摘要
用8个形态性状比较我国南海4个野生青蛤群体间的形态变异。聚类分析表明,湛江群体和汕头群体的形态较接近,海口群体的趋异最大。主成分分析表明,主成分1、2、3的贡献率分别为26.441%、21.104%、15.884%,累计贡献率为63.428%。逐步判别分析结果显示,4个群体形态差异显著(P<0.01)。建立了4群体判别函数,其判别准确率P1为51.4%～71.4%,判别分析准确率P2为60%～75.8%,4群体的综合判别率为64.3%。Mantel-test检验结果表明,我国南海青蛤不同地理群体的欧式距离与地理距离没有明显的相关性(r=0.4227,P=0.3910)。
Based on eight morphological characters of four populations of Cyclina sinensis from the South China Sea by means of multivariate morphometrics,the morphological variations of the four populations were studied.The results of Cluster analysis and Principal component analysis showed that clams from Zhanjiang and Shantou population were relatively similar in morphology,whereas Haikou population differed greatly from other populations in morphology.Through principal component analysis,three principal components were established,with their contributory ratio being 26.441%,21.104% and 15.884% respectively,and the cumulative contributory ratio being 63.428%.With the first principal component,the two dominant factors were the value of lBD /lAB and m0 /lAB,their contributory ratio being 63.7% and 62.7%;With the second principal component,the two dominant factors were the value of lAD /lAB and lEF /lAB,their contributory ratio being 71.7% and 68.3%;With the third principal component,the dominant factor was the value of lGK /lAB,its contributory ratio being 64.5%.The stepwise discrimination analysis revealed that the four populations differed significantly in morphology（P〈 0.01）.Then the discrimination functions of the four populations were set up,with the discrimination accuracy being 51.4%-71.4% for P1 and 60%-75.8% for P2,and the integrative discrimination accuracy was 64.3%.The result of Mantle-test showed that there was no obvious correlation between euclidian distance and geographical distance（r = 0.422 7,P = 0.391 0）.
出处
《广东海洋大学学报》
CAS
2010年第3期1-5,共5页
Journal of Guangdong Ocean University
基金
广东省海洋与渔业局科技推广专项项目(0909006)
关键词
青蛤
地理群体
形态差异
主成分分析
聚类分析
判别分析
Cyclina sinensis
geographical populations
morphological variation
cluster analysis
principal component analysis
multivariate discrimination analysis
