Natural selection mediated by seasonal time constraints increases the alignment between evolvability and developmental plasticity

Abstract

Phenotypic plasticity can either hinder or promote adaptation to novel environments. Recent studies that have quantified alignments between plasticity, genetic variation, and divergence propose that such alignments may reflect constraints that bias future evolutionary trajectories. Here, we emphasize that such alignments may themselves be a result of natural selection and do not necessarily indicate constraints on adaptation. We estimated developmental plasticity and broad sense genetic covariance matrices (G) among damselfly populations situated along a latitudinal gradient in Europe. Damselflies were reared at photoperiod treatments that simulated the seasonal time constraints experienced at northern (strong constraints) and southern (relaxed constraints) latitudes. This allowed us to partition the effects of (1) latitude, (2) photoperiod, and (3) environmental novelty on G and its putative alignment with adaptive plasticity and divergence. Environmental novelty and latitude did not affect G, but photoperiod did. Photoperiod increased evolvability in the direction of observed adaptive divergence and developmental plasticity when G was assessed under strong seasonal time constraints at northern (relative to southern) photoperiod. Because selection and adaptation under time constraints is well understood in Lestes damselflies, our results suggest that natural selection can shape the alignment between divergence, plasticity, and evolvability.

Keywords: Adaptation; G-matrix; developmental bias; genetic constraints; latitude; life history; phenotypic plasticity; time constraints.

Figure 1

(A) significant latitudinal variation in body mass and development time among seven populations reared at their native photoperiod length. Male values are represented by squares and female values are represented by triangles. Dots on the map show sample location of the populations studied. Standard errors are <0.01 and not shown. (B) geographic variation and developmental plasticity in body mass and development time. Shown are breeding values estimated from the Bayesian mixed model run on mean‐standardized traits (model specification in Supplementary Information 6). Breeding values for southern and northern populations are shown as red and blue points, respectively. Breeding values from populations raised at their native and nonnative photoperiod are filled and open points, respectively, and surrounded by 95% confidence ellipses drawn with full and broken lines, respectively. The northern photoperiod caused a significant shift in the orientation of G in the direction of multivariate developmental plasticity, seen in populations of both northern and southern origin. Northern populations have evolved stronger plasticity in response to the photoperiod cue relative to southern populations.

Figure 2

(A) The alignment between multivariate developmental plasticity and the major axis of genetic variation, G_max, given for northern (blue borders) and southern (red borders) populations raised at northern (blue background) or southern (red background) photoperiod. An angle = 0 indicates complete alignment and an angle = 90 indicates that the directions of multivariate plasticity and G_max are orthogonal. (B) Standardized evolvability in the direction of multivariate developmental plasticity. Circles show estimates based on empirical data and triangles show null expectations based on simulated unstructured G‐matrices with homogeneous variances and covariances set to 0. Given are Bayesian posterior modes and 95% credible intervals. See main text for further details.