92 REPORTS ON THE STATE OF SCIENCE. 
table, the second and fifth containing the number of times this particular 
average deviation occurred ; while the third and sixth columns contain 
the average number of earthquakes corresponding with the deviations. 
| | 
eet ; | | Aver 
| Deviations Number of | Average | Deviations | Number of | VSEEBe 
: Number of | Number of | 
| from Mean | Occurrences | Earthquakes from Mean Occurrences | Earthquakes | 
B hers Jaci Cee: | ri 
_ 28 i NS ech at a ia crea lei hpdane 
Hog) g Mart A te Re Rg) Te eas 
—18 10 | 65 +12 | 10 al 
= 1/6} 7 74 +17 8 17 
— 8 15 9-7 +22 6 | 20 
|} —8 23 13-9 +37 | 2 21 
+373 | 5 24 
The number of occurrences are given by way of contrast. It is 
obvious that they follow roughly the well-known law of grouping about a 
mean, the maximum being in the neighbourhood of zero deviation or mean 
deflection. But itis quite otherwise with the earthquake numbers, If there 
were no connection, direct or indirect, between the two phenomena, the 
earthquake numbers would be fairly constant throughout. There seems 
to be, however, a tendency toward greater values for higher deviations. 
For deviations up to +12 and —13 the averages total 66°5, or an average 
of 11-1. For deviations greater than these limits the averages total 126-6, 
or an average of 18:1. This conclusion lends a certain amount of support 
to Milne’s view that there is some connection between the occurrence of 
large world-shaking earthquakes and the movements of the earth’s pole. 
The mean curvature of the path of the pole is 30°5 per tenth year, or 
305 per annum. Hence the pole will make a complete revolution in 
365 x 360/305 days, or 432 days. ‘The value given by Chandler is 427 
days. It is well known that a rigid body of the size, figure, and mass of 
the earth will have a small precessional motion of period—305 days if the 
axis of the figure is not quite coincident with the axis of diurnal revolu- 
tion. To expliin the large discrepancy between the observed value 427 
and the theoretical value 305, Newcomb invoked the influence of elas- 
ticity in modifying the period of precessional rotation. His original 
calculation was admittedly approximate, and Hough! has worked out the 
problem in a more rigorous manner. Taking account of the elasticity 
only, he finds that the precessional period will have the value 427 days if 
the effective rigidity of the earth were a little greater than that of steel. 
Newcomb also pointed out that the mobility of the ocean would have the 
same effect of lengthening the precessional period. Further, if the effec- 
tive rigidity of the earth were to diminish all over, the precessional period 
would be increased. It is not easy to see what would be the immediate 
effect. of either a local diminution of rigidity or a local yielding to stresses 
such as takes place when an earthquake is originated. But it is at all 
events not unreasonable that some effect will be produced. This is prob- 
ably the direction in which we must look for the connection imagined by 
Milne. 
1 «Rotation of an Elastic Spheroid,’ Phil. Trans., vol. clxxxvii. A, 1896. 
