Abstract: By the methods of phenotypic identification and SRAP makers amplification, the g enetic diversity of twenty-five local tea cultivars in Guangdong Provi nce and five contrastive cultivars from other regions was assessed and classifie d, and the phenotypic traits of the cultivars were clustered by Pearson correlati on and Farthest neighbor methods. The coefficient of variation of the phenotypic traits was averagely 3215%. Fine-hair had the highest coefficient of variati on (4241%), while the growth period of bud leaves had the smallest one (1852 %). Based on the cluster analysis of phenotypic traits, the test 30 tea cultivars co uld be clustered into 4 groups, 17 cultivars in the first group, 10 cultivars in the second group, 2 contrastive cultivars Yunnan-dayezhong and Lingyun-baimao ch a in the third group, and 1 contrastive cultivar Hainan-dayezhong in the fourth group. After the amplification with 21 SRAP primers, a total of 127 fragments we re detected, among which, 114 fragments were polymorphic, accounting for 8867% of the total. The amplified fragments and polymorphic fragments per primer combi nation were averagely 605 and 543, respectively. At the genetic distance of 0 39 cm, the tea cultivars could be classified into three groups A, B and C, and 8 333% of the cultivars were belonged to group A. At the genetic distance of 0.3 1 cm, group A could be further classified into three sub-groups Ⅰ, Ⅱ and Ⅲ, 13 cultivars in subgroup Ⅰ, 2 cultivars in subgroupⅡ, and 10 cultivars in subgro up Ⅲ. It was not exactly the same between the clustering based on SRAP markers amplification and the performance of phenotypic traits.

Key words: offshore ecosystem health, assessment index, assessment method.