374 HENRY S. WASHINGTON 



in those of the leucitic rocks. The ciminites and vulsinites 

 reach no higher than 6o per cent, of silica, and lie in general 

 between 55 and 58 per cent., while the toscanites are no lower 

 than 64 per cent., and are in general much higher. Correspond- 

 ing differences may also be noted in the other constituents, 

 especially FcgOg, FeO, MgO and CaO. It is also seen that 

 transitional chemical forms are of limited occurrence, though 

 the anal37ses are not as numerous as in the leucitic group. The 

 analyses, then, of both groups are clustered about certain points 

 and do not form a gradual series from the most basic to the 

 most acid, as Brogger understands a series to be constituted. A 

 similar clustering of the analyses may be seen at such volcanic 

 localities as the Yellowstone Park, Montana, Cape Verde Islands, 

 and yEgina and Methana ; and the study of complete series of 

 analyses would probably reveal the same state of affairs else- 

 where. Indeed the fact may be said to be generally known, 

 though perhaps not definitely formulated. An explanation 

 which may be suggested is that this clustering of analyses is 

 due to a quite complete course of differentiation, assuming the 

 correctness of the differentiation hypothesis. 



Each main group of rocks, therefore, may be subdivided 

 chemically, and mineralogically to a certain extent, into two 

 subgroups (classing the ciminites and vulsinites together), a 

 basic and an acid. Further, while the leucitic subgroup is 

 more basic in each case than the corresponding trachydoleritic 

 one, yet the variations in each are similar. These are best seen 

 on tabulating the means of the best of the various preceding 

 analyses. It is scarcely worth while to do so in this place, but the 

 main features may be briefly pointed out. There is a difference 

 in each case of 8 to 10 per cent, of silica; and iron, magnesia, 

 lime, and to a less extent alumina, are higher in the basic group, 

 while alkalies show less regularity. The molecular ratio of soda 

 to potash varies in my separate analyses from .3 to 1.2, being 

 fairly constant (.5 to .6) in the leucitic groups, and with higher 

 soda at the extremes of the trachydolerites. It is probable that 

 with more numerous analyses the resemblances would be even 



