ON THE FACTS AND THEORY OF EARTHQUAKE PHENOMENA. 85 
The great earthquake of 1819, which extended its influence right across 
this peninsula from Calcutta to Cutch, and during which the Ullah Bund was 
elevated, and the Runn of Cutch submerged—the former a low mass of sand 
and clay seventy miles long, about fifteen miles wide, and elevated about 
10 feet ; and the latter an area of subsidence of about 2000 square miles— 
had a great general line of horizontal propagation of shock, as shown by the 
heavy red line, of nearly from W. to E., a few degrees to the S.E.; yet at 
Calcutta it was felt from N.E. to S.W., and at many places along this immense 
line—situated between the Aravulla and Vindhya chains of mountains, as 
for example at Rampura—the great shock was felt in directions quite trans- 
verse to the principal line. 
So also the general line of horizontal direction of the great earthquake of 
1833, whose origin was far beneath the Himalayas to the E. and N., had a 
great general direction about that shown by the long red arrow line. At 
Katmandu, in the mountains, the shocks were more directly E. to W., and 
also (reflected shocks probably) from the ranges to the N., which had a 
direction nearly N.E. to S.W., while in the great plain of the Ganges the 
observed directions were various, and, without a more complete knowledge of 
the geology and surface-configuration of the country, perfectly unanalysable, 
in some places N. to S., and at others, sixty miles off, from E. to W. 
While we must regard many of these observations as deserving of little 
stress as to accuracy, enough remains to prove that perturbations in the 
main directions of emergence at the surface of the normal earth-wave, due 
to heterogeneity of structure in depth, and to inequality of surface, prin- 
cipally, are of such a nature, as to render a special choice of district neces- 
sary in attempting any seismometrical researches (even with perfect instru- 
ments) which have in view the determination of the position of the focus 
of disturbance. This choice, according to our present knowledge, must be 
determined by the following conditions :— 
1. The whole surface-area of observation, and to as great a depth as 
possible, must be uniform in geological structure. 
If of stratified rock, not greatly shattered and overthrown, but 
(viewed largely) level or rolling only. The harder and more dense 
and elastic the formations, the better, but neither intersected by 
long and great dykes, nor by igneous protrusions of magnitude, nor 
suddenly bounded by such formations. 
2. The surface must not be broken up into deep gorges, and rocky ranges, 
and valleys. Seismometry, in a high and shattered mountainous 
country, can scarcely lead to any result but perplexity. If the surface 
be deeply alluvial all over, it is less objectionable than valley-basins, 
and pans of deep alluvium, with rocky ribs between them. 
3. The size of the area chosen for observation must bear a relation to the 
4 force of the shocks experienced in it. Moderate shocks are always 
4 best for observation, and, in large areas of the most uniform character 
fh of formation and surface, will give the most trustworthy indications. 
_ 4, If several seismometers be set up in the area, they should be all 
placed on corresponding formations, either all on rock, or all on deep 
alluvium. The rock, when attainable, is always to be preferred. 
Three seismometers, at as many distant stations, will be generally 
found sufficient, if the object be chiefly to seek the focal situation and 
depth. 
‘ 
Having now cleared the way by stating the difficulties of seismometric 
observations, Ist, as respects the instruments themselves, 2nd, as respects 
a 
