H. 8. Washington — Rhyolites of Lipari. 455 



The typical Sant' Angelo rock is thus decidedly lower 

 in silica than the rhyolites, somewhat lower in alkalies 

 (though their relative proportions are about the same), 

 with but slightly higher iron oxides, (in which the domi- 

 nance of ferrous over ferric oxide is to be noted), and 

 higher lime and alumina. From the two silica determina- 

 tions reported by Bergeat, it is clear that the Sant' 

 Angelo lavas are quite variable, and that further study 

 of them is called for. 



Classification. — The norm of the Perrera rock is as 

 follows : 



Quartz 13.20 



Orthoclase 21.68 



Albite 26.20 



Anorthite 28.63 



Hypersthene 5.77 



Magnetite 1.39 



Ilmenite 1.37 



Apatite 0.67 



This norm places the rock in amiatose, with the symbol 

 1.4(5). 3.3., showing thus that there is a decided excess of 

 silica, even though no quartz is seen in the rock itself. 



In the current classifications these lavas have been 

 called andesite by Bergeat and by Sabatini, and this 

 would appear to be justifiable from consideration of the 

 mode, the rock being composed essentially of andesine, 

 pyroxene, magnetite, and glass. The analysis and the 

 norm, however, reveal the fact that, had the magma 

 solidified as a noncrystalline rock, a very notable amount 

 of quartz would have been present, in which case the rock 

 would have to be classed systematically as a dacite. The 

 normative quartz is " occult ' in the glass base, to use 

 Iddings' term. 14 



The choice of a suitable name for such rocks, those in 

 which there is considerable excess silica occult in the glass 

 base, has been a mooted question for many years, espe- 

 cially since the chemical composition of igneous rocks 

 has become a matter of some importance. 



The name dacite was applied to them originally, and 

 still denotes, generally, a quartz-bearing andesite, one in 

 which the quartz is present modally. For the modally 

 quartz-free "andesites" which, however, contain so 

 much silica that notable amounts of quartz would crystal- 

 lize were they noncrystalline, Lacroix 15 has recently pro- 



14 J. P. Iddings, Igneous Kocks, Vol. II, p. 19, 1913. 



15 Lacroix, C. E., 168, 298, 1919. 



