Both tarsocrural joints of 4 horses were inoculated with 1.5 X 10(5) colony-forming units of Staphylococcus aureus. On days 1, 3, and 6, each horse had one tarsocrural joint lavaged with a balanced electrolyte solution and had the contralateral tarsocrural joint lavaged with 0.1% povidone-iodine solution. All horses were orally administered trimethoprim (5 mg/kg)/sufadiazine (25 mg/kg) combination twice daily and phenylbutazone (2 g) once daily for the duration of the study (21 days). On days 0, 1, 3, 6, 9, 14, and 21, synovial fluid specimens were collected and analyzed for color, clarity, total protein concentration, WBC count and differential, and mucin clot-forming ability. Synovial fluid specimens collected on days 1, 3, 6, 9, 14, and 21 were bacteriologically cultured. On day 21, all horses were euthanatized, the tarsocrural joints were opened and examined, synovial membrane specimens were collected, bacteriologically cultured, and histologically evaluated, and articular cartilage specimens were histochemically evaluated. Repeated measures analysis of variance were used to evaluate differences between lavage solutions and among days for objective measurements. A paired t test was used to evaluate differences between solutions for the indices of synovial membrane inflammation and articular cartilage staining intensity with safranin-O-fast green. To be considered significant, the probability of a type-I error was less than 0.05. Significant differences were not found between joints lavaged with electrolyte solution vs povidone-iodine solution for synovial total protein concentration, WBC count, results of synovial fluid and membrane bacteriologic culture, synovial membrane inflammation, or articular cartilage glycosaminoglycan concentration.(ABSTRACT TRUNCATED AT 250 WORDS)