PROF. TYNDALL has several times called attention to a point in regard to the height of the snow-line, which seems to be steadily overlooked by those who speculate on the causes of the great prevalence of snow during the glacial epoch. It is of course well known that the height of the snow-line at any place is determined mainly by two things, viz., the depth of annual snow-fall, and the temperature of the place. If the amount of snow falling over the whole earth is to be increased, the evaporation must also be increased. (“Heat as a Mode of Motion,” Pp. 206–7. New York, 1866.) This would also raise the temperature, but the snow-line might nevertheless descend. We have a case of exactly this kind in the Himalayas. On the warm southern side of these mountains the snow-line is, nevertheless, 3,000ft. lower than on the northern side, where the temperature is very much colder. This is evidently due to a difference in the amount of annual precipitation. Assume that the sun was at one time much warmer than now, and that since then it has been steadily cooling; and I believe you have the key to the solution of the questions asked by J. H. Rohrs, as well as to such questions as the widespread occurrence of tropical vegetation during the past ages.