Morphological variance in cave and surface populations of Asellus aquaticus challenges the fluctuating selection and greater male variability hypotheses

Anna Biro^1,2,3*Gergely Balázs^1,2Gergely Horváth^1,2Žiga Fišer⁴Serban M Sarbu^5,6Cene Fišer⁴Gábor Herczeg^1,2

• ¹Department of Systematic Zoology and Ecology, Institute of Biology, Faculty of Science, Eötvös Loránd University, Budapest, Hungary
• ²HUN-REN-ELTE-MTM Integrative Ecology Research Group, Budapest, Hungary
• ³Doctoral School of Biology, Institute of Biology, ELTE Eötvös Loránd University, Budapest, Hungary
• ⁴Department of Biology, Biotechnical Faculty, University of Ljubljana, Ljubljana, Slovenia
• ⁵Academia Romana Institutul de Speologie Emil Racovita, Bucharest, Romania
• ⁶Department of Biological Sciences, California State University, Chico, United States

Introduction: Fluctuating selection, the changes in the strength and/or direction of natural selection stemming from fine-scale spatio-temporal environmental variation, is a fundamental, yet rarely tested explanation for the maintenance of phenotypic variation within populations. Asellus aquaticus is an aquatic isopod inhabiting both surface (highly variable, complex environment) and cave (stable, simple environment) habitats, providing an excellent model to test the fluctuating selection hypothesis. The inclusion of multiple populations allows to explore variance patterns among-populations. We predicted that within-population variance (Vwp) is lower in caves due to negligible environmental variation. Regarding among-population variance (Vip) the study was exploratory as in caves, (i) lower variance could be expected due to their high environmental similarity, but (ii) higher variance could be expected due to the mutual isolation of caves. Finaly, we tested whether males exhibit higher morphological variance than females following predictions of the greater male variability hypothesis. Methods: We analysed the variance of 12 functional morphological traits of eight cave and nine surface A. aquaticus populations. Two statistical approaches were applied. Trait variances were determined for each sex using Bayesian double-hierarchical models. Then overall Vwp and Vip was estimated for each sex and habitat using Bayesian generalised linear models. Results: Overall Vwp did not differ between the habitats but in some traits, variance was larger in cave than in surface populations (12.5% of cases). Overall Vip was larger in caves. There were no sex-related Vwp patterns, but we found higher Vip in males than in females in the cave habitat. Discussion: Our results reject both the fluctuating selection and greater male variability hypotheses. We suggest that equal Vwp between habitats is a result of several processes overriding the marked habitat difference in environmental heterogeneity. Higher Vip in caves suggests the importance of isolation allowing local adaptation to cave specificities. Higher Vip in males in caves suggests the role of sexual selection in among-cave population divergence. These results underscore unmeasured differences among the caves and challenge the general view that caves are uniform replicates of simple and stable environments. For solid evolutionary inference, we recommend considering patterns of variance besides means.