a weight of five. 

The first three requirements, with a combined weight of 
seventy-five, relate to the production of good concrete, while 
the remaining requirements, with a combined weight of twenty- 
five, pertain to economy in use. In short, the first requisite 
is that a machine shall be capable of producing a perfect mix- 
ture; then the machine that accomplishes this result at the 
lowest cost per cubic yard is the best. The choice of a machine 
will depend frequently on the character of the work to be done, 
as some machines can only be used economically where large 
quantities of concrete are to be used in a restricted area, while 
others are particularly adapted for portable plants. 

ART. 43. CONCRETE MIXING PLANTS AND COST OF MACHINE 

MIXING 

337. Coosa River Improvement. The concrete plant used 
at Lock No. 31, Coosa River Improvement, 2 was erected in a 



1 Engineering News, Aug. 27, 1903. 

3 Major F. A. Mahan, Corps of Engineers, U. S. A., in charge. 



CONCRETE PLANTS 227 

three-story shed. The top story served as a cement storage 
room and two hoppers were arranged in the floor to receive 
the cement for the mixers below. Level with the floor of the 
second story were two other hoppers immediately below the 
cement hoppers, to receive the sand and broken stone, while in 
the first story or basement the mixers were suspended ' at a 
height sufficient to allow concrete cars to pass under them. 
The following description is from the report of the designer, 
Mr. Charles Firth, U. S. Asst. Engineer: 1 

"The cars used in handling the sand and broken stone are 
of the side dump pattern and are brought into the charging 
room on either side of the hoppers. The cement is drawn from 
the cement room overhead in proper quantities, through verti- 
cal chutes arranged somewhat on the principle of the old- 
fashioned powder flask. 

"The water is added to the materials as they enter the mix- 
ers, and the quantity, which will probably be variable with 
the temperature, is controlled by valves on the mixing floor, 
the operators being governed by indicators, which show the 
quantity used. The mixers are cubical boxes four feet on each 
side, inside measurement, made of steel plate five-sixteenths of 
an inch thick, with 2 by 2%-in.ch angle irons. Each mixer is 
provided with a door in one corner, twenty- two inches square, 
fastened with a tempered steel spring catch, and held open 
when required with a hinged screw bolt. The shaft which 
revolves the mixers is three inches square. It is securely 
fastened to them by trunnion castings at diagonally opposite 
corners. The whole is driven by a 10 by 16 inch horizontal 
engine, and thrown in and out of gear by ordinary friction 
gearing with friction and brake levers. 

"After a sufficient number of revolutions in the mixers, the 
concrete is dumped into the concrete cars below, which are of 
the center dump pattern." 

The method given of measuring the cement is not recom- 
mended, as the charge of cement, if not a full barrel, should 
always be weighed. The three-story arrangement by which 
the materials were handled almost entirely by gravity was 
made possible by the high bank at the side of the lock pit. 



1 Annual Report, Chief of Engineers, U.S.A., 1894, p. 1292. 



228 CEMENT AND CONCRETE 

The total cost of the plant, exclusive of the boilers, is stated 
to have been about $8,000, and the average output about two 
hundred cubic yards of concrete per day of eight hours. The 
cost of mixing, depositing and ramming 8,710 cubic yards of 
concrete in the construction of lock walls was at the rate of 
$0.884 per cubic yard. 

338. Portland, Maine, Defenses. In the construction of the 
defenses at Portland, Maine, 1 a five-foot cubical mixer was 
used. Sand and stone were delivered, by bucket conveyors, 
in bins directly over the mixer. " Immediately under these 
bins were two measuring hoppers for stone and sand, respec- 
tively, and an additional hopper for cement. From these meas- 
uring hoppers the charge was dumped into the mixer and 
thence, when mixed, into a car immediately under it. This car 
delivered the mixed batch by means of a hoisting engine and 
an inclined track to the site of the battery under a fifty-five 
foot derrick, which placed it in the work at the point required. 
Two barrels of cement, sixteen cubic feet of sand, and thirty- 
two cubic feet of stone constituted a batch. * * * The usual 
number of men engaged in the operation of mixing and placing 
was as follows: Two master laborers, three steam engineers, 
two stokers and twenty-five laborers." It is said that 200 
barrels of cement, or 100 batches, could be mixed and placed in 
a day of eight hours. This would make the labor cost of this 
portion of the work 50 or 60 cents per cubic yard. The cost 
stated, however, varies greatly according to the amount of detail 
in construction, and the lowest cost given for " labor of mixing 
and placing" is $1.15 per cubic yard. 

339. San Francisco Defenses. A cubical mixer used in the 
construction of the defenses at San Francisco 2 mixed 250 cubic 
yards per day with seven men, engineer, fireman, and five men 
to feed and dump mixer, at a labor cost of $14.67 per day, or 
about six cents per cubic yard, exclusive of cost of transporta- 
tion and ramming. The materials and concrete were handled 
on cars run almost entirely by gravity. 

340. Buffalo Breakwater. In the construction of the Buf- 



1 Report of Charles P. Williams to Maj. Solomon W. Roessler, Corps of 
Engrs., U. S. A., in charge. Report Chief of Engineers, 1900, p. 745. 

2 Maj. Charles E. L. B. Davis, Corps of Engineers, U. S. A., Report Chief of 
Engrs., 1900, p. 980. 



CONCRETE PLANTS 



229 



falo Breakwater, 1 the mixing plant, consisting of a cubical 
mixer with necessary engines and boilers and two derricks, was 
mounted on a dismantled lake schooner which could be placed 
beside the section of the breakwater under construction. The 
broken stone was delivered in a canal boat which could be tied 
up alongside the schooner, and outside of the canal boat lay the 
material scow. The latter was made from an old dump scow, 
the decked pockets serving as bins for cement, sand and gravel. 
Into a steel bucket on the scow were loaded, by wheelbarrows, 
the following materials: 

5.4 cu. ft. (H bbls.) cement. 
10.8 cu. ft. sand. 

5.4 cu. ft. gravel. 
21.6 cu. ft. total. 

Into a similar bucket on the canal boat 21.6 cubic feet of 
broken stone were shoveled. As these buckets were filled, they 
were hoisted by one of the derricks and dumped into the cubical 
mixer. The latter discharged the mixed concrete into a skip 
and a derrick deposited the concrete in place. The cost of labor 
per cubic yard of concrete is as follows: 



ITEMS. 


No. 
MEN. 


COST 

PER 

HOUR. 


Cu. YDS. 

PER 

HOUR. 


COST OF 
LABOR 

PER 

Cu. YD. 


Loading material into buckets from scows 
Mixing, including engine men and derrick 


18 
11 


$3.17$ 
2.35 


18.2 
18.2 


$0.174 
0.129 


Placing, including foreman . ... 


13 


2.65 


182 


146 












Total labor 


42 


$8 \1\ 


182 


$0449 













The above does not include cost of fuel, nor of transporting 
materials from the storehouses or yards to the site of the work, 

341. Quebec Bridge. The plant used in the construction 
of the Quebec Cantilever Bridge a consists of a No. 5 rotary 
stone crusher, with a maximum capacity of thirty cubic yards 
per hour, discharging into a bucket conveyor which delivered 
the crushed stone in a small storage bin directly over the con- 
crete mixer. The latter was of the cubical form, five feet on a 
side, with a capacity of two cubic yards of concrete per batch. 



1 Emile Low, U. S. Asst. Engr., Engineering News, Oct. 8, 1903. 

2 Engineering News, Jan. 29, 1903. 



230 CEMENT AND CONCRETE 

The cement warehouse and the sand supply were near the 
mixer. Cement and sand were hoisted to the top of the ma- 
chine in boxes, with bottoms inclined at forty-five degrees, 
each holding a batch, and dumped into the charging hopper of 
the mixer as required. The mixer was elevated sufficiently to 
permit dumping the concrete directly in a skip on a car, the 
latter being run to the work. The skip was handled by guy 
derricks. This plant made the remarkable record of two hun- 
dred eighty-five batches in ten hours, and on one occasion 
turned out one hundred fifty batches in five hours, or, if all 
were two-yard batches, at the rate of sixty yards per hour. 

342. Galveston. For the construction of the Galveston sea 
wall two concrete mixing and handling machines were designed, 1 
each consisting of a double-deck car, on eight wheels, with two 
revolving derricks, one on either side for handling materials 
and concrete, respectively. The materials are delivered on 
tracks beside the mixer car track which is parallel to the sea 
wall. One derrick hoists the loaded skips from the material 
cars and deposits them on the upper deck of the mixer car, 
whence they are delivered in measured quantities to the Smith 
Rotary Mixer located on the lower deck. When mixed, the 
concrete is dumped into a skip, which is handled by the second 
derrick and dumped into the forms. 

343. For work having similar requirements to that just 
described, namely, for retaining walls on track elevation, Chicago 
& Western Indiana R. R. at Chicago, the problem was met in 
a somewhat different manner. 2 An ordinary flat car was double 
decked and the space between decks inclosed to protect the 
machinery, including the Drake Concrete Mixer. Cars contain- 
ing cement, sand and stone were coupled in the rear of the mixer 
car. These material cars were fitted with removable wheeling 
platforms, making a complete runway along the sides of the 
cars. The materials were delivered at the mixer car in wheel- 
barrows and dumped into measuring boxes, and thence fed to 
the mixer. The concrete was delivered on a belt conveyor 
mounted on a boom with turntable permitting nearly half of a 
revolution. The outer end of the conveyor could be raised or 
lowered as desired, and the concrete was thus deposited where 



1 Engineering News, Jan. 15, 1903. 
Ibid., Feb. 28, 1901. 



COST OF CONCRETE 



231 



needed in the work. To permit the mixer train to move along 
the track, the two. ends of a cable were made fast to anchorages 
placed about a thousand feet apart, one in front of, and the 
other behind, the train. As this cable had about eight turns 
around a winding drum on the mixer car, the train could be 
propelled forward or backward at will. 

A somewhat similar form has been used for street work, 
where the mixer and electric motor are mounted on a truck 
with a swinging conveyor for the delivery of concrete anywhere 
between the curbs. A pair of wheels in the rear serve to carry 
an inclined runway for wheelbarrows by which the materials 
are delivered to the mixer. 

344. The data for the following items concerning the cost of 
mixing concrete for culverts on railroad work are taken from 
an article in Engineering News. 1 

"The plant is located on a hillside with the crusher bins 
above the loading floor or platform that extends over the top 
of the mixer, so that crushed stone can be drawn directly from 
the chutes of the bins and wheeled to the mixer. The sand is 
hauled up an incline in one-horse carts and dumped on the 
floor, and is also wheeled in barrows to the mixer." The capa- 
city of the cubical mixer used was seven-eighths cubic yard. The 
cost of mixing and placing was as follows: 



ITEMS. 


COST 

PER 

DAY. 


COST 

PER 

Cu.YD. 


One foreman assumed at $2 50 per day 


$2.50 




Three men supplying mixer at -$1.50 per day ..... 
One eugineman assumed at $2.00 per day 
Fuel and supplies assumed at 


4.50 
2.00 
200 










Cost of mixing 40 cu. yds 

Two men loading wheelbarrows at $1.50 
Four men wheeling wheelbarrows at $1.50 


$11.00 

$3.00 
6.00 


$0.275 


Cost of wheeling 40 cu. yds. 100 feet 


$900 


0225 


Four men ramming at $1.50 . .... 


$600 


0150 


Four men wheeling in and bedding large stone in concrete at 
$1.50 


6.00 


150 








Total cost mixing and placing 




$0.800 



1 Location and Construction of the Ohio Residency, Pittsburg, Carnegie 
& Western R.R., Engineering News, May 21, 1903. 



232 CEMENT AND CONCRETE 

It is not explained why six men are required to load and 
wheel forty cubic yards one hundred feet in ten hours, but it 
may be that these men assisted in other operations. 

Another contractor on the same work used a different form 
of mixer with much lower loading platform and handled the 
mixed concrete with skips and derrick. The cost is estimated 
as follows: 

1 man feeding mixer $1.50 

1 engineman assumed at . 2.50 

1 derrick man assumed at 2.50 

2 tagmen swinging boom and dumping 3.00 

6 barrowmen supplying mixer 9.00 

2 men tamping 3.00 

Fuel, supplies, etc 1.50 

Cost of mixing and placing 50 cu. yds. . . $23.00 
Cost per cu. yd., 46 cents. 

ART. 44. COST OF CONCRETE 
345. QUANTITIES OF INGREDIENTS IN A CUBIC YARD. As 

has already been indicated, the rational method of proportion- 
ing concrete is to use just sufficient mortar to fill the voids in 
the stone, or possibly a very small excess to allow for imperfect 
mixing; and in ordinary practice this rule should not be de- 
parted from unless it be for some special reason. When so pro- 
portioned, a cubic yard of concrete will contain approximately 
a cubic yard of stone, depending on the method of measure- 
ment. If we know the percentage of voids in the broken stone 
or gravel, and consequently the percentage of mortar which 
should be found in a cubic yard of the finished concrete, we 
may readily obtain the approximate cost per cubic yard of the 
latter for a given quality of mortar and given unit prices. 

Thus, suppose we have stone in which the voids are such 
that the mortar will amount to forty per cent, of the finished 
concrete, and we wish to have the mortar composed of three 
volumes of loose sand to one volume packed natural cement, 
unit prices being as follows: 

Cement, $1.25 per barrel of 300 pounds net, 3.75 cubic feet. 

Sand, $1.00 per cubic yard. 

Stone, $1.75 per cubic yard. 
., As in 290, we .find the ingredients in one cubic yard of 



COST OF CONCRETE 233 

mortar to cost $3.33. Since forty per cent, of the concrete is 
to be composed of mortar, the mortar in one cubic yard of 
concrete will cost forty per cent, of $3.33, or $1.33, and one 
yard of stone at $1.75 will make the total cost of the materials 
in the concrete $3.08 per cubic yard. 

The diagram herewith may be used to get the approximate 
cost of the concrete after having obtained the cost of the mortar 
as before. Thus, if we enter the diagram with the cost of mor- 
tar $3.33, and follow it to the diagonal line marked forty per 
cent., we find this is on the ordinate $2.33, the cost of the in- 
gredients in one cubic yard of concrete when the stone costs 
one dollar per cubic yard. Hence, $2.33 plus $0.75 equals 
$3.08, the approximate cost of the materials in a cubic yard of 
the concrete as desired. 

346. The usual method, however, of stating proportions in 
concrete is to give the volumes of sand and stone to one volume 
of cement. Thus, one of cement, three of sand and six of stone 
would usually mean one volume of packed cement, three vol- 
umes of loose sand and six volumes of loose broken stone. To 
arrive at the cost of concrete when proportions are thus ar- 
bitrarily stated, involves a greater amount of work. From the 
tables already given (Art. 36), we can determine the amount of 
mortar which a given quantity of dry ingredients will make, 
and the consequent cost of the mortar per cubic yard. Then a 
knowledge of the voids in the broken stone will permit of a 
close estimate of the amount of concrete made, whence we can 
determine the cost of the latter. 

For example, suppose it is desired to determine the cost of 
the materials in a cubic yard of natural cement concrete under 
the following conditions: 

1 bbl. cement containing 280 pounds net, at $1.00 per bbl. 

3 bbls. sand weighing 100 pounds per cu. ft., at $.75 per cu. yd. 

6 bbls. loose broken stone, having 45 per cent, voids, at $1.25 per cu. yd. 

1 bbl. cement = 3.75 cu. ft. = .139 cu. yd., cost $1.000 

3 bbls. sand = 11.25 cu. ft. = .417 cu. yd., cost .313 

6 bbls. stone = 22.50 cu. ft. = .833 cu. yd., cost 1.041 

Total cost $2.354 

From Table 61, 286, we find that it requires 2.03 barrels of 
cement to make one cubic yard of one-to-three mortar, when 



234 



CEMENT AND CONCRETE 



CONCRETE MAKING 
Cost of Concrete, Dollars per Cu. Yd. 
Stone Assumed to Cost $1.00 per Cu. Yd. 










""'' >'> : *!"!!!BHl! 






EXAMPLES OF COST 235 

proportions are stated as above; then one barrel of cement 

would make ^-^ = .493 cu. yd. As forty-five per cent, of the 
^.Uo 

stone is voids, the amount of solid stone in six barrels would 
be 07~~ x -55 = .458 cu. yd. Then the mortar plus solid 

stone would be .493 + .458 = .951 cu. yd. It has been found 
by experiment that the amount of concrete will exceed the sum 
of the mortar and solid stone by from two to five per cent.; 
hence we may assume in this case that the amount of concrete 
made with the above materials would be .95 +- .03 = .98 cu. 
yd., and 2.354 -*- .98 = $2.40, the cost of materials in one cubic 
yard of finished concrete. To obtain the actual cost of con- 
crete in place, the cost of mixing and deposition must be added 
(see Arts. 41 and 43). When the volume of mortar used is not 
greater than the voids in the loose stone, then the amount of 
rammed concrete made may be less than the volume of loose 
broken stone. 

347. EXAMPLES OF ACTUAL COST OF CONCRETE. The fol- 
lowing data are given concerning the cost of concrete on several 
works where sufficient details have been published to be of 
value. 

Defenses Staten Island. 1 Cubical box mixer ; proportions by vol- 
ume, 1 cement, 3 sand, 5 broken stone; 5,609 cu. yds. of concrete. 



ITEMS. 


COST PER Cu. YD. 
CONCRETE IN 
PLACE. 


Cement, Portland, at $1.98 per bbl. 


$2.546 


Broken trap rock . . ' . 


1041 


Sand drawn from beach 


0.225 


Receiving and storing materials .... 


0.149 


Mixing, placing and ramming 


0.879 


Forms, lumber and labor 
Superintendence and miscellaneous .... 


0.477 
0.190 


Total cost per cu. yd 


.$5.507 



It is stated that hand mixing for a portion of the concrete 
used in another emplacement cost fifty-six cents more per 
cubic yard than machine mixing. 



1 Major M. B. Adams in charge. Report Chief of Engineers, U.S.A., 
1900, p. 837. 



236 CEMENT AND CONCRETE 

348. Defenses Tampa Bay, Florida. 1 Cockburn-Barrow 
mixer, with cableway for placing concrete. Shell concrete 
made up of 1 cement, 3| sand and 5 shell. 

1.31 bbls. cement, at $2.42 per bbl. ...... $3.17 

.71 cu. yd. sand, at 21 cents per cu. yd .15 

1.08 cu. yd. shell, at 50 cents per cu. yd 5.4 

$3.86 
Labor mixing $0.28 

Labor placing and tamping .33 

Labor on forms .155 .765 



Total cost per cubic yard $4.625 

The above does not appear to include costs of running ma- 
chinery, fuel, repairs, and depreciation of plant. 

At the same battery 2 in the following year the cost of broken 
stone and shell concrete was as follows: 

.9 bbl. cement at $2.46 (including $0.59 per bbl. 

storage) $2.214 

.28 cu. yd. shell, at $0.45 per cu. yd 128 

.47 " sand, at 0.12 " " 056 

.80 " stone, at 2.95 " " 2.360 

Total materials $4.758 

Mixing and placing $0.623 

Forms 370 

Total .993 



Total cost per cubic yard ...... $5.751 

349. Defenses San Francisco, Cal. 3 Cubical mixer; ma- 
terials drawn from bins into measuring cars, hoisted by elevator 
and dumped into hopper of mixer. Mixer given twelve to four- 
teen turns and concrete dumped into cars, pushed by hand out 
on trestles, and dumped in place. Average capacity plant, 280 
cubic yards per day of eight hours. Itemized cost of 8,328 
cubic yards of concrete in place was as follows: 



1 Report Lieut. Robert P. Johnson, Corps of Engineers, U. S. A., Report 
Chief of Engineers, 1899, p. 906. 

2 Report Lieut. Frank C. Boggs, Corps of Engineers, U. S. A., Report 
Chief of Engineers, 1900, p. 931. 

3 Maj. Charles E. L. B. Davis, Corps of Engineers, in charge. Report 
Chief of Engineers, 1900, p. 980. 



EXAMPLES OF COST 



237 



.758 bbl. Portland cement, at $3.03 per bbl . . . $2.298 

.887 cu. yd. rock, at $1.80 per cu. yd 1.597 

.41 cu. yd. sand, at 0.73 per cu. yd. . . ... . .299 

Water .016 

Cost of materials $4.210 

Concrete plant, erection per cu. yd. concrete . . $0.269 
Concrete plant, running expenses per cu. yd. . . .022 
Concrete plant, taking down .020 

Cost for plant, exclusive of purchase 

price .311 

Forms materials $0.272 

Forms labor in erecting .346 

Forms labor taking down .079 

Cost forms .697 

Labor mixing, placing and ramming .626 

Total cost per cubic yard $5.844 

350. In the building of a concrete dam for the enlargement 
of the head of the Louisville and Portland Canal/ comparison 
of cost of hand and machine mixing is given by Asst. Engr. J. 
H. Casey. 

1.63 bbls. natural cement, at $0.635 per bbl. . . $1.034 

2 volumes sand, .47 cu. yd., at .87 per cu. yd. . . .408 

5 volumes broken stone, .89, cu. yd. at $.84 . . .756 

Cost of testing cement .081 

Forms, material for .107 

Forms, labor making and setting up .168 

Cost materials and forms $2.554 

Hand mixed concrete: 

Cost of mixing $1.917 

Cost of placing and tamping .791 

Cost of mixing and placing $2.708 

Total cost hand mixed concrete per cu. 

yd. in place $5.262 

Machine mixed concrete: 

Charging and running mixer $0.864 

Placing and tamping .585 

Cost mixing and placing $1.449 

Total cost machine mixed concrete per 

cu. yd. in place $4.003 

Difference in favor machine mixing $1.26 



1 Capt. George A. Zinn, Corps of Engineers, U. S. A., in charge. 
Chief of Engineers, 1900, p. 3467. 



Report 



238 



CEMENT AND CONCRETE 



Since the above concrete was placed in large masses, the 
costs of labor are considered high, and it is probable the work 
was done with exceptional care. 

351. In the construction of the lock at the Cascades Canal 1 
the concrete plant was so arranged that the materials did not 
have to be elevated, but much of the work of transportation 
was done by gravity. The mixing of about eighteen hundred 
yards by hand permits a comparison to be. made with machine 
mixing by which method about seven thousand eight hundred 
yards were made. The costs were as follows: 



ITEMS. 


COST PER Cu. YD. OF CONCRETE. 


HAND MIXED AND 
PLACED BY DERRICK. 


MACHINE MIXED AND 
PLACED BY CHUTE. 


AMOUNTS. 


TOTAL. 


AMOUNTS. 


TOTAL. 


.805 bbl. Portland cement at 
$4.08 . 


$3.29 
.47 
.60 

.54 
'.15 
.22 


$4.90 

.37 
1.09 
.79 


$3.29 
.47 
.60 

.54 
!l5 
.22 

'.39* 
.04 

'.41 

.05 


$4.90 

.37 
.43 
.46 


.456 cu. yd. sand at $1.04 . . 
.579 cu. yd. gravel at $1.04 . . 
.317 cu. yd. broken stone at 
$1.70 ... 


Cost materials in concrete . 
Timbering 


Testing cement and general 
repairs . 


Forms and tests 


Mixing, labor 


1.07 
.02 

.60 

.19 


" repairs and fuel . . . 
Total cost mixing 
Placing, labor . . 


" fuel, tramways, etc. 
Total cost placing . . . 


Total cost concrete per cu. yd. 




$7.15 


S6.16 



352. In the construction of the retaining walls for the 
Chicago Drainage Canal, 2 a special plant was designed for the 
work on account of the large quantities of concrete required, 
and this, combined with the low cost of materials and the char- 
acter of the work, resulted in a very low cost concrete. On 



1 Maj. Thomas H. Handbury, Corps of Engineers, U. S .A., in charge. 
Report Lieut. Edward Burr, Report Chief of Engineers, 1891, Vol. v. Ab- 
stracted, Engineering News, June 2, 1892. 

2 " Construction of Retaining Walls for the Sanitary District of Chicago," 
by Mr. James W. Beardsley, and discussion by Mr. Charles L. Harrison. 
Jour. W. Soc. Engrs., Dec., 1898. 



EXAMPLES OF COST 239 

Section 14 the stone was selected from the spoil banks along the 
canal and could usually be obtained within one hundred feet. 
This stone, which was delivered, to the crusher by wheelbarrows, 
required some sledging to reduce it to crusher size. An Austin 
jaw crusher was mounted on a flat car with the Sooysmith 
mixer. "The cement, sand and stone were raised from their 
respective bins by means of belt conveyors running at the 
same rate of speed, but carrying buckets spaced proportional to 
the required ingredients." "The cost of a second hand plant 
used on this section was estimated at $9,600, including two 
crushers and two mixers at $1,500 for each machine. Common 
labor cost $1.50 per day; firemen, enginemen, and carpenters 
from $2.00 to $3.00 per day. The itemized cost is as follows: 



ITEMS. 



COST, CENTS 
PER CU.YD. 



General, including superintendent, blacksmith, water boys, etc. 

Quarrying, i. e., delivering stone to crusher 

Crushing 

Transportation, delivering sand and cement to mixer by teams 

Forms, exclusive of lumber 

Mixing 

Placing and tamping 



7.8 

30.3 

7.3 

14.2 
15.0 
12.1 
10.8 



Total 

Cost of plant (no salvage allowance) 
Cost of cement and sand .... 
Total cost concrete per cubic yard 



97.5 

40.7 

163.3 



$3.015 



The amount of concrete used on this section was 23,568 cu. yds. 
353. On Section 15 of the same work the conditions were 
somewhat different. The stone had to be quarried within about 
a thousand feet of the crusher. The stone, after being broken 
to crusher size, was delivered on the tipping platform of the 
No. 7 Gates crusher in cars drawn by a cable hoist. "The 
average output of the crusher for a day of ten hours was about 
210 cubic yards." The materials were transported to the 
mixer in four and one-half yard dump cars drawn by a light 
locomotive. The mixer was of the spiral screw type and de- 
posited the materials on a rubber belt conveyor. The mixer 
and operating machinery were mounted on a car which pro- 
pelled itself by means of rope and winch. The plant for this 
section was new and estimated to cost $25,420, including $12,000 
for one crusher. 



240 CEMENT AND CONCRETE 

The detailed cost is as follows: 



ITEMS. 



COST 

PER Cu. YD. OF 

CONCRETE, 

CENTS. 



General, including superintendent, blacksmith, teams, etc. 
Quarrying (exclusive of 8.3 cents for explosives) . . . 

Crushing 

Transportation, delivering cement, sand and stone on a 

platform beside the mixer 

Forms, exclusive of timber 

Mixing, including shoveling materials from platform to 

mixer 

Placing and tamping 



Total 

Cost of plant (no salvage allowance) 
Powder for quarrying .... 
Cement and sand . 



8.2 
19.2 
12.8 

8.1 
14.2 

25.0 

11.6 



99.1 

56.7 

8.3 

158.6 



Total cost concrete per cu. yd. 



$3.227 



The amount of concrete used on this section was 44,811 
cubic yards. 



CHAPTER XV 

THE TENSILE AND ADHESIVE STRENGTH OF CEMENT 

MORTARS AND THE EFFECT OF VARIATIONS 

IN TREATMENT 

ART. 45. Tnfc TENSILE STRENGTH OF MORTARS OP VARIOUS 
COMPOSITIONS AND AGES 

354. THE PROPORTION OF SAND. The rate of change in 
the strength of mortars as the proportion of sand is increased 
varies greatly for different cements. The fineness and chemical 
composition of the cement, and the quality of the sand, are the 
most important factors influencing this rate of change upon 
which the question of the relative economies of different mor- 
tars is so largely dependent. 

Table 67 gives the results of tests with two brands of Port- 
land cement mixed with from two to ten parts of river sand, 
the age of briquets being six months and two years. It is of 
interest to notice that the strengths of the mixtures are ap- 
proximately in the inverse ratio of the number of parts of sand 
used. Thus the strength with six parts sand is approximately 
two-sixths of the strength with two parts, while with ten parts 
sand, the strength is nearly two-tenths of that with mortar 
containing two parts. 

TABLE 67 

Rate of Decrease in Strength with Addition of Sand 
PORTLAND CEMENT; RIVER SAND, " POINT AUX PINS " 





TENSILE STRENGTH, LBS. PER SQ. IN. 




TO 1 

CEMENT BY 


6 MONTHS. 


2 YEARS. 


PROPORTIONATE 
STRENGTH, 
Two YEARS, IF 




H 


R 


H 


R 


Mean. 


1 TO 2 100. 


2 


512 


504 


534 


548 


541 


100 


3 


390 


335 


363 


355 


359 


66 


4.09 


295 


261 


296 


288 


292 


54 


6 


175 


144 


191 


174 


182 


35 


8 


113 


96 


132 


132 


132 


24 


. 10 


64 


74 


104 


116 


110 


20 



241 



242 



CEMENT AND CONCRETE 



355. In Table 68 similar results are given for two samples 
of Portland cement and two kinds of sand, neat cement speci- 
mens being included in the comparison. The one-to-one mor- 
tars give a higher strength than neat cement, and even the 
mortar containing two parts of the limestone screenings is 
stronger than the neat specimens. From the one-to-one mor- 
tars the strengths decrease rapidly as more sand is added, until 
five parts sand are used, but the strengths then decrease less 
rapidly as larger additions of sand are made. 

TABLE 68 
Rate of Decrease in Strength with Addition of Sand 

PORTLAND CEMENT, BRAND R; SAND, CRUSHED QUARTZ AND LIMESTONE 

SCREENINGS 





TENSILE STRENGTH, POUNDS 
PER SQ. IN. 


PROPORTIONATE 
STRENGTH IF STRENGTH 
1 TO 1 MORTAR =. 100. 


PARTS SAND 










TO 1 


Sample 


Sample 






CEMENT BY 


Cement H H, 


Cement II, 






WEIGHT. 


Crushed Quartz 


Limestone 


Crushed 


Limestone 




Sand, 20-30. 
Age Briquets, 
6 Months. 


Screenings, 20-30. 
Age Briquets, 
6 Months. 


Quartz. 


Screenings. 





689 


686 


82 


78 


1 


840 


881 


100 


100 


2 


521 


703 


62 


80 


3 


368 


508 


44 


58 


4 


236 


335 


28 


38 


5 


203 


267 


24 


30 


6 


156 


178 


19 


20 


8 


104 


138 


12 


15 


10 


78 


98 


9 


11 



356. In Table 69 two samples of natural cement are treated 
in a similar manner, from one to eight parts river sand being 
used in the mortars. With Sample II the strength is dimin- 
ished rapidly until five parts sand have been added, but with 
further additions of sand, the strength is decreased more slowly. 
Sample 18 S gives quite a different curve, as the one-to-two 
mortar is stronger, and the one-to-three mortar is but little 
weaker than the one-to-one. With four parts sand the mortar 
shows a marked falling off in strength, but further additions of 
sand diminish the strength more slowly. 

357. INCREASE m TENSILE STRENGTH WITH TIME. In 

Table 70 are given the results obtained in tests of tensile strength 



STRENGTH OF MORTARS 



243 



TABLE 69 

Rate of Decrease in Strength with Addition of Sand. Natural 
Cement, Brand Gn; River Sand, "Point aux Pins" 







TENSILE STRENGTH, POUNDS PER SQUARE INCH. 


SAND TO 1 
CEMENT 
BY WEIGHT. 


AGE. 


G MONTHS. 


2 YEARS. 


Proportionate 
Strength, Two 
Years if 1 to 2 
= 100. 


Sample 
Gemeiit 


II. 


18 S. 


18 S. 







380 








1 




297 


308 


280 


86 


2 




260 


314 


324 


100 


3 




183 


280 


294 


91 


4 




128 


193 


187 


58 


5 




81 


161 


165 


51 


6 




69 


142 


172 


53 


7 




56 


119 


156 


48 


8 




53 


101 


114 


35 



with twelve samples of Portland cement, illustrating the rates 
of increase in strength from seven days to three years. It is 
seen that rich mortars gain strength rapidly, neat and one-to- 
one mortars showing usually eighty to ninety per cent, of their 
ultimate strength in twenty-eight days. Mortars containing 
not more than four parts sand to one cement give practically 
their ultimate strength at six months. It is also of interest to 
notice that the variations in strength among the several sam- 
ples are not very great. The lowest strength at the end of 
two to three years is seventy-five to eighty per cent, of the 
highest. 

358. In the case of natural cements, results for ten brands 
of which are given in Table 71, only fifty to seventy per cent, 
of the ultimate strength is gained in the first twenty-eight 
days; with mortars containing three parts sand to one cement 
the average result at twenty-eight days is less than forty per 
cent, of the strength at two years. Most of the samples gain 
some strength after six months, but two samples fail at two 
years which had given a fair result at six months. The varia- 
tions in strength among the several samples are very much 
greater than with Portland cements; even omitting the two 
samples that failed, the strength of the highest is two or three 
times the strength of the weakest sample at two years. 



244 



CEMENT AND CONCRETE 



H 
H) d 

PQ ^ 



^911190 i 
<n puisg 9 



!*,9in90 I 



OF SAND TO ONE CEMENT 
BY WEIGHT. 



SAND TO ONE 
BY WEIGHT. 



ONE PA 
CEME 



FINENES 
PER CE 
PASSIN 



8X89 A" 



8X89 A" 



, CO O .00 . .CO .05 
00 00 CO CO -^ 

<M <M ' (M ' * <N ' C^ 



8X89 A" g 



sora 9 



SA'Bp 



8X89^ g 



soui 9 



stop 



8X891" g 



8x139 A" 55 



soui 9 



82 



'r-' II-H 

cococococococococococo 



O (M 
<J<S >O 

cq co 



OUQOCCOiCDi (OOCO 
GO OO OO t^- t^- O ^O iO CO 

COCOCOCOCOCOCOCOCO 



00 

-^ 



I- 
t- OO 

O CO 



Oi it (MO5>O^- 
OOOti-Hr-IQOCO 



CO 

Oi co 

t^ t^ 

00 



OCOOT^OOt 



i-H O (M 
O O O1 



OO5i lTOCO'^Tt<i-H'^5OCOO 
iO CO OS CO *-O O i"^ C^ "^ CO i^ CO 
CO CO ^O *O "^ CO CO CO O CO CO iO 



^ CO 
iO T* 
CO CO 



t-HCOO5Dt^t-^(MOJCOaoai 
OOO^TtiTtiTtiiOCOO^Tti 



OOOSrHiOCOi i^OOO<M 

COi-H 

t~- tr- 



COCOCOOOCOOOOt-t~i-tO 
O^-^CO'^'NOSCC^OOCO 

r--:ot-cocDr~t^cot-coi>- 



co 

00 






t-OO^i-HOOO 



9900' 



.COt^COOOiOO 



gjdureg 



ptrejg 



STRENGTH OF MORTARS 



245 



THREE PARTS SAND To 
ONE CEMENT. 



Two PARTS SAND 
ONE CEMKNT. 



PART SAND TO ON 
CEMENT. 



soiu 9 



soiu g 



stop 80 



'SOU! 9 



stop 



stop 



soui 9 



soui g 



stop 83 



stop 



soui 9 



stop 82 



stop 



top 



t~iOO5(NOOOOOOt^- CO t^- t~ 
i itOO5l"-OOCOt-Tl<M<7<l lO-^QO 
r-iC^CiO^r-H^COCOCO (MCOC^ 



0000 00 

1 OOi-"'NOOO'TOO O4 QO <M 1^- 
i ii ilMCOCqcNCMtNCN-M (M GS| <N rH 



t- 05 01 05 00 



T i i tO *"" >O t^* OS O O ^ 

OS O O O C/j O CO * ^O ^ 
^Tjif-iOCC' 00 O >O CO 



CO 
T^I ^ 10 
OCOr-4 



T*! C^-l O CO "^ 



COCOfMt^OSOOCOOO^OS 

* o 25 55 35 3 oo ? 

F (i li ijvji '(NrHi (iiC<l 



O^Dr^ 

i- oo i- 



(jq OS OS O CO 
I CO CO OS Tt< 



iCCDt-COCOt^ 



i 1 - i O t^ O 
** -rt<?OCO<7^ 



<M-<tlO 
ROC 



OCO 
COO 



THOSOCOOSOSOQO C^ICO^OS 
OOl^QO^O'MQO'-'O QOOSOOCO 

I li ! r-lr-l<M<M!-l<MCM t-H l-l rH l-H 



CO CO CO O * l>- 00 W OS 

^COOOt^'-H-^CDt^ 



ooosiWT^ojWOsaooj 
00 OS O "* CO --* O> O OS 



00 yf CO 



o o o o 

-fl 



246 



CEMENT AND CONCRETE 



359. Table 72 shows in some detail the rate of increase in 
strength of a sample of natural cement when the specimens 
are maintained in air and in water. This table has several 
points of interest. When hardened in water, the cement gained 
steadily in strength up to six months, when it began to fall 
off, and at two years this cement failed, as is shown in Table 
71. The neat cement specimens hardened in air are very 
irregular, as usual. These specimens showed high strength at 
three months, suffered a marked falling off at six months and 
one year, but showed a remarkable strength, equal to neat 
Portland cement, at two years. The strength developed at 
one year by specimens of this sample containing two parts 
sand to one cement and hardened in air, is also equal to that 
shown by similar mortars of Portland cement. 

TABLE 72 
Rate of Increase in Strength in Water and Air 



AGE OF BRIQUETS. 


TENSILE STRENGTH, POUNDS PER SQUARE INCH. 


NEAT CEMENT. 


MORTAR Two PARTS SAND 
BY WEIGHT TO ONE 
CEMENT. 


Water. 


Air. 


Water. 


Air. 


1 day 
7 days 
14 days 


81 
192 
232 
305 
390 
437 
432 
395 


152 

254 
315 
473 
551 
372 
314 
731 


135 


142 


28 days ..... 
3 mos 
6 mos 
1 year 
2 years 


232 
367 
409 
249 


271 
459 
475 

537 



All cement, Brand Hn, Sample 26 S, which fails in water after two years, 
see Table 71. 

ART. 46. CONSISTENCY OF MORTAR AND AERATION OF CEMENT 

360. EFFECT OF CONSISTENCY OF MORTAR ON TENSILE 
STRENGTH. The results in Table 73 are from briquets of 
Portland cement with two parts " Standard" crushed quartz. 
The consistency of the mortars varied from a "trifle dry/' in 
which water rose to the surface only after continued tamping, 
to a wet mortar which would just hold its shape when placed 
in a heap on the slab. Half of the briquets were immersed, 



CONSISTENCY OF MORTAR 



247 



while the remainder were stored in the air of the laboratory. 
The air hardened specimens gave higher results in all cases 
than those hardened in water. The highest strength was given 
in general by the dryest mortar, but the differences in strength 
decrease as the age of the specimen increases. 

TABLE 73 

Effect of Consistency on the Strength of Portland Cement Mortar 
Hardened in Water and Air 



AGE OF BRIQUETS. 


TENSILE STRENGTH, POUNDS PER SQUARE INCH. 


Consistency of Mortar. 


Briquets Hardened in Fresh 
Water. 


Briquets Hardened in Air 
of Laboratory. 


a 


b 


c 


d 


e 


a 


b 


c 


d 


e 


7 days . . . 


340 


310 


226 


191 


158 


407 


341 


263 


230 


202 


28 days . . . 


383 


378 


314 


291 


249 


506 


463 


345 


393 


302 


3 months . . 


515 


535 


514 


429 


411 


665 


593 


638 


597 


451 



Cement: Brand R, Sample 18 R, with two parts "Standard" sand. 
Consistency: a, trifle dry; b, O.K.; c, moist; d, very moist; e, would just 
hold shape. 

361. Tables 74 and 75 give similar results for Portland and 
natural cement mortars, respectively, all specimens having 
hardened in water for three months. The amount of water 
used in gaging had a wide range, giving mortars of all consis- 
tencies from very dry to very moist. The richness of the mor- 
tar was also varied, from neat cement to five parts sand. A 
comparison of the results in these two tables indicates that the 
highest strength is usually given by mortars a trifle dryer than 
that considered right for briquets; that an excess of water is 
less deleterious to rich mortars than to lean ones, and to Port- 
land cement than to natural cement. 

362. Conclusions. Although all of these tests indicate the 
superiority of dry mortars, in considering the effect of consis- 
tency from a practical standpoint, one must not fail to consider 
the difference between the conditions existing in the actual use 
of mortars and in laboratory tests. When mortar is used in 



248 



CEMENT AND CONCRETE 



TABLE 74 

Variations in Consistency of Mortar 
EFFECT ON STRENGTH OF PORTLAND MORTAR AT THREE MONTHS 





TENSILE STRENGTH, POUNDS PER SQUARE INCH, FOR 


PARTS SAND TO . 


CONSISTENCY NUMBER. 


1 CEMENT 




BY WEIGHT. 






















1 


2 


3 


4 


5 


6 


7 


8 


9 





608 


635 


763 


744 


708 


707 


729 


C85 




1 


513 


543 


618 


588 


594 


613 


566 


566 


538 


2 






429 




447 


398 


393 


382 




3 




289 




322 


329 


310 




279 




5 


. . . 


208 




230 


201 


189 




167 





Consistency 
Significance of numbers : 
Increasing per cent, water 
used for higher numbers. 



1 Very dry ; little or no moisture appeared on 

surface of briquets. 

5 About proper consistency for briquets. 
9 Very moist ; mortar would barely hold shape 

and shrank in molds in hardening. 



TABLE 75 

Variations in Consistency of Mortar 
EFFECT ON STRENGTH OF NATURAL CEMENT MORTAR AT THREE MONTHS 



PARTS SAND TO 
1 CEMENT 
BY WEIGHT. 


TENSILE STRENGTH, POUNDS PER SQUARE INCH FOR 
CONSISTENCY NUMBER. 


1 


2 


3 


4 


5 


6 


7 


8 


9 



1 
2 
3 
5 


. . . 


239 


372 
312 
255 
217 
150 


373 
314 

283 
208 
155 


277 
206 
125 


305 

286 
258 
183 
101 


281* 
242 


268 
241 
204 
139 

74 


263 

207 
176 







Consistency Significance of numbers: 

1 Very dry; little or no moisture appeared on surface briquets. 
5 About proper consistency for briquets. 

9 Very moist ; mortar would barely hold shape and shrank in 
hardening. 

masonry, the stones or bricks, even though they be dipped, 
or sprayed with a hose before setting, are very likely to press 
out or absorb considerable moisture from the mortar. To 
realize this one has only to raise a heavy stone just after it has 
been bedded; and the greater ease of setting either stone or 



AERATION OF CEMENT 



249 



brick/ and obtaining a full mortar bed, with a rather wet mor- 
tar, is appreciated by all masons. In the laying of concrete, 
the difficulties of obtaining a compact mass with a dry mortar are 
also not to be overlooked, but this point is discussed elsewhere. 

363. Effect of Aeration on the Tensile Strength of Cement. &- 
Portland cements that are not perfect in composition and; 
burning, and that therefore contain free lime, may sometimes 
be rendered sound by exposing them to air, and such exposure; 
was at one time considered almost essential in Portland cement 
manufacture. 

Fresh Portland cements that are slightly defective may have 
their properties quite radically changed by such treatment; 
their rate of setting becoming first more rapid, and then, by 
further aeration, slower, and their tendency to expand over- 
come or ameliorated. Portland cements that are perfectly 
sound suffer some loss in specific gravity by the absorption of 
carbonic acid and water from the atmosphere, but moderate 
aeration has no radical effect upon their strength, and Port-; 
lands deteriorate but very slowly by storage, provided the. 
cement is kept dry and does not cake in the package. 

Natural cements, however, usually suffer by aeration, and 
this is illustrated by tests on several samples of one brand; 
given in Tables 76 and 77. Of the four samples in Table 76, 

TABLE 76 
Effect of Aeration on Four Samples of Same Brand Natural Cement 





TENSILE STRENGTH, POUNDS PER SQUARE INCH. 


NUMBER' 




WEEKS 






CEMENT 


Age of Briquets, 6 Months.to 7 Months., 


Age Briquets, 2 Years.; 


AERATED. 








Sample QQ 


ss 


NN 


00 


NN 


OO ' 





242 


183 


343 


340 


316 


306 ;:( 


2 ' 


237 


269 


357 


506 


368 


432 ::.,, 


5 . 


256 


403 











7 


268 


358 


. . 




: . . . ' 


= .: 


10 


. . ''. 


. . . ' * 


J226 


212 


246 - 


284 /- 


11 , 


313 


279 




. 






13 


. . . 


. . . 


213 


218 


260 


258 '" 



Cement: Brand Gn; Sand, two parts crushed quartz to one cement. 
All 1 briquets of one sample were made by one molder and same percentage 
water used. 



250 



CEMENT AND CONCRETE 



NN and OO showed an improvement by two weeks' exposure 
to air, spread out in a thin layer, but longer exposure resulted 
in a serious loss of strength. Of the other two samples, SS was 
greatly improved by five weeks' aeration, but longer exposure 
was detrimental, while sample QQ showed a continuous im- 
provement up to the limit of eleven weeks' exposure to air. 

In Table 77 the effect of aeration on five samples of the same 
brand is shown. One of these samples was overburned and 
was rendered practically worthless by fourteen weeks' exposure 
to air. Nearly all of the samples in this table were seriously 
affected by six weeks' aeration. 

TABLE 77 
Natural Cement. Effect of Aeration 









TENSILE STRENGTH, POUNDS 






ft. 


CKMENT. 


PARTS 




PEB SQUARE INCH, CEMENT 











SAND 


AGE OF 


AERATED. 






5 




TOl 


BRI- 








3 w 

W > H 9 


Brand. 


Sam- 
ple. 


MENT. 




4 to 5 
days. 


11 to 12 

days. 


45 to 51 
days. 


99 
days. 


a 


b 


SNfeK 
02 35 

a 


Gn 


84 


2 


6 mo. 


414 


321 


208 


216 


80.5 


54 


3.01 


tt 


83 


tt 


it 


463 


392 


211 


235 


85.9 


41 


3.11 


ii 


82 


tt 


tt 


445 


350 


217 


266 


85.6 


34 


3.09 


tt 


U' 


tt 


tt 


383 


354 


273 


274 


87.8 


23 


2.95 


" 


0' 


tt 


" 


263 


293 


277 


52 


89.7 


97 


3.14 



a Fineness expressed as per cent, passing holes .0046 inch square. 
b Time setting fresh cement, time to bear ^ inch i Ib. wire. 

ART. 47. REGAGING CEMENT MORTAR 

364. The Effect of Thorough Gaging. The value of thor- 
ough gaging is a point frequently overlooked in the preparation 
of mortars and concretes. Table 78 gives a few of the results 
obtained in experiments to determine the effect of thorough 
work in mixing. The tests are made with two brands of natural 
and one of Portland, with two parts sand to one cement by 
weight. The two minutes' mixing with hoe and box method 
gave a more thorough gaging than could have been accom- 
plished in the same time with a trowel, and represented 
about the amount of work put on mortars for testing. We are 
not, therefore, comparing well mixed and poorly mixed mortars, 
but: rather well gaged and better gaged. The- effect of the 
additional work is shown in all cases; to double the time spent 



REGAGING MORTAR 



251 



in gaging, increases the strength of the resulting mortar about 
five per cent., while to quadruple the time adds twenty -six 
per cent, to the strength. 

TABLE 78 
Effect of Thorough Gaging 





CEMENT. 


SAND, Two PARTS 
TO ONE CEMENT. 


TENSILE STRENGTH, LBS. 
PER SQ. IN., FOR MORTAR 
GAGED, 


REF. 










Kind. 


Brand. 


Kind of Sand. 


2 Min. 


4 Min. 


8 Min. 


1 


Natural 


Gn 


( Pt. aux Pins ) 
\ Pass #10 Sieve J 


352 


356 


482 


2 


u 


it 


Standard 


418 


459 


572 


3 


II 


An 


j Pt. aux Pins ) 
I Pass #10 Sieve f 


368 


376 


421 


4 


Portland 


R 


( Pt. aux Pins ) 
\ Pass #10 Sieve f 


525 


554 


616 


Mean 


416 


436 


523 


Prop 


ortional ... 


100 


105 


126 





365. REGAGING. When more mortar is mixed at one 
time than is required for immediate use, there is always a 
temptation to retemper the mass and use it, even though it 
may have been standing for some time. The practice is 
usually prohibited by specifications and strenuously opposed 
by engineers. The tests recorded in Tables 79 to 83 were 
made to determine the effect of regaging on the resulting 
strength of the mortar. 

366. The results obtained with two brands of Portland ce- 
ment are given in Table 79. The first result in each line of the 
table is the strength attained by the mortar when treated as 
usual. The severity of the treatment of the mortar as regards 
regaging is shown by the letters heading the columns and the 
corresponding foot notes. The first general statement to be 
made concerning the results in this table is that in no case is 
the effect of regaging Portland mortars containing sand shown 
to be seriously deleterious to the tensile strength. Neat cement 
mortar is not improved by regaging, and if allowed to stand 
more than one hour, and then made into briquets without any 
further addition of water, the strength is considerably decreased. 
If water is added and the mortar frequently regaged, however, 



252 



CEMENT AND CONCRETE 






K 





- 


. . . <N ^ CO CD 




- 


. . . . CO <M CO <N . 

U3 O CO, CD ' ' 








A 


** 


' -SS 




Pd 
H 


- 


CCt^CDOS ' CDOt-^rH 


o< 






p 


* 


' -So 











S 


--g . 









H 


<u 




S 






03 






S 


ts 








-sti 


S 





t^ t>- >O 1 t Tt^ . -^ 

t~- i ' <N QO (X) O 










>Q 


rH CD 










e 


CO ^O *O O Oi T-H |>- Tf4 OO 'M ^" GO 


1 


ll 


S^St^S^* >i" St>sS^ 
CDrHCCrH^ (N COT-H?Oi-l 




Parts 
tol 

Cement. 





.02 


Q 

3 




d . . . . i^j 

- i 

PH 

PH 







^^5^W^^.^v.-v.^ 


H 
fe 

M 


CO 




M 
- w 
u 


a 






n 


" 1 


c 


I 


1-1 I-H 1-1 




nd 
std 3 
stand 5 
that 
stand 3 
that 
that 
stan 
d 



as usual. 
allowed to 
allowed to 
allowed to s 
allowed to s 
except 
wed to 
except 
cept 
d to 
to s 



d 
r 



ar 
ar 
rtar 



as 
r allow 
as 
as 



alow 
allow 



a Trea 
6 Mort 
c Mort 
d Mort 
e Mo 
/ 
g 
h 



me 
me 
rt 



Mo 
Sa 
Sa 
i : Mo 
K Mo 



REGAGING MORTAR 



253 



even neat cement mortar does not suffer a great decrease in 
strength by three to six hours standing. Rich mortars, con- 
taining one part sand, are not seriously affected by standing 
three hours if regaged frequently. Poorer mortars, with two 
to four parts, sand, show an actual increase in strength as the 
effect of such severe treatment as standing five hours, if re- 
tempered with more water once an hour. These two brands 
were slow setting Portlands, beginning to set in forty minutes 
to two hours. The increase in strength of the regaged mortars 
is doubtless due, at least in part, to the more thorough gaging 
which they received. 

Table 80 gives similar results of briquets one year old made 
at another time with two parts river sand. The fact that dur- 
ing the delay between, the making and use of the mortar it 
should be frequently re tempered with water to make up for 
the loss by evaporation,, is plainly shown. 

TABLE 80 

Regaging Portland Cement Mortar 



TENSILE STBENGTH, POUNDS PER SQUABE INCH, FOB VABYING TBEATMENT. \ 


a 


c 


d 


e 


/ 


h 


i 


j ;: 


579 


565 


569 

-, 


570 


568 


. . . 


. . . 


. . . 


554 


579 






. . . 


627 


624 


560 



Cement: Portland, Brand R, Sample 42. M. Sand: 2 parts "Point aux 
Pins/' passing No. 10 sieve. Age of briquets, 1 year. 
Treatment:: a L-Iolded as soon as gaged. 

c Mortar let stand 1 hour, regaged and briquets made. 

d Mortar let stand 3 hours, regaged each hour. 

h Mortar let stand 3 hours, regaged each hour and water 

added to restore original consistency. 
e Mortar let stand 5 hours, regaged each hour. 
i Mortar let stand 5 hours, regaged each hour and water 

added to restore original consistency. 
/ Mortar let stand 5 hours, regaged and briquets made. 
/ Mortar let stand 5 hours, regaged and briquets made ; 
water added to restore original consistency. 

367. Similar tests with natural cements are shown in Table 
81, and it appears that cements of this class, especially if mixed 
neat, will not stand the same severe treatments without injury. 



254 



CEMENT AND CONCRETE 



H 


- 


: :ggs? : : : : 

. <N CO (M . . . . 


* 


- 


' QO OO C- 
<M CO <M 


" H 


- 


(M <N <M CO <N -H 1-1 


11 


S 


. <N ^ CO 


/-? 






P^H 






"0 






H ~i 






*3 




oo 






^ co 


ga 
ag 





<7<l CO 


HO 




1-1 GO 


^S 




CN S 


M 






8 


8 


^^OiOCOOt^Oii-i 
OOiO>^'^"<*'OiOO 
(NG^-^^COCO(Nr-i(N 


Hfc 



BRIQUETS. 


OB 05 05 05 00 

S>>2>->5>>c>-} 

COrHO(NCOCOi-lCOrH 




Parts to 
1 Cement. 


(MS^OOOiMlM 1 ^^ 


G 


5 


O 

O* T3 

s U l|ss..3 

OH* 


9 


Sample. 


ot ^ 


9 




Brand. 


d - - - dfl --^ 


J3 


a 


rH(MCO^'0>t>.OOOi 



111.1 

3-33 s 




REGAGING MORTAR 



255 



$ 

I 

41 

S 

J 



iJ (d 

n ^ 
5 



Regaging 





J 


* ' " OS 









: : ; ;! 




3 










... 


g 


4 


co 









"^ 




5S 








^ 

g 



MI 


s : 


W U9 OS* 
of <N* ~ - 


fc . 

4 




.^ .CO . 


Ifill 






o a 
fcg 





.J2 .'Si 

CO ^ 


rt".S S) SD 


1-1 H 
., ^ 




. 


^^ 53 bD bC 


33 




* O ' t- ' 


^"o "o 


DH 

*B 

si K 

h 





.'co .w ; 


Illl 

illi 

d c c 


s 


:I : : : 


TH, POUNDS : 
:EIVING DIFF 






3^^^ 

?jff 

2 IT IT 

Illl 


V 


^ OS 


< 


^ ' O * 1-1 
O 'CO <M 


9 S 






_^ fl 


CNSILE STREN 
R] 


u 


CO .CO .CO 


UH 4_i H 

o ^ o o 

2 S -s ^ 

>f It 

s - ^ ^ 

C3 nH C 


- 


Sill! 


P 


e 


OS ?D 10 ^ 

1 O QC tC 00 O 
CO CO CO CO -^ 


T3 ,3 .H 

^ ii ti 








^3 QJ ^ "^ 




S3 

32 


<N <N <N :N :N 


*iffrl 


a 
jq 


S3 




r S c 2^ 2^ gf 

II^II 

'C _ "* <N <N -^ 




o 
6 L 


00 


d 


^ 1 


S-r,^-^^ 




3 


&( jj 


3 *9 9 3 S Si 

g 2 "S to 






^ ? 


O -3 ^S ^^ jj 

c S -2 JS ^ ^ ' 






P-I 


^ I* >H t_l H fe 

23333 




. 




9 v 4B w V 

e -e -e -e -e 




^^ 


^%* ^* 


o o o o o 






^ ^ CO i ^jj 


sssssc 


i 




^ 


1 1 1 1 1 


a 






G -O O "X3 ' 


3 





a- fl- a 


I 




I 


O <l " O " 


+3 


-H 


* 




nsistency o 
nsistency o 
nsistency o 
cy 




regaging 
regaging 
aging 
aging 
briqu 
in 




te 
wate 
wate 
water 
hour, 
ours, 



as b exce 
as c, except 
as d, except 
as e, except 
r let stand 1 
let stand 4 



e 
e 
e 
rta 
rta 




/ 
g 
h 

* 
k 



256 



CEMENT AND CONCRETE 



Neat cement mortars of these two brands appeared more plastic 
when they were retempered with more water after standing one 
hour (column /), but if allowed to stand three hours (column 
t), they had then become quite hard set. Mortars containing 
two parts sand that had stood sixty to ninety minutes with 
intermediate retempering, showed a. slight increase in tensile 
strength, but more severe treatment was deleterious. 

In Table 82 the mortars all contain two parts sand to one 
of cement, by weight. The only cases of any serious results of 
retempering are for mortars standing four hours and regaged, 
at intervals of one-half hour or one frour, with no water added 
to the original mortar. Briquets made from mortar that had 
been gaged every half hour, and was molded two honrs after 
first mixed, showed a somewhat higher strength than briquets 
made of fresh mortar. 

Table 83 shows that the behavior of regaged natural cement 
mortars, as shown in the preceding tables, is not an eccentricity 
of one or two brands. Mortars containing two parts sand do 
not appear to suffer in tensile strength by being allowed to 
stand two hours if regaged' hourly. 

TABLE 83 

Effect of Regaging on Tensile Strength, Five Brands Natural 

Cement 















TENSILE STRENGTH, POUNDS PER 


BTS SAND T 
CEMENT. 


AGE 
BRIQUETS. 


TIME 
ELAPSED 

BETWEEN 

FIRST GAG- 

1NG AND 

MOLDING. 


)TAL NUMBE 
GAGINGS. 


INTERVAL 
BETWEEN 
SUC- 
CESSIVE ' 
GAGINGS. 


SQUARE INCH. 


BRANDS. 












-< 
PH 






H. 


t :r 


H: En 


An 


Dn 


Kn 


Hn 


2 


28 days 




1 




68 


171 


231 


178 


174 


2 


" 


2 hours 


3 


1 hour 


109 


168 


310 


178 


190 


2 


6 months 




1 


... I 228 


328 


306 


361 


273 


2 


K 


2 hours 


3 


1 hour 1 284 


382 


307 


416 


347 


4 


28 days 


. 


1 




39:. 


49 


149 


23 


58 


4 


ii 


2 hours 


3 


1 hour 


33 


70 


146 




56 


4 


6 months 


. 


1 




IjM 


146 


188 


216 


129 


4 


u 


2 hours 


3 


1 hour 


97 


147 


184 


227 


137 



NOTES: Sand, Point aux Pins, passing No. 10 sieve. 
In general, each result mean of five briquets. 
All briquets made by one molder and stored in one tank. 
All mortars appeared about same consistency when molded. 
No water added in regaging except Brand Kn, 1 to 2 mortar, 
standing two hours. 



MIXING CEMENTS 257 

368. Conclusions. The conclusions to be drawn from these 
tests appear to be as follows: The cohesive strength of mortars 
of neat cement is appreciably diminished if they are allowed to 
stand a considerable length of time after gaging before they 
are used. Sand mortars, especially of Portland cement, usually 
develop a higher tensile strength under moderate treatment of 
this kind; and if regaged frequently, with sufficient water added 
to keep them plastic, mortars of slow setting cements may be 
used several hours after made without serious detriment to the 
tensile strength. Portland cements withstand severe treatment 
better than natural cements. 

The effect of regaging on the adhesive strength is shown in 
Table 117, 405. These tests were quite severe and pointed to 
the conclusion that the adhesive strength is diminished by stand- 
ing and regaging, rich mortars and natural cement mortars 
being most affected. 

The effect of regaging on a given sample should be investi- 
gated before it is permitted to any great extent, or in the most 
careful work. Regaged mortars are said not to give good re- 
sults in sea water, and it may be expected that quick setting 
cement will be injured by regaging. 

ART. 48. MIXTURE OF CEMENT WITH LIME, ETC. 

369. Mixture of Portland and Natural Cements. For cer- 
tain uses mortar is sometimes made from a mixture of Portland 
and natural cement, with the idea of retaining some of the 
properties of the Portland without involving the expense of 
using a clear Portland mortar. Several tests have been made 
to determine the rate of hardening and the ultimate strength 
of such mixtures. 

The mortars used in the tests given in Table 84 contained two 
parts sand to one of cement, and the cement was composed of 
one-eighth, one-quarter and one-half Portland to seven-eighths, 
three-quarters, and one-half natural. Mortars made with Port- 
land alone and with natural alone are included for comparison. 
It is seen that the mortars containing some Portland harden 
more rapidly than the natural cement mortar, so that the in- 
creased strength developed at short periods is more than pro- 
portional to the per cent, of Portland used. The results ob- 



258 



CEMENT AND CONCRETE 



tained at two and three years, however, indicate that mortars 
containing only a small proportion of Portland, as one-eighth 
or one-quarter, do not give a higher ultimate strength than is 
obtained with clear natural cement mortar. 

TABLE 84 

Tensile Strength of Mortars Made with Mixture of Portland and 

Natural 












TENSILE STRENGTH, POUNDS PER SQUARE INCH. 


8 


BRIQUETS. 


Per ( Portland 


100 


50 


25 


12.5 





I 




Cent. ( Natural 


00 


50 


75 


87.5 


100 


1 


7 days 




291 


205 


108 


75 


24 


2 


28 days 




357 


264 


219 


190 


123 


3 


6 months 




550 


425 


378 


300 


322 


4 


1 year 


. 


574 


441 


360 


336 


291 


5 


2 years 




543 


449 


375 


343 


393 


6 


3 years 




592 


501 


428 


370 


429 



NOTES. Portland cement, Brand R, Sample 42 M. 
Natural cement, Brand Gn, Sample 54 R. 
Sand, two parts of " Point aux Pins," pass No. 10 sieve, to one 

part cement by weight. 

All briquets made by one molder and immersed in one tank. 
Each result, mean of ten briquets. 

370. In Table 85 four kmds or mixtures of cement are 
used, Portland, natural, an " Improved cement" or a cement 

TABLE 85 
Comparisons of Portland, Natural, and " Improved " Cements 





***** 




TENSILE STRENGTH, POUNDS PER SQUARE INCH. 




H j H fc M 


AGE OF 




REF. 


"Jap 


WHEN . 

BROKEN. 


Portland, 
Brand U. 


" Improved," 
Brand Nn. 


Portland,20%, 
Natural, 80%. 


Natural, 
Brand, Mn. 


1 


None 


7 days 


547 


206 


250 


199 


2 


it 


28 " 


586 


293 


341 


270 


3 


One 


7 " 


458 


169 


200 


165 


4 


" 


28 " 


569 


253 


331 


234 


5 


1C 


7 months 


702 


550 


578 


517 


6 


" 


2 years 


577 


563 


534 


497 


7 


Two 


2 " 


522 


510 


573 


529 


8 


Three 


7 days 


176 


52 


80 


49 


9 


" 


28 " 


272 


122 


143 


114 


10 


K 


6 months 


389 


301 


282 


255 


11 


It 


2 years 


371 


346 


356 


342 


12 


Mean 


2 years 


490 


473 


488 


456 



LIME WITH CEMENT 259 

sold as a mixed cement, and a sample made by mixing twenty 
per cent, of the Portland with eighty per cent, of the natural. 
The first point noticed is that the ''Improved" cement does 
not exhibit the early hardening properties due to the Portland 
cement in its composition (if any), as strongly as the sample 
containing twenty per cent. Portland. In only two tests did 
the " Improved" cement give a higher strength tlmn the clear 
natural. The results of the two-year tests are of interest as 
showing how nearly the same ultimate strength is shown by 
the four samples. The sample of natural cement is of excep- 
tional quality. 

371. Conclusions. It appears from these tests on the effect 
of mixing Portland and natural cements that, in general, the 
full strength of both cements is developed in the mixture; that 
in the early stages of hardening, the mixture sometimes ex- 
hibits more nearly the properties of the Portland, gaining 
strength quite rapidly, but that the ultimate strength of mix- 
tures containing small amounts of Portland are sometimes as 
low as mortars made with natural cement alone. It cannot be 
stated that all samples of Portland and natural cement will 
give as good results in combination as those obtained in the 
above tests, and any extended use of such mixtures should be 
based on full tests of mixtures of the brands that are to be used 
in combination. 

372. Free Lime in Cement. The presence of free lime in 
cement is known to be a serious defect. Table 86 gives the 
results obtained by adding ground quicklime to Portland cement 
in one-to-two mortars. It appears that eight per cent, quick- 
lime reduces the strength at six months about twenty-five per 
cent., and smaller amounts of lime produce approximately 
proportional decrements. The seven-day results, both hot 
and cold, show greater proportional effects. The free lime 
occurring in cements as a result of defects of manufacture is 
likely to be much more dangerous in character than the lime 
used in these tests. 

373. THE USE OF SLAKED LIME WITH CEMENT. A small 
quantity of Portland cement is frequently added to lime mortar 
to hasten the hardening and improve the strength. The ad- 
dition of a small amount of slaked lime to Portland cement 
mortar is also practiced. This not only cheapens the mortar 



260 



CEMENT AND CONCRETE 



TABLE 86 
Mixture of Ground Quicklime with Portland Cement 



BRIQUETS STORED 
IN WATER. 


AGE OF 
BRIQUETS. 


TENSILE STRENGTH OF MORTARS IN 
POUNDS PER SQUARE INCH. 


Lime as Per Cent, of Total Lime and Cement. 





2 


4 


6 


8 


Hot 80 C 
Hot 80 C 
Ordinary tank . . 
Ordinary tank . . 


3 days 
7 days 
7 days 
6 months 


269 
367 
348 
604 


223 

297 
321 
545 


207 
266 
273 
489 


194 
223 
241 
495 


159 
191 
220 
454 



NOTES. Cement: Portland, Brand R, Sample 83 T. 

Lime: Quicklime ground to pass No. 100 sieve (holes .0065 

in. sq.). 
Sand: Standard crushed quartz, 600 grams, to 300 grams of 

cement plus lime. 
Per cent, of lime given replaced the same weight of cement; 

thus: for "4 per cent, lime " the mortar contained 288 grams 

cement, 12 grams lime and 600 grams sand. 
All briquets made by one molder; each result, mean. of five 

briquets. 

but renders it much more plastic, or less "brash," in mason's 
parlance. It is very difficult to lay bricks in a full mortar bed 
with Portland cement mortar containing two or three parts 
sand to one cement, and to use a richer mortar is usually too 
expensive. The work is very much facilitated by mixing a 
little slaked lime paste or powder with the mortar. 

374. The tensile strength of such mixtures is shown by the 
tests in Tables 87 to 89. In the mortars of Table 87 a sample 
of Portland cement is mixed with slaked lime in two forms, 
paste and powder. When the briquets are hardened in open 
air the addition of ten to twenty per cent, of CaO in the form 
of lime paste decreases the strength about twenty-five per cent. ; 
seven per cent, of lime in the form of slaked, dry powder has, 
however, no deleterious effect, and even twenty-eight per cent, 
gives no serious decrease in strength. For water-hardened 
specimens the addition of twenty to thirty per cent, of lime in 
the form of paste appears to increase the strength twenty per 
cent, and no deleterious effect is shown by the addition of 
forty per cent. Also for water-hardened specimens, seven to 



LIME WITH CEMENT 



261 



twenty-eight per cent, of CaO in the form of slaked powder in- 
creases the strength nearly twenty per cent. It thus appears 
that the addition of lime gives better results in mortars that are 
to harden in water, and that for air-hardened mortars lime 
powder should be used in preference to lime paste. Similar 
tests of seven-day briquets showed the lime paste to retard 
the hardening of the mortar. 

TABLE 87 

Slaked Lime in Portland Cement Mortars 









TENSILE STRENGTH, 






PROPORTIONS. 


POUNDS PER 
SQUARE INCH., SAMPLE 








STORED IN 


~f>~K>r> 


LIME IN FORM 






K,EF, 


OF 


















CaO in Lime 












Cement, 
Grams. 


Paste or 
Powder, 


Sand, 
Grams. 


Open Air. 


Water 
Laboratory. 








Grams. 








1 


Paste 


200 





600 


404 


382 


2 




200 


20 


600 


308 


426 


3 




200 


40 


600 


292 


450 


4 




200 


60 


600 


224 


462 


5 




200 


80 


600 


219 


384 


6 


Powder 


200 





600 


382 


371 


7 




200 


14.3 


600 


385 


443 


8 




200 


28.6 


600 


316 


451 


9 




200 


42.8 


600 


338 


431 


10 




200 


57.1 


600 


325 


440 



Cement : Portland, Brand R. Sand : Crushed Quartz, 20-30, or "Standard." 
Age of briquets, 6 months. 

375. In Table 88 only lime paste is used, but both Portland 
and natural cement are tested, and the specimens are hardened 
in dry air and damp sand. In the first column of results 
are given the strengths attained by Portland cement mortar 
containing three parts sand to one of cement without lime. 
In the second column, ten per cent. CaO in form of paste is 
added to the cement. In the third, fourth and fifth columns, 
respectively, ten, twenty-five and fifty per cent, of the cement 
is replaced by CaO. 

It appears that ten per cent, of the cement in a one-to- 
three Portland mortar may be replaced by lime made into paste 
without diminishing the strength, if the mortar hardens in 
damp sand. Even in dry air exposure, it is only at one year 



262 



CEMENT AND CONCRETE 



that the lime shows any deleterious effect. To replace twenty- 
five per cent, or more of the cement with lime, however, dimin- 
ishes the strength of the mortar in a marked degree. 

In the case of natural cement, replacing ten per cent, of the 
cement with lime is decidedly beneficial, and even twenty- 
five per cent, lime gives enhanced strength, except for speci- 
mens hardened in dry air. 

Table 89 gives similar results for one-to-four mortars and 
different percentages of lime, the briquets being hardened in 
dry air and damp sand. 

TABLE 88 

Use of Lime Paste in Cement Mortars Containing Three Parts 

Sand 











TENSILE STRENGTH, LBS. PER SQ. IN. 










Cement, gm. 


200 


200 


180 


150 


100 























FERENCJ 


CEMENT. 


BRIQUETS 
STOKED. 


00 

1 

!Sf 



Lime Paste, " 





60 


60 


150 


300 


CaO in Lime 
Paste, gm. ) 





20 


20 


50 


100 


a 






PQ 


Amt. CaO exO 




















pressed as % 1 
of Cement j 





9 


10 


25 


50 










plus Lime.J 


















Kind. 


Brand. 






Sand, gm. 


600 


600 


600 


600 


600 


1 


Port. 


X 


Dry air 


28 da. 




201 


242 


238 


168 


57 


2 






Damp sand 







294 


330 


309 


238 


95 


3 






Dry air 


3 mo. 




236 


265 


264 


171 


70 


4 






Damp sand 


U 




350 


410 


398 


309 


125 


5 






Dry air 


lyr. 




384 


377 


317 


215 


98 


6 






Damp sand 


(( 




430 


445 


442 


332 


171 


7 


Nat. 


An 


Dry air 


3 mo. 




310 


338 


359 


251 


69 


8 






Damp sand 


(i 




267 


344 


327 


318 


93 


9 






Water 


it 




222 


301 


319 


293 


79 



In all of the above tests the mortars containing much lime 
paste were not only more plastic, but somewhat wetter than 
the corresponding mortars of cement and sand alone, on ac- 
count of the water contained in the paste. 

376. The conclusion to be drawn from these tests appears 
to be that the addition of a small amount, ten to twenty per 
cent., of slaked lime to cement mortars containing as much as 
three parts sand, not only renders them more plastic, but 
actually increases the tensile strength, especially if the mortars 
are kept damp during the hardening. It also appears that for 



PLASTER PARIS WITH CEMENT 



263 



TABLE 89 

Use of Lime Paste in Cement Mortars Containing Pour Farts 
Sand to One Cement 



COMPOSITION OF MORTAR. 


TENSILE STRENGTH OP MORTAR, 
POUNDS PER SQUARE INCH. 


Cement. 


Lime 


Lime 
in 


Sand, 


Stored in Damp 
Sand. 


Stored in Dry 
Air. 


Kind. 


Grams. 


Grams. 


Paste, 
Grams. 


Grams. 


Fresh 
Lime 


Old 
Lime 


Fresh 
Lime 


Old 
Lime 












Paste. 


Paste. 


Paste. 


Paste. 


Portland, C 


240 


00 


00 


960 


176 


180 


254 


244 


Brand X, 1 


240 


80 


27 


960 


212 


200 


280 


250 


Sample 1 


200 


120 


40 


960 


198 


212 


227 


237 


41 S I 


180 


180 


60 


960 


204 


194 


232 


184 


Natural, 

BrandAn, -j 
Sample L [^ 


240 

240 
200 
180 


00 
80 
120 

180 


00 
27 
40 
60 


960 
960 
960 
960 


150 
160 
166 
140 


133 
154 
173 

166 


127 
162 
131 
124 


142 

150 
170 
154 



NOTE. All briquets three months old when broken. 

mortars exposed to the open air the lime should be in the form 
of slaked powder rather than paste. It may be added, that in 
all cases care should be taken that the lime is thoroughly slaked 
before use, and all lumps should be removed by straining or 
sifting. Further results on this subject are given in connection 
with the tests on adhesion of cement mortar to brick (Art. 5). 
377. EFFECT OF PLASTER OF PARIS ON THE COHESIVE 
STRENGTH OF MORTARS. 

The use of plaster of Paris, or calcium sulphate, in the man- 
ufacture of cement to regulate the time of setting, has already 
been mentioned. The amount of such additions at the factory 
are usually small, the German Cement Makers' Association limit- 
ing it to two per cent. 

Tests on three brands of Portland cement, showing the effect 
of small additions of plaster Paris, are given in Table 90. All 
of these mortars hardened in water. It is not known whether 
any of the cements had received additions of plaster Paris be- 
fore leaving the factory. It is probable that brands R and X 
had been so treated, since they are German cements, but it is 
not probable that the other brands of Portland had received 
any addition of plaster. 

It appears that with these brands the addition of from one 



264 



CEMENT AND CONCRETE 



to three per cent, of plaster Paris hastens the hardening and 
increases the strength of the mortar at ages of six months to 
two years. Six per cent, plaster sensibly retards the hardening, 
but, in all cases except one, Brand S, neat, six months, the 
mortars containing six per cent, plaster, gave higher results 
on long time tests than did the corresponding mortars to which 
no plaster had been added. 

TABLE 90 

Plaster of Paris in Portland Cement Mortars, Hardening in "Water 





, 






to 




ri 


w ^ 


H H 




B 


TENSILE STRENGTH, POUNDS PER SQ. 




52 


il 


TEMPERA- 
TURE WATER 

IN WHICH 


S S 


IN., WITH PER CENT. OF CEMENT 
REPLACED BY PLASTER 
OF PARIS. 


E 


w 


"H 


BRIQUETS 


?|g 




S 


i.3 


E ^ 


STORED. 














w 


g j 


02 




1 





l 


2 


3 


6 


! 


s 


o - 


60to65Fahr. 


7 da. 


487 


626 


600 


519 


380 


2 









6 mos. 


743 


746 


754 


742 


660 


3 




2 




7 da. 


323 


388 


360 


289 


182 


4 




2 




6 mos. 


492 


530 


547 


607 


663 


5 




2 




lyr. 


487 


515 


610 


588 


647 


6 




2 




2yrs. 


533 


586 


612 


659 


684 


7 


;, 







7 da. 


562 


608 


726 


709 


432 


8 









6 mos. 


745 


751 


799 


804 


795 


9 




2 




7 da. 


288 


347 


372 


352 


165 


10 




2 




6 mos. 


532 


538 


624 


638 


642 


11 




2 




lyr. 


591 


595 


643 


645 


666 


12 




2 




2yrs. 


590 


623 


680 


673 


666 


13 


X 







7 da. 


351 


368 


405 


450 


204 


14 









6 mos. 


560 


606 


580 


645 


797 


15 




2 




7 da. 


227 


258 


261 


282 


96 


16 




2 




6 mos. 


494 


546 


591 


574 


563 


17 




2 




lyr. 


572 


580 


586 


583 


652 


18 


^ 


2 




2 yrs. 


592 


575 


592 


592 


667 


19 


s 


2 


176 Fahr. 


5 da. 


296 


307 


362 


391 


422 


20 


R 


2 


140 " 


5 da. 


403 


440 


416 


495 


442 


21 


X 


2 


140 " 


5 da. 


361 


334 


390 


452 


474 



NOTES. Sand, Point aux Pins (river sand) passing No. 10 sieve, except 
for hot tests, where standard sand was used. Cement and 
plaster of Paris passed through No. 50 sieve before using. 
Plaster Paris had no apparent effect on consistency mor- 
tar at first, but after making first three briquets of batch 
of five, the mortar containing plaster Paris dried out 
somewhat. 
Each result, mean of five briquets. 

Similar tests of natural cement mortars hardening in water 
.are given in Table 91. One of the brands is not much affected 



PLASTER PARIS WITH CEMENT 



265 



TABLE 91 
Plaster of Paris in Natural Cement Mortars, Hardening in Water 













TENSILE STRENGTH, POUNDS PER 




CEMENT, 


SAND, 


TEMPER- 


AGE OF 


SQUARE INCH, WITH PER CENT . 
OF CEMENT REPLACED BY 


REF. 


NATURAL 
BRAND. 


PARTS TO 
ONE 
CEMENT. 


ATURE 
WATER 
WHERE 
STORED. 


BRIQUETS 
WHEN 
BROKEN. 


PLASTER OF PARIS. 

























1 


2 


3 


6 








Degrees F. 














1 


An 





60-65 


7 da. 


233 


225 


213 


235 


a 


2 









6 mo. 


422 


449 


438 


441 


324 


3 




2 




7 da. 


111 


109 


97 


144 


a 


4 




2 




6 mo. 


418 


416 


435 


409 


133c 


5 




2 




1 yr. 


415 


451 


430 


454 




6 




2 




2 yrs. 


478 


476 


489 


514 




7 


Gi 







7 da. 


146 


156 


115c 


a 


a 


8 









6 mo. 


383 


3986 


323 


312e 


234/ 


9 




2 




7 da. 


62 


80 


94 


a 


a 


10 




2 




6 mo. 


374 


312 


355 


86/ 


151/ 


11 




2 




1 yr. 


448 


395 


408 


131/ 


107/ 


12 




2 




2 yrs. 


456 


437 


397 


172/ 


a 


13 


An 


2 


140 


5 da. 


319 


365 


405 


402 


203 


14 


Gii 


2 


" 


' ' 


359 


351 


189 


138 


100 



NOTE. Sand, Point aux Pins (river sand) passing No. 10 sieve, ex- 
cept for hot tests, where standard sand was used. 
a Found badly swelled and nearly disintegrated after a few 

days in tank. 

b Surface cracks, 1 inch section swelled to 1 ^ inches. 
c Surface cracks, 1 inch section swelled to 1 -fa inches. Had 

nearly disintegrated after 2 days. 
d Surface cracks. 
e Badly cracked on surface. 
/ Badly cracked on surface, and 1 inch section swelled to 

about l^s inches. 

by additions of one to three per cent., but the other brand is 
practically ruined by the addition of more than one or two per 
cent., and both brands are rendered quite unsound by six 
per cent, plaster. 

378. The briquets reported in the preceding tables were 
hardened in water, as usual. Table 92 gives some of the results 
obtained by adding plaster Paris to mortars that are hardened 
in dry air. The effects on the two samples of the same brand 
of Portland, one quick setting and one slow setting, are quite 
different. The strength of the quick setting sample is increased, 
two per cent, giving the best results, while that of the slow 



266 



CEMENT AND CONCRETE 



etting sample is diminished by the addition of plaster. Both 
brands of natural cement appear to be notably improved by 
the plaster, the best result being given by three per cent, Such 
an addition to one brand results in a remarkable increase in 
strength of 250 per cent. 

TABLE 92 

Plaster of Paris in Cement Mortars, Hardening in Dry Air. Effect 
on Different Samples, Portland and Natural 









TENSILE STRENGTH POUNDS PER 




CEMENT. 


OB 

g 

a s 


SQUARE INCH, WITH 
PER CENT. OF CEMENT REPLACED BY 


REF. 




o* 
^5 


PLASTER OF PARIS. 




Kind. 


Brand. 


Sample. 








i 


2 


3 


6 


1 


Port. 


R 


26 R 


6 mo. 


443 


443 


560 


529 


493 


2 


tc 


R 


23 R 


t! 


559 


483 


419 


436 


337 


3 


Nat. 


An 


L 


(I 


162 


220 


282 


286 


272 


4 


ti 


In 


28 S 


|| 


76 


110 


151 


269 


240 



NOTES. Sample 26 R, Portland, quick setting, bears T V inch wire in 18 

minutes. 
Sample 23 R, Portland, slow setting, bears T V inch wire in 244 

minutes. 

Sand, two parts Point aux Pins (river sand) to one cement. 
All briquets stored in air of laboratory until broken. 
Each result, mean of five briquets. 

For the effect of plaster of Paris on the adhesive strength of 
mortar, see 407. 

379. Conclusions. It is evident from the above tests that 
the addition of small amounts of plaster Paris affects different 
samples of cement in quite different ways, and it is necessary 
to bear this in mind in the application of general conclusions 
to special cases. The indications are that the addition to 
cement of from one to three per cent, of plaster of Paris or 
sulphate of lime generally hastens the hardening and will not 
usually result in decreased strength; that some natural cements, 
however, are sensibly injured by more than one per cent., 
especially if used neat. The presence of as much as six per 
cent, plaster of Paris retards the hardening (although hastening 
the initial set) and is quite apt to ruin either Portland or natural 
cements. The addition of plaster Paris usually gives better 
results in air hardened than in water hardened specimens. 



CLAY WITH CEMENT 267 

ART. 49. MIXTURES of 1 CLAY AND OTHER MATERIALS WITH 

CEMENT 

380. EFFECT OF CLAY ON CEMENT MORTAR AND CONCRETE. 

Clay may occur in cement mortar or concrete due to the 
use of sand or aggregate that is not clean. As the plasticity of 
cement mortar is increased by the presence of clay, small 
amounts are sometimes added to produce this effect, and clay 
is also sometimes used to render mortar stiff enough to with- 
stand immediate immersion in water. In the case of concrete, 
the presence of a certain percentage of clay renders it easier to 
compact the mass by tamping, though if too much clay is pres- 
ent, the mass becomes sticky. 

A number of tests have been made to determine the behavior 
of such mixtures of clay and cement. In all of these tests the 
clay was first dried, pulverized and sifted, and then a weighed 
quantity equal to a given per cent, of the weight of the cement 
was added to the latter. In the writer's first tests of this kind 
small percentages of clay were used, less than ten per cent., but 
it was found that with lean mortars much larger percentages must 
be used to determine the point where clay began to be injurious. 

381. Table 93 shows the effect of clay on the time of setting 
and soundness of neat cement. The effect of small percentages 
of clay on the time of setting of Portland cement is not very 
marked, but with natural cement even ten per cent, of clay 
retards the setting in a marked degree. As to the effect on 
soundness, Portland cement pats disintegrate with more than 
twenty-five per cent, of clay added, while the natural cement 
is affected if more than ten per cent, of clay is present. 

382. Table 94 shows the tensile strength of neat cement 
mortars to which clay to the amount of 10 to 100 per cent, of 
the cement has been added. Some of the Portland briquets 
were immersed as soon as molded, while others were left the 
customary twenty-four hours in moist air before immersion. 

It is seen that to mix clay with neat Portland cement results 
in a decided decrease in strength, the results obtained with 
twenty-five per cent, clay being only about sixty or seventy 
per cent, of the strength of the mortar without clay. With 
natural cement the presence of clay seriously retards the hard- 
ening and results in decreased strength, though it does not 



268 



CEMENT AND CONCRETE 



TABLE 93 

Effect of Pulverized Clay on the Time of Setting and Soundness 

of Cement 









TIME TO BEAR jV INCH WIRE IN MINUTES, 








AND THE CONDITION 


1 




CLAY. 


OF PATS AFTER FIVE MONTHS. 




CEMENT. 






1 






Clay as Per Cent, of Cement. 


B 













Kind. 


Brand. 


Sam- 
ple. 


Kind. 





10 


25 


50 


100 


i 


Portland 


X 


41S 


Red 


285 


318 


328 


328 


450 


i 


i 




i 





Good 


Fair 


Good 


Bad a 


Bad a 


2 


i 




i 


Blue 


288 


286 


300 


305 


306 


2 


i 




i 


it 


Good 


Fair 


Good 


Bad 


Bad a 


3 


Natural 


Gn 


KK 


Red 


69 


123 


195 


345 


445 


3 


' 




* 


it 


Fair 


Fair 


Bad 


Bad 


Bad a 


4 


4 







Blue 


98 


173 


215 


350 


415 


4 


' 




i 


ii 


Poor 


Poor 


Bad 


Bad 


Bad a 























NOTE. Results marked a, pats cracked badly in air and were not im- 
mersed. 

TABLE 94 
Effect of Clay on Tensile Strength; Neat Cement Paste 









Alt 




TENSILE STRENGTH, LBS. PER 








B 5 

O O 




SQUARE INCH. 




CEMENT. 


KIND 


a 5 S 


J 






REF. 




OF 

CLAY. 


jjl 


IB 


Clay Expressed as Per Cent, of 
Cement. 










H 










V ~ 


cq 






Kind. 


Brand. 


Sample. 




Sag 

a H sj 







10 


25 


50 


100 


1 


Port. 


X 


41S 


Red 


24 


3 mo. 


6c8 


535 


474 


336 


253 


2 


u 


it 


l( 


*' 





(4 


660 


587 


476 


318 


255 


3 


Nat. 


An 


D 


" 


24 


28 da. 


389 


280 


138 


60 


22 


4 


it 


An 


D 


II 


24 


3 ino. 


376 


365 


323 


219 


176 



have as deleterious an effect as it does with Portland. The mix- 
ing of clay with neat cement is of course very severe treatment. 

In Table 95 the mortars contain equal parts cement and sand, 
and the clay is from 50 per cent, to 200 per cent, of the weight 
of cement. It appears from this table that clay in as large 
amounts as 50 per cent, of the cement is injurious to one-to- 
one mortars of either Portland or natural cement. 

383. The mortars in Table 96 are all of Portland, and con- 
tain three parts sand to one cement. Smaller percentages ' of 



CLAY WITH CEMENT 



269 



TABLE 95 

Effect of Large Amounts of Clay in Mortars Containing Equal 
Parts Cement and Sand 













TENSILE STRENGTH, LBS. 


8 


CEMENT. 




HOURS 




PER SQUARE INCH. 


g 




KIND 


ELAPSED 


i 






w 
i 
p 
K 




OF 

CLAY. 


BETWEEN 

MOLDING 
AND IM- 


|o| 



CLAY EXPRESSED AS PER 
CENT. OF CEMENT. 








M 

Mi 










MERSING. 


PQ 




Hq 


Kind. 


Brand. 


Sample. 



































50 


100 


150 


200 


1 


Port. 


X 


41 S 


Red 


24 


3 mos. 


747 


512 


337 


239 


193 


2 


ti 


" 


" 







ii 


720 


549 


321 


242 


189 


3 


Nat. 


Gn 


KK 




24 


3 mos. 


454 


241 


212 


183 


140 


4 


M 


ti 


" 







u 


413 


231 


206 


157 


128 


5 


ii 


An 


D 




24 


II 


442 


259 


194 


167 


152 


6 


u 


it 


it 







II 


440 


274 


184 


141 


125 


7 


I( 


ii 


u 




24 


6 mos. 


488 


335 


268 


217 


184 



clay are used, namely, 10 to 40 per cent. The mortars harden- 
ing in water show a decided improvement due to the presence 
of clay, but the briquets hardening in the open air indicate that 

TABLE 96 

Effect of Clay in Portland Cement Mortar Containing Three Parts 
Sand to One Cement 



REFERENCE. 


BRIQUETS STORED. 


AGE OF 
BRIQUETS. 


TENSILE STRENGTH, LBS. PER SQUARE 
INCH. 


CLAY ADDED AS PER CENT. OF CEMENT. 





10 


20 


40 


1 


Tank, Laboratory 


6 months. 


385 


435 


489 


533 


2 


u it 


2 years. 


375 


412 


478 


593 


3 


Open Air 


6 months. 


381 


403 


394 


418 


4 


it U 


2 years. 


660 


624 


631 


570 



NOTES. Cement, Portland, Brand R, Sample 83 T. 

Sand, three parts crushed quartz f $ to one cement by weight. 
Clay, red clay dried, pulverized, and passed through No. 100 

sieve. 
Clay added to mortar, amount cement and sand remaining 

constant. 



270 



CEMENT AND CONCRETE 



at two years the mortar without clay is stronger. It may be 
noted in passing that these results, obtained at two years, with 
one-to-three mortars hardened in open air, are very high. 

The effect of clay on mortars containing four parts sand to 
one cement is shown in Table 97. In this case the addition of 
clay equal to the weight of the cement almost invariably re- 
sults in increasing the strength of the mortar. Briquets im- 
mersed as soon as made were especially benefited by the pres- 
ence of clay, except in one case, the red clay did not appear to 
increase the ability of the natural cement Gn to withstand 
early immersion. The red clay appears to give better results 
than the blue with Portland, while the reverse is true with at 
least one brand of natural. Whether this difference is a chemi- 
cal or physical one is not known; the red clay is a good pud- 
dling clay, while the blue clay is not, but appears to contain 
some very fine sand. 

TABLE 97 

Effect of Large Amounts of Clay in Cement Mortars Containing 
Four Parts Sand to One Cement 













TENSILE STRENGTH, LBS. 





CEMENT. 




HOURS 




PER SQUARE INCH. 


1 




KIND 


ELAPSED 


, i 






H 
1 




OF 

CLAY. 


BETWEEN 

MOLDING 
AND IM- 


o& 
<2 

(6 


CLAY AS PER CEKT. OF 
CEMENT. 








PH 










MERSING. 


PQ 






Kind. 


Brand. 


Sample. 



































50 


100 


150 


200 


1 


Port. 


X 


41 S 


Red 


24 


3 mos. 


271 


348 


305 


239 


193 


2 








Blue 


24 




227 


304 


250 


179 


145 


3 








Red 


00 




156a 


320 


324 


200 


192 


4 








Blue 


00 




149a 


270 


215 


148 


114 


5 


Nat. 


Gu 


KK 


Red 


24 


3 mos. 


138 


155 


146 


164 


133 











Blue 


24 




118 


167 


200 


167 


134 


7 








Red 


00 




83 


87 


39 


86 


72 


s 








Blue 


00 




49 


127 


147 


136 


106 


9 








Red 


24 


2yrs. 


194 


348 


306 


256 


190 


10 




Au 


D 


Red 


24 


3 mos. 


138 


218 


174 


174 


190 



NOTES. Sand, crushed quartz fg, ("Standard"), four parts to one 

cement by weight. 

Clay, dried, pulverized and passed through sieve before using. 
All briquets immersed in tank in laboratory as usual. 
Each result, mean of five briquets. 

Results marked "a," briquets disintegrated some on face from 
early immersion. 



CLAY WITH CEMENT 



271 



384. Table 98 gives the results of tests by other experimenters, 
showing the effect of clay on one-to-three mortars of Portland 
And natural cement. 1 The amount of clay used in these tests 
.appears to be stated as percentage of the total ingredients in- 
stead of as a percentage of the cement as in the preceding 
tables. The mortars were mixed quite dry for these experi- 
ments. The Portland cement mortar seems to be improved by 
the addition of clay to the amount of twelve per cent, of the 
mortar. The hardening of natural cement mortar is some- 
what slower with twelve per cent, clay than with three to six 
per cent., but at the age of twelve weeks the mortars containing 
clay were all stronger than that without clay. 

TABLE 98 

Effect of Clay on the Tensile Strength of One-to-Three Mortars 









TENSILE STRENGTH, POUNDS 


CEMENT. 


PARTS 
SAND TO 


AGE OF 
BRIQUETS 


PER SQUARE INCH. CLAY EXPRESSED AS 
PER CENT. OF MORTAR. 




CEMENT. 


BROKEN. 





3 


6 


9 


12 


Portland 


3 


2 weeks 


202 


267 


280 


318 


333 


K 


3 


4 weeks 


362 


301 


334 


381 


353 


ti 


3 


12 weeks 


451 


506 


521 


522 


547 


Natural 


2 


1 week 


68 


117 


101 


99 


65 


ii 


2 


4 weeks 


152 


199 


219 


170 


146 


" 


2 


12 weeks 


170 


214 


252 


230 


211 



NOTE. Tests by Messrs. J. J. Richey and B. H. Prater. 

385. Conclusions. Always keeping in mind the limitations 
to be observed in drawing general conclusions from experi- 
ments having a limited range, it may be said that the indications 
are as follows: Neat cement and rich mortars are injured by the 
addition of clay, the rate of hardening and the ultimate strength 
being diminished. Lean mortars containing three to four parts 
sand to one cement are usually improved by the addition of 
clay to the amount of 40 to 100 per cent, of the cement, or 
10 to 25 per cent, of the combined weight of cement and sand, 
and the ability of such mortars to withstand early immersion 
may be greatly enhanced by such additions. It is evident 
from the above tests that the expense which should be incurred 
in washing sand to remove a small percentage of clay is limited, 



1 Messrs. J. J. Richey and B. H. Prater, Technograph, 1902-3. 



272 



CEMENT AND CONCRETE 



and for certain uses there is no question that mortar may be 
improved by the addition of clay. 

(For the effect of clay on the compressive strength of con- 
crete, see Art. 55.) 

386. Powdered Limestone, Brick, etc. Various foreign sub- 
stances are sometimes used with cement, either in lieu of sand ; 
or to make the mortar more plastic. Such foreign ingredients 
may also occur in mortar as impurities in the sand used. Pow- 
dered limestone, slaked lime, powdered brick and clay are some 
of the materials experimented with in this connection. A few 
tests of the effects of such mixtures on the setting time of ce- 

TABLE 99 
Foreign Substances in Cement Mortar 



. 


CEMENT. 


Bl. 




TENSILE STRENGTH, POUNDS PER 
SQUARE INCH. 


fc 

W 

c 


. 






H H 


AGE OP 
BRIQUETS 
WHEN 


Composition of Mortar. 


1 


Kind. 


Brand. 


Sam- 
pie. 





BROKEN. 
















& 








*B 




a 


b 


c 


d 


e 


f 


l 


Port. 


R 


JJ 


None 


3 months 


705 




674 


583 


615 


667 


2 








3.75 


5 days, H 


152 


217 


164 


175 


240 


198 


3 








3.75 


3 months 


259 


367 


297 


284 


311 


304 


4 








3.75 


1 year 


309 


365 


367 


333 


438 


438 


5 


Nat. 


An 


G 


None 


3 months 


286 




203 


307 


154 


203 


6 








4 


5 days, h 


86 




105 


94 


132 


164 


7 








4 


3 months 


185 




214 


157 


239 


215 


8 








4 


1 year 


210 


... 


234 


238 


263 


264 



NOTES. Sand, "Standard." Materials added to mortar were first pul- 
verized and passed through No. 80 sieve, holes .007 inch 
square. 

j 5-day results, H = immersed in hot water, 80 C. 
t 5-day results, h = immersed in hot water, 60 C. 
Composition of mortars: 

a No foreign substance. 

b No foreign substance, but additional amount cement added, 

making mortar 1 to 3 instead of Ito 3.75. 
c Kelleys Isd. Limestone, equal to 25 per cent, weight of ce- 
ment added to mortar. 
d Slaked lime powder, equal to 25 per cent, weight of cement 

added to mortar. 
e Red clay, equal to 25 per cent, weight of cement added to 

mortar. 

/ Red brick, equal to 25 per cent, weight of cement added 
to mortar. 



FOREIGN SUBSTANCES WITH CEMENT 



273 



merit indicated that the rate of setting of Portland cement was 
not appreciably affected by the addition of twenty-five per 
cent, of any of these substances, but the setting time of natural 
cement appeared to be sensibly hastened by such additions. 
None of these materials had any appreciable effect on the 
soundness of either Portland or natural. 

Table 99 shows the effect on the tensile strength of mortar 
of adding twenty-five per cent, of each of the four substances 
mentioned. It appears that the strength of neat cement mor- 
tar, either Portland or natural, is usually diminished by the 
presence of such materials, but in almost every case mortars 
containing about four parts sand to one cement are improved 
by the addition of the substances in question to an amount 
equal to twenty-five per cent, of the cement. Pulverized clay 
and brick give the best results, the increased strength amounting 
to from twenty to forty per cent. 

387. Sawdust. Where a very light and porous mortar is 
desired for use in floors and similar purposes, the incorporation 
of sawdust in the mortar is suggested by a similar use in clay 
building materials. The results in Table 100 show that the 
use of sufficient sawdust to materially diminish the weight, 
practically ruins the cohesion of the mortar, even ten per cent, 
of sawdust materially diminishing the strength. 

TABLE 100 

Sawdust in Cement Mortar 



t 

I 

w 

ti 
1 

2 
3 
4 
6 
6 


CEMENT. 


PARTS SAND TO 
ONE CEMENT. 


BRIQUETS 
STORED. 


AGE OF 
BRIQUETS WHEN 
BROKEN. 


TENSILE STRENGTH POUNDS PER 
SQUARE INCH. 


Kind. 

Port, 
it 

a 
n 

Nat. 

4( 


Brand. 


Sawdust as Per Cent, of Cement. 





10 


20 


25 


50 


100 


X 

M 

U 
it 

An 
(i 





2 
2 

2 


Tank 
Dry air 
Tank 
Dry air 
Tank 
Tank 


lyr. 


799 
674 
502 
452 
433 
313 


409 
492 


. . . 


169 
103 


44 
28 
32 
14 
38 
58 


31 
a 

32 
a 
b 

20 


253 


129 
108 


104 



NOTES. Sand, crushed quartz, f. Sawdust from white pine, passed 

through sieve with one-quarter inch meshes. 
a Briquets broken in applying initial strain. 
b Briquets disintegrated in tank. 



274 



CEMENT AND CONCRETE 



388. Use of Ground Terra Cotta as Sand. A light weight 
mortar may also be made by using as sand or aggregate, ma- 
terials of burned clay, such as brick or terra cotta. The tests 
in Table 101 were made to determine the value of ground terra 
cotta for use in place of sand, and it appears that this material 
gives excellent results. The strength given with one of the 
brands of natural cement is especially high. 

TABLE 101 
Use of Ground Terra Cotta as Sand in Cement Mortar 



REFERENCE. 


CEMENT. 


AGE OF 
BRIQUETS. 


TENSILE STRENGTH, LBS. PER SQ. IN. 


Parts Ground Terra Cotta to One Cement, 
by Weight. 


Kind. 


Brand. 


1 


2 


3 


4 


6 


1 


Portland 


X 


3 months 


523 


406 


332 


257 


174 


2 
3 


Natural 


It 

An 


1 year 
3 months 


604 

284 


518 
338 


429 
346 


337 
347 


266 
224 


4 

5 


u 


En 


1 year 
3 months 


262 
291 


360 
303 


351 
184 


361 
136 


186 


6 


" 


" 


1 year 


340 


434 


284 


161 


. . . 



NOTES. Terra Cotta tile, of medium burn, ground and passed through 
No. 20 sieve, and used in place of sand. 

ART. 50. THE USE OF CEMENT MORTARS IN FREEZING 
WEATHER 

389. It is frequently desirable to use cement in freezing 
weather, but to ensure good work under these circumstances it 
is necessary to take certain precautions. If mortar is frozen 
immediately after mixing, setting cannot take place until it 
has again thawed. In the practical use of cement it is always 
gaged with a larger quantity of water than is required for the 
chemical combination, and if this excess water is frozen after 
the setting is somewhat progressed, the consequent expansion 
may be sufficient to disrupt the partially set mortar. By warm- 
ing the materials or by lowering the freezing point of the water 
by the addition of salt, glycerine, or some other substance hav- 
ing this effect, it is sought to prevent the mortar freezing until 
the work is protected by another layer of mortar, or otherwise, 
and thus to avoid the expansion. Salt is generally used much 



EXPOSURE TO FROST 275 



too sparingly to prevent freezing. The freezing point of water 
is lowered about one and a half degrees Fahr. for each per cent, 
of common salt added; thus a twenty per cent, solution would 
freeze at about two degrees Fahr. 

390. The following tests are selected as showing typical re- 
sults of a large number of experiments made under the author's 
direction to determine the effect of exposing cement mortars to 
frost, and to indicate what treatment will alleviate the delete- 
rious effects of low temperature. In making tests with small 
specimens, it is difficult to approach the conditions existing in 
the actual use of mortars in freezing weather. A small mass 
of mortar exposed to the air on all sides sets more quickly than 
the interior of a large mass; and on the other hand, the effect 
of frost on a small specimen must be more severe and more 
quickly apparent. Many of the results are more or less contra- 
dictory, and the conclusions that have been drawn are such as 
appear to be indicated by the majority of the tests. The 
treatment of the briquets, and the conditions existing, are given 
in some detail, that the limits of applicability of such conclu- 
sions may be seen. 

391. Exposure to Frost of Mortars Already Set. In the 
tests recorded in Tables 102 to 104 the briquets were allowed to 
remain one or two days in the laboratory. It is evident that 
these results are of but limited practical importance, since it 
is seldom that mortars which are made in winter can be allowed 
to set in a warm place before exposure; they are given, how- 
ever, for what they are worth. Tables 102 and 103 give the 
results obtained with Portland cement briquets exposed to a 
severe temperature twenty-four to forty-eight hours after made. 
The most important deduction, and the one most clearly indi- 
cated by these tables, is that Portland cement mortar made with 
fresh water may be subjected to very low temperatures twenty- 
four to forty-eight hours after molded, without seriously de- 
creasing the tensile strength given at six months to two years. 
It also appears that solutions containing as much as fifteen 
per cent, salt are deleterious, and smaller percentages are not 
advantageous under these conditions. 

Table 104, giving the results of similar tests with natural 
cement mortar, indicates that this brand gives good results if 
allowed to set in warm air before exposure to frost. Solutions 



276 



CEMENT AND CONCRETE 



TABLE 102 

Exposure of Portland Cement Mortars to Low Temperatures after 
Twenty-four Hours in Laboratory 







AGE 


TENSILE STRENGTH, POUNDS PER SQ. IN. 




DATE 








MADE. 


BROKEN. 


a 


b 


c 


d 


e 


/ 


9 


Standard . . . 


1-15 


6 mo. 


772 


960 


816 






524 


507 


Standard . . . 


1-15 


21 mo. 


796 


882 


766 






443 


642 


Ft. aux Pins, ) 


1-18 


6 mo. 


651 


630 


. 


769 


463 


443 




pass, sieve ^10 J 


1-18 


21 mo. 


760 


780 





711 


543 


447 





NOTES. Cement, Portland, Brand R. One part sand to one cement. 

Briquets made in laboratory, temp., 64 to 66 Fahr. ; materials 

about 65 Fahr. 

Temperature, open air, Jan. 16 to Jan. 19, 4 to 15 Fahr. 
Treatment of briquets: 

a Fresh water, briquets stored in water in laboratory. 
b Fresh water, briquets stored in open air after 24 hours. 
c Fresh water, briquets alternated, two days in open air 
and then two days in air laboratory, for fifty-two 
days, then left in open air. 

d Water 5 per cent, salt; briquets stored in open air. 
e Water 15 per cent salt; briquets stored in open air. 
/ Water 25 per cent, salt; briquets stored in open air. 
g Water 25 per cent, salt; briquets stored in water in lab. 

TABLE 103 

Exposure of Portland Cement Mortars to Low Temperatures, 
Twenty-four to Forty-eight Hours after Made 



SAND, 
KIND. 


DATE 
BRIQUET 
MADE. 


AGE 
WHEN 
BROKEN. 


TENSILE STRENGTH, POUNDS PER SQ. IN. 


a 


b 


c 


d 


e 


/ 


g 


h 


i 


J 


Std. . 
Std. . 

P. P. 
P.P. 


1-16 
1-16 
1-19 
1-19 


6 mo. 
21 mo. 
6 mo 


415 
602 


372 
372 


401 

438 


262 

384 


202 
326 






















381 
638 


394 
430 


360 
418 


371 
375 


233 
344 


21 mo. 

























NOTES. Cement, Portland, Brand R. Three parts sand to one cement. 

Briquets made in laboratory. Temp . : Air and materials, 64 to 
67 Fahr. Open air, Jan. 10 to 20, -15 to +18 Fahr. 

Treatment of briquets: a, b, c, d and e mixed with water con- 
taining 0, 10, 15, 20 and 25 per cent, salt, respectively; 
a to d, inclusive, air laboratory 24 hours, water laboratory 
16 hours, air laboratory 12 hours, then in open air. 

/, g, h, i and /, mixed with water containing 0, 10, 15, 20 and 
25 per cent, salt, respectively. 

c to /, inclusive, put in open air after about 24 hours in 
air of laboratory. 



EXPOSURE TO FROST 



277 



TABLE 104 

Exposure of Natural Cement Mortars to Low Temperatures, 
Twenty-four Hours after Made 



PARTS 
SAND TO 
ONE 


DATE MADE. 


AGE WHEN 
BROKEN. 


TENSILE STRENGTH, POUNDS PER SQ. IN. 














CEMENT. 






a 


b 


c 


d 


e 


/ 


2 


1-20 


6 mo. 


297 


404 


319 




297 


176 


2 


1-20 


lyr. 


305 


390 


343 


. . . 


273 


217 


4 


1-20 


6 mo. 


222 


318 


319 


344 




114 


4 


1-20 


1 yr. 


223 


259 


339 


205 




150 



NOTES. Cement, Natural, Brand Gn. Sand, " Ft. aux. Pins" (river 
sand). 

Temp, materials and air of laboratory where briquets were 
molded, 65 to 68 Fahr. Temp, open air Jan. 21 to 23, 
= - 1 to +29. 

Treatment of briquets: a, briquets stored in water in labora- 
tory, b to /, inclusive, briquets stored in open air after 
twenty-four hours in air of laboratory. 

a and 6, fresh water used for gaging mortar. 

c, d, e and /, water used in gaging had 5, 10, 15 and 25 per 
cent, salt, respectively. 

containing more than ten per cent, salt are deleterious for such 
treatment. Briquets of another brand of natural cement, a 
one-to-one mortar of which gave about 450 pounds tensile 
strength at one year, failed entirely when placed, one hour 
after made, in open air for three days, and then immersed in a 
tank in the laboratory. A 7.4 per cent, solution of salt used 
for gaging assisted very materially in preserving the mortar 
under the same severe treatment, although this amount of salt 
was not sufficient to lower the freezing point of the water below 
the temperature to which the briquets were subjected. 

392. Effect of Salt in Mortars Hardened in Water and Air. 
In the tests recorded in Table 105 the materials used were at a 
temperature of forty degrees Fahr., and the briquets were 
molded in an open warehouse where the temperature was usually 
below twenty-three degrees Fahr., though for a few of the 
tests the temperature of the air at time of molding was as high 
as twenty-seven degrees. The temperature of the mortar 
when briquets were finished was usually but little above thirty- 
two degrees Fahr. The briquets were left in a warehouse for 
three days, when part of them were immersed in cold water 



278 



CEMENT AND CONCRETE 



I 

.s 
si 

H O 



Cent. Salt in Water Used in 
Gaging. 



lt in Water Used 
Gaging. 



O .00 .O 
1-1 OS i i 



.iO .O . <M 



. >O .1^ .00 

COQOt 

! IQ ' O ' ^ 



. 
<M 

" CO 



OS . O 
t _ CO 

^ " GO 



. r-i .CO .O 
_ i 1 i I O 
' CO " l>- " CO 



CO .O .CO . O .CO .CO . 
t-(Mr-IOOCOO5 
Tfl 'CO 'CO 'O 'CO 'O 



1-1 .CD . rt< 
OIOOS 
O 'GO ' > 



.OS .CO . 

GOi I 
'O 'CO ' 



.GO .OS 
C*4l 

*tO ^jl 



' lO ' TP O 



. . 

GOGst- 



T-H .co .cs 

cocoO 

' 



CO . * . i i 
<Mi-HOS 



CO '^ 



. <M .1-1 

>-OO 
* -^ ' Tfi 



Naxoug 



any 



aavjv 



I* 



3XVQ 



TURES, 
FAHR 



s 



t-OOr-COCOOSOOOSCO(MCO<MOSOOS 



^ilr^cOt^OOOOOOOOCDCOCO 
COCOCOCOCOTt<COTtHCOCOCOCO 

1 1 1 1 1 1 1 1 1 I I 1 
(N'NfMCNtNCDOJCOTtl^^fN 

co co co co co co co co co co co co 



OSt OSt^COCOOOCOCOt 

T^T^ 

iaos 



222223222322 

Tp ^ 



CO^)COCOOOt-OOt-OS OS 
<??-r<Ni-trH(N^I<Ni-lOiT-lOS 



- -- 



- --- o- -. - PS- 

- - - --- 3 *S 3 S 3 



EXPOSURE TO FROST 



279 



(under ice), and the remainder stored in open air on a shelf 
covered by a rough board roof, but with front left open to the 
weather. All mortars contained two parts river sand to one of 
cement by weight. The water used in gaging varied from 
tresh to a twenty-five per cent, solution. 

The results indicate that Portland mortars made in low 
temperatures, to be immersed in cold water, are improved by 
fifteen to twenty per cent, salt in the water of gaging, but that 
more than five per cent, salt is deleterious for mortars exposed 
to the air only. The very high results given by the air-hardened 
specimens are worthy of notice. 

A similar series of tests of natural cement gave results from 
which no definite general conclusions could be drawn. The 
effect of freezing and of the use of salt varied greatly for dif- 
ferent samples. For any given sample the treatment, as re- 
gards the use of salt, giving good results in open air, was usually 
the reverse of that giving good results in cold water. The 
conclusions indicated for rich mortars were sometimes the re- 
verse of those shown by lean mortars. 

393. The results obtained with five brands of natural ce- 

TABLE 106 
Effect of Low Temperatures on Five Brands of Natural Cement 













S 


__. 


v 










Q 


2 W 


4 


S 


a Z 




t 


W 
Q 


K 


S5 

4 


|J 


00 K 


1 


5 i S 


MEAN TENSILE STRENGTH, 


EFEREl 


1 


H 

| 


OQ 

1 


1 1 


ijj 


H 


|| 


BRAND. 



















a, 


H | 


K M 
i 




1 


I 


Gn. 


An. 


Kn. 


Hn. 


Jn. 




a 


6 


C 


d 


e 


/ 


g 


h 


i 


j 


k 


I 




Mo. Da. 






Deg. 






Days 












1 


2 20 


N 


1 


9-11 


18 


Canal 




234 


322 


285 


423 


284 


2 


2 22 


N 


1 


16-19 





it 




201 


327 


326 


302 


219 


3 


2 20 


S 


1 


9-11 


18 


Open air 





344 


416 


412 


321 


292 


4 


2 22 


S 


1 


16-19 











367 


305 


480 


360 


311 


5 


2 20 


s 


1 


9-11 


18 


u 


7 


274 


306 


413 


244 


304 


6 


2 22 


s 


1 


16-19 





t' 


7 


292 


338 


426 


311 


304 


7 


2 21 


N 


2 


7-14 


19 


Canal 


. 


161 


318 


329 


348 


238 


8 


2 23 


N 


2 


9-9 





n 




160 


217 


355 


258 


186 


9 


2 21 


S 


2 


9-14 


19 


Open air 


'o 


288 


289 


382 


282 


319 


10 


2 23 


S 


2 


9-9 











338 


275 


422 


423 


367 


11 


2 21 


S 


2 


9-14 


19 


it 


3 


268 


271 


340 


240 


295 


12 


2 23 


S 


2 


9-9 





' ' 


7 


317 


333 


414 


345 


356 



NOTE. All briquets broken when six and a half months old. 



280 



CEMENT AND CONCRETE 



ment are given in Table 106. The briquets were made in a 
temperature of nine to nineteen degrees Fahr. Half of the 
briquets were made with fresh water, and half with water con- 
taining enough salt to lower its freezing point below that of the 

TABLE 107 

Portland Cement Mortar in Low Temperatures 
Effect of Heating Materials 



REFERENCE. 


PARTS SAND TO 
ONE CEMENT. 


TEMPERATURE 
DEGREES FAHR. 
AIR WHERE 
MOLDED. 


h 

a 






WHERE 
STORED. 


H 

5* 

& 


- 


AGE OF BRIQUETS 
WHEN BROKEN. 


TENSILE STRENGTH, POUNDS 
PER SQUARE INCH. 


Cold 
Ma- 
terials, 
40. 


Warm 
Ma- 
terials, 
110. 


Cold 
Ma- 
terials, 
40. 


Warm 
Ma- 
terials, 
110. 


1 
2 
3 
4 
5 
6 
7 
8 

Means 

9 
10 
11 
12 
13 
14 
15 
16 

Means 
Grand 


1 
1 
1 
1 

2 
2 
2 
2 


1-5 
8-9 
1-5 
8-9 
14-16 
23-24 
14-16 
23-24 


23 

23 

14 

14 



Canal 

u 


Wet 
i 

i 
i 


Mo. 
6 

a 

# 

81 

it 

18i 


590 


593 


582 

Vii 

542 
549 


598 
734' 
550 
597 


770 
460' 
467' 


737 
476 
541 


572 

Vii 

628 
673 
543' 


587 
724 

ei4 

657 
495 


596 
469 


620 
450 


1 
1 
1 
1 

2 
2 
2 
2 


4-6 
9-10 
4-6 
9-10 
15-18 
24-25 
15-18 
24-25 


23 


23 

14 

14 



Open air 

i 


Dry 

(4 

Wet 


Dry 



Wet 
it 


<|i 


487 
507 
422 


470 

542 
453 


639 
605 


622 
605 


471 
534 


479 

549 


Means 























NOTES. Cement, Portland. Sand, " Point aux Pins." 

When warm materials used, the temperature mortar after briquets 
finished, 63 to 71 Fahr. 

When cold materials used, the temperature mortar after briquets 
finished, 32 to 39 Fahr. 

When salt water used for mixing, water was 23 per cent, salt for 
1 to 1 mortars and 14 per cent, salt for 1 to 2 mortars. 

Briquets stored in canal were left in cold air three days before 
immersion. 

Part of briquets stored in open air were immersed in tank in labo- 
ratory one week just before breaking, while others were broken 
dry as indicated. 

In general, each result is mean of five briquets. 



EXPOSURE TO FROST 



281 



air where the briquets were made. The results are chiefly of 
interest as showing the strength that may be attained by natural 
cement mortars under these severe conditions. 

Higher results are usually given by the air-hardened speci-^ 
mens than by those immersed in cold water, though this de- 
pends somewhat upon the brand. Salt is usually beneficial if 
the briquets are immersed, and detrimental for open air ex- 
posure. 

394. Effect of Heating the Materials. The tests in Table 
107 were made to determine the effect of heating the materials 
when working in low temperatures, and thus delaying for a 
time the freezing of the mortars. The details of the tests are 
fully given in the table. The conclusion indicated is that the 
ingredients may be used cold or warm indifferently. A gain of 
only four per cent, is indicated for warm materials in mortars 
mixed with salt water and hardened in cold fresh water. In 
practical work, however, the use of warm materials may so delay 
the freezing as to permit thorough tamping before the mortar 
freezes. Table 108 gives similar results with one brand of natural 
cement, from which it appears that warm materials have a slight 
advantage for either cold water or cold air. hardening. 

TABLE 108 

Natural Cement Mortars in Freezing Weather 
Effect of Heating Materials 





o . 




f. 


TENSILE STRENGTH, POUNDS PER SQ. IN. 


g 


S 




P - 




fe 

a 
9 


w 


TURE AlR 

WHERE 


Wffl 


STORED IN CANAL. 


STORED IN OPEN AIR. 


a 


OD^ 


BRIQUETS 


fe sz 






a 


a 
z 


MOLDED. 


OH 


Materials. 


Materials. 


Materials. 


Materials. 




UH- 




<^ 


32 F. 


100 F. 


32 F. 


100 F. 






Deg. Fahr. 












1 


3 


15 to 16 


6 mos. 


140 


151 


311 


372 


2 


3 


15 to 19 


9 " 


175 


203 


. . 


. . 


3 


2 


22 to 24 


9 " 


167 


204 


355 


361 



NOTES. Cement, Brand Gn, Natural. All mortars made with fresh water. 
Briquets made with warm materials were frozen in from 15 to 24 

minutes after made. 
' Each result, mean of ten briquets. 



282 



CEMENT AND CONCRETE 



395. Consistency of Mortars to Withstand Frost. Since the 
injury due to frost is caused by the expansion of the water 
used in gaging, it would be expected that mortars mixed wet 
would suffer most. This conclusion is confirmed by the tests 
in Table 109. The superiority of dry mortars is especially 
shown in mortars that harden in the air. The treatment to 
which these briquets were subjected was very severe, yet the 
results are excellent. 

TABLE 109 

Consistency of Mortars as Affecting Ability to Withstand Low 

Temperatures 



AGE OF 
BRIQUETS 
WHEN 
BROKEN. 


TENSILE STRENGTH, POUNDS PER SQUARE INCH. 


STORED IN CANAL. 


STORED IN OPEN AIR. 


a 


6 


c 


d 


e 


/ 


g 


h 


6 mos. 


414 


414 


372 


501 


601 


571 


521 


674 


9 mos. 


474 


468 


431 


527 


727 


622 


525 


563 



NOTES. Cement, Portland, Brand R; sand, " Point aux Pins," passing holes 
.08 inch sq. Two parts sand to one cement by weight. Each 
result, mean of five briquets. 
Temperature, air where briquets were molded, 13 to 14 Fahr.; 

materials used, 40 Fahr. 

Temperature mortar when molding completed 32 to 36 Fahr. 
Briquets made with fresh water had frozen after 30 minutes. 
Treatment briquets: a to d, stored in canal (under ic2). 

e to h, stored in open air, January, North- 
ern Michigan. 

Water used: a and e, 10.4 per cent, fresh water. 
6 and /, 11.9 per cent, fresh water. 
c and g, 13.3 per cent, fresh water. 
d and h, 11.9 per cent, water containing 15 per 

cent. salt. 

396. Fineness of Sand and Effect of Frost. The briquets 
reported in Table 110 were made from mortar containing one 
and two parts limestone screenings to one cement, the screen- 
ings varying from coarse to fine. In general, the results follow 
the rule applicable to mortars used in ordinary temperatures, 
namely, that the coarse sands give the best results; but it 
appears that the briquets made with fresh water and" exposed 



EXPOSURE TO FROST 



283 



O *H 









m a 



" 



P 

| 

i 



II 

<u a 

If 



. Oi .00 



03 ^pa^PQ^ 23^5 



OTt<r-rt<i5)M(MCCt^3SrH-^ 
Ci(NQOCOi lOCOi l 00 00 O <N 
CO rt< rf* < CO 



O -O CO >O 



PS 



ONIOV*) 

NI aaeQ Haxv^i NI 

H3J 



HHVJ - 
'aavj\[ 



aunxvaadwaj, 



aiawvg 



XN3W33 3NQ OX 
IHOg 
SXHVJ 



00 O t- O t^ 



SO (M (M <N CO <M <M (M 



CO CO GO Tfl O r-< (N 



5 5 PJ 

^ -S S 



r ^ 



^ .2 ^ 

OJ <D *_ 

s s o 

2* .2* <o 
'C C bo 

PQ <J 



284 CEMENT AND CONCRETE 

in open air reverse this rule, either the finest sand, f $, or the 
f giving the best result. 

397. CONCLUSIONS. The following conclusions concerning 
the use of cement mortars in freezing weather appear to be indi- 
cated by the foregoing tests: 

1st, Mortars should not be mixed wet for use in low tem- 
peratures. 

2d, Portland cement mortars made in cold weather usually 
develop a good tensile strength, especially when exposed to the 
open air. 

3d, Portland cement mortars for open air exposure may be 
benefited by the use of from three to seven per cent, salt in 
the water used in gaging, and from ten to twenty per cent, 
salt in the gaging water may prove beneficial for mortars hard- 
ening in cold water. 

4th, Warming the materials for Portland cement mortar 
appears to have but little effect on its frost resisting qualities. 

5th, Coarse sand usually gives the best results in Portland 
mortars made in cold weather, but fresh water briquets ex- 
posed in open air appear to give better results with fine sand. 

6th, Some natural cements give fairly good results in freez- 
ing weather, while others are practically destroyed by severe 
exposure. The effect of variations in treatment on different 
brands of natural cement is so varied that no general conclu- 
sions can be drawn from the above tests, but the indications 
are that salt water for gaging is beneficial if the mortar hardens 
in cold water, but detrimental for mortars exposed to the open 
air. 

ART. 51. THE ADHESION OF CEMENTS 

398. THE ADHESION BETWEEN PORTLAND AND NATURAL 
CEMENTS. The question sometimes arises as to whether Port- 
land cement will adhere to natural .cement already set, and 
whether fresh natural and Portland cement mortars may be 
used together, as in the case of a Portland facing mortar used 
with natural cement concrete. Tests bearing on these points 
are given in Tables 111, 112 and 113. 

In the tests in Table 111 fresh Portland cement mortar was 
applied to natural cement mortar that had set seven days. 
Natural cement briquets, made neat and with one to four 
parts sand, as seen in the headings of the columns of the table, 



ADHESION PORTLAND AND NATURAL 



285 



were broken at the age of seven days. The fresh Portland 
mortar was applied to the half briquets on the same day that 
the latter were broken, by placing the half briquet in one end 
of the mold, and filling the other half of the mold with fresh 
Portland mortar of the composition shown in the second column. 

TABLE 111 

The Adhesion of Portland Cement to Hardened Natural Cement 

Mortar 



REF. 


PARTS SAND TO 
ONE PART 
PORTLAND 
CEMENT IN 
FRESH MORTAR. 


ADHESION OF PORTLAND MORTAR TO HALF BRIQUETS OF 
HARDENED NATURAL CEMENT CONTAINING PARTS SAND. 





l 


2 


3 


4 


1 
2 
3 




1 
2 


246 

185 
63 


255 

210 
75 


197 

186 
60 


194 
152 
84 


161 
120 
85 



NOTES. Briquets of natural cement, containing parts sand indicated at top 
of columns, were broken at seven days. The half briquets 
were then placed in one end of briquet mold and the other 
end of mold was filled with fresh Portland mortar. 

Fresh mortar made of Portland cement, Brand R. 

Sand, " Point aux Phis," passing No. 10 sieve. 

In general, each result is mean of ten briquets. 

Nearly all briquets broke at juncture of Portland and natural 
mortars. 

It is seen that the neat Portland gave the highest results in 
adhesion, the one-to-one mortar giving a comparatively low ad- 
hesive strength. It is also seen that the neat and one-to-one- 
mortars adhered best to the richer natural cement briquets, 
but the one-to-two Portland gave the greatest adhesi ve strength 
with the poorer natural cement mortars. All of the tests gave 
very irregular results. 

399. To make the briquets the results of which are recorded 
in Table 112, a plate was placed in the center of the mold, one- 
half of the mold was filled with fresh natural cement mortar, 
the plate was then withdrawn and the other half of the mold 
filled with fresh Portland mortar. Briquets in line 1 were 
made with Portland cement alone, while those in line 2 con- 
tained only natural cement, these briquets being made for pur- 
poses of comparison. The briquets containing both Portland 



286 



CEMENT AND CONCRETE 



and natural were made neat and with from one to three parts 
sand. By noting the number of briquets that broke at the 
juncture between Portland and natural, it was found that, in 
general, the adhesion of rich Portland mortar to rich natural 
cement mortar is greater than the strength of the natural, but 
that with the poorer mortars the adhesion is less than the 
strength of the natural. 

TABLE 112 
Adhesion between Fresh Mortars of Portland and Natural Cement 





PARTS SAND TO ONE OP 


ADHESIVE OR COHESIVE STRENGTH, POUNDS 


T> -_,_ 


CEMENT. 


PER SQUARE INCH. 


xvEF. 


IN PO RTLAND 
MORTAR. 


IN NATURAL 
MORTAR. 


28 days. 


3 months. 


6 months. 


1 year. 


l 


2 




278 


372 


410 


464 


2 




2 


164 


243 


268 


308 


3 








318 


358 


323 


380 


4 


1 


1 


252 


326 


376 


383 


5 





1 


229 


331 


356 


357 


6 


2 


1 


226 


196 


339 


298 


7 


2 


2 


128 


235 


265 


285 


8 


1 


2 


145 


213 


259 


271 


9 


1 


3 


103 


160 


185 


206 


10 


2 


3 


95 


176 


206 


197 


11 


3 


3 


63 


162 


197 


193 



NOTES. Portland Cement, Brand R. 
Natural Cement, Brand An. 
Sand, " Point aux Pins," passing No. 10 sieve. 
Both mortars mixed fresh and filled in opposite ends of mold. 

400. In Table 113 the natural cement mortar contained 
three parts sand to one of cement, while the richness of the 
Portland mortars varied from neat to four parts sand. Four 
combinations of different brands were used. Brand R, Port- 
land, and brand An, natural, appear to give the best results 
together. It is also seen from this table that the adhesion of 
the rich Portland mortar is greater than the cohesive strength 
of the natural cement, but when the Portland mortar contains 
three or four parts sand to one cement, the adhesion is less 
than the strength of the natural cement mortar. 

401. THE ADHESION TO STONE AND OTHER MATERIALS. 
Since cement mortars are usually employed to bind other ma- 
terials together, it follows that the adhesive strength is of 'the 



ADHESION TO VARIOUS MATERIALS 



287 



greatest importance. On account of the difficulty of making 
tests of adhesive strength, however, the data concerning it are 
very meager. Two methods have been employed by the au- 
thor in making such tests. One method, used for brick, is 
to cement two bricks together in a cruciform shape. The other 
method consists in placing small blocks of the substance to be 
used in the center of a briquet mold, and filling the ends of 
the mold with the desired mortar. 

TABLE 113 
Adhesion between Fresh Mortars of Portland and Natural Cement 





H 








a 


fc , 

la g 


NATURAL CEMENT 


i 


ADHESIVE STRENGTH POUNDS PER 
SQUARE INCH. 


i 


^ e 5 


MORTAR. 






H 


Q HO 




^ 


Parts Sand to One Cement in Portland 


p 

S 


Jg 




Parts 


P5 


Mortar. 


? 


o CQ 




Cement. 


One 
Cement. 


1 





l 


2 


3 


4 


1 


R 


An 


3 


3 mo. 


162 N 


173 N 


177 N 


162 J 


127 J 


2 


R 


Gn 


3 




147 N 


175 N 


167 N 


156 


151 


3 


A 


En 


3 




156 N 


157 N 


143 N 


136 N 


134 


4 


G 


Bn 


3 




88 N 


106 N 


115 N 


94 


91 J 


5 


R 


An 


3 


lyr. 


214 


206 N 


206 N 


201 N 


183 J 


6 


R 


Gn 


3 




167 N 


165 N 


180 N 


175 N 


169 


7 


A 


En 


3 




157 N 


158 N 


173 N 


166 N 


160 


8 


G 


Bn 


3 




108 J 


127 J 


126 J 


130 J 


114 J 



NOTES : In general, each result is mean of ten briquets. 

Results marked N, briquets broke through the natural cement. 
Results marked J, briquets broke at juncture of Portland and 
natural. 

The small blocks were made one inch square and about 
one-fourth inch thick, two opposite edges of each .piece being 
very slightly hollowed to fit, approximately, the side of the 
mold. These blocks being placed transversely in the center of 
the mold, and the ends of the latter filled with the mortar to 
be tested, formed two joints between the mortar and the block. 

402. Table 114 shows the adhesion of a rich Portland ce- 
ment mortar to various materials. The mortar adheres most 
strongly to brick, the adhesion exceeding the strength of the 
brick itself. A very high result is also obtained with terra 
cotta, and the adhesion to Kelleys Island limestone is high. 
The latter is a dolomitic limestone of the corniferous group, 
which is soft enough to be worked quite easily. The adhesion 



288 



CEMENT AND CONCRETE 



to Drummond Island limestone, which is a much harder stone 
belonging to the Niagara group, is considerably less, and the 
adhesion to the Potsdam sandstone is very low. A higher re- 
sult than would be expected is obtained with ground plate 
glass, but the hammered bar iron gives the lowest result of any 
of the substances tried. 

TABLE 114 
Adhesion of Portland Cement Mortar to Various Materials 



EFERENCE. 


KIND 

OF 

SAND. 


PARTS 
ND TO ONE 
CEMENT. 


AGE 

OF 

SPECI- 
MENS. 


(HESION OF 
MORTAR. 


ADHESION, POUNDS PER SQUARE INCH, 
TO MATERIALS. 





















oS 







a 


6 


c 


d 


e 


* 


9 


1 


Cr. Qtz. 20-30 


1 


28 days 


742 


91 


78 


211 


100 


241 


223 


290 


2 


U ( ( 


1 


6 mos 


775 


103 


122 


201 


252 


284 


310 


395 



NOTES: Cement, Portland, Brand R. 

Adhesion Blocks,. 1 in. X 1 in. X in. inserted in center mold. 
Materials: a Hammered bar iron. 

6 Potsdam sandstone, cleavage surface. 

c Drummond Id. limestone, cleavage surface. 

d Ground plate glass. 

e Kelleys Id. limestone, sawn surface. 

/ Soft terra cotta, filed surface. 

g Soft red building brick, sawn surface. 

403. The Adhesion of Neat and Sand Mortars. Table 115 
shows the cohesive and adhesive strengths of different mortars, 
the adhesion blocks being all of the same material, Kelleys 
Island limestone. The Portland mortar giving the highest ad- 
hesive strength at six months is that containing one-half part 
sand to one part cement, though the greatest cohesive strength 
is given by the one-to-one mortar. With natural cement the 
one-to-one mortar gives the highest strength, both in adhesion 
and cohesion. The ratio of the adhesive strength to the co- 
hesive strength is greater for natural than for Portland. It 
also appears that between twenty-eight days and six months 
the adhesive strength increases more than the cohesive strength. 

404. Effect of Consistency on Adhesion. Table 116 gives 
the results of tests to show the relative effects of the consis-. 
tency of the mortar on the adhesive and cohesive strength. It 



EFFECT OF CONSISTENCY 



289 



TABLE 115 
Adhesion of Mortars Containing Different Amounts of Sand 











COHESIVE OR ADHESIVE STRENGTH, LBS. 


o 


CEMENT. 


AGE 

OF 


COHESION 

OR 


PER SQUARE INCH, OF MORTARS WITH 
SAND, PARTS BY WEIGHT. 


fa 




k-PECI 


ADHESION. 













Kind. 


Brand. 






None. 


Half Part 
Sand. 


One Part. 


Two 
Parts. 


1 


Port. 


R 


28 days 


Cohesion 


686 


710 


747 


467 


2 


i 


j) 


M 


Adhesion 


270 


233 


221 


169 


3 


' 


i 


6mos. 


Cohesion 


631 


787 


816 


551 


4 


1 


4 


li 


Adhesion 


335 


346 


287 


209 


5 


Nat. 


An 


28 days 


Cohesion 


183 


198 


218 


186 


6 


c 


' 


it 


Adhesion 


94 


104 


116 


66 


7 








6mos. 


Cohesion 


263 


334 


383 


376 


8 


' 


' 


" 


Adhesion 


228 


222 


233 


171 



NOTES : Sand, crushed quartz, 20 to 30. 

Adhesion blocks, 1 in. X 1 in. X ^ in., Kelleys Id. limestone, sawn 
surface, saturated before used. 

is seen that the effect of consistency on the adhesive strength 
is less than on the cohesive strength, but that the best results 
in adhesion are given by a mortar that is considerably more 
moist than that which gives the highest strength in cohesion. 
The practical bearing of this point on the use of mortars is evi- 
dent. 

TABLE 116 
Adhesion of Mortars. Varying Consistency 



H 

i 

H 

- 
H 

1 

1 

2 
3 
4 
5 

6 

7 
8 


CEMENT. 


AGE 

OF 

SPECI- 
MENS. 


COHESION 

OR 

ADHESION. 


COHESIVE OR ADHESIVE STRENGTH, LBS. 
PER SQUARE INCH, MORTAR OF 
CONSISTENCY: 


Trifle 
Dry. 


Trifle 
Moist. 


Quite 
Moist. 


Very 
Moist. 


Kind. 


Brand. 


Port. 

Nat. 


R 
t 

An 

c 
i 


28 days 

u 

6mos. 

(C 

28 days 
K 

6mos. 
it 


Cohesion 
Adhesion 
Cohesion 
Adhesion 
Cohesion 
Adhesion 
Cohesion 
Adhesion 


541 
148 
697 
191 
239 
96 
397 
146 


502 
160 
660 
209 
212 
96 
385 
165 


443 
145 
616 
228 
151 
87 
314 
164 


372 
136 
539 
192 
112 
70 
285 
126 



NOTES: Sand," Point aux Phis," pass No.10 sieve, one part to one cement 

by weight. 

Adhesion blocks, 1 in. X 1 in. X { in., Kelleys Id. limestone, 
surfaces filed smooth, saturated with water before used. 



290 



CEMENT AND CONCRETE 



405. Effect of Regaging on Adhesive Strength. The tests 
given in Table 117 were designed to show the effect of regaging 
on the adhesion of cement mortar to stone. A comparison is 
made between mortars used fresh and those that were allowed 
to stand three hours and gaged once an hour. There are but 
few tests from which to draw conclusions and the treatment is 
very severe, but it appears that while the regaging to which 
these mortars were subjected usually resulted in a slight in- 
crease in cohesive strength, the adhesive strength was consid- 
erably impaired. The decrease in adhesive strength was greater 
for natural cement than for Portland, and greater for rich than 
for poor mortars. The effect of regaging on the cohesive strength 
is treated in Art. 47. 

TABLE 117 

Effect of Regaging on Adhesive Strength 



CEMENT. 


ADHESION OR 
COHESION. 


ADHESION OR COHESION, LBS. PER SQ. IN. 


ONE PART SAND TO ONE 
CEMENT. 


THREE PARTS SAND TO 
ONE CEMENT. 


Fresh. 


Regaged. 


Fresh. 


Regaged. 


Portland, Brand X 

U ' i( 


Adhesion 

u 


178 
202 


141 
170 


62 
59 


41 
61 


it It 


Cohesion 


718 


764 


327 


343 


Natural, An 


Adhesion 


142 


90 


17 




u tt 


K 


180 


120 


31 


28 


u u 


Cohesion 


352 


361 


235 


227 



NOTES: Sand, crushed quartz, f$. Each result, mean of two to five speci- 
mens, broken at age of six months. 

In adhesive tests, pieces Kelleys Id. limestone, 1 in. X 1 in. X in., 
placed in center mold and two ends mold filled with mortar. 

Results in columns headed "Fresh" from mortar treated as usual. 

Results in columns headed "Regaged" mortar allowed to stand 
three hours before use, mortar being regaged each hour. 

406. Character of Surface of Stone. In the tests recorded 
in Table 118 all of the adhesion blocks were of Kelleys Island 
limestone, but part of them were finished with smooth filed 
surfaces, while the others were grooved with a coarse rasp. In 
the twenty-eight-day tests there is but little difference in the 
adhesion to the different surfaces, but at six months the adhe- 
sion to the smooth surfaces appears to be slightly greater, ex- 
cept in the case of one-to-two natural cement mortar. 



EFFECT OF PLASTER PARIS 



291 



TABLE 118 
Adhesion of Mortars. Effect of Character of Surface of Stone 



COHESION OR ADHESION 

AND 

CHARACTER OF SURFACE. 

u 


AGE OF 
SPECIMENS. 


ADHESION OR COHESION, 
LBS. PER SQ. IN. 


Portland Brand R. 


Natural BrandD. 


PARTS SAND TO ONE CEMENT. 


1 


2 


1 


2 


Cohesion 


28 days 
ti 

u 

6 mos. 
u 

ti 


539 
151 
152 
714 

238 
223 


377 
. 85 
115 
503 
176 
164 


343 
138 
129 
387 
141 
115 


289 
113 
98 
304 
68 
96 


Adhesion, smooth surface . . 
u grooved surface 
Cohesion 


Adhesion, smooth surface . . 
" grooved surface 



407. The Effect of Plaster of Paris on the Adhesion of Mortar 
to Stone. The results in Table 119 show the effect on the 
adhesive strength of adding small percentages of plaster of 
Paris to cement mortars of Portland and natural cement. The 
Portland cement used was a quick setting sample, neat cement 
pats of which began to set in eighteen minutes. The effect of 
plaster of Paris on the cohesive strength of mortars from these 
samples hardened in dry air, is shown in Table 92, 378. It is 
seen that the addition of from one to three per cent, plaster 

TABLE 119 

Effect of Plaster of Paris on the Adhesive Strength of Cement 

Mortars 











ADHESIVE STRENGTH, LBS. PEK 


REF. 


CEMENT. 


PARTS P.P. 
SAND TO 
ONE 


AGE OP 

SPECIMENS. 


SQ. IN., OP MORTARS IN WHICH 
PER CENT. OP CEMENT REPLACED 
BY PLASTER OF PARIS. 




Kind. 


Brand. 


Sample. 


CEMENT. 







l 


2 


3 


6 


1 


Port. 


R 


26 R 





1 year 


263 


311 


376 


291 


89 


2 


' 


R 


i< 


2 


K 


130 


107 


144 


157 


34 


3 


Nat. 


An 


L 





{( 


88 


97 


87 


133 


a 


4 


it 


An 




2 


n 


64 


74 


89 


82 


93 



NOTES : Adhesion pieces between two halves of briquet were of Kelleys Id. 

limestone, sawn surfaces, saturated with water before used. 
Cement and plaster Paris passed through No. 50 sieve. 
All briquets stored in tank in laboratory. 
Each result, mean of four to ten briquets. 
o Found badly cracked and separated from limestone prisms after 

three days. 



292 CEMENT AND CONCRETE 

has no deleterious effect on the adhesive strength of these 
samples at one year. Six per cent, plaster, however, ruins the 
Portland and the neat natural cement. 

408. THE ADHESION OF CEMENT MORTAR TO BRICK. 
Tests of the adhesion of cement mortar to brick were made by 
cementing pairs of brick in a cruciform shape, with a, one-fourth 
inch joint of mortar. The brick were placed together flatwise, 
with the bed down, so that in the case of stock brick, one 
stock mark, or depression in one side, was filled with mortar. 
The mortar was made more moist than was ordinarily used for 
briquets, but not so moist as would be used in brickwork. The 
top brick of each pair was slightly tapped to place with the 
handle of a pointing trowel, and the excess mortar cut away. 
About forty-eight hours after cemented, the pairs of brick were 
packed in damp sand in a large box prepared for the purpose, 
and the sand was kept in a moist condition by a thorough daily 
sprinkling. For pulling the bricks apart, a special clip was de- 
vised to equalize the pull on the two ends of each brick, and a 
simple lever machine was used to measure the force required. 

409. Tensile tests were made of briquets from mortars 
similar to those used in the adhesive tests and stored in damp 
sand, and the results are used for comparison with the adhesive 
tests. The cohesive strength given by the briquets is not 
strictly comparable with the adhesive strength shown in the 
tests with brick, because of the great difference in the area of 
the breaking sections in the two cases. It has been well estab- 
lished in tensile tests of cohesion that briquets of large cross- 
section break at a lower strength than those of small section. 
It is quite possible also that even with the special clip devised, 
cross-strains were more likely to occur in the adhesive tests 
than in the briquet tests. An opportunity was furnished of 
comparing the tensile strength of neat natural cement mortar 
under the two conditions, for in one case six joints "broke di- 
rectly through the mortar, the adhesion being greater than the 
cohesion. It was found that the strength per square inch 
given by the briquets was at least six times that given by the 
large joint. This difference should be kept in mind in making 
comparisons in the tables between the cohesion and adhesion 
as given. It should also be noted that some of the highest 
results of adhesive strength represent in reality the strength 



ADHESION TO BRICK 



293 



of the brick rather than the adhesive strength of the mortar, as 
chips were pulled from the brick, leaving the mortar joint 
undisturbed. The brick were of a rather poor quality, but 
selected with a view to obtaining those of a uniform degree of 
burning. 

TABLE 120 
Adhesion of Cement Mortar to Brick. Variations in Richness 

of Mortar 









TENSILE STKENGTH, POUNDS PER 








SQUARE INCH, OF MORTARS CONTAINING 


CEMENT. 


AGE OF 


ADHESION 

OR 


PARTS SAND TO ONE CEMENT. 




MORTAR. 










COHESION. 


None. 


i 


1 


2 


3 


Portland, X, 41 S 


28 days 


Cohesion 


632 


596 


589 


409 


270 


t % 


tt 


Adhesion 


48 


42 


24 


20 


11 


i i 


3 months 


Cohesion 


676 


728 


694 


423 


325 


i t 


tt 


Adhesion 


64 


52 


41 


24 


12 


: i 


6 months 


Cohesion 


723 


764 


679 


524 


374 


i t 


1 1 


Adhesion 


50 


56 


39 


20 


14 


Natural, Gn, KK 


3 months 


Cohesion 


180 


240 


317 


279 


181 


tt tt tt 


tt 


Adhesion 


46 


52 


42 


28 


15 


tt It tt 


6 months 


Cohesion 


276 


444 


388 


331 


236 


t t tt tt 


tt 


Adhesion 


44 


52 


50 


38 


18 



NOTES : Bricks were cemented together in pairs in cruciform shape and 

kept in damp sand until time of test. Briquets for cohesion 

tests stored in same manner. 
Each result in cohesion, mean of five briquets. 
Each result in adhesion is in general mean of six results, three 

with common die cut brick and three with sand molded stock 

brick. 
When adhesion exceeded 50 pounds per square inch, bricks were 

about as likely to break as the joint between brick and mortar. 

410. Adhesion of Neat and Sand Mortars of Portland and 
Natural. Some of the results of these tests are given in Table 
120. The most noteworthy point developed is that for mor- 
tars containing more than one-half part of sand to one part 
cement, the adhesion of the natural cement is greater than 
that of the Portland with the same proportion of sand, al- 
though the Portland mortar was much the stronger in cohesion. 
The mortars giving the highest adhesive strength are. those 
containing not more than one-half part sand to one part cement. 

The addition of sand lowers the adhesive strength more 
rapidly than it does the cohesive strength. This point would 



294 



CEMENT AND CONCRETE 



be shown still more clearly if the true adhesive strength of the 
richest mortars was obtained, as we may be certain that the 
adhesion of these mortars would be shown to be considerably 
greater if the brick were strong enough to allow this strength 
to be developed. With natural cement mortars containing not 
more than two parts sand to one cement, the adhesion is one- 
sixth to one-ninth the cohesion, and with Portland mortars 
containing not more than one part sand, the adhesion is about 
one-fifteenth the cohesion. (But see 409 in this connection.) 

411. Effect of Lime Paste on Adhesive Strength of Cement 
Mortars. A number of tests were made to determine the 
effect, on the adhesive and cohesive strength of mortars, of 
mixing lime paste with the cement. Tables 121 and 122 give 
the results of a few preliminary tests on this subject. 

For the tests recorded in Table 121 the mortars were al- 
lowed to harden in dry air. From the cohesive tests it is seen 
that lime in form of paste to the amount of ten per cent, of 

TABLE 121 

Adhesion of Cement Mortar to Brick. Effect of Lime Paste in 
Mortar Hardened in Dry Air 









TENSILE STRENGTH, POUNDS PER 








SQUARE INCH. 


CEMENT. 


OF 

MORTAR. 


COHESION 

OR 

ADHESION. 


Composition of Mortar. 








A 


B 


c 


D 


E 


Portland, X, 41S 


3 months 


Cohesion 


97 


99 


101 


46 


59 


u n n 


4 


Adhesion 


18 


29 


20 


22 


13 


44 (4 a 


6 


44 




24 


20 


19 


11 


Natural, Gn, LL 


3 


Cohesion 


18 


38 


21 


22 


68 


44 44 44 


4 


Adhesion 


39 


32 


36 


28 


11 


44 44 44 


6 


(4 


26 


31 


25 


27 





NOTES: Brick, sand molded stock. 

All briquets and brick stored in dry air. 
Composition of mortars: A 

Grams P. P. river sand. 480 

Grams cement, 120 

Grams lime paste, 

Grams lime contained in lime paste, 
Lime in paste expressed as per cent. 

of cement plus lime, 
Consistency about same as mason's mortar. 



B 

480 

120 

40 

14 



C 

480 
90 
30 
10 



D 

480 
60 
60 
20 



E 

,480 


120 
41 



10 10 25 100 



EFFECT OF LIME PASTE 



295 



the cement had little effect on one-to-four Portland mortars r 
but that a larger amount of lime was very deleterious for dry 
air exposure. The sample of natural cement used did not harden 
well in dry air, and the highest result is given by the lime mor- 
tar without cement. It appears that the adhesive strength of 
the Portland mortar was slightly increased by the addition of a 
small amount of lime paste, but the adhesive strength of natural 
was not greatly affected. The adhesive strength of the natural 
cement is, in general, higher than the Portland. The nat- 
ural cement appeared to harden better in the joints than in the 
briquets, and we have, as a peculiar result, the adhesive strength 
exceeding the cohesion. This illustrates a statement already 
made, that to store briquets in dry air does not approach very 
nearly the ordinary conditions of use. 

412. In Table 122 are given a few tests of mixtures of Port- 

TABLE 122 

Adhesion of Cement Mortar to Brick. Effect of Lime Paste in 

Portland Cement 







TENSILE STRENGTH, POUNDS PER SQUARE INCH. 


MORTAR HARDENED IN 


COHESION 

OR 


COMPOSITION OF MORTAR. 




ADHESION. 








A 


B 


C 


D 


E 


Tank in Laboratory 


Coh'n. 


177 


203 


183 


158 


82 


Dry air, 


it 


167 


180 


167 


150 


81 


Damp sand, u 


" 


173 


198 


171 


154 


88 


Dry air, " 


Adh'ii. 


15 


36 


40 


33 


26 


Damp sand, " 


it 


15 


33 


35 


32 


27 



NOTES: Bricks cemented together in pairs in cruciform shape. 

Age of all mortars when tested, three months. 

Cement, Portland, Brand R, Sample 14 R. Sand, " Point aux Pins." 

Lime paste slaked about six months before use. 

Each result in cohesion, mean of five to ten briquets. 

Each result in adhesion, mean of eight to sixteen pairs of bricks. 

Half of pairs were hard burned brick and half soft burned. 

Composition of mortars: A B C D E 

Grams P. P. (river) sand, 960 960 960 960 960 

Grams cement, 240 240 200 180 120 

Grams lime paste, 80 120 180 360 

Grams lime contained in paste, 27 40 60 120 

Lime as per cent, lime plus cement, 10 16.7 25 50 



296 CEMENT AND CONCRETE 

land cement and lime paste, the mortars being hardened in dry 
air and in damp sand. Cohesive tests are also given of briquets 
hardened in damp sand, water and dry air. It appears that 
the addition of ten per cent, of lime in the form of paste to 
mortars of this sample of Portland increases the tensile strength, 
the effect being least when the mortars harden in dry air. The 
substitution of lime for one-sixth of the cement in a one-to-four 
mortar has little effect on the tensile strength. Larger propor- 
tions of lime result in decreased strength, and if one-half of the 
cement is replaced by lime, the resulting strength is only about 
one-half that given by the cement mortar without lime. The 
results of the adhesive tests show that if half of the cement in 
the mortar is replaced by an equal weight of lime in the form 
of paste, the resulting strength is increased by nearly 100 
per cent., and that if smaller amounts of lime are used, the 
adhesive strength is increased by about 150 per cent, over 
that given by the cement mortar without lime. 

413. The results of a more complete set of tests on this sub- 
ject are given in Table 123. The mortars used included one 
made with four parts sand to one of cement by weight; one in 
which about ten per cent, of lime by weight, which had pre- 
viously been made into lime paste, was added to the mortar; a 
third in which lime, in the form of paste, was substituted for 
one-sixth of the weight of the cement used in the first mortar; 
a fourth, in which lime was substituted for one-fourth of the 
cement; and finally, a mortar composed of lime paste and sand 
only. 

The adhesive strengths of the mortars are given in the 
table. The difference in the adhesion of Portland cement 
mortar to hard brick and to soft brick is not clearly brought 
out. Neither is the strength of air-hardened specimens much 
different from that of the mortars stored in damp sand. The 
use of lime paste with Portland cement in the amounts tried 
here more than doubles the adhesive strength of the mortar. 

The first point to notice in the case of natural cement is 
that the adhesive strength of this mortar without lime is nearly 
double the adhesive strength of Portland mortar without lime. 
The adhesive strength of mortars hardened in damp sand is 
somewhat greater than the strength of similar mortars hard- 
ened in dry air. The addition of a small amount of lime paste 



EFFECT OF LIME PASTE 



297 



PIO 



'pireg duiBQ 



o oot^:o os T-HI-^ 



iO O O <M_ 0? O CO X 

OS' O CO' O ^' I-H t-' O 
I-H -^00-* CO rfiCOO? 



t^ (MCOO O 

r-' os <M' os' ^' cd co <M' 

-H CO-*CO CO OOCOCO 



-2 



2.2 

5 OQ 



s 



- 

B 



'OKVg 



'axsvj 



SUTBJQ 



CO lOr-HiO 

o 

<M 



OS COCO*O 

t-' eo os os 

CO * CO CO 



t- rt| OS t- 

*O CO (M* O* 

T-> CO * rt< 



CO CO CO CO 
*O t^ O CO' 
CO CO CO CO 



(N (MOt>- 

(N CO Tti O 
<M 



O COOO 



* CD CO <M CO OOCOCO 
OS O t^* OS OS O CO i* 
t i TjHTt^CO CO -^COCO 



10' o <N o 

-i ^<*^ 



10 CO CD CO OS CO t- CO 



t^ cococo 



co oscoos 
o co -*' CD' 



t^ OCO^t< CO COrHl^ 
Tji CD CO uO -*' <N CO 



OOSt- (M 



co cs -*ti i>. Os' i i co' ^' 

r-t (NCOCO CO Tt^COCO 



<M CO-^C<> (M 

O ^' CO O CO Tt<' CO 
GN OOCO 



c !> co r*- 



000 o o o o 
CO SV| CO O CO S 00 



' = 



298 CEMENT AND CONCRETE 

increases the adhesive strength somewhat, and when as much 
as twenty-five per cent, of the cement is replaced by lime in 
the form of paste, the adhesive strength of the natural cement 
mortar is not usually diminished. The effect of lime paste, 
however, on the adhesive strength is not nearly so great as it 
is in the case of Portland mortars. 

The following conclusions may be briefly stated: The ratio 
of adhesive to cohesive strength is much greater with natural 
cement than with Portland. If a high adhesive strength is 
desired, Portland cement should not be mixed with more than 
two parts sand unless lime paste is added to the mortar, as 
the use of lime paste materially increases the adhesive strength 
of lean mortars. Tests of cohesion of similar mortars contain- 
ing lime paste are given in Art. 48. 

414. THE ADHESION OF CEMENT TO RODS OF STEEL AND 
IRON. The tests recorded in Tables 124 and 125 were made 
to determine the adhesion of cement mortar to iron rods, or 
the strength of a bolt anchorage secured with cement mortar, 
and the style of rod and kind of mortar which would give the 
best results. The bars were made in an ordinary concrete 
mold, ten inches by ten inches by four and one-half feet. The 
rods or bolts were placed in a row along the center of the box, 
being spaced about nine inches apart, and the mortar was 
rammed about them. After being allowed to set in a warm 
room for twenty-eight days, the rods were pulled by means of 
two hydraulic jacks, a special grip being used to grasp the free 
end of the rod, and an hydraulic weighing machine serving to 
measure the pull required to start it. The supports against 
which the hydraulic jacks were braced bore at points on the 
concrete bar #bout three or four inches from the center of the 
rod which was being tested. 

415. The rods given in Table 124 were imbedded in mortar 
composed of one part of Portland cement to two parts lime- 
stone screenings. The rods were cut from bar iron and were 
perfectly plain, without nuts or fox wedges. The results in- 
dicate that the force required is proportional to the area of 
contact. Comparing the different styles and sizes of plain rods, 
no difference in favor of one style or size can be determined; 
the apparent higher resistance per square inch offered by one- 
inch rods would probably disappear in a large number of tests. 



ADHESION TO STEEL RODS 



299 



TABLE 124 

Resistance to Pulling of Iron Rods of Various Forms Imbedded 

in Mortar 





h 
o 










POUNDS PULL. 










PERIM- 


DEPTH 




REF. 


W 

g 


MORTAR, 
BAR No. 


DESCRIPTION OF ROD. 


ETER 
OF ROD, 
INCHES. 


IM- 
BEDDED, 
INCHES. 


Per In. 
Depth 
Im- 


Per Sq. 
In. Area 
in Con- 




fe 










bedded. 


tact. 


1 


3 


2, 6, 7 


Plain. 1" diameter 


1.57 


8 to 10 


700 


447 


2 


3 


" 


1" 


3.14 


" 


1750 


656 


3 


3 





Jl" 11 


3.93 


" 


2060 


524 


4 


3 


ii 


|" square 


2.00 


ti 


1085 


543 


5 


4 


2, 5, 6, 7 


j// . . 


4.00 


it 


2250 


562 


6 


3 


2, 6, 7 


H" " 


5.00 


it 


2170 


434 


7 


3 


4,5 


( Twisted 1" square, ) 
) 1 turn in 8" length f 


4.3 1 


9 


2595 


608 


8 


3 


tt 


j Twisted 1" square, ) 
( 2 turns in 8" length f 


4.3* 


9 


2215 


516 


9 


3 


ii 


j Twisted 1" square, ) 
1 3 turns in 8" length j 


4.3 1 


9-9.5 


2405 


561 



NOTES: Cement, Portland, Brand R. 

Sand, limestone screenings passing f inch slits, two parts by 

weight to one cement. 
Mortar one month old when tension was applied to rods. 

The rods given in lines seven to nine were made by twisting 
a piece of one inch square bar iron. The twisted portion was 
eight inches in length. Comparing the plain one inch square 
bolts with the twisted bolts, it appears that the former offered 
a resistance of 2,245 pounds per inch depth while the latter 
gave 2,405 pounds, an increase of less than eight per cent. 

416. In the tests recorded in Table 125, the ordinary river 
sand used in construction was employed. The mortar was 
made neat and with two and four parts sand to one of cement. 
The depth the rods were imbedded varied from two inches to 
ten inches. The one-to-two mortar gave nearly as good results 
as neat cement, but the one-to-four mortar gave much lower 
results. The resistance seems to vary directly as the area of 
contact without reference to the depth imbedded, except as 



1 In computing adhesion, or shear, or pounds pull per square inch of area 
in contact, perimeter considered circumference of a circle of diameter equal 
to the distance between opposite edges of rod after twisting. A core of mor- 
tar of this diameter, was torn from bar in pulling. 

To perceive effect twisting, compare pounds pull per inch depth imbedded. 



300 



CEMENT AND CONCRETE 



this enters in obtaining the said area. The results obtained in 
this table do not compare favorably with those obtained in 
Table 124, where limestone screenings were used. 

TABLE 125 

Resistance to Pulling of Iron Rods Imbedded in Mortar. Variations 
in Depth Imbedded and in Richness of Mortar 



PARTS 
SAND TO 
ONE 
CEMENT. 


ADHESION, POUNDS PER SQUARE INCH OF SURFACE IN CONTACT FOR DIFFERENT 
DEPTHS IMBEDDED. 


Depths ) 
Imbed- J 
ded.In J 


1.9-2.2 


3.2 


4 


4.5-4.8 


6.8-6 


7.8-8 


8.8 


9.6-10 


No. 
Re- 
sults. 


Mean. 



2 
4 




340 

272 
74 


294 


346 
270 
119 


262 


313 
255 
117 


247' 
100 


228 


340 
275 
142 


5 
15 
10 


313 
264 
111 









NOTES : Cement, Portland, Brand R. 

Sand, "Point aux Pins," river sand. 
Mortar one month old when rods pulled. 
Rods, round, 1 inch diameter. 

417. Tables 126 and 127 are from similar tests made by 
Messrs. Peabody and Emerson. 1 The rods in Table 126 were 
of various shapes and included some having rivets through 
them. The inch by 1 inch bars gave lower adhesion per 
square inch than the square and round rods. When two rods 
are twisted together and imbedded in a small specimen, the 
tendency is to split the specimen. The bars containing rivets 
broke before the adhesion was overcome, although the depth 
imbedded was but six inches. 

In Table 127 neat cement paste and concretes of several 
compositions are tried. These results are of interest as show- 
ing that concretes show as great adhesion to steel rods as do 
mortars. The very low result obtained with neat cement in 
this table is not explained and is in opposition to the results 
in Table 125. 



Engineering News, March 10, 1904. 



ADHESION TO STEEL RODS 



301 



TABLE 126 

Adhesion of Mortar to Steel Rods of Various Shapes, Imbedded 
about Six Inches 



No. 

OF 

TESTS. 


DESCRIPTION OF ROD. 


PERIMETER 
OP ROD, 
INCHES. 


POUNDS PULL. 


Per Inch 
Depth 
Imbedded. 


Per Square 
Inch Area 
in Contact. 


4 

3 
4 
4 

4 

4 


Plain, \" square. 
Plain, \" square. 
Plain, |" round. 
Twisted, " square, 
i" by 1 ". 

Two \" rods twisted 
together. 


1.0 
2.0 
1.57 

2.5 


369 
864 
804 
1259 
744 


369 
432 
512 

293 


Three specimens split. 
One rod broke at 8,000 Ibs., or when 
adhesion was 1,250 Ibs. per inch 
depth. 


4 


i" by 1" with f " rivets 
through. 


One specimen split. 
Three rods broke at first rivet with 
9,800 to 10,500 pounds, or when 
adhesion was 1,500 to 1,660 Ibs. 
per inch depth. 



NOTES: Tests by Messrs. George A. Peabody and Samuel W. Emerson. 

Mortar composed of one part cement (Portland) to three parts sand. 
Specimens approximately 6 inch cubes. One rod imbedded in 

each, 6 to 6 inches. 
Rods pulled forty and eighty days after mortar was made. 

TABLE 127 

Adhesion of Mortars and Concretes of Various Compositions to 
One Inch Square Steel Rods Imbedded about Ten Inches 





COMPOSITION OF MORTAR OR 


POUNDS PULL. 


No. 


CONCRETE. 




OF 








TESTS. 




Per Inch 
Depth 
Imbedded. 


Per Square 
Inch Area 
in Contact. 


Cement. 


Sand. 


Stone. 


Gravel. 


4 


1 











1112 


278 


4 


1 


3 








1644 


411 


4 


1 


3 


6 





1912 


478 


4 


1 


3 





6 


2062 


516 


4 


1 


2 


4 





2348 


587 


4 


1 


2 





4 


2187 


547 



NOTE : Tests by Messrs. George A. Peabody and Samuel W. Emerson. 



CHAPTER XVI 

THE COMPRESSIVE STRENGTH AND MODULUS OF 
ELASTICITY OF MORTAR AND CONCRETE 

ART. 52. COMPRESSIVE STRENGTH OF MORTAR 

418. The compressive strength of cement mortar is from 
five to ten times the tensile strength. As the result obtained 
in tests of either compression or tension depends upon the shape 
and size of the specimen, no definite value can be assigned to 
the ratio of compression to tension. Comparative tests have 
indicated in a general way that the cements giving the best 
results in tension show also the highest compressive strength; 
but with variations in treatment, different kinds and brands of 
cement do not give the same variations in the ratios of the two 
kinds of strength. 

Mortar is not usually employed alone in large masses. It 
more frequently forms the binding medium between fragments 
of other substances, such as brick and stone. The dependence 
of the strength of masonry upon the strength of the mortar 
increases with the roughness of the stone or brick, and the 
thickness of the bed joints. In fine ashlar masonry this depend- 
ence is comparatively small, in brickwork it is important, and 
in concrete any increase in the strength of the mortar increases 
the strength of the concrete in nearly the same ratio. 

Piers of brickwork may give a crushing resistance either 
greater or less than the strength of cubes made from mortar 
of the same composition as that used in building the piers. 
Thin beds of mortar between strong materials resist high com- 
pressive stresses, while in walls or piers built with weak blocks, 
the mortar is destroyed by the cracking of the blocks at a lower 
stress than the mortar would withstand in a cube pressed 
between steel plates. Since in brick and stone masonry the 
mortar forms but a small part of the structure, it is not econom- 
ical to use a poor quality of mortar with good brick and stone. 

302 



COMPRESSIVE STRENGTH MORTAR 



303 



419. Ratio of Compressive to Tensile Strength. M. E. 

Candlot has made many experiments showing the effect of 
certain variations in the preparation of mortars upon the com- 
pressive and tensile strength. A few of the results of one 
series are presented in Table 128. The reduction from the 
metric system has been made, and a column added giving 
approximately the number of parts of sand to one of cement by 
weight, the accurate proportions appearing in the form of 
weight of cement to one cubic yard of sand. These results in- 
dicate that the ratio of the strength in compression to that in 
tension increases with the age of the mortar and also with its 
richness. 

TABLE 128 

Resistance of Cement Mortars to Tension and Compression, -with 

Varying Proportions of Normal Sand 

SPECIMENS HARDENED IN FRESH WATER 

[From Ciments et Chaux Hydrauliques, par M. E. Candlot.] 



I1ft^ 


POUNDS CEMENT 
PER CUBIC YARD OF 
SAND, ACTUAL. 


RESISTANCE IN POUNDS PER SQUARE INCH IN TENSION AND 
COMPRESSION. 


1 


7 days. 


28 days. 


1 year. 


2 years. 


3 years. 


T. 


C. 


T. 


C. 


T. 


C. 


T. 


C. 


T. 


C. 


10.8 
6.4 
4.6 
3.5 
2.9 
2.6 
2.0 
1.6 


250 
420 
590 
760 
930 
1100 
1350 
1690 


27 
128 
139 
233 
251 
349 
368 
443 


266 
643 
1040 
1520 
2110 
2630 
3360 
3310 


38 
143 
234 
393 
462 
551 
550 
561 


408 
1164 
1940 
3080 
3690 
5020 
5020 
5070 


70 
212 
337 
435 
490 
594 
713 
767 


507 
1730 
2980 
4020 
5580 
5820 
7750 
7670 


74 
209 
284 
400 
490 
557 
805 
907 


572 
1630 
2930 
4400 
5680 
6060 
7860 
8800 


108 
219 
341 
462 
557 
616 
784 
815 


738 
1775 
3080 
4590 
6060 
6480 
8710 
9180 


6.8 
8.1 
9.0 
9.9 
10.9 
10.5 
11.1 
11.3 



From a study of the results of nearly three thousand tests 
made by Professor Tetmajer, the late Professor J. B. Johnson 
concluded that for mortars containing three parts sand to one 
cement the ratio of the compressive strength to the tensile 
strength is equal to 8.64 4- 1.8 log. A, where A is the age of 
the mortar in months. It is shown above that the ratio in- 
creases with increasing proportions of sand. 

420. Table 129 gives some results obtained at the Water- 
town Arsenal in tests of cement mortar cubes. 1 The mortars 



Prepared by Mr. George W. Rafter for the State Engineer of New York. 



304 



CEMENT AND CONCRETE 



TABLE 129 
Compressive Strength of Cement Mortar. Portland and Natural 

TESTS OF 12 INCH CUBES, TWENTY MONTHS OLD, MADE AT WATERTOWN 
ARSENAL FOR STATE ENGINEER OF NEW YORK 



METHOD OF STORAGE OF 
CUBES. 


CEMENT. 


CONSISTENCY 

OP 

MORTAR. 


CRUSHING STRENGTH, LBS. PER 
SQUARE INCH, FOR MORTARS 
CONTAINING PARTS SAND TO ONE 
CEMENT BY VOLUME: 


Kind. 


Brand. 


1 


2 


3 


4 


Mean 








r Dry 


3479 


2200 


1154 




2278 


Water 3 to 4 mo., 
then buried in sand. 


Nat. 


Buffalo 


! Plastic 
I Excess 


2795 
2161 


1783 
1698 


1000 
776 




1859 
1545 








I Mean 


2812 


1894 


977 


. . . 


1894 


Covered with burlap; 






C Dry 


3347 


2000 ! 


961 




2103 


kept wet for several 
weeks, then exposed 


Nat. 


Buffalo 


J Plastic 
1 Excess 


2476 
2070 


1294 
1358 


692 
738 




1487 
1389 


to weather. . . 






L Mean 


2631 


1551 


797 




1660 








f Dry 


2844 


2051 


987 




1961 


In cool cellar . . 


Nat. 


Buffalo 


1 Plastic 
1 Excess 


2514 
2159 


1256 
1386 


883 
678 


. . . 


1551 
1408 








I Mean 


2504 


1564 


849 


. . . 


1640 


Fully exposed to 
weather . . . 


Nat. 


Buffalo 


f Dry 
! Plastic 
1 Excess 


3272 
2667 
1996 


1879 
1356 
1311 


1054 
822 
669 


. . . 


2068 
1615 
1325 








I Mean 


2645 


1513 


848 




1669 








f Dry 


3236 


2032 2 


1039 




2102 


Means .... 






j Plastic 


2613 


1421 


849 




1628 








[ Excess 


2097 


1438 


715 


. . . 


1417 


Grand mean . . 








2649 


1630 


868 




1716 


Water 3 to 4 mo., 
then buried in sand 


Port. 


Empire 


f Dry 
\ Plastic 


. . . 


3897 
3642 


2494 
2168 


1782 
1717 


. . . 


Covered with burlap 


















kept wet for severa 


















weeks, then exposed 
to weather. . . 


Port. 


Empire 


f Dry 

| Plastic 


. . . 


3880 
3672 


2492 
2168 


1489 
1726 


. . . 


In cool cellar . . 


Port. 


Empire 


f Dry 
1 Plastic 


. . . 


3397 
3313 


2132 
2164 


1614 
1679 


. . . 


Fully exposed to 
weather . . . 


Port. 


Empire 


J Dry 

1 Plastic 


. . . 


4059 

3589 


2450 

2270 


1715 
1465 










j Dry 




3808 


2392 


1650 




Means 






|^ Plastic 


. . . 


3554 


2193 


1647 






contained one, two and three volumes of sand to one of natural 
cement, and two to four parts sand to one volume of Portland. 



1 Result interpolated. 

* 2,043 omitting interpolated result. 



COMPRESSIVE STRENGTH CONCRETE 305 

The proportions of water used were such as to give mortars of 
different consistency, "dry," like damp earth, "plastic," of the 
consistency usually employed by masons, and "excess," quak- 
ing like liver with slight tamping. The specimens were twelve 
inch cubes and four methods of storage were used, as indicated. 

Comparing the results with similar tests of tensile strength, 
it appears that the strength in compression decreases more 
rapidly as sand is added than does the tensile strength. The 
same conclusion was drawn from Table 128. 

The strength of the Portland mortar with four parts sand 
is about equal to the strength of the natural with two parts. 
The dry mortar gives the highest strength with natural cement, 
but with Portland the "dry" and "plastic" give about the 
same result. 

Concerning the consistency, it has already been pointed but 
that the conditions of the actual employment of mortar are 
such as to favor, in general, the use of a wetter mixture than 
that which gives the best results in laboratory tests of mortars. 
As to storage, the specimens kept in water for three or four 
months after made, give the highest results with natural ce- 
ment. There seems to be no choice between the other three 
methods of storage. 

ART. 53. COMPRESSIVE STRENGTH OF CONCRETE WITH VARIOUS 
PROPORTIONS OF INGREDIENTS 

421. With the increasing use of concrete in arch bridges, 
in foundation' piers and in columns of buildings, and especially 
in connection with steel in beams, etc., the compressive strength 
of the material becomes of the greatest importance. Moreover, 
the composition of concrete may vary so much, the range of 
available aggregates is so wide, and the methods of manipula- 
tion are so diverse, that many tests must be studied before 
one can judge of the probable strength of a given mixture. 

For any very extended work, it may be found economical 
to make a series of tests using the materials available, and 
combining them as nearly as possible in the manner proposed 
in actual construction. This practice has been followed in 
several important works, and the data thus accumulated have 
added much to our hitherto somewhat vague notions of the 
probable strength of different mixtures under varying condi- 



306 



CEMENT AND CONCRETE 



tions of use. It is possible here to abstract but a few of the 
more reliable and complete tests of this kind, selecting those 
which indicate the value of certain special kinds of aggregate 
or the effect of certain variations in manipulation. 

422. In connection with the design of the Boston Elevated 
R. R., Mr. George A. Kimball, Chief Engineer, prepared a series 
of concrete cubes of mixtures usually employed in practice, 
and wjth the materials available for the work in hand, and these 
cubes were tested at the Watertown Arsenal in 1899. A por 
tion of the results of these experiments are given in Table 130, 
where the details concerning character of the materials and 
the preparation of the specimens are shown. As each result 
in the table is the mean of at least twenty specimens, the ir- 
regularities frequently appearing in compressive tests have 
been largely eliminated, and the results are worthy of much 
confidence. 

TABLE 130 

Compressive Strength of Concrete 
TESTS OF 12 INCH CONCRETE CUBES FOR BOSTON ELEVATED RAILROAD. 



COMPOSITION OP CONCRETE BY 
VOLUME. 


CRUSHING STRENGTH, POUNDS PER SQUARE INCH, 
AT AGE, 


Cement. 


Sand. 


Stone. 


7 days. 


1 month. 


3 months. 


6 months. 


1 


2 


4 


1525 


2440 


2944 


3904 


1 


3 


6 


1232 


2063 


2432 


2969 


1 


6 


12 


583 


1042 


1066 


1313 



NOTES: 

Materials : Cement, mean results with four brands Portland, two Ameri- 
can, two German. 

Sand, coarse, clean, sharp, voids 33 per cent, loose. 
Broken stone, conglomerate passing 2^ inch ring, voids 49 

per cent, loose. 

Mixing: Sand and cement turned twice, mortar and stone turned twice. 
Storage : Cubes removed from molds three or four days after made and 

buried in wet ground until about a week before testing. 
Each result, mean of twenty or more tests. 

Tests made at Watertown Arsenal, for George A. Kimball, Chief Engineer, 
Boston Elevated R.R. "Tests of Metals," 1899. 

At the time of making these tests some cubes were crushed 
with a die having a smaller area than the face of the cube. 



COMPRESSIVE STRENGTH CONCRETE 307 

With a die 8 by 8^ inches on one compression face, the area of 
the die being thus about .46 of the area of the cube face, the 
strength per square inch under the die was about twenty-five 
per cent, higher than when the entire face of the cube was 
pressed. This is in line with the behavior of all brittle sub- 
stances under compression, as shown by Professor Bauschinger 
in testing sandstone specimens. 

423. Tables 131 and 132 give a summary of a part of a very 
valuable series of tests of concrete cubes prepared by Mr. George 
W. Rafter and tested at the Watertown Arsenal for the State 
Engineer of New York. 1 

The results summarized in Table 131 are those obtained with 
four brands of Portland cement made in the State of New 
York, namely, Wayland, Genessee, Empire and Ironclad. Tests 
were also made with a sand-cement, and with one brand of 
natural, but these results are not included in the table. The 
aggregate was sandstone of the Portage group, broken by hand 
to pass a two inch ring. 

The mortars used in making the cubes were of three degrees 
of consistency: (a) In the dryest blocks the mortar was only a 
little more moist than damp earth, and much ramming was 
required to flush water to the surface, (b) In another set the 
mortar was about the consistency of ordinary mason's mortar, 
(c) In the third set, the mortar was wet enough to quake like 
liver under moderate ramming. 

424. The mortar was composed of one volume of loose ce- 
ment to two, three or four volumes of loose sand. Other pro- 
portions were also employed, but in this table only those re- 
sults are included in which the series of tests was complete as 
to variations in consistency and storage. 

The voids in the stone were about forty-three per cent, 
when the measure was slightly shaken, and thirty-seven and 
a half per cent, when rammed without mortar. The amount 
of mortar used was made either thirty-three per cent, or forty 
per cent, of the volume of the loose stone. 

Four methods of storage were used as follows: 1st, blocks 
immersed in water as soon as they were removed from the 
molds, and after three or four months they were buried in sand; 



Report of State Engineer of New York, 1897. 



308 



CEMENT AND CONCRETE 



TABLE 131 

Compressive Strength of Concrete 

MEAN RESULTS WITH FOUR BRANDS PORTLAND CEMENT, ILLUSTRATING EFFECTS 

OF PROPORTIONS, CONSISTENCY, AND METHODS OF STORAGE. TESTS OF 

CONCRETE CUBES, ABOUT TWENTY MONTHS OLD, MADE FOR 

STATE ENGINEER OF NEW YORK 



CO O i-l CO 

. 1-1 "<* CO CO 
' - 1- IO T* VO 

OI OI OI OI 



GSI T 1 T 1 T 1 

(N GNJ <>J <M 



OI OJ OI OI 



r- Tt< co Oi 

OI CO rH CO 

T-H O O O 
OI OI OI OI 



CO Oi Tt< 

CO T-H i-H 

OI OI OI 



CO QO CO I-H 
<M t^ TJH O 
t^ CO O CO 



OI Oi U3 Oi 
T-H t^ OI CO 
CO CO CO CO 



CO ^ Oi 
Oi CO OI 

10 o 10 



S 



t^ -^ O <M 
CD CO CO t- 



Tj< 

UO 



S w 



OI < CO O 
CO >O TH CO 
' -^ CO CO CO 
OI OI OI OI 



OI Oi O U3 

TH OI CD CO 

CO Oi Oi 

01 TH TH T-H 



i 



Oi H^ 
*O CO 



rH GO CO Oi 



CO CO 10 
T-H CO Oi Oi 

01 Oi t^ CO 



lO CO CO OI 
CO T-H rH t- 
I>- lO T-H HH 

CO CO CO CO 



TH Oi Oi Oi 

70 01 01 01 



CO TH rH 

t~ t- CO 
OI OI OI 



Oi 
tO UD 

01 OI 



O OI CO 

O O ^ 
CO 01 01 



CO Oi t 
T-H O >O OI 
iO CO "^ O 

01 01 01 01 



02 bfl 



PER C 
GGREG 



a 
ter 3 to 4 mos., then 
buried in sand. 



four 
orage 
cent- 



o 
mo 



a 






s 

II 
SB 



n resu 
method 
and tw 
of 



COMPRESSIVE STRENGTH CONCRETE 309 

2d, blocks covered with burlap and wet frequently for several 
weeks, after which they were exposed to the weather; 3d, kept 
in a cool cellar from the time of fabrication until shipped for 
testing, and 4th, fully exposed to the weather throughout. 

425. In Table 131 each result is the mean of four cubes, one 
of each brand. The mean results are so arranged as to show 
the effects of variations in the amount, the richness, and the 
consistency of the mortar, and of the different methods of storage. 

Taking up first the question of consistency, it appears from 
column "/" that the use of plastic mortar, marked "mason's," 
gave from 92 to 97 per cent, of the strength given by the dry 
mortar of about the consistency of "moist earth;" and that 
the "quaking" concrete gave from 89 to 95 per cent, of the 
strength of that marked "moist earth." From the three lines 
at the bottom of the table it is seen that in the poor concrete, 
one-to-four mortar, the wettest mortar gave nearly as good 
results as the dryest, while in the rich concrete, one-to-two 
mortar, the strength of the wet was but 89 per cent, of the dry. 
The explanation of this may be found in the fact that in the 
poor concrete the mortar was "brash," and the concrete did 
not ram well with a dry mortar, while the rich mortar was 
"fuller" and more plastic, so that the excess of water was not 
needed to make a compact mass. 

426. Turning to the question of the amount of mortar, it is 
plainly shown that the concrete containing forty per cent, is 
but little better than that containing thirty-three per cent. 
This is in line with what has been said elsewhere, that an excess 
of mortar, as well as a deficiency, may be an actual detriment 
to the strength of the concrete. In this case the thirty-three 
per cent, mortar was not quite sufficient to fill the voids in 
the stone, and forty per cent, was a very slight excess. 

Some interesting conclusions are indicated by the results in 
the line marked "ratios," near the bottom of the table. The 
ratios of the strength of the concrete containing thirty-three 
per cent, mortar to the strength of that containing forty per 
cent, are 91.6 per cent., 98.5 per cent, and 102.6 per cent., re- 
spectively, for one-to-two, one-to-three, and one-to-four mor- 
tars. That is, with a rich mortar forty per cent, may be used 
to advantage, but if the mortar is of poor quality, the strength 
of the concrete is not increased by an excess of mortar. 



310 



CEMENT AND CONCRETE 



Finally, as to the strength developed under different con- 
ditions of storage, column " k " shows that for these cements the 
highest strengths are attained by immersing the concrete in 
water. In comparison, the strength developed by the concrete 
covered with wet burlap is 84 per cent. ; in cool cellar, 82 per cent. ; 
and in the open air fully exposed to the weather, 81 per cent. 

427. The results given in Table 132 are the mean crushing 
strengths obtained in the same series of tests as described above, 
so arranged as to bring out the effect of the richness of the 
mortar. Although several brands were tested, only the results 
obtained with a single brand of Portland, namely, Millen's 
"Wayland," are included here, since the series was not com- 
pleted with other brands. From similar tests with concretes 
containing one-to-two and one-to-three mortars only, it was 
found that three other brands of Portland gave from 91 to 102 
per 'cent, of the strength obtained with the Wayland, and a 
brand of sand-cement gave 66 per cent. 



TABLE 132 

Compressive Strength of Concrete. Effect of Richness of Mortar 
MEAN RESULTS, FOUR METHODS OF STORAGE 







MORTAR, PROPORTIONS CEMENT TO SAND. 


VOLUME MOR- 






TAR .\s 


CONSISTENCY 












PER CENT. OF 


OF 


1-1 


1-2 


1-3 


1-4 


1-5 


VOLUME OF 


CONCRETE. 












AGGREGATE. 










Crushing Strength, Lbs. per Sq. In. 


( 


Moist Earth . 


4267 


2888 


2056 


1810 


}537 


33 1 


Mason's . . . 


4072 


2777 


2207 


1600 


1568 


I 


Quaking . . . 


3764 


2847 


1723 


1767 


1441 


/ 


Moist Earth . 


3966 


3404 


2179 


1671 


1559 


40 


Mason's . . . 


4123 


2960 


2027 


1750 


1465 


( 


Quaking . . . 


3256 


3168 


2016 


1670 


1400 




Mean 


3908 


3007 


2035 


1711 


1495 




Proportional . 


100 


77 


52 


44 


38 



NOTES: One brand Portland cement. 

Aggregate, Portage sandstone, broken to pass two-inch ring. 
Age of cubes about twenty months. 
Each result, mean of four cubes. 



COMPRESSIVE STRENGTH CONCRETE 



311 



Each result in the table is the mean of four cubes, each 
stored in a different manner. Tests with four brands (Table 
131) where the concretes were made with one-to-two, one-to- 
three and one-to-four mortars, indicated that the percentages of 
the mean strength developed in the several methods of storage 
were as follows: If stored in water, the cubes developed 115 per 
cent, of the mean result; covered with burlap kept wet, the 
cubes developed 97 per cent.; stored in a cool cellar, 95 per 
cent.; and fully exposed to weather, 93 per cent, of the mean 
strength. 

The mean results given at the bottom of the table represent 
each a mean of twenty-four cubes made with two different 
.amounts of mortar, three degrees of consistency, and four 
methods of storage. By applying the percentages given above 
the probable corresponding result for any set of conditions 
may be obtained. The last line of the table shows the propor- 
tions that the strength of the concretes made with poorer mor- 
tars, bear to the strength obtained with one-to-one mortar. 

428. Table 133 gives the results of a series of tests made by 
J. W. Sussex at the University of Illinois. 1 The materials used 
were " Chicago AA Portland cement, sand containing a small 

TABLE 133 

Compreasive Strength of Concrete. Relative Strength of Dry, 
Medium, and Wet Mixtures 







TENSILE STRENGTH, POUNDS PER 








SQUARE INCH, AT AGE OF 


PROPOR- 


CONSISTENCY. 


TAMPING. 




TIONAL VALUE 












AT 3 MoS. 






7 days. 


1 month. 


3 months. 




Dry ... 


Light 


1200 


1750 


2500 


82 


Medium 


it 


2200 


2290 


2150 


71 


Wet ... 


i 1 


1040 


2230 


3040 


100 


Dry ... 


Hard 


1340 


1960 


2600 


86 


Medium 


( 


1330 


2565 


2580 


85 



NOTES: Concrete composition: 

Cement, Portland, one volume. 

Sand, containing some fine gravel, three volumes. 

Six volumes broken limestone passing one-inch mesh. 
Specimens, six-inch cubes. 
Results by J. W. Sussex, Univ. of 111. 



Technograph, 1902-03. 



312 CEMENT AND CONCRETE 

percentage of fine gravel, and crushed limestone which would 
pass through a sieve with one-inch mesh." The proportions 
were three parts sand and six of broken stone to one volume of 
loose cement. The cubes were six inches on a side. The 
treatment during storage is not stated. The consistency of 
the concrete was as follows: "Dry," water 6.0 per cent., as 
moist as damp earth, no free water flushed to surface in ram- 
ming. "Medium/' 7.8 per cent, water; water flushed to surface 
and concrete quaked only after being well rammed. "Wet," 
water 9.4 per cent., concrete quaked in handling and could be 
tamped but lightly. 

Each result in the table is the mean of three cubes. The 
concrete was tamped in layers about one inch thick with a 
rammer weighing 11 pounds and dropped six inches. Ten 
blows of the rammer constituted "light" tamping and twenty 
blows "hard" tamping. The results show that the "medium" 
concrete gains its strength more rapidly than the "wet," but 
that at one month the "wet" concrete has a higher strength 
than the dry, and that at three months the wet surpasses in 
strength both the dry and the medium. 

ART. 54. CONCRETES WITH VARIOUS KINDS AND SIZES OF 
AGGREGATES 

429. It has already been stated that the character of the 
aggregates is second only to the quality of the mortar in its 
effect on the strength of concrete. The materials available for 
aggregate in different localities are so varied that only a general 
idea of their relative values may be obtained from a limited 
number of tests. 

The results given in Table 134 are from tests made at the 
Watertown Arsenal, 1 and show the compressive strengths of 
concretes made with broken trap and gravel of different sizes. 
The concretes are all very rich, and the strengths correspond- 
ingly high, although the oldest specimens have hardened less 
than three months. The results are somewhat irregular, and 
the conclusion to be drawn concerning the best size for the 
aggregate is not very clearly brought out. The one-inch trap 
gives uniformly good results, as do the mixtures of two or 



'Tests of Metals," 1898. 



COMPRESSIVE STRENGTH CONCRETE 



313 



more sizes. The trap rock gives a higher result than the gravel, 
the mortar being sufficient to fill the voids in the trap, and in 
excess for the gravel. 

TABLE 134 

Compressive Strengths of Rich Concretes at Different Ages 
TESTS OF TWELVE-INCH CUBES 



CHARACTER OF AGGREGATE. 


WT. PER 
CU.FT.OF 
CONCRETE 
WHEN 
ABOUT 1 
Mo. OLD, 
IN LBS. 


COMPRESSIVE STRENGTH, POUNDS PER 
SQUARE INCH, AT AGE, DAYS, 


7-8 


19-23 


29-34 


61-76 


Trap \" 


148.6 
148.5 
159.8 
159.2 
160.2 
158.4 
159.8 


1391 
1900 
3390 
3189 
2400 
2800 
2800 


2220 
2769 
4254 
4006 
4143 
3786 
4156 


2800 
3200 
4917 
4562 
4140 
4349 
4800 


5021 

5272 
2583 
4523 
4544* 
5542 


4^ ....... 


U" 


2*" 
J"-l, 2i"-2. . . . 
i"-l, 1"-1, 2J"-1 . 


Mean results, trap rock alone 




2553 


3619 


4110 


4581 


Pebbles f" . 

" H" . . 


148.2 
151.0 
150.3 
147.8 


1298 
2276 
1994 
1486 


2600 
3186 
3023 
2676 


2992 
3817 
3800 
3000 


3870 
4018 
3490 
3800 


r-i, ir-2. . . 
$"-i, i"-i, ir-i 


Mean, pebbles alone . . . 




1764 


2871 


3402 


3794 



NOTES: Tests made at Watertown Arsenal, "Tests of Metals," 1898. 

All concretes composed of one cubic foot of Alpha Portland cement, 
weight 96| to 106 Ibs. per cu. ft., one cu. ft. of bank sand, 
weight 93| to 104 Ibs. per cu. ft., and 3 cu. ft. of aggregate, 
weighing from 93 to 105 Ibs. per cu. ft. 

The size of aggregate indicated gives the larger of the two screens 
used in separating it into different sizes; thus, "f inch" 
means passing f inch mesh and retained on \ inch mesh. 

The compressive strength of twelve inch cubes of one-to-one mor- 
tar alone was 3,833 Ibs. per sq. in. at seven days, and 4,800 
Ibs. per sq. in. at seventy-five days. 

430. In 1896-97 Mr. A. W. Dow 2 prepared a number of 
twelve-inch cubes of concrete for the Engineer Commissioner 



1 Not fractured. 

2 Report Operations, Engineer Department, District of Columbia, 1897.. 
Also Baker's "Masonry Construction," p. 112 r. 



314 



CEMENT AND CONCRETE 



of the District of Columbia. These cubes are of interest as 
showing the strength of natural cement concrete as well as 
Portland, and the results are abstracted in Table 135. 

TABLE 135 

Compressive Strength of Concrete 

TESTS OF TWELVE-INCH CUBES FOR THE ENGINEER COMMISSIONER OF THE 
DISTRICT OF COLUMBIA 



REF. 


COMPOSITION OF CONCRETES. 


PER 

CENT. 
VOIDS IN 
AGGRE- 
GATE. 


CRUSHING 
STRENGTH, LBS. 
PER SQUARE 
INCH AT ONE 
YEAR. 


Cement. 


Sand. 


Broken Stone. 


Gravel. 


Port- 
land. 


Natural. 


Coarse. 


Average. 


Average. 


Small. 


1 
2 
3 
4 
5 
6 


1 
1 
1 
1 
1 
1 


2 
2 
2 
2 
2 
2 


6 


6' 
6 1 

3 
4 






45.3 
45.3 
39.5 
29.3 
35.5 
36.7 


1850 
3060 
2700 
2820 
2750 
2840 


829 
915 
800 
763 
841 
915 


6 


' 3 ' 
2 



NOTES: Materials: 

Cement, Portland, "Atlas" (American), 104 Ibs. per cu. ft/.; Natural, 

"Round Top," 70 Ibs. per cu. ft. 
Sand, 15 per cent, retained on No. 8 mesh, 75 per cent, between 8 and 

40 mesh, 10 per cent, passing 40 mesh. Sand was used damp, and 

weighed in that condition 90 Ibs. per cu. ft. 

Stone, Bluestone, "Average," 93 percent, between J inch and 2 inches. 
"Coarse," 89 per cent, between 1 inches and 2| inches. 
Gravel, "Average," 90 per cent, between inch and 1 inches. 

"Small," 90 per cent, between inch and f inch. 
Granolithic, 92 per cent, between T V inch and inch. 
Mixing, thorough by experienced man. 

Tamping, light, in 4 inch layers, just sufficient to bring mortar to surface. 
Storage, cubes thoroughly wet twice a day. 
Age of specimens when broken, one year. 

The concretes all contained two parts sand and six parts 
aggregate to one cement, but the character of the aggregate 
varied as shown. The natural cement concrete gave from one- 
quarter to one-third the strength of the Portland concrete. 
The best result seems to be given by the average size broken 
stone, which was in reality a mixture of various sizes, ninety- 



1 Mixture of one part granolithic size to one of concrete stone. 



COMPRESSIVE STRENGTH CONCRETE 



315 



three per cent, of it being retained on a one-third inch mesh 
and passing a two-inch mesh. The mortar was probably in- 
sufficient to fill the voids in the stone for the first three cubes 
in the table, and under these conditions the gravel, with its 
smaller percentage of voids, makes a good showing. This illus- 
trates what we have already said, that the relative value of 
broken stone and gravel for aggregate depends upon the pro- 
portion of mortar used. 

TABLE 136 

Compressive Strength of Concrete. Portland Cement 

TESTS OF SIX-INCH CUBES OF VARIOUS MIXTURES 



|i 




PAUTS BY VOLUME TO 


CRUSHING STRENGTH, POUNDS PER SQUARE 


fc 

K 


ONE CEMENT. 


INCH, AT AGE OF, 


H 






h 

H 

A^ 


Sand. 


Gravel. 


Broken 

Stone. 


7 days. 


30 days. 


90 days. 


1 











3412 


5318 


6140 


2 





3 1 





1077 


1908 


2517 


3 


1 


2 


2* 


1430 


2215 


2903 


4 


2 


2 


3 


420 


2117 a 


1324 


5 


2 


3 


4 


640 


1199 


1290 


6 


2* 


5 





566 


1385 


1609 


7 


2* 





5 


739 


2033 


1783 


8 


2* 


2 i 


24 


792 


1482 


2014 


9 


3 








767 


1345 


1409 


10 


3 


3 


4 


714 


1028 


1848 


Means, Actual 


1056 


2003 


2284 


Means, Per Cent 




46 


88 


100 





NOTE: Results of Messrs. Ketchum and Honens. 

431. The results in Table 136 were obtained by Messrs. R. 
B. Ketchum and F. W. Honens at the laboratory of the Uni- 
versity of Illinois, 3 and illustrate the rate of gain in strength 
of several mixtures. The cement used was Baylor's Portland, 
fine and of good quality. The sand and gravel were composed 
principally of silica, with 10 to 30 per cent, of limestone. About 
60 per cent, of the sand passed a "number thirty " sieve. The 
unscreened gravel had about 42 per cent, caught on a " num- 
ber five" sieve and eighteen per cent, of it passed a "number 



1 Unscreened. 

7 Result irregular. 

8 Technograph, 1897-98. 



316 CEMENT AND CONCRETE 

thirty." Except in one mixture, however, the gravel and 
broken stone were screened, and only that portion passing a 
two-inch ring and retained on a " number five" sieve was used. 
The stone was a magnesian limestone. 

The concrete was mixed dry, so that considerable tamping 
was required to bring water to the surface. The cubes were 
first kept under a damp cloth for one day, immersed six days, 
and then stored in air in a room until broken. In crushing, 
"the direction of the force applied was parallel to the tamped 
surface." 

432. Each result in the table is the mean of six specimens. 
Comparing number 2 with number 9 indicates that the strength 
obtained with one part cement to three parts unscreened gravel 
is much higher than with mortar of one part cement to three 
parts sand. Comparing 9 and 10 indicates that seven parts 
gravel and stone may be mixed with one-to-three mortar and 
give higher strength than the mortar alone. A comparison of 
6, 7, and 8 shows that in case there is sufficient mortar to fill 
the voids in the aggregate, angular fragments give a somewhat 
higher strength than rounded ones, but that a mixture of broken 
stone and gravel is better than either alone. One of the most 
important points brought out by the tests is that the strength 
at seven days is 46 per cent., and at thirty days is 88 per cent., 
of the strength attained at three months. 

ART. 55. CINDER CONCRETE, ETC. 

433. For such purposes as floors for buildings, cinders are 
used in concrete to a considerable extent on account of their 
light weight. Cinder concrete weighs only from two-thirds to 
three-fourths as much as broken stone or gravel concrete. 
The strength, however, is correspondingly less, and whether for 
a given strength a floor may be made lighter by the use of 
cinders will depend upon the conditions of use and the charac- 
ter of the reinforcement. 

Table 137 gives the results of the tests of eight-inch cylinders, 
fifteen inches high, made by Mr. George Hill. 1 In these cylin- 
ders, cinders, broken stone, and gravel were used as aggregates. 
The character of the materials is shown in the foot-note of the 



1 Trans. Am. Soc. C. E., Vol. xxxix, p. 632. 



STRENGTH CINDER CONCRETE 



317 



table. As the specimens were but one month old when tested, 
the results are low, but since in the construction of floor arches 
the centers are usually removed in less than one month, the 
strength developed in a short time has a special interest. 

TABLE 137 
Compressive Strength of Concrete about One Month Old 

TESTS OP CYLINDERS, EIGHT INCHES DIAMETER, FIFTEEN INCHES HIGH 





PROPORTIONS BY VOLUME. 


AGE, 


COMPRESSIVE STRENGTH, 
LBS. PER SQ IN. 


AGGREGATE. 


Cement. 


Sand. 


Aggregate. 


Days. 


American 
Portland 
Cement. 


Slag 
Cement. 


Cinders. 




3 


6 


33 


246 




it 




3 


6 


18 


292 




u 




2 


5 


33 


305 




( 




2 


5 


33 


464 




u 




2 


5 


32 


490 




tl 




2.4 


6 


32 


590 




u 


1 


1.7 


4.2 


30 




342 


II 


1 


1.6 


4 


30 




330 


II 


1 


1.6 


-4 


31 




765 


II 


1 


1.6 


4 


31 




765 


Stone. 


1 


3 


6 


30 


398 






1 


2.4 


4.1 


30 


503 






1 


2.4 


4 


33 


. 


645 




1 


2.4 


4 


30 




730 


Gravel. 


1 


3 


6 


30 


oi? 


618 




1 


2.4 


4.8 


30 


. 


650 




1 


2 


7 


25 


880 






1 


1.6 


6.5 


31 




730 


Stone and gravel, ) 
graded . . . . $ 


1 


2 


10 


30 


625 





NOTES: 

Cement, American Portland, tensile strength 624 Ibs. per sq. in., neat, seven 
days. 

Slag cement, a little less than 400 Ibs. per sq. in., neat, seven days. 
Sand, clean, sharp, bank sand of mixed sizes, from moderately fine up to 

some pebbles size of bean. 
Cinders, ordinary steam, dust to inch size. 
Stone, broken trap, nearly uniform size passing 1 inch ring. 
Gravel, clean, washed, | in. to 1^ in. 
Abstract of tests by Mr. George Hill, M. Am. Soc. C. E., Vol. xxxix, p. 632. 

It is evident that cinder concrete should not be loaded very 
heavily within a month after made. The gravel gives a better 
result than broken stone. 



318 



CEMENT AND CONCRETE 



434. In Table 138 are given the results of some tests of 
twelve-inch cubes of cinder concrete made at the Watertown 
Arsenal for the Eastern Expanded Metal Companies. Steam 
cinders were used, practically as they came from the furnace, 
only the larger clinkers being broken. Two proportions were 
used and the specimens were broken at one month and three 
months. It is seen that the one-one-three mixture is about 
twice as strong as the one-two-five with all brands. The varia- 
tions between the several brands are also very great. 

TABLE 138 

Crushing Strength of Cinder Concrete. Portland Cement 
TESTS OF TWELVE-INCH CUBES AT WATERTOWN ARSENAL 



BRAND 

OF 

CEMENT. 


STRENGTH, POUNDS PER SQUARE INCH. 


Mixture A, 1-1-3. 


Mixture B, 1-2-5. 


Age of Specimens. 


Age of Specimens. 


1 month. 


3 months. 


1 month. 


3 months. 


A 
B 
C 
D 


2329 
1602 
1438 
1032 


2834 
2414 
1890 
1393 


940 
696 
744 
471 


1600 
1223 
880 
685 



NOTES: 

Concretes mixed rather dry, 10 to 12 pounds of water per cubic foot of 

concrete. 

Mixture "A," one part cement, one part sand, three parts cinders. 
Mixture "B," one part cement, two parts sand, five parts cinders. 
Weight of concrete, 104 to 116 pounds per cubic foot. 
Tests for Eastern Expanded Metal Companies. Data from "Tests of 
Metals," 1898. 

435. Table 139 gives the results of other tests in the same 
series, using a single brand of cement and five mixtures, the 
richest containing three parts cinders and one part sand to one 
volume cement, and the poorest six parts cinders and three 
parts sand to one cement. The weight per cubic foot of the 
several concretes is also given. 

Tests of cinder concrete prisms made by the late Prof. J. B. 
Johnson at Washington University l indicated that the mixture 



; Materials of Construction," p. 628. 



CONCRETE WITH CLAY 



319 



containing one part sand and three parts cinders to one volume 
cement gave the highest strength, or about twelve hundred 
pounds per square inch, at one month. The same mixture gave 
the highest values for the ratios of strength to cost, and of 
strength to weight per cubic foot. 

TABLE 139 

Crushing Strength of Cinder Concrete. Various Proportions with 

Germania Portland Cement 
TESTS OF TWELVE-INCH CUBES AT WATERTOWN ARSENAL 



PROPORTIONS IN CONCRETE. 


WEIGHT PER 


CRUSHING STRENGTH, 
POUNDS PER SQUARE INCH, 




Cu. FT. AT 98 


AT AGE, 




TO 102 










DAYS, POUNDS. 






Cement. 


Sand. 


Cinders. 




29 to 39 days. 


98 to 102 days. 


1 


1 


3 


110.4 


1466 


2001 


1 


2 


3 


112.8 


1098 


1634 


1 


2 


4 


107.9 


904 


1325 


1 


2 


5 


105.3 


769 


1084 


1 


3 


6 


103.5 


529 


788 



NOTE: Tests for Eastern Expanded Metal Companies, "Tests of Metals," 
1898. 

436. Clay in Concrete. The effect of clay on the tensile 
strength of mortars has already been shown (Art. 49). Aggre- 
gates available for concrete frequently contain a certain amount 
of clay, and the question arises whether such aggregate must 
be washed, or whether certain small percentages may be per- 
mitted in the concrete, using, perhaps, a trifle richer mortar. 
The results in Table 140 were made to determine the effect of 
clay on the crushing strength of concrete. 1 

The test specimens were six-inch cubes, and were broken 
when one week to twelve weeks old in an Olsen machine. The 
proportions were two parts sand and six parts gravel by weight 
to one of Portland cement, or two parts sand and four parts 
gravel by weight to one of natural cement. The clay is ap- 
parently expressed as the per cent, of total aggregates. It is 
seen that while six or twelve per cent, clay retards the harden- 
ing of both Portland and natural cement concrete, the strength 
of the Portland concrete after four weeks is increased by six per 



1 Tests by Messrs. J. J Richey and B. H. Prater, Technograph, 1902-03. 



320 



CEMENT AND CONCRETE 



cent, clay, while at the same age the strength of the natural 
cement concrete is not greatly affected. The ramming of con- 
crete is facilitated by the presence of a small amount of clay, 
but larger amounts may render the mass sticky and difficult 
to ram. 

TABLE 140 
Effect of Clay on Crushing Strength of Concrete 

SIX-INCH CUBES 



CEMENT. 


PROPORTIONS BY 
WEIGHT, No. PARTS 
TO ONE CEMENT. 


AGE OF CUBES 

WHEN 


CRUSHING STRENGTH, POUNDS PER SQ. 
IN.; CLAY AS PER CENT. OP CONCRETE, 




Sand. 


Gravel. 


BROKEN. 





6 


12 


Port. 


2 


6 


1 week 


1030 


1001 


692 


14 


2 


6 


4 weeks 


1398 


1525 


1287 


II 


2 


6 


12 " 


2110 


2760 


1865 


Nat. 


2 


4 


1 week 


208 


131 


81 


it 


2 


4 


4 weeks 


428 


364 


283 


ti 


2 


4 


12 " 


786 


722 


480 



ART. 56. THE MODULUS OF ELASTICITY OF CEMENT MORTAR 

AND CONCRETE 

437. With the increasing use of concrete and steel in com- 
bination, the modulus of elasticity of cement mortar and con- 
crete assumes a new importance, since the ratio of the stresses 
in the two materials depends upon the relative moduli of elas- 
ticity. Some of the earlier determinations of the modulus of 
mortar gave very high values. This may have been due to the 
use of richer mixtures, and the exercise of greater care in the 
manipulation, than are employed in actual construction, and 
also to the fact that the determinations were based upon the 
deformations resulting from the application of very limited 
loads. 

It is now considered that the ratio of stress to strain is not 
constant, even for moderate loads, but that the modulus of 
elasticity decreases with increasing stress, and this fact is brought 
out in the following tables. The tests cited bring out a wide 
range of values for concretes and mortars made from a variety 
of sand and aggregate and of various compositions and ages. 

438. Modulus of Elasticity of Natural and Portland Cement 
Mortars. Table 141 gives the modulus of elasticity of mortars 
as determined by tests of twelve-inch cubes at the Watertown 



MODULUS OF ELASTICITY 



321 



PROPORTIONS CEMENT TO SAND IN MORTAK BY VOLUME. 


t 


Modulus of Elasticity in Thousands, between Loads, in Pounds per Square Inch, of 


100-1000 


o 

t- 

QO fN 


1 


, -.. . . uo iO 

.... 2S 
. . . . CO <N 


2 


I 


o . 
01 ' 

.... II 


100-1000 


00 OS 


100-600 


<N t-- tO >O O 

Tt< QO CO O O . 

SS^ * 3 . 


2 


i 


(M CO CO 

i . 

|H . . . i-l <M 


! 


^^ ^^ ^ CO O CO 
CO O 00 CO O O 
CO <N i-l <M O Oi 
i-H r- 1 rH i 1 CM i 1 


! 


1-1 O O ^ i-l CO 

t~ o 10 cq t~ 01 

rt< <M (M CO *O OS 

1-1 1-1 1-1 T-H CO I-l 


3 


! 


OS- O CO il " " 
OO (M CO 00 

i I i 1 i 1 . 


100-1000 


Ol O O OS 
O t-- O O 

^ oo oo t^ 

1 1 1 1 II II . 


I 


t^ O CO t- . 

CO O t~- Tfl 

CO iO (N I-H ' 
rH cq <M (M . . 


6 j 

ml 

1 1 


.2 |g d .2 

!! ^1 

"pnW^ M S 


CEMENT. 


Brand. 


^o g 
- - '&- 

1- - - s- 

K 


2 

2 


1,,, 1 

r^ r 



322 



CEMENT AND CONCRETE 



Arsenal. 1 These specimens were a portion of those prepared 
by Mr. Rafter, the compressive strength being given in Table 
129. As each value is the result of but one determination, the 
results are not as regular as might be desired. In general the 
strength and the modulus decrease together as the amount of 
water used in mixing is increased. The modulus also decreases 
with the strength as the proportion of sand increases. 

439. Modulus of Concretes One Month to Six Months Old. - 
In the compressive tests of twelve-inch concrete cubes made 
for Mr. George A. Kimball and abstracted in Table 130, many 
of the specimens were also gaged for compression under load to 
determine the modulus of elasticity, and a part of the results 
are presented in Table 142. 

TABLE 142 
Modulus of Elasticity of Concrete 

TESTS MADE ON TWELVE-INCH CUBES OF PORTLAND CEMENT CONCRETE AT 
WATERTOWN ARSENAL FOR BOSTON ELEVATED RAILROAD 





CONCRETE 1-2-4. 


CONCRETE 1-3-6. 


CONCRETE 1-6-12. 


AGE 








OF CUBES 
WHEN 
CRUSHED. 


Modulus of Elasticity in Thousands, between Loads, 
in Pounds per Square Inch, of 




100-600 


100-1000 


1000-2000 


100-600 


100-1000 


1000-2000 


100-600 


100-1000 


7 days 


2592c 


2053c 


1351a 


1869c 


15296 








1 month 


2662c 


2444c 


1462c 


2438 


2135 


1219a 


1376 




3 months 


3670 


3170 


2157 


2976 


2656 


1805 


1642 


1363 


6 months 


3646 


3567 


2581 


3608 


3503 


1868 


1820 


1522 



NOTES: Results marked "a" are means of five or more tests of one brand. 
Results marked "b" are means of five or more tests on each of 

two brands. 
Results marked "c" are means of five or more tests on each of 

three brands. 
Results not marked are means of five or more tests on each of 

four brands, two American, two German. 
For compressive strengths of similar cubes, see Table 130. 

It is seen that the modulus increases with the age and rich- 
ness of the specimens, and decreases as the load increases. For 
one-two-four concrete the modulus at one month, for loads 
between a hundred and a thousand pounds, is about two and 



Tests of Metals," 1899. 



MODULUS OF ELASTICITY 323 

one-half million, and for six months, three and a half million. 
The corresponding values for the one-three-six concrete are two 
million and three and one-half million. When the ultimate 
strength is approached, the modulus of elasticity decreases 
rapidly, and between loads of one thousand and two thousand 
pounds per square inch, the richest concrete gives only about 
one and one-half and two and one-half million at one month 
and six months, respectively. 

440. Modulus of Concrete Dependent on Richness of Mortar. 
The results in Table 143 are abstracted from the extensive 
tests made at the Watertown Arsenal for the State Engineer 
of New York. Although several brands were tested, the results 
in the table are from one brand only, namely, "Wayland" 
Portland. These cubes were all stored in the same manner, 
namely, in water three to four months, and then buried in damp 
sand until broken at the age of twenty months. The mean 
ultimate strengths of similar cubes stored according to four 
methods are given in Table 132. 

Since in all of these mixtures the quantity of mortar was a 
given percentage, either thirty-three or forty, of the volume of 
aggregate, the effect of the richness of the mortar may be studied. 
While the proportional strengths of the concretes made with 
mortars containing from one to five parts sand are 100, 77, 52, 
44, and 38, the corresponding proportional moduli of elasticity 
are 100, 92, 77, 60, and 55, the modulus decreasing less rapidly 
than the strength, with the addition of sand. 

441. Gravel and Trap Aggregates. Table 144 gives the re- 
sults of the determinations of the modulus of elasticity of con- 
crete specimens made and tested at the Watertown Arsenal, 1 
the strength of which was given in Table 134. As these are all 
rich concretes, the moduli and the strengths are high. The 
values of the modulus for the gravel concretes are about 70 
per cent, of those for the trap, but the strengths of the gravel 
concretes are in general about 80 per cent, of those obtained 
with concretes having trap aggregate. In a general way, how- 
ever, the modulus and strength vary together. 

442. Modulus of Cinder Concrete. The modulus of elas- 
ticity of cinder concrete prepared for the Eastern Expanded 



'Tests of Metals," 1898. 



324 



CEMENT AND CONCRETE 



CO -*i CO 
I rH -^1 

CO Oi SO 



<-0 T* CO 

T^ O . Tjl 

O I-H O 



CM O Oi O CO 

CO O QO O rH 

O <N CO (M CO 



CO rH (71 
rH I :O 
CO T* O 






CO ^^ CO 
OC t- CO 
O -^ O 



CO Oi ' rH 
CO CN . rH 



O O t ^ 

01 <M GO 00 rH 

rH (M rH rH (N 



t^ t- t Oi rH 



CO rH 

^ CO 

rH t~- CO CM 

Cq rH rH CM GN 



O O ' 
O O 



CO t- CO 00 t- CO 

t^ co oo t^ cc t 

(M CO I- I- rH ?q 

(M rH rH CM -HH (M 










PROPORTION 
CEMENT TO 
SAND IN MORTAR. 




axvoaHoov 

-iN3 

sv UVXHOJ^ 



MODULUS OF ELASTICITY 



325 



TABLE 144 

Modulus of Elasticity of Rich Concretes with Gravel and 
Trap as Aggregates 

TESTS OF TWELVE-INCH CUBES AT WATERTOWN ARSENAL. 1-1-3, ALPHA 

CEMENT 



CHARACTER OP AGGREGATE. 


MODULUS OF ELASTICITY IN THOUSANDS, BE- 
TWEEN LOADS OF 100 AND 1,000 
POUNDS PER SQUARE INCH, AT AGE, DAYS, 


7-8 


19-23 


29-34 


61-76 


Trap \" S 


1875 
3214 
4091 
4500 
3214 
5000 
3461 


2500 
2368 
6429 
5625 
5625 
4500 
4500 


3750 
5625 
5625 
5000 
4500 
7500 
5625 


3750 


3ff 


1" 

H" 

2*" 
*"-!, 2^"-2 .... 
$"-1, 1"-1, 2"-l . . 


5625 
4091 
7500 
5625 
7500 


Mean Results, trap rock alone 


3622 


4507 


5375 


5682 


Pebbles f " 


1800 
3750 

2812 
1800 


3750 
4091 
3461 
3214 


3461 
3750 
4091 
3461 


3214 
3000 
4500 
3214 


1$" 


" f"-l, l"-2 . . . 

" I"- 1 * t"-i. H"-i . 


Mean Results, pebbles alone 


2540 


3629 


3691 


3482 



NOTES: 

Tests at Watertown Arsenal, "Tests of Metals," 1898. 

For crushing strength of these concretes, see Table 134. 

The modulus of elasticity of twelve-inch mortar cubes, one volume cement 
to one volume sand, was, for loads between five hundred and one thou- 
sand pounds per square inch, 3,461,000 at seven days and 5,000,000 at 
seventy-five days. 

Metal Companies is given in Table 145. The results are seen 
to be low, as is the crushing strength. The permanent set in 
five-inch gaged length for a load of six hundred pounds per 
square inch is also shown in the table. 



326 



CEMENT AND CONCRETE 





H 

H 














||| 


*g 












^ 


G 


0000 


i 1 


QO Tfi C5 t CO 


CO 




ISP 


|S 


0000 


8 


00000 







S -, & 


fg 

O02 


00000 


o 


000000 


o 


3 


fe 










a 

-5 

T- 


W 

H 


f 


i 








J 


fe 


Z 
< a 

o 


i 


05 O i-H 
(M C5 iO . . 

^ i-H CO 




:::.::: 


H 
C 
fe; 


o 


5 a 


o 








1 


H 
_ tf 





rt 








G 


fi K 

8 5 


fcl 


g 








'c 

e 


k 

fl ^ 

3 s 


|| 

is 


00 and 1 


*O (7^1 QO O CO 

CM CO 1-1 T-I 1-1 


1 


GO iO CO -^ 
(7^1 Oi 00 QO ' 




_c 


Js " 


o 


1 








'( 


i 


S w 


i 








i 

1 


2 5 n 


gj 

p g 


d 

H 


|co||t2 


(N 


i H t^- t^- O O QO 

t^ CO 00 C* CO O 


GO C 


B 1 
I * S 


ga 





S5S28 


<M 


^ O O I~H l^* CO 
rH i I T I i 1 


^ ~ 

I 


3 S a 
<^ o g 

H 3 g 

rt H 

W r 


MPRESSIVE 
FRENGTH. 


|| 


O QO O t^" C>i 

1>- CO ^O CO OO 

(M (N (M 1-1 i i 


OS 


HUH 


T 

" ! 

I 


*o 


OOQ 


^ ^ 










P 


O 


CO 








c 















1 

: 


3 a 




ao 








t 

: 


d fc 


s 


a> 


CO CO CO CO CO 


CO 


U5 KQ >-0 10 10 


>0 ^ 


i J 







5 








I 


a 


EH 2 












p 


is 


c' 








1 


h 


o o 


02 













aO 












BB 

S 


| 
p 


C 








1 




j^ 


D 










H 


> 


a 








^ 




a 














ii 

~~ - 


p 

2 




fl 

- 




S 




03 S 

|o 


m 




i 




1 



CHAPTER XVII 

THE TRANSVERSE STRENGTH AND OTHER PROPERTIES OF 
MORTAR AND CONCRETE 

ART. 57. TRANSVERSE STRENGTH 

443. TENSILE, TRANSVERSE AND COMPRESSIVE STRENGTHS 
OF MORTAR COMPARED. The tests given in Tables 146 and 
147 were designed to compare the strengths of cement mortars; 
in tension, bending and compression, and to show the relative 
effect on the three kinds of strength of certain variations in 
manipulation. 

The tensile specimens were briquets of the ordinary form, 
made in brass molds. The transverse and compressive speci- 
mens were made in wooden molds, the bars for transverse tests 
being two by two by eight inches and molded horizontally, 
while the specimens for compressive tests were two-inch cubes. 
Specimens of the three forms were made from the same batch 
of mortar to obviate, as far as possible, variations due to differ- 
ence in gaging. Two cubes, two briquets and one bar were 
usually made from one gaging of mortar. 

The briquets were broken in the usual manner on a Riehle 
cement testing machine. The bars were broken on a home- 
made lever machine. Two fixed knife edges were placed five 
and one-third inches apart, and the breaking stress was applied 
through a third knife edge at mid-span. The lengths of the 
lever arms of the testing machine were in the ratio of one to 
twenty-five, and water was allowed to run gently into a vessel 
at the end of the longer arm. The span of five and one-third 
inches was chosen because at this length the modulus of rup- 
ture, for a two inch square specimen, has the same numerical 
value as the center load applied. 

The cubes were crushed in a crude machine, improvised for 
the purpose, consisting of two iron plates, two hydraulic jacks, 
with hydraulic weighing gage and proper framework. The 
upper plate was fastened to the base of the framework by 

327 



328 CEMENT AND CONCRETE 

means of two bolts which worked freely in the lower plate, and 
the latter was connected to the weighing gage at the top of the 
framework by two bolts which worked freely in the upper 
plate. An hydraulic jack was placed under either end of a 
yoke, at the middle of which was supported the weighing gage, 
While the tensile and transverse tests are doubtless good, the 
compressive tests are lacking in accuracy because of the crude 
method of crushing. 

444. Table 146 shows the comparative tensile, transverse 
and compressive strengths of two samples of cement, one of 
Portland and one of natural, with different proportions of sand. 
It is seen that the modulus of rupture, or stress on the extreme 
fiber in transverse tests, computed by the ordinary formula, is 
considerably greater than the strength obtained in direct tensile 
tests. The ratio of the transverse to the tensile strength varies 
from 1.25 to 1.90 for Portland and from 0.95 to 2.19 for natural. 

These tests indicate that the ratio of the compressive strength 
to the tensile strength diminishes with the addition of sand, 
but the reverse has been found to be true in other series of 
tests where the facilities for making compressive tests were 
better. The result obtained here may be attributed to the fact 
that richer mixtures gave cubes with smoother and more regular 
faces, and thus less subject to eccentric loading. The com- 
pressive strength increases between three months and one year 
much more than the tensile and transverse strengths. Tests on 
ten brands of Portland and ten brands of natural showed that 
in general the brands giving the highest strength in tension 
gave also the highest strength in transverse and compressive 
tests. 

445. A few results to show the effect of consistency of the 
mortar on the three kinds of strength are given in Table 147. 
With Portland cement the highest strength in transverse and 
compressive tests is given by a wetter mortar than that giving 
the highest strength in tension, but with natural cement the 
compressive strength is lowered more than the tensile strength 
by an excess of water. All of the specimens were one year 
old when broken. 

446. TRANSVERSE TESTS OF CONCRETE BARS. The effect 
on the strength of concrete of variations in manipulation and 
treatment is most satisfactorily investigated by tests of large 



TRANSVERSE STRENGTH 



329 



$ > 

* 1 

w 



MEAN STRENGTH, POUNDS PER SQUARE INCH, FOR VARYING RICHNESS OF MORTARS, 
PARTS SAND TO ONE OF CEMENT BY WEIGHT. 


"6 

02 

&H 


> 







' O O " ' t^ *O 

- <N <M . t- rH 

>O I- CO -^ 


NOTE: Cement P = Portland, Brand R ; N = Natural, Brand An; Sand, crushed quartz, passing No. 20, retained on 
No. 30 sieve. 


X 



H 


' so Gi ' ' O "* 
. . cc o - . r- <N 

<N CO rH OJ 


g 
a> 
H 


t^ "N O CO 

QO O "t Tt< 

. TH (M . . 1-1 I 1 


1 

I 
1 

H 


IH 

a, 

1 


O QO O O !N iO 
Ci CO rJH TO i I r^ 
O O <M i^ ' & SO ' 


I 
H 


t~ t^ *-< C3 " CC I ' 

^ Ci 't 00 C: ?M 

(M CO >O iO i I CO 


1 


(M t^ 00 Ci . O t 

00 t- CO t>- i I "^ 

i I S^l CO CO t (N 


a 

02 

PL, 
| 


I 


o o o o t- 

(N >O Tfl CO O <N 
** ' Ci ' CO r- r-i <M 

i 1 . rH . T 1 rH 


30 

B 

C-| 


t^ ' ^ ' QO <M to (M 

. SO t~ ,-( SO t~ 
^ t- <M Tt< ^ 


30 
1 


-* .'-H . ^ 00 Ci CO 

Ci Ci SO <M r-l rH 

(M . Tj< . i I CO CO 


T5 

1 

1 




& 


(M CO '-O to -CO ' 
t~ * iQ (N rH 
SO CO <M OO O 'CO ' 
(N CO CO kO . rH . 





t-- >O rH O O ' C5 

rH (N CO -^ CO 
l^> CI rH rH rH SO 


/: 

1 


^H 1C CO .CO . 
CO CO (N Ci 

Tt< SO t~ t- . Tj< . 


Neat Cement. 


2* 
1 


cc r~ t- ^ 

IO rH ^ * CO rH . 
T* ^5 -^ ' Ci CO T 
CO Tt< * rH (M . 


OJ 


\Q t- ' t^ O 
rH CO "^ CO t> rH 
rH (M CO (M TjH lO 
rH rH rH . 


H 


CO Ci CO " CO rH GO 
O SO I- rH (M (N . 


8JJ3WID3dg 
SO 

aoy 


e3 e8 2 K eS e3 2 ^' 

^Tj a >> 73 TS a >> 

t^ CO CO rH t- QO CO rH 


iNtawaQ 


^pHPn'^ 2;^ 


h 

9 



T-KNCO-* 0^QO 



330 



CEMENT AND CONCRETE 



sized specimens either in compression or bending. In the prep- 
aration of such large specimens the conditions of actual con- 
struction may be closely reproduced, and the results, although 
likely to be quite irregular, as the strength of concrete in struc- 
tures is not uniform throughout, are nevertheless very valu- 
able On account of the expense connected with such tests, 
the number of specimens is usually so limited that the natural 
irregularities in strength mask the true conclusions. 



TABLE 147 

Comparative Tensile, Transverse and Compressive Tests, 
of Varying Consistency of Mortar 



Effect 







WATER 


MEAN STRENGTH, POUNDS PER 


TRANSVERSE AND 
COMPRESSIVE 






AS PER 


SQUARE INCH. 


STRENGTH AS PER 


REF. 


CEMENT. 


CENT. OF 
DRY 




CENT. OF TENSILE. 






INGRE- 
DIENTS. 


Tensile. 


Trans. 


Comp. 


Trans. 


Comp. 


1 


P 


9 


516 


837 


1731 


162 


335 


2 


P 


12 


533 


987 


2173 


185 


408 


3 


P 


15 


467 


850 


2498 


180 


533 


4 


P 


18 


461 


966 


2823 


209 


612 


5 


P 


21 


430 


1022 


2487 


239 


578 


6 


N 


12 


272 


447 


2270 


164 


835 


7 


N 


14 


325 


516 


2141 


158 


659 


8 


N 


16 


319 


519 


1481 


163 


464 


9 


N 


20 


304 


509 


1512 


167 


497 


10 


N 


24 


315 


462 


1317 


147 


418 



NOTES: Cement, P = Portland, Brand R; N = Natural, Brand In; 

Sand, "Point aux Pins," pass No. 10 sieve. Age of specimens, 
one year. Two parts sand to one cement by weight. 

In Tables 148 to 156 are given some of the results obtained 
in testing over two hundred concrete bars at St. Marys Falls 
Canal. l"he molds for making the concrete bars were ten 
inches square by four and one-half feet long inside. The con- 
crete was rammed into the mold with a light wooden rammer. 
The bars were, in general, covered with moist earth soon after 
completed, to await the time of breaking. To break them they 
were supported on knife edges placed four feet apart, and the 
load was applied at mid-span through an iron bolt laid across 
the bar. In the earlier tests a direct load was imposed by means 
of a platform which was gradually loaded with one-man stone, 
but in the later tests the load was applied by means of hydraulic 



TRANSVERSE STRENGTH CONCRETE 



331 



jacks, an, hydraulic gage being used to measure the force. In 
many cases the half bars were again broken at a later date 
with a twenty-inch span, as shown in the tables. 

447. Variations in Richness of Mortar. In Table 148 sev- 
eral concretes made with mortars having different proportions 
of sand are compared, and the results of briquet tests on similar 
mortars are also given. Although the briquets were not broken 
at the same age as the bars, the tests on the latter at the differ- 
ent ages show that they were not gaining strength rapidly, 
and the results may therefore be compared without serious 
error. 

TABLE 148 
Transverse Tests of Concrete. Variations in Richness of Mortar 











2 










olS 


- 5j fe 


MODULUS OF RUPTURE. 


No. 
BARS. 


DATE 
MADE. 


CEM- 
ENT. 


III 


HSis 


Four Foot Span. 


Twenty Inch Span. 








U* 


lifts 

o a M 






No. 
Tests. 


Age. 


Mean. 


. No. 
Tests. 


Age. 


Mean. 




Mo. Da. 










Yr. Mo. 






Yr. Mo. 




76-77 


11-2 


Port. 





717 


2 


1 7 


593 


4 


2 9 


600 


78-79 


<4 


u 


1 


790 


2 


i 


689 


4 




698 


80-81 


II 


H 


2 


595 


2 


i 


538 


4 




577 


82-83 


14 





3 


432 


2 


' 


489 


3 




415 


84-85 


11-3 


I* 


4 


335 


2 


i 


379 


4 




385 


86-87 


H 


(t 


5 


252 


2 


l 


284 


4 




316 


88-89 


If 


" 


6 


218 


2 


i 


262 


4 




279 


90-91 


11-4 


Nat. 


1 


483 


2 


i 


420 


4 




450 


92-93 


ii 


" 


2 


396 


2 


' i 


332 


4 




387 


94-95 


14 


ii 


3 


330 


2 


1 


240 


4 




224 


96-97 


(I 


" 


4 


237 


2 


i 


186 


4 




205 



NOTES: 

Portland, Brand R, Sample 82 M. 

Natural, Brand Gn, Sample 83 T. 

Sand, from '' Point aux Pins" (river sand). 

Stone, Potsdam sandstone, retained on inch square mesh, and no pieces 

larger than 3 inches in one dimension. 
Amount mortar used in each case equal to voids in stone measured loose, 

except in case 1-2 natural, when mortar exceeded voids by seven per 

cent. 
The fracture showed concrete very compact in nearly all cases. 

The results obtained with natural cement show that the 
tensile strength of the mortar in pounds per square inch was 
greater than the modulus of rupture obtained for the concrete. 



332 



CEMENT AND CONCRETE 



This is also the case with rich mortars of Portland cement, 
but for Portland mortars containing more than three parts sand 
to one of cement the concrete ^gives the higher result. The 
strength of the concrete with one-to-four mortar is fifty-five per 
cent, of the strength with one-to-one mortar for Portland, and 
forty-five per cent, for natural. The decrease in strength due 
to larger proportions of sand in the mortar is usually greater 
than the decrease in cost. 

TABLE 149 
Transverse Tests of Concrete. Variations in Quantity of Mortar 







* 
5*| 


-_- - 
o * 


MODULUS OP RUPTURE. 


No. 
BAR. 


DATE 

MADE. 


111 


<3|?s 


Four Foot Span. 


Twenty Inch Span. 








EH P5p2 










<! i^ 




No. 

Tests. 


Age. 


Mean. 


No. 
Tests. 


Age. 


Mean. 




Mo. Da. 














Yr. Mo. 




42 


7 3 


31 


88 


1 


lyr. 


247 


2 


1 10 


363 


37-40 


7 1 


38 


92 


2 




284 


3 


" 


447 


38-41 


7 1 


47 


104 


2 


ti 


350 


4 





596 


39-43 


7 1,3 


60 


112 


2 


It 


346 


4 


u 


589 



NOTES: 

Cement, Portland, Brand R, Sample 64 T. 

Sand, " Point aux Pins," three parts by weight dry to one cement. 
Stone, Drummond Island limestone, passing 1 inch slits and retained on 
| inch slits. 

448. Variations in Amount of Mortar Used. Bars 37 to 
43, Table 149, were all made with the same kind and quality 
of stone and the same quality of mortar, three parts sand to one 
cement by weight, but the amount of mortar varied; thus, in 
bars 41 and 38 sufficient mortar was used to fill the voids in 
the stone, while the bars above were deficient in mortar, and 
those below contained an excess. It is seen that the highest 
result is given by the bars in which the mortar was just suf- 
ficient to fill the voids in the stone, though the bars containing 
an excess of mortar gave practically the same result, while a 
deficiency of mortar resulted in decreased strength. 

449. Variations in the Amount of Sand for Fixed Quantities 
of Cement and Stone. In Table 150, bars 68 to 75 were all 
made with the same kind and quantity of cement and stone, 
but the amount of sand, and consequently the quantity and 



TRANSVERSE STRENGTH CONCRETE 



333 



quality of the mortar, varied. The highest strength is given by 
the concrete in which the weight of the sand was three times 
the weight of the cement; this quantity of sand gave sufficient 
mortar to fill the voids in the stone. The richer mortars, 
though stronger, were deficient in quantity, while four parts 
sand made an oxcess of mortar having a lower strength. 

TABLE 150 

Transverse Tests of Concrete. Variations in Quantity of Sand for 
Fixed Quantities of Cement and Stone 





1 




o 


95 a . 
o J2 o u 


Pi 


MODULUS OF RUPTURE. 







g 


|| 


Q Z 


H fc, O 


^< en <J . 




B 






m 


* 


rag 


OQ W 


^Jfia 


^ W S fc 


Four Foot Span. 


Twenty Inch Span. 


a 

i 


o 


CEMENI 


1* 


.^ O 
Pu 


PI! 


-<oo 






H 

tf 


No. 
Tests. 


Age. 


Mean. 


No. 

Tests. 


Age. 


Mean. 
















Yr.Mo. 






Yr.Mo. 






74-75 


65 


65 


i 


16 


95 


2 


1 8 


299 


4 


2 10 


295 


a 


72-73 


65 


130 


2 


24 


101 


2 


u u 


335 


4 


U 44 


303 


6 


70-71 


65 


195 


3 


32 


104 


2 


u u 


324 


4 


44 44 


354 


c 


68-69 


65 


260 


4 


42 


110 


2 


u u 


322 


4 




321 


d 



NOTES: 

Cement, Portland, Brand R, Sample 768. 

Sand, " Point aux Pins." 

Stone, Potsdam sandstone, screened with f inch mesh, and all pieces larger 

than 3 inches in one dimension rejected. 
Appearance of fracture: a, very porous; 6, many voids; c, some voids; 

d, few voids. 

450. Consistency of Concrete. The bars, the results of 
which are given in Table 151, were made to show the effect of 
the consistency of the concrete on the strength obtained. It is 
seen that the highest strength is given when the consistency 
is such that a little moisture is shown when ramming is com- 
pleted ; the decrease in strength from an excess of water is much 
less than that caused by a corresponding deficiency. The re- 
sults of briquet tests on similar mortar are also given in the 
table, and it appears that the highest result is given by the 
mortar containing the least water, which shows the familiar 
fact that the mortar for concrete should be more moist than 
that which gives the best results in briquet tests. 

451. Value of Thorough Mixing. Bars 182 to 189, Table 
152, were made to show the effect of thorough mixing of the 



334 



CEMENT AND CONCRETE 



TABLE 151 

Transverse Tests of Concrete. Variations in Consistency 











Q M 


MODULUS OF RUPTURE. 




eg 






PRO- 


lg 


We . 




brf 


EH 




CEM- 


PORTIONS. 




S^ W 






O 


w 




ENT, 




|B 


^ 7 s w 


4 Foot Span, 


20 In. Span, 


U 





BAR. 






o 


fa 


13 Months 


2 Years. 


E-i 


i 




Kind. 






if 


& W w 

PS 






i 


a 






Cem- 
ent, 
Lbs. 


Sand, 
Lbs. 


H 


g g p 


No. 
Tests. 


Mean. 


No. 
Tests. 


Mean. 


5 


TENSII 


138-139 


Port. 


120 


237 


0.61 


7.31 


2 


354 


2 


289 


a 


509 


136-137 


it 


120 


237 


0.83 


7.12 


2 


450 


3 


482 


6 


404 


140-141 


!< 


120 


237 


1.03 


7.00 


2 


450 


4 


442 


c 


415 


142-143 





120 


240 


1.16 


7.12 


2 


385 


4 


417 


d 


400 


146-147 


Nat. 


115 


230 


0.83 


7.64 


2 


180 


4 


156 


a 


267 


144-145 


u 


115 


230 


1.03 


7.31 


2 


223 


4 


282 


b 


187 


148-149 


u 


115 


230 


1.16 


7.12 


2 


234 


4 


256 


c 


145 


150-151 


u 


115 


230 


1.35 


7.12 


2 


202 


4 


177 


d 


127 


152-153 


u 


115 


230 


1.51 


7.12 


2 


155 


3 


170 


e 


116 



NOTES: Portland cement, Brand R, Sample M. 

Natural cement, Brand Gn, Sample 88 T. 

Sand, " Point aux Pins" (river sand). 

Stone, Potsdam sandstone, 7 cubic feet to each batch. 

Results in last column give tensile strength at one year of briquets 

made from similar mortar. 
Consistency : a, very dry ; no moisture shown on ramming. 

6, slight moisture appeared at surface after continued 
ramming. 

c, quaked somewhat. 

d, quaked and water rose to surface in ramming, 

e, too wet to ram. 

TABLE 152 
Transverse Tests of Concrete Bars. Value of Thorough Mixing 







MODULUS OF RUPTURE. 


No. BAR. 


MIXING OF CONCRETE. 


Four Foot Span. 


Twenty Inch Span. 






No. 
Tests. 


Age. 


Mean. 


No. 
Tests. 


Age. 


Mean. 


182-186 


Turned once and back 


2 


lyr. 


290 


4 


21 mo. 


373 


183-187 


" twice " " 


2 





294 


4 


it 


353 


184-188 


" 3 times " 


2 


u 


306 


4 


u 


444 


185-189 


"4 " u u 


2 





328 


4 


" 


474 



NOTES: Cement, Portland, Brand X, 200 Ibs. 
Sand, " Point aux Pins," 600 Ibs. 
Stone, Potsdam sandstone, 15 cubic feet. 



TRANSVERSE STRENGTH CONCRETE 



335 



concrete. Comparing the concrete turned once or twice, and 
back, with that turned three or four times, and back, it is seen 
that the mean strength of twelve tests with the former is 32& 
pounds per square inch, while the mean strength of the same 
number of tests with the more thoroughly mixed concrete is 
388 pounds per square inch, an increase of eighteen per cent. 

TABLE 153 
Transverse Tests of Concrete. Variation in Size of Aggregate 





K 


STONE. 


tei 


fc K H 

- W 


MODULUS OF RUPTURE. 
LBS. PER SQ. IN. 


No. 
BAR. 


KMENT SAS 


Kind. 


Sizes. 


PER 
CENT. 
VOIDS 
IN- 
COM- 


AMOUNT C 

iCT STONE 

CUBIC FE 


ls# 
111! 

< Kr 

rt u 


One Bar, 4 
Ft. Span, 
Age 1 Yr. 


Half Bar, 
20 In. 
Span, Age, 
21 Mo. 











PACT. 


OH 








202 


XR6 


a 


V 


45 


3.75 


3.75 


259 


367 


199 


i 1 


a 


|V,fF 


43 


3.75 


3.75 


259 


347 


201 


a 


a 


M 


44 


3.75 


3.75 


216 


269 


200 


" 


.{ 


4 each, 
V, F, & M 


[40 


3.75 


3.75 


245 


292 


198 


(C 


'1 


Vsach, 
,F,&M 


i 36 


3.75 


3.75 


288 


390 


196 


XM3 


d 


V 


32 


3.75 




216 


311 


195 


" 


d 


F 


33 


3.75 


3.86 


186 


302 


197 





d 


M 


34 


3.75 


3.75 


131 


208 


194 


" 


-1 


$ each, 
V, F, & M 


i30 






207 


302 







NOTES: 

All mortar, three parts sand to one part Portland cement by weight. 
Quantity of mortar about one-third volume of compact stone. 
Stone: a = Potsdam sandstone; d = gravel. 
Size : K = y , T inch to inch. 

v = i " f > 

F = \ " 1 " 
M = 1 " 2 " 

452. Variations in Size of Stone and Volume of Voids. The 
bars given in Table 153 were all made with mortar composed 
of three parts sand to one of Portland cement by weight. The 
stone for these bars was sorted into different sizes, and these 
were recombined in the proportions indicated in the table. 
The sizes are denoted as follows: that passing one-half inch 
mesh and retained on one-quarter inch mesh, is called V; one- 
half inch to one inch is called F; one inch to two inches, M; 
two inches to three inches, C; and coarse sand, one- ten th. inch 
to one-quarter inch, is called K. 



336 CEMENT AND CONCRETE 

The first five bars were made with broken sandstone, and it 
is seen that the coarsest stone, size one inch to two inches, 
gave the lowest result. The size V, one-quarter inch to one- 
half inch, although containing no smaller percentage of voids, 
gave a much higher strength. The highest result was given 
by the bar made with a mixture of four sizes, the voids in this 
mixture being only thirty-six per cent. 

The bars containing gravel as aggregate indicate that the 
strength decreases as the size of stone and volume of voids 
increase, but a mixture of three sizes gives nearly as good a 
result as the fine grave? alone. Comparing the results with 
similar sizes of the two kinds of aggregate, it appears that the 
broken sandstone gives somewhat better results than gravel, 
notwithstanding that the proportion of voids in the former 
exceeds that in the latter. 

453. Sandstone and Bowlder Stone Compared. The results 
given in Table 154 are from a series of tests made for the Mich- 
igan Lake Superior Power Company by Mr. H. von Schon, 
Chief Engineer, 1 and show the strength of concretes made with 
two kinds of aggregate available at Sault Ste. Marie. Two 
samples of Portland cement, one made from marl and one from 
limestone, a slag cement, and a natural cement, are used in 
these tests. 

The two samples of Portland cement give nearly the same 
result, the slag less than half the strength, and the natural 
quite weak. The ratio of the strength obtained with crushed 
bowlders to that made with sandstone is about 1.6 with 
Portland, and the superiority of the former is shown with all 
cements. 

454. Various Kinds of Aggregate. Table 155 gives the re- 
sults obtained at St. Marys Falls Canal in using various kinds 
of stone. In bars 25 to 30, three kinds of stone are compared. 
The superiority of the Kelleys Island Limestone " shavings " 
from the stone planers is evident. The shape of the pieces may 
have had a considerable influence on this result, the planer 
shavings being flat, or lenticular in form. Bar 34 was made 
with a hard limestone from Drummond Island, 33 with gravel, 
and 31 and 32 with -gravel and stone mixed in equal propor- 



Tests reported by H. von Schon in Trans. A. S. C. E., Vol. xlii, p. 135. 



TRANSVERSE STRENGTH CONCRETE 



337 



TABLE 154 

Transverse Strength of Concrete with Crushed Sandstone and 

Bowlders 



AGGREGATE. 


MIXTURE No. 


MODULUS OF RUPTURE, POUNDS PER SQUARE INCH. 


Portland. 

(Marl.) 


Portland. 
(Rock.) 


Slag. 


Natural. 


Sandstone . . . 


( 

u 
u 


1 
2 
3 
4 
5 


328 
283 
220 
178 
106 


312 

265 
178 
173 
186 


122 
161 
118 
74 
131 


43 
34 
40 

'35' 


Mean, Sandstone 


223 


223 


121 


38 


Bowlder Stone . 

u u 
11 It 
It It 


1 
2 
3 
4 
5 


407 
377 
332 
327 
333 


397 
395 
374 
351 
325 


145 
167 
176 
146 
123 


36 
67 
55 
52 
60 


Mean, Bowlder Stone .... 


355 


368 


151 


54 


Ratio of f Bowlder Stone j 


1.59 


1.65 


1.25 


1.42 


Moduli J Sandstone f 



NOTES: Cross breaking tests of 6 in. by 6 in. by 24 in. bars made for 
Michigan, Lake Superior Power Co. 

Materials: Cement, representative brands of each of four classes. 

Sand, river sand, " Point aux Pins," mostly quartz, 96^ Ibs. per cubic foot. 
Voids, 41.7 per cent. Fineness, 96 per cent, passing No. 20 sieve, 

39 per cent, passing No. 40 sieve. 
Stone, Sandstone, broken Potsdam 1 to 1 inch size. 

Bowlder stone, broken gneiss and granite bowlders, 

1 to 1J inch size. 

Proportion in mortar, 1 part cement to 2.4 parts sand by volume. 
Mixing: Consistency, plastic; cement and sand mixed dry, then wet and 

mixed; mortar added to wet aggregate and concrete mixed by hand. 
Storage : Bars stored in shed, protected from rain, fully exposed to air. 

Age of specimens when broken, sixty days. 
Mixture : 1. Mortar 15 per cent, in excess of quantity required to fill voids. 

2. Mortar 10 per cent, in excess of quantity required to fill voids. 

3. Mortar 5 per cent, in excess of quantity required to fill voids. 

4. Mortar just sufficient to fill voids in stone. 

5. Mortar 15 per cent, in excess of amount required to fill voids in 

stone, but this 15 per cent, excess made with lime instead of 
cement. 



338 



CEMENT AND CONCRETE 



TABLE 155 
Transverse Tests of Concrete. Value of Different Kinds and Sizes of Aggregate 


MODULUS OF RUPTURE. 


Twenty Inch Span. 


9 






Aggregate : a = Potsdam sandstone ; size, f inch to 1 inch. 
6 = Drummond Island limestone ; size f inch to 3 inches. 
c = Kelleys Island limestone; shavings from stone planers; size, | inch to 3 inches, 
d = Gravel. 
e = Broken brick. 


O CO CO CO t- "^ 

*^H to O O OS O 

CO CO ^ -* CO SO 




| 


d 


' s ^ 


* * ell 


C^ 1 ^ 
<N cq 


4 




CO 






Four Foot Span. 


1 


39S3S83SS 


1 


'- 


T-H 


4 






PROPORTIONS. 


ifi 


05 CO 00 


<M CO t^ t- 


a 


Oi T* rH O O 


^ r 


b 


o ^... oo 




aivoaaooy 


CrOOSfdClfO'^S ^^W 


CEMENT. 


a 


H g 


d 


_c 


*J *' 


fe " 


. - f "/ ' 


r- O5 

10 kO 

TH 1 1 



TRANSVERSE STRENGTH CONCRETE 



339 



tions The gravel and hard limestone gave about the same 
result, but it is seen that the mixture gave a higher strength. 
Bars 154 to 159 were made to test the value of broken brick for 
use in concrete. It is seen that the strength obtained with 
brick is considerably lower than that obtained with the soft 
limestone. Had a poorer mortar been used, the brick would 
doubtless have given a better comparative result, since with 
the one-to-two mortar, the brick are not strong enough in them- 
selves to utilize the full adhesive strength of the mortar. 

TABLE 156 

Transverse Tests of Concrete. Use of Screenings with Broken 

Stone 







SAND AND 










STONE. 


STONE 
TO 80 LBS. 


a ^ 


vt 


MODULUS OF RUPTURE. 






CEMENT. 


-5 P 


S . 

O H 




No. 
BA.R 








aT 


* 


s 


FOUR FOOT 


TWENTY INCH 




2 
o 


iss 


if 

ftjg 


Js 


si 


M 


SPAN. 
AGE, 11 Mos. 


SPAN. 
AGE, 2 YRS. 






'S>3 


oS 


~ 3 





g 








03 

A 

5 


r.s 




F 

02 




> 


No. 
Tests. 


Mean. 


No. 
Tests. 


Mean. 


114-115 


a 


49 


240 


7.0 


3.43 


3.05 


2 


233 


4 


237 


124-125 


b 


48.4 


243 


7.0 


3.39 


3.05 


2 


196 


4 


210 


112-113 


c 


44 


240 


7.0 


3.08 




2 


194 


3 


236 


116-117 


d 


44 


138 


7.8 


3.43 


2.16 


2 


227 


3 


311 


118-119 


e 


40.5 


243 


7.0 


2.83 


3.10 


2 


201 


4 


219 


120-121 


f 


38.8 


243 


7.0 


2.72 


3.05 


2 


122 


4 


164 


122 


g 




243 


7.0 




3.05 


1 


130 


2 


141 



NOTES: 

Cement: Natural, Brand Gn, Sample 92 T, 80 Ibs. 
Stone: a = Drummond Island limestone, screened. 

b = 10 parts screenings to 100 parts stone. 

c = 17 parts screenings to 100 parts stone. 

d = 17 parts screenings to 100 parts stone. Screenings replacing 
equal amount sand. 

e = 50 parts screenings to 100 parts stone. 

/ = 100 parts screenings to 100 parts stone. 

g = Screenings only, no broken stone. 

455. Use of Screenings with Broken Stone. Table 156 gives 
the results of a number of tests made to show the effect of 
mixing screenings with the broken stone. A smaller amount 
of mortar is required to fill the voids in a given volume of stone 
and screenings mixed than is required for the same volume of 



340 CEMENT AND CONCRETE 

stone. It is seen that, with natural cement, when the same 
volume of mortar is used in the two cases, the presence of 
screenings to the amount of one-third of the total aggregate 
does not make a material change in the strength of the result- 
ing concrete, but when the screenings are allowed to take the 
place of a part of the sand in the mortar, as in bars 116 and 
117, a much stronger concrete results. Natural cement mixed 
with sand and screenings alone, bar 122, does not make a strong 
concrete, but Portland cement with screenings without sand 
was found to give excellent results. 

456. Deposition in Running Water. A few tests were made 
to show the effect of depositing concrete in rapidly running 
water. The molds were placed in the stream and weighted 
down in twelve inches of water. The concrete for two bars 
was deposited as soon as mixed, that for two other bars was 
allowed to stand in the air three hours before deposition, until 
it should have acquired an initial set, and two bars were made 
after the mortar had been allowed to stand five hours and 
twenty minutes before deposition; by this time the mortar had 
set quite hard. No attempt was made to ram the concrete, 
which was deposited by lowering it carefully into the water 
with shovels, the molds being filled as rapidly as possible. A 
very large amount of the cement was washed out by the current 
in all cases. After a few months the bars were removed from 
the stream and covered with earth as usual. The tests at eleven 
months did not appear to show any advantage in allowing the 
mortar to stand some time before deposition, but the tests at 
two years showed a distinct advantage in this treatment. 

457. Use of Concrete in Freezing Weather. Table 157 gives 
the results obtained with Portland cement concrete made in 
the open air during cold weather. The conditions as to tem- 
peratures and the character of the materials are fully given in 
the table. The experiments are too limited to permit of draw- 
ing definite conclusions, but the following points are indicated 
by the results obtained. The use of warm water, 100 to 156 
Fahr., in freezing weather appears to give somewhat better 
results than cold water. Salt should not be used unless the 
temperature is below the freezing point, but in very cold weather 
the use of enough salt in the water to lower its freezing point 
below the temperature of the air seems to hasten the harden- 



TRANSVERSE STRENGTH CONCRETE 



341 



TABLE 157 

Transverse Tests of Concrete Bars. Use of Concrete in Low 
Temperatures 



*1 






tf a 


_ ^-^v~ . . ~ 




S^ 


CO Tt< <N CO OS CO i-H 
CO O-l "^ i 1 rH CO *O 

i I GN GN1 CO rti CO CO 


t^OSt--OOSi>.OOOOT ICO 

9 . 8 3. & 9 9 9 8 ' S 9 S 


o & ^. fl . 


<M. O r~- CO CO i 1 C^ 

O -^ (N (7<l IT- >O CO 


lllSlgSIHI 


w 


s ^^7^01^^^ 


2S as S O5 ?3 Oi ?a os 2n ao ^r' 00 "^ OSCOOSOO5CO!3:><:o 




? 




sxsax 'ON 




. Z^!l^!lZl^Z^Zl!!,^II^l 


Ijjj 


to 

iN 00 

. rH r\ 


s : o : o : ,, : g 


a- P 

3 hT^? 

ID W Si- 1 
H r^ 


*H O >O "* * ^1 

T * 1 1 


ra S S S3 S : : 

. 1 1 


8-d 

fe ^ < 
K t>2 
Q ^"0 

H .^w 


OS CO (M CO CN OS t- 


t--t--iOCOTjTt<CO(NiOOt^ 

(MCqCOCOCOrH^^Hr-HG^c^ 


w tl 
M 


OS -* CO CO . . 00 

O "" 00 . . O 


^00 ^ CO CO 

^ rt *. S ~ i- CO CO - 

O O o O 




(M O 

CO ^ ^ ^ ^ ^ O 


-''---CO--'-* 


01 ^ . 

o 5 ^ 
1 ^^ 


O ----- 
G<1 (M 


T 1 


2 

J 







o 






i 

1 


il = ^ s ^ s 




P 


s 


I i 1 1 



REMARKS: 

a, Concrete frozen after 10 to 20 min.; 6, frozen in 45 min.; c, began to 
freeze in 15 min.; d, frozen hard following morning after molding; 
e, concrete still soft 9 A.M. morning after molding; /, bar defective. 
NOTES: Cement, Portland, Br. R. Sand, "Point aux Pins." Stone, 
Potsdam sandstone. 



342 



CEMENT AND CONCRETE 



ing as well as to increase the ultimate strength, 
mortars in freezing weather, see Art. 50.) 



(For tests of 



ART. 58. RESISTANCE TO SHEAR AND ABRASION 

458. Shearing Strength. The shearing strength of mortars 
and concretes is of importance not only because of its intimate 
relation to the compressive strength, but because of the shear- 
ing stresses to which these materials are subjected in structures 
reinforced with steel. But few tests of shearing strength have 
been made, however, partly because of the lack of appreciation 
of their value, and partly because it is difficult to subject a 
specimen to a purely shearing stress. It is frequently stated 
that the shearing strength is somewhat in excess of the tensile 
strength, perhaps as much as twenty per cent. 

Table 158 gives the results of a series of tests made by Prof. 
Bauschinger in 1878. 1 The values in shear are very closely 
twenty per cent, in excess of the tensile strengths of similar 
mortars tested at the same time. 

TABLE 158 

Shearing Strength of Portland Cement Mortar Cubes Hardened 

in Air 







SHEARING STRENGTH, POUNDS PER 








SQUARE INCH. 


TENSILE 


CEMENT. 


*g| 


Age of Mortar. 


OF SIMILAR 
MORTARS 




PS D W 








s 2jO 










WEEKS. 




tt 


1 Aveek. 


2 weeks. 


4 \veeks. 


8 weeks. 




Quick setting Port- ( 


None 


225 


270 


257 


259 


210 


land, mean results } 


3 


108 


128 


154 


196 


169 


of four brands. ( 


5 


67 


94 


112 


168 


139 


Slow setting Portland, ( 


None 


301 


323 


341 


377 


256 


mean results of four 1 


3 


124 


164 


199 


237 


181 


brands. ( 


5 


78 


122 


138 


199 


169 



NOTE : Cement, each result mean of four brands. Sand, medium grain, 

clean. Mortars hardened in dry air. 
Tests by Professor Bauschinger, 1878. 

459. A distinction should be drawn between the resistance 
offered by a thin mortar bed to the sliding of one stone or brick 
on another and to shear of the mortar itself. The forme'* re- 



Quoted by Mr. Emil Knichling in a Report on Cement Mortars. 



M1EAR AND ABRASION 343 

sistance involves the adhesion, of the mortar to the surface of 
the brick or stone, and the values for this resistance are usually 
much less than the shearing strength, and not greatly in excess 
of the adhesive strength. The one is of importance in the de- 
termination of the stability of masonry dams, retaining walls, 
etc., but the latter is the resistance in question in the design 
of monolithic concrete structures. 

460. Resistance to Abrasion. The resistance of cement 
mortar to abrasion depends on the quality of the sand as well 
as the cement. The abraiding surface wears away the cement 
or pulls the particles of sand out of their beds in the cement 
matrix. If the adhesion to the sand grains is strong, the sand 
particles receive the wear and withstand it until nearly worn 
away. With hard sand particles, therefore, the resistance to 
abrasion should increase as the proportion of sand increases, 
until the volume of the cement matrix becomes relatively too 
small to thoroughly bind the sand grains together. This limit is 
reached, however, when the mortar contains not more than two 
parts sand. With soft sand grains, the neat cement will usually 
give the highest resistance to abrasion, at least in the case of 
Portland. It has been found that specimens hardened in the air 
are brittle and wear more rapidly than those hardened in water. 

461. Table 159 gives the results of several tests made to 
determine the relative wearing qualities of different mortars for 
such uses as sidewalk construction. The specimens were two- 
inch cubes, hardened in water and dried for a few hours just 
before grinding. An emery plate, set horizontally, was used 
in most of the tests. The results in any given line of the table 
are comparable, but, owing to changes in the grinding plate 
and in the methods used, the results in different lines are not 
all intercomparable. It is seen that when soft sand is used, 
such as limestone screenings, the greatest resistance to abrasion 
is offered by the neat cement mortar, and the resistance de- 
creases constantly as the amount of sand is increased. When 
hard sand, such as the siliceous river sand, from Point aux 
Pins ("P.P." in the table) is employed, the greatest resistance 
is offered by mortars containing about equal parts of sand and 
cement. A comparison of lines 5 and 10 indicates that rich 
natural cement mortars lose about twice as much as similar 
mortars of Portland, but natural cement mortars containing 



344 



CEMENT AND CONCRETE 






WEIGHT OF SPECIMEN IN GRAMS 
OF SURFACE SUBJECTED TO ABRA 



i 



sq^i'ua 
ui -bg a 



uo pasn 



s 



OS CO CO 

10 co 



CO *O O <M 
!>' CO rH CO 



COO OS rH CM t- 

id o TtJ t-' <M' CN co 



rH lO <* rH CO CO O 
Tti Tji Tf' t-' i-i (jq CO* 



cq o o 

00 OS* T}1 

<N 



t^ rH OS 
T*J OS "*' 



O i- <N 



OrHOlO 



1 O t- 

CO r-i CO 



OSCO-*D 
O r-i O O 



<7q OS_ OS 
CO O <M' 



O rH MI CO 
SO rH O Tf< 



<M_ t- 
!>' O rH 



coosco 
<M' o <N 



o o co d co r-f 



OCOt^uO (751 b-O 
(N r-i O rH (N O CO 



COCO CO CO MH CO 



O O 

^> co 



OO 

coo 



* s 
d 




pui r/2 C/2 QC oJ '/: Oi CO 

n] CO rH rH oJ CO rH rH 



-o pdN 



rH <?5 CO M* CO t- CO OS 



EXPANSION AND CONTRACTION , 345 

more than two parts by weight of sand do not give relatively 
as good results. 

ART. 59. THE EXPANSION AND CONTRACTION OF CEMENT MOR- 
TAR. AND THE RESISTANCE OF CONCRETE TO FIRE 

462. Change in Volume during Setting. Cement mixtures 
shrink somewhat when hardened in air, while specimens stored 
in water expand a trifle during hardening. Although several 
experiments have been made on this subject the specimens used 
have been so small that the results obtained by various author- 
ities do not agree, and the effect of variations in the character of 
the mixtures has not been thoroughly investigated. The impor- 
tance of the question is found in the necessity of providing ex- 
pansion joints in long walls or sheets, and in the effect of such 
changes in volume in producing initial stresses in concrete or 
steel where these materials are used in combination. 

Certain general conclusions are well established and may be 
stated as follows: 1st. The shrinkage of mortar and concrete 
hardening in air is considerably greater than the expansion of 
similar specimens hardening in water; 2d. The amount of 
change in volume increases with the proportion of cement used 
in the mixture; 3d. The change in volume is continuous up to 
one year, but about one half of the change occurs in the first 
week, and it is very slow after 3 to 6 months. 

The following values of the change in linear dimensions are 
derived from the results of several experimenters, and show in a 
general way what changes are to be expected at the end of three 
months. 1 Variations in the character of the cement and the 
consistency of the mortar will affect the result. 



COMPOSITION: PARTS SAND 
TO ONE PORTLAND 
CEMENT. 


SHRINKAGE OF MORTARS 
HARDENED IN AIR. 


EXPANSION OF MORTARS 
HARDENED IN WATER. 


CHANGE IN LINEAR DIMENSIONS, ONE UNIT IN 


Neat cement .... 
One part sand .... 
Three parts sand . 


300 to 800 
600 to 1200 
700 to 1200 


500 to 2000 
1200 to 3000 
3000 to 5000 



1 For more detailed results the reader is referred to the following authori- 
ties: Dr. Tomei, Trans. A. S. C. E., Vol. xxx, p. 16. Mr. John Grant, 
Proc. Inst. C. E., Vol. Ixii, p. 108. Prof. Bauschinger, Trans. A. S. C. E. 
Vol. xv, p. 722. 



346 CEMENT AND CONCRETE 

463. The Coefficient of Expansion of Cement and Concrete. 

Concerning the coefficient of expansion of cement mortars of 
various compositions, we know but little. The result obtained 
by M. Bonniceau, giving the coefficient of neat Portland cement 
as about .000006 per degree Fahr., is frequently quoted. This 
is very nearly the value for iron and steel, and has formed a the- 
oretical basis for combining these materials. In the case of 
cement mortars and concretes, however, it is highly probable 
that the coefficient follows quite closely the behavior of the sand 
and stone used in the mixture, and is much less dependent upon 
the coefficient of the cement. This was indicated by the results 
of M. Bonniceau who obtained a value of about .000008 for con- 
crete. 

464. A number of experiments to determine the coefficient 
of expansion of cement concretes were carried out under the 
direction of Prof. Wm. D. Pence by students of Purdue Uni- 
versity. 1 As a mean of seven tests with one-two-four concrete 
of Bedford oolitic and Kankakee limestones combined with 
Portland cements of two well-known brands, the mean result 
for the coefficient was .0000055, the lowest result being .0000052, 
and the highest result .0000057. The coefficient of a bar cut 
from the Kankakee limestone was .0000056, the same result as 
obtained from the mean of three tests of concrete containing 
broken stone of this variety. 

The average result of four tests of gravel concrete composed 
of one part Portland cement, two parts sand and four parts 
screened gravel, or one part Portland cement to five parts un- 
screened gravel, gave .0000054 as the coefficient of expansion. 

These values differ from the coefficient of steel enough to 
indicate that in positions where the range in temperature is 
great, the resulting stresses in the concrete and steel may be 
considerable, and worthy of attention. 

465. THE FIRE-RESISTING QUALITIES OF CONCRETE. The 

value of concrete as a material to be used in the construc- 
tion of the walls and floors of buildings, is largely dependent 
on its fire-resisting qualities. That its use for such purposes 
is rapidly extending, is some evidence that these qualities are 



1 Paper read before the Western Society of Engineers, Engineering News, 
Nov. 21, 1901. 



RESISTANCE TO FIRE 347 

as satisfactory as in other classes of materials devoted to the 
same use. 

Under favorable circumstances, a fire in a building filled with 
combustible materials may reach a temperature of 2,000 to 
2,300 Fahr. If a small specimen of cement mortar or concrete 
is subjected to a temperature approaching this intensity, the 
cement loses its water of crystallization and becomes friable. 
If cooled suddenly in water, the specimen cracks and disinte- 
grates. If cooled gradually, the outer edge of the specimen 
crumbles away. From such tests on small specimens some very 
erroneous conclusions have been drawn as to the value of con- 
crete as a fire-resisting material. Such conclusions have done 
much to prejudice the public mind against concrete, and to re- 
tard its introduction in buildings designed to be fireproof. 

466. Conductivity. The great value of concrete as a fire 
resistant is due to its low conductivity of heat, and while the 
surface of a mass of concrete exposed to an intense flame for 
some time is ruined, and may be flaked off by the application 
of a strong stream of water from a fire hose, the depth to which 
the heat penetrates is very limited. Steel is said to lose ten 
per cent, of its strength at about 600 Fahr. and fifty per cent, 
at about 750 Fahr. The importance of protecting the steel 
framework of a building, not only from warping and complete 
destruction due to flames, but from loss of strength from over- 
heating, is therefore evident. 

Among engineers and architects it is recognized that the 
term " fireproof construction" is only relative, although the 
lay mind js apt to give a definite and literal meaning to the 
term. It is well known that fireproofing tile, whether hard or 
porous, will fall to pieces if subjected to a temperature above 
that employed in its manufacture. The practical question then 
is, what type of construction will withstand long continued 
intense flame, and subsequent quenching with water, with the 
least injury to the strength of the structure. The results of 
fire tests that have been conducted in several places, and no- 
tably those made by the Department of Buildings of New York 
City, have shown that floor arches properly constructed of 
concrete-steel are equal to any style of floor with which they 
come in competition. 

The low conductivity of concrete is shown by the fact, 



348 CEMENT AND CONCRETE 

stated by Mr. Howard Constable in connection with the dis- 
cussion of fire tests of concrete floor arches, 1 ' ' that in some 
thirty-five cases where the temperature ranged frpm 1,500 to 
2,400 degrees, the time of exposure being from one to six hours, 
the temperature of the upper flanges of six-inch to ten-inch 
beams might be approximately placed at not much above 200 
degrees." He also says "in one case, where the beam was pro- 
tected by three inches of concrete, the fire was maintained for 
five hours, and the temperature went as high as 2,300 degrees, 
and there was no practical or permanent set produced in the 
beams." 

467. Behavior in Conflagrations. As to the behavior of 
concrete-steel arches in an actual fire, a board of experts was 
appointed by the insurance companies to investigate the causes 
and extent of damage to the fireproof buildings in the Pitts- 
burg, Pa., fire of May 3, 1897. This board stated in their 
report that they believed that in important structures of this 
class "the fireproofing should be in itself strong and able to 
resist severe shocks, and should if possible, be able to prevent 
the expansion of the steel work "; and continued, "There seems 
to be but one material that is now known that could be utilized 
to accomplish these results, and that is first-class concrete. 
The fire-resisting qualities of properly made concrete have been 
amply proven to be equal, if not better than fire clay tile, as 
shown by the tests carried on by the Building Department of 
the City of New York." 

468. In a report on the Baltimore fire, Captain Sewall, 2 
Corps of Engineers, U. S. A., says concrete "undergoes more or 
less molecular change in fire; subject to some spalling. Molecu- 
lar change very slow. Calcined material does not spall off 
badly, except at exposed square corners. Efficiency on the 
whole is high. Preferable to commercial hollow tiles for both 
floor arches or slabs and column and girder coverings. In form 
of reinforced concrete columns, beams, girders and floor slabs, 
at least as desirable as steel work protected with the best com- 
mercial hollow tiles. Stone concrete spalls worse than any 



1 Trans. A. S. C. E., Vol. xxxix, p. 149. 

2 Report to the Chief of Engineers, U. S. A., by Capt. John Stephen 
Sewall, Corps of Engineers. Published in Engineering News, March 24, 1904. 



RESISTANCE TO FIRE 349 

other kind, because the pieces of stone contain air and moisture 
cavities, and the contents of these rupture the stone when 
hot. Gravel is stone that has had most of these cavities elimi- 
nated by splitting through them, during long ages of exposure 
to the weather. It is therefore better for fire-resisting concrete 
than stone. Broken bricks, broken slag, ashes and clinker all 
make good fire-resisting concrete. Cinders containing much 
partly burned coal are unsafe, because these particles actually 
burn out and weaken the concrete. Locomotive cinders kill 
the cement, besides being combustible. On the whole, cinder 
concrete is safe only when subjected to the most rigid and 
intelligent supervision; when made properly, of proper ma- 
terials, however, it is doubtful whether even brickwork is much 
superior to it in fire-resisting qualities, and nothing is superior 
to it in lightness, other things being equal." 

469. Aggregate for Fireproof Work. Since air is a poor 
conductor of heat, the more porous concretes are the better 
protectors against fire. On this account, as well as because of 
its lightness, cinder concrete is preferred for fireproofing. Care 
should be taken that cinders to be used in fireproofing concrete 
do not contain any appreciable amount of unburned coal; in 
concrete to be used next to steel members the cinders should 
also be practically free from iron rust. (See 473.) 

The strength of cinder concrete is much inferior to that 
made with the ordinary aggregates, and there should be no 
difficulty in making a porous concrete with the latter. In fact, 
in many other classes of construction it has been seen that 
great precautions must be taken to avoid porosity. By the 
use of insufficient mortar to fill the voids in the stone, voids 
may be left in the concrete, though at the expense of dimin- 
ishing somewhat the strength of the mixture. In adopting such 
an expedient one should not lose sight of the fact that in order 
to preserve the imbedded steel from corrosion, it must be fully 
covered with the mortar. 

470. Broken bricks are excellent for fireproofing concrete. 
The bricks themselves are fire resistant, porous and light, while 
the adhesion of cement mortar to bricks is so great that unless 
a very weak mortar is used, the strength of the concrete is 
limited only by the strength of the brick employed. 

Sandstones, especially those with siliceous cementing ma- 



350 CEMENT AND CONCRETE 

terial, are also well adapted for this purpose. Limestone, on 
account of the low temperature at which it is broken up, is 
not good, though as to just how far a limestone concrete would 
be disintegrated by the heat of an ordinary building fire has 
not, so far as the author knows, been fully investigated. It is 
known, however, that limestone masonry is calcined to a cer- 
tain depth in a conflagration. 

Granite in large pieces is cracked by only a moderate degree 
of heat, and spalls badly. Just how much danger there might 
be of a similar action in concrete aggregates of this material is 
not known, nor whether small pebbles or fine gravel would 
have this property in the same degree, though it is believed 
they would not, and this view has been confirmed by observa- 
tions of the Baltimore ruins. 

Before adopting a given aggregate for fireproof work, one 
should satisfy himself by ( actual test as to the suitability of the 
materials available, but such tests should be conducted upon 
concretes containing the proposed aggregates, rather than upon 
fragments of the materials not incorporated with mortar. 

ART. 60. THE PRESERVATION OF IRON AND STEEL BY MORTAR 

AND CONCRETE 

471. The rusting of steel members in modern buildings and 
other engineering structures is one of the most serious menaces 
to their permanence. The introduction of concrete-steel con- 
struction has given rise to some discussion, especially among 
those unfamiliar with the properties of concrete, as to the 
effect of the concrete upon the steel. 

472. Action of Corrosion. The rusting of iron takes place 
only in the presence of moisture, air and carbon dioxide. In 
perfectly dry air, or in perfectly pure water, iron does not rust. 
Under the proper conditions, however, the iron, water and 
carbonic acid combine to form ferrous carbonate, which at once 
combines with oxygen from the air to form ferric oxide, the 
carbonic acid being liberated to act on a fresh portion of the 
metal. It is seen that only a very small amount of the carbon 
dioxide is necessary. If, however, the carbon dioxide or other 
acid filling the same role, is neutralized by the presence of an 
alkaline substance, the foregoing reactions cannot take place. 
As cement is strongly alkaline, it thus furnishes an almost per- 
fect protection against rusting. 



EFFECT ON CORROSION OF METAL 351 

473. Tests of Effect of Concrete on Corrosion of Metal. - 
To determine the cause of occasional rusting of steel surrounded 
by cinder concrete, and consequently the proper methods of 
applying cement mortar or concrete to steel, Prof. Chas. L. 
Norton, engineer in charge of the Insurance Engineering Ex- 
periment Station at Boston, made tests on several hundred bri- 
quets in which steel was imbedded in mortars and concretes 
of various compositions. 1 The briquets were subjected to air, 
steam and carbon dioxide, others to air and steam, to air and 
carbon dioxide and to the ordinarily dry air of a room. At the 
end of three weeks it was found that neat Portland cement had 
furnished a perfect protection in all cases. The corrosion of 
the steel in other specimens was always at a point where a void 
existed in the concrete, or where a badly rusted cinder had lain. 
In every case where the concrete or mortar had been mixed wet, 
and the surface of the steel had been thus coated with a thin 
layer of grout, no rust spots occurred. 

In the first tests made by Professor Norton the specimens 
were thoroughly cleaned before being imbedded in the concrete, 
but later tests indicated that in specimens that had begun to 
corrode before treatment, the rusting was arrested by the coat- 
ing of cement mortar or concrete. After from one to three 
months in tanks holding steam and carbon dioxide, specimens 
which had' been in all stages of corrosion before being im- 
bedded in the concrete had not suffered any sensible change 
in weight or size except when the concrete had been poorly 
applied. 

474. The results of these experiments showed that the steel 
need not necessarily be freed from rust before being imbedded 
in the concrete; that the concrete to be applied next the steel 
should be mixed wet, or that the steel should be first coated 
with grout by dipping or brushing; and it appeared that the rust- 
ing sometimes found in cinder concrete is due to the rust in the 
cinders rather than to the sulphur, and that if proper precau- 
tions are taken, cinder concrete is nearly as effective as stone 
concrete in preventing corrosion. Prof. Norton says,- "In the 
matter of paints for steel there is a wide difference of opinion. 
I cannot believe that any of the paints of which I have 



1 Report III of Insurance Engineering Exp. Station, Boston, Mass. 



352 CEMENT AND CONCRETE 

any knowledge can compare with a wash or painting with 
cement." 

475. Sulphur in Cinders. The conclusions drawn by Booth, 
Garrett and Blair from a series of tests made for the Roebling 
Construction Co., were that cinders from anthracite pea coal 
contained about two-tenths per cent, of sulphur which they 
considered sufficient to cause corrosion of unprotected iron- 
work, more or less rapidly, depending on the presence or absence 
of moisture; but they further concluded that a "full" concrete 
(one in which the voids in the cinders were entirely filled by 
mortar of cement and sand) would fully protect the steel. 

In a paper read before the Associated Expanded Metal Com- 
panies, Prof. S. B. Newberry has this to say concerning cinder 
concrete: l " The fear has sometimes been expressed that cinder 
concrete would prove injurious to iron, on account of the sulphur 
contained in the cinders. The amount of this sulphur is, how- 
ever, extremely small. Not finding any definite figures on this 
point, I determined the sulphur contained in an average sample 
of cinders from Pittsburg coal. The coal in its raw state con- 
tains rather a high percentage of sulphur, about fifteen per 
cent. The cinders proved to contain only 0.6 per cent, sulphur. 
This amount is quite insignificant, and even if all oxidized to 
sulphuric acid, it would at once be taken up and neutralized 
in concrete by the cement present, and could by no possibility 
attack the iron." 

476. Precautions. While so far as the corrosion of steel 
is concerned, the above experiments by Prof. Norton show that 
'the rusting is corrected by the concrete, yet it is quite possible 

that the adhesion of cement to steel may be impaired by a coating 
of rust. The cleaning of the steel may be accomplished by 
first brushing with wire brushes to remove all scales, followed 
by treatment with hot dilute sulphuric acid, and finally apply- 
ing an alkaline wash such as hot milk of lime to neutralize all 
traces of the acid. Oxalic acid may be used in place of the 
sulphuric, and the application of the milk of lime dispensed with, 
since the acid oxidizes. The crystals of oxalic acid as purchased 
commercially should be mixed with about seven parts hot water 
and the solution applied with a brush or sponge. When the 



1 Engineering News, Apr. 24, 1902. 



PRESERVATION OF STEEL 353 

adhesion of the mortar or concrete to the steel is of any impor- 
tance, as it is in all concrete-steel construction where the stresses 
are divided between the steel and concrete, any of the ordinary 
oil paints will not only be quite unnecessary, but may be a very 
serious detriment to the construction. 

The experiments quoted indicate the importance of having 
the steel covered with an unbroken coating of cement or cement 
mortar. To insure this the steel must either be coated with a 
layer, preferably of neat Portland, by dipping or brushing, or 
the mortar placed next the steel must be wet enough to insure 
intimate contact throughout. It may be added also, that the 
addition of a small amount of thoroughly slaked lime to Port- 
land cement mortar or concrete will not only render the mate- 
rial more alkaline, but will make the mortar more plastic, and 
thus insure a better coating of the steel. Such small additions 
have no deleterious effect on the mortar. 

477. Practical instances of the preservation of iron by con- 
crete are not wanting. The writer has stored in water, briquets 
with small iron plates imbedded in Portland cement mortar, 
and at the end of six months the plates were found moist, but 
entirely free from corrosion except where they projected beyond 
the mortar. A concrete-steel water main built on the Monier 
system at Grenoble, France, was taken out and examined after 
fifteen years service in damp ground. The metal imbedded in 
the mortar showed no signs of corrosion, and the mortar could 
only be detached from it by hammering. 

Mr. W. G. Triest l relates that in breaking up cast-iron, con- 
crete-filled pillars, a wrench was found that had been buried 
in the concrete for twenty-two years. The wrench had main- 
tained its metallic surface in the concrete, while a part of it that 
had been imbedded in coal ashes had corroded badly. 

Similar instances showing the action of concrete on steel 
and iron might be multiplied, but it is sufficient to state that 
the preservation of iron or steel properly imbedded in Port- 
land cement mortar or concrete is now seldom questioned, and 
the use of cement paint, in place of the ordinary oil paints, 
as a steel preservative, has been adopted in many places. 



1 Trans. A. S. C. E., April, 1894. 



354 CEMENT AND CONCRETE 

ART. 61. POROSITY AND PERMEABILITY; EFFLORESCENCE; 
POINTING; USE IN SEA WATER 

478. The porosity and permeability of mortars have been 
thoroughly investigated by M. Paul Alexandre, who has pub- 
lished his results in"Recherches Experimentales Sur Les Mortiers 
Hydrauliques." 1 The results and conclusion in the following 
notes on the subject are largely a resume of the systematic 
investigation made by M. Alexandre. 

The two qualities, porosity and permeability, should not be 
confused, nor should it be thought that a porous mortar is 
always very permeable, or that a permeable mortar must of 
necessity be very porous. Porosity is measured by the amount 
of water which will be absorbed by a specimen after drying, 
while permeability is measured by the amount of water which 
will pass through a specimen in a given time under certain de- 
nned conditions of thickness, water pressure and area of face. 

479. Porosity. The porosity of mortars is due to, and in 
fact is measured by, the volume of the voids contained. These 
voids may be divided into three classes, according to the causes 
to which they may be attributed, as follows: 1st, apparent 
voids, due to the mortar not being properly compacted; 2d r 
latent voids, due to the imprisonment of air in the mortar when 
made; and 3d, voids resulting from the evaporation, during har- 
dening, of a portion of the water used in gaging. 

480. Apparent voids may occur as the result of using insuf- 
ficient cement to fill the voids in the sand, or, in the case of 
concretes, insufficient mortar to fill the voids in the aggregate. 
They may also be due to improper manipulation as to tamping, 
or improper mixing, giving an excess of matrix in one place and 
a deficiency in another. It was found by experiment that 
mortars made with coarse sand had the largest volume of ap- 
parent voids. 

It has been shown elsewhere that if dry sand be moistened 
and agitated, the bulk of the sand is increased. This is caused 
partially by the imprisonment of air bubbles in the mass, and 
if a measure of sand so treated is filled with water, the bubbles 
will rise to the surface on jarring the vessel. Latent voids in 
mortar are due to a similar action, and hardened mortars con- 

1 Extrait des Annales des Fonts et Chaussees, September, 1890. 



POROSITY AND PERMEABILITY 355 

taining such voids refuse to absorb water to replace the air 
bubbles, at least for a long time. 

481. A portion of the water used in mixing mortar enters 
into chemical combination with the cement, another portion is 
absorbed by the sand grains, and a third portion goes, to moisten 
the sand. The quantity absorbed by the grains depends upon 
the character of the sand, and the amount required to moisten 
the sand depends upon the superficial area of the grains in a 
given volume, being greatest for fine sands and least for coarse 
ones. At least one fourth of the water ordinarily used in 
mixing neat cement is given off later, if the hardened mor- 
tar is allowed to remain in dry air. The water required to 
moisten the sand, and at least a part of that absorbed by 
the sand grains, also dries out, leaving voids of the third class 
mentioned. 

The apparent voids may be reduced to a very small per- 
centage by care in the proportions and preparation of the 
mortar. The latent voids may amount to six or seven per 
cent, of the total volume. The evaporation of water may 
leave from six to eighteen per cent, of voids in the mass. 

482. The conclusions drawn from M. Alexandre's experi- 
ments are briefly as follows: The porosity varies between wide 
limits according to the fineness of the sand and the richness of 
the mortar. It may be as low as thirteen per cent, and may 
exceed thirty-one per cent. 

With sand of the same degree of fineness, the porosity di- 
minishes as the proportion of cement in the mortar increases. 

With the same quantity of cement per voltime of sand, the 
porosity increases with the fineness of the sand. This is es- 
pecially marked in rich mortars, where the increase in porosity 
may reach 50 to 100 per cent., while in lean mortars the use 
of a fine sand may not increase the percentage of voids more 
than 20 per cent. 

The least porous mortars are those rich in cement and made 
with coarse sand. Mortars made with fine sand are relatively 
very porous, even when made rich with cement. 

Mortars gaged dry are more porous than those of ordinary 
consistency, and mortars gaged wet are also likely to be more 
porous, unless the manipulation is such as to allow the excess 
water to rise to the surface of the mortar. 



356 CEMENT AND CONCRETE 

483. Permeability The degree of permeability of mortars 
is a more important property than the porosity, since not only 
does it affect the suitability of the mortar for certain uses, 
but the life of the structure may depend upon the difficulty 
with which water may percolate the mass. 

The permeability of mortar decreases as the proportion of 
cement is augmented, 'and in the case of concretes the per- 
meability diminishes as the percentage of mortar increases, at 
least to the point where the latter is in excess of the voids in 
the stone. 

From experiments made at the Thayer School of Civil En- 
gineering, Messrs. J. B. Mclntyre and A. L. True found that a 
five-inch layer of concrete containing from 30 to 45 per cent, 
of one-to-one Portland cement mortar, and some of the speci- 
mens containing 40 to 45 per cent, of one-to-two mortar, were 
impermeable with pressures of 20 to 80 pounds per square 
inch, maintained for two hours. 

484. Mortars made with fine sand are much less permeable 
than those made with coarse sand. This difference is so marked 
that a less permeable mortar is made with one barrel of cement 
per cubic yard of fine sand, passing a sieve having, say, fifty 
meshes per inch, than with two barrels of cement per cubic 
yard of very coarse sand in which the grains are, say, one- 
tenth inch in diameter. Mortars made with sands composed 
of a mixture of grains of various sizes are neither very porous 
nor easily permeated. 

Mortars mixed very dry or very wet have greater permeabil- 
ity than those of the ordinary consistency, and in the case of 
concretes, it would probably be found that a deficiency of 
water would result in a much more permeable mass than the 
use of what might be considered an excess. 

All of the above conclusions indicate that a mortar may be 
quite porous, and yet so long as the voids are very minute, the 
percolation of water through it will be slow. This is especially 
shown by the fact that mortars of coarse sand, not porous, 
are more permeable than the porous mortars of fine sand. 

485. When water is permitted to percolate continuously 
through a mass of mortar, the interstices gradually become filled, 
and the permeability decreases in marked degree. M. Alex- 
andre found that a volume of water which passed a certain 



WA TER-PROOFING 357 

mass of mortar in twenty minutes at the beginning of the ex 
periment, required five hours to percolate the mass at the end 
of a month. M. R. Feret has obtained similar results in making 
extensive experiments l on the subject of permeability, and 
considers that fine particles of cement or lime are carried along 
by the water, forming efflorescence at the surface and tending 
to stop the flow. 

486. The Preparation of Water-Proof Mortar and Concrete. 
To enumerate briefly the precautions necessary to attain 
water-tightness in mortars and concretes, it may be said that 
different brands of cement present different characteristics in 
this regard. Fine grinding is a prime requisite, and sand 
cement or silica cement, containing as it does very fine grains of 
sand intimately mixed with cement particles of extreme fine- 
ness, is admirably adapted to such uses. 

The sand should, if possible, be composed of a mixture of 
grains of various sizes, because such a mixture gives a mortar 
not only little permeable, but one that is not porous, and that 
has, besides, a good strength. The amount of cement in the 
mortar should be in excess of the voids in the sand, not less, 
in general, than three barrels of cement per cubic yard of sand. 

In concrete the volume of mortar should exceed the volume 
of voids in the aggregate, and to obtain this result without 
too great expense, the aggregate should be so selected as to have 
a minimum of voids. Gravel concrete properly proportioned 
may be made water-tight somewhat more easily than broken- 
stone concrete, but a mixture of gravel and broken stone will 
give good results not only in this regard, but in the matter of 
strength as well. 

487. To make a compact mortar for use where the facilities 
for tamping are ordinarily good, the consistency should be 
neither very wet nor very dry. When the mortar is struck 
with the back of a shovel, moisture should glisten on the surface, 
but in a pile the mortar should appear but little moister than 
fresh earth. This is the consistency which, with a moderate 
amount of tamping, gives the least volume of mortar with 
given quantities of dry materials. In places difficult of access, 



* "La Capacitt des Mortier Hydrauliques " Annales des Fonts et Chausstes, 
July. 1892. 



358 CEMENT AND CONCRETE 

or in the preparation of concrete, better results will be obtained 
with a mortar somewhat wetter than the above, since large 
voids will be less likely to occur in the more plastic mass. In 
fact, unless the supervision is very close, it is advisable to use 
a rather wet mixture in preparing concrete where water-tight- 
ness is desired. 

488. Washes. The application of certain washes to the 
surfaces of walls intended to be water-proof, and the introduc- 
tion of foreign materials into the mortar or concrete to make 
it less permeable, have been practiced to some extent. Alter- 
nate coatings of soap and alum solutions are applied with a 
brush, not only to concrete, but to brick and stone masonry 
surfaces. These penetrate the pores of the masonry, forming 
insoluble compounds which prevent percolation. Washes of 
grout, composed of cement, or of cement and slaked lime, are 
used for a similar purpose. 

" Sylvester's Process for Repelling Moisture from External 
Walls " consists in applying first a solution of three quarters 
of a pound of soap to one gallon of water, followed, after twenty- 
four hours, by the application of a solution containing two 
ounces of alum per gallon of water. Both solutions are applied 
with a brush, the soap solution boiling hot, and the alum so- 
lution at 60 to 70 Fahr. The applications are alternated, 
with twenty-four hours intervening each time. Experiments at 
the Croton Reservoir l indicated that four coats of each wash 
were required to render brickwork impervious to a head of forty 
feet of water and the cost of the four double applications was 
about ten cents a square foot. 

In Reservoir Number Two of the Pennsylvania Water Co., 
two washes of each solution were used on the walls at a cost 
for materials and labor of twenty-three cents per hundred 
square feet, and the results were said to be good. 

A modified recipe for such a wash in which but one solution 
is made is given as follows: 2 A stock solution is prepared of 
one pound lye, five pounds powdered alum, dissolved in two 
quarts water. One pint of the stock is used to a pail of water 
in which ten pounds Portland cement has been well mixed. 



1 Trans. A. S. C. E., Vol. i, p. 203. 

2 J. H. G. Wolf, Engineering News, June 30, 1904. 



WATER-PROOFING 359 

489. In a few cases the use of alum and soap solutions in 
the body of the mortar has been tried with apparently success- 
ful results. Mr. Edward Cunningham, 1 in making experiments 
on water-proof concrete vessels, used powdered alum equal to 
one per cent, of the combined weight of the sand and cement, 
mixing this with the dry ingredients. To the water used in 
mixing, one per cent, of yellow soap was added. The results 
were said to be very satisfactory. In the above proportions,, 
however, the amount of alum is made to depend upon the 
amount of cement and sand used, while the soap added depends 
upon the amount of water, whereas the soap should bear a de- 
finite ratio to the alum. 

In experiments with mortar composed of one part cement 
to two and one-half parts of bituminous ash, Prof. W. K. Hatt 2 
found that the alum and soap mixed with the mortar at the 
time of gaging increased the strength and hardness of the ash 
mortar about fifty per cent., and diminished the absorption by 
the same percentage. One half of the water used for gaging 
was a five per cent, solution of ground alum, the other half 
being a seven per cent, solution of soap. The alum solution 
was used first and the gaging completed with the soap solution. 

Mr. W. C. Hawley 3 employed a stock solution of two pounds 
caustic potash, five pounds powdered alum, and ten quarts 
water, and used in the finishing coat three quarts of this solu- 
tion in each batch of mortar containing two bags of cement. 
The mortar was mixed with two volumes of sand to one of ce- 
ment and covered forty-eight square feet to a depth of about 
one-half inch. The extra cost for materials and preparing so- 
lution was only about nine and a half cents per hundred square 
feet. With less than two parts sand to one cement, it was 
found the finishing mortar checked in setting. It was also 
found that any organic matter in the sand was softened by the 
potash, and an excess of potash caused checking, although an 
excess of alum had no deleterious effect. 

490. Use of Lime, etc. The introduction of slaked lime in 
mortars designed to be water-proof is suggested by the fact 



1 Trans. A. S. C. E., Vol. li, p. 128. 
a Trans. A. S. C. E., Vol. li, p. 129. 
1 Journal New England Water-Works Association, 1904. 



360 CEMENT AND CONCRETE 

that the permeability of mortar diminishes if water is allowed 
to percolate it for some time, the theory being that fine par- 
ticles of cement and lime are dislodged by the passage of the 
water to form a deposit at or near the surface, and check the 
flow. This suggestion, however, needs experimental confirma- 
tion, since it seems quite possible that the introduction of a 
substance containing such a large proportion of water as does 
slaked lime, may increase the percentage of voids in the mor- 
tar, if not the permeability. 

The use of pulverized clay and pozzolanic materials for a 
similar purpose has been suggested. It has already been shown 
that moderate doses of clay have no deleterious effect on the 
strength of mortars for ordinary exposures. The action of the 
pozzolanic substances has been found by Dr. Michaelis and 
M. Feret to be not mechanical alone, but chemical, and the 
effect on the strength of the resulting mortar depends upon the 
exposure to which it is subjected, such admixtures being dele- 
terious for mortars hardened in air. 

491. Efflorescence. The white deposit sometimes formed 
at the surface of brick and masonry walls is usually due to the 
filtration of water through the mortar, dissolving out salts of 
potash, soda, etc., and depositing these salts on the surface by 
evaporation or by the formation of sodic carbonate. The ab- 
sorption of water from the atmosphere may also account for 
this deposit in some degree, especially near the sea. The same 
term is applied to a more harmful deposit, sulphate of calcium, 
which may be supplied by the filtrating water or may come 
from the cement, either from the addition of gypsum or from 
the fuel used in burning. The crystallization of this salt in the 
pores of the masonry at the surface may cause disintegration. 

On the other hand efflorescence may be quite harmless, as 
when it is formed by washing out from the mortar an excess of 
hydrate of lime. A portion of the latter may then be changed 
to carbonate of lime near the surface of the wall and actually 
stop up the pores or voids, a.nd prevent further filtration. 

492. The discoloration of brickwork and fine masonry by 
efflorescence is sometimes serious. To ameliorate these condi- 
tions, the use of water-proof mortars, and careful pointing of 
the work, are precautions to be recommended. General Gill- 
more, in " Limes, Hydraulic Cements and Mortars/ 7 suggests 



EFFLORESCENCE 361 

the use of about ten pounds of animal fat to one hundred pounds 
of lime and three hundred pounds of cement; the object of the 
fat being to saponify the alkaline substance, the lime in form 
of paste serving only as a vehicle for the fat. A more practical 
method, however, would seem to be the application of soap 
and alum washes on the surface, or the use of soap and alum 
in the preparation of the mortar to be used near the face of the 
wall, and especially for pointing. The remedy to be adopted, 
however, will depend upon the cause of the efflorescence. 

493. Pointing Mortar. Pointing serves the double purpose 
of making the joint practically water-tight at the edges, and giv- 
ing a finish to the face of the wall. If the edge of the joint is 
not well filled, moisture collects there either from the face or 
from seepage through the wall. Subsequent freezing or the 
crystallization of certain salts may spall the stones or loosen 
them from their bed. 

In laying cut-stone masonry, the joints should be raked out 
for about two inches back from the face to be pointed. 
Pointing mortar should be prepared from fine sand and the 
best Portland cement. The proportion of sand should not 
exceed two parts by weight to one cement, and in the highest 
class work, equal parts of cement and sand are sometimes 
used. No advantage is gained, however, by using a mortar 
richer in cement than the one last mentioned. The use of fine 
sand and rich mortars are specified not only because such mor- 
tars are practically water-tight, but because they take a fine 
finish. 

494. The tools required for pointing are a bent iron to rake 
out the joints (though this should be partially done while the 
mortar is green), a mortar board and small trowel, a calking 
iron and wooden mallet, a brush for moistening the joint, and 
one or more beading tools. After raking out the joint it is 
moistened by the brush, and the mortar, which is mixed quite 
dry, is filled in with the trowel. When enough mortar is in 
place to fill half the depth of the joint, it is tamped with the 
calking iron and mallet much as a ship's seam is calked with 
oakum. The joint is then filled to the face, and again tamped. 
The bead is then formed by running the beading iron back 
and forth over the joint. This beading iron is of steel with the 
handle parallel to, but some two or three inches out from, the 



362 CEMENT AND CONCRETE 

line of the blade forming the bead. The blade is three to five 
inches long and " hollow ground" or finished with a smooth 
concave surface. Only such a length of joint is pointed at one 
operation as may be quickly carried to completion. The wall 
must be kept moist for some time after the pointing is done, 
and it should be protected from the direct rays of the sun, as 
fine cracks are very likely to appear in this rich, finely finished 
mortar. If possible, pointing should be done in moderate 
weather and must be entirely suspended in temperatures ap- 
proaching the freezing point. 

495. Cements in Sea Water. The theory of the action of 
sea water upon cements is not fully understood. It is known 
that some cement structures exposed to the worst conditions 
have given most satisfactory results, while others have failed 
in greater or less degree. It may be said at once, however, 
that many of the most eminent and conservative engineers 
consider that the failures that have occurred in the use of Port- 
land cement in sea water are due to improper specifications, 
proportions and manipulation, rather than to any defect in 
Portland cements as a class. 

496. It is thought the following represents, in the main, the 
most generally accepted theory of the chemical action. In the 
setting of cements that are rich in lime, the whole of the lime 
is not engaged in stable compounds, and when placed in the 
sea the sulphate of magnesia of the sea water is able to com- 
bine with the lime, forming calcic sulphate, the magnesia being 
precipitated. The discovery of magnesia in decomposed mor- 
tars led, at first, to the supposition that the cause of failure 
was the presence of magnesia in the cement when used. If 
the water level about the structure changes frequently, as is 
usual, or if the wall is at times subjected to a greater head on 
one side than on the other as in tide docks, the percolation of 
water through the wall is stimulated, and the sulphate of lime 
may then be washed out if the mortar is quite pervious, and 
more will be formed from a fresh supply of sea water attacking 
the lime of the cement, until the latter is destroyed. If, how- 
ever, the sulphate of lime is .not washed out, it may crystallize 
and thus cause swelling of the mortar. 

497. It would appear from the above that for successful 
use in sea water the hydraulic index of the cement should be 



ACTION OF SEA WATER 363 

high; that is, that the lime should be comparatively low in or- 
der that the lime compounds may be more stable. For this 
reason it is not impossible that some of our natural cements, 
which are so much more nearly uniform than the Roman ce- 
ments of Europe that have been condemned for this reason, 
may give fairly good results in sea water. The fact that the 
mortars of natural cement are more permeable than those of 
Portland, is, however, a serious defect. 

Following a similar reasoning, Dr. Wm. Michaelis has ad- 
vanced the theory that if trass, or other pozzolanas of proper 
composition, be mixed with Portland cement subsequent to 
the burning, the hydrate of lime which separates from the ce- 
ment in hardening will at once combine with the pozzolanas, 
forming a stable compound. This view, however, has been 
vigorously opposed by the Society of German Portland Cement 
Manufacturers, as well as by many engineers, especially of 
France, and the discussion is not yet at an end. 

M. Candlot l says that, from the experiments of various en- 
gineers, "we have arrived at this conclusion, that the only 
remedy to adopt against decomposition is to prevent the sea 
water from penetrating the mortar. We are led thus to dis- 
miss the chemical reactions of sea water on mortars and to 
consider their action from a purely physical standpoint." 

498. To resist the attacks of sea water the mortar should not 
only be impervious, but also as little porous as possible. The 
cement should be finely ground and should not contain free 
lime. The content of magnesia and of sulphuric anhydride 
should be as low as possible, the latter not exceeding one and 
five-tenths per cent. The proportion of lime should not be 
too high, and above all, special pains should be taken with the 
manufacture to insure proper comminution and mixing of the 
raw materials, and uniform burning. The addition of sulphate 
of lime to regulate the setting is believed to be injurious for 
cements to be used in sea water; even two or three per cent, 
is said to cause rapid disintegration, and in the specifications 
for recent extensive works in dock construction, the addition 
of gypsum or other foreign matter was entirely prohibited. 



1 "Le Ciment," September, 1896, quoted by F. H. Lewis, M. Am. Soc. 
C. E. Trans. A. S. C. E., Vol. xxxvii, p. 523. 



364 CEMENT AND CONCRETE 

Although slag cements have given good results in the sea 
for a short time, it is considered that they will not, in general, 
resist the action of sea water for long periods. 

499. Sand or aggregate containing argillaceous or soft cal- 
careous matter should be avoided for works in the sea. Two 
instances of failure of sea walls in which shells were used as the 
aggregate are mentioned by Col. Wm. M. Black, 1 and although 
the failures are not definitely traced to the calcareous matter in 
the concrete, the fact that experiments have shown that cal- 
careous sands do not withstand the action of sea water, makes 
it probable that this was an important cause of the failure. 

Fine sands that give porous mortars, though not easily per- 
meable, are to be strictly avoided. Coarse sands giving per- 
meable, though not porous, mortars are better, but still leave 
much to be desired as to immunity from decomposition. The 
best sands are those containing various grades of sizes of par- 
ticles from coarse to fine, as mortars made with such sands are 
not only compact, but practically impermeable. 

500. Since the mortar and concrete should be made as com- 
pact as possible, the precautions mentioned under the head of 
water-proof mortar and concrete should be taken in the prep- 
aration of mortars and concretes for use in the sea. That is, 
the proportion of cement should exceed the voids in the sand 
and the mortar should exceed the voids in the aggregate. 

M. Alexandre has found that the mortars mixed to the or- 
dinary consistency are attacked least by sea water. When 
specimens are merely immersed in the water, those mixed dry 
suffer the most, but some tests indicate that if mortars are 
submitted to the filtration of water soon after made, those 
mixed wet are most easily decomposed. As to whether fresh 
or salt water should be employed in mixing mortars to be used 
in sea water, although Mr. Eliot C. Clarke, M. Paul Alexandre 
and many others have investigated this subject, the conclusions 
are not definite and it is probable that either may be used as 
convenient. 



1 Trans. A. S. C. E., Vol. xxx, p. 601. 



PART IV 

USE OF MORTAR AND CONCRETE 
CHAPTER XVIII 

CONCRETE : DEPOSITION 

501. Concrete may be molded into blocks which are allowed 
to set and then are transported to the structure and laid as 
blocks of stone. This is the block system of construction. The 
adaptability of concrete to being built in place, however, is 
one of its chief merits, and consequently the monolithic method 
of construction is far more common Since it has been found 
that expansion and contraction, due to changes in temperature, 
affect concrete walls as they do any other walls of masonry, 
it has become customary to mold the concrete in sections, usu- 
ally alternate sections of equal size and shape being built first, 
and the omitted sections built in later. This method of con- 
structing a long wall is also called monolithic, since the blocks 
are of large size and are built in place. 

502. When concrete is deposited either in air or in water, 
molds must be provided to keep the mass in the desired shape 
until it has lost its plasticity and acquired sufficient strength 
to stand alone. In foundations, the earth at the sides of the 
excavation may supply the place of a mold, and sometimes 
the mold forms a part of the permanent structure, as in the 
case of masonry piers with concrete hearting, and in steel cylin- 
der piers filled with concrete. 

ART. 62. TIMBER FORMS OR MOLDS 

503. The construction, placing and removal of forms fre- 
quently represent a considerable percentage, from five to thirty 
per cent., of the total cost of the concrete, and it is therefore 
evident that an improper design may result in a considerable 
waste of money, as well as in marring the appearance of the 

365 



366 CEMENT AND CONCRETE 

work. The character of the form will of course depend on the 
character of the work; in the construction of a large number 
of small blocks of the same shape, where one mold may be used 
over and over, the thickness of the pieces should not be stinted, 
and the ease of knocking down the mold should be carefully 
considered. When a form can be used but once, the size of 
pieces should be no larger than necessary to give the requisite 
stiffness, and the ease of first construction is a main considera- 
tion. Forms should be left in place forty-eight hours to allow 
the concrete to set, and in the case of arches and beams a 
much longer time is necessary, so that the concrete may assume 
considerable strength before it is called upon to support its 
weight. 

504. Sheathing. Forms for massive walls of .monolithic 
construction usually have vertical posts, with iron ties across, 
or braced by battered posts outside. The sheathing planks 
are then placed horizontal. In a few cases horizontal wales 
have been placed within the posts and vertical sheathing laid 
against the wales. 

The strength of the sheathing must be sufficient to stand 
the pressure transmitted to it through the concrete when the 
rammer is used close to the face of the mold. The concrete is 
seldom built up fast enough to bring upon the sheathing a great 
head of fluid pressure, but the ramming brings a heavy local 
pressure upon it. If supported at intervals of four feet, two- 
inch lumber dressed to one and three-quarters inches thick is 
usually sufficient; for spans of more than 5 feet, 2 J inch lum- 
ber is required to make a perfect face. Boards seven-eighths 
inch thick are suitable only when supports are not more than 
about 2 feet apart. In placing concrete in molds under water 
there is more danger of bursting the mold by the weight of 
semi-fluid concrete, and if the work is to be built up rapidly, 
this must be guarded against by sufficient bracing. 

505. For exposed faces, the duty to be performed by the 
lagging includes leaving as smooth a finish as possible on the 
concrete after the removal of the forms. If green lumber is 
employed, the boards may shrink before use, leaving openings 
between the sheathing that will show plainly on the face of the 
work. A slight tendency of this kind may be checked -by 
keeping the boards well wet with' a hose until the concrete is 



TIMBER FORMS 367 

placed. On the other hand, thoroughly seasoned lumber will 
swell when the concrete is placed; to obviate this difficulty the 
lower edge of each sheathing plank may be beveled on the outer 
edge; the thin edge on the inside will then crush when the 
planks swell. 

The use of tongue and grooved lagging has been tried, but 
is not usually satisfactory, as there is no opportunity to expand, 
and the planks are particularly hard to place a second time. 
To give a good face in work under water, however, tongue and 
groove sheathing will assist in preventing washing of the cement. 
Yellow pine lumber is found to be excellent for sheathing; on 
account of the large amount of pitch contained, it absorbs 
water slowly and holds its shape. For a similar reason, fir 
timber would be suitable. 

In order that the face of the mold shall be perfectly smooth, 
it is necessary to size and dress the plank on at least one side 
and two edges. 

As it is almost impossible to avoid having some line of de- 
marcation shown in the concrete at the joints of the sheathing 
planks, care should be taken that the lagging is of uniform 
width throughout, and laid horizontal so that consecutive sec- 
tions show the joint continuous. The sheathing may be placed 
for the entire form before concreting is commenced, or the 
plank may be raised on the posts as the work advances. The 
former method will usually give the neater appearance, but is 
too expensive for high walls. 

506. Lining. The appearance of the finished concrete is 
much improved, and the labor of preparing the forms probably 
not increased, since less care may be taken in surfacing, by 
lining the mold with thin sheet iron. Iron of number twenty 
gauge (.035 inch thick, 1.42 pounds per square foot) has been 
used for this purpose, but where the same lining is used several 
times, a heavier iron is preferable. The joining of one sheet of 
lining to another may present greater difficulties than the join- 
ing of planks, but joints will occur less frequently. 

In the construction of the Marquette Breakwater, Mr. 
Clarence Coleman, Asst. Engineer, used sheet steel one eighth 
of an inch thick for lining molds for building monolithic blocks. 
Concerning the use of the steel, Mr. Coleman says: 1 "Very 



Report Chief of Engineers, U. S. A., 1898, p. 2254. 



368 CEMENT AND CONCRETE 

smooth surfaces were produced on the slopes of the concrete 
and the work of the molders was greatly facilitated on account 
of the comparative ease with which the concrete was compacted 
under the slope pieces of the molds. The steel effectually pre- 
vented the aggressive friction of the sharp particles of broken 
stone on the wooden surfaces of the molds, thus increasing the 
life of the molds and decreasing the cost of molding the con- 
crete." 

507. Oiling the Forms. Oiling or greasing the face of the 
mold, in order that the latter may be removed without detach- 
ing particles from the concrete face, is usually advisable. Soap, 
crude oil, linseed oil, bacon fat, are some of the materials that 
have been used for this purpose; the first mentioned probably 
gives the best results, and if not applied too freely will have no 
injurious effect upon either the finish or strength of the work. 
Applying shellac to the molds improves the appearance of the 
concrete surface. When the forms are lined with steel, the 
adhesion of the concrete to the lining is more difficult to over- 
come. In this case the ordinary oils are not entirely success- 
ful, but fat salt pork has been found to give satisfactory 
results. 

508. Joints and Corners. If desired, triangular strips may 
be nailed to the inside of the forms in such a way as to block 
off the face to represent stone masonry, and in this way the 
marks of joints between planks or between strips of lining may 
be avoided. Square corners should not be allowed on exterior 
angles, as it is difficult to so tamp the concrete as to make the 
corner perfect, and they are so likely to be chipped off. Tri- 
angular strips or moldings should be tacked along the corners 
of the mold as a fillet to cut off the corner by a plane making 
equal angles with the adjacent faces. This plane may be from 
one inch to two inches wide. 

To form water drips on projecting ledges, such as door caps 
and sills, abutment copings, etc., a small half-round should be 
nailed to the upper surface of the mold a short distance back 
from the projecting face. This leaves a ridge at the edge of 
the under side of projection so that the water must drip from 
the edge, and not follow back to the main wall face. 

509. POSTS AND BRACES. The sizes of posts and braces 
must be such as to make a practically unyielding support to 



TIMBER FORMS 369 

the sheathing. With one and three-quarters inch lagging, posts 
may be four feet apart; if five feet four inches apart (three to 
each sixteen foot length), some yielding of the sheathing may 
be expected if it is less than two and three quarter inches. 
If sheathing is four inches thick, the distance between posts 
may be six or seven feet. 

Fir, yellow pine, and Norway pine are suitable for posts. 
Three-inch by eight-inch is an ordinary size, and a post of 
these dimensions should be supported, either by ties or braces, 
at intervals of four to six feet. Where the posts are four inches 
by ten inches, supports may be six to eight feet centers, while 
with six-inch by twelve-inch posts, the distance between cen- 
ters of supports may be eight to ten feet. Posts should be 
sized and dressed on the side which is to receive the sheathing, 
in order that the alignment may be perfect. 

510. Methods of Bracing. The general plan of the mold 
may vary according to conditions, the following methods hav- 
ing been employed on heavy work to support the vertical posts: 
1st, With outside inclined braces, leaving the interior of the 
mold unobstructed. 2d, Tie rods across the interior of the 
mold connecting opposite posts at frequent intervals. 3d, Each 
post trussed vertically and tied across at top and bottom only. 
4th, Horizontal trussed wales outside of posts, spaced four to 
five feet apart in the vertical and tied across at the ends. 

511. Inclined Braces. The sizes of inclined braces depend 
on their lengths, the inclination to the vertical, and the amount 
of shoring used. An approximate rule for the size of braces 
under usual conditions and using ordinary dimension stuff, 
not boards, is that the number of square inches area of cross- 
section of brace should equal length of span in feet. If thin 
planks are employed, they should be in pairs, one on either 
side of the vertical post, and made to act together by cross- 
pieces nailed to the two planks. 

The aim should be to make the whole form practically un- 
yielding under the action of the tampers, as it has been found 
that this action is usually more severe than the mere pressure 
of the concrete in a semi-liquid condition. The sizes of pieces 
cannot, therefore, be accurately computed, but the above sizes 
are derived from the general result of experience as to what 
has proved satisfactory. 



370 CEMENT AND CONCRETE 

The advantage of the form of construction just described is- 
that the interior of the mold is left entirely unobstructed. On 
high walls, however, the amount of timber required for braces 
is excessive, and the braces may be almost as objectionable as 
tie rods, since the former prevent the laying of tracks along 
the side of the form. 

512. Tie Rods. When the vertical posts are supported by 
tie rods across the mold and the wall is thin, it may be possible 
in removing the mold to withdraw the bolts or rods if they 
have been thoroughly greased or wrapped with stiff paper be- 
fore the concrete is placed. If it is designed to leave the rods- 
in the concrete, they should be provided with sleave nuts near 
the end, which, when unscrewed, will leave the end of the rod 
within the concrete mass not less than two inches from the face. 
The hole left by the nut should be carefully filled with mortar 
after the mold is removed. 

With vertical posts four feet apart, this method of support 
is objectionable, as it leaves a network of ties within the forms 
interfering seriously with the operation of a skip and with the 
ramming. It is not necessary, however, to place all of the tie 
rods to the top of the mold before beginning the concreting, 
as it is sufficient to keep one or two rods in place above the 
plane where tamping is being done. 

A modification of this method is t6 use wires of large diam- 
eter with an eye at the end just inside the finished face of 
the concrete. A short bolt, with hook at one end and threaded 
at the other, passes through the post, hooks into the eye of the 
wire, and is tightened by a nut on the threaded end outside 
the post. After removing the nut, the rod is unhooked and 
the hole in the face filled with mortar, the wire remaining in 
the concrete. 

513. Trussed Posts. The third method of support, where 
the posts are trussed and provided with heavy tie rods at the 
top, and held at the bottom either by tie rods or some other 
means, seems to have fewer objections than the methods just 
described. Less timber will usually be required to build this, 
form than for that where inclined braces are used, and the 
obstruction to operations will usually be less than with either 
of the other styles. This mold is also very readily taken down,, 
though the posts are heavier and more difficult to handle. 



TIMBER FORMS 371 

To secure the bottoms of the posts, they may be set in the 
ground, or rest against sills braced to some other portion of 
the structure, or to piles. A suitable support may also be 
obtained by dumping a mass of concrete around the bottom 
of each post and allowing it to set. Forms erected on rock 
may have the posts rest against blocks bolted to the rock. 

514. Trussed Wales. The fourth method of supporting the 
posts is particularly applicable where the work is divided into 
blocks of moderate size in horizontal cross-section, say twenty 
feet square. In longer lengths the horizontal trussed wales 
become rather heavy for convenient handling. Within these 
limits, however, this is an excellent form. In the construction 
of Lock No. 2 between Minneapolis and St. Paul, 1 a form of 
this kind was used for blocks about twelve by fifteen feet at the 
bottom. The sheathing was one and three-quarters inches, 
lined with No. 20 galvanized iron. Verticals were four by 
twelve, spaced about two feet centers. The trussed wales were 
twelve by twelve inch, trussed with one and one-quarter inch 
rod, the king-post being of twelve by twelve inch about two 
and one-half feet long, making depth of truss three and one- 
half feet. The ends of opposite wales were connected by one 
and one-quarter inch rod passing outside of the sheathing. 
Each pair of the longitudinal wales was just above the corre- 
sponding pair of transverse wales, so that they did not inter- 
fere at the corners. The mold was twenty-nine feet in 
height. 

In describing this mold, Mr. Powell says: "One complete 
form weighs twenty-eight tons; each piece about seven tons. 
Each piece is moved separately by the cable-way in forty-five 
to sixty minutes. The operation of removing one complete 
form requires from three to four hours time. After being 
moved, a small crew of men occupy nearly a day in plumbing 
and bolting together the form." "The boxes containing 1.7 
cubic yards of concrete are landed on top of the form by cable- 
way and tipped from that position. Although the jar and 
strain is severe, the forms have shown no ill effects therefrom, 
remaining tight and secure." 



1 Major Frederic V. Abbot in charge. Mr. A. O. Powell, Asst Engr., 
Report Chief of Engineers, 1900, p. 2778. 



372 CEMENT AND CONCRETE 

ART. 63. DEPOSITION OF CONCRETE IN AIR. 

515. Transporting to Place of Deposition. In. depositing 
the concrete in place,. care must be taken not to undo the work 
of mixing. If the concrete is allowed to fall freely a distance 
of several feet or to slide down an inclined plane, the stones 
will be 'likely to separate from the mass, and the result will be 
a layer of broken stone followed by a layer of mortar. If the 
concrete is deposited in a pile, the stone will roll down the out- 
side of the cone. This action is especially bad in concrete that 
is mixed rather dry. The author has seen a pavement founda- 
tion in which the limits of each wheelbarrow load of concrete 
could be distinguished, the foundation presenting the appear- 
ance of the cross-section of a honeycomb, made up of irregular 
hexagons outlined by broken stone having a deficiency of 
mortar. In all such cases, if the action cannot be avoided by 
some other method of dumping, then care must be taken to 
remix the concrete. 

There is one method by which the concrete may be deposited 
by gravity without separation of the materials. This consists 
in allowing the material to slide down a tube, but the tube 
must be kept continually full, the concrete being allowed to 
run out at the bottom only as fast as it is filled in at the top. 
This method is only applicable where the mixing is continuous, 
as in the case of machine mixers. 

516. Sometimes it will be found possible to mix the con- 
crete so near to the place of deposition that it may be shoveled 
directly into place. In mixing by hand this is practicable, as 
the mixing platforms may usually be easily moved, and this 
method of deposition is carried out even in street work where 
the concrete is in thin layers and hence requires much moving 
of platforms. 

Where a machine mixer is used that is so mounted as to be 
portable, the concrete may be delivered in place by a belt 
conveyor. Such an arrangement for the building of walls and 
for foundations of pavements, has already been described in 
Chapter XIV. 

The conditions are usually such, however, as to preclude the 
possibility of mixing the concrete so close to the work that it 
may be shoveled into place or handled economically on a con- 



PLACING IN AIR 373 

veyor of the style mentioned. The next cheapest method is 
to use a derrick to handle skips or bottom-dump buckets, pro- 
vided the work is sufficiently concentrated to have one posi- 
tion of the derrick serve to place a large quantity of con- 
crete. The skips should hold about a cubic yard, and if a batch 
mixer is used, the skip should hold a batch, whatever that 
may be. 

517. If the concrete is mixed on the same level and within 
less than two hundred feet of the work, wheelbarrows may be 
used, but for greater distances, carts, or, what is usually cheaper, 
cars running on a track, should be employed. 

For large masses of concrete a cableway may be employed 
to advantage, provided there is sufficient use for it to repay 
the high original cost of plant. The selection to be made from 
among these common methods is dependent on economy as in 
handling other material, the only requirements being that the 
concrete shall be conveyed to place quickly, and that the ma- 
terials shall not be allowed to separate as a result of any of the 
manipulation. In laying large quantities of concrete, the dif- 
ference between success and failure from a financial standpoint 
may easily rest in the proper transportation of the materials 
to and from the mixer. 

518. Ramming. The concrete should be deposited in hori- 
zontal layers about six inches thick, leveled with a shovel and 
thoroughly rammed. The length of time ramming should be 
continued and the vigor with which it should be done depend 
largely on the degree of plasticity of the concrete. If the con- 
crete is made of such a consistency that when struck a smart 
blow with the back of a shovel a film of moisture will just show 
on the surface, it should have vigorous ramming to insure a 
compact mass. A flushing of water to the surface will then 
indicate when to cease tamping. 

With a little more water there is less danger of the larger 
stones " bridging" and leaving large voids in the mass, and 
less work will be required to flush water to the surface. With 
such a consistency, cutting the mass with a spade before start- 
ing, the ramming may assist in expelling air bubbles and pre- 
venting voids. With still wetter mixtures ramming becomes 
difficult, as the concrete will soon begin to quake, after which 
the ramming should not be long continued as the mass is then 



374 CEMENT AND CONCRETE 

semi-fluid, and the stones may gradually work themselves to 
the bottom of the layer, forcing the mortar to the top. 

519. Rammers are frequently made of wood, but those of 
iron are believed to be better. The weight of a rammer is 
limited by the capacity for work of the man who wields it. 
They are usually made to weigh from twenty to forty pounds. 
If a man lifts and drops a forty-pound rammer with forty square- 
inch face twenty times a minute, he is doing less good to the 
concrete than if he dropped a twenty-pound rammer with twenty 
square-inch face forty times a minute. If the face of the ram- 
mer exceeds thirty-six square inches, the result is apt to be a 
mere patting of the surface of the concrete, unless the rammer 
is so heavy as to require two men to operate it. Iron rammers 
with face, say, three by six inches, and weighing twenty to 
thirty pounds, are believed to be the most efficient. Still thinner 
rammers than this may be necessary in work involving such 
detail as for filling in between iron beams, and are desirable 
for tamping near the face of the mold. 

520. Rubble Concrete. In massive work the embedding of 
stones of " one-man size," or larger, in the concrete is a practice 
that has long been in vogue. The objection is sometimes made 
that this interferes with the homogeneity of the wall and that 
variations in expansion may result in injury to the work. It 
is thought, however, that in large masses this danger is more 
theoretical than real, and the author sees no objection to this 
form of construction for many purposes if properly carried out, 
and it is frequently permitted in important works. Thin walls, 
the arch rings of bridges, shallow foundations, etc., should not 
of course be built in this way, because the stresses to which such 
structures are subjected should be met by a uniform resistance, 
to avoid the effects of eccentric or irregular loading. In such 
structures as dams, lock walls, breakwaters, retaining walls, 
and in many cases bridge piers and abutments, the work may 
be considerably cheapened without sacrificing the fitness of the 
structure. The stones thus embedded should be perfectly 
sound and should not lie nearer one to another than six inches, 
nor should they lie nearer than this to the face of the wall. 
The concrete should be mixed rather wet, and much care taken 
that each stone is completely surrounded by a compact mass of 
concrete. The stones should be settled into the concrete al- 



PLACING IN AIR 375 

ready laid far enough to assure their having a full bed. Stones 
used in this manner are sometimes called " plums." 

521. Another class of rubble concrete differing from the 
above more in degree than in kind, is formed by placing large 
stones in the work, and filling the joints between them with a 
rather wet concrete in which spalls may be rammed if desired. 
The difficulty of obtaining a compact wall by this method is 
perhaps a little greater than, when smaller stone are used, but 
in either case if really water-tight work is desired, the inspec- 
tion must be thorough. 

The saving in cost by the use of rubble concrete depends 
upon the local conditions, but under ordinary circumstances 
when broken stone is employed, the cost of crushing the stone 
and the cost of cement, for a volume of concrete equal to the 
volume of the stone imbedded, are practically saved. 

522. Joints in Concrete. In the construction of large 
masses of concrete in place, joints cannot be avoided; that is, 
it is not possible to make the entire mass monolithic, as force 
enough could not be employed to carry up the entire struc- 
ture at once. Even if this were possible, it would not be de- 
sirable, since the changes in length of the wall due to changes 
in temperature would probably result in cracks which would 
be irregular in outline and mar the appearance of the wall, 
if they had no more serious effect. 

When the concrete is subjected to vertical forces only, as 
in foundations for buildings, horizontal joints are less objec- 
tionable than vertical joints. But in the construction of con- 
crete lock walls, dams, and breakwaters, vertical joints are de- 
sirable to confine the cracks to predetermined planes. In the 
building of such structures, therefore, the method has been 
adopted of dividing the work into sections of such horizontal 
dimensions as may be thought best, and completing each sec- 
tion as a monolith. This will sometimes require the contin- 
uous prosecution of work for twenty-four or forty-eight hours. 
Whether this method, involving work at night, which is always 
more expensive and usually less thorough, is justified by the 
end sought, depends upon the character of the structure. 

523. If this method is not adopted, and a horizontal plane 
of weakness is a serious defect, special means should be pro- 
vided for avoiding this plane of weakness. Such provision may 



376 CEMENT AND CONCRETE 

be made by iron dowels set in the concrete at the end of the 
days work and projecting above the surface to be covered by 
the concrete placed the next day; steps or hollows, or grooves 
parallel to the length of the wall, may be left to be filled by 
the next layer. Large stones weighing a hundred pounds or 
more are frequently imbedded one half their depth in the 
last layer of a days work to form a bond with the following 
layer. 

In any case special care should be taken to thoroughly wash 
and clean the surface of hardened concrete before continuing 
the work, using preferably wire brooms for this purpose and 
removing any stones at the surface that appear to be loose. 
A thin layer of rich cement mortar should then be laid upon it, 
into which the first layer of fresh concrete is well rammed. 

If the appearance of the finished face is of importance, special 
care must also be exercised in joining at this point. Before 
leaving a layer which is to be allowed to harden before contin- 
uing the work, the line limiting the height of the concrete at 
the face should be made perfectly horizontal, for a slight crack, 
or at least a noticeable line, may be expected at this point, and 
if not straight it will be the more unsightly. 

524. If for any reason a layer of concrete cannot be carried 
over the whole area of the wall or foundation, it should neve r 
be allowed to taper off to a wedge, but a plank equal in width 
to the thickness of the layer should be set on edge, firmly se- 
cured, and the concrete tamped against it. In the construction 
of arches, culverts and sewers, such stop planks may well be set 
normal to the surface of the intrados instead of vertical. In 
case more than one layer is left incomplete, they should be 
stepped back, making an offset for each layer of at least one or 
two feet. The concrete should never be built up on a smooth 
batter if new concrete is to be joined to it later. 

525. Keeping Concrete Moist. All concrete should be kept 
moist from the time it is in the wall until it has become well 
hardened. Surfaces exposed to the air should therefore be 
sprinkled frequently for at least several days after placing. 
An excellent practice is to cover the surface with burlaps which 
may be kept saturated, as this not only furnishes the necessary 
moisture, but protects the work from the direct rays of the 
sun. The interior of a large mass will probably take care of 



SURFACE FINISH 377 

itself in this regard, but the precaution has sometimes been 
taken of leaving vertical holes or wells in the mass, which are 
kept filled with water for some weeks and are then filled with 
concrete. 

526. FINISH. Some of the precautions that must be taken 
to secure a good finish to the face of concrete work have al- 
ready been mentioned in considering the forms and the meth- 
ods of deposition. These are usually supplemented, however, 
by certain special means when the appearance is of much im- 
portance. 

We must say first, that the application of a plaster of ce- 
ment mortar to a finished and set concrete face will almost 
never be permanent. It is seldom that it will adhere with suf- 
ficient strength to prevent scaling due to differences in expan- 
sion of the materials of different composition and age. If 
plaster must be used on the face of a wall, it should be applied 
before the concrete has set, but it is safer to avoid plastering. 
It is of course advisable to fill with rich mortar any voids that 
may appear in the face of the work, but such places should be 
few. 

If the molds are removed while the concrete is still moist, 
the face may be coated with a thin grout and then immediately 
scraped off with the edge of a trowel. This results in filling the 
small voids in the face of the work, but does not leave a coat of 
plaster on the surface to scale off. 

527. A good finish may be obtained when the molds are 
smooth if the workmen will force the blade of a spade or shovel 
between the fresh concrete and the mold, and pull the handle 
away from the mold. This has the effect of forcing the large 
stone back from the face and allowing the mortar to flow in. A 
layer of mortar is thus left next the mold with no marked line 
of junction between mortar and concrete, as may be the case in 
using a mortar facing. A similar effect may be produced by 
throwing the concrete against the face of the mold with such 
force that the ' larger pieces of aggregate rebound. In very 
finely finished work this may mar the surface of the sheathing, 
but ordinarily this method is effective. 

528. When a special layer of mortar is used for facing, there 
is more danger, perhaps, of making the layer too thick than too 
thin. As to the richness of the mortar, two parts sand by 



378 CEMENT AND CONCRETE 

measure to one volume packed cement is usually sufficient 
though a more glossy finish may be made if desired, by using 
equal parts of cement and sand. It is better to avoid too great 
a variation between the richness of the mortar used for facing 
and that used in the body of the concrete. 

One of the best ways of applying such a layer is to prepare a 
sheet of steel of width equal to the thickness of one layer of con- 
crete, usually six to eight inches, with two handles on the upper 
edge to facilitate moving it. At the ends of the sheet, on the 
side next the mold, rivet short pieces of 1 in. by 1J in. or 2 in. 
by 2 in. angle iron. This sheet of iron with the projecting legs 
of the angles against the face of the molds, forms, with the 
latter, a space one and one-half or two inches thick, which 
is to be entirely filled with the finishing mortar made rather 
moist and tamped lightly with edge rammers. The concrete is 
filled in behind the iron, after which the latter is withdrawn by 
means of the handles, and the whole mass is thoroughly rammed. 
The end sought is that the finishing mortar shall have some 
approximately definite thickness, and that the stones of the 
concrete shall be tamped into the finishing mortar, but not 
through it, and thus destroy any sharp line of demarcation 
"between mortar and concrete, and ensure a perfect bonding of 
the two. It is evident that this can only be accomplished by 
placing the mortar and concrete at the same time. 

529. One other cautionary remark concerning the use of fin- 
ishing mortar. With the present state of our knowledge con- 
cerning the rates of expansion of mortars and concretes of dif- 
ferent composition, it is not considered wise to use too many 
'combinations in the same structure. To illustrate, a pavement 
or surfacing of a large concrete structure was once built in layers 
as follows: first, thick natural cement grout was placed on the 
concrete foundation; second, natural cement concrete; third, 
Portland cement concrete; fourth, a richer Portland cement 
concrete; fifth, Portland granolithic; sixth, rich Portland mortar; 
and seventh, floated with dry Portland cement and sand. We 
cannot be absolutely sure that this is bad practice, but it would 
seem that this structure might have served its purpose with 
fewer varieties of -material, and it is usually considered very 
doubtful whether Portland cement mixtures will always ad- 
here well to mixtures of natural cement, although the author 



SURFACE FINISH 379 

knows of instances where they have been used in juxtaposition 
apparently with good results. 

530. Granolithic is a facing or surfacing mortar composed 
of crushed granite and cement. The granite is usually specified 
to contain no particles larger than f inch to one inch, and about 
one and one-half to two and one-half parts are used to one vol- 
ume of cement. This is more frequently used for foot walks 
and other places where resistance to wear is required, but may 
also be used to surface walls, to line reservoirs, etc. It will be 
mentioned again in connection with cement sidewalk construc- 
tion. 

531. Exposed concrete surfaces frequently present a patchy 
appearance. This may be the result of lack of care in placing 
the concrete next the mold, or it may be due to variations in 
the purity of the sand or in the amount of water used in mixing. 
On mortar-faced work this lack of uniformity is less noticeable. 
The use of slag sand, or of a little fine pozzolanic material, may 
be advantageous, and a small amount of lampblack in the facing 
mortar also tends towards uniformity in appearance. 

A very pleasing finish may be given by applying to the set 
concrete a thin wash of cement and plaster of Paris, though the 
permanence of such a wash may be open to question. The 
sheathing should be removed as early as it is perfectly safe to 
do so, and the concrete surface cleaned from any oil or grease 
that may have come from the mold planks. The wash, which 
should be very thin, may be applied with a whitewash brush. 
A mixture of equal parts Portland cement and plaster of Paris 
gives a very light gray finish, and one part plaster to three parts 
cement gives a trifle darker shade. 

532. A rubbed finish of excellent appearance may be given 
by removing the sheathing before the concrete has set very 
hard, say after twenty-four to forty-eight hours, and rubbing 
the surface with white brick or with a wooden float. If there 
are small voids in the surface, it may be covered with a thin 
grout of equal parts of cement and sand and then rubbed 
hard with a circular motion. The grout should not leave a 
scale on the work, the object being only to fill surface imperfec- 
tions. 

If the mold boards are removed at just the proper time, a 
good finish may be given by rubbing with a wooden float, with- 



380 CEMENT AND CONCRETE 

out the coating of thin grout. A somewhat similar effect is 
produced by brushing the surface with brooms or stiff brushes. 

533. " Pebble-dash." What is called a pebble-dash finish 
was used in the construction of a bridge in the National Park at 
Washington, D. C. 1 Eighteen inches of the concrete next the 
face was made of one part cement, two parts sand, and five parts 
of gravel and rounded stone from one and one-half to two inches 
in their smallest diameter. After the removal of the forms the 
cement and sand were brushed from around the face of the gravel 
next the surface exposed to view. It was found by experiment 
that the brushing should be done when the concrete was about 
twenty-four hours old. At twelve hours the gravel was displaced 
by the brushing, and after thirty-six hours the mortar had be- 
come so hard as to be removed from the surface of the stones 
with difficulty. The forms were therefore designed so that 
sections of the lagging could be removed as desired. The cost 
of the brushing was said to be about sixty cents per square yard. 

A somewhat similar method is employed in giving to con- 
crete the appearance of cut stone. The materials used in the 
surfacing mortar are Portland cement and crushed rock, the 
character of the rock depending upon the color and texture de- 
sired in the finish. The molds having been removed after the 
proper time has elapsed, the mortar covering the face of the 
particles of crushed rock is removed by brushing or by washing 
the surface with a weak acid solution, followed by clean water, 
and finally by an alkaline solution to prevent any further action 
of traces of the acid which might be left on the face. This last 
method is said to be patented, "the patent covering the obtain- 
ing of a natural stone finish for concrete by mechanical, chemi- 
cal or other means." 2 It is hoped that such a blanket patent is 
somewhat less formidable than it appears. 

If the sheathing planks of the molds can be removed about 
twenty-four hours after the concrete is placed, the same effect 
may be produced without the use of acid. By using plenty of 
water the cement and finer portions of crusher dust in the face 



'Capt. Lansing H. Beach, Corps of Engrs., U. S. A., in charge. Work 
described by Mr. W. I. Douglas, Engr. of Bridges, D. C., Engineering News, 
Jan. 22, 1903. 

2 Engineering News, May 21, 1903. 



SURFACE FINISH 381 

may be washed out with a stiff corn broom, leaving the facets 
of the crushed rock exposed. 

534. Pointing and Bush-hammering. If the molds have 
been left in place until the concrete is set hard and it is found 
that the face of the concrete is not what is desired, it may still 
be improved although it may not be plastered. With this ob- 
ject the face is sometimes tooled with the stone cutter's point to 
give the appearance of rough pointed or rock face masonry. 
Grooves may be cut to block off the work into rectangles of the 
proper size, then a draft of one to two inches may be left along 
all of these artificial joints, and within the draft line the rough 
pointing may be done. 

A cheaper method, however, is to bush-hammer the entire 
face, and this tends to mask any lack of uniformity in color or 
smoothness. Bush-hammering may be done by ordinary labor- 
ers at a small cost, as one man can go over from fifty to one 
hundred square feet in ten hours, making the cost from If cents 
to 3 cents per square foot, with labor $1.75 per day. Where it 
is decided beforehand to bush-hammer the work, less pains need 
be taken in dressing the lagging of the forms. 

535. Colors for Concrete Finish. The addition of coloring 
matter to cement and concrete is not at present widely prac- 
ticed, and consequently experience has not been sufficient to in- 
dicate just what colors may be used without detriment to the 
work. Lampblack has been most commonly employed, giving 
different shades of gray according to the amount used. In any 
large work where the use of coloring matter is desirable and 
there is not time to institute thorough tests, the advice of a 
cement chemist should be sought. The dry mineral colors, 
mixed in proportions of two to ten per cent, of the cement, 
give shades approaching the color used. Bright colors are diffi- 
cult to obtain and would not be in keeping with a masonry 
structure except in architecture. 

When mixed with an American Portland cement mortar, 
containing one part cement to two parts by weight of a yellow 
river sand, the particles of which are largely quartz, the colors 
indicated in the following table are obtained. 

With no coloring matter added, the mortar was a light green- 
ish slate when dry. Ultra marine green, in amounts up to 8 
per cent, of the cement, had no apparent effect on the color of 



382 



CEMENT AND CONCRETE 



this mortar. Variations in character of cement and sand will 
affect the result obtained in using coloring matter. The colors 
indicated below are for dry mortars ; when the mortar is wet 
the shades are usually darker. None of the materials mentioned 
in the table seems to affect the early hardening of the mortar, 
though very much larger proportions might prove injurious. 
With red lead, however, even one per cent, is detrimental, and 
larger proportions are quite inadmissible. 

COLORED MORTARS. 

Colors Given to Portland Cement Mortars Containing Two Parts 
River Sand to One Cement. 







|d 


DRY 


WEIGHT OF DRY COLORING MATTER TO 100 POUNDS OF CEMENT 


0,0 

~ 


MATERIAL 




l 


USED. 




*l 












J Pound. 


i Pound. 


2 Pounds. 


4 Pounds. 


11 










Dark Blue 




Lamp Black 


Light Slate . 


Light Gray 


Blue Gray . 


Slate . 


15 


Prussian 


Light Green 


Light Blue 




Bright Blue 




Blue . . 


Slate . . 


Slate . . . 


Blue Slate . 


Slate . 


50 


Ultra Marine 




Light Blue 




Bright Blue 




Blue 




Slate 


Blue Slate 


Slate 


20 


Yellow 












Ochre 


Light Green . 






Light Buff 


3 


Burnt 


Light Pinkish 




Dull Laven- 






Umber . 


Slate . . 


Pinkish Slate . 


der Pink 


Chocolate . 


10 


Venetian 


Slate, Pink 


Bright Pinkish 


Light Dull 






Bed . . 


Tinge . . 


Slate . . . 


Pink . . 


Dull Pink 


2J 


Chattanooga 
Iron Ore . 


Light Pinkish 
Slate . . 


Dull Pink . . 


Light Terra 
Cotta . . 


Light Brick 
Red . . 


2 










Light Brick 




Red Iron Ore 


Pinkish Slate 


Dull Pink . . 


Terra Cotta 


Red . . 


8* 



536. In some cases it may be sufficient to color the surface 
of the work by painting. Ordinary oil paints are sometimes 
applied after washing the surface of the wall with very dilute 
sulphuric acid, one part acid to 100 parts water, but the per- 
manence of such a finish seems very questionable. 

The method of obtaining a gray finish by painting with a 
thin grout of cement and plaster of Paris has already been de- 
scribed ( 531). Similar methods may be used with the dry 
mineral colors, and, while their permanency cannot be vouched 
for, it seems a more reasonable procedure than to paint a can- 



PLACING UNDER WATER 383 

crete surface with oil paints. One pound red iron ore to ten 
pounds cement mixed dry, and then made into a very thin grout 
and applied to a well cleaned concrete surface with a white- 
wash brush, gives a pleasing brick-red color; and a rich dark 
red is given by one pound red iron ore to three pounds cement. 
The earlier this is applied after the concrete has set, the more 
likely is it to remain permanent. 

ART. 64. PLACING CONCRETE UNDER WATER 

537. In building a concrete structure under water where the 
site cannot be coffered, it must be expected that the expense 
of the work will be increased, and the quality of concrete poorer. 
The methods employed for subaqueous construction are: 1st, 
the laying of freshly mixed concrete in roughly prepared forms; 
2d, placing the fresh concrete in bags of burlap or canvas which 
are deposited while the concrete is still soft; and 3d, molding 
in air concrete blocks which are placed in the work when well 
set. 

In the first method some cement will certainly be washed 
out of the concrete, the extent of this loss depending upon the 
condition of the water in which the work is done (i.e., its depth 
and the amount of current and wave action) and the care with 
which the concrete is lowered to place. Tamping cannot be 
done with this method, and any movement of the concrete to 
level it will cause further loss of cement. 

In the second method the loss of cement will be much less, 
but the adhesion between the different masses will be slight. In 
the third method there is no loss of cement and the concrete 
can be well rammed; but if small blocks are used, there may 
be difficulty in so placing them under water as to make a solid 
structure, while if large blocks are used, special hoisting ma- 
chinery is required to handle them. 

538. DEPOSITING IN PLACE. The first method mentioned 
above, depositing fresh concrete in place, is usually the cheap- 
est and most expeditious method, though it is not likely to 
give the best results. When concrete is lowered through water, 
there is a tendency for the cement to separate from the sand 
and stone. This tendency seems to be exhibited in a more 
marked degree with some cements than with others. In con- 
nection with the construction of the concrete foundations of 



384 CEMENT AND CONCRETE 

the Charlestown bridge, a test was devised for determining the 
relative values of the different lots of cement for depositing in 
water. 1 Concrete was laid, through a small chute, in a cement 
barrel placed in a hogshead filled with salt water. It was found 
that while some specimens would retain their form after twenty- 
four hours when the barrel was removed, others showed but 
little cohesion after twenty-four to forty-eight hours. In the 
former, the cement and gravel remained well distributed through- 
out the mass, but in the latter much of the cement had sepa- 
rated from the gravel, and settled in the bottom of the barrel, 
where it remained in an inert state, while the central portion of 
the concrete, robbed of its cement, had many voids. As a 
result of this test, some lots of cement were not accepted for 
use. 

The finest portion of the cement is very liable to separate 
from the remainder as the concrete passes through the water, 
and if subjected to the action of waves or a current, much of 
the cement will be washed away. In exposed situations it is 
especially necessary to inclose the site of the work with sheet pil- 
ing or cribs, or a wall constructed by the bag or block method. 
When the water level outside the form is constantly changing, 
the flow of water through the joints in the sheathing is especi- 
ally effective in washing out the cement, and in such conditions 
the sheathing should be made as nearly water-tight as possible. 
To this end tongue and groove lagging may be used, or the face 
of the mold may be covered with tarred felt, or canvas, tacked 
in place. 

539. Laitance is the term applied to the whitish spongy 
material that is washed out of concrete when it is deposited in 
water. Before settling on the surface of the concrete, which 
it does slowly, it gives to the water a milky appearance, hence 
the name. In fresh water this semi-fluid mass is composed of 
the finest flocculent matter in the cement, containing generally 
hydrate of lime. It remains in a semi-fluid condition for a 
long time and acquires very little hardness at the best. In sea 
water the laitance is more abundant and is made up of silica, 
lime and magnesia, with carbonic acid and alumina, its exact 



1 Report of Mr. William Jackson, Chief Engineer. Third Annual Report 
Boston Transit Commission. 



PLACING UNDER WATER 385 

composition depending upon the character of the cement. This 
interferes seriously with the bonding of the layers of concrete, 
and when it has settled it should be cleaned from the surface 
before another layer is placed. 

540. The Tremie. A method frequently employed to pre- 
vent, as much as possible, the loss of cement, is to make use of 
a large tube of wood or sheet iron, made in sections so that 
its length is adjustable, and provided with a hopper at the 
upper end. Such a tube is called a tremie. The hopper is 
always above water, and the lower end of the tube, which may 
also terminate in a hopper, rests upon the bottom of the founda- 
tion. 

The tremie is first filled with concrete, a box placed over the 
lower end serving to prevent the escape of the concrete while 
the tube is being lowered until the end rests upon the bottom. 
The tube is then lifted from the bottom sufficiently to allow 
the concrete to escape as fast as fresh concrete is added at the 
top. The surface of the concrete in the tube should be kept 
continuously above the v/ater surface. The tremie may be 
held in position by a crane, or it may be so supported as to al- 
low of two motions at right angles to each other. Such an 
arrangement was used in building the piers for the Boucicault 
Bridge, the tube traveling along a platform, which in turn 
could move on a track at right angles to the first motion. In 
using a tremie the thickness of a layer may be regulated at will. 

In the construction of the Charlestown Bridge * a tube was 
used fourteen inches in diameter at the bottom, and about 
eleven inches in diameter at the neck, above which was a hopper 
to receive the concrete. When the attempt was made to place 
too thick a layer at one operation, it was found that the charge 
was likely to be lost, and the best results were believed to be 
obtained with layers two feet to two and one-half feet thick. 
Some experiments were made with a plug designed to keep the 
water from flowing up through the concrete when the tube 
was being refilled after a loss of the charge. This plug was 
made with a central core of wood and sides of canvas expanded 
by steel ribs. It worked fairly well, but its use was not con- 
tinued. 



Third Annual Report, Boston Transit Commission. 



386 CEMENT AND CONCRETE 

541. This principle was employed by Mr. Daniel W. Mead 
in placing concrete in a small shaft in ninety feet of water. 1 
An eight inch, wrought iron pipe was screwed together in sec- 
tions, and provided with a hopper at the upper end and a wooden 
plug at the lower end. After lowering the pipe into the shaft, 
the pipe was filled with concrete and it was expected that its 
weight would force out the plug at the bottom when the pipe 
was raised. On the first attempt, however, the plug failed ta 
drop out, and on raising the pipe the cause was apparent. The 
plug had evidently leaked, and as the first concrete was dropped 
into the pipe it had separated, the broken stone being at the 
bottom, the sand next, and the cement above had so plugged 
the pipe as to support the weight of the concrete. The second 
attempt, when a tighter wooden plug was used and a small 
pipe placed inside the larger one to assist in loosening the plug 
if necessary, was successful. 

542. The Skip. Since in submerged work the concrete 
should be deposited in as large masses as possible, the use of a 
large skip will probably give better results than the tremie. 
A box form may be used with hinged lids at the top to permit 
filling, and two hinged doors at the bottom which may be 
opened from the surface by a tripping rope when the box has 
reached the place for depositing the concrete. 

A convenient form of skip is made in two halves, each half 
having a cross-section either of a right angled triangle or a 
quadrant of a circle. The two boxes are hinged at their upper 
inside corners and the pieces through which the hinge rod 
passes are prolonged upward, the lowering cables being at- 
tached to their ends. Two opening cables are fastened to the 
outer corners of the boxes. Two sheets of iron may be used 
as covers to the boxes, being attached to the hinge rod that 
serves for the two halves of the skip. 

It is seen that the skip will work on the principle of a pair 
of ice tongs. While being filled with concrete the box is sup- 
ported by the lowering cables, and the hinged lids are kept up 
by some simple contrivance. When full, the lids are closed 
and the skip lowered till it rests on the bottom; the skip being, 
then hoisted slowly by means of the opening cables, the con- 

1 Trans. Assn. of Civil Engineers of Cornell University, 1898. 



PLACING UNDER WATER 387 

crete is gently deposited in place. Such skips are supplied by 
the makers of concrete machinery. 

543. In depositing concrete by means of skips it is well to 
have the latter of large size, holding not less than a cubic yard, 
and preferably two cubic yards or more. The larger quantity 
will compact itself better on account of the greater weight, 
and the surface which is subjected to wash will have a lesser 
area in proportion to the volume of the mass. The skip should 
be completely filled with concrete and tightly closed while it is 
being lowered. It. is important also that the skips be lowered 
slowly, in order that the inclosed air may be replaced by water 
without commotion. 

544. The Bag. Mr. Wm. Shield * devised a bag for de- 
positing concrete under water which is said to work very satis- 
factorily. The top of the bag is closed, and has a three-quarter 
inch wrought iron bar fastened across the end with a loop to 
receive the hook of the lowering line. The mouth of the bag 
is slightly larger than the upper end, to facilitate the discharge 
of the concrete. The bag is inverted to be filled, and the mouth 
is then secured by a turn of a line provided with loops through 
which a small tapering pin is passed. This pin is attached to 
a tripping line, and when the bag has reached the pla.ce of 
deposition, a pull of the tripping line releases the pin; when 
the bag is gently lifted, the concrete is deposited in place with 
such slight commotion that but little cement is said to be lost. 

545. Other Methods of Depositing in Situ. For deposition 
under water the materials for concrete are sometimes mixed 
dry, but this is not good practice. The mere soaking of water 
into cement does not form a compact mortar; the moistened 
materials need to be thoroughly mixed and, if possible, rubbed 
together in order to obtain perfect adhesion. Then, too, if the 
dry materials are lowered to place and water is suddenly al- 
lowed access to the mass, much of the cement will be washed 
away in the disturbance caused by the sudden inrush of water. 

M. Paul Alexandre 2 found by experimenting on mortars of 
"dry" (stiff), "wet" and "ordinary consistency," that mortars 



1 "Subaqueous Foundations," London Engineering, 1892. Abstract in 
Engineering News, Vol. xxviii, p. 379. 

2 "Recherches Experimentales sur les Mortiers Hydrauliques," par M. Paul 
Alexandre, pp. 93-96. 



388 CEMENT AND CONCRETE 

mixed "dry" suffered the greatest decrease in strength by im- 
mersion in running water. Mortars mixed "wet" suffered the 
least loss, though their resistance was less than those mixed to 
the ordinary consistency, since when not subject to the current 
of water, the wet mortars gave much lower results than those 
of ordinary consistency. 

546. In order to avoid the washing out of the cement, the 
concrete is sometimes allowed to partially set before deposition. 
Mr. Robert W. Kinipple has used this method and advocates 
its adoption. 1 In employing this method, the concrete, which 
should be deposited when of the consistency of stiff clay, re- 
quires careful watching that it does not set so hard as not to 
reunite after being broken up. Under ordinary supervision, 
this will probably not prove as successful as some of the other 
devices, but it may be found valuable under certain circum- 
stances. The writer made a few experiments with this method 
on a small scale in swiftly running shallow water. Much of the 
cement appeared to be washed out by the current, but the 
results were somewhat better than were obtained when the 
concrete was deposited fresh. (See 456.) M. Paul Alexandre 
made some short time experiments on this point, which indi- 
cated that but little advantage was gained in allowing the 
cement to partially set before deposition. 

547. DEPOSITING CONCRETE IN BAGS. The second method 
of depositing concrete under water, namely, by placing the freshly 
mixed concrete in coarse sacks and immediately lowering them 
to place, is very convenient under certain conditions. This 
method is of especial value in leveling a foundation to receive 
concrete blocks, or to form a base for concrete deposited in situ. 
Small bags of concrete have been successfully used in filling the 
spaces between pile heads which were to support an open caisson. 
In such a case the bags should be lowered to a diver who places 
and rams them. If the bags be properly leveled and the earth 
firm, a part of the load is thus transmitted to the material be- 
tween the pile heads, while if the earth be very unstable, the 
bag construction compels the piles to act together, giving lateral 
stiffness to the foundation and tending to prevent over turning. 



"Concrete Work under Water," Proc. Inst. C. E., Vol. Ixxxvii. See 
also "Notes on Concrete," by John Newman, pp. 116 and 117. 



PLACING UNDER WATER 389 

548. The bag method was successfully used in replacing 
with concrete the timber superstructure of the breakwater at 
Marquette, Mich. 1 The main portion of the breakwater was 
built of monolithic blocks on the rock-filled timber substruc- 
ture. After removing a portion of the rubble filling, a bed was 
made for the monolithic blocks by laying concrete in place two 
feet thick, extending from one foot below to one foot above 
low water datum. This method was afterward replaced by 
the use of concrete in bags, which made it safe to remove a 
lesser amount of the rock filling of the crib at the center, and 
thus decreased the expense of the work. The bags were of 
eight ounce burlap made 6 feet long and 6 feet 8 inches in cir- 
cumference, and held about one ton of concrete. They were 
filled while lying on a skip specially constructed, so that when 
the skip was in place it could be tripped and the bag placed in 
its exact position in the work. 

549. In connection with this work a practical indication of 
the character of the concrete deposited in this manner was 
given by some small bags of concrete that were laid to protect, 
during the winter storms, a portion of the crib filling. Mr. 
Coleman says of this, 2 "Only one layer of these sacks, laid 
slightly overlapping from the lake side of the crib, was used. 
The sacks were so lightly filled that when laid as described, the 
average thickness of the concrete covering was not more than 
six inches. The crib was storm swept many times without- 
displacing a single sack', and when they were removed in the 
following spring to facilitate the work, they came away, when 
pulled up with the floating derrick, a dozen or more at a time, 
so firmly were they cemented together, and in many cases 
large rubble stones were lifted up along with them, because of 
the adhesion of the cement to their surfaces." 

550. The Cost of the concrete in bags was as follows : 

Materials, cement, sand, stone, burlaps, etc $5.281 

Mixing concrete and filling bags 912 

Transportation 157 

Depositing 408 

Total cost per cubic yard $6.758 

Or, cost in bags, exclusive of materials . . . . , 1.477 



Major Clinton B. Sears, Corps of Engineers, in charge; Mr. Clarence 
Coleman, Asst. Engineer. 

a Report Chief of Engineers, U. S. A., 1897, p. 2620. 



390 CEMENT AND CONCRETE 

The cost of the first plan, placing a two foot layer of con- 
crete in situ, where different methods of handling were em- 
ployed, was, for labor: - 

Loading scow with materials $0.411 

Mixing concrete 846 

Depositing . .524 

Cost in situ, exclusive of materials $1.781 

551. When concrete bags are used in forming a foundation, 
the lower layers should usually cover a considerably greater 
area than that required for the top. Especially is this true if 
building upon insecure earth. This increased area at the bot- 
tom may be obtained by building the sides on a batter, or by 
the use of. footing courses. If the latter are used, they should 
be so designed that in any case the projection beyond the course 
next above is not greater than the thickness of the layer. 

Before filling the concrete into the bags it should be thor- 
oughly mixed, as for deposition in the ordinary manner. The 
practice of using dry concrete for this purpose is reprehensible 
for the same reason as has been given in 545. It has also been 
found that if the concrete is mixed and filled into the bags in 
a dry state, a layer of concrete on the outside may cake before 
the water has had time to reach the interior portion. ' The 
bags should be filled about three-fourths full, leaving the mass 
free to adjust itself to inequalities in the rock, or to the irreg- 
ular surface of the previously deposited layer. When strength 
and compactness are desired, the bags should be placed by a 
diver and gently rammed. In this way the mass may be well 
bonded by " breaking joints." 

552. Large Masses in Sacks. Very large bags of concrete 
are sometimes employed, as in the construction of a breakwater 
at New Haven, England. 1 "The top of the breakwater has a 
width of thirty feet, is ten feet above high water, and is sur- 
mounted by a covered way and parapet running along the outer 
side, both sides battering one in eight. The breakwater is 
unsheltered from the force of the Atlantic, is founded on the 
rough, natural sea bottom, and the foundation course has a 



1 From London Engineering, quoted in Engineering News, Vol. xxvii, 
p. 551. 



PLACING UNDER WATER 391 

width of fifty feet; the lower portion of the structure, from the 
bottom up to a level of two feet above low water, consists of 
one-hundred-ton sacks of concrete deposited while plastic. 
The canvas with which the concrete was enveloped was of 
jute, weighing about twenty-seven ounces per square yard. 
The sacks were dropped into place by a specially designed 
steam hopper barge. The 'sack-blocks' in the finished work 
became flattened to a thickness of about two feet six inches. 
With the exception of this sack work the breakwater is built 
of plastic concrete laid in situ." Similar sack-blocks of one 
hundred sixty tons have been employed in breakwater con- 
struction. 

It is evident that this method of depositing concrete in 
large sacks is peculiarly suited to forming a foundation on a 
soft bottom, since, if the bags are made to project well beyond 
the sides of the molded concrete to be deposited above, they 
act in the double capacity of a mattress to prevent scour, and 
a foundation for the upper part of the structure. 

553. Other uses for Bags of Concrete. In the construc- 
tion of the Merchants' Bridge at St. Louis, bags of concrete 
were used to check the scour which occurred beneath the up- 
stream cutting edge of one of the caissons while it was being 
grounded. The bags were thrown into the river at such a dis- 
tance above the pier that they settled to the bottom at the 
point where the scour was taking place. 

Burlap bags were used at St. Marys Falls Canal for laying 
concrete under water next the face of the form to prevent 
washing of the cement in building concrete superstructure for 
canal walls. As the bags were placed by hand they were made 
to hold only about two cubic feet of concrete. 

554. Paper Sacks. Paper sacks are sometimes employed 
instead of jute bags. Dr. Martin Murphy * describes the meth- 
ods employed in filling steel cylinders for the substructure of 
the Avon Bridge as follows: "Bags made of rough brown paper 
well stiffened with glucose, were employed and slipped into the 
water over the required place of deposition. Each bag held 
about one cubic foot of concrete; smaller ones were used be- 



"Bridge Substructure and Foundations in Nova Scotia," by Martin 
Murphy. Trans. A. S. C. E., Vol. xxix, p. 629. 



392 CEMENT AND CONCRETE 

tween dowels. The bags were quickly made up and dropped 
one after another, so that the one following was deposited 
before the cement escaped from the former one. The paper 
was immediately destroyed by submersion, and the cement 
remained; it could not escape. The bags cost one dollar thirty- 
five cents per hundred, or thirty-five cents per cubic yard." 
The success of this method will depend upon the character of 
the sacks, for in some experiments on a small scale with sacks 
of stiff manila paper the author found that the bags were not 
destroyed, and that no adhesion took place between the separate 
sacks. 

555. THE BLOCK SYSTEM OF CONCRETE CONSTRUCTION. - 
The advantage of the block system of construction lies in the 
fact that the individual blocks may be made with the greatest 
care, and as they are allowed to harden thoroughly before being 
put in place, the loss of cement incident to the other systems 
is avoided. There is, however, the difficulty of forming a joint 
between adjacent blocks. The joints are of great importance 
when small blocks are employed, since the latter may not have 
sufficient weight to escape being washed out of the work. Large 
blocks may make a very solid structure by being simply super- 
imposed, but special hoisting machinery will be required to 
place such blocks. 

Sometimes a large bed of mortar is laid in coarse sacking 
and carefully lowered and spread on the block last laid, the 
next block being placed upon it immediately. A very rich 
mortar should be used for this purpose. Usually, however, it 
is not attempted to place mortar in the horizontal joints in 
concrete block work laid under water, but it is considered that 
all vertical joints should be filled with rich Portland cement 
mortar when the work is to be exposed to wave action. If 
settlement is anticipated, and large blocks are used, no attempt 
should be made to break joints in the direction of the longer 
dimension of the work, but the blocks should bond in a direc- 
tion transverse to the wall. Concrete blocks may be advan- 
tageously employed to form the faces of a structure built under 
water or exposed to wave action, the concrete hearting or 
backing being built in situ. 

556. For convenience in handling, a groove to receive a 
chain or cable should be left down two sides and across the 



PLACING UNDER WATER 393 

bottom of the blocks to enable them to be placed close together 
and to facilitate the withdrawal of the hoisting chain. These 
grooves may afterward be filled with concrete; such recesses 
are sometimes molded for the sole purpose of filling them with 
fresh concrete when in place, and thus binding the work to- 
gether. The molds for forming the blocks should be carefully 
made in order that the finished blocks may have good bearings 
one upon another. If the corners are rounded, they are less 
likely to be chipped off in handling or by having an undue 
strain come upon the corner when in place. 

If any recesses are desired in the blocks, the pieces placed 
in the mold to form them should be trapezoidal in cross-section 
with the longer parallel face against the side of the mold. If 
such filling pieces are made rectangular, difficulty will be ex- 
perienced in removing them when the concrete has set. The 
molds should, of course, be so constructed as to be readily 
taken apart to be used again. The opposite sides may be kept 
from spreading by rods which pass through the mold, but such 
rods are an inconvenience in packing the concrete into the 
mold, and it is therefore better to truss the mold outside. If 
such tie rods are used, they may be left imbedded in the con- 
crete, or removed with the mold, as desired. 

557. Cost of Molding Blocks. An illustration of the use of 
the. block method is furnished in the United States breakwater 
at Marquette. 1 The general plan of this breakwater has already 
been briefly noted and two methods of laying a two foot layer 
of subaqueous concrete, as a foundation for monolithic blocks 
forming the superstructure proper, have been described. A 
third method was to mold footing blocks, seven feet by five feet 
in section and two feet high, which were afterward laid flush 
with the lake side of the substructure cribs and filled in behind 
with concrete laid in place. The footing blocks thus assured 
a good quality of concrete beneath the toe of the monolithic 
block on the lake side where it was most necessary to provide 
a good foundation, and also served as a protection behind which 
the remainder of the two foot layer could be placed with greater 
facility. 

Many of these blocks were built during the winter in a shed 



Report Chief of Engineers, 1897, p. 2624. 



394 CEMENT AND CONCRETE 

artificially heated, the materials being thawed out as required. 
The molds were of six by six inch and four by four inch pine, 
lined with two by eight inch plank dressed on one side. Strips 
of trapezoidal cross-section, nailed inside the mold, provided for 
two parallel grooves on the bottom and two sides of the block 
to receive hoisting chains. A dovetail at the back of the block 
was also formed by three wedge-shaped pieces placed against 
the back face of the mold. The cost per cubic yard of making 
forty blocks is as follows: 

1.42 bbls. Portland cement, at $2.75 $3.90 

.45 cu. yd. sand, at $0.45 20 

1.0 cu. yd. stone screenings passing f" sieve, at $1.10, 1.10 

Cost materials in concrete per cu. yd $5.20 

Superintendence, labor and watchman $2.21 

Fuel 31 

10 per cent, of cost of warehouse and molds 52 

Total cost of making per cu. yd 3.04 

Total cost per cu. yd. of blocks ready to place 

in work $8.24 



CHAPTER XIX 

CONCRETE-STEEL 

558. The ratio between the compressive and tensile strengths 
of steel is nearly unity. The same thing is approximately true 
of wood and some other materials of construction. In cement 
and concrete, however, the conditions are quite different, the 
strength in compression being from five to ten times the strength 
in tension. Concrete cannot, therefore, be economically used to 
resist tension, and in structures requiring transverse strength 
concrete is at a great disadvantage. 

559. The idea of supplementing the tensile strength of con- 
crete by the use of iron in combination with it, seems to have 
been suggested independently by a number of men. It is 
known that combination beams were tested by Mr. R. G. Hat- 
field as early as 1855. In 1875 Mr. W. E. Ward, 1 M. Am. Soc. 
Mech. Engrs., constructed a dwelling entirely of "beton," the 
floors, roofs, etc., being reinforced with light iron beams and 
rods. These early uses of the combination have some bearing 
upon the ability of patentees to cover in their blanket patents 
more than the peculiar form of the steel member which they 
advocate in their particular system. 

ART. 65. MONIER SYSTEM 

560. A much more picturesque beginning of the concrete- 
steel industry is furnished in the story, quite true, that about 
1876, a French gardener, Jean Monier, used a wire netting as 
the nucleus about which to construct his pots for flowers and 
shrubs, and seeing that the practice might be extended, he 
called to his aid engineers and capitalists who developed the 
Monier system of construction. 

This system consists of imbedding in the concrete two sets 
of parallel rods at right angles to each other, the rods of the 
two sets being tied together with wire at all intersections. 



1 Proc. Am. Soc. Mech. Engrs., Vol. iv, p. 388. , 

395 



396 CEMENT AND CONCRETE 

The larger wires run in the direction of the greater tensile stresses 
and are usually spaced two to four inches apart. The rods at 
right angles to these main tension members are to assist in dis- 
tributing the stress to the main members and may be of smaller 
diameter. 

The iron rods in this system are designed primarily to resist 
the tension only, and the form of the bars is not such as will 
stiffen the structure while the concrete is fresh. In an arch, 
two systems of netting are used, one near the intrados and one 
near the extrados. 

561. The main advantages which this system has over some 
of its competitors are the simple shapes required, that is, round 
rods, which may always be obtained without difficulty, and 
the fact that these may be so readily put together by ordinary 
workmen under supervision. This system is especially adapted 
to vertical walls, whether curved or straight, and found its 
first extensive use in the construction of tanks and reservoirs. 
It has been extended, however, to the construction of sewers, 
floors, roofs, and arch bridges. 

One of the practical disadvantages of the system is that the 
nets are somewhat difficult to handle and keep in position, and 
in thin sections it has not been found practical to imbed the 
nets in concrete containing broken stone of the ordinary size. 
The use of cement mortar, usually one part cement to three 
sand, has been found necessary in order to get a perfect con- 
nection between the wires and the body of the work. This, 
of course, increases the cost. Another objection has been 
urged against it, namely, that the transverse rods do not in 
general have any duty to perform, and are simply a waste of 
material so far as the final strength of the structure is con- 
cerned. While this may be so in certain forms of construction, 
it may be met by the statement that these cross-rods may be 
made as small as desired if they are to act merely as spacers 
for the main rods. In slabs, walls, etc., however, these cross- 
rods have a purpose, and in some other systems members are 
supplied to fulfill this necessary function. 

562. Some very bold arches have been built on the Monier 
system, including three bridges in Switzerland having 128 foot 
span, 11 foot rise, and a thickness of but 6J inches at the crown 
and 10 inches at the abutments. 



PATENTED SYSTEMS , . 397 

A Monier arch of 32.8 foot span, rise one-tenth of span, 
width 13.2 feet, in which the mortar at the crown was six inches 
thick and eight inches at the abutments, was tested in Austria 
in 1890. It held a fifty- three ton locomotive on half the arch, 
and finally failed at the abutments under a load of 1,700 pounds 
per square foot over half the span. 

563. Pipes are now made by this system at yards and trans- 
ported to the place of use. It has also been used as a substi- 
tute for iron in cylinders for bridge piers. A novel use of this 
system consists in making a pipe covering for piles exposed to 
marine borers. The pipe, which is long enough to reach from 
above the water surface to below the bed of the waterway, is 
slipped over the head of the pile and settled a short distance 
into the mud or silt of the bottom with the aid of a water jet. 
A question, however, has been raised as to the action of con- 
crete and iron in combination in sea water on account of the 
possible setting up of galvanic action. 

ART. 66. WUNSCH, MELAN, AND THACHER SYSTEMS 

564. Wiinsch System. This system, which was invented 
in 1884 by Robert Wiinsch of Hungary, consists of two iron 
members of angle irons and plates imbedded in concrete, the 
lower member being arched and conforming to the outline of 
the soffit, while the upper one is horizontal and continuous. 
The two members are riveted together at the crown, and at the 
abutment are rigidly connected by a vertical member. The 
several systems of rib bracing thus constructed are connected 
laterally at the abutment by channel bars running transverse 
to the arch and riveted to the bottom of each vertical in the 
abutment. Assuming that the abutments are stable, it is evi- 
dent that we have here not simply an arch, but also some ele- 
ments of the cantilever. The spandrels being built up solid of 
concrete, there is no definite arch ring, and the quantity of 
material required, especially in long spans, is likely to be much 
greater than in other systems. On the other hand, the great 
depth at the springing permits the use of concrete only moder- 
ately rich in cement. 

565. A bridge of this type, built at Neuhausel, Hungary, 
consists of six spans of about 56 feet each, rise 3.7 feet, thick- 
ness at crown 9.8 inches, and at springing line 54.3 inches. 



398 CEMENT AND CONCRETE 

The total width of the arch was 19.7 feet and contained thir- 
teen systems of arch ribs. Concrete in the abutments below 
water was made mainly of Roman cement. Above water it 
was composed of one part Portland to eight or ten parts sand 
and gravel. Ten to twelve inches of the arch was built of 
strong Portland concrete rammed in layers at right angles to 
radial lines of the arch, special care being taken with that part 
below the bottom arched member. An arch was usually com- 
pleted in one day, and the centers remained in place thirty to 
forty days, the greatest settlement on the removal of centers being 
two-thirds of an inch. This bridge contained 1,346 cubic yards of 
concrete and 88,180 pounds of iron, and cost, complete, $13,700. 

566. Melan System. This system, invented by an Austrian 
engineer, Joseph Melan, consists of arched ribs between abut- 
ments as in bridges, or between beams or girders as in floor 
construction, the space between the ribs being filled with con- 
crete. Steel I-beams curved to the proper form are usually 
employed for the reinforcement, though angle iron flanges with 
lattice connections have been used in some of the large bridges. 
The steel members extend into the piers or abutments and are 
there connected by angles or other shapes, and firmly imbedded 
in the concrete. 

567. This system as adapted to bridge construction has 
probably met with greater favor among American engineers 
than any other form. Perhaps this is because of the stiffness 
of the form of iron beam used, and because by assuming a 
rather high fiber stress for steel the reinforcement may be de- 
signed to withstand the entire bending moment without exces- 
sive dimensions for the steel members. There is thus a feeling 
of security in its use that is not felt in the same degree with 
other systems. The arch dimensions are determined by com- 
puting the forces and required thickness of arch ring after 
assuming certain safe working stresses for the steel and con- 
crete; but if desired, the size of steel members may then be 
increased slightly where necessary to such dimensions that 
with unit stresses of, say, one-half the elastic limit, the entire 
bending moment shall be taken by the steel. Some of the 
largest bridges built after this system in the United States are 
the five-span bridge at Topeka, Kan., and the three-span bridge 
at Paterson, N. J. 



PATENTED SYSTEMS 399 

568. Thacher System. A modification of the Melan system 
Is that invented and patented by Mr. Edwin Thacher. Steel 
bars are used in pairs and imbedded in the concrete near the 
intrados and extrados of the arch and extending well into the 
abutments. The bars of each pair may be connected by bolts 
or stirrups, though Mr. Thacher's original idea seems to have 
been to have no connection between two bars of a pair ex- 
cept through the concrete. The bars are provided with pro- 
jections which may be in the form of rivet heads, lugs, or 
bolts, to increase the resistance of the bars to slipping in the 
concrete. 

569. Mr. Thacher has more recently designed a special form 
of rolled bar having projections that serve the same purpose 
as the rivet heads mentioned above. Several bridges have 
been built on this system, one of the most notable of these being 
the Goat Island bridge at Niagara Falls, one span of which is 
110 feet in length. 

570. In the construction of arch bridges many of the other 
systems are simply modifications of the Melan. The shapes of 
the steel members may have different forms, and the connec- 
tions between the pairs of bars forming the arch ribs may vary 
to suit the idea of the inventors. But though these systems 
lose their identity in long-span arches, their distinctive features 
are more apparent in the construction of floors, roofs, columns, 
etc. 

ART. 67. OTHER SYSTEMS OF CONCRETE-STEEL 

571. The Hennebique System. The rods are here arranged 
in pairs, one above the other, in a vertical plane. In girders, 
the bar in the tension side is straight, while the other one of 
the pair is horizontal for a short distance along the center of the 
span, the ends being inclined upward near the ends of the 
beam. The two bars are connected by bent straps or U-bars 
so that the steel reinforcement may be compared to a queen 
post truss within the concrete. This system has been used in 
the construction of bridges, both arch and girder, floors, roofs, 
stairways, etc., but it is in beams and girders that its distin- 
guishing characteristics are best displayed. 

572. A beautiful arch on this system is the bridge over the 
river Vienne at Chatellerault, France, consisting of three spans, 



400 CEMENT AND CONCRETE 

the central one of which is 164 feet long, with rise of 15 feet, 
8 inches. Four arch ribs 20 inches deep support the roadway ; 
25 feet wide, by posts forming a skeleton spandrel. 

573. Kahn System. In this system, which is somewhat 
similar to the Hennebique, the distinguishing feature is the care 
taken to provide against shear, or against that combination of 
tension and shear which tends to cause failure in a beam by 
cracks that extend diagonally upward toward the center of 
span from near the points of support. The steel plates forming 
the tension members are sheared longitudinally at intervals, 
and short ends are bent up at an angle of forty-five degrees 
with the horizontal. These ends, which may be compared to 
the tension diagonals of a truss, are thus a part of the main 
steel member, and the stress is transferred directly to the 
latter without dependence on the concrete. 

The advantages are the great resistance offered by the bar 
to being pulled out of the concrete and the thorough manner 
in which all tension stresses may be provided against. The 
main disadvantages would seem to be the necessity of detailed 
shop work for each size of girder, the inconvenience of shipping 
the steel in its complete form and the difficulty of thoroughly 
tamping the concrete around the diagonals. 

574. The Ransome System. One of the earliest patents to 
be issued in this country for a method of using concrete and 
iron in combination was that issued to Mr. E. L. Ransome in 
1884. The valuable and distinctive feature of this system is 
the use of a square bar that has been twisted cold. This twist- 
ing not only insures a good bond between the concrete and iron, 
but actually somewhat increases the strength of the bar. 

In building beams with twisted bars as tension members, the 
latter are given a slight inclination from the center upward 
toward the ends. For use in buildings, as in floors and columns, 
and for covers to areaways, and similar uses, this system is 
largely employed. 

575. Roebling System. As its name implies, wire forms 
the main feature of this system, and in a general way it resem- 
bles the Monier. Its application thus far is found principally 
in floor construction, two distinct methods being used. In the 
arched floor a wire netting, stiffened by round steel rods woven 
through it is sprung between the lower flanges of the main 



STRENGTH CONCRETE-STEEL 401 

I-beams of the floor. This netting, further stiffened and held in 
place by iron rods running parallel to the axis of the arch, forms 
a permanent center for the placing of the concrete, which fills 
all of the space to the level of the top of the I-beams. A level 
ceiling below is obtained by a similar netting laid flat against 
the under side of the I-beam and fastened thereto. This acts 
as a wire lath to receive a coat of plaster. If the level ceiling 
is not necessary, the plaster may be applied to the under side 
of the arch netting, in which case the lower flange of the I- 
beam should be encased in concrete to protect it from corrosion 
and fire. 

576. For lighter loads, flat bars are placed at suitable in- 
tervals above and below the I-beams and clamped to the flanges. 
To these bars the wire netting is attached, a thin layer of con- 
crete laid on the upper wire incasing the bars, and plaster ap- 
plied to the lower netting forming the ceiling. Cinder concrete 
is usually employed with this system. 

577. Expanded Metal. The use of whac is commonly known 
as expanded metal lath has been extended to concrete-steel 
construction. As in the Monier and Roebling systems, the 
strength and stiffness of the structure are increased by the use 
of steel rods in connection with the expanded metal, the chief 
duty of the latter, where great strength is required, being that 
of a distributing member. Expanded metal is made from 
sheet steel by shearing short slits parallel to the grain, and 
extending the sheet at right angles to the slits, resulting in a 
network of diamond shaped openings. The metal used is of all 
weights up to one-quarter inch thick with meshes six inches long. 

578. The steel bars used in connection with expanded metal 
by the St. Louis Expanded Metal Fireproofing Co. are square, 
with frequent corrugations surrounding the bar. These corru- 
gations serve only to prevent the slipping of the bars in the 
concrete without adding to the strength. 

The applications of this system include conduits, sewers, 
and walls of buildings, as well as floors and roofs. 

ART. 68. THE STRENGTH OF COMBINATIONS OF CONCRETE AND 

STEEL 

579. While we have in this country been somewhat slow in 
acknowledging the worth of concrete-steel construction, there 



402 CEMENT AND CONCRETE 

is now a strong interest displayed in the subject; many experi- 
ments are being made in our educational and commercial labo- 
ratories and the theory of the action of concrete and steel in 
combination is being rapidly developed. It is natural that in 
the investigation of a form of construction permitting so many 
variations in methods of preparation, that the opinions now 
advanced, based on insufficient data, should be more or less 
conflicting. 

580. Experiments. The experiments of M. A. Considere, 
made in France between 1898 and 1901, which have been made 
more available to us through the translation and collection of 
his articles on the subject by Mr. Moisseiff, 1 are exceedingly 
valuable. The effect of the quality of the steel and the con- 
crete, of repeated loads, of changes in volume in hardening, 
and many other points are carefully analyzed by experiment 
and theory. 

One of the most important deductions drawn by M. Con- 
sidere is that fibers of concrete within what may be called the 
sphere of influence of a reinforcing rod of iron or steel, is capable 
of enduring very much greater elongations without visible frac- 
ture than similar concrete without reinforcement. The expla- 
nation advanced for this is that the steel so distributes the 
stress throughout the length of the concrete in tension that 
the development of insipient fractures or excessive elongations 
at the weaker sections of the concrete is prevented until each 
section has taken its maximum load. The conclusion to which 
this theory leads is that the resistance of the concrete through- 
out the area of influence of the steel reinforcement, is main- 
tained far beyond that degree of deformation which, in concrete 
not reinforced, would cause its rupture. 

581. Neglect of Tensile Strength. Notwithstanding these 
conclusions, it is believed that it is sufficient in most cases of 
design to neglect the tensile strength of the concrete in concrete- 
steel combinations. This course may be defended by the fol- 
lowing considerations. The tensile strength of concrete is, at 
best, not usually above two hundred to four hundred pounds 
per square inch. If the stress on the extreme fibers of a beam 



1 " Reinforced Concrete," by Armand Considere, McGraw Publishing Co., 
New York. 



STRENGTH CONCRETE-STEEL 403 

is three hundred pounds, and we consider that this stress de- 
creases uniformly toward the neutral axis, the mean stress is 
but one hundred fifty pounds per square inch. Again, if we 
disregard M. Considered conclusions, we find that since the 
modulus of elasticity of steel is, say, fifteen times that of con- 
crete, the former is only stressed to forty-five hundred pounds 
per square inch when the imbedding concrete has reached its 
ultimate strength. 

582. The resistance of concrete to tension may easily be 
destroyed or impaired by accident, especially when fresh. The 
properties of concrete vary so much with the materials, the 
proportions, and the manipulation, and the investigation of 
the behavior of concrete and steel under stress is as yet so in- 
complete, as to make refinements in theoretical treatment not 
only unwarranted but really undesirable for practical purposes, 
since they lead to the appearance of greater accuracy than is 
in reality attainable. 

It is true that by the judicious selection of values for the 
constant appearing in formulas for the strength of concrete- 
steel beams, the results of such formulas sometimes show a re- 
markable agreement with the results of that series of actual 
tests for which the constants have been selected; but one has 
only to recall his experience in other lines, hydraulics for in- 
stance, to realize the importance of the almighty constant. 
The opinion sometimes advanced, that the strength of a given 
concrete-steel beam may be as accurately derived by formula 
as can the strength of a steel beam, the writer does not believe 
to be tenable, at least in the present state of our knowledge 
concerning the behavior of concrete. 

583. To neglect the tensile strength ,of the concrete will 
result in a slight increase in the required area of steel reinforce- 
ment, and, in so far as the tensile strength of the concrete 
may be developed, will tend to make the compression side of 
the beam weaker than the tension side. The only objection 
to this is that the failure of the beam, though at a higher 
load, may be more sudden. This possibility, however, seems 
less serious than the error of depending on the tensile strength 
of the concrete only to find it lacking at the critical mo- 
ment. 

Since the aim here is to develop a formula that may be used 



404 CEMENT AND CONCRETE 

with safety in the design of structures, and since to neglect 
the tensile strength of the concrete is to add an unknown, 
though probably small, factor of safety, the tensile strength 
will not be considered in the following analysis. 

ART. 69. CONCRETE-STEEL BEAMS WITH SINGLE 
REINFORCEMENT 

584. Definitions. In this discussion the word strain has 
its technical meaning, the relative change in length of a piece 
under stress. It is usually expressed as the ratio of the elonga- 
tion (or shortening if in compression) to the original length of 
the piece. But for our purpose it is the ratio of the increment 
of change in length, occasioned by a given increment of stress, 
to the length of the piece before the increment of stress was 
applied. These two expressions for strain are usually consid- 
ered equivalent, since, according to Hooke's law, the ratio be- 
tween stresses and corresponding strains, for a given material, 
is constant within the elastic limit. But in dealing with con- 
crete it is found that, even before the stresses become excessive, 
Hooke's law does not hold true. Bearing in mind, then, the 
meaning of the word strain, we represent as usual the ratio of 
stress to strain by E, the modulus of elasticity, or 

,., _ stress 
strain 

Let E, = modulus of elasticity of steel. 

E e = modulus of elasticity of concrete in compression. 
/ = tension in steel, Ibs. per sq. in. 
/ = compression in concrete, Ibs. per sq. in. 
a = thickness of steel considered as a flat plate, or the area of imbedded 

steel bars per inch of width of beam z. 
yi = distance the extreme fiber of concrete in compression is from the 

neutral axis. 
7/2 = distance the center of the steel reinforcement in the tension side 

of the beam is from the neutral axis. 
i = depth of concrete below reinforcement. 
d = yi + 2/2 and h = d + i. 

\i = unit compression of extreme fibers of concrete in compression. 
X 2 = unit elongation of steel in tension. 

E f 

Represent -^ by R, and j- by r. 

&c Ic 



SINGLE REINFORCEMENT 



405 



585 . Formulas for Constant Modulus of Elasticity. The 

cross-section of the beam, the graphical representations of the 
Strains and of the stresses are shown in the following diagrams: 




FIG. 10. 
CROSS-SECTION. 



STRAINS 



FIG. 13. 
STRESSES. 



Figure 12 shows the conditions when the stresses are so small 
that the modulus of elasticity of the concrete may be considered 
constant, and this case will be first considered. 

In the strain diagram, A. t represents the deformation of the 
extreme fiber of concrete in the compression side of the beam, 
and \2 the deformation of the steel. Since a section plane be- 
fore bending is considered to be plane after bending, the steel 
is considered not to slip in the concrete, and NN is the neutral 
axis, 



= ; but 
' 



, and E c = 



or 



and 



= f s and 



/ E, 

i, K 



\8 2/2 






fs EC r 

y-2 = T ^2/i=p2/ 
fc Hi s n, 



(Eq. 1.) 



In the stress diagram the triangle NAB represents the 
total compressive stress on the concrete for unit width of beam, 

and is equivalent to a single force ' c ^ 1 applied at the center of 

2t 

gravity of the triangle. 

The total compressive stress for section of width z is 



The total tension in the steel is T = zaf s . 



406 CEMENT AND CONCRETE 

As we disregard the tensile strength of the concrete, and as the 
total normal compression and total normal tension on a section 
must be equal, as they are the two forces of a couple, we have 

P = T, or zf c =zaf,, 



whence a = = (Eq. 2.) 

h *r 

2 
The point of application of the force P is = y l above the neu- 

f O 

tral axis, while the point of application of T is y 2 below the 
neutral axis; the arm of the couple is therefore f ^ y l + y z \ 

and the moment of resistance is equal to either force into this 
arm, 

Jlf. = */. 
substituting the value of y 2 given in (1) and reducing, 



586. Formulas for Varying Modulus of Elasticity. The fore- 
going formulas are based on the supposition that the compres- 
sive stress in the extreme fiber of the concrete has not passed 
the point beyond which equal increments of stress no longer 
produce equal increments of strain or deformation. They are 
based, in other words, on the common theory of flexure, except 
so far as we have departed from the application of this theory 
in neglecting the tensile strength of the concrete. It is well 
known that even for steel and wooden beams this common 
theory does not, and is not meant to, apply outside the elastic 
limit. In the case of concrete, however, it has been found that, 
even for quite moderate stresses, the modulus of elasticity is 
not constant (Art. 56), but that after a certain stress is reached 
the modulus decreases with increasing stress. The effect of 
this upon the internal forces may be illustrated by the curve 
N B in Fig. 13. The extreme fiber is supposed to be subjected 
to the stress / c ; the fibers nearer the neutral axis have a smaller 
stress per square inch, and the modulus of elasticity for this 
smaller stress is greater; but in order that a section that is 



SINGLE REINFORCEMENT 407 

plane before flexure shall be plane after flexure, the strain must 
be proportional to the distance from the neutral axis. It fol- 
lows, then, that the stresses in the inner fibers do not decrease 
according to the ordinates of the triangle, but are greater than 
indicated by such ordinates. The exact form of the curve 
B N is not known, but the examination of a number of de- 
formation curves has indicated that it is parabolic, and for the 
purpose of this discussion it may be considered a parabola 
with axis A B without serious error, although it is known the 
axis does not coincide with A B for stresses below the elastic 
limit of the concrete. 

587. While the formulas derived in 585 may represent, then, 
the conditions existing in a beam subjected to very moderate 
stresses, it appears that beyond the limit of stress at which 
the modulus of elasticity of concrete becomes variable, they 
should be so modified as to take into account this variable 
modulus. 

Then if A B in Fig. 13 now represents f c and M S = } 8 , we have 
as before, 

-**-* (Eq - 4 -> 

The total stress on the concrete above the neutral axis is 

2 

now represented by the area within the parabola, or - f c y^ , and 

o 

the total compression on section of width z is 



and the total tension 

T' = zaf.. 
As these are the two forces of a couple 

-zyj c = zaf s ; 
whence a = </i = - (Eq. 5.) 



The point of application of P' is on a line through the center 
of gravity of the parabola, or - y l from the neutral axis, while 

o 

the point of application of T' is at distance y 2 below the neutral 



408 CEMENT AND CONCRETE 

axis: the arm of the couple is, therefore, 5- y^ -h y z , and the 

o 

moment of resistance 

9 



= ^y* + i z fyM- (Eq.Ga.) 

Substitute value of y 2 given in (4) 



(Eo, 6.) 

In applying these formulas, it must be remembered that 
(1), (2), and (3) are applicable where the stresses are below 
the point at which the modulus of elasticity of the concrete 
begins to diminish, while (4), (5), and (6) illustrate the con- 
ditions for stresses above that limit. 

588. Example. Design a beam of 10 foot span to carry a 
load due to 20 feet head of water. 

Load per square foot = 20 X 62.5# = 1250#. 
Total load per foot width of beam = 12,500 \bs. = W v 
First, using Eqs. 1, 2, and 3. 

M = -5- = 187,500 inch-lbs. on beam 1 ft. wide, (z = 12). 

8 

Assume 

/. = 12,500, f c = 500, r= f s = 25; 

Jc 

E, = 28,000,000, E c = 2,000,000, R = -^ - = 14. 
From (3) 

M = 187,500 = 12 X 500 Q + 

y? = 25.5, y i = 5.05 inches. 

From (2) 

v< 5.05 



az 



= .101 X 12 = 1.21 sq. in. of steel for beam 12 in. wide. 



SINGLE REINFORCEMENT 409 

From (1) 

T 25 

y 2 = y l= x 5.05 = 9.02 inches. 

If i = thickness of concrete below center of steel bars = 2 inches, 
h = total depth beam = 5.05 + 9.02 4- 2.00 = 16.07 inches. 

Second, using Eqs. 4, 5, and 6. 
Assume 

/. = 50,000, / e = 2,000, r = = 25; 

/c 

E. = 28,000,000, E c = 1,400,000, R = (' = 20. 

1C 

As the stresses per square inch given above are approxi- 
mately the breaking strengths of the materials, we must supply 
a factor of safety, say 4; i.e., design the beam to withstand four 
times the required bending moment before the stresses assumed 
above are attained. 1 
From (Eq. 6) 

M = 4M = 4 x 187,500 = 12 x 2000 (jg + 1 X |jW; 

187,500 X 4 x 12 

whence M* = - - - = 25 

12 x 2000 X 15 



or, yi =5. 

From (Eq. 5) 



2 5 

3 25 



and az = 1.6 square inches of steel for 12-inch width of beam. 

1 The method of using the breaking strengths of the materials, and com- 
puting the ultimate resistance equal to a certain number of times the desired 
strength, is considered inferior to that of assuming safe working stresses and 
computing directly the safe load. These safe working stresses should be 
fixed with reference to the elastic limit of the materials, rather than with 
reference to ultimate strength. The use here of the term factor of safety is 
for the momentary purpose of emphasizing the fact that the conditions 
assumed in deriving equation (6) are such as are supposed to exist under 
comparatively high stresses; but the formulas may evidently be applied to 
the safe working stresses the same as equations (1), (2) and (3), and in the 
present example the same size beam will result by eliminating "factor of 
safety" and using working stresses equal to one-fourth the values of the 
stresses assumed. 



410 CEMENT AND CONCRETE 

From (Eq. 4) 

2/2 = T^ 1 = l^ 1 = L25 x 5 " = 6 ' 25 inches> 

If i = 2 inches as before, 

/i = total depth beam = 5.00 + 6.25 + 2.00 = 13.25 inches. 

It is seen that equations 4, 5 and 6 give, for the assumption 
made, a lesser depth of beam with more reinforcement than 
is given by equations 1, 2 and 3 with the corresponding as- 
sumptions as to stresses and moduli. 

589. An inspection of the equations shows that to increase 
the amount of steel reinforcement in the tension side of the 
beam tends to move the neutral axis nearer to the tension 
side, and bring a greater area of cross-section of concrete into 
compression. If we arbitrarily decrease the depth of the beam 
which must withstand the same bending moment, it will in- 
crease the required area of reinforcement, and if carried too 
far will eventually raise f c beyond a safe value. On the other 
hand, if we take the beam as designed in accordance with equa- 
tions 1, 2 and 3 and subject it to a greater bending moment 
than that for which it is designed, then so long as R remains 
constant, r also remains constant, that is, the steel and con- 
crete are equally overstressed; but since R increases with the 
load, r will also increase, that is, the increment of stress in 
steel will be relatively greater than that in concrete. 

590. Excessive Reinforcement. In the solution of the above 
example if we introduce the requirement that the total depth 
of the beam shall be but 12 inches, while the quality of the con- 
crete is not improved, we may assume, as before, E s = 
28,000,000 and E c = 1,400,000. Let us introduce the same 
factor of safety, 4, by using f c = i&gJL = 500 pounds instead 
of designing the beam for four times the required bending 
moment; as we have seen, this does not affect the result. 
Since the depth of the beam is fixed, /, and r cannot be as- 
sumed, but must be found, together with a. 

We have 

d = 2/ A -f !/ 2 = 12 2 = 10 inches, and y 2 = 10 y v 
From (6 a) 
Mo = A X 12 x 500 y* + f X 12 X 500 y, (10 - y,) = 187,500, 



EXCESSIVE REINFORCEMENT 411 

Solving, we have y t = 6 inches nearly, 
and 2/ 2 = 10 6 = 4 inches. 

From (4) ,7 = ff ' 

2/1 /c^a 

Substituting values of y 2 , y^ f e , E c and E s , we have 

/, = 6,667 Ibs. per sq. in. 
Fro m (5) = v,- 



and az = 3.6 sq. in. of metal to each foot width of beam. This 
is more than double the amount of reinforcement required for 
a 13.25 inch beam, while the steel is stressed only 6,667 Ibs. 
per square inch. 

It may be asked why not use a smaller area of metal, say 
2 sq. in., stressed to 12,000 Ibs. per square inch, giving the same 
total tension; but a moment's consideration shows that in order 
that the metal should assume this higher stress, its elongation 
must increase proportionally, involving a corresponding in- 
crease of strain in the concrete in compression with an accom- 
panying increase in stress beyond the assumed safe limit of 
500 Ibs. per sq. in. 

591. To pursue this subject of excessive reinforcement a 
little further, let us examine some tests of concrete-steel beams 
made by Prof. Gaetano Lanza and reported in Trans. Am. 
Soc. C. E. for June, 1903. 

In these beams the width z = 8 inches, h = 12 and d = 10 
inches nearly. The span was 11 feet. Proportions in concrete 
by volume 1 part Portland cement, 3 parts sand, 4 parts broken 
trap that would pass 1 inch ring, and 2 parts of the same rock 
that would pass J inch ring. Both plain and twisted square 
steel bars were used as reinforcement, the plain bars having a 
tensile strength of about sixty thousand pounds per square 
inch and the twisted steel about eighty thousand pounds per 
square inch. 

If we assume the ultimate strength of the concrete to be 
2,000 pounds per square inch, the modulus of elasticity at this 
high stress to be 1,400,000 and the modulus of the steel to be 
28,000,000, we have, 

28,000,000 _ 
R ~ 1,400,000 = " 2 ' 



412 CEMENT AND CONCRETE 

and for twisted bars, 

80,000 
r= 2^ = 4 ' 

From Eq. (4) y 2 = ^y, = 2 y lt 

.'. 3y 1 = 10 inches, y 1 -Cinches. 

o 



From Eq. (5) a= = - X ^- X = = .055, and az = .444. 

That is, .444 sq. in. of twisted steel reinforcement is required 
in the beam 8 inches wide in order that the stresses in concrete 
and steel shall simultaneously reach the values of 2,000 and 
80,000 Ibs. per square inch, respectively. 



Fro m( 6) M = 8 

= 311,100 inch-pounds. 

One beam having .56 sq. in. reinforcement, or an area very 
close to the theoretical amount called for above, broke under a 
bending moment of 470,000 inch-lbs. Eight other beams hav- 
ing a greater area of reinforcement gave moments of 355,000 
to 443,000 inch-lbs., and the average of the nine bars was 403,000, 
or 30 per cent, greater than the value derived by formula. 

592. Included in the series of tests were three beams, in 
each of which were placed two 1J inch twisted rods. As we 
have seen, the correct amount of steel to develop the full strength 
of both steel and concrete is about .444 sq. in.; the three bars 
mentioned had 3.12 sq. inches of steel, or a large excess of 
reinforcement. To determine the theoretical moment of re- 
sistance of these beams, assume as before: 

E 8 = 28,000,000, 
E c = 1,400,000, 
/ e = 2,000. 

From (4) y, = y, = *, ' () 



STRENGTH OF BEAMS 413 

2 2 000 

From (5) a = .39 = - -^--y l9 (b) 

" / 

2/ 2 =10-2/i. (c) 

Solving (a), (b) and (c), we obtain 
/, = 22,000, 2/1 = 6.45 inches, and y 2 = 3.55 inches; 

whence r =-/= 11, 

/ c 
and from (6), 

M = 8 x 2000 (TO + o x oTj) ( 6 - 45 ) 2 = 522,000 inch-pounds. 

These three beams developed the following moments of re- 
sistance: 553,550, 663,700 and 783,500, mean 667,000 inch-lbs., 
or 28 per cent, greater than that derived by formula. None of 
them failed, however, by crushing of the concrete at the top 
of the beam, but by longitudinal shearing "at or a little above" 
the reinforcing rods. 

593. It appears, then, that by increasing the area of steel 
reinforcement over 600 per cent., or from .44 sq. in. to 3.12 
sq. ins., the strength of the beams was increased about 68 per 
cent, by theory, or 66 per cent, according to the few tests cited. 
The cost of the beam, however, was increased about one hun- 
dred per cent. 

This method of increasing the moment of resistance of a 
beam is not economical; it is better to improve the quality of 
the concrete. It may, however, be necessary at times to use 
excessive reinforcement on account of restrictions on the size 
of beam, but one may easily carry this so far that he passes 
outside the true theory of concrete-steel construction, and it 
becomes a question of the steel being sufficient to carry the 
entire load. In such cases double reinforcement may be adopted. 

594. Tables of Strength. In Table 160, equations (5) 
and (6) have been reduced to simpler forms by the introduc- 
tion of values of E s and /,. Selecting in the table the division 
corresponding to the modulus of elasticity of the concrete 
which is to be used, and the line opposite the assumed stress in 
the concrete, M = quantity in column a times the square of the 
depth of beam, d; and the area of steel in a beam of 12 inch 
width, i.e. 12 a, equals quantity in column 6 times the depth 



414 



CEMENT AND CONCRETE 



a 

(U 

P 



a 



a 



02 flj - 3 

Sl^l o 

tt S "S 

1 Id ^ 

d g -S cu a; 



^<% -54 

lftl 

1^3 

O..S- c -3, 
3 lif 

d a & ^ 
| S ^ S 

C 5 O ., 

^ e .S " 








lg 
U 3l 




O 00 t^ (M rH Tj< 

O rH CO CD OS <N 






'3 


^8 

003 




rH 




. 


i 


a 








c? 




^ ^ *o3 ei 




rH T^ -!fl OS OS 




i 




r*S 


* 


(7<l Tjt t- O Tj< 

_ O O O rH rH 






/ 

1 


8 


^ ^ ^ ^ ^ 


*3 .S 
^ cr 

5 02 


^ CO ^ CO rH OS 
rH CO ^O QO O 
rH 






(3 S d 






H OJ 






S o 

^^'-i 





CO O t^ rH O 
_ <M_ <*<_ r- r-i O 


a 




1 


^o^ 






ff:9 


. 


BQ 


a ao * 






8 


?5 




g o3 r2 2 




CO c ^"" ) 




g 




^5Jc 





SCO 00 
rH rH 


'S^ 


eo" 




w w - 






03 ^ 




. 






iO " 




||| 


8 


OS OS C<| rH rH O 

o o <M' oo' t-' os 

<M T}H CC 01 <M 


S o" 




^ 






| * 






d S s 












111 


o 


I-H 00 ^O 00 t*~ OO 
i 1 CO !> rH O rH 


o" 




1 


O>*M a> 
^OCO 




i 1 rH <2<l 


O> || 


. 


3 

cc 


ai on 






o ^ 


O 




^ ^ 'cS c 




CO O O <M O OJ 


<*H 







^I^H 


iC 


O O O rH Ol C<1 


1 


<N~ 




CO W^ 






,> ^ 










o o 




a 

III 

o o-H 


8 


(N (N t- (?q O O 

oo 06 -^ cd o d 

(> iO 00 CM O 


T3 - 

<u a 
3 S 




S Wi ? 




i-H rH 








cga 

S 

^ 


o 


i i to OS TJI O iO 


W) 

- . c 






<CH 






^ 




i 


fU 5C 






1 





02 


<n an g 






" 2 


8 




G^'Ss 




00 O Tf< -^ O OS 


H 


| 








O O rH rH OJ CO 


^ 




||8 




rH rH rH O O O 






gw 


8 


i-H Tj< S rH Oi 
rH rH i-H 




V W 

5S 

3S 


Ifo'lll 


Is 


! 




As 


^oSS" 


3* 







STRENGTH OF BEAMS 415 

of beam, d. Column c gives the area of cross-section of steel 
expressed as the per cent, of the area of section above the center 
of steel reinforcement. 

595. For example, suppose we wish to know the strength of 
a beam ten inches deep (d = h i = 10 in.) and the amount 
of steel required to develop a stress in the concrete of 400 Ibs. per 
square inch when the stress in steel is 10,000 Ibs. per sq. in., 
and the modulus of elasticity of the concrete is assumed at 
3,000,000. In column a under 3,000,000 modulus, and opposite 
400 Ibs. stress, we find 68.1; then the moment of resistance of a 
beam one inch wide is 68.1 inch-lbs. X 10 X 10 = 6,810 inch- 
Ibs., and the resistance of a beam 12 inches wide is 6,810 foot- 
Ibs. The area of steel required in 12 inches width of beam is 
.092 d or 0.92 sq. in. This beam is reinforced with .77 of one 
per cent, steel. Similar tables may be prepared for other values 
of E, and /, if desired. 

596. In Table 161 the equations have been completely 
solved for certain typical values of E c and /., assuming the 
values for E s and /, of thirty million and ten thousand respec- 
tively, as in Table 160. Having computed the bending mo- 
ment, and fixed upon the probable safe working stress and 
modulus of elasticity of the concrete which it is proposed to 
use, it is only necessary to take from the table the required 
depth of beam and the amount of steel reinforcement required. 

For example, a girder 10 feet long supported at the ends 
carries two loads of 5,000 pounds, each load being 2.5 feet from 
a support. 

If the width of girder is 15 inches, working stress of concrete 
300 Ibs. per sq. in. and modulus of elasticity of concrete 1,500,000, 
what is the required depth of girder and area of steel in tension 
side? 

The maximum bending moment (neglecting weight of beam) 
is 12,500 ft.-lbs. throughout the central five feet. The required 

12 

moment of resistance for twelve inches in width is^of 12,500 

10 

= 10,000 ft.-lbs. Looking in the table for this bending moment 
under 300 Ibs. stress and 1,500,000 modulus, we find it is be- 
tween d = 12 and d = 14, or at about d = 13 inches. If we 
.allow 2 inches below center of steel reinforcement, we have 
total depth of beam, /&=13 + 2 = 15 inches. In the same 



416 



CEMENT AND CONCRETE 






"O O O O O 
l OOOO 
t^Tt<OCCOi 






<M<N<MCOCO^Tt<iO 



O O O O O O O O 
>O-*ac:o'Mr-:^O 



(N (N CO CO CO "* * 'O O t l~~ Ci. O r 7<j CO O 
' i-H r-i i-H T-( r-i (^' CN' 5<l' CO CO CO* 



t^OQOi tCl 






OrHrH(NCOCOTtllOOCOI>- g OiOrHOOOl^6ii-lCO 
' T-H T-i ^H T-i ^H r-l cn' (N 



00 T-t 00 53 t b-*CO, ^J O O O O O O Q O Q O Q 

Tt*OiC<ICOOii iCOOCCt t^-Or- <O-^COI>-CO^S<I 



ft. W 
. O 

S z 

- 
J 



S 

|3 

&z 

H 
11 



Oi'*oD 
OT-(r-l 



T-H T-H i-H (7^ CO CO "^ 



STRENGTH OF BEAMS 417 

lines we find area of steel for 12 inch width between 1.08 and 
1.26, or, say, 1.17; then for 15 in. width the required area is-^ 
X 1.17 = 1.46 sq. in. The bars should not be more than 3 to 
6 inches apart. We may use, then, 5 bars inch square or 

- inch diameter, spaced three inches apart. In large beams it is 

o 

necessary to consider the bending moment occasioned by the 
weight of the beam after making a first approximation to the 
size required. 

597. The above tables are prepared on the assumption that 
the stress in concrete shall be equal to the value selected when 
the stress in the steel reinforcement reaches 10,000 Ibs. per sq. 
in. From the equations, other tables may be prepared if 
desired, in which the working stress in steel shall be 12,500, 
16,000 or any other assumed value. The tables are not suited 
to the computation of beams in which excessive reinforcement 
is used. 

As to actual tests of the performance of concrete and steel in 
combination, the possible variations in material are so diverse 
and the cost of experiments so great that the results thus far 
obtained appear somewhat fragmentary, but each investigator 
has selected a small branch of the subject for experiment. 
Among the more valuable tests in this line may be mentioned 
the following : 

Tests at Massachusetts Institute Technology, Prof. Gaetano 

Lanza, Trans. Amer. Soc. C. E., vol. 50, p. 486. 
Tests at Purdue University, Prof. W. K. Hatt, Jour. Western 

Soc. Engrs., June, 1904. 
Tests at Rose Polytechnic Institute, Prof. Malvard A. Howe, 

Jour. Western Soc. Engrs., June, 1904. 
Tests at University of Illinois, Prof. A. N. Talbot, Proc. Amer. 

Soc. for Testing Materials, 1904. 

Tests at University of Wisconsin, Prof. F. E. Turneaure, Proc, 
Amer. Soc. for Testing Materials, 1904. 

ART. 70. CONCRETE-STEEL BEAMS WITH DOUBLE REINFORCE- 
MENT 

598. We have seen that when the depth of a beam is limited 
by structural considerations we may increase the normal load 



418 



CEMENT AND CONCRETE 



by excessive reinforcement, but that this method results in low 
stresses in the steel and is not usually economical. We may 
now consider the effect of placing reinforcing rods in the com- 
pression side of the beam as well as in the tension side. 




the position of the 



FIG. 14. FIG. 15. FIG. 16. 

CROSS-SECTION CROSS-SECTION STRAIN DIAGRAM. 

(Single Reinforcement.) (Double Reinforcement.) 

Let Fig. 14 represent the cross-section of a beam reinforced 
on the tension side with sufficient steel, area a, to develop the 
proper working stresses in the materials, and let the position 
of the neutral axis be N N. If at distance x from the neutral 
axis we add an area of steel A' in the compression side, the 
position of the neutral axis would be changed for similar load- 
ing; but if at the same time we place in the tension side an ad- 

A x 
ditional area of steel A such that -p = > 

neutral axis will be unchanged. Let // = stress in steel in 
compression; then since the steel must suffer the same deforma- 

tion as the surrounding concrete 77 = Multiplying the last 

fs x 

two equations, we have, f s A = f, A', that is, we have added 
equal forces to the two sides of the beam, and have increased 
the moment of resistance by } s A (x 4- i/ 2 ) inch-pounds. 

599. To illustrate the application of this principle we may 
take the beam considered in 591, in which z ='8, R = 20, 

444 20 

r = 40, az = .444, a = - - = .055, /. = 80,000, y t = - - inches, 

o o 

T/! = in., and M = 311,100 inch-pounds. 

o 

When the area of reinforcement in the tension side of this 
beam was increased to az = 3.12 sq. in. or a = .39, the theo- 
retical bending moment was increased to 522,000 inch-pounds 
( 592). What will be the result of a similar increase in steel 
distributed between the two sides of the beam? 



DOUBLE REINFORCEMENT 419 

Let k distance from top of beam to center of reinforce- 
ment on compression side = 2 inches, 



" 



then x = yi 2" = -^ 2 = - inches, 

o o 

+*- *& A -L \) f-\ A *-\ A A f 

-77 = - = o -*- TT = 0.4 or A = 0.4 A . 
A y 2 o o 

A+ A' = .39 - .055 = .335 

1.4 A' = .335 A' = .24 A' = 1.92 

vl = .095 Az = .76 

whence a .055 as = .44 

A + a = .150 Total steel, 3.12 sq. inches. 

//.* .4 /. = 32,000. 

i/2 

Added moment of resistance equals 

24 

Az (x + ?/ 2 ) /. = .095 X 8 X x 80,000 = 486,400 in.-lbs. 

o 

And total moment of resistance equals 

311,100 + 486,400 = 797,500 inch-pounds. 

None of the bars in the series mentioned in 591 had as large 
an area of reinforcement as 1.92 sq. in. on the compression side. 

It is noticed, first, that the double reinforcement gives bet- 
ter results than such excessive reinforcement on the tension 
side; second, that the stress in steel on the compression side is 
less per square inch than that in tension; and third, that in 
case a large addition of steel is made, this results in a greater 
area of steel in compression than the total area of steel in ten- 
sion. In practice the area of steel in compression is usually 
made equal to, or less than, the area in tension, but beams with 
double reinforcement are seldom accurately designed. 

ART. 71. SHEAR IN CONCRETE-STEEL BEAMS 

600. There are several methods of failure of concrete-steel 
beams other than those considered above, direct tension in the 
steel or direct compression in the concrete due to the bending mo- 
ment. These other methods of failure are popularly called failures 
in shear, although some of them cannot properly be so classed. 

601. We have seen that the shearing stress of concrete is 
usually considered to be somewhat in excess of the tensile 



420 CEMENT AND CONCRETE 

strength (458) and that the latter is one-fifth to one-tenth the 
compressive strength. With a beam having only a normal 
amount of reinforcement, then, there is little danger to be 
feared from simple vertical shear, and as a matter of fact, tests 
have not developed instances of such weakness. In compara- 
tively short spans, however, failures have occurred near the 
quarter points, in cracks starting at the under side of the beam 
and extending upward in a direction inclined toward the center. 
This method of failure has the appearance of being due to a 
combination of shear and tension in the lower section of the 
beam, since the cracks are approximately at right angles to the 
theoretical " lines of direct tension." Such failures, however, 
are almost always accompanied by a slipping of the steel bar in 
the concrete, and may frequently be prevented by taking 
proper precautions against such slipping. 

602. A more frequent cause of failure is a longitudinal shear 
in the plane near the steel reinforcement and on that side of it 
lying nearer the concave side of the beam. 

It is evident that a failure caused by slipping of the bar in 
the beam, although caused primarily by shearing forces, is 
really a failure in adhesion, yet the two forms of weakness are 
so closely connected that it is simpler to consider them together. 

603. Comparison with Plate Girder. In a steel plate girder 
the lower flange is considered to carry the tension, the upper 
flange the compression; the web connects the two flanges, caus- 
ing them to act together as one beam, and we may think of 
the web as preventing the ends of the compression flange sliding 
beyond the ends of the tension flange. When the web is not 
able to accomplish this without buckling, it is stiffened by 
vertical angles. 

In a concrete steel beam we have considered the entire 
tension to be carried by the steel reinforcement, and the entire 
compression to be carried by the concrete on the other side of 
the neutral axis. The connecting web is also concrete. This 
web is thick and not liable to buckle, but it may shear in a lon- 
gitudinal plane as a wooden beam may do when short and deep. 
All of the tension in the steel reinforcement must be trans- 
mitted through the surrounding concrete. If there are no pro- 
jections on the steel bar, the adhesion of the concrete to it 
may, under certain circumstances, be not strong enough to 



SHEAR IN BEAMS 421 

safely carry this stress; and if the adhesion is sufficient, then 
the shearing strength of the concrete may be too low to transmit 
the stress to contiguous fibers or layers. 

604. Illustration. Let us consider a concrete-steel beam 
twelve inches wide, twelve inches deep and of ten foot span, 
supported at the ends; reinforcement, one square inch of metal 
properly distributed in a plane two inches above the bottom 
of the beam. Let us suppose this beam carries a uniform load 
of 600 pounds per foot, giving a maximum bending moment of 
90,000 inch-lbs., and a stress in steel of 10,000 pounds at the 
center. The ends of the steel bars are of course without stress. 
Since the bending moment at any section of such a beam is 
proportional to the product of the segments into which the 
section divides the span, the bending moment one foot from 
the ends will be 

X 90,000 = 32,400 inch-pounds. 



o X o 

Let us consider the neutral axis in the same position at the 
end of the beams as near the center. (This is not strictly true, 
because of the lighter stress near the ends of the beam, but 
the error made by such an assumption will be unimportant for 
our present purpose.) Then the tension in the steel will have 
the same proportion, or, tension in steel one foot from the end 
^ X 10,000 = 3,600 pounds. 

The stress in steel, then, which is zero at the end, has in- 
creased to 3,600 Ibs. in one foot of length. To provide against 
poor contact near the end, consider two-thirds of this length, 
or eight inches, to be operative. If the reinforcement consists 
of four one-half-inch square bars, the necessary adhesion per 

Q rt f\f\ 

square inch is ^ - - = 57 Ibs. per sq. in.; but if only one bar 

o X o 

is used one inch square, the required adhesion is 114 Ibs. per 
sq. in. The latter would not be good practice, not only be- 
cause of high adhesion required, but because the steel is not 
properly distributed. 

Where the stress in adhesion is greater than can be safely 
relied upon for plain rods, it is necessary to use some kind of 
deformed bar, or to anchor the bar securely at the end. This 
may be done by passing the end of the tension bar around a 



422 CEMENT AND CONCRETE 

rod transverse to the beam near the end. Care should be taken 
that the safe value of adhesion is not assumed too high. 

605. Value of Shear. The same total stress of 3,600 Ibs. 
must be transferred through the concrete immediately above 
the bar. If the reinforcement is so distributed that the entire 
width of the beam has practically the same stress, and we 
consider, as before, that two-thirds of the length of the end 



Q 

foot is operative, we have mean shear = ~ - = 37.5 Ibs. 

iZ X o 

per sq. in. The value of stress in shear should not exceed one- 
tenth the safe value in compression, and there is a general ten- 
dency to use not more than one-twentieth. 

If the same form of beam had a span of but five feet with 
same bending moment, the value of the shearing stress by this 
method becomes 75 Ibs. per sq. in., and it will be necessary to 
provide against this stress coming upon the concrete. 

Another approximate method is the ordinary one for rect- 
angular beams, viz. to consider the shear in horizontal plane 
just above the steel reinforcement to be f times the total shear 
at any section, divided by the area of vertical section of the 
beam. 

606. Provision is sometimes made for relieving the concrete 
of all shearing stresses. In this case the beam is divided into 
imaginary panels of length equal, say, to the depth of the 
beam, and the diagram of maximum shear is drawn. The 
shear in each imaginary panel is then provided for by a vertical 
or inclined bar of the proper dimensions. Or, what is usually 
better, the shear bars are all of one size and the proper num- 
ber of them are distributed throughout each panel length; the 
spacing of the shear bars thus becomes wider near the center 
of the beam. 

607. Resistance to Shear. When provision against shear 
is made by using small steel rods placed either vertical or in- 
clined downward toward the center of the beam, as mentioned 
above, these rods may well be made in the form of inverted 
U-shaped stirrups, with their ends securely fastened to the 
reinforcing metal in tension. 

In many cases all the provision necessary is given by the use 
of two longitudinal bars, parallel and close together near the cen- 
ter of the span, but one of them leading to a plane near the top of 



SHEAR IN BEAMS 423 

the beam at the supports. This system is very conveniently 
applied in concrete slabs supported by I-beams, one bar of the 
pair being hooked over the upper flanges of the I-beam and 
sagging toward the center. The Hennebique system (571) is 
a combination of the inclined bar and U-shaped stirrups. 

608. A modification of the single inclined bar is the Cum- 
mings system, wherein there are several pairs of bars of vary- 
ing lengths; these are all horizontal and near the bottom along 
the center of th,e beam; a short distance from the center the 
shortest pair turns up at an angle of about forty-five degrees; a 
littVe farther toward the end a second pair of bars is turned up, 
and so on, leaving a single pair to go through straight to the 
support. 

Another, and more radical modification, is the Kahn system 
(573), in which the bar is square with wings of metal on oppo- 
site corners which are sheared and bent up at angles of forty- 
five degrees, so that the outline of the steel work in a beam 
resembles the tension members of a Pratt truss. 



CHAPTER XX 

SPECIAL USES OF CONCRETE: BUILDINGS, WALKS, 
FLOORS AND PAVEMENTS 

ART. 72. BUILDINGS 

609. While the use of concrete and steel for the walls and 
floors of buildings is about fifty years old, yet it is only in com- 
paratively recent years that its value has become generally 
known. It is now applied to all classes of structures, ware- 
houses, factories, residences, station and office buildings, and 
it is anticipated that in the next twenty years concrete-steel 
will be as familiar in architecture as steel skeleton, stone, and 
brick are now. 

610. It happens that at present the concrete-steel building 
industry is largely in the hands of companies who are exploiting 
some particular form of steel rods or bars applied according to 
some one of the many " systems" of reinforcement. This con- 
dition has both good and bad features. A reputable concern of 
this kind will have in their employ engineers who should sat- 
isfy themselves that each design is a safe one, for the failure 
of a building will cast disrepute on their particular system. It 
is this fact that leads the companies to keep the construction 
entirely, and the design largely, in their own hands. Another 
advantage is that these concerns are able to perfect methods of 
construction by experience, and to lessen the expense of one 
structure by making use of the concrete plant and the molds 
that have been used on another. 

611. In making plans for a building, the owner is usually 
represented in the first instance by an architect whose business 
it is to dictate the design. If concrete-steel is considered, the 
architect may call an engineer in consultation and they may 
together harmonize the features of utility and appearance with 
economy and strength, but in letting the contract it is found 
that the competition is limited to one or two companies using 
the particular system which the engineer considers the best 
adapted to the particular conditions in question. 

424. 



BUILDINGS 425 

On the other hand, the architect will hesitate to go to the con- 
struction company for assistance, since he must first select the 
system he shall use, a question upon which his ideas may be neither 
clear nor well grounded, and he is then having the prospective 
contractor assist in the design. Under these circumstances the 
architect will usually consider concrete-steel construction as 
something he wishes to avoid if possible. But this condition 
will correct itself in time, for owners will demand a considera- 
tion of this form of construction, engineers will become fa- 
miliar with its use and will be employed to design the engineer- 
ing features, while reliable contractors in every city will obtain 
permission to build in accordance with any " system" under 
the supervision of a competent engineer. 

612. Roof. While a pitch roof is sometimes built of con- 
crete-steel, this form of construction is particularly adapted to 
so called flat roofs. The roof is constructed much the same as 
a floor slab (Art. 65-67), except that expansion joints are some- 
times provided, and the roof is covered with tar and gravel, 
or some of the patent roofings ordinarily used. While the 
roof loads are usually light, permitting a greater span of slab 
between beams than for floor construction, it will seldom be 
economical to introduce these longer spans because of the 
changes necessary in the molds. In most buildings it is neces- 
sary to provide against condensation, and for this purpose a 
flat ceiling may be suspended at the level of the under side of 
the beams giving an air space. 

613. Floor System. The floors may be constructed in con- 
formity with the principles stated in Chapter XIX. The 
strength of short span arches, such as are used for floors, 
where the haunches are built up level with the top of crown 
of arch, is a matter of experiment and cannot be accurately 
determined theoretically. Empirical formulas may be derived 
for a certain system based on a sufficient number of tests. 
The principles underlying the strength of slabs may be con- 
sidered the same as those applying to beams (Art. 69), although 
if the length of slabs is not much greater than the span, they 
are not strictly applicable, but will err on the safe side. 

614. A decision must first be made as to the size of bays into 
which the floor space is to be divided. This will of course de- 
pend on the use of the building, the engineering features con- 



426 CEMENT AND CONCRETE 

forming to requirements of utility. If the bays are not square, 
the girders should usually take the shorter span between columns. 
This length is then divided into the number of slab spans that 
will give maximum economy. The shorter these spans the less 
the amount of material required in slabs and the greater the 
number and cost of floor beams. Computations should be 
made, therefore, for two or three arrangements to determine 
this point. As this distribution for maximum economy will 
vary with the loads to be provided for, it is well, if the floors 
are not all to carry the same load, to take for this computation a 
load intermediate between the heaviest and lightest, and use if 
possible the same arrangement of spans throughout the building. 
The strength of slabs for given bending moments may be 
taken directly from Table 161, after deciding upon the working 
stress to be allowed in the concrete and the probable modulus 
of elasticity. The beams and girders, if single reinforcement 
is used, are taken from the same table or computed by the 
methods of Art. 69. 

615. In some instances it may be found economical to use 
concrete-steel slabs for floors supported by concrete protected 
steel beams and girders. One advantage of this system is that 
the forms for building the protecting concrete and for the 
floor slabs may be hung from the steel girders and beams. For 
this method of construction the enveloping concrete should 
not be less than one and one-half inches thick over the edges of 
flanges, and wire fabric or metal lath wrapped about lower 
flanges of beams will insure the concrete remaining in place. 
This is not properly concrete-steel construction, but simply 
concrete protected steel, and except in case the concrete ex- 
tends well above the steel, forming an independent compres- 
sion flange, no added strength should be computed for the con- 
crete covering. 

616. Columns. In the foundations of buildings of moder- 
ate height the supporting columns may be built entirely of 
concrete. Since, however, the pressure on the concrete, even 
when it is constructed with the greatest care, should not ex- 
ceed two hundred to three hundred pounds per square inch, 
the required area of cross-section in the lower stories is usually 
so great as to preclude the use of columns built entirely of 
concrete. 



BUILDINGS 427 

617. Concrete Filling and Covering. A steel column of 
any of the ordinary styles, built up of steel shapes may be 
used, and protected from corrosion and fire by filling and cover- 
ing with concrete. This not only serves as a protection against 
rust, but materially increases the stiffness and permits the use 
of a somewhat higher working stress in the steel. The concrete 
filling should be mixed quite wet in order that it shall work 
into all angles. The edges of the metal should not approach 
nearer than one and one-half inches to the exterior of the con- 
crete, and flat surfaces of metal should have a covering of at 
least two and one-half inches. Where it is necessary to cover 
large, flat surfaces, they should be first covered with expanded 
metal or wire fabric, locked on by twisting around the edges 
of the plate or channel. 

618. COLUMNS OF CONCRETE STEEL. Concrete-steel col- 
umns differ from the above in that the main dependence is 
placed on the concrete rather than on the steel. For such 
columns longitudinal reinforcement has generally been employed. 
Steel bars extending from end to end of the column are dis- 
tributed throughout the cross-section, and are tied together at 
intervals of four to twelve inches by smaller bars forming loops 
to hold them in place. The splicing of the bars is effected by 
placing a small tube over the upper end of the lower bar and 
projecting above it, and then setting the lower end of the upper 
bar within the tube resting on the lower bar. Where this is 
done it is essential that the two ends be planed perfectly square, 
and it is much better to avoid splices in a column between 
lateral supports. In a building the reinforcing rods project up 
through the floor above and are spliced into the bars of the 
columns in the next story. 

619. Strength of Columns. When a column reinforced with 
longitudinal bars is subjected to pressure, the concrete and 
steel must shorten together. The relative stresses in the two 
materials will then be proportional to their moduli of elasticity. 
From this follows the formula, 

P = f c (C +RS) 

where P = total pressure on column, 

f c = stress in concrete, 
C and S = areas of concrete and steel respectively, 



428 CEMENT AND CONCRETE 

pi 
and R = =?, or ratio of the modulus of elasticity 

&c 

of the steel to that of the concrete. 

In a series of tests of twenty-one columns made by Prof. 
Gaetano Lanza, 1 but three failed under a lower stress than that 
computed by the above formula. The columns were eight to 
ten inches square, six to seventeen feet long and reinforced 
with either one or four bars, the latter being from } inch to 1J 
inches square. 

The lowest breaking load was fifty tons on an 8 by 8 inch 
column with one bar one inch square, and the strongest column, 
10 by 10 inches, with four j inch longitudinal bars, was not 
crushed with a load of one hundred fifty tons, the limit of the 
testing machine. The lowest result was twenty per cent, less 
than that given by the formula, and the greatest excess strength 
over the theoretical was fifty per cent. 

620. While longitudinal reinforcement undoubtedly strength- 
ens a long column against flexure, as well as adds to the resist- 
ance to crushing, yet the added strength is gained at the ex- 
pense of considerable additional cost. Suppose we have a ten 
inch square column, twelve feet long, made of concrete with a 
breaking load of 1,800 Ibs. per square inch, or 180,000 Ibs. 
total breaking load. Suppose eight } inch square bars to be 
built into this column as longitudinal reinforcement, and that 
the modulus of elasticity of the steel is ten times that of the 
concrete. Then the strength of the reinforced column would 
be, by the formula above, 

P = 1,800 (95.5 + (10 X 4.5)) = 252,900. 

The longitudinal reinforcement has thus resulted in an increase 
of strength of 40 per cent., while by the addition of 180 
pounds of metal, the cost of the column has risen from about 
$3.00 to say $8.50, an increase of about 180 per cent, without 
counting the cost of lateral ties, and the additional trouble in 
building a reinforced column. 

621. Hooped Concrete. In extended experiments on what 
he has called " hooped concrete," M. Considere 2 has shown that 



1 Trans. A. S. C. E., Vol. 1, p. 487. 

2 Comptes Rendus de I' Academic des Sciences, 1898-1902. Translation, 
"Reinforced Concrete," by Armand Considere, translated by Leon S. Mois- 
seiff, McGraw Publishing Co., New York. 



BUILDINGS 429 

reinforcement is much more important and beneficial in a 
transverse or circumferential direction than if longitudinal. 
This may be accounted for by the fact that the natural method 
of failure of concrete prisms, is by splitting along planes parallel 
to the direction of pressure, and the ordinary method of failure 
by shear along inclined surfaces is induced by the friction of 
the plates transmitting the pressure to the prism. It was also 
shown that while concrete reinforced by longitudinal bars with 
the ordinary amount of lateral ties breaks suddenly, hooped 
concrete fails gradually under a much heavier load. 

622. M. Considere concluded from his experiments that the 
circumferential ties should not be farther apart than one- 
seventh to one-tenth the diameter of the column, even when 
longitudinals were used to assist in completing the network, 
and that the results were more successful the nearer together 
the hoops or ties were placed. He found that spirals were 
better than individual single ties and that longitudinals were of 
value chiefly in assisting to confine the concrete, transmitting 
the bursting pressure at a given plane to the contiguous spirals 
above and below. 

623. M. Considere says 1 that the "compressive resistance of 
a hooped member exceeds the sum of the following three ele- 
ments : 

"1. Compressive resistance of the concrete without rein- 
forcing. 

"2. Compressive resistance of the longitudinal rods stressed 
to their elastic limit. 

"3. Compressive resistance which could have been produced 
by imaginary longitudinals at the elastic limit of the hooping 
metal, the volume of the imaginary longitudinals being taken 
as 2.4 times that of the hooping." 

To subject hooped concrete to a practical test, M. Considere 
constructed, in 1903, a truss bridge of sixty-five foot span with 
parabolic top chord of seven and one-half feet rise, 2 the com- 
pression members being of hooped concrete, and the tension 
members of concrete-steel with longitudinal reinforcement, or 
concrete protected steel. A central panel of the truss was con- 



" Reinforced Concrete," p. 159. 
2 Engineering News, May 5, 1904. 



430 CEMENT AND CONCRETE 

structed with a reduced section of top chord about eight inches 
diameter reinforced by eight longitudinal bars .43 inch in 
diameter and a helix 6J inches in diameter of .43 inch metal 
coiled to a pitch of about one inch. This reduced top chord 
section showed signs of failure when the computed stress reached 
about 5,000 pounds per square inch. 

624. FORMS FOR BUILDINGS. One of the most serious prob- 
lems in the construction of concrete-steel buildings is the de- 
signing of the forms. They must be as light as is consistent 
with strength to facilitate handling. They should be of simple 
construction so that they may be set up and removed without 
too much supervision, and they should be so assembled with 
bolts and screws that they may be used repeatedly. In erect- 
ing a large building sufficient forms are usually provided to set 
up one floor complete, including columns, beams, girders and 
floor slabs. After placing the reinforcement, the concrete is 
filled in as rapidly as possible, making the slabs, girders and 
columns practically monolithic. 

The forms for the girders usually rest upon the column 
molds and are supported at intermediate points by posts rest- 
ing on the completed floor below. While column molds are 
sometimes filled from the top, better work is assured by having 
one side of the mold built up as the concrete is filled in from 
the side. 

The mold to receive the concrete forming the floor slab is 
either a part of, or is supported by, the pieces forming the 
sides of the girders and beams. Provision is sometimes made 
for leaving supports at intervals under the completed beams 
and girders after removing the forms from the sides of the beams 
and the bottom of floor slabs. This is done by making the 
bottom piece of the girder mold separate, and attaching the 
side pieces to it by screws which may be removed without dis- 
turbing the bottom. The caps of the supporting posts are then 
made long enough to permit the lower edges of the side pieces 
to rest directly on them. This method was adopted in build- 
ing the Central Felt and Paper Company's factory at Long 
Island City. 1 



1 Wight-Easton-Townsend Company, Contractors, Engineering Record, 
Jan. 16, 1904. 



BUILDINGS 431 

625. In the same building the walls were built with molds 
three feet high and sixteen feet long, placed in pairs on oppo- 
site sides of the wall. When one section was completed, the 
molds were " lifted until the lower edges were two inches below 
the top of the concrete. In the new position they were sup- 
ported by horizontal bolts through their lower edges, across the 
top of the concrete; the upper edges were tied together by 
transverse wooden strips nailed to them about three feet apart, 
and they were braced to the false work supporting the roof 
and column molds." "The bolts passed through sleeves which 
were left permanently embedded in the walls. At first, iron 
pipes were used for this purpose, but afterwards it was dis- 
covered that pasteboard tubes were equally efficient and much 
easier to trim and point after the molds were removed." 

626. An excellent system of molds was used in the con~ 
struction of the Kelley and Jones Company's factory at Greens- 
burg, Pa. 1 The floor molds were especially convenient, being 
made collapsible by a hinge joint at the top along the longi- 
tudinal center line. . These floor molds were in reality cores 
between adjacent floor beams; when in place the top surface 
was horizontal, to form the under side of the floor slab, and 
the vertical side pieces formed the sides of the floor beams. 
When the concrete had set sufficiently, the lower edges of the 
form were made to approach each other, thus coming away 
from the concrete gradually. A special light wooden framework 
or tower, with a working platform six feet below the floor, and 
a rope sling to receive and lower the floor mold, permitted of 
removing the molds rapidly and without injury. A special truck 
was also used for moving the floor molds about the building. 

627. A convenient adjunct for the construction of concrete 
wall forms consists of a short section of - I-beam having a width 
between flanges equal to the thickness of the plank to be used. 
These plank holders are laid in pairs, with web horizontal, one 
on either side of the wall, and connected by a bolt passing 
through them and through the wall. 2 Two rows of planks on 
edge are first placed around the building so as to inclose the pro- 



1 Mr. E. L. Ransome, Architect and Engineer, Engineering Record, Feb. 
6 and 13, 1904. 

2 Patented by Thomas G. Farrell, Washington, N. J. 



432 CEMENT AND CONCRETE 

posed wall. At the upper side of each junction between two 
planks in the same horizontal row is placed one of these plank 
holders. Another horizontal row of planks may now be placed, 
with the iron plank holders at the joints as before. As the 
wall is built up, the lower planks and holders may be removed 
and placed on top, and thus few forms are required. Tees and 
L-forms are provided for partition walls and corners. 

When an air space is desired in a wall a special terra cotta 
tile or building block may be built into the wall, but this is 
quite expensive, and an interior collapsible form may be made 
of timber by the use of two planks held apart by a wooden 
brace which may be knocked out. Special means of handling 
the interior plank should be provided, and the building of a 
high wall cannot be continuous with this method. 

628. New York Building Regulations. While city building 
regulations are not always criteria of good practice, yet the 
Regulations of the Bureau of Buildings of the Borough of 
Manhattan concerning the use of concrete-steel construction are 
exceptional. Emanating from a bureau that has been dis- 
tinctly hostile to concrete-steel, they are naturally conservative, 
but are, on the whole, excellent, and work conscientiously done 
in accordance with them will not bring discredit on concrete 
construction. 

It is specified that the cement shall be only high grade 
Portland standing certain tests, that the sand shall be clean 
and sharp, aggregate, broken trap, or gravel of a size that will 
pass a three-quarter inch ring, and that the proportions used 
shall be one cement, two sand and four of stone or gravel, or 
that the concrete shall have a crushing strength of two thou- 
sand pounds per square inch in twenty-eight days. Only the 
best quality of concrete is thus permitted. 

629. The Regulations concerning the design are then stated 
as follows : 

"Concrete-steel shall be so designed that the stresses in the 
concrete and the steel shall not exceed the following limits : 

Extreme fiber stress on concrete in compression, 500 Ibs. per sq. in. 

Shearing stress in concrete 50 " " 

Concrete in direct compression 350 " " 

Tensile stress in steel 16,000 " " 

Shearing stress in steel 10,000 " " 



BUILDINGS 433 

i 

"The adhesion of concrete to steel shall be assumed to be 
not greater than the shearing strength of the concrete. 

"The ratio of the moduli of elasticity of concrete and steel 
shall be taken as one to twelve. 

"The following assumption shall guide in the determination 
of the bending-moments due to the external forces: Beams and 
girders shall be considered as simply supported at the ends, no 
allowance being made for the continuous construction over 
supports. Floor plates when constructed continuous and when 
provided with reinforcement at top of plate over the supports, 
may be treated as continuous beams, the bending-moment for 

uniformly distributed loads being taken at not less than ; 

the bending-moment may be taken as in the case of square 

floor plates which are reinforced in both directions and sup- 
ported on all sides. The floor plate to the extent of not more 
than ten times the width of any beam or girder may be taken 
as part of that beam or girder in computing its moment of 
resistance. 

"The moment of resistance of any concrete-steel construc- 
tion under transverse loads shall be determined by formulas 
based on the following assumptions: 

"(a) The bond between the concrete and steel is sufficient 
to make the two materials act together as a homogeneous solid. 

"(6) The strain in any fiber is directly proportionate to the 
distance of that fiber from the neutral axis. 

"(c) The modulus of elasticity of the concrete remains con- 
stant within the limits of the working stresses fixed in these 
Regulations. 

"From these assumptions it follows that the stress in any 
fiber is directly proportionate to the distance of that fiber from 
the neutral axis. 

"The tensile strength of the concrete shall not be considered. 

"When the shearing stresses developed in any part of a 
construction exceed the safe working strength of concrete, as 
fixed in these Regulations, a sufficient amount of steel shall be 
introduced in such a position that the deficiency in the resist- 
ance to shear is overcome. 

"When the safe limit of adhesion between the concrete and 



434 CEMENT AND CONCRETE 

steel is exceeded, some provision must be made for transmitting 
the strength of the steel to the concrete. 

" Concrete-steel may be used for columns in which the ratio 
of length to least side or diameter does not exceed twelve. 
The reinforcing rods must be tied together at intervals of not 
more than the least side or diameter of the column. 

"The contractor must be prepared to make load tests on 
any portion of a concrete-steel construction, within a reasonable 
time after erection, as often as may be required by the Super- 
intendent of Buildings. The tests must show that the con- 
struction will sustain a load of three times that for which it is 
designed, without any sign of failure." 

ART. 73. CONCRETE WALKS 

630. One of the most important uses of concrete is in the 
construction of street and. park walks. It has not only driven 
stone flagging almost out of use, but it is being employed to a 
large extent in towns and villages where board walks have 
formerly been used almost exclusively. 

A concrete walk is made up of a sub-base or foundation, a 
base, and a wearing surface. 

631. Foundation. As in other structures, one of the most 
important essentials for success lies in the preparation of the 
foundation, and the care that must be bestowed on it will 
depend upon the character of the soil and the climate. In the 
higher latitudes of the United States, frost may soon destroy a 
walk the foundation of which is not well drained. 

The excavation should be made to the sub-grade previously 
determined upon, any objectionable material such as loam or 
organic matter being removed, and the bottom of the excava- 
tion smoothed and well rammed. Upon this is laid the sub- 
base, its thickness varying from nothing to twelve inches. In 
a sandy soil with good natural drainage and little danger from 
frost, and where light traffic is expected, it may be unnecessary 
to provide any special sub-base, since the soil itself furnishes a 
good foundation for the concrete, but in clay soil in northern 
climates, twelve inches of sub-base may be required. The best 
material for this sub-base is broken stone varying in size from 
one-half inch to two and one-half inches. Usually broken 
stone is considered too expensive, and gravel, coarse sand, 



CONCRETE WALKS 435 

cinders, or broken brick is employed. A layer four inches thick is 
usually sufficient for good materials, but six to twelve inches of 
cinders are sometimes required. It should be well rammed to 
a level surface, and when completed should be firm but porous. 

The most important point is that this course shall have 
good drainage, otherwise it may be a menace to the walk. If 
it is more porous than the retaining soil, it will naturally drain 
this soil, and if the water is not able to escape into the sewer 
or elsewhere, it may be frozen and heave the walk. An ex- 
cellent plan sometimes adopted is to lay at intervals of twenty 
to twenty-five feet, a blind stone drain from the walk founda- 
tion to the foundation of the curb. In exceptional cases it may 
be necessary to lay a tile drain in the sub-base to lead the water 
away from the walk. 

632. Base. The base is the body of the walk giving stiff- 
ness to the structure. Its functions are to furnish a solid 
foundation for the wearing surface and to give transverse 
strength to the walk, transmitting the pressure uniformly to 
the sub-base. The base is of concrete, which need not be very 
rich for ordinary traffic. A proportion of one part packed 
Portland cement to two and one-half volumes of dry sand 
and six volumes broken stone is excellent, and proportions of 
one, three and seven parts cement, sand and stone, respectively, 
will usually be found sufficient, though the richer the concrete in 
the base the better will the top dressing adhere to it. 

The broken stone for this concrete should be of a size not 
exceeding one and one-half inches in any dimension, some cities 
requiring three-quarters inch or less. Crushed granite and trap 
are excellent, though limestone or any other moderately hard 
rock may be used that is suited to making concrete for ordinary 
purposes. If of a hard rock, the screenings may well be left 
in the broken stone, and when this is done, the dose of sand 
should be diminished. (See Art. 37.) 

The thickness of the layer of concrete should not be less 
than three inches. Four inches is much better and is recom- 
mended for general use in sidewalks, while in exceptional cases 
six inches is required. The top of the concrete base should 
be finished to a plane parallel to the proposed surface of the 
walk and at a distance below it equal to the proposed thickness 
of the top dressing. 



436 CEMENT AND CONCRETE 

633. Wearing Surface. The preparation and application 
of the wearing surface require much care if satisfactory result 
are to be obtained. The most evident service of this layer i 
to withstand wear, and it should therefore be made of rich 
Portland cement mortar. With a sand consisting principally 
of quartz particles, it is found that a mortar composed of equal 
parts cement and sand gives about the best results in tests of 
abrasion. If the mortar is used richer than this, it is likely to 
check or crackle in setting, marring the appearance of the walk. 
Mortar containing two parts cement to three parts sand gives 
nearly as good results, and two parts sand or fine crushed 
granite to one of Portland cement is usually satisfactory. The 
sand for the mortar should be quartz if possible, or crushed 
granite or trap. It should be screened through a quarter 
inch mesh, and there should not be a large proportion of 
very fine particles. 

The thickness of the layer of top dressing is usually about 
one inch, and this is probably the maximum thickness ever 
required. One-half inch of top dressing is believed to be suf- 
ficient when the wear is not excessive, provided the base has 
been carefully leveled. 

634. The Construction of the Walk. If the walk has not a 
considerable longitudinal slope, it should be given a transverse 
slope of about a quarter inch to the foot to provide for draining 
the surface. 

Stakes for grade and line having been given, a maitre cord 
is stretched along the line stakes to mark the sides of ike exca- 
vation. After the material has been excavated to the proper 
sub-grade and all soft material in the bottom removed, the 
bottom of the trench is well rammed. If tile drain is necessary, 
it is laid with open joints on this foundation. The material to 
form the sub-base is now wheeled in and rammed to the proper 
thickness, water being used freely if it facilitates the packing. 
The top of the sub-base is brought to a level plane'at the proper 
distance below the grade stakes. 

The molds for the walk are now to be laid. These are made 
of two by four or two by six inch scantling, sized and dressed 
on at least one side and one edge. Stakes are first securely 
driven, about five or six feet apart, with their faces two inches 
back from the side lines of the proposed walk, and their tops 



CONCRETE WALKS 437 

at grade. Against these stakes the scantlings are placed on 
edge with dressed side toward the walk, and smooth edge level 
with the grade stakes. These molds are held in place by nail- 
ing through the supporting stakes into the scantling, and if 
these nails are not driven "home," they may easily be pulled 
to release the mold when the work is completed. On the upper 
edges of the mold are then marked off the sizes of blocks de- 
sired, being careful that the marks defining a joint are exactly 
opposite each other on the two scantlings. 

635. The concrete materials having been previously deliv- 
ered near the work, the concrete is mixed, either by hand or 
machine, according to the methods already given, and rammed 
in place after the sub-base has been well wet down to receive 
the concrete. The concrete should be just short of quaking, 
and in ramming care must be taken not to disturb the molds. 
For tamping next the molds, the makers of cement working 
tools offer a light rammer with square face at one end and 
blunt, chisel shaped tamper at the other. The surface of the 
base is brought to a plane parallel to the proposed finished sur- 
face of the walk, and at a distance below it equal to the thick- 
ness of the top dressing. A straight edge, long enough to span 
the walk and notched out at the ends so that when placed on 
the molds the straight edge will define the correct grade of the 
base, is a convenience here. 

636. The concrete is now cut into blocks exactly corre- 
sponding to the proposed blocks in the top dressing. For this 
purpose a straight edge is laid across the walk in line with 
marks previously made on the molds to define the joints, and 
with a spade or special tool the concrete base is cut entirely 
through to the sub-base. This division is necessary to allow for 
expansion and contraction, and prevent cracks in the top dressing 
elsewhere than at the joints. This joint in the base should 
then be filled with clean sand. If preferred, these joints in the 
base may be made by placing thin steel strips across the molds 
to be removed after the concrete for the next block is in place. 

The end block made from a given batch of concrete should 
be limited by a cross mold set exactly on line of a proposed 
joint. When the base is continued, this cross mold is removed. 
A part of a block should never be molded and then built on 
after having stood long enough to begin to set. Any concrete 



438 CEMENT AND CONCRETE 

left over from finishing a block should either be mixed in with 
the next batch, if this is to follow in a very short time, or it 
should be wasted. A disregard of this rule will probably result 
in a crack in the top dressing above the line of division between 
adjacent batches. 

637. When a block of base is finished, the top dressing or 
wearing surface should be applied immediately. The lack of 
adhesion between the base and wearing surface is one of the 
most frequent causes of failure in cement walks. The mortar 
should not merely be laid on in a thick layer and then struck 
off to grade, but it should be worked and beaten into close con- 
tact with the concrete at every point. The mortar should be 
tamped with a light rammer and beaten with a wooden batten, 
and to accomplish this properly the mortar must not be very 
wet. The surface is then to be struck off with a straight edge 
bearing on the top of the mold planks. Some hollows or rough 
places will remain, and the straight edge should be run over a 
second or perhaps a third time, a small amount of rather moist 
mortar, made from thoroughly screened sand, having been first 
applied to such places. 

When the surface film of water is being absorbed, the surface 
is worked with a wooden float. The exact time when the work 
should be floated will soon be known by experience. After the 
floating is completed, the trowel may be used to give a smoother 
surface, but this makes the walk so slippery that it is not usually 
desirable. 

638. If the top dressing is worked too long, the cement is 
brought to the surface, robbing the next lower layer of its ce- 
ment and resulting in scaling. The top dressing is now cut 
entirely through on exact line above the joints in the base. 
This may be done by a trowel working against a straight edge, 
but special tools are made for cutting through the mortar and 
rounding the edges of the joint at one operation. A quarter- 
round tool is also run along the edges of the mold to give a neat 
finish. When desired, an imprint roller run over the walk 
gives it the appearance of having been bushhammered. 

It is important that the top dressing be applied before the 
concrete has begun to set, and it must not be applied to a por- 
tion of a block and then some time allowed to elapse before 
applying the remainder. The edge of the top dressing must 



CONCRETE WALKS 439 

be cut off squarely at the end of the block. If desired, the 
wearing surface may be colored by the use of lamp black in the 
mortar, giving a uniform gray color to the walk. ( 535.) 

639. When the walk is completed, it should be fenced off 
so that animals may not walk over it while still fresh, and it 
should be protected from a hot sun. The surface should be 
kept moist, and this may be done after the first twenty-four 
hours by spreading a layer of damp sand over the walk and 
wetting the sand with a rose nozzle as often as may be needed. 
The walk may be opened to light travel after about'_four days, but 
it is better to remain covered with the damp sand for a week. 

640. Cost of Concrete Walk. The cost of concrete walks 
varies from ten cents to twenty-five cents per square foot. A 
fair price for a walk of average quality where there are no 
special difficulties is twelve to eighteen cents per square foot. 

As an instance of a walk built with special care, the one 
constructed about the top of the bank of the Forbes Hill Reser- 
voir may be mentioned. 1 The sub-base of this walk was of 
stone and twelve inches thick, the layer of concrete was five 
inches thick at the center of the walk and four inches at the 
sides. The top was of granolithic finish one inch in thickness. 
The walk was laid in separate blocks about six feet square. 
The average gang employed on the concrete consisted of six 
men and one team, while the finishing was done by two masons 
and one tender. The amount laid per day was about forty 
square yards. The cost per square yard was as follows: - 

| cu. yd. stone in foundation or sub-base, at $.40 per cu. yd. . $0.133 
Labor, placing stone at $1.50 per day ........... 502 

Total cost stone foundation per sq. yd. of walk . . . $0.635 

.158 bbl. cement, at $1.53 per bbl. . . . ........ $0.242 

.065 cu. yd. sand, at $1.02 per cu. yd. ..... ...... 066 

.109 cu. yd. stone, at $1.57 per cu. yd ............ 170 

Labor, mixing and placing concrete ............ 450 

Total cost concrete base per sq. yd ......... $0.928 

.11 bbl. cement, at $1.53 per bbl ............. $0.168 

.022 cu. yd. sand, at $1.02 per cu. yd ............ 022 

Lampblack ...................... 008 

Labor, preparing and finishing surface ........... 149 

Total cost top dressing or wearing surface ..... $0.347 



Total cost walk per s q . yd. | ;?' ; ; ; ; 



C. M. Saville, M. Am. Soc. C. E., Engineering News, March 13, 1902. 



440 CEMENT AND CONCRETE 

641. The following is given as an estimate of cost of items 
in a walk built with six inch cinder sub-base, four inch concrete 
base and one inch top dressing . 

COST PER SQ. YD. OP WALK 

MATERIALS LABOR 

Preparation of foundation, excavation and ramming . . . $0.20 

Sub-base, 6 in. cinders cu. yd., at $0.40 cu. yd $0.07 

Placing and ramming cinders 0.04 

cu. yd. concrete, at $3.00 per cu. yd. for materials alone . 0.33 

^ cu. yd. concrete, placing, at $1.80 per cu. yd 0.20 

Top dressing -fa cu. yd. mortar, at $9.00 per cu. yd . . . . 0.25 

Placing top dressing and finishing walk 0.25 

Superintendence and molds 0.10 



Totals $0.65 $0.79 

Total cost per sq. yd., $1.44, or 16 cents per sq. ft. 

642. As an example of a low priced walk, the concrete walks 
in San Francisco 1 are but three inches thick, two and one-half 
inches of concrete composed of one part Portland cement, two 
parts beach gravel, and six parts of crushed rock of size not 
exceeding one inch; the top dressing being one-half inch thick 
of equal parts Portland cement and beach gravel. With ce- 
ment $2.50 per bbl., crushed rock and gravel from $1.40 to 
$1.75 per cu. yd., and wages twenty cents an hour for laborers 
and forty cents for finishers, this walk is constructed at from 
nine to ten cents per square foot. It is stated that a gang of 
three or four men will lay 150 to 175 square feet per day. 

ART. 74. FLOORS OF BASEMENTS, STABLES AND FACTORIES 

643. The principles governing the laying of walks apply also 
in a general way to the construction of floors that rest directly 
on the ground. 

For residences, basement floors may be laid with three inch 
base of concrete and one-half inch wearing surface. The thick- 
ness of sub-base will depend upon the character of the soil. 
Where natural conditions do not assure good drainage of the 
foundation, this should always be provided for by either a blind 
stone or tile drain laid around the outer edge of the building 
and leading to the sewer or other outlet. The finished surface 
of the floor should always have a slight slope toward the center 



Engineering News, March 4, 1897. 



FLOORS 441 

or one corner of the basement, and a trapped sewer connection 
set at this lowest point in such a way that it is accessible for 
repairs and cleaning. 

644. Wet Basements. Where much ground water is en- 
countered, and especially where a basement is subjected to a 
head of water from without, special precautions must be taken 
in building the floor. The concrete must be made thick enough 
so that its weight and the arch action set up, shall be able to 
withstand the upward pressure of the water. In building such 
a floor it is necessary to keep a sump hole, preferably in the 
center, towards which the construction proceeds from the sides. 
A pipe placed in the sump hole permits pumping until the con- 
crete is laid about the pipe, when the latter may be filled with 
rich cement mortar. In such cases the side walls of the base- 
ment should be plastered with Portland cement mortar on the 
outside and special care taken in joining the floor to the wall. 

645. Size of Blocks. As the changes in temperature in a 
building are usually much less than in open air, the blocks of 
concrete may be of much larger size, say ten feet square, and 
many basement floors are laid without any joints, though 
sooner or later they will probably crack if so laid. In factories 
for certain purposes, however, the floors may be subjected to 
greater changes in temperature than walks laid in the open 
air. In such cases the blocks should not be more than three 
or four feet on a side, and the joints may well be filled with 
asphalt, especially if water- tightness is desired. 

646. Stable floors may be made of six inch cobble or broken 
stone sub-base, six inches of concrete made with mortar con- 
taining three parts sand to one cement, and one inch of top 
dressing containing three parts sand (mixed sizes) or crushed 
granite to two parts cement. 

Factories having heavy machinery with much vibration re- 
quire strong floors. Such a floor may be made of six inches 
of cobble stone sub-base covered by six inches of a lean concrete 
made with one-to-four mortar, and above this, three to five 
inches of rich concrete made with mortar containing two and 
one-half parts sand to one cement, and one inch of top dressing, 
equal parts cement and sand or cement and crushed granite. 

647. Example and Cost. In the construction of the new 
printing building for the Government Printing Office at Wash- 



442 CEMENT AND CONCRETE 

ington, the basement floor is nine inches thick, made as fol- 
lows : l 

1. Concrete sub-base, six inches thick of one part natural 
cement, two parts sand and four and one-half parts broken 
brick. 

2. Concrete base, two and one-half inches thick of Portland 
cement one part, sand two parts and fine broken gneiss four 
parts. 

3. Top dressing, one-half inch in thickness, of two parts 
sand to one part Portland cement. 

The cost of this floor was about $1.50 per square yard, or 

about seventeen cents per square foot. 



ART. 75. CONCRETE IN PAVEMENTS AND DRIVEWAYS 

648. PAVEMENT FOUNDATIONS. The principal use of con- 
crete in connection with city pavements has been as a founda- 
tion, the wearing surface being of some other material, as brick, 
asphalt, cedar blocks, etc. 

Concrete for pavement foundations should not be less than 
six inches in thickness, and a greater thickness will be required 
where the ground is insecure. The excavation having been 
made to the required sub-grade, and all loose soil removed and 
the places refilled with broken stone, the earth is thoroughly 
rolled to a smooth surface parallel to the surface of the proposed 
pavement. Drainage for the foundation should be provided 
where necessary by broken stone or tile drains beneath the 
curb. Before beginning the placing of concrete, stakes may be 
driven in the foundation, with their tops at grade, at intervals 
of five to ten feet over the entire pavement, to assist in securing 
the proper grade of concrete surface. 

649. The stone for the concrete should be broken so that no 
piece is larger than two and one-half inches in its greatest di- 
mension. If the stone is of good quality, it need not be screened 
except to remove the finest dust, if this is present in consider- 
able quantities. Sufficient mortar should be used to fill the 
voids in the stone, this mortar being composed of about two 
parts sand to one of natural cement, or better, two and one- 
half or three parts sand to one of Portland cement. This con- 



x Report of Capt. John S. Sewall, Report Chief of Engineers, 1896.- 



PAVEMENTS 443 

crete is thoroughly rammed in place, care being taken that 
adjacent batches as laid in the street mingle with each other 
so as to show no line of demarcation. In stopping work for 
the night, the concrete should cut off sharply oil a straight 
line parallel to the direction of the proposed joints in the wear- 
ing surface. Joints extending across the street should be left 
at intervals of thirty to forty feet to allow for expansion and 
contraction. 

650. The concrete is finished to a surface parallel with the 
proposed street surface, a templet being employed to secure 
this. The concrete should be kept damp for a few days, and 
no traffic allowed upon it until the wearing surface is laid. If 
the wearing surface is of brick or wooden blocks, a layer of 
sand about one inch thick is first spread over the concrete. 

The advantages of a concrete foundation for street pave- 
ments are its strength and durability and water- tightness. 

651. CONCRETE PAVEMENT. Concrete has not been a popu- 
lar material for a street surface except for short driveways and 
in courts where both vehicles and pedestrians must be accom- 
modated. One reason for this is that concrete is slippery, and 
another, that owing probably to carelessness or ignorance, the 
wearing qualities have not been good. The first objection may 
be largely removed by cutting the surface into blocks, four by 
eight inches, by deep grooves, or by the use of a deep imprint 
roller on the wearing surface. As to wearing qualities, there 
seems to be no good reason why a concrete cannot be made 
tough enough to withstand heavy traffic. It will of course be 
necessary to divide the work into blocks of twenty to twenty- 
five square feet, with expansion joints of sand, asphalt, or 
tarred paper between. A third objection is the glare of the 
surface in summer. A partial remedy for this may be had by 
placing some coloring matter, such as lamp black, in the top 
dressing. 

652. The sub-base may consist of a six inch layer of broken 
stone, or twelve inches of cinders, well drained and thoroughly 
compacted by rolling. For exceptionally heavy wear it may be 
advisable to use a five inch layer of lean concrete for the sub- 
base, after rolling the bottom of the excavation and providing 
drainage. 

Upon the sub-base should be laid a base, composed of four 



444 CEMENT AND CONCRETE 

inches of concrete made with first class stone, such as granite, 
trap or hard limestone crushed to pass a ring one and one-half 
inches in diameter, and containing enough mortar, one part 
Portland cement to two or three parts sand, to fill the voids in 
the stone. The top dressing, a layer of granolithic one and one- 
half or two inches thick, should then be immediately applied. 
This mortar should be made with one or two parts granite, trap, 
or other hard rock crushed to pass a five-eighths inch screen, to 
one part Portland cement. 

These two layers are placed in much the same manner as 
that described for laying concrete sidewalks, but the joints in 
base and top dressing should run at angles of forty-five degrees 
with the curb to prevent ruts following the lines of the joints. 
A roller making deep imprints is then run over the finished 
surface to furnish a foothold for horses, or, for this purpose a 
special roller may be used to mark the top dressing into blocks 
approximately four by eight inches, with deep (one-half inch) 
grooves. 

When completed, the pavement should be kept moist, pref- 
erably by a layer of damp sand, and no traffic should be al- 
lowed upon it for at least a week or ten days. 

653. Concrete pavement laid in Belief ontaine, Ohio, was 
found to be in good condition after ten years' service; 1 the 
only serious defect apparent being that, since the blocks were 
marked off parallel to the curb, ruts have sometimes formed 
along these joints. This pavement was made with four inches 
base concrete, laid directly on sub-grade where foundation 
is gravel, sand or porous soil; or if soil is impervious, the 
base was laid on four inches of broken stone or cinders. The 
top layer was two inches thick, equal parts cement and sand or 
pea granite. Sub-drains of three inch tile were laid inside each 
curb line, and the curb is formed as part of the outer blocks. 
Both the base and top dressing were cut through in squares, 
five feet on a side. The cost of the pavement is said to have 
been $2.15 per square yard, and very few repairs have been 
found necessary. 

In Germany a cement macadam, made with six inch sub- 



1 Municipal Engineering, December, 1900, and Engineering News, Jan. 7, 
1904. 



CURBS AND GUTTERS 445 

base of broken stone or gravel, with a wearing surface of hard 
macadam stone mixed with cement, has been successfully used. 

ART. 76. CURBS AND GUTTERS 

654. The use of concrete for curbs and gutters is rapidly 
increasing. Curbing is sometimes molded and afterward put in 
place like stone curbing, but the greatest advantages in the use 
of concrete for this purpose are only attained by molding in 
place the curb and gutter as one structure. 

The Parkhurst combined curb and gutter is a patented form 
that has proved very satisfactory. This form has a projection 
of about one inch at the back and another along the bottom 
just below the curb, this feature being patented. 

A combined curb and gutter may consist of a curb four to 
six inches wide at the top, and five to seven inches at the bot- 
tom, and have a face of six to seven inches above the gutter. 
The upper face corner of the curb and the angle between curb 
and gutter should be rounded with a radius of one and one- 
half to two inches. The- gutter is sixteen to twenty inches 
wide, and from six to nine inches thick, with top surface con- 
forming to the grade of the street. 

655. The sub-base should consist of a layer of broken stone 
six inches thick, or six to twelve inches of cinders thoroughly 
rammed. The preparation of the foundation should be similar 
to that required for a pavement, care being taken that the 
sub-base be thoroughly drained, tile being used if necessary. 
Forms to receive the concrete are held in place by stakes, the 
molds being carefully set to grade. The sub-base may now be 
covered by a layer of four to six inches of Portland concrete of 
only moderate richness, as one to three to six, and the concrete 
to form the curb and gutter placed upon it before it has set, 
or a six inch layer to form the gutter may be placed directly 
on the sub-base. 

656. Concrete to form the curb and gutter should be of good 
quality, not more than two and one-half parts sand to one part 
Portland cement being used for the mortar, and sufficient mor- 
tar used to entirely fill the voids in the stone. The broken 
stone for this concrete should be rather fine, with few, if any, 
pieces larger than one inch in greatest dimension. The ex- 
posed faces receive a top-dressing, or wearing surface, of one- 



446 CEMENT AND CONCRETE 

half inch to one inch of granolithic containing not more than 
one and one-half parts of trap or granite, pea size, to one part 
Portland cement. This coating is applied as soon as possible 
after the concrete is placed, as in sidewalk work. The surface 
is troweled or floated, but a smooth, glossy finish is avoided. 

The curb and gutter may well be laid in alternate blocks 
about six feet long, but a somewhat neater appearance is se- 
cured by making the work continuous, and cutting it entirely 
through at intervals of six feet to provide for slight movement. 
As the molds may be used repeatedly, they should be sub- 
stantially made. Special forms are of course required at corners, 
catch basins, etc. As in other concrete construction, the work 
should be protected from injury and kept moist for at least a 
week. 

657. On business streets it is desirable to build the sidewalk 
close to the curb, with only a joint between, the grade of the 
walk conforming to the curb and sloping up toward the build- 
ing line one-quarter inch to t^e foot. On residence streets the 
walk should be separated from the curb by a park strip, the 
walk being high enough to give drainage toward the curb. 

Steel facing is sometimes used for curbs subjected to excep- 
tional wear, as in front of shipping warehouses and freight 
sheds. Where these are applied, they should cover the top 
and the upper part of the face of the curb and must be well 
anchored, by bolts or special webs, to a substantial mass of 
concrete, otherwise they will work loose and defeat the object 
for which they are used. 

658. Cost of Concrete Curb and Gutter. At Champaign, 
111., 1 a curb was built seven inches high and five inches thick, 
the gutter, six inches thick, extending nineteen inches into the 
roadway from the face of the curb. The foundation consisted 
of six inches of gravel or cinders well rammed. The concrete 
was composed of one part Portland cement to five parts fine 
gravel, and the finishing coat, one inch thick, was of one part 
Portland cement to one part clean, sharp, coarse sand. The 
cost per foot was thirty-nine cents, including all excavation. 

A similar curb at Urbana, 111., was 4 inches thick at the 
top, 5 inches at the base and 7 inches high; the gutter being 



1 W. H. Tarrant, Engineer, Proc. 111. Soc. Engr. and Surveyors, 1899. 



STREET RAILWAY FOUNDATIONS 447 

5 inches thick and extending 18 inches into the roadway. The 
foundation was composed of eight inches of cinders or gravel. 
The concrete was of one part Portland cement to five parts 
clean gravel, and the finishing coat was one inch thick, com- 
posed of one part Portland cement to two parts sharp sand. 
The price per linear foot, including the excavation, removal of 
old curbing, and refilling, was forty-six cents. 

At South Bend, cement curb alone, 6 inches wide at top, 
7 inches at bottom and 16 inches depth, with the upper half 
composed entirely of one to two Portland cement mortar, has 
been constructed for eighteen cents per linear foot. 

ART. 77. STREET RAILWAY FOUNDATIONS 

659. The heavy motor cars used on city and urban electric 
railways subject the track to very severe service. As the head 
of the rail must be practically flush with- the pavement on city 
streets, cross-ties, when used, are so far beneath the surface 
that they decay rapidly and their renewal entails the tearing 
up of the pavement. As there is not the same necessity for a 
cross-tie on street tracks as on railroads, since the rails are held 
to gage by the pavement, these objections to the cross-tie have 
led to the adoption of a concrete girder under each rail. The 
rails and ties (if ties are used) should not only rest upon the 
concrete, but should be imbedded in it. Track in which the 
rails rested upon concrete, but were not imbedded in it, has 
been found to yield laterally and get out of alinement, while 
on the other hand, if the ties rest upon earth or gravel and are 
filled between with concrete, the track is likely to settle, break- 
ing the bond of the concrete. 

660. The method of placing concrete beams for street rail- 
way tracks in Minneapolis was as follows : l The rails were first 
spiked to cross-ties at intervals of six to eight feet, and the rail 
joints cast-welded. In laying the street pavement foundation 
of natural cement concrete, a rough groove, fifteen inches wide 
at the bottom and eighteen to twenty inches at the top, was 
left under each rail. This groove was immediately filled be- 
tween ties with concrete made of one part Portland cement, 



1 F. W, Cappelen, M. Am. Soc. C. E., Engineering News, Oct. 14, 1897; 
Municipal Engineering, November, 1896. 



448 CEMENT AND CONCRETE 

two and one-half parts sand, and four and one-half parts broken 
stone. 

The rails were tied together every ten feet with wrought 
iron tie bars, three-eighths inch by two inches, set on edge. These 
tie bars were rounded at the ends, threaded and attached to 
the web of the rail by two nuts, one on either side of the web. 
The rails were then spiked to the concrete beam, the temporary 
wooden ties removed, and the spaces left by them filled with 
concrete, completing the beam. As the concrete beam was 
eight inches thick and the rail five inches, the sub-grade was 
thirteen inches below the top of the rail. 

On the gage side of the rail were' placed toothing blocks of 
granite, 3J by 9 inches by 4J inches deep, held away from the 
rail 1J inches by temporary wooden strips. After removing 
these strips, cement grout was poured into the groove to fill 
2J inches over the base of the rail, the remaining 2J inches to 
the top of the rail being filled by asphaltic cement which re- 
mained soft enough to permit a flange groove to be made by the 
first car over the track. The asphalt wearing surface was laid 
against the rail on the outer side. Mr. Cappelen, in describing 
this construction, says that a rail six inches high with six-inch 
base should be used, with granite toothing blocks, six by nine 
inches by five and one-half inches deep. 

The cost per foot of rail for the concrete beam construction 
only, was twenty-six to twenty-seven cents, and for the filler, 
five cents per foot. The cost per mile of double track, exclusive 
of rails and pavement, was about $8,670.00. 

Somewhat similar methods have been employed in Toronto 
and Montreal. Canada, Indianapolis, Ind., and Scranton, Pa., 
Denver, Detroit and Cincinnati. 

661. At Scranton, Pa., 1 the rails were laid directly on the 
six-inch concrete base of the pavement. This thickness was 
increased to twelve inches under the joints (which were rein- 
forced by an inverted rail four feet long) and under steel cross- 
ties spaced ten feet centers and formed of old girder rails in- 
verted and riveted through the flanges at the intersection. 
Flat steel tie bars, threaded at the ends, spaced ten feet centers, 
were also used here as at Minneapolis. 



1 Description of the systems employed in several cities are given in En- 
gineering News, Dec. 26, 1901. 



STREET RAILWAY FOUNDATIONS 449 

The concrete mixing plant was mounted on a car running 
on the track; the materials were delivered to the machine by 
hand measuring boxes, and the Drake mixer deposited the con- 
crete directly into the trench. The total cost per foot of track 
is given as $2.65, $1.17 of which was for grading, rolling, con- 
creting and brick paving at $1.97 per square yard, and for extra 
concrete at joints and ties at $0.72 per square yard. 

662. At Toronto, Canada, the six-inch concrete base of the 
pavement is increased to eight inches in thickness for twenty 
inches width under each rail, and the base of the latter is im- 
bedded one inch in the concrete. A 6|-inch grooved girder 
rail is used, with mortar rammed between the web and the 
adjacent paving blocks. 

663. At Cincinnati the bottom of the concrete stringer is 
nine inches below the base of the nine-inch grooved girder rail, 
and the concrete is built up from three to six inches on the web, 
according to the thickness of the wearing surface of the pave- 
ment. The space between the upper part of the web and the 
adjacent paving is then filled with cement mortar, thus sup- 
porting the head of the rail as well as protecting the web from 
corrosion. 



CHAPTER XXI 

SPECIAL USES OF CONCRETE (CONTINUED). SEWERS, SUB- 
WAYS, AND RESERVOIRS. 

ART. 78. SEWERS 

664. There seems to be no very good reason why concrete is 
not more generally employed in the construction of all large 
sewers. With sizes less than two or two and one-half feet in 
diameter the difficulty of removing the centers prohibits the use 
of concrete in the ordinary way, and although some appliances 
have been devised for building these small sewers as monoliths 
by a mold that advances as fast as the concrete is tamped in 
place, they have not proved popular. The difficulty of obtain- 
ing a perfect grade, and the undesirable feature of leaving the 
green concrete unsupported, are probably reasons sufficient for 
this lack of popularity. 

For the larger size sewers concrete has several advantages 
over brick. First may be mentioned the very smooth finish 
that may be obtained on the invert, appreciably increasing the 
velocity of flow over that usually obtained with brick inverts. 
Cheaper labor may be employed in concrete work with less 
danger of annoyances from strikes. The cost is from one- 
third to one-half less than for brick. 

665. METHODS OF CONSTRUCTION. The City of Washing- 
ton was one of the first to use concrete extensively in sewer 
construction 1 . For sizes up to twenty-four inches internal diam- 
eter the concrete is used only as a foundation and bedding 
for the ordinary sewer pipe. For a twenty-four inch sewer 
the pipe rests in a bed of concrete twenty-seven inches wide 
at the bottom, flaring to forty inches wide at the level of the 
center of the pipe, and then carried up with plumb sides for 
six inches, and finally finished by planes tangent to the upper 
curve of the pipe. At the joints there are bands of concrete 



1 Described by Capt. Lansing H. Beach, Corps of Engrs., U. S. A. Report 
Operations District of Columbia, 1895. 

450 



SEWERS 451 

extending over the top, so that at these places the pipe is en- 
tirely inclosed. Similar forms are used for the smaller sizes 
with corresponding decreased dimensions. For all sewers be- 
tween ten inches and twenty-four inches the sub-grade is six 
inches below the exterior of the pipe, and in all cases the band 
about the joint is four inches thick at the top. 

666. The method of laying these sewers is as follows: The 
trenches are 2J to 3 feet in width, with "headers" about 2 feet 
wide, left at intervals of 10 to 16 feet, which are tunneled 
through. The grade and line pegs are placed in the headers 
at the ground surface, and a cord is stretched on the sewer line 
over at least four stakes, at a convenient height above the 
grade, and thus parallel to the bottom of the sewer. 

When the trench is to the required grade, a six inch layer of 
concrete, made with one barrel natural cement, two barrels 
sand and four barrels gravel, is placed. This concrete is rammed 
with iron rammers weighing sixteen pounds, and having eighteen 
square inches ramming surface. The pipe is then laid upon 
this bed and each section is tested for line and grade. For the 
former, a plumb bob is used with its cord held against the 
grade cord already mentioned, and for testing the grade a grad- 
uated pole is used, with a projection at the bottom which sets 
on the interior of the pipe, just within the open end. 

Concrete is then lowered in buckets, deposited on top of 
the pipe and allowed to fall down on the sides so as not to 
disturb the alinement. When enough concrete to secure the 
pipe has been thus placed, it is rammed and the concreting 
continued until the required form is obtained, as already de- 
scribed. The concrete in the bands carried over the joints is not 
rammed but is beaten with wooden paddles and heavy trowels to 
compact it and bring it to the desired form, four inches thick 
and four inches wide at the top, and flaring to twelve inches 
wide (in the direction of the sewer) at the top of the pipe. 

667. Cost. The quantities of concrete materials required 
to lay one hundred linear feet of pipe sewers as described 
above are given as follows : 

Size of sewer 8 inch 12 inch 18 inch 24 inch 

Cement, bbls 6.76 10.58 14.77 19.14 

Sand, cu. yd 2.07 3.23 4.52 5.85 

Gravel, cu. yd 4.16 6.47 9.04 11.70 



452 CEMENT AND CONCRETE 

With natural cement costing $0.79 per barrel in sacks, sand 
$0.47 per cu. yd., gravel $0.75 per cu. yd., and laborers $1.50 
to $1.75 per day, foremen, masons and inspectors $4.00 per 
day, the average cost of laying pipe sewers in this manner was 
approximately as follows, exclusive of the cost of the pipe: 
8-inch, $1.11; 12-inch, $1.14; 15-inch, $1.46; 18-inch, $1.60; 
21-inch, $1.67; 24-inch, $2.32 per foot. 

668. Sewers at Chicago. In the construction of some 
17,000 feet of sewers for the Chicago Transfer and Clearing 
Yards, 1 concrete was used for all sewers of thirty-six inches 
diameter and over. The excavation was mostly in blue clay 
and done by steam shovel to a depth of twenty feet, the re- 
mainder being removed by hand shovels and swing derrick. 
The material was such that in general the bottom of the trench 
could be trimmed to the form of the exterior of the sewer. 
The thickness of the ring of concrete was 8 inches for 36 and 
42-inch sewers, 10 inches for 48-inch, and 12 inches for 84 and 
90-inch sewers. 

The concrete was composed of one part " Steel Pozzolana" 
(slag) cement, three parts sand and five parts broken stone. 
The cement was of course very finely ground and showed high 
seven-day tests. The cost was $1.30 per barrel delivered. 
The sand was the Chicago " torpedo" sand, coarse and of good 
quality, and cost about ninety cents per cubic yard delivered. 
The stone was a limestone from Summit, 111., crushed in two 
sizes, namely, 1 to 2J inches and ^ to'lj inches. These two 
sizes of stone were mixed in proportions one part of the coarser 
to two of the finer. The cost of stone was about $0.80 per 
cubic yard delivered. 

The concrete was mixed by a rotary mixer of the continuous 
type provided with radial blades. The mixer was mounted on 
a flat car, with engine and upright boiler. Three cars of stone, 
the mixer car, two cars of sand and one of cement made up the 
concrete train, which ran on a track laid close to the trench 
and was kept near the work by a small locomotive. The mixer 
was supplied by wheelbarrows running from the material cars 
on plank runways attached to the cars. The concrete was also 
transported in wheelbarrows from the mixer to the trench. 



1 E. J. McCaustland, Trans. Assoc. C. E., Cornell University, 1902-. 



SEWERS 453 

669. The bottom of the trench being cut to form, the con- 
crete for the invert was laid directly on the sub-grade, tamped in 
layers carried up until the invert occupied about one hundred 
forty degrees of arc. The form of the inner face of the invert 
was maintained by template, grade stakes being set 12^ feet 
apart along the trench. The remainder of the sewer was laid 
on centers resting on the invert. The ribs for this centering 
were made in a complete circle, of three thicknesses of one by 
twelve inch boards nailed together and cut to a true circle. 
Ribs were placed four feet center to center, and covered with 
lagging two inches thick and three inches wide, planed to radial 
joints. The strips of lagging were held in place at each end of 
a section by a j\ by 2 inch iron band passing over all of the 
strips, and turned in at the ends, forming a hook in which rested 
the lower lagging strip, the other strips being supported by this 
one. The lower part of each rib rested on the invert, the upper 
portion being cut to a diameter four inches less (that is, smaller 
by twice the thickness of the lagging). While the trench was 
near enough to the outside of the sewer ring not to measurably 
increase the amount of concrete over and above the desired 
thickness, the trench served as the outside form. Above this 
point, planks were inserted and braced to the sides of the trench. 
From the haunches to the crown the exterior was finished with 
a template. 

When completed, the exterior form planks were removed, 
and a light covering of earth placed on the surface to protect 
it from drying too rapidly. This was especially necessary in 
this case on account of the kind of cement used. The centers 
were removed usually after forty-eight hours, by swinging the 
ribs about the vertical diameter and removing the lagging. As 
soon as the centering was removed, the inner surface was plas- 
ered with a mortar composed of three parts lake sand to one 
part cement. 

670. Cost. The company furnished the materials used in 
the sewer ring and manholes, and delivered it on the work, 
while the contractor furnished all tools and labor to dig the 
trenches, complete sewer and manholes, and do the back filling. 
The contract prices per foot are given by Mr. McCaustland, the 
resident engineer, as follows : 



454 



CEMENT AND CONCRETE 



36-inch sewer in trench averaging 11 feet deep, 3,340 feet, at $2.30. 
42 " " " 14 " 2,660 " 3.00. 

48 " " " 17 " 4,540 " 3.57. 

84 " " " 22 " 1,000 " 5.91, 

90 " " " 24 " 5,400 " 6.68. 

From the data given we have computed the approximate 
quantities of concrete per foot of sewer, and assuming the cost 
of the materials for a cubic yard at $3.00, we obtain the follow- 
ing approximate costs : - 







MATERIALS. 






SIZE 
SEWER. 


DEPTH 
TRENCH. 




CONSTRUCTION 
CONTRACT 
PRICE PER 
FOOT. 


ESTIMATED 
TOTAL COST 
PER FOOT. 


Approximate 
Cubic Yards 


Approximate 
Cost 






Concrete. 


Concrete. 






36 in. 


11 feet. 


.285 


$ .85 


$2.30 


$3.15 


42" 


14 " 


.325 


.97 


3.00 


3.97 


48" 


17 " 


.47 


1.41 


3.57 


4.98 


84" 


22 " 


.93 


2.79 


5.91 


8.70 


90" 


24 " 


.99 


2.97 


6.68 


9.65 



671. Special Molds for Small Sewers. In the construction 
of a thirty inch sewer at Medford, Mass., 1 Mr. William Gavin 
Taylor made use of a very convenient form. The lower 240 
degrees of the sewer was of concrete, the upper 120 degrees 
being of brick. To construct the concrete portion as a mono- 
lith, the forms were constructed in lengths of ten feet, separat- 
ing on a vertical line into two halves. The two halves were 
connected by clamps, and held at the proper distance apart by 
dog irons in the end ribs of each form. After smearing the 
forms as usual, the concrete was deposited and rammed. When 
it had partially set, the dog irons were removed and turn-buckles 
used to slowly pull the two halves together. This method pre- 
vented the green concrete being broken, although the concrete 
extended up on the sides thirty degrees above the horizontal 
diameter. 

672. The centers used for the brick arch were also ingen- 
iously arranged, and since they might have been used for a con- 
crete arch they may be described here. These centers were 
also in ten foot lengths. The ribs, of two inch plank, were 



1 Abstract from Annual Report of City Engineer, Engineering Record, 
NOT. 7, 1903. 



SEWERS 455 



spaced two feet centers, with lagging J inch thick by 1 inch 
wide, with one bevel edge to make a tight upper surface. The 
rear end of each center was supported by wedges securely 
fastened to the outer end of the preceding section, the forward 
end being supported by a screw jack. 

After turning the arch, these centers were removed by the 
aid of a special truck the axles of which were bent at such an 
angle as to make the cast iron wheels fit the concrete invert. 
The axle of a roller was first fastened to the outer rib of the 
center to be removed; the truck was then run back a foot or so 
under the center and the screw jack supporting the forward 
end of the center released. This allowed the forward end to 
drop a short distance, the roller resting on the running board 
of the truck. The latter was then pulled into the sewer far 
enough to let the roller run off the end of the truck and lock 
itself. The truck being then pulled out of the sewer toward 
the finished end, drew the center away from the wedges sup- 
porting the rear end, allowing the form to drop on the truck 
and be wheeled out of the sewer. By this method the centers 
were successfully removed without injuring the concrete. 

673. Cost. From data given, the cost of this sewer 
about sixteen hundred feet in length is approximately as 
follows, labor costing twenty-five cents an hour: 

1.25 cu. yds. excavation and back fill, at $.59 $0.74 

.15 cu. yd. concrete, at $6.70 1.00 

.037 cu. yd. brick masonry, at $12.05 44 

Cost of linear foot, exclusive of manholes, estimated at . . $2.18 
The total cost per linear foot is given as $2.39 

674. New York Sewers. In connection with the construc- 
tion of the New York Rapid Transit Railway, some of the 
sewers were built of concrete. This work was done with ex- 
ceptional care, and on a large scale, and it was found that the 
concrete sewers cost one-third less than similar sewers of brick. 

The method of construction of one section may be described 
as follows : l The forms for the invert of the straight lengths 
of sewer were twelve feet in length, consisting of a strong frame- 
work covered with closely matched lagging, planed smooth and 



Engineering News, March 6, 1902. 



456 CEMENT AND CONCRETE 

greased with machine oil. After the trench was prepared, con- 
crete was placed and rammed until the top of the concrete was 
within about one-half inch of the flow line of the invert. To 
accomplish this, a straight edge was used, bearing on the fin- 
ished invert in the rear and a template secured to the trench 
timbering just ahead of the section under construction. 

The invert centers were then placed, resting on the finished 
invert at the rear and on a solid foundation accurately set to 
grade at the forward end. Mortar composed of equal parts 
Portland cement and sand was then tamped between the invert 
form and the bottom concrete already laid. When the flow 
line had been thus accurately formed, the center was braced 
and vertical planking set to form the outside of the walls. The 
concrete was then rammed in place. 

Joists of two inch by four inch scantling laid along the 
center of the top of each side wall of the invert section, formed, 
when removed, a mortise into which the fresh concrete of the 
arch section was rammed to form a bond. Similar mortises 
were also made in the forward end of each section as built. 
After twenty-four hours or more the forms were removed, and 
a thin cement wash was applied to the interior, sufficient only 
to fill any slight imperfections in the surface. 

The arch centers, similar in construction to the forms for 
the invert, were put in place and plastered with one inch of 
rich Portland mortar. Concrete was then placed sufficient to 
make the arch eight inches thick, the outside of the walls being 
formed by inclined boards braced to the trench, and the top of 
the extrados was formed by hand. 

675. Steel Forms. Two novel types of centering have been 
devised, in which the surface next the concrete is of steel. In 
one of these l the forms are in sections about three feet long. 
Two of the pieces of steel are of a width suitable to reach from 
the bottom of the sewer to just above the spring line of the 
arch, while a third piece forms the arch center. The strips are 
bent at an acute angle at the sides, thus projecting into the 
sewer along an element of the surface where the plates join ; the 
two sides of adjacent plates, which flare away from each other, 
are then connected by a continuous U-shaped clip of steel slipped 



1 Engineering Record, Jan. 9, 1904. 



SUBWAYS AND TUNNELS 457 

on from the end of a three foot section, and the intervening 
space in the clip filled with clay or melted paraffin. The form is 
assembled outside the trench, and after the paraffin is in place, 
the center may be handled. When the sewer is completed, 
the paraffin is melted by a suitable heater, or the clay is washed 
out, and the form may be collapsed and removed. 

676. In the other form l the steel plates are in continuous 
strips about six inches wide and are applied by setting up the 
wooden form on an improvised axis, revolving the form and 
wrapping the steel sheet about it as it is revolved. The wooden 
form is in two parts, upper and lower, firmly connected while 
in use, but the two parts may be made to approach each other 
by driving out the wedges between them. After the winding, 
the center, with its sheet steel jacket, is lowered into the trench. 
When the concrete is completed, the form is collapsed and 
removed, leaving the spiral of steel in place to support the con- 
crete until the latter is well set. The steel is then removed by 
simply pulling on one end. As it comes away from the concrete 
it is wound into a coil, and is then ready to be rewound on the 
wooden form. Both of the above styles have been patented. 

ART. 79. CONCRETE SUBWAYS AND TUNNEL LINING 

677. The advantages of concrete in subway construction 
and in tunnel lining are now well established. In subways 
built in open cut, the side walls and invert are of concrete built 
in place, while the roof is frequently made with I-beams with 
concrete arches turned between them. The I-beams are sup- 
ported directly on the side walls, which are usually made mono- 
lithic with the invert. 

678. Special precautions have to be taken to exclude water 
from a subway, and for this purpose tarred felt and Portland 
cement plaster are employed. 

The specifications for the New York Rapid Transit Subway 2 
were carefully framed to secure a waterproof construction. On 
the sub-grade was placed a layer of concrete, smooth and level 
on top. This was covered by alternate layers of hot asphalt 
and felt, from two to six layers of each being used as deemed 



1 Engineering News, Feb. 18, 1904. 

a Abstracted in Engineering News, Feb. 13, 1903. 



458 CEMENT AND CONCRETE 

necessary for the conditions encountered. The remainder of 
the concrete forming the floor was then laid upon the top layer 
of asphalt. In dry, open soil the felt was not required, and in 
dry rock excavations above water level both the asphalt and 
felt were omitted. Similar provisions were made for water- 
proofing the side walls and roof, resulting in a complete layer 
of asphalt and felt imbedded in concrete about the entire tun- 
nel, the waterproofing being protected both inside and out by 
concrete. 

679. In the construction of the Boston Subway l the por- 
tion built in open cut was made as follows: The work was di- 
vided into sections of convenient length, about twelve feet, 
so that work on a section could be carried on continuously until 
completed. Upon the prepared grade were laid three thick- 
nesses of tarred felt with six-inch lap joints, well pitched be- 
tween the layers, and the top of the upper layer thoroughly 
covered with the pitch. When the latter had hardened, the 
invert was laid over the entire width of the section. 

At each side a back wall six inches thick was built up to a 
convenient height and braced. The forms were then removed 
and the face of this back wall was plastered with rich Portland 
cement mortar. The main side walls were then built up be- 
tween this layer of plaster and the forms defining the interior 
face of the wall. This portion of the subway had an arch 
roof, two feet thick at the crown, which was laid on wooden 
centers. The exterior of the roof was plastered like the side 
walls, and then covered with four inches of concrete to protect 
the plaster from injury. The centers were removed after from 
ten to thirty days; the span of the arch was about twenty- 
three feet. 

680. Tunnel Lining in Firm Earth. In building tunnels in 
earth that is sufficiently firm not to require extensive timber- 
ing, concrete is well adapted for lining. An instance of this is 
furnished by the extensive system of tunnels constructed for 
telephone and telegraph service under the streets of Chicago. 2 
The trunk conduits for this system are about thirteen by four- 



1 Annual Report Boston Transit Commission, 1900; also described in 
Engineering News, April 4, 1901. 

2 'Mr. George W. Jackson, Engineer, Proc. W. Soc. Engrs., 1902; also in 
Engineering News, Feb. 19, 1903. 



SUBWAYS AND TUNNELS 459 

teen feet inside, and the laterals about six by seven feet, all of 
the five center horseshoe form. 

The excavation was in hard clay which stood up well. Shafts 
were located in basements of buildings rented for the purpose, 
and in these basements were placed the compressed air plants, 
material bins, concrete mixers, etc. The large air locks, some 
of which would hold ten small construction cars, were placed 
at the bottoms of the shafts. Work was done in three shifts, 
working eight hours each. The two night shifts could excavate 
about twenty-one feet of lateral tunnel in the sixteen hours, 
and the day shift placed the lining. 

681. The concrete was in general composed of five parts of 
broken stone and screenings, or of mixed gravel and sand, to 
one part Portland cement. For intersections but four parts 
aggregate were used. This should make a very strong concrete. 
The centers for the smaller conduits were made of three-inch 
channels, each rib being in five parts bent to the proper form 
and connected by flange plates bolted to the inside of the chan- 
nels at the ends. These ribs were placed three feet apart, and 
two-inch plank used for lagging. 

The ribs for the trunk sewers were of similar construction, 
but with heavier channels braced with angles. Steel lagging 
was used, made of plates about twelve by thirty-six inches, 
stiffened by 1J inch angles on four edges. There were also 
provided bulkheads or steel end plates of voussoir shape, twelve 
inches along the intrados and twenty inches high, for the pur- 
pose of retaining the end of each section of lining and permit 
thorough tamping. These bulkhead sheets, or "end flights," 
were also stiffened along three edges, and could be attached to 
the webs of the channel ribs by short bolts. 

The concrete was mixed at the shaft head and conveyed to 
the work in cars twenty inches wide and four feet long, running 
on a fourteen-inch gage track. The floor of the tunnel was 
first laid in the excavation, the steel ribs then put in place on 
the floor, and the lagging placed at the bottom and built up 
the sides just ahead of the concrete. When near the crown, 
short pieces of lagging three feet in length covering but two 
ribs were used, and the concrete rammed in from the end of 
these short sections until they were complete, and then another 
row of short pieces placed and the operations repeated. 



460 CEMENT AND CONCRETE 

The concrete floor of laterals was designed to be thirteen 
inches , and the sides and arch ten inches thick, but in all cases 
the entire space between the lagging and the sides of the ex- 
cavation was filled with concrete. 

682. In such work as this only the best materials should be 
used, and, as early strength is desired, the use of Portland 
cement is general in order that the centers may be removed 
within a reasonable period. The ends of the sections into 
which the work is divided should, if possible, be brought up 
square, the bulkhead sheets described above being an ingenious 
and effective method of providing for this. Where it is not 
practicable to finish with a square end over the entire area of 
section, then the work on the sides should be stepped back 
from the bottom toward the crown, each step being bounded 
by planes corresponding to coursing and heading joints in a 
masonry arch. 

683. Tunnel Lining in Soft Ground. For tunnels in soft 
ground requiring the use of a shield, some difficulties in using a 
concrete lining are apparent. The principal one of these lies 
in the fact that the fresh concrete is not capable of taking the 
thrust of the jacks used in forcing the shield ahead. Attempts 
have been made to overcome this difficulty by so constructing 
the centers that the jacks may bear against them instead of 
on the fresh concrete. 

Another difficulty is that in materials requiring almost con- 
tinuous support, the temporary timbering is in the way of the 
centering for the concrete construction; and still another is the 
difficulty of properly tamping the arch at the crown where the 
tail of the shield confines the working space. Concrete blocks 
were tried in the construction of sewers in Melbourne, but 
without entire success. Such blocks were successfully em- 
ployed in the underground road system of Paris, though at- 
tempts to use fresh concrete in shield tunneling for this work 
proved a failure. 

684. East Boston Tunnel. In the construction of the East 
Boston Tunnel Extension of the Boston Subway, however, a 
monolithic concrete lining has been successfully built, the 
tunnel being excavated by shield. 

This tunnel is about twenty by twenty-four feet for double 
track electric line. The arch ring and the walls are thirty- 



SUBWAYS' AND TUNNELS 46 1 

three inches in thickness, while the invert is twenty-four inches. 
Two side drifts, eight feet square, were first driven a certain dis- 
tance and timbered. The bottoms of these drifts were then 
excavated, and the side foundations of concrete were placed in 
lengths of sixteen to twenty feet. When the foundations had 
set, the interior forms for the side walls were placed upon them, 
supporting the caps, the exterior plumb posts removed, and 
the concrete side walls, three feet thick, built up to within 
sixteen inches of the springing line of the arch. This work was 
kept about one hundred feet in advance of the shield. 

The shield, provided with live rollers, rested upon these side 
walls, the rollers running in a flanged plate placed on top of the 
walls. The shield was forced ahead thirty inches at a time, 
and sections of the arch thirty inches in length were turned 
directly behind the shield. 

685. The centers of the arch were of curved, ten inch steel 
channels spaced thirty inches apart, and the lagging, four 
inches thick, was placed from the bottom toward the key as 
the concrete was built up. Each section of arch is keyed with 
concrete pressed through two holes in the rear girder at the 
top of the shield, special rammers being used to tamp the con- 
crete into the space at the crown of the arch, the concrete being 
directed into place by curved sheet-iron troughs. 

In each section of arch sixteen cast iron bars, three and one- 
quarter inches in diameter and thirty inches long, are built 
into the concrete in position to receive the thrust of the shield 
jacks. Wooden bulkheads on the jack plungers serve to con- 
fine the fresh concrete, but the reaction is taken on the cast 
iron bars which, being butted end to end in successive sections 
of the arch, carry the stress back to concrete that is able to 
sustain it. As the shield advanced, the space left over the 
completed arch by the tailpiece of the shield was filled with 
grout under pressure. The centers remained in place thirty 
days. The invert was excavated and laid in ten-foot sections 
about twenty-five feet in the rear of the shield. The concrete 
was mixed at the bottom of the shaft and passed through the 
air lock on cars. The concrete cars ran on a higher level than 
the muck cars, in order not to interfere with the excavation. 

686. Lining Tunnels in Rock. If the rock through which 
a tunnel is driven is seamy and insecure, concrete is in most 



462 CEMENT AND CONCRETE 

cases the cheapest and best lining. The cost of the lining is, 
of course, less if it can be built in connection with the excava- 
tion, but it is frequently difficult to foresee how a given rock 
will stand exposure to the air and water, and it becomes an 
exceedingly nice question to determine at the time of building 
a tunnel whether lining is required. In many cases this ques- 
tion is settled in the affirmative by other considerations than 
the character of the rock, as the resistance to flow, in water- 
works and sewers, or the ease of ventilation and the necessity 
of a good appearance, as in street or steam railway tunnels. 

687. New York Subway. In the construction of portions 
of the rapid transit subways of New York, a traveling center 
which served also to support a working platform was carried 
on six wheels running on a track laid on the footing courses of 
the side walls. This center carried at the side, sections of lag- 
ging curved to the required form of the side walls. This lagging 
was adjusted in place, and braced from the platform or center 
by means of wedges. Directly behind this traveling center was 
a similar platform carrying a derrick; and behind this, the 
traveling center carrying the lagging for the roof. This third 
platform was jacked up to place the roof lagging at the correct 
elevation, and firmly supported by wedges. 

The concrete was brought in skips on cars that ran on the 
floor level and stopped beneath the derrick platform. The 
derrick hoisted the skips through a hole in the platform and 
placed them on cars on either the side wall or the roof platform, 
so that the concrete was delivered either to the side wall forms 
in advance, or the roof forms in the rear as required. The 
concrete was rammed in a direction transverse to the tunnel 
axis until the roof was completed, except for a space about 
five feet wide at the crown. The arch was then keyed by 
tamping the concrete in from the end of the form. The two 
platforms carrying the forms were each forty feet long, and 
the derrick platform was eighteen feet. 

688. The excavated rock was crushed for the concrete on a 
working platform erected over and around the shaft head. 
Cars delivered the excavated material at the shaft in steel skips, 
which were hoisted to the working platform, set on push cars 
and dumped into bins, from which stone was delivered to the 
crusher; these cars then passed under the crushed stone bins, 



SUBWAYS AND TUNNELS 463 

were loaded with broken stone, run back to the shaft head, 
and the broken stone dumped into bins mounted over the 
mixer. The skips were then lowered into the shafts by the 
derricks, to be run to the headings and reloaded. The stone 
and sand were fed to a measuring box by means of a hopper, 
the measuring box discharging directly into a cubical mixer, 
which was high enough above the tunnel floor to dump directly 
into skips on the cars. 

689. Cascade Tunnel. In the construction of the Cascade 
Tunnel of the Great Northern Railway a somewhat different 
arrangement was used. * The working platform in the tunnel 
was erected five hundred feet in length, and cars hauled by 
cable up an incline to the platform. The side walls were built 
in alternating sections, eight to ten feet in length, the support 
of the arch timbering being thus gradually transferred from 
the plumb posts to the concrete of the side walls. Arch sections 
were built in twelve foot lengths, the centers being made of 
four by sixteen inch plank without radials, so as to leave a 
clear way for concrete cars on the working platform. The 
latter were high enough to allow the material cars to run be- 
neath them. 

690. Concrete vs. Brick. There are frequent instances in 
engineering construction where brick masonry might well have 
been replaced by concrete, and the use of brick for tunnel lining 
is still adhered to in many cases. This is partly because some- 
what less elaborate centers can be used for brick arches, and 
the centers may be struck somewhat earlier, and partly be- 
cause of extreme conservatism on the part of the designer, 
although without doubt there are cases where the use of brick 
is entirely warranted. 

An interesting instance of the greater adaptability of con- 
crete under unforeseen conditions, however, is presented by the 
Third Street concrete and brick lined tunnel at Los Angeles, 
Cal. 3 This tunnel was excavated mostly through an argilla- 
ceous sandstone. The side walls were of concrete up to the 
haunches, the upper part of the arch being of six courses of 
brick. A streak of yellow clay was encountered, and it "was 

1 Mr. John F. Stevens, M. Am. Soc. C. E., Engineering News, Jan. 10, 
1901. 

2 J. H. Quinton, M. Am. Soc. C. E., Engineering News, July 18, 1901. 



464 CEMENT AND CONCRETE 

soon demonstrated that the six ring brick arch, which occupies 
the central portion of the roof, was not strong enough to hold 
up the immense weight above it, and the temporary timbering 
was crushed and broken in a most alarming way." The strength 
of the arch was increased by using nine rows of brick instead 
of six until the clay seam was passed. In such portions of the 
six ring arch as had cracked, it was found that the inner ring 
of brickwork had separated from the second ring, and in places 
the second ring had separated from the third. The concrete 
walls had shown no evidence of weakness. 

To repair the brickwork, steel concrete beams or arches were 
inserted in the brickwork at intervals of four feet, and extend- 
ing from one concrete wall to the other. These beams were 
twelve inches wide and eight to twelve inches deep, made of 
rich concrete, and had imbedded in each beam two pieces of 
three inch by three-quarter inch steel. The steel ribs were set 
in recesses cut out of the brickwork, and rested at the ends upon 
the concrete of the side walls. Substantial centers were used 
for building the concrete beams, and when the latter had set, 
the defective brickwork between adjacent beams was cut out 
and replaced by rich concrete. 

691. Aspen Tunnel. Another illustration of the adaptabil- 
ity of concrete when unexpected difficulties arise, is furnished 
by the construction of the Aspen Tunnel on the Union Pacific 
Railroad. * The original design provided for sets of timbers to 
support the excavation, spaced about three feet, center to center, 
but for nine hundred feet of the tunnel such pressures were 
encountered that in places a solid wall of twelve by twelve inch 
timber was forced in. For a portion of this section the lining was 
builfr of a combination of concrete with steel ribs. The latter 
were 12-inch, 55-pound I-beams spaced from twelve to twenty- 
four inches, center to center, curved to conform to the interior 
of the tunnel. The concrete was built up around and between 
the beams, the inner flange being covered by from four to seven 
inches, and the total thickness of the walls two to three feet. 

692. The Perkasie Tunnel of the Philadelphia and Reading 
Railroad was constructed through a firm rock, which, however, 
was intersected by several strata of seamy rock. As trouble 



1 W. P. Hardesty, Engineering News, March 6, 1902. 



SUBWAYS AND TUNNELS 465 

was experienced from rock falling from these strata, it was 
decided to line the tunnel at such places. This lining had a 
minimum thickness of eighteen inches at the crown and twelve 
inches at the sides. Traffic through the tunnel was not ob- 
structed during the work of placing the lining. In laying 
about five hundred cubic yards of concrete, the cost was about 
ten dollars and eighty cents per cubic yard, exclusive of cost 
of centering and dry filling. 1 

693. Water Works Tunnel. The lining of portions of the 
Beacon Street Tunnel of the Sudbury River Aqueduct was 
undertaken some fourteen years after its excavation, and at a 
time when it was necessary to use the tunnel intermittently to 
supply water to the city of Boston. The methods employed 
are described by Mr. Desmond FitzGerald in Transactions 
American Soc. C. E. for March, 1894. 

A substantial track of 2 feet 1J inch gage was laid from a 
manhole furnishing access to the sewer to the portion of the 
tunnel to be lined. The rails, weighing thirty-six pounds to 
the yard, were supported on small but substantial trestles, 
built of three by four inch spruce joists, and placed eight feet 
between centers. Every third trestle was braced from the 
sides and roof of the tunnel to prevent the track being floated 
when the tunnel was in use. The trestles also carried five rows 
of planks for the workmen to walk on in pushing the cars. 
The track was elevated by these trestles, so the work was not 
seriously interfered with by a small amount of water in the 
tunnel. The track cost about eighty-seven cents a foot. 

Cars to run on these tracks to deliver materials and concrete 
had frames five feet by one foot nine inches, with twenty inch 
wheels, and cost about fifty-six dollars each. 

694. Centers. The centers were in three parts, two for 
side walls and one for roof. The ribs were of three thicknesses 
of two by ten spruce plank, without interior bracing for the 
roof section. The side sections had each an inclined brace. 
Wedges were inserted between the tops of the side sections 
and the bottoms of the roof ribs to hold the latter in place. 
The lagging was two by four inch spruce, in eight foot lengths, 
with beveled edges and planed both sides. The centers were 



1 P. D. Ford, M. Am. Soc. C. E., Trans. A. S. C. E., March, 1894. 



466 CEMENT AND CONCRETE 

spaced four feet apart, and seventy-five full centers were built; 
these, with the lagging, contained 14,000 feet B. M. of lumber, 
and cost $1,460.55, or $104.30 per thousand feet B. M. 

695. Methods of Work. Broken stone, sand and cement 
were stored in shanties over and around the manhole leading 
to the tunnel, and arrangements made by which the materials 
could be delivered through chutes down the manhole to the 
cars. As it was found more convenient to work in winter, 
special provision was made for storing large quantities of ma- 
terial in the shanties. The sand was piled around an iron lined, 
wooden bulkhead, in the center of which was a large stove. 

The concrete was mixed within the tunnel as close to the 
work as possible, and in places where the cross-section had 
been sufficiently enlarged by falls of rock to permit easy work- 
ing. The materials, delivered to the material cars down the 
chutes already mentioned, were pushed to the mixing platforms 
and combined in the proportions of 18.56 cubic feet of crushed 
stone and 7.35 cubic feet of sand to one barrel of Portland 
cement, being approximately 1 to 2 to 5J. The above quanti- 
ties of materials made 20 to 21 cubic feet of concrete. When 
mixed, the concrete was shoveled into cars, conveyed to the 
work and then shoveled into place. 

The tamping was done principally with oak rammer five 
inches square, twelve inches long, with a short wooden handle 
in one end. In tamping the key of the arch, long-handled iron 
rammers were used. Much care was requisite here to prevent 
the aggregate separating from the mortar and lodging next 
the lagging, as it always has a tendency to do, thus resulting 
in voids in the face of the work when the lagging is removed. 
The concrete was built up on the sides in horizontal layers and 
stepped back by inserting bulkheads, so that the adjacent 
sections bonded together. 

696. Cost. The cost of this concrete lining, which was 
built under great disadvantages, amounted to $16.15 per cubic 
yard. This cost must be considered reasonable in view of the 
fact that the materials had to be transported an average dis- 
tance of more than one-half mile on small push cars, and the 
work in the tunnel was suspended for three days of each week 
to allow the tunnel to be used to maintain the water supply of 
the city. 



RESERVOIRS 467 

ART. 80. RESERVOIRS: LININGS AND ROOFS 

697. Although the choice of the material with which to 
construct a reservoir may in some cases be varied by local 
conditions, it is found that under ordinary circumstances con- 
crete offers the greatest advantages for a minimum cost. For 
the side walls of small reservoirs, concrete furnishes the requi- 
site strength and water- tightness with a moderate thickness; 
earthen embankments and floors may be made practically im- 
pervious with concrete and mortar, combined with asphalt 
when considered necessary; while for the roofs, groined arches 
or beams and slab construction, with supporting piers, all of 
this material, make a neat, permanent, and altogether satis- 
factory covering, at a smaller expense than would be required 
for brick or stone masonry. 

698. Details of Construction. In the walls and floors, 
water-tightness is a prime consideration, and this is best at- 
tained by a layer of mortar on the inner surfaces or between 
two layers of concrete. 

As in floors, walks, etc., the necessity of providing for ex- 
pansion and contraction will depend upon the extremes of 
temperature to which the surface is to be subjected. In covered 
reservoirs which are to be almost constantly filled with water, 
or in very equable climates, the blocks may be large, say twenty 
feet square, while under more severe conditions the blocks 
may not contain more than twenty square feet. The joints 
between the blocks may well be wide enough to be filled with 
asphalt. This furnishes an elastic joint which is compressed 
as the blocks expand, and swells when the blocks again con- 
tract. 

699. Reservoir Floors. One of the principal difficulties ex- 
perienced in the construction of floors is from settlement of the 
foundation. The floor should, therefore, have strength enough 
to bridge any small irregularities in the foundation that may 
result from inequalities in settlement. For a similar reason, it 
is not well to make the blocks too large, as smaller blocks with 
compressible joints will more readily conform to an uneven 
surface without permanent injury. In order that the reser- 
voir shall not leak even if the foundation settles, the concrete 
and mortar may be covered with one or more layers of asphalt. 



468 CEMENT AND CONCRETE 

in building the floor lining, alternate blocks are sometimes 
placed first in molds and the intermediate blocks built in later. 
In other cases the blocks are laid consecutively. The advan- 
tage of the former method seems to lie principally in the ease 
of construction, as access may be had to all sides of the 
block. 

700. In hard clay soil not liable to settlement, four inches 
is sufficient thickness for the floor, the concrete to be covered 
before it has set with a half-inch layer of rich Portland mortar, 
troweled to a smooth surface. If the reservoir when empty 
will be subjected to hydrostatic pressure from without, the 
floor must be designed to resist this pressure. In this case, if 
seepage from without into the reservoir is objectionable, a layer 
of mortar may be placed over the first layer of concrete and 
protected by the concrete laid upon it. This outside pressure 
may be provided for in a covered reservoir by making the floor 
of inverted arches between piers, the weight of the floor, piers, 
roof, and earth filling over the roof, being made sufficient to 
balance the upward pressure on the floor. If there is no ob- 
jection to the water from without being led into the reservoir, 
a porous layer of broken stone or gravel beneath the floor may 
be connected with the interior of the reservoir through pipes 
provided with check valves, and the outside pressure be thus 
removed. Where it can be accomplished, it will usually be 
better to lead this ground water through a pipe to a sewer or 
a lower level rather than into the reservoir. 

701. Walls. The thickness of the wall is determined by 
methods similar to those used in designing a retaining wall or 
a dam according as the pressures are greater from the embank- 
ment without or the water pressure within. In the case of a 
covered reservoir, the thrust of the roof arches may convert 
any vertical section of the wall into a beam, the earth pressure 
from without being supported by the floor at the bottom and 
the roof at the top. Or in case there is no back pressure from 
earth filling, the thrust of the roof may be added to the inner 
water pressure. In. circular covered reservoirs the arch thrust 
is usually taken by steel bands laid in the concrete and en- 
circling the reservoir near the top of the wall. In narrow 
reservoirs rectangular in plan, tie rods may be used, or the 
wall may be buttressed to take the roof thrust. Concrete side 



RESERVOIRS 469 

walls are usually built vertical, or nearly so, on the inside, and 
with a batter on the outside. 

702. Linings. Linings of sloping earthen embankments are 
laid the same as the floors, and similar precautions are required. 
There is greater danger of settlement of embankments than of 
the floor foundation, and the blocks, therefore, may well be made 
smaller. Some difficulty may be experienced with laying hori- 
zontal asphalt joints on a sloping face, and some sliding of the 
lining may be expected under ordinary conditions, the asphalt 
joints being compressed. For this reason it would seem to be 
better to use asphalt in the inclined joints only, and a mastic 
in the horizontal joints. Another method which would probably 
prove satisfactory is to lay first a tier of blocks next the floor, 
and when these have set, apply a very thin coat of asphalt to 
the upper edges of these blocks, following with another tier, 
and so on. 

703. ROOFS. Where it is necessary to cover, a reservoir, 
either to prevent the formation of ice, or the growth of algae, 
or for other reasons, the groined arch is an excellent design for 
the roof on account of the small amount of concrete required, 
the clear head room given, and the ease of ventilation. The 
extending use of reinforced concrete will also probably enter 
this field to a greater extent in the future than it has here- 
tofore. 

The determination of the stresses in a groined arch roof is 
complicated not only by the peculiar form of the arch itself, 
but by the fact that the spandrels of the arches are filled with 
concrete over the piers to the level of the extrados at the crown. 
This evidently results in making of any given unit of the roof, 
having a pier as its center, a cantilever, and the arch action is 
interfered with. Unless, however, tension members of steel are 
laid in the concrete near its upper surface, it is not wise to count 
on the strength of the cantilever except to consider it a factor 
of ignorance on the safe side. If one wishes to depart from 
the ordinary and tried dimensions for groined arches in concrete, 
such departure had better be based on some special experi- 
ments and tests on full sized sections. Some of the dimensions 
that have been used are given in the examples cited below. 

704. Forms. The preparation of forms or centers for 
groined arches is one of the most difficult and expensive details 



470 CEMENT AND CONCRETE 

of the construction of such a roof. It will probably be best to 
have each section of the form cover the space, square in plan, 
between four piers. The ribs of the centering may well be 
built up of planks, nailed together and sawed to proper form. 
The lagging should be planed to size, and have radial joints to 
make a smooth and even top surface. Care is necessary to make 
a neat fit along the valley extending diagonally between piers, 
and a small fillet may well be fitted into this valley to avoid 
a sharp corner on the finished concrete, as well as to cover up 
possible imperfections in the joints. The forms should, of 
course, be designed to take the thrust of the adjacent com- 
pleted arches, and if sufficient forms are not built to cover the 
entire reservoir, and thus transmit the thrust to the walls, the 
piers at the border of the forms must be thoroughly braced to 
the opposite side walls or the piers will be toppled over and 
the roof wrecked. This accident occurred to one reservoir 
roof during .construction, the pier braces having been removed 
without the knowledge of the engineer. 

705. In laying the concrete, joints between the work done 
on consecutive days should cut the arches at right angles to 
their axes, and bulkheads should be used to make such a joint 
a vertical plane. The covering of each unit between four piers 
is made monolithic, and care is necessary to prevent the stones 
working to the bottom of the mass and thus becoming exposed 
when the forms are removed. This may be prevented by plas- 
tering the forms with mortar and placing the concrete upon it 
before the mortar has begun to set. 

706. A roof consisting of a network of concrete-steel beams 
intersecting at right angles, supported by piers and covered by 
concrete-steel slabs, makes a very simple design. The forms 
are much easier to construct, and forms for only a limited area 
need be erected at one time. An excellent article on " Covered 
Reservoirs and Their Design," by Mr. Freeman C. Coffin, M. 
Am. Soc. C. E., is contained in the July, 1899, number of the 
Jour, of the Assn. of Engr. Soc. An article on the " Groined 
Arch," by Mr. Leonard Metcalf, Assoc. M. Am. Soc. C. E., ap- 
pears in Trans. A. S. C. E. for June, 1900; and Mr. Frank 
L. Fuller presents an article on " Covered Reservoirs/ 7 in Jour. 
Assn. Engr. Soc. for Sept., 1899. 

707. Examples of Concrete Reservoirs. Wellesley. The 



/ RESERVOIRS 471 

reservoir at Wellesley, Mass./ a part of the water supply sys- 
tem, was designed by Mr. Freeman C. Coffin. It is eighty-two 
feet in diameter, walls fifteen feet high, four feet thick at bottom 
and two feet at top. The walls are of concrete and rubble 
masonry. In the construction of the walls, concrete was used 
containing three parts sand and five parts of stone to one 
of cement, one cubic yard of concrete containing about 1.2 
barrels of cement. The bottom of the walls, which were de- 
signed to be built of concrete three feet four inches thick, were 
actually built of rubble four feet thick, as a large quantity of 
bowlders was at hand. The excavation was in hard clay con- 
taining but little water, and the floor was made only four inches 
thick, of concrete of the same quality as that used in the 
walls. 

The floor and side walls were plastered with two coats, the 
first, one-half inch thick, of mortar containing two parts sand 
to one of Portland cement, and a coat about one-eighth inch 
thick, of neat Portland carefully rubbed and smoothed with 
trowels. Such a plaster coat should be applied before the con- 
crete has set. The two plaster coats cost twenty cents per 
square yard. 

708. The piers to support the groined arch roof were two 
feet square, and built of brick. The span of the arches was 
12 feet, rise 2.5 feet, and the concrete 0.5 foot thick at the 
crown. A channel iron ring or band was set in the concrete 
walls at the springing of the roof arches to take the thrust of 
the latter. The centers were placed over one-fourth of the 
area at a time, the piers being braced to take the thrust of the 
arches until the roof was completed. The concrete in the roof 
was composed of two and one-half parts sand and four and one- 
half parts broken stone to one part Portland cement. The 
centering cost twenty-two and one-half cents per square foot of 
area covered. The spandrels were filled in level with top of 
concrete at crown. On top of the concrete roof was placed six 
inches of clean gravel for drainage and to prevent the earth 
freezing to the concrete. This gravel was drained by four 
inch vitrified pipe discharging at the toe of the slope walU 



1 Engineering News, Sept. 30, 1897; Jour. Assn. Engr. Societies, July 
1899; Trans. A. S. C. E., June, 1900. 



472 CEMENT AND CONCRETE 

One foot of earth filling and one foot of loam were placed upon 
the gravel. 

709. Astoria. The reservoir for the Astoria City Water 
Works 1 was designed and built by Mr. Arthur L. Adams, M. 
Am. Soc. C. E. The reservoir has a capacity of six and one- 
fourth million gallons, walls twenty feet high. The excavation 
was in hard clay and sand mixed with clay, which in some places 
resembled a soft sandstone. The embankment was in general 
about five feet, the remainder of the depth being in excavation. 

The floor consisted of six inches of concrete, f inch cement 
mortar, one coat liquid asphalt and one coat harder asphalt. 
The slope lining was of six inches concrete, one coat asphalt, 
one layer of brick dipped in asphalt and laid flat, and a final 
finishing coat of asphalt. The concrete was composed of one 
barrel Portland cement, one-tenth cubic yard sand, five-tenths 
cubic yard gravel and nine-tenths cubic yard of crushed stone, 
these quantities of the ingredients making one cubic yard of 
concrete. Here we have an instance of the use of a mixture 
of broken stone and gravel, a practice which has already been 
commended as resulting in a small amount of voids. 

The concrete of the floor was laid in blocks twenty feet on a 
side, molds of two by six inch plank forming the outside edges 
of a block, and serving as a guide to the straight edge used in 
finishing, as in concrete walk construction. The finishing coat 
was of two parts fine sand to one of Portland cement and was 
applied, before the concrete base had begun to set, by two 
finishers with smoothing trowels. When the next block was to 
be laid, the plank were replaced by one-half inch weather 
boarding. When the concrete had thoroughly set, these boards 
were removed and the joints so formed were run full of asphalt, 
when the first layer of this material was spread. 

The concrete on the sides was also six inches thick and laid 
in sheets ten ' feet wide, extending up and down the slopes, 
expansion joints being provided on the inclined joints only. 
The finishing coat of mortar was not used here, but all inequali- 
ties in surface were smoothed by using a little mortar from the 
next batch of concrete. 

710. Each concrete gang was composed of twenty men and 



Trans. A. S. C. E., December, 1896. 



RESERVOIRS 473 

one water boy. All concrete was mixed by hand on movable 
platform in half-yard batches. On the entire work 1.84 cubic 
yards of concrete per day were mixed and placed per man 
employed, and on the floor alone this quantity was increased 
to 2.35 cubic yards, an excellent showing for this class of work. 
The cost of concrete per cubic yard, without profit, was as 
follows: 

On Slopes: Cement, at $2.45 per bbl $2.82 

Other materials 1.94 

Labor 1.07 

Total per cubic yard for 600 yards $5.83 

On Floor: Cement, at $2.45 $2.64 

Other materials 1.92 

Labor .68 



Total cost per cubic yard for 680 yards $5.24 

The costs of the slope lining and floor complete, per square foot, 
are given as follows : - 

Slope: 6 inches concrete $0.1187 

.649 inch asphalt 0100 

Brick in asphalt 0889 

.851 inch asphalt 0131 

Chinking crevices 0030 

Ironing 0036 

Total cost per square foot of slope $0.2373 

Bottom: 6 niches concrete $0.1031 

Cement mortar finish .0113 

.537 inch coat asphalt 0077 

.573 inch coat asphalt 0082 

Total cost of bottom per square foot $0.1303 

711. Forbes Hill. The Forbes Hill reservoir * forms a part 
of the distribution system of the Metropolitan Water Works of 
Boston and was built under the direction of Mr. Dexter Brack- 
ett, M. Am. Soc. C. E. The reservoir is two hundred eighty by 
one hundred feet, -partly in embankment. The soil under the 
embankment was first stripped to a depth of two and one-half 



1 Described by Mr. C. M. Saville, M. Am. Soc. C. E., Division Engineer, 
before the N. E. Water Works Assn. Abstracted in Engineering News, March 
13, 1902. 



474 CEMENT AND CONCRETE 

feet at the toe, increasing to five feet stripping at the inner edge 
of the slope. The material was hard pan, and the embank- 
ments were built in four inch layers, rolled with four thousand 
pound rollers, so made as to leave a slightly corrugated surface. 
The bank was extended one foot inside of the finished line to 
assure a compact face, and afterward trimmed to grade. 

712. The bottom of the reservoir was covered first with a 
layer of concrete about four and one-half inches thick, com- 
posed of one part natural cement, two parts sand, and five 
parts stone. The sand was of good quality; the stone came from 
the excavation and was washed before crushing. This layer of 
natural cement concrete was covered by a layer of Portland 
cement mortar one-half inch thick, made of two parts sand to 
one cement, and finished with a richer mortar, one part sand to 
four of cement. 

This half-inch layer was laid in strips four feet wide and 
finished like a cement sidewalk. Although this mortar coat 
was kept well moistened, some cracks developed which were 
filled with grout before applying the second layer of concrete. 
If no joints were used in the lower layer or base concrete, and 
joints in the coat of mortar were provided in one direction only, 
as appears to have been the case, the cracking should have been 
anticipated. At any rate, the value of the mortar coat be- 
tween the two concrete layers was greatly impaired by this 
cracking, and the experience points to the advisability of plac- 
ing the upper layer of concrete on the mortar before the latter 
has set, thus avoiding the expense of finishing the mortar layer. 

The upper layer of concrete was composed of one part Port- 
land cement, two and one-half parts sand and four parts broken 
stone, and was laid in blocks ten feet square. These blocks 
were laid alternately each way. 

. The slope was first lined with Portland concrete of 1 to 2J 
to 6i, then one-half inch layer of mortar as for the bottom. 
The upper layer of concrete on slope was same as the upper 
layer of the bottom lining, but the blocks were eight by ten 
feet and finished with one inch of granolithic, in which stone 
dust and particles smaller than three-eighths inch were sub- 
stituted for the one and one-half inch stone of the concrete. 
... ,. 713. Cost. The cost of lower layer of concrete on bottom, 
natural 1 to 2 to 5, was as follows: 



/ RESERVOIRS 475 

1.25 bbl. natural cement, at $1.08 $1.350 

.34 cu. yd. sand, at $1.02 347 

.86 cu. yd. stone, at $1.57 1.350 

Materials in concrete $3.047 

Forms, lumber, at $20.00 per M , . . $0.090 

Forms, labor 0.100 



Total forms ' .190 

General expenses $0.08 

Mixing and placing 1.17 



1.250 
Total cost per cu. yd $4.487 

Cost of lower layer on slopes, Portland 1 to 2J to 6J, was as 
follows: 

1.08 bbls. Portland cement, at $1.53 $1.652 

.37 cu. yd. sand, at $1.02 377 

.96 cu. yd. stone, at $1.57 1.507 

Materials in concrete $3.536 

Forms, lumber, at $20.00 per M $0.016 

Forms, labor . 0.121 



Total forms .137 

General expenses $0.177 

Mixing and placing 1.213 

1.390 

Total cost per cubic yard $5.063 

The cost of the upper layer on bottom and slopes, including 
the finish on slopes, Portland 1 to 2J to 4, was as follows: 

1.37 bbls. Portland cement, at $1.53 $2.09 

.47 cu. yd. sand, at $1.02 .48 

.745 cu. yd. stone, at $1.57 1.17 

Materials in concrete $3.74 

Forms, lumber, at $20.00 per M $0.25 

Forms, labor 0.26 



Total forms .51 

General expenses $0.15 

Mixing and placing 1.53 

1.68 

Total cost per cu. yd $5.93 



476 CEMENT AND CONCRETE 

The cost of the half-inch plaster coat between the layers of 
concrete was twenty cents per square yard. 

714. Rock ford. A reservoir for the city of Rockford, 111., 1 
was built almost entirely of concrete after plans prepared by 
the City Engineer, Mr. Chas. C. Stowell. The soil was a loose 
gravel, and after excavation was completed, parallel lines of 
drain tile were laid in trenches nine to ten feet centers and 
leading to a fifteen inch vertical sewer pipe carried to the sur- 
face of the street and capped. This sewer pipe served as a 
sump for a pump should it be found necessary at any time to 
repair the bottom. These trenches were filled with broken stone 
and the whole area of the foundation covered with three inches 
of clay. The concrete bottom was in two layers, eight inches 
and seven inches thick, respectively, and composed of two 
parts sand and five parts stone to one of Portland cement. 

The walls were of similar concrete for the bottom twelve feet, 
natural cement being used in the upper eight feet of the walls. 
The thickness at the bottom was 4^ feet, walls being straight on 
outside with one to ten batter on inside. The entire inner 
surface of floor and walls was plastered with one-half inch of 
Portland mortar, one to one. The cost of concrete in the work 
averaged $6.50 for Portland and $4.00 for natural, and the 
finishing coat cost seventy-five cents a square yard. 

715. The roof was of concrete, expanded metal lath, and 
steel rods, the finished thickness being but two inches. This 
was supported by ribs of concrete, each twelve inches thick at 
the crown and having a seven-inch channel on the under side. 
The springing line of the ribs was eight feet below the top of 
the walls, giving a good depth at the skew back. Ribs were 
spaced about seven feet centers. The span of the roof was 
about fifty-five feet, and rise about eleven feet. The cost of 
roof was less than twenty-five cents a square foot. 

716. Concord. The groined arch roof of the Concord, 
Mass., 1 sewage storage well, designed by Mr. Leonard Metcalf, 
Assoc. M. Am. Soc. C. E., was fifty-seven feet diameter and 
contained about one hundred cubic yards of masonry. The 
cost of the roof per square foot of surface was as follows: 



Described in Engineering News, Feb. 22, 1894. 



RESERVOIRS 477 

Centering ...... $0.18 

Concrete materials ............ .15 

Labor and supervision 05 



Total $0.38 

717. Albany. The Albany filter plant roof, 1 designed by 
Mr. Allen Hazen, Assoc. M. Am. Soc. C. E., was also of the 
groined arch type, the arches having the same span and rise as 
the Wellesley reservoir. The cost of the roof per square foot 
of area was as follows : - 

.029 cu. yd. concrete, at $6.30 $0.182 

Piers 054 

Earth filling and seeding 014 

Manholes, entrances, etc 027 



Total cost per sq. ft $0.277 

For a list of reservoirs and filter beds, in the roofs of which 
groined arches have been used, giving in tabular form the 
general dimensions, the proportions used in the concrete, and 
in several cases the cost of the roof per square foot of reservoir, 
the reader is referred to Engineering News of December 24, 
1903. 



1 Trans. A. S. C. E., June, 1900. 



CHAPTER XXII 

SPECIAL USES OF CONCRETE (CONTINUED) 
BRIDGES, DAMS, LOCKS, AND BREAKWATERS 

ART. 81. BRIDGE PIERS AND ABUTMENTS AND RETAINING 

WALLS 

718. The use of concrete in large bridge piers was at first 
confined to the hearting or backing of stone masonry shells. 
It was soon found, however, that in many cases the concrete 
was able to withstand the effects of frost and ice better than 
was the variety of stone available for building the masonry 
shell, and many important bridges are now supported by piers 
built entirely of concrete. 

As an example of this use may be mentioned the bridge 
across the Red River l in Louisiana, which has six concrete 
piers of heights from forty-four to fifty-three feet. The pivot 
pier is twenty-seven feet in diameter, with vertical sides. The 
draw rest piers are seven feet wide under the coping, nineteen 
feet between shoulders and twenty-six feet long over all. The 
sides have a batter of one-half inch to the foot. The coping is 
of limestone. 

719. In the construction of the Arkansas River Bridge 2 of 
the K. C. P. & G. R. R., ten piers and two abutments were 
built of concrete. The piers varied in height from twenty to 
sixty-five feet, some of them containing over six hundred cubic 
yards of concrete. The entire work was completed in eleven 
months, although many difficulties were met. The concrete 
was composed of one part Portland cement, two and one-half 
parts coarse, sharp sand, and five parts of clean, broken stone. 

The lagging for the forms was of two-inch yellow pine, sur- 
faced one side and sized to one and three-quarters inches. On 



1 George H. Pegram, Consulting Engineer. Walter H. Gahagan, Engi- 
neer for Contractors. 

2 Engineering News, Aug. 25, 1898. 

478 



BRIDGE PIERS 479 

the straight part of the pier this lagging was laid horizontal and 
supported by four by six vertical posts set four-feet centers, 
posts on opposite sides of the pier being tied together by three- 
quarter inch bolts passing through one and one-half inch gas 
pipes spaced five feet vertically. The gas pipe was allowed to 
remain in the finished pier, the bolts being withdrawn. 

The lagging for the semicircular ends was of two by six with 
bevel joints, placed vertical, and supported at five-foot inter- 
vals by segmental ribs of double two by twelve planks. At 
the ends of the ribs were bolted short pieces of eight by eight 
inch angle irons with edge horizontal. These angle irons were, 
in turn, bolted to four by six verticals at the corners or shoul- 
ders of the pier. 

720. The foundation piers of the Lonesome -Valley Viaduct, 1 
thirty-six piers and two abutments, are entirely of concrete. 
The piers are four feet square on top with batter of one inch 
to the foot, and are from five to sixteen feet in height. The 
total concrete laid was 926 cubic yards at a contract price of 
about $7.00 per cubic yard. The piers were finished on top with 
a steel plate, four feet square and one-half inch thick, taking 
the place of coping stones. Where rock foundations were not 
found, the lower portion of the pier was given an increased 
batter to secure such a cross-sectional area at the bottom that 
the unit pressure on the earth did not exceed one ton per square 
foot. The cost of the concrete for this work has already been 
given ( 322). 

721. Steel Cylinders. Steel shells filled with concrete have 
been used to good advantage, especially for bridge approaches. 
Such shells are usually in pairs placed abreast, one under each 
truss of the bridge or viaduct. The two cylinders of a pair 
are usually connected by lateral bracing, and if desired in heavy 
work, this bracing may be inclosed in a concrete wall and thus 
protected from injury by running ice, etc. The thickness of 
metal in the shells need not be great, three-eighths of an inch 
usually being sufficient, though this depends on the height, the 
stresses, and the liability to injury. In soft ground requiring 
piles, most of the piles are sawn off below the limit of scour, or 
below the water line for land piers, but one or more may be 



1 Gustave R. Tuska, Trans. A. S. C. E., September, 1895. 



480 CEMENT AND CONCRETE 

allowed to project up into the cylinders and the concrete filled 
in around the heads, thus anchoring the pier. In foundations 
on rock if the cylinders require an anchorage, this may be pro- 
vided with bolts fox-wedged or cemented in the rock and pro- 
jecting into the cylinder. (For details of methods adopted in 
this class of construction, see "Bridge Substructure and Foun- 
dations in Nova Scotia/' by Martin Murphy, Trans. A. S. C. E., 
September, 1893.) 

722. Repair of Stone Piers. Where masonry piers are being 
destroyed by the abrading or expansive action of ice, or by 
other causes, concrete is successfully used to arrest such action, 
the entire pier being incased in a layer, one to three feet 
thick, of Portland cement concrete of good quality. 

The piers of the Avon River bridge, 1 originally built of ashlar 
masonry, failed entirely to withstand the deteriorating in- 
fluences of an extreme range in tide coupled with the severe 
temperature of a Nova Scotia winter. Five of them were sub- 
sequently incased in concrete, as follows: A form was made of 
ten by ten inch spruce timber surrounding the ashlar masonry 
of the piers and forming a mold to receive the concrete and 
retain it in place until set. The thickness of the concrete casing 
was two and one-half feet at the bottom and one and one- 
third feet at the top, which was three feet above high water. 
The concrete was composed of one barrel Portland cement, 
one and one-half barrels clean sand, one barrel of clean gravel, 
and in it was placed by hand four parts of common field stone 
weighing from eight to twenty pounds each. This treatment 
was entirely successful in preventing further disintegration. 

723. An efficient cutwater for bridge piers is made by placing 
old rails vertically on the upstream nose of the pier, anchoring 
them to the masonry and filling between with concrete, leaving 
only the wearing surface of the rail head exposed. 

724. Pneumatic caissons are usually filled with concrete, the 
filling over the working chamber being carried up fast enough 
to keep the work above water as the caisson is sunk. The 
filling of the working chamber calls for special care in tamping 
under and around the longitudinal and cross-timber braces. A 
space of about three inches next the roof of the chamber is 



1 Trans. A. S. C. E., December, 1893. 



RETAINING WALLS AND ABUTMENTS 481 

filled with a rich concrete, containing no stone larger than one 
inch, mixed quite dry and solidly tamped with special edge 
rammers. 

725. RETAINING WALLS AND ABUTMENTS. Concrete is used 
very largely for constructing retaining walls and bridge abut- 
ments. The foundations of a retaining wall should be of ample 
width, and if the wall is not founded on rock, some settlement 
and outward movement may be expected if the common for- 
mulas are used in computing the dimensions. 

If this movement is not equal throughout the wall, cracking 
is likely to take place, and to confine these cracks to prede- 
termined vertical planes, it is well to construct the wall, if a 
long one, with vertical joints at intervals of fifteen to thirty 
feet. Such a joint is made by building one section and fol- 
lowing with another, without special precautions to make a bond 
between. 

If there is an opportunity for water to accumulate, care 
should be taken to drain the earth back of the wall, either by 
drains leading around the ends, or by pipes passing through 
the wall. The latter may result in discoloration of the face. 

726. Coping. The face of a retaining wall or abutment is 
preferably given a batter, and a coping is provided to improve 
the appearance. The coping should have a slight inclination 
toward the back to prevent the discoloration of the face by 
dripping. It should be divided by vertical joints into blocks, 
not more than six to eight feet in length. The projection of 
the coping will depend upon the dimensions of the wall. Wing 
walls are preferably built with a sloping top or coping, but this 
should be made monolithic with the wall by special molds 
(729). 

727. Rules for Use of Concrete in Abutments. In the use 
of concrete for abutments and piers, the practice of the Illinois 
Central Railroad, as set forth in their specifications, can hardly 
be improved upon. The engineer of bridges and buildings on 
this road, Mr. H. W. Parkhurst, M. Am. Soc. C. E., is one of 
those engineers who early recognized the value of concrete in 
bridge work, and as the result of his extensive experience along 
this line, he is widely known as an able and conservative au- 
thority. 

These specifications are printed nearly in full in Engineering 



482 CEMENT AND CONCRETE 

News of July 18, 1901, from which the following extracts are 
made: 

728. Natural and Portland Cement : Where used : 

Natural cement concrete "may be used where foundations 
are entirely submerged below low-water mark, or where there is 
no risk of the same being exposed to the action of the weather 
by cutting away the surrounding earth. It, however, shall be 
used only where a firm and uniform foundation is found to 
exist after excavations are completed. In all cases where 
foundations are liable to be exposed to the action of the water, 
or where the material in the bottom of excavations is soft or 
of unequal firmness, Portland cement concrete must be em- 
ployed for foundation work. 

"The natural cement concrete shall usually be made in the 
proportions (by measure) of one part of approved cement to 
two parts of sand and five parts of crushed stone, all of char- 
acter as above specified. For Portland cement concrete foun- 
dations, one part of approved cement, three parts of sand and 
six parts of crushed stone may be used. Wherever in the 
judgment of the engineer or inspector in charge of the work, a 
stronger concrete is required than is above specified, the pro- 
portions of sand and crushed stone employed may be reduced, 
a natural cement concrete of 1,2 and 4, and a Portland cement 
concrete of 1, 2 and 5 being substituted for those above speci- 
fied. 

"Portland Cement Concrete: Concrete for the bodies of 
piers and abutments, for all wing- walls for same, and for the 
bench walls of arch culverts, shall generally be made in the pro- 
portions (by measure) of one part of cement, two and one-half 
parts of sand and six parts of crushed stone. Where special 
strength may be required for any of this work, concrete in the 
proportions of 1, 2 and 5 may be used; but all such cases shall 
be submitted to the judgment of the engineer of bridges, before 
any change from the usual specification is to be allowed. 

"For arch rings of arch culverts and for parapet head walls 
and copings to same, Portland cement concrete, in proportions 
of 1, 2 and 5, shall generally be used. Concrete of these pro- 
portions shall also generally be used for parapet walls behind 
bridge seats of piers or abutments, and for the finished copings 
(if used) on wing- walls of concrete abutments, also for. arch 



USE OF CONCRETE IN ABUTMENTS 483 

work in combination with I-beams or in combination with iron- 
work for transverse loading. 

" Bridge seats of piers and abutments and copings of con- 
crete masonry which are to carry pedestals for g