CLASSIFICATION OF IGNEOUS ROCKS 687 



which each object can be easily and promptly thrust and dock- 

 eted, and any method which fails to achieve this and relieve 

 them of considerable mental effort in classifying rocks will not, 

 in their estimation, be a proper one. 



Unfortunately the difficulty in this respect lies not in the 

 method, but is inherent in the subject itself. Rocks grade into 

 one another in all directions, chemically, minerally, and textur- 

 ally, and these again overlap in the different modes of geological 

 occurrence. 



Therefore, unless the future should reveal new properties of 

 rocks, as yet unknown and unsuspected, which may be used as 

 bases of classification, every method so far devised, or which 

 can be devised, must have artificial lines of division, and cause 

 the petrographer trouble in certain cases in deciding where his 

 rocks belong. The erection of the monzonite group by Brogger 

 is a good example of this. Where rocks contained about equal 

 amounts of orthoclase and plagioclase it was formerly difficult 

 to say whether they should be classed as syenites or diorites. 

 The formation of the monzonite group avoided this difficulty, 

 but in its turn gave rise to two others whose sum total is greater 

 than the original one ; for it is clear that it will be equally diffi- 

 cult now to decide whether a syenite with considerable plagio- 

 clase is still a syenite or a monzonite, and the same difficulty is met 

 in diorites with considerable orthoclase. Thus, the formation of 

 new groups simply multiplies the difficulty ; it does not remove it. 



The system which we propose does not, therefore, meet this 

 trouble ; it does not pretend to, for any system based on our 

 present knowledge of rocks which should claim this would be in 

 the nature of a mere nostrum, and would simply profess to do 

 what it could not perform. In the ultimate analysis, every 

 system will and must throw on the petrographer the mental 

 responsibility of deciding, in a considerable number of cases, where 

 rocks lie close upon divisional lines, into which of two divisions 

 they must be put. He may not like this, but it cannot be avoided. 



In different systems this difficulty is met in greater or lesser 

 degree in different parts of the systems. In the one we propose 

 it occurs chiefly in doubtful cases, at the beginning, and in the 



