118 H. S. JEVONS, H. I. JENSEN AND C. A. SUSSMILCH. 



unusual chemical constituents, or of some specific pecu- 

 liarity of the minerals in habit, colour, or inclusions. 



So far as the properties of mixtures of molten silicates 

 are known, the latter obey laws closely analogous to those 

 exhibited by solutions and mixtures of liquids at ordinary 

 temperatures. Until the contrary is proved it seems to me 

 very reasonable to suppose that the analogy is a near one; 

 and this assumption simplifies considerably the task of 

 finding the processes by which differentiation most probably 

 takes place. We may assert with confidence, in the first 

 place, for instance, that differentiation must be the result 

 of a change either of temperature (usually cooling), or of 

 pressure ; probably, usually, of both. 



The only process of differentiation which actually occurs 

 in mixtures of liquids without any connection with con- 

 solidation is what was termed by Durocher liquation. 1 

 Certain liquids, (say A and B) are miscible in all proportions 

 (consolute) above a certain temperature, but as they cool 

 they separate into two immiscible liquids, one a saturated 

 solution of A in B, the other a saturated solution of B in 

 A. As the temperature is lowered further, each liquid is 

 saturated with a smaller and smaller percentage of the 

 other, till one or both of them may become nearly pure. 

 When the separation first takes place an emulsion is formed; 

 but if the liquids differ in specific gravity, the heavier 

 gradually collects as a layer beneath, and distinct from the 

 other, the separation being quicker the greater the differ- 

 ence of density, and the less viscous the liquids. The 

 familiar example is phenol and water ; 2 and others are 



1 Enc. Brit. 9th ed. x, 223a, from translation in Haughton's Manual of 

 Geology, 1866, p. 16. 



2 Teall, Q.J.G.S., 1901, p. 79. Whetham, Solution and Electrolysis, 

 1895, p. 17; Nernst, Theoretical Chemistry (trans. Palmer, Macmillan, 

 1895) p. 411 ; W. D. Bancroft, The Phase Kule, (Cornell Univ., Ithaca, 

 N.Y., 1897). pp. 102 and 126. I should like to draw attention to the 

 important bearing of the information contained in the last two works 

 upon the physics of rock-magmas. 



