PORTLAND CEMENT TESTING. 179 



Tlie samples were well mixed and screened before testing; troweling was done 

 as uniformly as possible for exactly five minutes; the atmospheric and moist- 

 closet exposure was the same in all cases, except that there was a gradual change 

 in temperature from 27° at 8 a. m. to 29° at 2 p. m. 



After troweling sample F5-1 with 21 per cent of water for aboiit four minutes, 

 it suddenly became hot and dry, crumbling apart. No amount of patting would 

 cause the cement to stick together sufficiently to form a- pat. 



The same result v/as observed on repeating the operation, and a thermometer 

 placed in the mass rose 6° in four minutes. However, upon adding 22 per cent of 

 water to the same cement no rise in temperature Avas observed during troweling; 

 the resulting paste was sufficiently plastic to be easily molded into a jjat; and 

 the needle used for the initial set when first applied, sank about one-eighth of 

 an inch. However, five minutes after the pat was placed in the moist-air closet, 

 it began to Iieat and to drj' slightly, the initial set taking place in fifteen minutes. 

 This experiment was repeated with similar results. Twenty-three per cent of 

 water was then used. The plastic paste, when formed into a pat, acted normally 

 in every way and gave a satisfactory setting time. 



The results obtained with sample F5-3 and F5-5 were practically identical. 

 In the case of the latter, the excess percentage of water was reduced by 0.5 per 

 cent to determine if possible the minimum quantity necessary to effect so 

 profound a change. 



Two important facts become evident from the above data, namely, that 

 both tlie plasticity and setting time of a cement, such as was being tested, 

 are much affected after a certain quantity of water has been added by 

 the subsequent addition of even very small amounts of the solvent.^' 



We are not prepared to discuss fully these results at the present time, 

 but their analogy to the phenomenon of the crystallization of certain 

 salts from solution is striking. Many salts have a critical solution 

 factor. Under slow evaporation they will remain in solution until a 

 certain limiting percentage of the solvent has been reached, when the 

 salt will crystallize almost instantly, heat being generated during the 

 separation. A cement, the setting j)roperties of which are so profoundly 

 affected by the addition of even small quantities of Avater, may l^e said 

 to have a critical solution (or hydration) point. We would hesitate to 

 decide wdiether such a cement deserves to be approved. If tested accord- 

 ing to the United States Army specifications it would fail to pass the 

 setting test, but under those of the American Society the normal plasticity 

 method will give it sufficient water to cause it to set slowly. 



An engineer in these Islands related an experience illustrating the 

 practical importance of this problem. The mortar, after mixing, was 

 dumped into a car and transported to its destination by rail in five 

 minutes. Working with a large shipment of this cement no difficulty 

 was experienced for some time, but finallj' when one carload readied its 



" This same phenomenon is less delicately sliown in fig. 19 of Taylor and 

 Thompson "Concrete, Plain and Reinforced." It will be noticed that Portland 

 cement C (without gypsurp) reached its final set even in less than thirty 

 minutes with 20 per cent of water. With 25 per cent of water the initial set 

 took one hour and thirty minutes and the final set five hours. 



