A. W. Jackson — Nomenclature of Crystalline Rocks. 123 



source are liable only about the peripheries of the rock-groups, 

 to take up Eozenbusch's figure once more ; and the a 

 is to place the rock in a closely allied group where the mistake, 

 if it be one, can do the minimum of harm. 



I have shown thus far that un 7 would re- 



sult from a purely mineralogical nomenclature ; let us now 

 examine the adaptibility of such names. I have already called 

 attention to the variation within certain limits in the relative 

 proportion of the essential constituent minerals of any rock. 

 It is not worth while to express this variation in our nomen- 

 clature. In fact it is so irregular, so completely beyond the 

 control of any known or conceivable general law that it would 

 be altogether impossible to express it if we would. Names 

 founded upon mineral composition would ignore this variation 

 so long as it took place within the limits assigned for essential 

 constituents. Such names would therefore completely adapt 

 themselves to the natural conditions. They would have more- 

 over even a higher order of adaptability than this. I have 

 quoted Eosenbusch on the subject of rock-groups wherein 

 he asserts the substantia] unity of rock-types but of rock-types 

 that are themselves united by transitions due to the gradual re- 

 placement of one or more constituents by other minerals, or 

 by a gradual change of texture. Such names can be made to 

 adapt themselves most perfectly to the expression of the rela- 

 tions between the different members of a rock-group and be- 

 tween each and the central type. To this limited extent the 

 attempt could safely be made to be systematic in our nomencla- 

 ture ; for the relations between the closely allied rocks of each 

 group both chemically and geologically as well as mineralog- 

 - too evident ever to become questioned. 

 To illustrate with the granite group, Eosenbusch (Massige 

 Oesteine, p. 18) following and developing the suggestion of 

 Gustav Eose has divided' ami named the group in accordance 

 with these principles. Quartz, oi gioelase are 



present in even granite; associated with these are muscovite, 

 biotite and hornblende, sometimes one, sometimes two. When 

 muscovite alone is present we have muscovite granite ; when 

 biotite alone, biotite-granite (or granitite) : when hornblende 

 alone, hornblende granite, when muscovite and biotite together, 

 granite (in strict sense); when biotite and hornblende, biotite 

 hornblende granite, or, more conveniently, hornblende grani- 

 tite. Muscovite and hornblende do not occur simultaneously 

 with the type constituents, but if they did it would be a sim- 

 ple matter to make a name for the rock. 



The basalt group furnishes another good instance. From 

 basalt as a central type (plagioclase, augite, olivine) we pass to 

 nephelite basalt (-plagioclase + nephelite), or to leucite basalt 



