In vitro dermal absorption studies with rat, hairless guinea pig and human skin were conducted employing Bronaugh flow-through cells for two commercial formulations of the herbicide, 2,4-dichlorophenoxyacetic acid dimethyl amine (2,4-D amine) and in vivo studies were conducted with rats. The Clean Crop (CC) and Wilbur-Ellis (WE) 2,4-D amine formulations gave similar absorption values, ranging from 21 to 30% in vitro in all three species, except for absorption from WE in guinea pig skin [16 +/- 4.7% (n = 4)], which was significantly less (Student's t-test; P < 0.05) than absorption from WE formulation for human skin [28 +/- 2.6% (n = 4)]. A profound 'washing-in' effect of the soap wash conducted at 24 hr post-treatment was observed for all three species in vitro, and this effect was much more pronounced than in previous studies with 2,4-D where more dilute topical doses were employed. Dermal absorption of both 2,4-D amine formulations was significantly less for rats tested in vivo (16-17%) than for the skin from the same rats tested in vitro (26-30%). The lower percentage absorption observed in vivo was thought to be due to 'rub-off' of the dose application, as 68-78% of the dose was detected in the foam rubber patches used to protect the dose site. The inability to demonstrate greater permeability of animal skin to skin from the one human subject tested, the observation of a soap 'washing-in' effect, and the apparent overprediction of field exposure by both the in vivo and in vitro data were all factors considered to have important implications for accurate assessment of occupational exposure to 2,4-D.