282 GROVE KARL GILBERT— DAVIS [MEM0IES t * 1Txxi; 



clearness and keenness of penetration in an involved subject. It is not a little gratifying to see 

 that this able essay, like the two essays on the distribution of mining debris in California, both 

 of which were completed after his serious illness of 1909, are marked by all the intellectual 

 vigor of his earlier years. 



The geodesists had, largely as a matter of convenience in their calculations, assumed that 

 the density of the rock columns, which rise with differing heights from the level of compensa- 

 tion to stations at different altitudes on the continents, is inversely proportional to the height 

 of the columns; that the departures from the mean crustal density thus introduced were uniform 

 through the entire height of each column; that the depth of compensation is everywhere the 

 same; and that no variations of density occur below that depth. They recognized that these 

 assumptions were not necessarily correct, but their discussion gave prominence only to the 

 possible incorrectness of the first. Gilbert therefore, while accepting the conclusion that the 

 isostasy of the crust is nearly perfect, looked into the possible incorrectness of the other three 

 assumptions. Space can be here given only for his consideration of one of them. He argued 

 that, inasmuch as the assumption of isostatic compensation by variation of density in each 

 crustal column inversely as its height involves a horizontal change of density from column to 

 column, a vertical irregularity of density distribution in the various columns no greater in 

 degree than that of the horizontal change might be plausibly accepted; and then, still working 

 on the assumption that the average density of each column is in inverse proportion to its height, 

 he showed that if, at a station where gravity is slightly in excess, a slight excess of density be 

 assumed in the lower part of the local column with a compensatory defect of density in the 

 upper part, the gravity excess will disappear; and conversely, that if, where gravity is slightly 

 in defect, a slight excess of density be assumed in the upper part of the local column with a 

 compensatory defect in the lower part, the gravity defect will again disappear. Thus isostatic 

 equilibrium might become exact. But it is evidently not necessary or even expectable that 

 it should be exact; the special assumptions as to the vertical distribution of densities here 

 assumed need not be true. The crust may sustain residual inequalities of pressure by its rigidity. 



However, in order to learn whether the facts observable at the surface give support to 

 these special assumptions, Gilbert next examined the distribution of the calculated gravity 

 anomalies, as charted on a map prepared by the geodesists with anomaly contour lines for 

 every 300 rock-feet of excess or defect, and considered the anomalies with reference to facts of 

 geological structure and history. He then concluded, as others had also, that neither the 

 composition of the surface rocks, nor any recent changes due to erosion or deposition, nor any 

 reasonable underground extension of surface structures can account for the anomalies. For 

 example, if the axis of the Allegheny plateau be followed from the Catskills southwestward to 

 central Tennessee — an axis of similar geological structure and history for ages past— there is 

 found a defect of over 900 rock-feet in the Catskills, a small excess in southwestern Pennsylvania, 

 a moderate defect in West Virginia, and an excess of 900 rock-feet in central Tennessee, where 

 the plateau is not so high as in West Virginia. Neither the surface composition, nor the geo- 

 logical history, nor the present altitude of these several districts can account for such differences. 

 Similarly, if one travel from southwestern Alabama, a region of modern deposition, across the 

 Mississippi embayment of the coastal plain, also a region of modern deposition, to western 

 Arkansas, a region of modern erosion, the first and second regions, which are geologically much 

 alike, will be found to have, respectively, a defect of 1,200 rock-feet and an excess of 900 rock- 

 feet; while the first and the third regions, which are geologically unlike, both have the same 

 defect of 1,200 rock-feet. Here again, neither composition, history, nor altitude will explain 

 the anomalies. In view of these and other similar facts, Gilbert reached the conclusion that 

 the. correspondences of "visible structure and physiography" with the gravity anomalies of the 

 map "are so slight that they may be regarded as accidental, and the general relation is that of 

 independence and discordance." He thereupon concluded further that, as the anomalies can 

 not be accounted for by the surface facts nor by the extension of those facts to a depth of a 

 few kilometers, they must probably be due not only to the irregularity in the vertical distri- 

 bution of density in the crustal columns down to the 122-kilometer level of compensation as 

 above assumed but also in part to heterogeneity in the nucleus below that level. This may 

 be regarded as a geological in contrast to a geodetic conclusion. 



