The availability of a drug from a topical preparation is dependent on many factors, one of the most important being the composition of the vehicle. Glycerol, propylene glycol and polyethylene glycol 200 are widely used as ingredients in topical formulations. The aim of this study was to examine how these glycols affected some of the fundamental factors involved in percutaneous absorption of ethyl- methyl- and glycol salicylate. To do this, certain drug-vehicle, drug-vehicle-skin and vehicle-skin interactions were investigated. Drug-vehicle interactions were evaluated using solubility - and rheological measurements, equilibrium dialysis, diffusion coefficient - and release rate determinations. In particular, the study showed that it is important to consider the viscosity contribution of a cosolvent to the vehicle and that it is essential not to over solubilise a drug in the vehicle. The evaluation of drug-vehicle-skin interactions involved both in vitro and in vivo determinations. The in vitro study consisted of solubility- and partition coefficient measurements together with determination of diffusion coefficients and penetration rates for the drugs and the glycols using a two compartment cell in which nude mouse skin was the rate controlling barrier. In vivo, the concentrations of the glycols were determined in the individual layers in the skin and in the plasma of nude mice. The flux of the salicylates was largely unaffected by the various solvent concentrations and the different solvents, except when high concentrations of propylene glycol and glycerol were employed. The measurements of glycerol and propylene glycol in the skin and plasma suggested that the amount of the glycols penetrating into the skin from a topical application were unlikely to have any effect on the partition coefficient of a drug between the vehicle and the skin or the diffusivity of the drug in the skin phase. A comparison of the in vivo and in vitro results indicated a good correlation between these studies. In vivo and in vitro histological assessment were employed to evaluate vehicle-skin effects. Applications of glycerol and PEG 200 had no effect on the skin, whereas increasing concentrations of propylene glycol caused progressive disintegration of the stratum corneum. Some `penetration enhancers' showed unacceptable levels of skin damage and/or irritancy. Future studies may correlate these findings with their penetration enhancing properties.