28 CLARKE— STATISTICAL METHOD 



[April 18 



surface conditions, and at five or ten miles below the surface it must 

 assume a different and unknown value. The coefficient becomes 

 larger as the temperature rises, but the influence of increased pres- 

 sure is undetermined. In short, the data as they stand to-day are 

 inapplicable to discussions of this kind ; although it is conceivable 

 that future discoveries may enable us to eliminate the difficulties that 

 now exist. When the physical properties of rocks and minerals 

 shall have been measured under widely differing conditions of 

 temperature and pressure, we may be able to apply the data to such 

 averages as I have given ; and so assist in the solution of geophysical 

 problems. 



Professor Van Hise^ has attempted to combine my third average 

 for the igneous rocks with the composite analyses of the sedimen- 

 taries, in order to determine the redistribution of the elements dur- 

 ing the processes of metamorphism. In so doing he has assumed 

 that shales, sandstones and limestones exist respectively in the pro- 

 portions of 65, 30 and 5 per cent, of the sedimentary rocks; and 

 upon recombining the data with allowances for matter contained in 

 the ocean, he found various excesses and deficiencies. These differ- 

 ■ences from the parent rock he seeks to explain; but it seems to me 

 that his efforts are premature. The data are not yet sufficiently 

 precise to justify so elaborate a discussion, as a comparison of Mr. 

 Washington's average with mine will show. Furthermore, as I 

 have already observed, several important factors in the problem 

 remain to be determined. The metamorphosed sediments and the 

 oceanic deposits are not covered by the composite analyses, and that 

 relative to the shales is otherwise imperfect. It represents only 

 78 rocks, a number which is quite inadequate. A much larger 

 amount of more varied material must be studied before the analyses 

 of the igneous and sedimentary rocks can be well fitted together. 

 Only the broadest relations are ascertainable now, as I have sought 

 to show in the preceding pages. We can determine, roughly, the 

 maximum relative mass of the sedimentaries, and also something of 

 their proportional quantities ; but much farther than that we are 

 hardly yet ready to go. One suggestion, however, merely as a sug- 



^"A Treatise on Metamorphism," pp. 947-1002. ^ 



