打开文本图片集
摘要：【目的】探討不同地理环境下大鲵的形态特点及利用多元统计分析法进行分析的可行性，找出不同大鲵种群间的主要形态差异，为大鲵群体形态学研究提供科学依据。【方法】选取野生大鲵、汉中大鲵、恩施大鲵及张家界大鲵各30尾，分别测量12个形态学特征参数，经校正后得到11个形态比例性状，然后采用主成分分析、聚类分析及判别分析等多元分析法对不同大鲵群体进行形态差异分析。【结果】主成分分析结果表明，前3个主成分（PC1、PC2和PC3）的特征值较大，累积方差贡献率为61.13%。其中，第一主成分（PC1）主要受X1（头长/全长）、X5（头宽/全长）和X7（体高/全长）的影响，第二主成分（PC2）主要受X11（前肢长/全长）的影响，第三主成分（PC3）主要受X3（尾长/全长）的影响。4个大鲵群体被分为两大类，其中，汉中大鲵先与恩施大鲵聚为一类，再与张家界大鲵聚为一大类;野生大鲵则单独聚为一类。采用建立的判别方程式对所有大鲵样本进行判别，结果发现4个大鲵群体的总判别准确率均为77.5%，其中，张家界大鲵的准确率最高，达83.3%，其次是野生大鲵（准确率为80.0%），汉中大鲵和恩施大鲵的准确率为73.3%。【结论】不同地域的大鲵种群在形态上具有一定差异，其中野生大鲵与养殖大鲵在头长/全长、头体长/全长、头高/全长和体高/全长4个形态比例性状上存在显著差异，通过判别分析可进行初步区分。
关键词： 大鲵;形态差异;主成分分析;聚类分析;判别分析
中图分类号： S966.6 文献标志码： A 文章编号：2095-1191（2019）12-2819-07
Morphological differences and discriminant analysis of Chinese giant salamander（Andriasd avidianus） in four different populations
XIE Yi-xing1， MAO Pan1， DENG Zhi-yong1， HE Ping1*， LI Jun-tao2，3*
（1Hunan Zhangjiajie Andriasd avidianus National Nature Reserve Management Office， Zhangjiajie， Hunan 427499， China; 2Institute of Tropical Bioscience and Biotechnology， Chinese Academy of Tropical Agricultural Sciences， Haikou 571101， China; 3Hainan Provincial Key Laboratory for Functional Components Research and Utilization of Marine Bio-resources， Haikou 571101， China）
Abstract：【Objective】To explore the morphological characteristics of giant salamander（Andriasd avidianus） under different geographical environments and the feasibility of multivariate statistical analysis， find out the main morphological differences among different populations， provide scientific basis for the study of population morphology of giant salamander in different geographical areas. 【Method】To select wild giant salamander，Hanzhong giant salamander， Enshi giant salamander and Zhangjiajie giant salamander 30 each， and 12 morphological characteristic parameters were measured and 11 morphological proportional characters were obtained after correction， then principal component analysis， cluster analysis and discriminant analysis were used to analyze the morphological differences of different populations of giant sa-lamander. 【Result】The results of principal component analysis showed that the first three principal components（PC1， PC2 and PC3） had larger eigenvalues and their cumulative variance contribution rate was 61.13%. The first principal component PC1 was mainly affected by X1 （head length/total length），X5 （head width/total length） and， X7 （body height/ total length）， the second principal component PC2 was mainly affected by X11 （forelimb length/total length） ， and the third principal component PC3 was mainly affected by X3 （tail length/total length）. The four groups were divided into two main groups， with the Hanzhong giant salamander and Ensch giant salamander into a group， and then with the Zhangjiajie giant salamander into a group， and the wild giant salamander was a separate group. All samples of giant salamanders were identified by using the established discriminant function. The results showed that the total discriminant accuracy of four groups of giant salamanders was 77.5%，of which， the accuracy of the Zhangjiajie giant salamander was the highest （83.3%） ， followed by the wild giant salamander（80.0%）， the Hanzhong and Ensch giant salamanders were 73.3% accuracy. 【Conclusion】The populations of giant salamanders from different regions have some differences in morphology， among which there are significant differences in head length/total length， head-body length/total length， head height/total length and body height/total length between wild and cultured giant salamanders， a preliminary distinction can be made by means of discriminant analysis.
