The effects of three vehicles, phosphate-buffered saline (PBS), ethanol (50% in PBS w/w) and propylene glycol (50% in PBS w/w) on in vitro transdermal penetration of testosterone was investigated in the horse. Skin was harvested from the thorax of five Thoroughbred horses after euthanasia and stored at -20 degrees C until required. The skin was then defrosted and placed into Franz-type diffusion cells, which were maintained at approximately 32 degrees C by a water bath. Saturated solutions of testosterone, containing trace amounts of radiolabelled [14C]testosterone, in each vehicle were applied to the outer (stratum corneum) surface of each skin sample and aliquots of receptor fluid were collected at 0, 2, 4, 8, 16, 20, 22 and 24 h and analysed for testosterone by scintillation counting. The maximum flux (Jmax) of testosterone was significantly higher for all sites when testosterone was dissolved in a vehicle containing 50% ethanol or 50% propylene glycol, compared to PBS. In contrast, higher residues of testosterone were found remaining within the skin when PBS was used as a vehicle. This study shows that variability in clinical response to testosterone could be expected with formulation design.