ROOSTS OF WINTERING CROWS.--The collections of pellets from roosts of wintering crows in western Harvey County and northeastern Reno County differed in having a higher percentage of plant material. Sorghum, corn, and wheat predominated in early autumn, while sorghum, sunflower seed, and corn predominated in the winter. Ants were utilized to a much greater extent in early autumn. For grasshoppers and beetles, frequency of occurrence was high but percentages were low. Most of the standing water in the sand dune country had dried approximately one year before, and the aquatic component of the diet was almost entirely lacking. The two principal food items taken by crows in the winter of 1953 were grain sorghum and sunflower seed. Censuses in late November and late December, 1953, showed that feeding was mostly in harvested sorghum fields and corn fields, but alfalfa fields, wheat fields, plowed fields, and native pasture were also utilized. Sorghum and sunflower seeds were also the staple foods during December, 1952. Oats and wheat showed higher percentages than in 1953, perhaps because different foods were available in these two winters or because of differences in locality. The pellets collected in 1952 were from western Harvey County, whereas most of those collected in 1953 were from northeastern Reno County. The collection taken in February, 1954, showed a large percentage of oats in the diet. Newly sown oat fields were probably a major source of food at that time. Economic and Ecologic Significance The chief factors that determine the economic bearing of crows locally are: the yearly diet, the time of year in which each food item is taken, and fluctuation in the population density at different times of year. In the study here reported upon, the yearly diet was computed by averaging the percentages of each item determined for each biweekly period. Of the twenty-one collecting periods shown in the tables, six are overlapping pairs; that is to say, each includes one collection from eastern Harvey County and one from the western part of the study area. The average of these pairs was used in computing the yearly average. The yearly average is therefore based upon eighteen separate samples. The percentages are weighted toward the food items taken in summer and autumn, since many biweekly periods in late winter and early spring are not represented. Of the collecting periods represented, two were in spring, six were in summer, seven were in autumn, and three were in winter. Pellets collected at a number of different localities are averaged together as a percentage; consequently the figures obtainable do not necessarily represent the diet of any one group of crows. Nevertheless the percentages obtained by this method are perhaps valid as a general indication of the diet of the crows in this area.