PORTLAND CEMENT TESTING. 155 



by the wooden molds. Therefore, a moist-air closet formed of several 

 small, insulated compartments, each with its own pan of water, is best 

 adapted for the purpose. Such a moist-air closet has been constructed 

 for this laboratory. 



Standard cement specifications should include a definite form of moist- 

 air closet with a complete description of the materials for its construction, 

 its dimensions, and directions for its use; otherwise one source of the 

 so-called personal error will persist. 



TIME OF SETTIiSTG. 



The American Society has adopted the Vicat needle method for de- 

 terminations of tlie time of setting. It seems to be the general impres- 

 sion that the Gillmore method does not insure the desired accuracy, and 

 many cement testers will regret that such a convenient and time-saving 

 process has been supplanted hj a more cumbersome one; still the Gillmore 

 method, if properly regulated, can be made accurate, reliable, and im- 

 partial and at the same time retain its simplicity, even though the meager 

 directions in the United States Army specifications do not insure uniform 

 results between different operators and at times imposes unjust tests 

 upon some good cements. 



Merz, Meyer, Schiffner, Bohme and many others have each pointed 

 out that to determine the time of setting of a cement it should be gauged 

 with a quantity of water proper to it.^' It has often been demonstrated in 

 this laboratory that 20 per cent of water is not enough to meet the re- 

 quirements of the fineness, specific gravity, chemical composition, and 

 physical properties of many good Portland cements sufficiently to produce 

 a paste plastic enough to be molded into a pat. The resulting paste is 

 often so diy and non-cohesive that it will not stick together or to the 

 glass j)late; and yet 1 to 3 per cent of water in addition will produce 

 the desired plasticity and cohesiveness. 



The whole phenomenon of the manufacture of artificial stone from 

 finely powdered cement is one of solution, hydration and subsequent 

 crystallization. The addition of sufficient water is essential for proper 

 solution and hydration. The addition of too much water is to be avoided 

 because- of its effect upon the subsequent crystallization, and because 

 the density of the paste must allow of proper manipulation. 'Therefore, 

 it is very evident that plasticity and not a given percentage of water 

 should be the condition regulating the paste for cement pats. 



The insistence of the United States Army engineers ujion a paste 

 gauged with 20 per cent of water seems to be a striking illustration of 

 Spalding's assertion that ^^ "tests may be imposed which in nearly all 



"Soc. Chem. Industry (1801), 10, 928. 

 "76irf., 87. 



