292 R. H. Rastall — Differentiation and Ore -deposits. 



a simple case on which we possess some solid information as a basis. 

 In the smelting of metals from their ores the metal sinks to the 

 bottom, while the non-metallic impurities and fluxes rise to the top 

 as a slag. This is not merely sinking by gravity, but depends on 

 the principle of immiscible liquids. Vogt has shown that at any given 

 temperature above the freezing-point of both liquids there is a certain 

 amount of mutual solution, the metal dissolving some of the slag 

 and the slag dissolving some of the metal, and that the solubility 

 in each fraction increases with the temperature. It may be inferred, 

 therefore, that at a certain very high temperature there would be 

 complete mutual solubility in any proportions. Such temperatures 

 have not, however, been attained in ordinary furnace practice. 

 Now, if we invert the argument and imagine a rock-magma consisting 

 of mutually dissolved silicates and sulphides at a very high 

 temperature, but constantly losing heat by conduction, a time will 

 come when the magma must begin to separate into two layers, one 

 rich in silicates, the other in sulphides, and the sulphide portion being 

 the heavier will occupy a lower position, thus leading to a segregation 

 of sulphides at the bottom of the magma basin. In other instances 

 the metals may be present as oxides or fluorides, instead of sulphides, 

 but the result will be the same, namely, a stratification of the magma. 

 It is possible that in some cases the metallic compounds may be 

 lighter than the silicates, and therefore form the upper layer, but 

 the principle is the same. The amount of one of the constituents 

 present in the rock formed from the other fraction when solidified, 

 e.g. the amount of sulphide in normal gabbro or norite, will depend on 

 the solubility of the sulphide in the gabbro magma at its temperature 

 of consolidation. Therefore, it would appear that there must be 

 a limit to the percentage of primary ore that can occur evenly 

 disseminated throughout a rock mass. If the amount of metal was 

 originally in excess of this proportion it would necessarily form a 

 richer segregation layer at the bottom. The foregoing statement 

 represents what must occur in the simplest possible case. When 

 the metals form volatile compounds the relations are more complex, 

 as will be explained in a later section. 



The argument may now be carried one stage further back. It has 

 been suggested by certain leading petrologists that some of the 

 phenomena of differentiation, as shown by the distribution of varying 

 rock-types in definite areas, can be explained on the supposition of 

 limited miscibility between magmas of different silica percentage, 

 or in other words between acid and basic magmas. The general 

 idea is that with a falling temperature the silicates of magnesia, 

 iron, and lime become less soluble in the silicates of alumina, potash, 

 and soda, and a separation occurs, the silica dividing itself in the 

 proper ratio between the two fractions, thus producing two partial 

 magmas. The magnesia, iron, and lime require less silica for complete 

 normal saturation than the alumina, potash, and soda, hence one 

 fraction is a basic magma, the other an acid one, in the ordinary 



