shape of hailstones is not impossibly due to the extremities on all sides being melted and to the fact that, as explained, particles either of moisture or of wind surround them evenly on all sides and in every quarter, when they freeze. '* Snow may be formed when a fine rain issues from the clouds because the pores are symmetrical and because of the continuous and violent pressure of the winds upon clouds which are suitable ; and then this rain has been frozen on its way because of some vToTent change to coldness in the regions below the clouds. Or again, by congelation in clouds which have uniform density a fall of snow might occur through the clouds which contain moisture being densely packed in close proximity to each other ; and these clouds produce a sort of compression and cause hail, and this happens mostly in spring. And when frozen clouds rub against each other, this accumulation of snow might be thrown off. And there are other ways in which snow might be formed r Dew is formed when such particles as are capable of producing this sort of moisture meet each other from the air : again by their rising from moist and damp places, the sort of place where dew is chiefly formed, and their subsequent coalescence, so as to create moisture and fall downwards, just as in several cases something similar is observed to take place under our eyes. And the formation of hoar-frost is not different from that of dew, certain particles of such a nature becoming in some such way congealed owing to a certain condition of cold air.