H. Stanley Jevons — nomenclature of Igneous Hacks. 315 



the contrary, it is the relative proportions of the constituents which 

 establishes the identity of the rock. For instance, a kalitheralite 

 (e.g. malignite) may contain the same minerals as an alisyenite. 

 The former may contain 25 per cent, orthoclase, 25 per cent, 

 nepheline, and 50 per cent, of ferromaguesian minerals, and the 

 latter 90 per cent, orthoclase, 5 per cent, nepheline, and but 

 5 per cent, of ferromagnesian minerals. They may be composed 

 of exactly the same minerals, and yet in such different proportions 

 that they belong to widely separated families. 



Another example will show how unsatisfactory is this practice of 

 merely giving a qualitative description of a rock. I have, of late, 

 frequently had to read descriptions of rocks in order to determine 

 to which subdivision of a family they should be consigned in cases 

 where the subdivisions are distinguished by the most abundant ferro- 

 magnesian mineral. In one description, after two or three pages had 

 been devoted to an amphibole and two pyroxenes which the rock 

 contained, it was just mentioned at the end that the rock contained 

 biotite. The impression given was that the amphibole and pyroxenes 

 were the important ferromagnesian minerals, whilst biotite was quite 

 subordinate in quantity. On examining a slice of the rock, however,^ 

 I saw at a glance that the biotite was far more abundant than both 

 the amphiboles and pyroxenes taken together, and that it had received 

 so little attention merely because it presented no point of special 

 interest. Now that an easy method of determining the relative 

 proportions of constituents is known, namely, that of Rosiwal already 

 referred to (see footnote, p. 305), it is to be hoped that there will be 

 speedy reform in this matter. 



Although I believe that the names given to the rocks in the 

 above table generally represent their composition correctly, I cannot 

 vouch for the fact. Owing to the vagueness of many descriptions 

 in the above-mentioned respect, and because I have not thought it 

 worth while merely for the purpose of illustration in a preliminary 

 notice to spend much time in consulting literature not to be obtained 

 in Cambridge, it may be that a few of my names will bear correction. 

 That cannot, however, affect the value of the system of nomenclature 

 itself; and I trust that authors who are acquainted with any rocks 

 which I may have misnamed will be so kind as to set me right. 



IX. The Naming of Newly Discovered Rocks. 

 In conclusion, I may perhaps offer some suggestions as to how 

 newly discovered rock-species should be named in the future. The 

 first step in identifying a rock is to determine its essential con- 

 stituents, that is to say, those which make up more than 5 per 

 cent, by weight of the whole rock. According to their nature, 

 whether they are of the alkaline type or not, the rock is first 

 assigned either to the Alkaline or the Calc-alkaline Series. Amongst 

 the essential constituents it will next usually be possible to find 

 two or three which coincide with the index minerals of some 

 particular family, and the rock can then be forthwith placed in 

 that family. Should the essential constituents include the index 



