Difficulties in obtaining and using human skin have tempted many workers to employ animal membranes for percutaneous absorption studies. We have investigated the suitability of two species of snake (Elaphe obsoleta, Python molurus) for this purpose and compared our in vitro experimental results for human skin and for hairless mouse, a currently popular model. The effects of long-term hydration on the membranes were investigated over 8 d using tritiated water as a model permeant. The initial permeability coefficients of all the membranes were similar (0.74-2.2 X 10(-3) cm 2h-1). Although the human and squamate skins did not change significantly over the test period, the permeability of hairless mouse skin increased 37 times. The actions of typical enhancers on the permeabilities of the membranes to a model penetrant 5-fluorouracil (5-FU) were tested using 3% Azone in Tween 20/saline, propylene glycol (PG), 2% Azone in PG, and 5% oleic acid in PG. While the data from snake membranes tended to underestimate the enhancer effects, those from hairless mouse skin greatly overestimated the changes. None of the membranes was a completely reliable model for assessing human percutaneous absorption as modified by accelerants. Pretreatment with acetone did not significantly change the permeability of human or squamate skins to 5-FU, although that of hairless mouse increased twentyfold. An overall conclusion is that, wherever possible, human skin should be used in absorption studies and not hairless mouse or snake skin; otherwise, misleading results may be obtained.