An in vivo fish preparation was used that allowed a direct measure of the transport rates of 14 different organic chemicals across the gills of rainbow trout (Salmo gairdneri). The chemicals, all C14 labeled, were selected from five classes, encompassing a range of octanol-water partition coefficient (log P) values, from 0.23 (ethyl formate) to 7.5 (mirex). The uptake efficiency (extraction efficiency) of each chemical was determined by monitoring the inspired and expired water of trout exposed to each chemical over an exposure period of 1 to 6 hr. The mean gill extraction efficiency for all chemicals tested varied from a low of 7% to a high of 60%, extracted in a single pall of the chemical across the gills. The extraction efficiency of chemicals with log P or 1 or less were low and showed no relationship to log P. These low extraction efficiencies seen at log P of 1 and below with molecular weights below 100 were indicative of aqueous pore transport. The mean extraction efficiency for chemicals with log P values of 1 to 3 seemed to vary directly with log P, to a maximum of slightly greater than 60%, suggesting that uptake was controlled by the lipid membrane. The mean extraction efficiency for chemicals with log P of 3 to 6 was independent of log P and remained at 60%, which suggested that gill uptake was controlled by aqueous diffusion rates rather than gill membrane permeability. The mean extraction efficiency with mirex (log P = 7.5) decreased to 20%. This behavior was consistent with previous physical models which predict that high log P and melting point (low water solubility) may substantially reduce the rate of accumulation in fish. These studies support the passive diffusion model for the uptake of organics at the gill-water interface.