A novel methodology for establishing a pharmacological dose-effect relationship of methyl nicotinate, hexyl nicotinate and nicotinic acid acting as peripheral vasodilators in the skin following topical application is investigated. This methodology involves the estimation of the unbound drug concentration in the aqueous compartment at the site of action in tissue, termed C(*), which was evaluated as the pertinent concentration responsible for the pharmacological effect. Blood capillaries next to the epidermis-dermis boundary were postulated to be the relevant site of action. C(*) was estimated from drug transport parameters for different layers of human cadaver skin determined in vitro. Immunohistochemical studies showed that the plane of separation of skin achieved by heat treatment was between the basal cells of the epidermis and the lamina lucida, confirming the integrity of the epidermis and the dermis used in the experiments. The permeation rate for epidermis increased drastically with increasing lipophilicity of the drug. Dermis permeability was roughly the same for all three compounds. The epidermis represented the major transport barrier in vitro for methyl nicotinate and nicotinic acid but not for hexyl nicotinate. The esters were metabolised to nicotinic acid during tissue permeation to an extent that was rather limited for the epidermis but very pronounced for the dermis. Nonspecific alpha-naphthylacetate-esterase activity was predominantly located in the dermis, which was in agreement with the metabolism results. The drugs were applied each at three different concentrations in vivo to the ventral forearm of healthy human volunteers and vasodilation was evaluated based on skin erythema which was quantified by measuring colour change of reflected light. Area under the curve of the change of colour co-ordinates as a function of time was used as a measure of pharmacological effect. The pharmacological effect of all three drugs was comparable when similar C(*) values were considered, even though the concentrations applied to the skin differed by orders of magnitude. The effect showed a strong positive dependence on C(*). Methyl and hexyl nicotinate showed identical, nearly sigmoidal effect/C(*)-profiles, while the profile for nicotinic acid was linear, suggesting a possible difference in the intrinsic pharmacological potency between the esters and the acid. These results demonstrate the validity of C(*) as the relevant drug concentration for the cutaneous pharmacological effect of the topically applied drugs and underline the usefulness of the presented methodology for establishing dose-response relationships in dermal therapy and expressing bioavailability.