Efficacy of otolith morphometry for rapid discrimination of cryptic fishes

Abstract

The cryptic external morphology of fishes poses challenges for accurate species identification. This study assessed whether the morphometry of otoliths (and body length) could be used as a method to distinguish among closely related, morphologically similar species that are highly valued by commercial or recreational fishers. The otolith morphometry of fishes within four genera (Centroberyx, Lethrinus, Polyprion and Sillago) were compared. Canonical Analysis of Principal Coordinates that included combinations of morphometric variables (i.e., otolith length, width, thickness, weight, and fish length) produced highly accurate species prediction rates ranging from 92.5 to 99.9% across all genera. When different combinations of morphometric variables were modelled, changes in species prediction accuracy rates were minimal (i.e., up to 3%) for the Centroberyx, Polyprion and Sillago species. The exclusion of otolith weight from the multivariate analysis did not significantly decrease prediction accuracy within all four genera. In contrast, the prediction accuracy of the two lethrinids reduced by 7–25% when fish length was excluded, which may be attributable to the larger size attained by L. nebulosus in comparison to its congener L. punctulatus, as well as the very similar ovate shaped otoliths in younger fish of both species. Simplifying data collection to only length-based otolith measurements (and fish length for lethrinids) would provide a more rapid and cost-effective method for discrimination and be more practical when sampling at sea where obtaining accurate weight measurements of otoliths is typically unfeasible. Spatial variation in otolith morphometrics was evident for C. gerrardi, L. punctulatus, P. oxygeneios and S. vittata, which might be attributed to the differences in length distributions of fish samples from different regions. The misclassified specimens within the species groups were notably smaller individuals, suggesting the complexity of otolith shape increases with ontogeny. Hence, further investigation of variations in morphological characteristics, other than otoliths, may be required for the discrimination of cryptic species during the juvenile stage. This study highlights the robustness of multivariate analysis of otolith morphometric data for discriminating closely-related, cryptic species.