462 PROCEEDINGS OF THE AMERICAN ACADEMY. 



, In the last two varieties (he ratios correspond to the symbols : — - 



VIII. VIII. VIIL 



Si^ O3 (R, R) . mfi 

 do. . 2H.,0 



do. . H^O. 



Here again it will be noticed that the differences in the definiteness 

 of these ratios correspond to the differences of hygroscopic power de- 

 scribed above. In the case of Hallite, the ratios are almost precisely 

 those of even molecules, while in the case of Jefferisite the a<rreement 

 is much less close. 



In conclusion, we consider that the follo\^ing general results may be 

 deduced from this investigation : — 



First. That all the vermiculites are unisilicates. 

 • Secondly. That these minerals combine with water in several de- 

 finite proportions, thus confirming the opinion advanced by the writer 

 in his previous paper on this subject, that the water in the vermiculites 

 is water of crystallization. 



Thirdly. That all these minerals may be reduced to the condition 

 expressed by the ratio 2:2:1, which we regard as the normal ratio of 

 the vermiculites. 



Fourthly. That the only essential difference between the different 

 varieties of vermiculites is in the ratio between the sesquioxide and 

 protoxide bases. 



