474 Reports and Proceedings — British Association. 



in a magma which still contains the iron of the magnetite in solution 

 plagioclase and augite may be comparatively soluble and magnetite 

 comparatively insoluble, but that when magnetite has already 

 crystallised out from the magma the plagioclase and augite may be 

 comparatively insoluble ; the experiments which are wanted are 

 experiments upon the solubility of certain minerals in magmas of 

 known composition under known conditions ; in these and similar 

 instances the order of separation is that of the solubility, but such 

 physical factors as the velocity of crystallisation (which varies very 

 considerably with the temperature), and the viscosity, may com- 

 pletely invert the order. 



Direct experiments made by Barns and Iddings upon the electric 

 conductivity of silicate magmas afford evidence that such magmas 

 contain dissociated as well as undissociated molecules, so that they 

 cannot be regarded as merely fused mixtures of certain minerals. 

 If two or more rock-forming minerals be fused together it may 

 happen that they form new compounds and crystallise out as different 

 minerals, or if one or the other remains unchanged it may crystallise 

 out in a different proportion. All this shows that double decom- 

 position goes on in the liquid. We cannot therefore expect, without 

 knowing the degree of dissociation, to make such use of the lowering 

 of the freezing-point in order to calculate the other factors in the 

 process of rock-formation. 



Doelter concludes that upon the whole the normal order of 

 crystallisation in rocks is in the main that laid down by Eosenbusch 

 long ago, namely, an order of increasing acidity, but that it is deter- 

 mined by the mutual affinities of the molecules in the magma, and 

 by the relative power of crystallisation of the components into which 

 they unite themselves, and that the physical factors which I have 

 already enumerated play a very important part in the process. No 

 one has endeavoured more systematically than Doelter to determine 

 for the rock-forming minerals the melting-points and the solubilities, 

 without which it is impossible to make much progress in our recon- 

 struction of the history of rocks. He has recentl}'^ shown us how the 

 mici'oscope may be used in the study of fused silicates at high 

 temperatures, and has so opened up a new field of research. 



Yogi's Applications of the Laws of Solutions. 



The work of Vogt has extended over many years, and is now 

 summarised in two remarkable memoirs recently published by him, 

 in which are expressed his mature opinions upon silicate magmas ; 

 the reasoning is based upon his own experiments, upon those of 

 Doelter, and upon the classic researches of Ebelmen, It is 

 now generallj' conceded that the particular minerals produced in 

 a silicate magma depend much more upon the chemical composition 

 of the magma than upon temperature and pressure ; Lagorio and 

 Morozewicz were led to this conclusion by their own experiments 

 upon fused silicates, Experiments upon slags at ordinary tempera- 

 tures and pressures may, therefore, be invoked to elucidate the 

 formation of rocks. 



