226 REPORT—1899. 
by their short duration and by the rapidity of their vibrations. The other: 
group, which in Tokio form about 5 per cent. of the whole, can be felt for: 
several minutes, and the period of movement is long. With many persons. 
earthquakes having this character produce feelings of nausea, and there is. 
abundant evidence to show that they represent undulations of the surface: 
of the ground. By the former of these groups, although they may some- 
times alarm a city, free horizontal pendulums, unless constructed like a 
bracket seismograph, are seldom disturbed, whilst the latter throw such 
instruments into violent and fitful motion which, rather than extending: 
over two or three minutes, continues for as many hours. 
One class of earthquake consists of what are practically elastic vibra- 
tions, which have a short life and do not travel to great distances from 
their origin, whilst the other class gives rise to surface waves which are 
propagated to very great distances. 
The earthquakes which are merely elastic shiverings may possibly be 
represented at their origin by a blow delivered on a small surface, whilst 
those which are shiverings accompanied by surface heaving are the result 
of collapse in and along an extensive region. 
If we divide earthquakes into these two groups, between which con- 
necting links, if they exist, are very rare, we then see an escape from the 
prevalent idea that as earthquakes radiate their duration apparently 
increases. 
Although we know that preliminary tremors outrace large waves, that 
both of these forms of movement increase in period, and that a single ~ 
wave at one station may at a more distant station be represented by two 
waves, all of which phenomena tend to the spreading out of a disturbance, 
it 1s difficult to realise that an earthquake recorded in Japan as having a 
duration of two or three minutes should, when it reaches this country, be 
represented by movements continuing over two or three hours. The 
circumstances which have led to this supposition are twofold. First, no 
distinction has been drawn between the two kinds of earthquakes ; and, 
secondly, the duration of a disturbance near to its origin has been deter- 
mined by a method very different from that by which it was determined at 
a distance. 
When these considerations are neglected the results we may arrive at 
are well illustrated in a paper on ‘ Earthquake Duration’ by Dr. E. Odone 
{‘ Atti della Reale Accademia dei Lincei,’ vol. iv. fas. 10, p. 425). We 
here find a list of twenty-four earthquakes, the origins of which were at 
distances varying between 25 and 11,170 kms. from Rocca di Papa, 
Rome, and Siena. At these places the duration of these shocks were- 
noted by fairly similar seismographs of the heavy pendulum type. A 
glance at this table apparently indicates that the durations of these earth- 
quakes had steadily increased with the distances of their origins from the 
observing stations. With an origin at a distance of, say, 25 kms., we find 
a duration of about 70 seconds, whilst if the origin was at a distance of 
9,000 kms. the duration becomes 4,800 seconds. 
Fer the first members of this series, which I will call local shocks, had 
the instruments employed been free horizontal pendulums it is very 
doubtful whether they would ever have been recorded, neither would they 
have been noted had the pendulums with their multiplying indices been at 
distances of a few hundred kilometres from their origins, The common 
experience, based on seismographic records of local shiverings in Japan, is 
that the duration of movement decreases with distance from an origin, and 
