plete crystals, and there may not be half so many. The greater number of perfect crystals is found in widespread storms, or blizzards, while the local storms produce most often granular or imperfect forms. So marked is this distinction that very often the character and extent of a storm may be in general determined by an examination of the crystalline forms obtained. Extensive storms produce smaller crystals, more uniform in size, less clustered in flakes, and in greater variety than local storms. Figs. 1 to 20, inclusive, are crystals from general storms, while Figs. 23 and 24 are those of local storms. When the temperature is very low while a local storm is raging, its crystals resemble those of the blizzard more closely.
Some forms are common to both classes of storms. Probably because identical conditions do not occur frequently, the crystalline forms of each storm during a winter may differ from each other, one type appearing abundantly in one storm, a different type in the next, and so on. Conversely, the types most common in a given storm may reappear after an interval of months or years—as, for example, those obtained during the great blizzard of March, 1888, were repeated in the storms of February 16, 1892, and March 3, 1896, and most of these were of forms such as Figs. 13, 14, and 15, while unusual types, such as Figs. 7, 9, and 17, occurred in the storms of February 24, 1893, and February 13, 1894.
Not only do different storms afford different types of crystals, but different parts of the same storm, if it be general, give different forms. In this region, the northern and western portions of the storm area produce more perfect crystals than the southern and eastern, and from this we infer a difference in the atmospheric conditions in these portions, the former being more quiet and otherwise favorable to crystallization.
What has been called granular snow is shown in Figs. 4, 6, and 23. In this very common form we find only loose, irregular, subcrystalline forms, which are larger and heavier than others. This is formed in the middle or lower cloud layers, and when these are disturbed by wind, or otherwise rendered unsuitable for crystallization. Sometimes, perhaps always, these granular masses have nuclei of true crystals. Granular snow may explain the origin of the great raindrops which often fall during a thundershower. It is probable that such drops have a snow origin. Most, if not all, hailstones also originate in granular snow, as their thin, opaque centers and concentric rings of opaque, snowlike ice show.
The superiority of photography over drawing in securing details of structure may be readily seen if one compares any of the accompanying figures with the ordinary drawings of snowflakes, or even with the finest illustrations hitherto published. It is unfortunate
