35, 34, 35, 37Meretrix meretrixs were collected from the four sea areas of Beihai, Fangchenggang, Haikou and Zhanjiang, respectively, in the South China Sea Coast from October 2013 to March 2014. Shell contours were extracted and digitized by image processing, and converted to binary image, converted elliptic Fourier descriptor (EFDs) by the SHAPE. The EFDs were then identified by using principal component analysis (PCA) and gradual discrimination analysis (GDA). The result indicated that the three principal components (PCs) accounted for 80.67% of the total variation, in which the first PC accounted for 49.72%. The GDA of EFDs indicated that exactitude distinguishing accuracy of four geographical populations (GPs) were among 65.7%-94.6%, and the cross validation were among 60.0%-94.6%. The lowest accuracy of the four areas appeared when discriminating the population in Haikou while the highest appeared when discriminating the population in Zhanjiang. The average exactitude distinguishing accuracy is 78%. And the cross validation were among 60.0%-94.6%. The lowest and the highest distinguishing accuracy appeared when discriminating populations in Haikou and Zhanjiang respectively. The average exactitude distinguishing accuracy is 74.5%. It indicated that outline method was effective in discriminatingMeretrix meretrix from the four sea areas.%2013年10月至2014年3月在北海、防城港、海口和湛江4个海区分别采集35、34、35、37只文蛤样本，用外形轮廓法对各群体进行判别。通过 SHAPE 软件采集傅里叶函数解析出文蛤的外形轮廓参数，并对其进行主成分分析和判别分析。结果表明，前3个主成分共解释外形轮廓信息量的80.67%，其中第1主成分占49.72%。判别分析的逐步判定中，对4个海区文蛤的正确判定率为65.7%~94.6%，判定率最低的是海口群体，判定率最高的为湛江群体，平均正确判定率达到78%；交叉判定中，正确判定率为60.0%~94.6%，判定率最低为海口群体，判定率最高为湛江群体，平均正确判定率达为74.5%。研究认为，外形轮廓法可较好地判定4个海区的文蛤群体。
展开▼
