2 4 o SCIENCE PROGRESS 



as free oxides, or as elements, or in some other state of association 

 or dissociation ? The question is quite distinct from that of 

 possible ionisation, which likewise has its importance. 



Some petrologists have explicitly supposed, and others seem 

 to assume by implication, that an igneous rock-magma is a 

 mixture of free oxides ; so that a chemical analysis in the 

 ordinary form will represent the constitution, as well as the 

 empirical composition, of the magma. It is not easy to imagine 

 the apparent suspension of chemical affinities implied in such a 

 mixture. This view involves also the supposition that the 

 various silicate compounds known to us as rock-forming minerals 

 are formed simultaneously with their crystallisation from the 

 magma ; which presents another difficulty, since, from analogy, 

 we should expect such nascent compounds to appear as 

 amorphous precipitates. The objections from the petrological 

 side are not less weighty. If rock-magmas were merely mixtures 

 of free oxides, and the variation which exists among rock- 

 magmas (by whatever physical causes produced) were a variation 

 in the relative proportions of the oxides, supposed to behave 

 independently, we should expect a much richer list of rock- 

 forming minerals than we have actually to record. Even from 

 the eight most abundant oxides, which appear in all chemical 

 analyses of igneous rocks, many minerals might be formed 

 which do not in fact occur in this connection in nature. Some 

 of these — such as wollastonite, monticellite, andalusite, cordierite, 

 periclase, and others — form readily, and are frequently produced 

 in the thermal metamorphism of sediments ; but they are not 

 found in igneous rocks, or are found only where there is reason 

 to believe that the magma has been contaminated by absorbing 

 sedimentary material — an exception of the kind which emphasises 

 the rule. 



Igneous rocks collectively present a very wide range of 

 variety as regards chemical composition, and there is often 

 great difference in this respect between parts of one and the 

 same rock-body. Nevertheless, this diversity is brought about 

 by the association in various proportions of minerals selected 

 from a very limited list. Further, a given mineral occurs in 

 very different kinds of rocks, and has therefore crystallised from 

 magmas differing greatly in total chemical composition. It is 

 scarcely possible to avoid the inference that the differentiation 

 of the parent rock-magmas, to which the diversity of igneous 



