I am inclined to think that the cooling of water, in ancient times, of which I have already spoken, is not to be ascribed so much to the boiling as to the jars being kept continually wet, and to the air to which it was exposed. A false opinion seems therefore to have prevailed respecting the cause; and because it was considered to be the boiling, many have not mentioned the real cause, which appeared to them only to afford a trifling assistance, though it has been remarked both by Galen and Athenæus. We know at present that coolness is produced by evaporation. A thermometer kept wet in the open air falls as long as evaporation continues[387]. With sulphuric æther, and still better with that of nitre, which evaporates very rapidly, water may be made to freeze even in the middle of summer; and Cavallo saw in summer a Fahrenheit’s thermometer, which stood at 64°, fall in two minutes, by means of æther, to +3, that is to 29° below the freezing-point[388]. On this principle depends the art of making ice at Calcutta and other parts of India, between 25° 30′ and 23° 30′ of north latitude, where natural ice is never seen unless imported. Trenches two feet deep, dug in an open plain, are strewed over with dry straw; and in these are placed small shallow unglazed earthen pans, filled with water at sunset. The ice which is produced in them is carried away before sunrise next morning, and conveyed to an ice-cellar fifteen feet deep; where it is carefully covered with straw to be preserved from the external heat and air. A great deal, in this process, depends upon the state of the atmosphere. When calm, pure and serene, it is most favourable to the congelation; but when the winds are variable, or the weather heavy and cloudy, no ice is formed; and the same is often the case when the nights are raw and cold[389]. It was once believed that this freezing was occasioned principally by the water having been boiled; but it seems to be owing much rather to evaporation[390]. It is not however said that the vessels are kept continually wet on the outside, but that they are unglazed, and so porous or little burnt, that the water oozes through them; and on that account their exterior surface appears always moist[391]. By vessels of this kind the trouble of wetting is saved. What has been said respecting the influence of the weather serves, in some measure, to confirm my conjecture. The more it favours evaporation, the ice is not only formed more easily, but it is better; and when evaporation is prevented by the wind or the weather, no ice is produced. The latest accounts how ice is made at Benares, say expressly that boiled water is not employed; and that all those vessels, the pores of which are stopped by having been used, do not yield ice so soon or so good. In porcelain vessels none is produced; and this is the case also when the straw is wet[392].