4o6 N. L. BOWEN 



crystals one can obtain a mass showing the last liquid as an upper 

 layer and also as interstitial material in part of the rest of the 

 mass. This is likely to give a pair of rocks showing that marked 

 contrast between a granophyre on the one hand and a diabase or 

 gabbro with only granophyric interstices on the other. Daly has 

 failed to grasp the significance of zoning in. producing grano- 

 phyric interstices and seeks to destroy the whole theory of 

 crystallization-differentiation at one fell blow by pointing out that 

 pyroxene, olivine, and plagioclase have not sunk out of granophyric 

 diabase.^ It is not necessary that they should, for the prevention 

 of reaction between crystals and the liquid from which they sepa- 

 rated may be accomplished, as we have seen, not only by a spatial 

 separation but also by a localized mechanical separation due to 

 zoning. 



There is, however, a third method of separation of liquid from 

 crystals that is probably the principal agent of production of those 

 discontinuous variations frequently seen in rock series and particu- 

 larly that shown in the gabbro-granite (granophyre, etc.) associa- 

 tion. It is the squeezing out of residual liquor at a stage when 

 the mass is largely crystalline, and it is many times more promising 

 as a possible cause of discontinuous variations than is any process 

 of limited liquid miscibihty, which, as we have seen, is a continuous 

 process requiring the slow action of gravity as an ally and not sig- 

 nificantly more competent than crystal settling to produce dis- 

 continuity. This squeezing out of liquid is a process that is at first 

 thought difficult to visualize, yet it is one that must be a very real 

 factor in igneous rock genesis. It is the sort of thing one may see 

 at any time when walking on the wet sand of a beach. The foot 

 leaves a slight imprint in the sand that is surrounded by an area 

 where the surface of the sand is particularly wet. The mass as a 

 whole is sensibly incapable of flow, so that the pressure of the foot 

 results merely in a more efficient close-packing of the sand grains 

 coming immediately under its influence, with a consequent squeez- 

 ing of water into the surrounding sand, where some of it exudes 

 upon the surface. On the other hand, in very wet sand or in mud 

 the whole mass may flow under the foot. The filter-press . action 



^ J our.Geol. , XXVI (1918), 121. 



