Two reaches of the North Fork of the Feather River in California were poisoned with rotenone to reduce nongame fish populations, one in 1977 and the other in 1981. Both reaches had been poisoned 10 years earlier as well. Standing crops, population structure, and growth rates were determined for rainbow trout (Salmo gairdneri, Sacramento sucker (Catostomus occidentalis), Sacramento squawfish (Ptychocheilns grandis), hardhead (Mylopharodon conocephalns), smallmouth bass (Micropterus dolomieni, and riffle sculpin (Cottns gnlosns) from fish killed in test sections. The nongame fish populations recovered from the previous poisoning operations, although the unusual age structure of the sucker population (skewed to large old individuals) in the 1977 section was attributed to the earlier treatment. In the reach treated in 1977, the rainbow trout population was small and the growth rates of smallmouth bass and nongame fishes were slow compared to populations in other streams. The section treated in 1981 had a substantial population of fast-growing wild rainbow trout that coexisted with a large sucker population, despite considerable angler pressure and repeated plants of hatchery rainbow trout. Differences between the two sections were attributed to differences in habitat combined with the effects of the severe 1976-1977 drought and not to interspecific competition. The drought caused greatly reduced flows and increased temperatures in 1977, producing conditions that probably stressed suckers and trout and favored squawfish, hardhead, and smallmouth bass. Because competition did not seem to be an important factor limiting the trout populations, habitat improvements, rather than poisoning operations, are recommended to improve wild trout fisheries in streams.