In order to study the dermal permeation of compounds through the skin, an in vitro model was developed which utilized pig ears perfused with autologous pig blood (de Lange, J., van Eck, P., Elliott, G. R., de Kort, W. L. A. M., and Wolthuis, O. L. (1992). J. Pharmacol. Toxicol. Methods 27, 71-77). In the present article we investigated to what extent the rate of permeation of xylene through pig ear skin is dependent on the perfusion medium used. Pig ears were exposed to xylene (10 cm(2) area) for a 4-hr period (30 degrees C, relative humidity of 40-60%) and the perfusate was analyzed for xylene using gas chromatography. The rates of permeation of xylene for whole blood, blood depleted of white blood cells, and a buffer containing 4.5% albumin were similar (+/-300 ng/min/cm(2)). The rate of penetration was fivefold higher when pig plasma was used and ninefold lower when albumin was excluded from the buffer. Using the buffer, we found that the rate of permeation of xylene was proportional to flow (constant protein concentration) and protein concentration (constant flow). Our data demonstrate that the measured permeation rate for xylene is, to a large degree, dependent on the effective protein concentration (mg/min) passing through the ear. Differences in this parameter could explain the variations in rates of permeation found using the different perfusion media. To avoid problems associated with the choice of receptor fluid for permeation experiments, we suggest that full blood remains the vehicle of choice, although the practical perfusion period is limited to about 6 hr. If longer perfusion periods are required, then it should be possible to reproduce results obtained with whole blood by choosing an appropriate buffer. (C) 1994 Academic Press, Inc.