STOCK IDENTIFICATION OF COILIA MYSTUS USING OTOLITH SHAPE ANALYSIS

YU Xin,CAO Liang,NAN Ou,ZHAO Bo,DOU Shuo-Zeng
2013-01-01
Abstract:In this study, the morphology of the sagittal otoliths of Coilia mystus from four stocks along the Chinese coastal waters (Liaodong Bay, Bohai Bay, Jiaozhou Bay and Changjiang River Estuary) were investigated using otolith shape analysis. Shape indices and Fourier coefficients were used as morphological variables and were statistically analyzed by both ANOVA and ANCOVA (otolith length as a covariate) to examine their differences among stocks. The results revealed that most of the morphometric variables of the otoliths examined in the study differed significantly among the four stocks in both statistical analyses. ANCOVA adjustment and nonparametric test were thereafter run to identify the effective variables for stock identification using discriminant function analysis. ANCOVA adjustment identified only otolith weight and 9 Fourier coefficients, whereas nonparametric test could keep more variables (otolith weight, roundness, fractal dimension, circularity, rectangularity and 19 Fourier coefficients) as the effective morphological variables of otoliths for identifying stocks. Accordingly, classification success (68.2%) of stock identification based on nonparametric test was higher than that (46.2%) based on ANCOVA adjustment, suggesting that the former statistical analysis could show stronger ability of otolith shape analysis to identify the stocks than the latter one.
What problem does this paper attempt to address?