432 H. S. S'ltinnters : 



of magnesia and iron cause the toti^l of the salic minerals in this 

 case to be moi-e than seven times the i'emic minerals. 



The Uralla analysis corresponds closely to the other analyses, 

 and would fall i'nto the group Almerose, but for the almost 

 entire absence of soda. It will be seen that the total alkalis are 

 about normal, and it is to be hoped that more analyses of 

 buttons from this area ma}^ be made to determine if this result 

 be in accord with others from the same locality. 



The analyses of the specimen from Pieman wanders farthest 

 from the t3q3e group. The large excess of lime over the alkalis 

 throws this analysis into Rang IV., otherwise it is fairly closely 

 related to Riesenose. 



The division line between Class I. and Class 11. is necessarily 

 an arbitrary one, so that with an increase in the number of 

 analyses of obsidianites we should probably get a perfect grada- 

 tion from Riesnose to Almerose. Even as it is we see that some 

 of the analyses belonging to the Almerose Sub-rang seem to be 

 more nearly related to the analysis of the Hamilton specimen 

 than to other analyses falling into the same group as them- 

 selves. 



By far the most imi)ortant result of the classification of the 

 analyses is, however, to demonstrate clearly that rocks having 

 compositions similar to those of the obsidianites are rarely met 

 with among the igneous rocks of the earth's surface. The im- 

 portance of this result becomes very evident when we come to , 

 consider the origin of the obsidianites. 



The Origin of Obsidianite.^s. 



Several hypotheses have been advanced to account for the 

 origin and distribution of obsidianites, and naturally their 

 chemical composition has an important bearing on the solution 

 of the problem. It has been stated: — 



(a) They are artificial products. 



(b) They represent ii peculiar form of volcanic ejectmenta. 



(c) They are meteoritic in origin. 



(a) Artificial Products. — 



At first sight the general apjjcarance of obsidianites may 

 incline one to the ])elief that thcv arc artificial in oricfin. l)ut a 



