The use which dealers and farmers make of these weather reports is, however, very far from having been reduced to science. Some of these persons, it is true, are frequently able, by a purely instinctive process of deduction, to make successful forecasts of crop yields from a close study of the weather, and others have worked out crude rules of their own for the same purpose; but the agricultural meteorologist approaches the problem in a different way. Immense progress has been made in the past decade in applying the mathematical theory of _correlation_ to this problem. This branch of mathematics, originally developed chiefly for statistical studies in biology by Galton, Pearson, and others, is now extensively used by meteorologists not only for studying the effects of weather on crops, but also for finding out what correspondences or relationships exist between variations of weather in different parts of the world, as well as between weather and sun spots, weather and vital statistics, etc. Sometimes, when the farmers do not disagree on the subject of favorable and unfavorable weather for the crops, they hold opinions in common that agricultural meteorology is unable to substantiate. An illustration is found in the idea that a good covering of snow during the winter is favorable to the yield of winter wheat. Apparently this is one of the host of popular ideas that are based merely on the delusive foundation of “everybody says so.” Smith has investigated the statistics of wheat for Ohio and C. J. Root those for Illinois. In both cases their results negative the prevailing opinion. Professor Smith finds “some evidence to indicate that wheat has a better prospect if it is not covered by snow during the month of January,” while Mr. Root states that, in general, “the winters of light snowfall are followed by good wheat yields and the winters of heavy snowfall by light yields.”