468 Professor S. J. Shand — A System of Petrography. 



The names of the aphanitic subfamilies in this system ought of 

 course to be drawn from existing names of lavas, just as those of the 

 phanerocrystalline rocks have heen derived from existing names of 

 plutonic rocks, and the same set of prefixes and suffixes may be used 

 to signify the degree of saturation, thus : andesite, subandesite ; 

 latite, sublatite, latoid, sublatoid ; and so on. A certain amount of 

 redefinition of names will be necessitated, and a number of new 

 names will be required. It is not proposed to coin these in advance : 

 it is better that the supply should wait upon the demand. In the 

 meantime it is always possible to describe any rock by means of its 

 co-ordinates in the system, thus : 03 L2; S 7 1 1 ; uMl, etc. 



6. With five divisions, about twenty effective groups in each 

 division, and two subgroups within each group, we have already 

 distributed all known igneous rocks into about 200 compartments. 

 Within each of these compartments there is still room for consider- 

 able variation as regards both the minerals themselves and their 

 proportions. A further use of the colour-ratio prefixes, Is to h and 

 m 5 to ni9, makes it possible to describe variations in quantity of light 

 and dark minerals. In the same way we may express different 

 degrees of oversaturation by means of prefixes, O u 2 , O3, etc. It 

 will still be necessary to discriminate between different mineral 

 associations. Compartment U 2, for example, holds both leucite- 

 albite rocks and nepheline-orthoclase rocks, and these will certainly 

 require different " specific " names. Again, in the u division we find 

 rocks with unsaturated magnesium and others with unsaturated 

 calcium ; and in W the melilite rocks (for melilite holds soda as well 

 as lime and magnesia) form a distinct facies. It is therefore to be 

 expected that from 300 to 500 " specific" names will be required in 

 all. These should be supplied, of course, partly by the readjustment 

 of existing names and partly by the coining of new ones as the need 

 arises. Still further subdivision is possible on a basis of texture; 

 but it is the writer's opinion that differences of texture, other than 

 the one fundamental difference already considered, should be expressed 

 by prefixes or descriptive adjectives, not by the coining of new 

 names. 



It will be noticed that no place has been found in this system for 

 the order of crystallization. This is one respect in which it is to be 

 hoped that the system may in future suffer modification. In the 

 meantime the discrepancies between theory and observation are so 

 great that it would only be misleading to give to this factor any place 

 in systematic petrography. 



In conclusion, let me say that I do not claim for the system 

 elaborated above that it is either a " natural classification" or a final 

 one, or that it will enable one to dispense with chemical analysis. 

 It is nothing more than a system of indexing, based as far as possible 

 on significant mineralogicaldata treated quantitatively. It lays down 

 lines along which the naming of rocks should proceed, and fixes 

 sharp boundaries which no rock-name should transgress if it is to be 

 more than a mere sack-name. The coining of unacceptable new 

 names is largely avoided by the use of simple prefixes and suffixes, 

 and where existing names have been used their connotation has been 



