The tissue distribution of amoxicillin in the bovine uterus in dependence of the administration route was examined using the in vitro model of the isolated hemoperfused bovine uterus. Four hours after intrauterine instillation or onset of perfusion with amoxicillin ("systemic" treatment), respectively, samples of the uterine tissue were taken from different locations of the organ. The tissue concentration of amoxicillin in the single layers of the uterine wall was determined via high performance liquid chromatography (HPLC) analysis. To compare the results with data achieved in vivo, healthy cows were treated post partum intrauterinely or systemically, respectively, and the concentration of amoxicillin was determined in endometrial biopsies. In vitro intrauterine instillation resulted in significantly higher levels of amoxicillin in the inner layers of the uterine wall compared to systemic treatment, whereas the drug is distributed more homogenously through all layers in the systemically treated organs. In the in vivo experiments the tissue concentration of amoxicillin is significantly lower after intrauterine instillation compared to the in vitro data. In contrast, the endometrial concentrations in the uteri perfused with amoxicillin in vitro are comparable to those measured in endometrium of systemically treated cows. The MIC for gram positive and negative bacteria was exceeded in the uterine wall of all organs in vitro and in vivo after intrauterine treatment. In systemic treated cows the MIC for gram negative bacteria was exceeded in two of three cases, but never in the hemoperfused uteri. The data indicate that combined systemic and intrauterine treatment results in the best distribution of sufficient antibiotic concentrations in the uterine tissue. The isolated hemoperfused bovine uterus is shown to be suitable as an in vitro model for pharmacokinetic examinations, especially with regard to detailed information on the tissue distribution of drugs.