406 11. S. WASHINGTON ISOSTASY AND ROCK DENSITY 



The relation is, however, not so simple, as the influence of varying values 

 of 8 comes into play. Inspection shows that the following relations hold 

 good mathematically : 



M decreases when A 8 (for constant A h) increases; 



M decreases when A h (for constant A 8) decreases. 



The problem becomes much more complex when both 8 and h Nary, 

 and when, also, varying numbers of extreme or intermediate data are 

 handled the final result is incapable of exact prediction. We shall see, 

 however, that the actual variation is not great when probable values are 

 assumed for 8 and h. 



It is a matter of difficulty to decide as to the respective merits of these 

 various values. Those derived from the normative densities can be re- 

 garded as but provisional approximations, in view of the tentative and 

 not wholly satisfactory character of some of their basal data. Their gen- 

 eral concordance within not very wide limits and their general agreement 

 with the values based on gravity determinations point to their approxi- 

 mate correctness, and also justifies the belief that the normative method 

 is based on sound principles and would be capable of yielding quite satis- 

 factory results were the chemistry and the hypsometry of the igneous 

 areas of the globe better known. 



Considering the quality of the data as to density and altitude which 

 are available at present, I regard the values based on a "ridge" Atlantic 

 depth as better than those which are based on the average Atlantic depth. 

 Furthermore, it would seem to be probable that the values arrived at 

 from the set of small areas and the ocean floors are nearer the truth than 

 those arrived at from the continents and the ocean floors, and still more 

 so than those derived from land areas alone. The value derived from 

 the United States alone is so small and so different from the others that 

 it may be neglected. With the small areas, we are dealing with sizes of 

 area which are more commensurate than are the continents with the pos- 

 sible isostatie units imagined by Harford, 80 who speaks of the importance 

 for future study of determining the maximum horizontal extent of areas 

 capable of showing isostatie compensation. 



Some considerations may point the way to a decision : First, the ma- 

 terial which composes the deep floors of the Atlantic, Indian, and Ant- 

 arctic oceans is unknown, although we can form some idea of its genera! 

 character, and it has not yet been taken into account. Second, the density 

 which is here assumed for the Pacific floor is almost certainly too low, 

 because of the undue proportion of "interesting" rocks which were 



80 Harford : The figure of the earth, 1909, p. 169. 



