210 REIBLING AND REYES. 



atmosphere. (3) Impurities in the water. (4) The fmeness of the 

 cement. (5) The retarders added to the cement. (6) The time and 

 manner of curing and storing it. (7) The quantity and the condition 

 of the free lime. (8) The temperature of the freshly molded cement 

 paste. 



The first three factors can be so carefully regulated as to render 

 their influence practically constant. The influence of the others is less 

 definite. They are interdependent and generally speaking but little 

 understood. 



The peculiar effect of g3^psmn or plaster of Paris upon the setting 

 properties of Portland cement has received much attention and many 

 theories have been proposed to explain its action. Of these we mention 

 only two. 



Dr. Rohland ascribes it to catalysis, considering calcium sulphate to be a 

 catalytic retarder. Candlot concluded that gypsum added to Portland cement 

 produces a double salt of basic calcium aluminate and sulphate. He believed 

 that he succeeded in producing this salt artificially. 



It would seem as if the action of gypsum was capable of a simpler 

 explanation than those already advanced. The salt is soluble in water 

 to the extent of 1 part of calcium sulphate to 490 parts of water at 15°. 

 In dissolving it is ionized and obviously the addition of calcium ions 

 Avould reduce the solubility of the other calcium ions already present. 

 The free lime would therefore be less rapidly dissolved and the hydrol- 

 ysis of the aluminates would be retarded, the result being a slowing 

 of the set. As the free lime changes to calcium carbonate it is removed 

 from solution and becomes inert. The loss of these calciimn ions of 

 course increases the solubility of the aluminates and the facility with 

 which they hydrolyze, the result being a quickening of the set. 



However, whatever the nature of this action, Candlot's experiments 

 show very conclusively that the presence of slaked lime increases the 

 efficiency of sulphates of calcium as retarders. His conclusions in this 

 respect were verified in the mechanical laboratory of Sibley College, 

 Cornell University. Experiments in the chemical laboratory of the 

 Bureau of Science have resulted in similar observations. 



The sixth factor mentioned above, namely the time and method of 

 curing, is dependent upon changes brought about by air and moisture. 

 According to Candlot the change ia the setting properties which some- 

 times occurs when cements are stored is due to the formation of calcium 

 carbonate. 



However, we do not believe that the change of the slaked lime in 

 Portland cement to calcium, carbonate is the usual or universal cause of 

 changes in setting properties. Such changes have been known to take 

 place within a few days in cements stored in paper-lined barrels or in 



