Parallel shifts of visual sensitivity and body coloration in replicate populations of extremophile fish

Abstract

Visual sensitivity and body pigmentation are often shaped by both natural selection from the environment and sexual selection from mate choice. One way of quantifying the impact of the environment is by measuring how traits have changed after colonization of a novel habitat. To do this, we studied Poecilia mexicana populations that have repeatedly adapted to extreme sulphidic (H2S‐containing) environments. We measured visual sensitivity using opsin gene expression, as well as body pigmentation, for populations in four independent drainages. Both visual sensitivity and body pigmentation showed significant parallel shifts towards greater medium‐wavelength sensitivity and reflectance in sulphidic populations. Altogether we found that sulphidic habitats select for differences in visual sensitivity and pigmentation. Shifts between habitats may be due to both differences in the water's spectral properties and correlated ecological changes.

Detecting (non)parallel evolution in multidimensional spaces: angles, correlations and eigenanalysis

“…When interest is in analysing a set of many lineages simultaneously (e.g. [37,40,41]), a convenient procedure is to construct a matrix of pairwise angles or correlations. Let x i denote phenotypic change vectors of p traits from n lineages (i = 1, …, n), each starting from its respective ancestor, and arrange these in rows of the n × p matrix X.…”

“…The angles between phenotypic change vectors of different lineages play an especially pivotal role in empirical analyses of parallel evolution (e.g. [33,[37][38][39][40][41][42][43]), because they are supposed to provide 'intuitive and mathematically formal' measures of (non)parallelism [37, p. 6].…”

Hybrid evolution repeats itself across environmental contexts in Texas sunflowers ( Helianthus )

“…In this context, a divergence angle (θ P ) of 0°is perfectly "parallel" divergence (repeated evolution, or similar evolution across home sites), 90°is completely orthogonal divergence, and 180°is completely anti-parallel (change in the opposite direction). To test for repeated evolution using these divergence angles, we tested whether the observed angles deviated significantly from orthogonal using a bootstrapping procedure, where divergence angles significantly lower than orthogonal are evidence for parallelism (Owens et al 2021b). At each common garden, we resampled within each generation × treatment × home site combination with replacement, calculated divergence vectors between generation 1 and generation 8 as above, and then calculated θ P. We repeated this for 1000 bootstrap replicates to determine the proportion of bootstrap replicates that were 90 degrees or greater, multiplied by 2 (two-tailed approach) (Stuart et al 2017).…”