LATER STAGES OF EVOLUTION OF IGNEOUS ROCKS 25 



results in the foregoing case will be referred to in the following as 

 maximal zoning. 



If the cooling is extremely rapid, the mixture M may undercool 

 to some temperature below N before crystallization begins, when it 

 will crystallize to a mixture of forsterite and pyroxene P. It has 

 already been shown that extremely slow cooling (with mixing) will 

 give the same final product, no zoning of the crystals resulting in 

 either case. There is one definite intermediate rate of cooling 

 which will give maximal zoning. If the cooHng is somewhat slower 

 than the definite rate which gives maximal zoning, there will be op- 

 portunity for partial change in the composition of the various zones, 

 and the mixture D, for example, instead of becoming completely 

 crystalline only at the temperature of the eutectic diopside- 

 silica, will be completely crystalline at a somewhat higher tempera- 

 ture, and the outer zone of the mix-crystals, instead of being pure 

 diopside, will be somewhat more magnesian. If the cooling is 

 somewhat faster than the rate which gives maximal zoning, the 

 same result will be accomplished, this time because of slight under- 

 cooling maintained throughout crystallization. Any intermediate 

 rate of cooling differing from the rate which gives maximal zoning 

 will, therefore, give some zoning, but with a smaller range of com- 

 position between the inner and outer zones, this being accompanied 

 by a smaller range of temperature of crystallization and a smaller 

 range in the progressive change in the composition of the liquid 

 than that shown in maximal zoning. It is obvious that the rate 

 of cooling has a fundamental effect on the course of crystalHzation 

 in the mixtures of the system investigated and indeed must have a 

 similar effect on the crystallization of any mixtures from which 

 mix-crystals separate. 



Crystallization when the crystals are free to sink in the liquid. — ■ 

 Even when the cooling is extremely slow, approaching or even 

 reahzing the rate required to produce perfect equilibrium, the 

 sinking of crystals is a factor entering into the problem and tending 

 to produce the same general result as the intermediate rates of 

 cooling which give rise to zoning. The continual downward 

 movement of growing crystals accomplishes their removal from 

 that part of the liquid from which they crystallized. The effect 



