A leukotriene antagonist [Ro 23-3544; 6-acetyl-7-[5-(4-acetyl-3-hydroxy-2-propylphenoxy)pentyloxy] -3,4-dihydro-2H-1-benzopyran-2-carboxylic acid; 1] was studied in vitro for its permeation through and retention in hairless guinea pig skin from various topical vehicles. Both the free acid and the sodium salt forms of the drug were used. The vehicles evaluated were polyethylene glycol 400, propylene glycol, dimethyl sulfoxide (DMSO), C12-C15 alcohol lactates, dimethyl isosorbide, butyrolactone, methylpyrrolidone, hexyl laurate, isopropyl myristate, and caprylic/capric triglyceride (Neobee M5). For the salt form of the drug, the highest permeability coefficient and retention were obtained from DMSO and methylpyrrolidone, respectively. For the acid form, however, the highest permeability coefficient and retention were obtained from hexyl laurate and DMSO, respectively. The highest permeation and retention values were not obtained from the same vehicle for either the salt or the acid form of the drug. This observation questions the validity of using permeation (flux) measurements to screen topical drugs and formulations. Although the precise reasons for this lack of correlation between permeation and retention are not known at this time, this study has shown that the solubility parameters of the drug and the vehicles used may play an important role. It seems logical to conduct skin retention studies rather than flux measurements in evaluating drug delivery from dermatological products.