Inductions of dicentric chromosome bridges in root-tip cells and of micronuclei in pollen tetrads by X rays and maleic hydrazide (MH) were studied in Tradescantia clone BNL 4430, in order to compare the accuracies of these endpoints for mutagenicity testings with that of somatic pink mutations in the stamen hairs of the same clone, the most accurate endpoint established in higher plant testers. The frequency of dicentric bridges in root-tip cells increased with X-ray doses with a slope of 1.249 on a log-log graph, the slope value being unexpectedly small (much smaller than the theoretical value of 2 for two-break events). The dose-response curve on a log-log graph for MH-induced dicentric bridges had a somewhat smaller slope of 1.188, and the bridges observed were predominantly of chromatid type, reflecting the nature of MH acting at the S period of cell cycle. As for X-ray-and MH-induced micronuclei in pollen tetrads, on the other hand, the dose-response curves could not be determined, excepting that for X-ray-induced micronuclei in pollen tetrads at earlier stage. Namely, the data obtained from pollen tetrads at later stage were not consistent with X-ray and MH doses, and those from pollen tetrads at earlier stage were also inconsistent with MH doses. These results were contrary to our expectation that the frequency of micronuclei would be more reliable in pollen tetrads at later stage than in those at earlier stage at least in clone BNL 4430, a hybrid clone. The micronucleus assay in pollen tetrads was also found to be unsuited for determining the effects of MH. Therefore, the assay of somatic mutations in Tradescantia stamen hairs is concluded to be the most accurate and most reliable mutagenicity test system with a high degree of reappearance.