In percutaneous penetration in vitro techniques, excised full-thickness skin with its stratum corneum, viable epidermis and dermis is often used. Since penetrants can be absorbed in vivo immediately below the viable epidermis, the dermal layer could act as an additional barrier in the in vitro experiments relative to the actual in vivo process. In the present paper, in vitro penetration studies through excised hairless rat skin devoid of its dermal layer are reported and compared with those previously carried out with the excised full-thickness skin of the same animal. A homologous series of 4-alkylanilines was used in all the studies, and the correlations found between permeability coefficients and n-octanol partition coefficients were analyzed. Correlations are bilinear in nature in both cases, but in the absence of the dermal layer the correlation line seems to tend to hyperbolicity, as assessed by a significant increment in permeability coefficients for the highly lipophilic compounds of the series and by a displacement of the optimal lipophilicity value (vertex of the correlation line) to a higher partition coefficient. It can be concluded that the heterogeneous nature of the skin, as far as absorption is concerned, may be due to the presence of the two anatomical hydrophilic layers, dermis and viable epidermis, rather than to the stratum corneum itself. A critical review of the results reported in the literature showed good agreement with these conclusions. The biophysical penetration model was identical and the optimal lipophilicity values very similar, so it may be that these features are independent of the type of epidermis used (rat, mouse or man) and also of the chemical composition of the penetrants.