The purpose of this study was to measure the absorption and intra-epidermal fate of 35S-radiolabelled sulphur mustard (35SM) in human breast skin in vitro. Skin (full-thickness or heat-separated epidermis) was placed into static diffusion cells and was exposed to droplets of liquid 35SM or saturated 35SM vapour. Amounts of 35SM penetrating the skin were measured from which skin absorption rates were calculated. Unbound radiolabel was washed from the surface, extracted from the skin and analysed to determine the identity of the radiolabelled species in order to measure the extent of hydrolysis of sulphur mustard. Penetration rates of liquid 35SM measured in vitro (71-294 microg cm(-2) h(-1)) were in agreement with those measured previously in vivo using human volunteers (60-240 microg cm(-2) h(-1)). Rates of liquid 35SM skin absorption under occluded, infinite dose conditions were highest through heat-separated epidermal membranes (294+/-58 microg cm(-2) h(-1)) and lowest through full-thickness skin (71+/-14 microg cm(-2) h(-1)). Fluxes of saturated 35SM vapour (110+/-75 microg cm(-2) h(-1)) through heat-separated membranes were similar to those previously measured through human forearm skin in vivo (162 microg cm(-2) h(-1)). Although hydrolysis of 35SM did occur, both on the surface and within the skin, it accounted for only a small percentage of the total applied dose (<2.7+/-1.2%). The difference in total amount of liquid 35SM penetrated between occluded and unoccluded conditions in vitro (79+/-14%) was similar to that lost as vapour from unoccluded skin in vivo (80%). A substantial reservoir of 35SM (14-36% of the applied dose) was measured within heat-separated epidermal membranes for up to 24 h which may have significant implications for the management of personnel exposed to sulphur mustard.