The tissue-to-unbound plasma distribution coefficients (Kpus) of 14 rat tissues after iv administration of nine 5-n-alkyl-5-ethyl barbituric acids, determined in a previous study, were used to identify a model of the relationship between tissue distribution and lipophilicity of the homologs, expressed in terms of their octanol to water partition ratio, P. Based on mechanistic considerations and assumptions, the parameter model was expressed as KpuT = fW, T[1 + aT(nPl, t)PbT], where fW, T is the tissue water content, (nPt, T) is the binding capacity of the tissue, n is the number of the binding sites, aT and bT are the parameters of the relationship KaT = aTPbT; and KaT is the binding association constant of each tissue. The parameter model was linearized and fitted to the predetermined Kpu values, yielding correlation coefficients ranging between .940 and .997. The predictive performance of the parameter model was evaluated using a leave-one-out procedure with subsequent computation of the mean prediction error (ME = measurement of the prediction bias) and the square root of the mean squared prediction error (RMSE = measurement of the prediction accuracy). The ME varied between −22.48 and 61.14%, indicating a slight tendency for overpredicting. The RMSE was between 24.73 and 102% for the individual tissues across the different homologs; and between 28.33 and 85.2% for the individual homologs across the different tissues. The apparently high Kpu prediction errors, when translated through the low sensitivity of the barbiturate whole-body physiologically based pharmacokinetic model, established previously, leads to predicted tissue concentration–time profiles within 5 to 20% of the original ones. Therefore, it is concluded, that the identified mechanistically based model is a good predictor of the tissue-to-unbound Kpus in the rat tissues.