cember, when the monthly means and the mean minima differ very little, but the mean maxima of Nukurs are higher by 7° and the absolute maxima by more than 10°. This is because snow rarely falls at Nukurs, and a covering lasting for several days seldom occurs. In the absence of snow there is nothing to interfere with the action of warm winds, and, in that latitude, the sun heats the ground sensibly, even in the middle of winter.
The influence of a bed of snow on the maxima should vary according as the temperature is below or above the freezing-point. In the latter case, the melting of the snow, involving the absorption of heat, would tend to prevent increase of temperature. Not only does it prevent high maxima, but it keeps the temperature near the freezing-point long after it has begun. This is why April is colder than October in Russia, Central Europe, Canada, and the Northern United States. Farther north, for the same cause, the mean temperature of May is below that of September. It can not be doubted that this cause of cold, or rather this conversion of heat into melting action, is proportional, all other things being equal, to the mass of snow that remains on the ground. Hence, in countries having cold winters, the principal obstacle to the rise of temperature in spring is found in the quantity of snow on the ground, and not in the previous cold of the winter. It is only in countries situated near the sea, or in the neighborhood of large frozen lakes, that the mean temperature of the winter has much influence on the depression of temperature in the following spring and summer, because a larger quantity of ice is found in such situations during cold seasons, and a greater sum of heat is necessary to make the melting complete than in average seasons. Snow is the only cause that can produce an analogous result in countries distant from seas and lakes. It may be concluded, from the observations on these subjects, that, at the moment when the mean temperature begins to rise above the freezing-point, everything depends upon the sum of cold existing in the form of snow and ice. The greater it is, the slower and more irregular will be the rise of temperature.
The time of the coming of the snow, its depth, and its extent, have also a very great influence on the beginning and duration of the winter frosts; and this influence is manifested, not only in the particular spot, but in the northern hemisphere far to the south of it. In short, we may say that snow gives duration to the cold, and prevents a rapid rise of temperature. If we knew the exact moment when a bed of snow was formed in the North in fall and winter, and if we could announce it by telegraph, we might predict the time of the freezing of rivers and canals, and thus serve the interests of vast territories as of most of Asiatic and European Russia, Scandinavia, British America, and the United States. The announcement of the closing of rivers by ice, even if it were only four or five days in advance, would prevent the considerable losses caused by premature
