Control of hardwoods is vitally important in the establishment of Douglas-fir seedlings on Pacific Northwest forest lands. In reducing the competition of these brushy species, herbicides have become nearly an indispensable tool. However, with the increasing awareness of the importance of a clean environment, it has become apparent that the safe use of these chemicals requires a knowledge of their behavior, particularly with respect to water contamination (NORRIS 1971). 2,4,5-T (2,4,5-trichlorophenoxyacetic acid) 2 —/ is a particularly effective brush control agent, but concern about its possible hazards has caused foresters to consider alternative chemicals. Dicamba (3,6-dichloro-o-anisic acid) is an auxin-type herbicide that has been suggested as a possible replacement for 2,4,5-T. Dicamba is both soil and foliage active and is relatively low in toxicity (SWANSON 1969; NEWTON 1972; BRADY 1971). However, before dicamba can be extensively used for brush control, its behavior in various parts of the forest environment must be determined. In this study, we measured the concentrations of dicamba in streams flowing from forest lands which had received aerial application of herbicide. Herbicide can enter forest streams through direct application to stream surfaces, in overland flow during periods of intense precipitation, or by leaching through the soil profile. Direct application to stream surfaces can be minimized through use of buffer strips between treated areas and streams (NORRIS 1967). The probability of overland flow or leaching of herbicide to streams depends largely on the persistence and movement chafacteristics of the chemical on and in the soil.