Rapid sperm length divergence in a polygynandrous passerine: a mechanism of cryptic speciation?

When populations become geographically isolated, they begin to diverge in various traits and at variable rates. The dynamics of such trait divergences are relevant for understanding evolutionary processes such as local adaptation and speciation. Here we examine divergences in sperm and body structures in a polygynandrous songbird, the alpine accentor (Prunella collaris) between two allopatric high-altitude populations, in Morocco and Spain. The populations diverged around 82,000 years ago, as estimated with a coalescence-based phylogenetic analysis of genome-wide single nucleotide polymorphisms. We found that birds in the two areas had non-overlapping sperm lengths. This suggests adaptation to divergent female reproductive tract environments. Sperm length also showed an exceptionally low coefficient of among-male variation, a signal of strong stabilizing selection imposed by sperm competition. The evolutionary rate of sperm length was almost twice the rates for the most divergent morpho..., Fieldwork was conducted in June-July 2012–2015 in the Picos de Europa mountain range in northern Spain, at altitudes above 2,000 m, and in June-July 2013–2015 in the High Atlas, central Morocco, at altitudes above 3,000 m. The data set on phenotypic traits (body size and sperm size components) contains measurements of 28 breeding males in the Spanish population and 9 males in the Moroccan population.
For each bird we measured the following body structures: right wing chord (straightened and flattened), length of the right primary wing feathers and tail length using a ruler (precision ± 0.5 mm); right tarsus length and bill length using a digital caliper (precision ± 0.1 mm); and body mass using a digital weighing scale (precision ± 0.1 g). The length of primaries were used to calculate three wing shape indices, isometric size, pointedness, and convexity, following the formulas in Lockwood et al. (1998), which are based on size-constrained component analysis (SCCA). These indices rely on...,