The accepted wisdom is that earthquakes are entirely independent of the weather. Nonetheless, there has been a recurring suspicion over the centuries that atmospheric conditions may sometimes be a seismic catalyst - that changes in atmospheric pressure, for example, may sometimes trigger the beginning of a tremor.
Charles Darwin was once of the first to think about the matter seriously. In The Voyage of the Beagle he writes: "There appears much probability in the view that when the barometer is low, and when rain might naturally be expected to occur, the diminished pressure of the atmosphere over a wide extent of country might well determine the precise day on which the Earth, already stretched to the utmost by subterranean forces, should yield, crack, and consequently tremble."
The idea surfaced again in 1925 after two minor tremors in the north-east United States. A local expert, one R.W. Sayles, outlined his views in the Boston Herald of March 21st that year:
"A deficiency of rainfall and low barometric pressure were doubtless important factors in precipitating the earthquakes of January 7th and February 28th in New England, for anything which would tend to make the land lighter would increase the strain along the fault line.
"The quake of January 7th was preceded by three months of very dry weather with a deficiency of 8.1 inches of rainfall. Then during February there was a deficiency again. Just before the quake of 28th the lowest barometric pressure for two years, 28.96 inches, was recorded in Boston, and the low moved in a north-easterly direction across New England.
"The weight of eight inches of rainfall over all New England is computed to be 32,260,244,000 tons; a deficiency of eight inches of rainfall, therefore, is equivalent to taking an enormous weight off the region, and this, together with the extremely low barometric pressure, might well have been the set of circumstances which `set off' the shock of February 28th."
Then a year or two ago, a Japanese seismologist called Masakazu Outake, looking back over centuries of records, noticed that all 13 major earthquakes in Japan between 684 and 1946 occurred during the autumn and winter months - a seasonal bias against whose happening by chance, it seems, the odds are something like 1,000 to one.
Outake, however, also noted that the average monthly atmospheric pressure in the region is about 10 millibars higher between August and February than it is in the other half of the year: he believes this may have something to do with the frequency of serious tremors on the winter months.