64 MINERALOGY AND LITHOLOGY, 



ence of alumina favors the formation of hornblende, although the differ- 

 ence in these two analyses in other respects is very wide, — the large 

 percentage of lime in the pyroxene being marked ; and it is known that 

 pyroxene much more uniformly possesses a larger percentage of lime 

 than does hornblende. The presence of alkali in the hornblende is also 

 noticeable. Hence we see that when these minerals are associated, 

 analysis shows them to be different chemical compounds. 



The occurrence of associated pyroxene and hornblende in the cavities 

 of the lava at Vesuvius has been described by Vom Rath.* He has 

 shown that these minerals have been deposited in these cavities by a pro- 

 cess of sublimation, and hence, both were formed under the same condi- 

 tions. His analyses were necessarily imperfect, since the very small 

 amount of material that he possessed did not allow of the determination 

 of all the ingredients ; yet the same distinctions between his analyses 

 of pyroxene and hornblende are prominent, — a larger percentage of 

 alumina, a smaller of lime, and the presence of alkali in the hornblende. 



Moreover, some of our igneous rocks contain pyroxene, and others 

 hornblende, and some both. Now, since these rocks form well defined 

 dykes, and possess those characters which make it perfectly evident that 

 they reached the surface in a molten condition, it might be inferred that 

 the minerals in them were formed under essentially the same conditions. 

 At Dixville Notch the traps are in part diorites, which contain horn- 

 blende in such large and well formed crystals that the separation of pure 

 material for analysis is easy. Now this compared with my analysis of 

 pyroxene, picked from the triassic trap of the Connecticut valley.f gives 

 us the results that follow : 



* PoSS- Annalen Erganzung, Bd. vi, p. 229. 



t American yournal of Science, iii, vol. ix, p. 187. 



