394 Dr. C. Davison—Founders of Seismology— 
on many subjects, some of which—such as earth-tremors, the 
diurnal wave, and the tilting due to tidal loading—have no apparent 
connexion with earthquakes. The sections on the distribution of 
earthquakes in time have been already noticed. Some others 
relating to distant earthquakes—such as earthquake-echoes, the 
‘nature of the large waves, and the dissipation of earthquake-motion 
as the waves radiate outwards—must here be passed over. In the 
remaining outline of Milne’s work I will confine myself to two 
important subjects —the time-curves of the different phases and 
the distribution of the earthquake-origins in space. 
It is interesting to trace Milne’s successive attempts to draw 
the average time-curves of the earthquake-motion. The first, 
which is really a velocity-curve, appeared in the report for 1897, but 
the observations (for earthquakes of 1891-6) were discordant, and 
the only result deduced was an increase in the velocity of the early 
tremors with the distance, whether measured along the are or the 
chord. In the following year an advance was made by separating 
the preliminary tremors and the large waves, and a curve was given 
showing very clearly that the average duration of the preliminary 
tremors increased with the arcual distance of the origin. In 1900— 
these years are dates in the history of seismology—the first time- 
curves proper were drawn. Distances from the epicentre in degrees 
of arc were measured along the horizontal axis, and the times of 
transit of the initial preliminary tremors and the maxima of the large 
waves to particular stations in the perpendicular direction. The 
time-curve for the large waves is practically a straight line, corre- 
sponding to a constant arcual velocity of 3 kms. per second ; that for 
the preliminary tremors shows that the velocity along the arc 
increases with the distance. Two years later the curves begin to 
assume their now familiar form ; and, as they are based on records 
from Milne’s seismographs only, the divergences from the average 
are less noticeable than before. It had been shown by Oldham in 
1900 that the preliminary tremors are divisible into two series, 
the first consisting mainly of condensational, the second of dis- 
tortional, vibrations. A curve of the duration of the first preliminary 
tremors was also added. In the 1903 report the bearing of these 
curves is discussed. Measured along the arc from 20° to 150° from 
the origin, the velocity of the first tremors increases from 10°5 to 
‘16°3 kms. per second. Measured along the chord the velocity is 
more nearly uniform, increasing from 10°5 to 12°0 kms. per second, 
and showing that there is apparently a slight increase in speed as 
the paths approach the centre of the earth. The interpretation 
which these observations require according to Milne is a “ globe with 
an approximately uniform nucleus not less than $2 of the earth’s 
radius covered by a shell which passes rapidly upwards into the 
materials which constitute the crust of the world ”’.1 
1Since 1902, many hundreds of new observations have been collected, and 
these have naturally modified Milne’s curves and the tables founded on them. 
