Skull shape of a widely distributed, endangered marsupial reveals little evidence of local adaptation between fragmented populations

The biogeographic distribution of diversity among populations of threatened mammalian species is generally investigated using population genetics. However, intraspecific phenotypic diversity is rarely assessed beyond taxonomy-focused linear measurements or qualitative descriptions. Here, we use a technique widely used in the evolutionary sciences-geometric morphometrics-to characterize shape diversity in the skull of an endangered marsupial, the northern quoll, across its 5,000 km distribution range along Northern Australia. Skull shape is a proxy for feeding, behavior, and phenotypic differentiation, allowing us to ask whether populations can be distinguished and whether patterns of variation indicate adaptability to changing environmental conditions. We analyzed skull shape in 101 individuals across four mainland populations and several islands. We assessed the contribution of population, size, sex, rainfall, temperature, and geography to skull shape variation using principal component analysis, Procrustes ANOVA, and variation partitioning analyses. The populations harbor similar amounts of broadly overlapping skull shape variation, with relatively low geographic effects. Size predicted skull shape best, coinciding with braincase size variation and differences in zygomatic arches. Size-adjusted differences in populations explained less variation with far smaller effect sizes, relating to changes in the insertion areas of masticatory muscles, as well as the upper muzzle and incisor region. Climatic and geographic variables contributed little. Strikingly, the vast majority of shape variation-76%-remained unexplained. Our results suggest a uniform intraspecific scope for shape variation, possibly due to allometric constraints or phenotypic plasticity beyond the relatively strong allometric effect. The lack of local adaptation indicates that cross-breeding between populations will not reduce local morphological skull (and probably general musculoskeletal) adaptation because none exists. However, the potential for heritable morphological variation (e.g., specialization to local diets) seems exceedingly limited. We conclude that 3D geometric morphometrics can provide a comprehensive, statistically rigorous phenomic contribution to genetic-based conservation studies.