To form subterranean ice, just as to form any other ice, two things are necessary: the first is cold, the second is water. Cold is supplied by the cold air of winter, and water must in some manner find its way into the cave while the cold air is there. The process is as follows: The cold air of winter sinks into and permeates the cave, and in course of time freezes up all the water which, in the shape of melting snow or cold winter rain or spring water, finds its way in; and once ice is formed it remains long after ice in the surrounding open country has melted away, because heat penetrates with difficulty into the cave. The only effect of the heat of summer is to melt the ice. The proofs, to my mind, of the truth of this view are: 1--Glacières are always found in parts of the world where, during part of the year at least, the temperatures of the surrounding country fall below freezing point. 2--All observations by reliable observers show that the temperatures of glacière caves vary, but in a much narrower thermometric scale, with those of the outside air: that the temperatures are lowest, and as a rule below freezing point, during the winter months; and that the temperatures are highest, and as a rule above freezing point, during the summer months. 3--Ice is never found far from the mouth of caves, but always near enough for the cold air to get in. 4--Evaporation, according to my observations, is, as in all other forms of ice in nature, connected mainly with the melting, not the freezing of the ice. _Geographical Distribution and Altitudes._--Glacière caves proper are found in various parts of Europe, Asia, and America, mostly in the smaller mountain ranges or in the outliers of the snowy mountain chains; generally in limestone and occasionally in basaltic formations. There are a good many in the Jura; a few in the Swiss and the Italian Alps; a number in the Eastern Alps of Tyrol and Carinthia. There are some in Hungary, several in Russia, one in Iceland, one on the Peak of Teneriffe, a number in Siberia, one in Kondooz in Central Asia, one in the Himálaya, one in Japan, and one in Korea. I have heard so far of over fifty glacières in North America, several of which are in Pennsylvania. From all over the world there are some three hundred places reported where subterranean ice is said to occur. This includes gorges, boulder heaps and freezing mines and wells, all of which exist in much the same localities as glacière caves. All the glacières which I know of, are situated in a latitude or at an altitude where ice and snow forms for part of the year in the surrounding open country. None are reported from India or Africa, or in fact from any low-lying places in tropical latitudes. Most of them are found in middle latitudes, and only where during part of the year, at least, there is a cold season, that is, where for some time the thermometer stands below freezing point.