162 SCIENCE PROGRESS. 



purely mechanical forces operating on a magma in which 

 crystallisation was in progress. 



This leads to a few remarks on the possible causes of 

 differentiation in rock-magmas, on which question various 

 speculations have been advanced. When the separation is 

 merely between crystallised minerals and the residual 

 magma, it is easy to see that the cause may be a purely 

 mechanical one. We have seen an illustration in the 

 filtering process described by Barrow, and this is of special 

 interest as supplying a possible link between differentiation 

 and crust-movements. Again, gravity must be taken into 

 account as a probable factor. Very little is known regarding 

 the relative densities of rock-magmas and crystals at very 

 high temperatures ; but it is to be expected that the more 

 basic minerals at least would sink in a magma from which 

 they had crystallised out, unless that magma possessed a 

 high degree of viscosity ; and even in this case, if the 

 viscosity followed the laws laid down by physicists, the 

 sinking, though slow, would be no less certain. Such a 

 process would give rise, in the lower layers of a magma- 

 reservoir, to a relative richness in the more basic minerals. 

 We have seen, however, that in some well-established cases 

 differentiation was certainly not posterior to the crystallisa- 

 tion of these minerals, and geologists have accordingly 

 turned their attention especially to possible causes of differ- 

 entiation by diffusion in a still fluid magma. 



Here again the question of viscosity would seem to be 

 an important one, and any experimental results bearing on 

 the viscosity of rock-magmas are of value. Vogt's 10 re- 

 searches on artificial slags are of especial interest. They 

 were made on slags having the general composition of 

 rock-magmas, without water or any special fluxes, and 

 under atmospheric pressure. Vogt found that the vis- 

 cosity of such a slag increases with its content of silica, 

 and increases rapidly if the silica exceeds 58 or 60 per 

 cent., i.e., in magmas corresponding to acid rocks. On 

 the other hand, ferrous oxide in a slag strongly promotes 

 fluidity; magnesia has the same effect, though less markedly; 

 and lime, and probably soda, act in the same sense. A 



