276 SCIENCE PROGRESS 



department of petrology in which so much progress has been 

 made in recent years, or in which there is such promise of still 

 further advances in the immediate future. 1 Yet, although the 

 actual phenomena that attend the consolidation of igneous rocks 

 are still obscure, there seems little doubt that in some cases, 

 at least, differentiation takes place in a partially crystallised 

 magma by the straining out, for some reason, of the still liquid 

 residue from the crystals that have already formed. The former 

 is usually eutectic in character, and therefore more or less definite 

 in chemical composition. The different minerals that have crystal- 

 lised out and the liquid residue may therefore be considered, for 

 some purposes, units of differentiation ; but these are, in many 

 respects, widely different from the mineral molecules that play 

 their part in the quantitative system of classification, and which 

 have been chosen because they lend themselves to a compara- 

 tively simple, but arbitrary, resolution into mineral groups. 



But the differentiation which appears to occur in a magma 

 that is still entirely liquid is at least equally important, and 

 in this case, too, there is reason to believe the oxides do not, 

 in all cases at least, act as independent units, but are associated 

 in larger molecules. 2 Alumina, for instance, may form with 

 the alkalies or lime compounds analogous to the spinels, and 

 these again appear to combine under certain circumstances to 

 form the simple felspars, or leucite or nepheline. There are, 

 however, no grounds for supposing that the minerals that 

 ultimately separate out necessarily represent the previous 

 molecular condition of the magma. 3 At present our information 



1 The whole subject has been reviewed by Mr. J. J. H. Teall in his Presidential 

 Address to the Geological Society {Quart. Journ. Geo/. Soc, vol. Ivii. [1901] 

 pp. lxx to Ixxxvi), and by Prof. H. A. Miers in his opening address in Section 

 C of the British Assciation in 1905. See also J. H. L. Vogt, " Physikalisch- 

 chemische Gesetze der Krystallisationsfolge in Eruptivgesteinen," Tscherm. Min. 

 Pet. Mitt., vol. xxiv. (1906) p. 437. 



2 See, inter alia, Rosenbusch, Elemente der Gesteinslehre, 2nd edition, Stuttgart 

 (1901), pp. 186-95. 



3 The fact that felspar crystals, having a composition intermediate between 

 those of albite and anorthite, crystallise out from a magma does not prove that 

 molecules, with the same complex composition, existed in the magma. What 

 evidence there is tells in the opposite direction, for the albite and anorthite 

 molecules appear to differentiate quite independently. But while the authors 

 of the quantitative classification are quite right in considering only the simple 

 felspars as units of classification, there appears to be no good reason for assuming, 

 as they do, that albite and anorthite are equivalent molecules which play nearly 

 the same part in the process of differentiation. 



