be cut off by drought, the larger such capacity, the longer before, by the reduced supply of water, the tension of the steam shall arrive at its maximum. It may be a period of many months, or only one of a few days, or it may be insensible altogether, if nullified by a constant and abundant supply of sea water, in addition to that from rain, &c.
In very dry countries, such as Asia Minor and Syria, (where they have long popularly connected the occurrence of shocks, with derangements in their annual rains), it follows from what has been stated, that a continuous and great increase in the usual rainfall, may be equally productive of increased seismic action; so that the conditions may be such, that either great drought, or great wet, may probably herald earthquake.
In so far then as the supply of rain is related, to the barometric pressure, to the direction of the wind, and to the temperature of the air, at a given point of surface and of time, and in so far as the supply of rain reacts upon the temperature of the air, by lowering it; in so far I think we may consider it highly probable, that future and more extended observation in volcanic and seismic countries will establish this connection, between meteorology and seismology; and that in such regions, unusually dry or wet seasons, may be found, (under geognostic conditions such as have been sketched) to prelude the earthquake, and if so may even ultimately give the means of attempting the prediction of its return.
The point here brought into notice,(so far as I am aware of for the first time,) of a distinctly observable relation between earthquake and rainfall, though on too narrow a
