86 REPORTS ON THE STATE OF SCIENCE. 
of a similar and similarly installed pendulum recording on a slowly 
moving surface. With very large earthquakes in which the preliminary 
tremors have comparatively large amplitudes, differences of this nature 
are not observable. Although attention is only called to the photo- 
graphic recording apparatus used with seismographs, it is evident that 
the character of the results obtained from other instruments may also to 
some degree be dependent upon the speed of recording surfaces. 
IV. Origins and Relationships of Large Earthquakes in 1906. 
The number of entries in the Shide register for 1906 is 207. Out of 
this number ninety-two may be regarded as megaseismic in character. This 
number is distinctly above the average. On the accompanying map the 
origins of seventy-one of these are shown, the districts of greatest activity 
being £ and F. The display of activity, however, to which public 
attention has been chiefly directed, occurred on the western shores of North 
and South America. On January 31 we had the Colombian earthquake, 
the origin of which was apparently sub-oceanic off the mouth of the 
Esmeralda River. This convulsion, which led to the interruption of 
several] submarine cables, was followed by shocks in the Antilles. On 
April 4 a disastrous shaking took place in the Kangra Valley, in North- 
West India. Ten days later the Formosan earthquake occurred, which 
ruined 5,556 houses. On April 18 San Francisco was destroyed. Not- 
withstanding the intensity of the initial impulse, which, we are told, sent 
earth-waves twice round our world, it is astonishing how very iittle 
damage was done merely by the shaking of the ground. The greatest 
destruction was occasioned by fire. The origin of this disturbance was 
along lines of faults, which can be traced for distances of several hundreds 
of miles. The strike of these faults 1s apparently fairly parallel to the 
coast line, or from N.N.W. to S.8.E. The seismograms obtained at 
distant stations lying to the east or west of California, or at right angles 
to the fault, were very pronounced, whilst at Cordova, in South America, 
records were extremely small. . Exactly the opposite occurred in‘ the 
Jamaica earthquake which took place on June 14. In this case the strike of 
the fault or faults was east to west, and the seismograms in Europe, 7.e., to 
the east of Jamaica, were extremely small. On August 17 Valparaiso 
and the towns in its neighbourhood were reduced to ruin. In Greenwich 
mean time the Valparaiso earthquake occurred at Oh. 41m. 2s. Thirty- 
three minutes before this, or at Oh. 8m., or Oh. 11m. G.M.T., a very large 
earthquake took place beneath the North Pacific to the north of the 
Sandwich Islands. The time taken for the second phase of this shock to 
travel from its origin to Valparaiso, a distance of 122°, would be about 
31 minutes. This time-interval suggests at least three possibilities : 
(a) The earth-waves from the North Pacific may have released a state of 
seismic strain in Chili; or (6) the earthquake in this latter country may 
represent an effort to establish a dynamical counterbalance consequent 
on a molar displacement in the North Pacific ; or (c) the two disturbances 
were due to some common influence. 
The fact that Jarge earthquakes so frequently occur in pairs or groups 
precludes the idea that these short intervals between megaseismic effects 
are merely matters of chance. 
