To compare levels of gamma-seminoprotein (gamma-Sm) assayed by original and revised assay systems, blood was obtained every 4 h over a 32-h period from 8 untreated prostate cancer patients. Serum levels of prostate specific antigen (PSA) were also examined. In 6 patients, the coefficient of variation (CV) of the serum levels assayed by the revised assay was significantly different from that of the intra-assay samples. In contrast, the CV of the gamma-Sm serum levels assayed by the original assay differed significantly from that of the intra-assay samples in only 2 patients. The fluctuations in gamma-Sm assayed by the revised assay were, at least in part, similar to those of the PSA serum levels in all patients. The mean CV of the gamma-Sm serum levels assayed by the revised assay was significantly larger than that for levels measured by the original assay. After treatment, the rate of decrease in gamma-Sm serum levels determined by the original assay differed from that in the serum levels of PSA and prostatic acid phosphatase. These results indicate that the original assay for gamma-Sm do not detect diurnal differences in serum gamma-Sm levels, even at levels below 20 ng/ml. These observations indicate that the analysis of data obtained using the original gamma-Sm kit should be interpreted with caution.