unfavourable to the existence of ice. I believe that the true explanation of the curious phenomena presented by these caves in general, is to be found in Deluc's theory, fortified by such facts as those which I have now stated. The mean temperature of the rock at Besançon, where the elevation above the sea is comparatively so small, renders the temptation to suggest some chemical cause very strong. The question of ice in summer where thaw prevails in winter, may fairly be considered to have been eliminated from the discussion of such caves as I have seen, in spite of the persistent assertions of some of the peasantry. The observations, however, in caverns in volcanic formations, and in basaltic débris, are so circumstantial that it is impossible to reject them; and in such cases a theory similar to that enunciated by Mr. Scrope[193] seems to be the only one in any way satisfactory, though I have not heard of such marvellous results being produced elsewhere by evaporation. One observer, for instance, of the cavern near the village of Both, in the Eiffel, found a thickness of 3 feet of ice; and in that case it was melting in summer, instead of forming. In some cases it has been suggested that the length of time required for external heat or cold to penetrate through the earth and rock which lie above the caves is sufficient to account for the phenomenon of summer frost and winter thaw. Thus, it is said, the thickness of the superincumbent bed may be such that the heat of summer only gets through to the cave at Christmas, and then produces thaw, while in like manner the greatest cold will reach the cave in mid-summer. But there is a fatal objection to this idea in the fact that the invariable stratum--i.e., the stratum beyond which the annual changes of external temperature are not felt--is reached about 60 feet below the surface in temperate latitudes,[194] while at the tropics such changes are not felt more than a foot below the surface. Humboldt calculated that in the latitude of central France the whole annual variation in temperature at a depth of 30 feet would not amount to more than one degree.[195] FOOTNOTES: