328 Shepherd, Rankin, Wright — Binary Systems of 



consider each separately first, and then to indicate briefly 

 wherein they overlap and are mutually confirmatory. 



Chemical data. — Modern research has shown that as a rule 

 rock minerals are not single, simple compounds but complex 

 mix-crystals containing various other mineral compounds in 

 solid solution. The investigation of such minerals and their 

 relations in rocks can only be satisfactorily carried on, there- 

 fore, after the -characteristics of the simple compounds have 

 been ascertained, as well as the extent and the effect of the 

 by-mixture. In the general investigation of problems of such 

 wide scope it is necessary to begin with the simplest condi- 

 tions, and after these have been thoroughly mastered, to work 

 up to the more complex. From a physico-chemical standpoint, 

 rocks are as a rule complex systems, too complex in fact to be 

 treated satisfactorily until the simple integral systems of which 

 they are made up have been studied in detail. The present 

 lime-alumina series is only one of a number of two-component 

 systems which mark the limits of larger three-component 

 systems, and these in turn lead to still larger systems. Such 

 systems eventually become exceedingly complex, and the only 

 hope the observer has of mastering them is to begin with the 

 simplest cases first and then with the experience thus gained 

 to proceed step by step to the more complex. The simplest 

 systems are the two-component systems, as the lime-alumina 

 series, and the chief function of the chemical work is to make 

 up preparations of definite composition and to guarantee their 

 purity throughout the investigation. Natural rock minerals 

 are almost never rigidly pure, in the sense of definite and 

 invariable chemical composition, and yet their investigation 

 from the standpoint of laboratory synthesis requires that at first 

 only chemically pure preparations of definite composition be 

 taken, and the properties of these determined accurately ; later 

 the actual minerals can be reproduced artificially and the effect 

 of solid solution of different substances in different proportions 

 can be studied and definite information obtained. Impurities 

 in solid solution tend only to veil the true relations of the 

 compound itself, and for the observer to allow such a variable 

 factor as impurity into the investigation at the very outset 

 would operate not only to increase the difficulties but also to 

 decrease the clean-cut aspect of the problem and the laws 

 underlying it. 



Other data which are of a physico-chemical nature, such as 

 questions of relative solubility, concentration and the like, may 

 properly be considered in a later paragraph. 



Thermal data. — The object of experimentation along these 

 lines is not only to reproduce rock- making minerals artificially, 

 but especially to study the conditions of their formation and the 



