If events at the beginning of the last glacial period took place in accordance with the cyclonic hypothesis, as outlined above, one of the inevitable results would be the production of snowfields. The places where snow would accumulate in special quantities would be central Canada, the Labrador plateau, and Scandinavia, as well as certain mountain regions. As soon as a snowfield became somewhat extensive, it would begin to produce striking climatic alterations in addition to those to which it owed its origin.[41] For example, within a snowfield the summers remain relatively cold. Hence such a field is likely to be an area of high pressure at all seasons. The fact that the snowfield is always a place of relatively high pressure results in outblowing surface winds except when these are temporarily overcome by the passage of strong cyclonic storms. The storms, however, tend to be concentrated near the margins of the ice throughout the year instead of following different paths in each of the four seasons. This is partly because cyclonic lows always avoid places of high pressure and are thus pushed out of the areas where permanent snow has accumulated. On the other hand, at times of many sunspots, as Kullmer has shown, the main storm track tends to be drawn poleward, perhaps by electrical conditions. Hence when a snowfield is present in the north, the lows, instead of migrating much farther north in summer than in winter, as they now do, would merely crowd on to the snowfield a little farther in summer than in winter. Thus the heavy precipitation which is usual in humid climates near the centers of lows would take place near the advancing margin of the snowfield and cause the field to expand still farther southward. The tendency toward the accumulation of snow on the margins of the snowfields would be intensified not only by the actual storms themselves, but by other conditions. For example, the coldness of the snow would tend to cause prompt condensation of the moisture brought by the winds that blow toward the storm centers from low latitudes. Again, in spite of the general dryness of the air over a snowfield, the lower air contains some moisture due to evaporation from the snow by day during the clear sunny weather of anti-cyclones or highs. Where this is sufficient, the cold surface of the snowfields tends to produce a frozen fog whenever the snowfield is cooled by radiation, as happens at night and during the passage of highs. Such a frozen fog is an effective reflector of solar radiation. Moreover, because ice has only half the specific heat of water, and is much more transparent to heat, such a "radiation fog" composed of ice crystals is a much less effective retainer of heat than clouds or fog made of unfrozen water particles. Shallow fogs of this type are described by several polar expeditions. They clearly retard the melting of the snow and thus help the icefield to grow.