METHODS OF STUDYING EARTHQUAKES 303 
direction of the movement was nearly perpendicular to one 
of the principal walls of the house in which the estimate was 
made. But this very fact, which seems to render the observa- 
tions valueless, turns out to be of service; for the impression of 
direction is most distinct in buildings whose walls are perpen- 
dicular to the true direction of the shock. The majority of the 
records naturally come from such houses, and thus the average 
of all the estimates collected gives a nearly accurate result. The 
mean directions for London and Birmingham, for instance, inter- 
sect almost exactly in the epicenter, and those for several 
counties pass within a short distance of this spot.* Thus the 
method of directions, if we give a somewhat different meaning 
to it from that intended by Mallet, may determine the position 
of the epicenter with a close approach to accuracy. 
It is doubtful whether the second method, depending on 
time-observations, can ever lead to any but very rough results. 
The chief reason for this is the difficulty of determining the time 
accurately to within a few seconds. But, supposing this were 
possible, there is also the uncertainty whether it is the same 
phase of the motion which is timed by observers in different 
places; for the vibrations which appear strongest to different 
persons do not necessarily come from the same part of the focus, 
and may come from parts which are separated by a distance that 
is considerable when compared with the dimensions of the dis- 
turbed area. While good time-observations may enable us to 
determine the surface velocity of the earth-wave, they can hardly, 
unless very numerous, afford information of much value with 
regard to the position of the epicenter, and still less with regard 
to the depth of the focus. 
There remains the third, and by far the most fruitful, method 
of inquiry —that which is founded on the intensity of the shock. 
“The Hereford Earthquake of 1896” (Cornish Bros., Birmingham,) pp. 265-270. 
I have applied this method to the Charleston earthquake of 1886, for which Captain 
Dutton’s well-known memoir supplies the materials. Here it was necessary to group 
together observations in separate states, the areas of which are to large to give good 
results. But, in several cases, the mean direction so obtained differs only by a few 
dezrees from the line joining the center of the state to the epicenter. 
