For this explanation to hold good, it is necessary that the level at which the ice is found should be below the level of the entrance to the cave; otherwise the mere weight of the cold air would cause it to leave its prison as soon as the spring warmth arrived. In every single case that has come under my observation, this condition has been emphatically fulfilled. It is necessary, also, that the cave should be protected from direct radiation, as the gravitation of cold air has nothing to do with resistance to that powerful means of introducing heat. This condition, also, is fulfilled by nature in all the glacières I have visited, excepting that of S. Georges; and there art has replaced the protection formerly afforded by the thick trees which grew over the hole of entrance. The effect of the second hole in the roof of this glacière is to destroy all the ice which is within range of the sun. A third and very necessary condition is, that the wind should not be allowed access to the cave; for if it were, it would infallibly bring in heated air, in spite of the specific weight of the cold air stored within. It will be understood from my descriptions of such glacières as that of the Grand Anu, of Monthézy, and the Lower Glacière of the Pré de S. Livres, how completely sheltered from all winds the entrances to those caves are. There can be no doubt, too, that the large surfaces which are available for evaporation have much to do with maintaining a somewhat lower temperature than the mean temperature of the place where the cave occurs. This had been noticed so long ago as Kircher's time; for among the answers which his questions received from the miners of Herrengrund, we find it stated that, so long as mines are dry, the deeper they are the hotter; but if they have water, they are less warm, however deep. From the mines of Schemnitz he was informed that, so long as the free passage of air was not hindered, the mines remained temperate; in other cases they were very warm. Another great advantage which some glacières possess must be borne in mind, namely, the collection of snow at the bottom of the pit in which the entrance lies. This snow absorbs, in the course of melting, all heat which strikes down by radiation or is driven down by accidental turns of the wind; and the snow-water thus forced into the cave will, at any rate, not seriously injure the ice. It is worthy of notice that the two caves which possess the greatest depth of ice, so far as I have been able to fathom it, are precisely those which have the greatest deposit of snow; and the ice in a third cave, that of Monthézy, which has likewise a large amount of snow in the entrance-pit, presents the appearance of very considerable depth. The Schafloch, it is true, which contains an immense bulk of ice, has no snow; but its elevation is great, as compared with that of some of the caves, and therefore the mean temperature of the rock in which it occurs is less