ON SEISMOLOGICAL INVESTIGATION. 65 



Origins in the East Indies {District F). 



. Durations 5 (441), 10 (505). 



8(3471, 10(435), 7(505). 

 „ ; 10 (324). 

 12 (377). 



10 (324), 9 (347), 8 (354) 9 (377), 

 7 (133).4 (134), 10 (324), 5 (347), 



.10 (355), 12 (404), 13 (460). 

 14 (505). 

 10 (324). 



Average values for the above durations which represent the intervals 

 in minutes by which the preliminary tremors have outraced tlie second 

 phase of motion at distances of 20, 30, 40, 50, GO, 70, 80, 90, and 100 

 degrees from their origins are, 4, 4 -5, G'o, 7*2, 8, 8-5, 9, 9 '5, and 9 '7 

 minutes. The materials to continue plotting the curve represented by 

 these figures, which is shown in fig. 3, indicate that ten minutes is 

 approximately the limit by which the second phase of motion is 

 outraced by the preliminary tremors, and this limit is reached at a 

 distance of ninety to 100 degrees from an origin. 



V. Time Curves /or the Earthquakes recorded during /our years 



ending in 1900. 



In the British Association Report for 1900, p. G7, time curves for 

 the large waves and preliminary tremors of earthquakes recorded at 

 long distances from their origins are given. These curves are based 

 upon records obtained from Milne horizontal pendulums prior to the 

 year 1900. In fig. 3 similar curves, together with a curve for the 

 second phase of earthquake motion, are gi\en for observations made 

 in 1900. The curve for the second phase of motion, which corresponds to 

 a curve given by Mr. R. T. Oldham,' was obtained by adding the time 

 ordinates of the lowest curve in this figure, discussed in the previous 

 section, to those of the time curve for the preliminai'y tremors. It is of 

 interest to note that this curve would closely correspond with a curve 

 representing the mean position of the signs x which are direct measure- 

 ments of the time taken by the second phase of motion to traverse 

 varying distances, and are not included in the materials upon which the 

 lowest curve is based. 



The individual observations relating to large waves are indicated by 

 small crosses ( + ), whilst those referring to the preliminary tremors are 

 marked by small circles. It is clear that the concordance between these 

 and their average position as represented by the curves is not so close as 

 could be desired. They do not dispose of the indication based on prior 

 observations that the apparent velocity of large waves is not unifoim, 

 but may be at its maximum in quadrantal regions. Also, as has been 

 shown in the time curves published in 1900, we see that the appaient 

 velocity of preliminary tremois may also be increased in regions (i0° to 

 90° from their origin. 



Dr. C. G. Knott writes about these observations as follows : — ' The 

 other day I took another and more careful look at your curves, and I must 

 confess that thej' bear out your old view of the large waves being surface 

 ■waves better than anything else. The large wave curve is a straight line, 



' rJdl. Trans. Boy. foe, p 1G3. 

 1902 jp 



