Dr. H. Warth — Chemical Classification of Igneous Rocks. 131 



An Appendix will follow, later on, with figures and descriptions 

 of the Cirripedia and Polyzoa. — E. M. B. 



EXPLANATION OF PLATES. 

 Plate VIII. 

 Fig. 13.— North BlufE, seaward aspect ; May, 1905. 

 ,, 14. — North Bluff, seaward aspect, showing (but only faintly, o\ving to shadow) 



the chalk roof over the clay pinnacle. 

 ,, 15. — October 2nd, 1905, showing the aspect from the head of the south bay the 

 day after the chalk roof was broken through. 



Plate IX. 

 Fig. 16.— October 2nd, 1905. South bay. 

 „ 17.— October 2nd, 1905. 

 ,, 18.— October 16th, 1905. 



VII. — A Method of Classifying Igneous Eocks according to 

 THEIR Chemical Composition. 



By Dr. Hugh Warth. 

 (WITH A FOLDING TABLE.) 



THE chemical classification of igneous rocks is rendered difficult 

 by the large number of substances which are present in them. 

 H. S. Washington, who based his system of classification upon the 

 composition of standard rock-forming minerals, found it necessary in 

 his great work ^ to divide his 2,880 rocks into no less than 167 

 final groups in order to ensure a close proximity between the rocks 

 within each group. 



The number of rocks in any system of classification must rise 

 so much more rapidly the greater the proximity of the individual 

 rocks to each other. In the case of only a single constituent the 

 deviation of individual rocks from the group average is inversely 

 proportional to the number of groups. A similar law prevails 

 when several constituents are considered at the same time, as will 

 be shown in the following. 



Five hundred rock analyses were selected at random for the 

 purpose of classification. The average composition of this whole 

 assembly of rocks was then calculated, and the mean deviation 

 of the several substances was found by deducting the percentage of 

 each substance present from the mean percentage of this substance 

 in the five hundred rocks. 



The differences obtained for all the rocks, positive as well as 

 negative, were then added together, and the sum-total divided by 

 500 gave the mean deviation of each substance. It requires to be 

 noted, however, that for the present purpose some of the substances 

 were taken two and two together, and their combined deviations 

 were thus ascertained. The following is the result : — 



500KOCKS. Si02 AI2O3 FejOs FeO MgO CaO NajO K2 



Average composition ... 57-0 15-5 3-9 3-9 4-8 5-8 3-6 2-9 



Mean deviation +9-6 3-0 4-0 7*0 2-6 



1 H. S. Washington, "Chemical Analyses of Igneous Rocks"; Washington 

 Government Printing Office, 1903. 



