摘要
运用单因子方差分析、聚类分析、主成分分析和判别分析对中国沿海的日本蟳7个地理群体进行形态学差异分析。单因子方差分析的结果表明日本蟳群体间的差异为不同地理种群的差异。聚类分析的结果显示日本蟳各群体大致按分布海区的不同而聚集,并且无与地理距离相关的明显聚类。主成分分析结果显示4个主成分的贡献率分别为47.639%、19.004%、12.270%、10.178%,其累计贡献率为89.090%。判别分析的结果表明,舟山群体与威海群体的判别率最低,均为40%,长江口群体的判别率最高,为73.3%,7个日本蟳地理群体的综合判别率为56.5%。根据判别分析获得两个典型判别函数,其贡献率分别为76.4%和14.2%,累计贡献率为90.7%,判别分析的结果与聚类分析结果基本一致。研究表明日本蟳渤海群体、黄海群体和东海群体间存在一定的形态学差异。
Morphological variation among seven geographic populations of Charybdis japonica was studied using one way-ANOVA analysis, principal component analysis, cluster analysis and discriminant analysis. One way-ANOVA analysis revealed the difference among C. japonica populations below geographical population levle. The cluster analysis showed that different C. japonica populations generally gathering in line with the distribution of different sea areas, and there was no apparent clustering related to the geographical distance. The principal component analysis resulted in four principal components, the contributory ratios of the four principal components were 47.639%, 19.004%, 12.270% and 10.178% respectively, and the cumulative con-tributory ratio was 89.090%. The discriminant analysis revealed that the discriminant accuracy of Weihai and Zhoushan population were the lowest and both were 40%, the discriminant accuracy of Yangtze estuary was the highest and was 73.3% , and the average discriminant accuracy was 56.5%. The contributory ratios of two＆amp;nbsp;canonical discriminant functions established by discriminant analysis were 76.4% and 14.2%, respectively, and the cumulative contributory ratio was 90.7%. The discriminant analysis was almost the same with cluster anal-ysis. This research showed that there were morphological differences between C. japonica populations in East China Sea, Bohai Sea and Yellow Sea.
出处
《浙江海洋学院学报（自然科学版）》
CAS
2013年第6期477-483,共7页
Journal of Zhejiang Ocean University(Natural Science Edition)
基金
国家自然科学基金项目(41006075)
关键词
日本蟳
形态学差异
单因子方差分析
主成分分析
聚类分析
判别分析
Charybdis japonica
morphological variation
one way-ANOVA analysis
principal componentanalysis
cluster analysis
discriminant analysis
作者简介
郑伟（1990-），男，湖南韶山人，硕士研究生，研究方向：渔业资源生物学．
通讯作者：韩志强．E-mail：d6339124@163.com
