That the ice is not formed in apparently static caves, by movements of air producing evaporation, seems to me proved by what I have noticed in regard to the atmosphere. The dryness or moistness of the air within a glacière cave is coincident with the state of freezing or thawing of the cave. When I have visited a large cave in June, everything was frozen tight, there were no drips nor mushy ice, the air was relatively dry and the sensation of cold not unpleasant. When I have visited a cave in August, the ice was soft and mushy, water was dripping from the roof, the atmosphere was moist, and the cold penetrating. It seems to me that the facts go to show that it is not evaporation which forms the ice, but the melting of the ice which fills the cave with moisture. If there are any draughts or movements of the atmosphere when above freezing point, then their tendency is to vaporize the ice. The process of the formation of ice in relation to the atmosphere is as follows: the cold air permeates the cave and freezes up all the drip: the atmosphere becomes dry: gradually warmer air gets in and the ice begins to melt: then the atmosphere gets charged slowly with the vapor of the melting ice. This process is the exact opposite of the formation of ice by evaporation; it is the atmosphere which is made humid by the vaporizing of the ice, and by the drip. When the air is thoroughly saturated with the vapor, being scarcely renewed from outside and but a few degrees above freezing point, it undoubtedly retards evaporation, acts like a blanket and lessens the rate of melting of the ice.[63] [63] See Part IV.: Thury, page 285; Fugger, page 296; Trouillet, page 298.