Have a personal or library account? Click to login
Differentiation of Five Species of Megachile (Hymenoptera: Megachilidae), based on Wing Shape Cover

Differentiation of Five Species of Megachile (Hymenoptera: Megachilidae), based on Wing Shape

Open Access
|Dec 2016

Abstract

In the present work we used landmark-based geometric morphometrics to compare the wing shapes of five species of Megachile (belonging to three subgenera) to confirm whether this technique may be used reliably for differentiation of this group. Analyses of wing shape by the use of principal component analysis (PCA), and canonical variate analysis (CVA) led to a clear differentiation among species. We found a close phenotypic similarity in wing shape between M. albisecta (belonging to the subgenus Creightonella) and M. picicornis (belonging to the subgenus Eutricharaea). According to the results of UPGMA, a higher degree of divergence between M. farinosa (belonging to the subgenus Pseudomegachile) and species belonging to other subgenera, was detected. The results of a cross-validation test indicated that geometric morphometrics is an effective technique to use for distinguishing between Megachile species. The reliability rate of this technique was between 85.71-100%. Using only two submarginal cell landmarks for generating shape variables, the cross-validation test correctly assigned individuals to their respective species, with a 92.85-100% reliability rate. Significant differences in wing size were obtained among the analysed species.

DOI: https://doi.org/10.1515/jas-2016-0015 | Journal eISSN: 2299-4831 | Journal ISSN: 1643-4439
Language: English
Page range: 41 - 50
Submitted on: Apr 26, 2016
Accepted on: Jul 5, 2016
Published on: Dec 8, 2016
Published by: Research Institute of Horticulture
In partnership with: Paradigm Publishing Services
Publication frequency: 2 issues per year

© 2016 Shahram Falamarzi, Behzad Habibpour, Mohammad S. Mossadegh, Alireza Monfared, published by Research Institute of Horticulture
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.