part 1] ANNIVEKSARY ADDRESS OF THE PRESIDENT. Ixi 



producing stresses set up by the sun and the moon. It is true 

 that many attempts have been made at different times to detect 

 some connexion between the frequency of earthquakes and the 

 position of the moon, and that no such connexion has yet been 

 established, but these attempts have all been based on very im- 

 perfect records. In time it ma}^, perhaps, be possible to apply to 

 an earthquake record that method of harmonic analysis which has 

 proved so fertile in the case of the ocean tides, but the day is long 

 distant when a record of sufficient completeness ^vi\\ be available. 

 Meanwhile, there are some simpler relations, of which a discussion 

 is feasible, and the most promising of these seems to depend on 

 the fact that the downward pressure is greatest at the time when 

 the attracting body is on the horizon, and least when it is on the 

 meridian. If, then, we divide an earthquake record into two 

 groups, one containing all shocks which occur within six hours 

 before a meridian passage, and the other all that happened within 

 six hours after, the first of the two groups wdll cover a period during 

 which the downward pressure is, on the average, increasing, while 

 the other will cover a period during which it is decreasing. As 

 the amount of the change so introduced is known, with sufficient 

 accuracy for the present purpose, and as it must, on the hypothesis 

 being used, influence the frequency of earthquakes, it follows that 

 we have here a method which should enable us to make an estimate 

 of the rate of growth of the strain to which fracture is due. 



Although simple in principle, the method is difficult in appli- 

 cation. To begin with, a record is required, of sufficient extent 

 and continuity to give a trustworthy average, not merely of the 

 general frequency, but also of the frequency in each of the two 

 sections into which it is divided ; and this means that the record 

 must contain at least two thousand shocks and ought to contain 

 double that number or more. Then it must be reasonably 

 accurate as to times and complete as to occurrences, or at least 

 must be fairly uniform in its incompleteness over the whole period 

 investigated. There are not many records which fulfil these 

 primary requirements, but there is another even more important. 

 In all records there is a noticeable variation in frequenc}'^ at 

 different times of the day ; moreover, the nature of this diurnal 

 variation has been found to vary in different regions, but appears 

 to be constant and chai-acteristic in each region over the period of 

 record. The cause of this periodicity" may be reasonably attributed 

 to some effect, meteorological or other, connected with the daily 

 course of the sun ; but its nature, no less than its variability. 



