_While round dark crags imprison'd waters bend Through rifted ice, in ivory veins descend._ CANTO III. l. 113. The common heat of the interior parts of the earth being always 48 degrees, both in winter and summer, the snow which lies in contact with it is always in a thawing state; Hence in ice-houses the external parts of the collection of ice is perpetually thawing and thus preserves the internal part of it; so that it is necessary to lay up many tons for the preservation of one ton. Hence in Italy considerable rivers have their source from beneath the eternal glaciers, or mountains of snow and ice. In our country when the air in the course of a frost continues a day or two at very near 32 degrees, the common heat of the earth thaws the ice on its surface, while the thermometer remains at the freezing point. This circumstance is often observable in the rimy mornings of spring; the thermometer shall continue at the freezing point, yet all the rime will vanish, except that which happens to lie on a bridge, a board, or on a cake of cow-dung, which being thus as it were insulated or cut off from so free a communication with the common heat of the earth by means of the air under the bridge, or wood, or dung, which are bad conductors of heat, continues some time longer unthawed. Hence when the ground is covered thick with snow, though the frost continues, and the sun does not shine, yet the snow is observed to decrease very sensibly. For the common heat of the earth melts the under surface of it, and the upper one evaporates by its solution in the air. The great evaporation of ice was observed by Mr. Boyle, which experiment I repeated some time ago. Having suspended a piece of ice by a wire and weighed it with care without touching it with my hand, I hung it out the whole of a clear frosty night, and found in the morning it had lost nearly a fifth of its weight. Mr. N. Wallerius has since observed that ice at the time of its congelation evaporates faster than water in its fluid form; which may be accounted for from the heat given out at the instant of freezing; (Saussure's Essais sur Hygromet. p. 249.) but this effect is only momentary.