For two contiguous sections of a Montana stream, the agricultural use of the floodplain was related to cover, stream morphology, and fish populations. In one section the vegetation of the floodplain had been reduced by clearing and intensive livestock grazing; in the other section, which had received light use by livestock, vegetation was relatively unchanged. This ungrazed section had 76 percent more cover (undercut banks, debris, overhanging brush, and miscellaneous) per acre of stream than the grazed section. Brown trout (+ 6 inches) were estimated to be 27 percent more numerous and to weigh 44 percent more per acre in the ungrazed section of the stream, although their rate of growth was similar in the two stream sections. The importance of shelter for trout in streams has been recognized for many years (Needham 1938). Efforts to elucidate the relationship included the habitat improvement studies of Tarzwell (1937, 1938), Shetter et al. (1946), and Saunders and Smith (1962). Boussu (1954) found that when undercut banks and overhanging brush of a stream were removed, trout populations, especially the larger fish, were adversely affected. Young beaver ponds with heavy willow cover and deep water generally support larger fish than associated stream areas (Rutherford 1955, Huey and Wolfrum 1956, Gard 1961, Knudsen 1962). Wamer and Porter (1960) stated that the removal of bank vegetation, overhanging banks, and other shelter destroyed some of Maine's finest trout streams. Loss of woody vegetation on stream banks and adjacent floodplain through land clearing and livestock grazing in Montana finally resulted in damage to the Rock Creek floodplain by a 1957 flood (Nelson and Hill 1960). The objective of the present study was 1 Study supported by the Montana Cooperative Fisheries Unit in conjunction with the Montana Fish and Game Department. 2 Present address: Washington Department of Fisheries, Fisheries Center, University of Washington, Seattle. to compare the stream morphology, bank cover, and fish populations of two areasone where the floodplain vegetation was reduced by clearing and grazing and the other where the vegetation was relatively unaltered. I wish to thank Dr. R. J. Graham who directed the study and assisted in preparing the manuscript. Donald Bianchi, who suggested the study, also gave valuable field assistance, as did Clint Bishop, Kenneth Tuinstra, William Gould, and several employees of the Missouri River Basin studies.