Radish(Raphanus sativus L.) is one of important vegetables in China and the premature bolting is a destructive problem in radish production. In this study, 73 radish accessions were selected as materials, whose seeds were treated at 4 ℃ for 21 d, to identificate bolting tolerance of radish germplasm in the field. It resulted that 9 foreign materials still not bolting after growing for 136d, which were classified as the level of high tolerance to bolting. Among them, 7 accessions were Raphanus sativus L. var. longipinnatus Bailey originated from Korea(6)and Japan(1) respectively, 1 Japanese accession belonged to R. sativus L. var. niger, 1 Russian accession belonged to var.radiculuse pers. 7 indexes were used to evalutate bolting characteristics of 64 bolting accessions. Frequency distribution of all indexes accorded with normal distribution, stem height at bolting stage and stem height at flowering stage slightly biased to low value area. Significance analysis showed that 7 indexes were at highly significant level, indicating that bolting features of tested accessiones expressed significant difference. Correlation analysis resulted that budding date and flowering date were different in highly significant level. Flowering date could be selected to effectively evaluate the bolting tolerance of radish germplasm. Bolting speed reached a highly significant correlation level with stem height at flowering stage, stem height at budding stage and number of bolting days. Bolting speed could well refelect bolting ability of radish germplasm. The stem height at flowering stage reached a very significant correlation with number of bolting days and bolting speed, which could be used to evaluate bolting ability conveiently. To evaluate the tolerance to bolting of radish germplasm using principal component analysis and subordinate function method, radish germplasm were well separated. 2 radish germplasm were screened to be highly tolerant accordingly. Only 1 radish germplasm was screened to be susceptible to bolting as principal component analysis.