The strip of country between the parallels of New York City and Richmond comprises the section wherein each winter storm is one large guess as to whether the precipitation is to be snow or rain. A compromise is usually affected in this way. Before the clouding up began the mercury may have stood at ten degrees below zero. As soon as the wind acquired an easterly slant the temperature increased. As it neared the freezing point the snow would begin, first in flakes of medium size which would enlarge until after a particularly heavy fall of a few minutes they would at once almost cease. Hail soon would succeed, the mercury still rising, and often the hail would have turned to rain before the freezing point of the air of the immediate surface of the earth had been reached, turning the snow already on the ground to slush and making a holiday for germs. One can always tell when this change to warmer is about to occur because the clouds which have been part and parcel with the obscuring snow suddenly show, not lighter but darker. The sudden increase in size of the flakes is another infallible symptom of increasing warmth in the atmosphere for each large flake is a compound of many smaller ones. When the temperature is low the flakes are very small, being grains and spicules in the severe blizzards of the west and falling as snow-dust in the Arctic. In the heavy storms of the guessing-belt the flakes are not necessarily small. I have noticed (in the latitude of Philadelphia) that our largest storms begin very leisurely indeed with small and regular-sized flakes. A quarter of an inch may not fall in the first hour. As the center nears the snow comes ever faster and larger, but not large, flakes are mixed with the original-sized flakes. Snow dust is apparent. At the height of the storm flakes of all sizes except the very large are falling, denoting great activity in the strata of air within the storm influence. In the ordinary storm an accumulation at the rate of an inch an hour denotes a storm of considerable intensity. The snow will likely keep on falling as long as the flakes are irregular in size. If they grow large and few or very small a cessation is likely, even though the wind is still blowing from an easterly quarter. The amount of snow likely to fall can be gauged not only by the process of flake-change but by the rate at which the wind rises. A storm's intensity is measured by the amount of wind. A storm can be a storm without a drop of rain or flake of snow if only there be enough wind. And as long as the wind in a snowstorm keeps rising the storm is likely to go on, probably increasing in volume of precipitation.