为探讨长江和珠江流域野生大眼鳜(Siniperca kneri)的外形差异, 本研究基于形态学和框架测量数据, 运用多元分析法比较了长江流域乌江沿河、锦江铜仁大眼鳜群体和珠江流域北盘江关岭大眼鳜群体的形态特征.结果表明, 铜仁段大眼鳜群体的眼最小, 沿河段大眼鳜群体的体型最薄, 而关岭段大眼鳜群体的体型最厚, 吻最短但尾柄最长.经数据标准化和参数选择后, 12个性状参数的数据被用来进行主成分分析(principal component analysis, PCA).主成分分析提取了2个主成分, 其累计贡献率为64.255%, 其中第1主成分主要受体宽、尾柄形态、眼间距等性状参数的影响, 而第2主成分主要受吻长和眼径大小的影响.应用逐步判别方法(discriminant function analysis, DFA)建立了这3个群体的特征判别函数, 其交互验证判别准确率为91.85%.在3个群体之间具有显著差异的12项性状的差异系数(coefficient of difference, CD)均未达到1.28这一亚种分化临界值.长江流域乌江沿河和锦江铜仁大眼鳜群体的外形较为相似, 而它们与珠江流域北盘江关岭大眼鳜群体的形态差异较大, 但本研究3个群体的形态变异仍为同一物种下的不同地理种群的形态变异, 该变异还没有达到亚种变异水平.%To analyze the morphometric differentiation of wild Siniperca kneri between the population from the Yangtze River Basin and that from the Pearl River Basin,the variation in morphological characteristics of S.kneri among three populations from three different sampling sections (Yanhe section of the Wujiang River, Tongren sec-tion of the Jinjiang River in the Yangtze River Basin, and Guanling section of the Beipanjiang River in the Pearl River Basin) were assessed using multivariate analysis methods based on the measured morphological data and the truss network. The results showed that among the three populations, the eyes of the individuals from the Tongren section were the smallest, the body width from the Yanhe section was the thinnest, and the population from the Guanling section had the thickest body width, the shortest snout length, but the longest caudal peduncle length. After standardization and parameter selection for original data, the data for 12 selected characters were used to conduct principal component analysis. Two principal components were constructed, which resulted in a cumulative contributory ratio of 64.255%. Principal component 1 was mainly affected by characters, including body width, caudal peduncle shape, and distance between the eyes, whereas principal component 2 was mainly affected by the length of the snout and size of the eyes. A stepwise discriminant method was used to establish the characteristic discriminant functions of the three stocks, which revealed that the total accuracy rate of discrimination by the cross-validated method was 91.85%. The coefficients of differences for the 12 characters among the three stocks did not reach the threshold value of 1.28, although the subspecies could be clearly discriminated. In short, the morphometric differentiation between the two populations from the Yanhe section and Tongren section in the Yangtze River Basin were low, whereas the morphological differences were higher between these two populations and the population from the Guanling section sampled from the Pearl River Basin. Nevertheless, the morphologi-cal variation of the three populations shown in this study represented morphological variation from different geo-graphic populations of the same species, which had not reached the level of subspecies variation.
展开▼