152 CARNEGIE INSTITUTION OF WASHINGTON. 



problem of the diversity of their deeper-seated relatives. The slowly 

 cooled subterranean masses of igneous rock are found to consist of 

 a mixture of minerals occurring as crystals. The one rock may be 

 composed entirely of a certain group of minerals, another may contain 

 the same minerals but in very different proportions, and still another 

 may be made up of a wholly different group of minerals. Indeed, the 

 profusion of rock types is so great that the.y have hitherto baffled 

 satisfactory systematic classification. 



From the finished products as he finds them, the field geologist must 

 do his best to infer the processes probably involved in their production ; 

 the experimental investigator devises means of testing in the laboratory 

 whether the supposed processes can take place and whether they will 

 produce the result observed in nature. 



As long ago as 1844, at a time when petrology was in its infancy, 

 Charles Darwin suggested that the sinking of crystals as they form in 

 the cooling molten rock may be a potent factor in producing diverse 

 rock types. The liquid rock as it cools does not acquire the crystalline 

 or solid state as a unit, but certain of its constituent minerals crystallize 

 at an early stage and others at a later stage. Plainlj'-, then, if the 

 earher-formed crj^stals are heavier than the molten material in which 

 they form, they should, under the action of gravity, sink in that liquid. 

 Equally plainly, these portions of rock mass into which a certain kind 

 of crystals moved during this process must become enriched in crystals 

 of that kind, and other portions must be correspondingly impoverished. 

 As we have already seen, it is principally a variation in the proportions 

 of the different kinds of crystals that gives rise to the different rock 

 types. It is but natural that Darwin and later geologists should have 

 held to the view that the sinking of crystals has been a factor in the 

 production of this diversity of types. 



It is natural also that the laboratory student should undertake experi- 

 mental investigations to determine what conditions are favorable to 

 the working of the process and whether the results observed in nature 

 are those which the process gives. 



Experiments were therefore instituted with a series of artificial sili- 

 cate mixtures containing pyroxenes and olivines, members of two 

 common groups of rock-forming minerals. These mixtures had for- 

 merly been completely investigated thermally, so that it was known 

 for each mixture at what temperature crystallization would begin, at 

 what temperature it would be complete, and the order of crystallization 

 of its various mineral components. By holding a mixture at a definite 

 temperature at which it was known that both liquid and crystals would 

 be present, opportunity was afforded for the sinking of the crj^stals. 

 It was found that sinking of both pyroxene and olivine crystals took 

 place quite readily, even in a small crucible. It was noted also that 

 the accumulation of crystals at the bottom increased as the time was 



