THE EVOLUTION OF IGNEOUS ROCKS. 163 



moderate proportion of potash or a large proportion of 

 alumina in an acid slag produces a notable increase of 

 viscosity. The experiments thus tend to show that a 

 normal acid rock-magma, rich in silica and usually in 

 potash, will be markedly viscous, while a basic one, poor 

 in silica and rich in lime, magnesia, and iron-oxides, will 

 be very fluid. Vogt remarks that the difference between 

 extreme examples of slags is very great : at the same 

 moderate temperature-distance above their respective 

 melting-points, highly basic slags flow like water, while 

 highly acid ones are as sluggish as tar. It is probable 

 that any differences between the slags and natural rock- 

 magmas will favour greater fluidity in the latter, the water 

 which probably forms part of every rock-magma doubtless 

 tending to reduce the viscosity. On the whole, then, we 

 should not expect viscosity to be a serious obstacle to 

 diffusion, at least in the more basic magmas. It is, in fact, 

 in the basic rocks that the actual phenomena of differentia- 

 tion are most strikingly exhibited. 



A not uncommon type of variation in a magma which 

 has been differentiated and consolidated in situ is that in 

 which the resulting rocks grow progressively more basic 

 from the centre to the margin of the mass. The phenomena 

 agree with the supposition that those constituents of the 

 magma which are normally the earliest to crystallise out as 

 the temperature falls, viz., the iron-oxides and, in a less 

 degree, the more basic silicates, have become concentrated 

 in the cooler marginal parts of the reservoir. Vogt u has 

 given an economic interest to this part of the subject by 

 his hypothesis, that certain rich bodies of iron-ore are the 

 products of an extreme " magmatic concentration" of this 

 kind. His results have been abstracted in English, and 

 are well known. Numerous instances of bodies of plutonic 

 rock or dykes becoming more basic in composition from 

 centre to margin have been described by other authors, and 

 several of these are cited in the new edition of Zirkel's 

 text-book. 



It is more especially in view of such cases as these that 

 several petrologists have tried to apply what is known as 



