Subrenal capsule assay was performed on 35 fresh cancer tissue samples (13 stomach, 7 colon, 8 breast and 7 others) against MMC, 5-FU, ADM, CPA, MTX and cDDP employing normal immunocompetent ddY mice following Bogden's original method. Evaluability was 86% (30/35). Sensitivity was between 32 and 46% for each of the drugs except cDDP. Some organ-specific variations were noted between gastro-colic cancers and breast cancers with regard to chemosensitivity spectrum and also individual cases of the same histological type did not show a uniform one. Gastric cancer had the highest activity. The assay/clinical correlations were evaluable in 10 cases where 6 cases on active agents achieved two PR, two NC and PD, while all 4 cases on non-active agents showed PD. Histological analysis of implants, however, disclosed the following inherent problems: First, there was heterogeneity in tissue fragments due to the variable amounts of stroma; Second, apparent host reactions were seen in the untreated groups on the 6th day, accompanying marked damage to cancer cells. In contrast, in drug-treated groups, host reactions in most of them were suppressed and cancer cells were seen frequently in considerable amounts. This fact implies the mechanism that SRCA is based on, in spite of the xenograft immune; Third, there was some discordance between the macroscopic and microscopic findings. SRCA is assumed to be appliable to individual clinical cases up to a certain level which should be determined by histological analysis as well as by accumulation of clinically correlated cases.