Analysis of Morphological Difference among Three Wild Populations of Perna viridis
Yuewen Deng
Abstract
Yuewen Deng
Abstract
Based on 5 morphological proportional characters, four multivariation analysis methods (difference coefficient test, principal component analysis, discriminant analysis and cluster analysis) were used to make comparative investigation of the morphological difference among three wild populations of Perna viridis in Bei Hai (BH), Zhan Jiang (ZJ) and SanWei (SW). The results showed that: (1) The values of difference coefficient test were from 0.024 to 1.134, showing that the populations have not yet reached the discriminant level of different subspecies. (2) In the principal component analysis, three principal components were constructed. The contributory ratio of the PC1, PC2 and PC3 was 40.134%, 27.956% and 20.614%, respectively and the cumulative contributory ratio was 88.704 %. (3) The stepwise discrimination analysis revealed that the three populations differed significantly in morphology. Then the discrimination functions of the three populations were set up, the discrimination accuracy being 73.0%-88.0% for P1 and 74.6%-86.8% for P2, the integrative discrimination accuracy being 79.7%. (4) The results of cluster analysis indicated that the morphological characters were similar between the populations of BH and SW, but quite different from that of ZJ. In this study, all the results have revealed some morphological difference among the three populations, which resulted from the joint effect of genetic components and environment.
OpenAlex reports 1 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Based on 5 morphological proportional characters, four multivariation analysis methods (difference coefficient test, principal component analysis, discriminant analysis and cluster analysis) were used to make comparative investigation of the morphological difference among three wild populations of Perna viridis in Bei Hai (BH), Zhan Jiang (ZJ) and SanWei (SW). The results showed that: (1) The values of difference coefficient test were from 0.024 to 1.134, showing that the populations have not yet reached the discriminant level of different subspecies. (2) In the principal component analysis, three principal components were constructed. The contributory ratio of the PC1, PC2 and PC3 was 40.134%, 27.956% and 20.614%, respectively and the cumulative contributory ratio was 88.704 %. (3) The stepwise discrimination analysis revealed that the three populations differed significantly in morphology. Then the discrimination functions of the three populations were set up, the discrimination accuracy being 73.0%-88.0% for P1 and 74.6%-86.8% for P2, the integrative discrimination accuracy being 79.7%. (4) The results of cluster analysis indicated that the morphological characters were similar between the populations of BH and SW, but quite different from that of ZJ. In this study, all the results have revealed some morphological difference among the three populations, which resulted from the joint effect of genetic components and environment.
Key concepts: Principal component analysis, Subspecies, Biology, Linear discriminant analysis, Morphological analysis, Significant difference, Perna viridis, Zoology