THE STRENGTH OF THE EARTH'S CRUST 31 



areally larger we must draw boundaries about large regions which 

 show a dominance of anomalies of one sign. These boundaries, 

 however, must be taken so as to give compact unit areas, so as not 

 to obtain an unreal result by the political expedient of gerrymander- 

 ing the districts. 



Select as a center the point whose geograj)hic co-ordinates are 

 lat. 42'^, long. 1.02°. Describe about this center a circle of 850 km. 

 radius. This includes an area equal to 29 per cent of the area of 

 the United States. It should be taken as including the negative 

 anomaly station 99 on its southern, border. This circle covers a 

 large positive region which could be made still more positive by an 

 extension of its boundaries to the northeast c)ver Wisconsin and 

 Michigan. Within this circle are distributed with a fair degree of 

 uniformity 31 of the 122 gravity stations of the United States. 

 The mean with regard to sign of the anomalies of these 31 stations 

 referred to the United States mean with regard to sign is H-o.oro 

 dyne. As the mean without regard to sign of all stations in the 

 United States excluding Seattle is only 0.018, it is seen that this 

 positive region stands out clearly from the general average. 



West of this circle and, on the south, to the west of long. 107'^ 

 there are 21 stations, including one of the two Seattle stations. 

 These mark a broad region of negative anomaly. The mean 

 anomaly with regard to sign is —0.017 dyne. There seems to be 

 no reason for completely omitting the exceptionally large Seattle 

 anomalies. One of them has therefore been retained, but if both 

 are omitted the mean is still —0.013. 'J'he value of —0.017 

 will here be adopted. The difference of the means of the central 

 and western regions is consequently 0.027 dyne. Let these be 

 regarded as the positive and negative phases of an harmonic wave 

 and the mean departure of the two phases becomes 0.0:35 from 

 each side of the mean plane. Now it may be computed for a 

 harmonic wave represented by the formula y=A. sin Bx that the 

 mean height of the wave above the mid-plane is 64 per cent of the 

 crest height. From mid-plane to crest of this wave-series is there- 

 fore 0.021. From the large negative anomalies along the Pacific 

 coast it would seem that this negative zone must extend somewhat 

 further. The wave-length of this series is consequ('ntIy between 



