The subrenal capsule assay in normal immunocompetent mice was used to test the responsiveness of ovarian cancer to combination chemotherapy. Of the assays, 42 were of untreated tumors and 19 of previously treated tumors. Fifty-nine (97%) of the assays were evaluable. The previously treated tumors were less sensitive than the untreated ones. Of the treated tumors 44% were sensitive, 33% intermediately sensitive, and 22% resistant versus 56, 44, and 0%, respectively, of the untreated tumors (P less than .01). Repeat assays for the tumors of seven patients were performed successfully after five to eight courses of therapy with the combination of doxorubicin, cyclophosphamide, and cisplatin. The responsiveness to this combination had weakened significantly (P less than .01); the response of only one tumor remained unchanged. The rates of resistance to the drug combinations doxorubicin-cyclophosphamide-cisplatin, doxorubicin-cyclophosphamide-tegafur, and cyclophosphamide-vincristine were 11, 10, and 21%, respectively; there was, however, considerable interindividual variation in tumor responses to these combinations. Of other combinations, hexamethylmelamine combined with 4-epidoxorubicin, aclarubicin, or chlorambucil and cisplatin had effect, whereas the combinations of cisplatin and etoposide and of tegafur and methotrexate or mitomycin were quite ineffective, as measured by the assay. The reliability of the subrenal capsule assay in normal immunocompetent mice is discussed, and it is concluded that the assay can be used to assess the response of ovarian cancer to chemotherapy, including multidrug therapy, without routine histologic control.