460 PROCEEDINGS OF THE AMERICAN ACADEMY 



10.55 per cent of its weight. The mean of these values is 10.66; and 

 regarding this as the water of crystallization of the mineral, and reduc- 

 ing Mr. Munroe's analyses accordingly, we obtain the following 

 results : — 



2.48 : 2.45 : 1.25 



It would appear, then, that Ilallite at 300° is in the same condition 

 of hydration which the other vermiculites examined assume at or about 

 100°. Now, corresponding to this, there is a very marked fact, indicated 

 by the tables given above, which is worthy of special notice. Air-dried 

 Jefferisite loses, at 100°, about ten per cent of its weight, while air- 

 dried Hallite loses only about three per cent, showing tliat it hohls its 

 water much more firmly than the first. In order to institute a just 

 comparison between the different vermiculites, it is obviously impor- 

 tant to seek for each variety the point at which the mineral assumes a 

 constant condition, and maintains it through a considerable variation 

 of temperature. Save only some practical couvenience, there is no 

 peculiar virtue in 100°, as the temperature at which a mineral should 

 be dried for analysis. As in the case of crystalline salts we should ex- 

 pect to find for e;ich hydrous mineral a certain point or points of tem- 

 perature at which it loses the whole, or a part, of its water of 

 crystallization, and certain limits between whi;;h it maintains a con- 

 stant composition. Moreover we slwuld expect that these tempera- 

 tures would be the more definite in proportion as what we may call 

 the hygroscopic power is the more marked; that, while in some cases 

 the miueral would lose its water at a nearly constant temperature, and 

 the intervals of definite hydration would be well marked, in others the 

 loss would extend over a considerable range of temperature, and it 

 would be more dilficult to secure the states of definite composition. 

 That such differences as these are conspicuous among the vermiculites 

 the tables given above abundantly show ; but, in addition to this evi- 

 dence, the difference in the behavior of the several varieties, wlien heated, 

 impressed upon us more strongly the principle we have stated than the 

 figures would indicate. Nevertheless, as the following table shows, we 

 have been able to bring all the vermiculites to essentially the same 



