556 ALFRED HARKER 



visibly cutting one another. In the peridotite this composite struc- 

 ture is more strikingly exhibited, and it can be detected in places in 

 the granite, which is a much more uniform rock. The several compo- 

 nent sheets are not disposed in an orderly fashion in accordance with 

 their various densities. Add to this evidence the fact that peri- 

 dotite, gabbro, and granite all make smaller separate intrusions, 

 some much too far away from the main complex to have any direct 

 connection with it, and it will appear beyond dispute that the 

 differentiation which yielded these various rocks was effected prior 

 to their intrusion, and therefore in some large reservoir at a deeper 

 level. 



Bowen does not refuse the conception of a deep-seated magma 

 basin stratified according to density; but he seems to think it an 

 absurdity that, on that hypothesis, the earlier intruded magmas 

 should be drawn from the lower levels (p. 73). I will try to remove 

 his objections. I have already urged 1 that in order that such a 

 basin may have a considerable degree of permanence, as it obviously 

 has, we must suppose some approach to thermal equilibrium 

 between it and the surrounding crust. This implies a temperature 

 gradient within the basin approximately like the normal gradient 

 in the earth's crust of the region. It implies, further, what I may 

 call a fusibility gradient corresponding with this normal temperature 

 gradient. Now, the separation and sinking of crystals, as pictured 

 by Bowen, goes with a cooling-down of the magma, which terminates 

 in complete solidification. Any intrusions drawn from the basin 

 must therefore be consequent upon remelting. The occasion of 

 this I presume to be a gradual rise of the isothermal surfaces, 

 which must accordingly become more closely spaced. In other 

 words, reheating implies a temperature gradient steeper than that 

 to which the fusibility gradient is adjusted, and it follows that 

 the lowest layers will first become fluid. I have not attempted to 

 develop this view of the matter, and should welcome criticism; but 

 Bowen's zeal for differentiation in place has caused him to pay little 

 regard to the possibilities of this alternative. 



I am wholly in accord with Bowen in the conviction that alka- 

 line and calcic rocks are derived from the same primitive magmas 



1 See especially Comptc Rendu XII Congr. Geol. Intern., Toronto, 1914, pp. 205-8. 



