290 NATURAL SCIENCE. 



JUNE, 



There is probably a local concentration of molecules before a crystal is 

 actually produced. This leads us to consider whether magmas may not 

 become differentiated quite apart from the actual separation of crystals. 



It follows, as a necessary consequence of Van't Hoft's theorem 

 (4) relating to osmotic pressure, that if the different parts of a dilute 

 solution are at different temperatures, the degree of concentration 

 will vary inversely as the absolute temperature. Thus, to take a 

 numerical example, the amounts of salt in a given mass corres- 

 ponding to temperatures of 80° and 20° C, after equilibrium has been 

 established, will be as 



(273 + 20) : (273 + 80) = 293 : 353 = I : I -205. 



In other words, a given weight of the colder portion will contain 

 about one-fifth more salt than the same weight of the warmer portion. 

 This had been experimentally established by Soret before Van't Hofi 

 proved it to be a consequence of a more general law, and it may 

 therefore still be conveniently referred to as Soret's principle. 

 Seeing that igneous magmas are of the nature of solutions in which 

 the mineral on the point of separating out may be supposed to be 

 dissolved in the rest of the magma, we have every reason to believe 

 that Soret's principle may have important geological applications. 



The general order in which constituents separate out in magmas 

 of intermediate composition is as follows. First, iron-ores, then the 

 ferro-magnesian constituents, and, lastly, felspars and quartz. 

 Among the felspars, the lime-soda members of the group precede 

 oligoclase, orthoclase, and microcline. From this we see that the 

 compounds liable to be concentrated by Soret's principle are, first, 

 those of the iron-ores, and, secondly, those of the ferro-magnesian 

 silicates. Later on we shall call attention to certain facts which 

 receive a simple interpretation by the application of this principle. 

 At present, however, we have to notice another cause which tends to 

 destroy the homogeneity of a solution. Gouy and Chaperon, by 

 applying a well-established principle of thermo-dynamics, have 

 deduced the conclusion that a solution standing at rest under the 

 influence of gravity will become more concentrated in the lower than 

 in the upper part. It is difficult to verify this conclusion experimen- 

 tally, because the height of the liquid column must be very great — 

 something like 100 metres — in order that any sensible differences 

 may be produced. Another cause which may operate to intensify the 

 concentration started in some other way is magnetic attraction ; at 

 least this is suggested by Professor \'ogt (4), although it cannot be 

 said to rest on evidence as satisfactory as that which can be 

 advanced in support of the principle of Soret or that of Gouy and 

 Chaperon. The idea is that the molecules of ferriferous compounds 

 are themselves liable to be affected by magnetic attractions, so that 

 if a local accunmlation of such molecules be produced in any way, 

 this attraction will intensify the concentration and extend the space 

 in which it is going on. 



