160 RKIBI.ING AND SALINGER. 



tory has adopted tlie standard portrayed and described by Taylor and 

 Thompson -" in order to insure a complete comprehension in this respect. 



Much diversity of opinion exists regarding the rejection oE a cement 

 wliicli fails to meet the boiling test,"^ but we regard such a cement as 

 dangerous if it is to be used in woi'ks exposed to the lieat of the tropical 

 sum. 



Excess of lime, coarseness of grinding, insufficient seasoning, and 

 underburning of a cement may cause it to fail to p)ass the soundness test. 

 If lime is the cause, storage may eliminate the defect, as the free lime 

 would thus be changed to the carbonate, or slaked, and so would not cause 

 subsequent expansion. 



Many engineers believe that failure to pass the hot test is not a proof 

 of inferiority, as the cement so failing, if mixed with sand or some other 

 aggregate, has produced durable masonry; it is also a known fact that 

 thoroughly slaked lime paste can be added to a Portland cement mortar 

 without injurious results. We suggest that, as is the case in determining 

 the time of setting, some sj)ecification be devised to test the mortar 

 mixture as well as the neat paste. 



TENSILE STRENGTH. 



The variation in the breaking streng-tli of both neat and sand briquettes 

 is a source of trouble to every cement tester, and despite every efEort to 

 eliminate this error, breaks continue to be variable with a persistence that 

 makes it necessary to double or treble the number of briquettes otherwise 

 required. We have made a thorough study of this variation and as a 

 result have come to the conclusion that only a portion of it is due to 

 the personal error of the operator, and that the remainder is caused liy 

 the characteristics of the cement itself. 



Personal error even with the most careful manipulation, maj^ be pro- 

 duced by (1) unavoidable variation in troweling; (3) diflEerence in the 

 force of tlie blows; (3) lack of equality in forming each layer of the 

 briquette; (4) variation in the size and shape of the mold; (5) difference 

 in the size and shape of the sand particles; (6) personal error in macMne 

 operation; (7) unavoidable internal strains and voids caused by the 

 manipulation which the specifications impiose (8) the impossibility of 

 seciiring a perfectly homogeneous mixture; (9) variation in dr3'ing. 



The errors caused by the natural characteristics of the cement, and 

 which need more extended explanation, are as follows : 



1. It is obvious that it is impossible to expose the same number of 

 cement particles to the action of the air for the same length of time in 



=» Concrete, Tlain and Reinforced. New York (1907), 103-107. 

 "^Loc. cit. 



