312 G.W. Tyrrell— Tertiary Dykes 



quartz and falls into the subrang lassenose (I', 4 (2). 3. 4, yellow- 

 stonose-lassenose). Furthermore, if the average hypersthene-andesite, 

 based on seventy-one analyses, be compared with the average hyper- 

 sthene-andesite computed by Daly on twenty analyses, it will be found 

 that the differences are quite insignificant. This testifies to the 

 distinctness and solidarity of this petrographic type, as distinguished 

 by many petrographers on the basis of mineral composition. There 

 can be little doubt but that a similar result would be arrived at for 

 many other rock-types. 



Relationships and Nomenclature of the Cumbrae Ttpe. 



It is very difficult to place this rock under the existing system of 

 nomenclature. Similar rocks from the North of England and else- 

 where have been called " augite-andesite ", "basaltic andesite", 

 " andesitic basalt ", " andesitic dolerite ", "tholeiite", "basalt", 

 etc., a list which sufficiently indicates the petrographers' perplexity. 

 The rock can be regarded neither as a true andesite nor a true 

 basalt. In chemical composition it approaches the andesites; but if 

 the term andesite be limited (as it should be if used specifically) to 

 volcanic rocks with predominant andesine felspar, the Cumbrae rock 

 clearly does not fall within this group. Its mineralogical expression, 

 on the other hand, approaches that of a basalt ; yet it cannot be 

 placed thereunder because its chemical composition, with 60 per cent 

 silica, is entirely unlike that of basalts. A similar difficulty of 

 nomenclature has been experienced with the related dykes of Mull, 

 and the Survey petrographers have solved the problem by giving the 

 rocks new locality names, such as leidleite and inninmorite. The 

 root of the difficulty is, of course, the abundant glass in the rock, 

 which, as has been shown, must contain the excess silica and alkalies 

 above that necessary to form the visible felspars. The probable 

 mineral composition had the rock been holocrystalline is given 

 approximately by the norm. There would have been about 47 per 

 cent of a felspar near Ab 3 An 2 (andesine), about 21 per cent pyroxenes, 

 12 per cent orthoclase, and 14 per cent quartz, i.e. the rock would 

 have been a pyroxene-dacite ; or, if the quartz had remained occult, 

 as it appears to do in most andesites, a pyroxene-andesite. 



Terms such as dolerite, basalt, andesite, dacite, and the like, have 

 become far too comprehensive for a more discriminative petrography 

 than frequently employed. If they are to be retained as more than 

 mere field names, or comprehensive "omnibus" names, they should 

 be frankly used as group or generic terms, comprising numerous 

 sub-groups and species which should be given new names. Lacroix 

 has recently begun the discrimination and renaming of ultrabasic 

 forms of basalt and nepheline-basalt, 1 whilst Iddings has attempted to 

 distinguish the numerous varieties of dacite, andesite, and basalt, 

 according to their predominant felspar. 2 



If, however, nomenclature is based on mineral composition, there 

 will always be a terminological difficulty with glassy rocks, for the 

 greater the amount of glass the more abnormal will be the constitution 



1 A. Lacroix, Comptes Rendus, vol. clxiii, pp. 177-83, 213-19, 253-8,1916. 



2 J. P. Iddings, Igneous Rocks, vol. ii, pp. 106, 191, 1913. 



