P ACINI, METAMORPHISM OF PORTLAND CEMENT 167 



4. To the decomposition of basic products present in the freshly set 

 cement, 



5. To equilibrium of calcium hydroxide with the siliceous constitu- 

 ents, and 



6. To the hydration of the double silicates and anhydrides of lime and 

 alumina. 



The two theories that have at present the greatest claim upon con- 

 sideration are that the strength of set cement is due to the progressive 

 crystallization of calcium hydroxide (80), and, in some respects diamet- 

 rically opposed, that this strength is due to the formation of a dense 

 complex colloid, soft at first but gradually adsorbing calcium hydroxide 

 and thus becoming harder and harder (64, 65). 



According to the latter theory, cement consists of a mixture of fused 

 compounds of silicic, aluminic and ferric acids with lime, together with 

 an excess of lime, partly dissolved and partly enclosed. Upon the addi- 

 tion of water to this system it is decomposed, and the water becomes a 

 supersaturated solution of salts, which react between themselves. The 

 compounds resulting from these reactions crystallize about the cement 

 grains in needle-shaped crystals. So far, the process is analogous to the 

 setting of plaster of Paris (45), and silica takes no part in these pre- 

 liminary reactions. 



A hydrogel begins to form about each grain, in which the crystals 

 become embedded. This hydrogel consists essentially of calcium hydro- 

 silicate, and to a minor degree of calcium hydroaluminate and calcium 

 hydroferrite. At first it is soft and plastic, but gradually becomes dense 

 and rigid by the adsorption of calcium hydroxide. The strength of 

 cement is mainly due to this process of coagulation. 



The calcium hydroxide may of course crystallize and lend additional 

 strength; but its crystallization is rather more likely to burst the har- 

 dened cell walls about each grain of cement, and thus admit liquids 

 later in the process which may be fatal to the integrity of the structure, 

 either by undesirable chemical reactions, or simply by dissolving away 

 the lime, with the formation of soft hydrates of silica, alumina and iron 

 oxide, instead of the desired hardened colloid (64, 65). 



Much corroborative evidence has been offered by supporters of this 

 view, and similarly by the exponents of the crystallization theory in 

 defense of that. The question is still at issue, and the main difficulty is 

 the microscopic recognition of the constituents of set cement (34, 78). 

 Unquestionably, colloidal materials result from the action of water on 

 silicates of this type, when the particles have been ground to the fineness 

 of Portland cement (23, 21, 95). This has been directly observed in the 



