176 Rev'mvs — A New Rock- Classification. 



the same as before. In classes iv and v eacli subclass is divided 

 into five orders, according to the ratio of pyroxenes and olivine to 

 iron-ores. By further particularisation of the constituent minerals 

 the authors establish successively suborders, rangs, subrangs, grads, 

 and subgrads. The reason given for the spellings ' rang ' and 

 * grad ' is that ' rank ' and ' grade ' are so frequently employed in 

 a general sense, but this consideration will not console French and 

 German readers. For our part, we are loth to see the words 

 ' class ' and ' order ' tied up to special meanings, and should have 

 preferred to borrow such terms as ' domain ' and ' nome ' from the 

 vocabulary of the older systematists. 



This leads us to notice that the new classification is expressed 

 throughout in terms of a new nomenclature. Given the one, we 

 must indeed grant the necessity for the other. As the authors justly 

 remark, petrology has suffered sufficiently from the redefining of 

 terms, a license which would not be patiently conceded in any other 

 branch of science. It is, moreover, a part of their design that each 

 term shall carry its own precise meaning on its face. To this 

 end they have boldly thrown aside the Greek lexicon, and con- 

 structed new words, each of which is a frank barbarism, but conveys 

 an explicit meaning on a certain mnemonical plan. Thus the two 

 groups of minerals, which we have distinguished above as the light 

 and the dark, are designated by what the late Lewis Carroll called 

 ' portmanteau- words ' ; salic, recalling s-ilica and aZ-umina, and/em«c, 

 suggesting ferro-magnesian. The five classes of rocks are named 

 persalic, dosalic, salfemic, dofemic, and perfemic. A like method is 

 pursued through all the minor divisions. In spite of its obvious 

 utility, it has at first a decidedly irritating effect, and will not 

 improbably prejudice the reception of the scheme. These uncouth 

 classificatory terms, however, need not be often used in speaking 

 of the rocks themselves, for each subdivision inferior to subclass 

 is provided with a proper name derived from some country or 

 locality where the rocks are typically exhibited. We have thus, for 

 a granodiorite from the Yosemite, the order Britannare, rang 

 Toscanase, subrang Toscanose. Different terminations are used to 

 mark divisions of diffei-ent magnitude or status, a feature which, we 

 think, must be commended. The termination -ite is at present 

 greatly overworked, and a novice might plausibly suppose Iherzolite 

 to be something of equal importance with granite. 



The only serious criticism that we have to make upon the new 

 classification, regarded as a purely empirical scheme, remains to be 

 noticed. The percentage mineralogical composition which is the 

 basis of classification is an ideal one, which may or may not agree 

 with the actual composition, and in some cases departs widely from 

 it. Thus the leucitite of Capo di Bove is assumed, for the purpose 

 of fixing its systematic position, to contain about 11 per cent, of 

 anorthite and 9 per cent, each of olivine and akermanite, although 

 none of these minerals is actually present in the rock. Some of the 

 actual constituents, melilite and biotite, are discarded, because their 

 complex formulas make them unsuitable to be included in the list of 



