432 WILLIAM BOWIE 



58.8 km. The greatest range for the mean anomalies with regard 

 to sign is for Zone and for mountainous regions where the 

 stations are below and above the general level. The range here is 

 from +0.006 to — 0.010, which is 0.016 in all. This range is 

 almost as large as the average anomaly for the United States 

 without regard to sign. The largest range for the other methods 

 of distribution is 0.007. 



We should expect the regional and local anomalies to be approxi- 

 mately the same for all of the stations not in mountainous regions, 

 but we should also expect that the local and regional anomalies 

 would differ for the stations where the relief is great. Where the 

 station is below the general level, the computed value of gravity 

 is less because some of the compensation of the mountains is dis- 

 tributed closer to the station than it would be by the local distribu- 

 tion. Where the station is above the general level, the computed 

 value of gravity is greater because some of the compensation of the 

 local topography is distributed under the contiguous valleys. 



We may conclude, I think, that the solution of the problem of 

 local distribution of compensation of the topography is indeter- 

 minate to a certain degree; that is, that any one distribution is as' 

 probable as any other one, out to a distance of 59 km. from the 

 station. It is possible and, in fact, probable that this uncertainty 

 may extend to a distance somewhat greater than 59 km., but it is 

 very probable that it does not extend to a distance of 167 km. from 

 the station. This conclusion is based upon the geodetic evidence, 

 as furnished by the gravity anomalies, and has no connection with 

 geological evidence. Any decision as to whether one method or 

 another is the more probable within the distance of about 60 km. 

 should be left to the judgment of geologists. It is of course possible 

 that, with more geodetic data available, geodesists may be able to 

 throw additional light on this subject. 



It should be remembered that when making the test for the 

 most probable method of distributing the compensation horizontally 

 the anomalies are treated as if they were due only to the method of 

 distribution. As a matter of fact, they are probably due to a num- 

 ber of causes, and this fact has some effect on the results of the com- 

 putation, but it is believed that its effect is small. 



