|
Morphological Classification of Wild Banana (Musa acuminate colla) by Classical Taxonomy and Cluster Analysis Technique in Lower Northern Thailand |
|---|---|
| รหัสดีโอไอ | |
| Creator | Det Wattanchaiyingcharoen |
| Title | Morphological Classification of Wild Banana (Musa acuminate colla) by Classical Taxonomy and Cluster Analysis Technique in Lower Northern Thailand |
| Contributor | Kumthorn Therakhupt, Surin Kangkun |
| Publisher | Department of Agriculture |
| Publication Year | 2547 |
| Journal Title | Thai Agricultural Research Journal |
| Journal Vol. | 22 |
| Journal No. | 3 |
| Page no. | 262-280 |
| Keyword | banana, classification, Musa acuminata Colla, morphology, cluster analysis, discriminant analysis. |
| URL Website | https://at.doa.go.th/journal |
| Website title | Thai Agricultural Research Journal |
| ISSN | 0125-8389 |
| Abstract | Three hundred and nine samples of wild banana (Musa acuminate Colla) were collected in the five national parks namely a wildlife sanctuary and a forest park during March 2002 to March 2003. All samples were examined through classical taxonomy, then major qualitative characteristics were classified into 7 groups. Meanwhile, Hierarchical and K-mean cluster analysis were numerical taxono being used to report examination with these characteristics. There are 7 characteristics of species identification criterion, while other 13 characteristics were subspecies. All samples were classified into 7 groups at 5 units of re-scaled distance with 80 % of similarity by Hierarchical cluster analysis. In addition, the similar classification results which were re-approved by K-means cluster analysis and non-significantly different at P=0.05 from Hierarchical cluster analysis. Discriminant analysis were shown the probability of precise separation and prediction into 7 groups at 5 units of re-scaled distance with 80 % of similarity by Hierarchical cluster analysis. In addition, the similar classification results which were re-approved by K-means cluster analysis and non-significantly different at P=0.05 from Hierarchical cluster analysis. Discriminant analysis were shown the probability of precise separation and prediction into 7 groups of both cluster techniques at 83-100 % with the original and cross-validated methods. Both results from numerical morphological classification were synchronized and non-significantly different from the result of classical morphological taxonomy with Pearson Chi-square test (P =0.05). Sample in the group 6 was highly different from M. acuminate Colla with its erect rachis position that should be a new wild banana group. It is very interesting to focus on re-identification of this sample through banana taxonomic system. The prominent strong convolute with significant 1/3 bright light green tip of male bud of the group 2 which was also interesting and might be a new promising subspecies of M. acuminate Colla. |
