Fire has always been an important factor in the ecological complex of grasslands, with man credited as the primary source of grassland fires (Sauer, 1944). Indians often used fire to attract game to the fresh grass or to drive game to the hunters (Hough, 1926). The entry of civilized man into the prairie has not changed the factors of the environment but has altered their emphasis and impact on the vegetation. Some arguments advanced for burning native pastures are weed control, reduction of patchy grazing, earlier grazing in the spring, and increased gains per animal (Anderson, 1946). Experimentally only the last reason has been proved completely valid (Smith et al., 1961), but while range burning may increase gains per head, it is likely to reduce forage yields and thereby increase the acreage required for each animal. This series of studies was initiated in 1928 by Aldous (1934), who reported the results from the first six years in Kansas Agricultural Experiment Station Technical Bulletin 38. The experimental area, located on the college pasture a mile north of the campus, was composed of two sets of five plots, one set burned annually and one biennially. The treatments were winter burned (about December 1), early spring burned (about March 20), midspring burned (about April 10), late spring burned (about May 1), and the unburned check. The range site in this nearly level area is an ordinary upland as described by Anderson and Fly (1955), with deep, friable soil. These studies were continued until wartime labor shortages forced their suspension in 1944. In 1953, the treatments were resumed, the plots formerly burned each two years now being burned annually along with those formerly burned each year. The forage yield data reported here were obtained only from the replication burned annually throughout the period. The unburned plot has produced the highest forage yield in 16 of the 26 years, averaging 272 pounds more forage per acre than the late spring burned plot, the second high yielder. These differences were shown by Duncan's multiple range test to be significant (Table 1). The late spring burned plot and the midspring burned plot were not significantly different in forage yield, but yield of the late spring burned plot was significantly higher than that of the early spring burned and winter burned plots.