Ostracod shell plasticity across longitudinal and bathymetric ranges

• Published in: Deep Sea Research Part I: Oceanographic Research Papers
• Main Authors: Karanovic, I, Huyen, PTM, Brandao, SN
• Format: Article in Journal/Newspaper
• Language: English
• Published: Pergamon-Elsevier Science Ltd 2019
• Subjects: Biological Sciences; Evolutionary Biology; Animal Systematics and Taxonomy; Nordic Seas
• Online Access: https://doi.org/10.1016/j.dsr.2018.11.005; http://ecite.utas.edu.au/134975

We combine geometric morphometric and molecular data to study shell shape changes in eurybathic ostracod Polycope pseudoinornata Chavtur, 1983 and discuss its occurrence in relation to environmental variables (of the sediment and deep water masses). Pairwise distances between 50 mitochondrial COI sequences indicate that populations from five sampled depths and nine localities in the Nordic Seas belong to one species, and haplotype network shows that populations are strongly interconnected. Datasets for geometric morphometrics consist of 209 valves with landmarks distributed on the outer surface and margins. Results support morphological homogeneity of shell throughout the study area. There are important differences between datasets in their dimensionality and directional asymmetry. Pores/ridges dataset is highly unstructured with 50 + eigenvectors describing 90% of variation in PCA and with no directional asymmetry. Outline dataset is highly structured with 5 eigenvectors describing 90% of variation with strong directional asymmetry, with the right valve showing more variation in shape than the left valve. Although pore homologization could be difficult, these landmarks might be more powerful in multispecies geometric morphometric analyses, or multivariate analysis of environmental influences on the ostracod shell. Also different valves can easily be matched, which is very important for fossil species where often one valve is missing. Permutation tests show a statistically significant negative correlation between size and depth, but predicted values are small. Partial least square analyses of covariation between the outline and landmarks on the surface reveal integration of the two compartments but the covariation is not sufficient to account for all the variation throughout the valves.