Abstract
The effect of the environment on fitness in natural populations is a fundamental question in evolutionary biology. However, experimental manipulations of both environment and phenotype at the same time are rare. Thus, the relative importance of the competitive environment versus intrinsic organismal performance in shaping the location, height, and fluidity of fitness peaks and valleys remains largely unknown. Here, we experimentally tested the effect of competitor frequency on the complex fitness landscape driving adaptive radiation of a generalist and two trophic specialist pupfishes, a scale‐eater and molluscivore, endemic to hypersaline lakes on San Salvador Island (SSI), Bahamas. We manipulated phenotypes, by generating 3407 F4/F5 lab‐reared hybrids, and competitive environment, by altering the frequency of rare transgressive hybrids between field enclosures in two independent lake populations. We then tracked hybrid survival and growth rates across these four field enclosures for 3–11 months. In contrast to competitive speciation theory, we found no evidence that the frequency of hybrid phenotypes affected their survival. Instead, we observed a strikingly similar fitness landscape to a previous independent field experiment, each supporting multiple fitness peaks for generalist and molluscivore phenotypes and a large fitness valley isolating the divergent scale‐eater phenotype. These features of the fitness landscape were stable across manipulated competitive environments, multivariate trait axes, and spatiotemporal heterogeneity. We suggest that absolute performance constraints and divergent gene regulatory networks shape macroevolutionary (interspecific) fitness landscapes in addition to microevolutionary (intraspecific) competitive dynamics. This interplay between organism and environment underlies static and dynamic features of the adaptive landscape.
Highlights
It is possible that some regions of the high-dimensional trait space may still connect scale-eater phenotypes to other regions of the morphospace through a fitness ridge (Gavrilets 1997, 1999)
Fitness landscapes in lake 2 were estimated from few survivors (n = 22); random permutation of survival in lake 2 showed that the probability of observing nonlinear selection equal to or more extreme than we observed (GAM χ2 ≥ 7.19) due to chance was P = 0.014 in the high-frequency enclosure and P < 0.000 (χ2 ≥ 21.36) in the low-frequency enclosure
We found no evidence of differences in survival or fitness landscape topography between treatments manipulating the frequency of rare transgressive hybrid phenotypes in two independent lake populations (Figs. 3, 4, and S8; Tables S6–S8; 97% power to detect 20% nonlinear survival differences)
Summary
It is possible that some regions of the high-dimensional trait space may still connect scale-eater phenotypes to other regions of the morphospace through a fitness ridge (Gavrilets 1997, 1999). To further explore the relative fitness of scale-eater hybrids, we visualized selection across all directions in the 30-trait morphospace by repeatedly sampling a random subset of 15 traits, calculating a discriminant axis for scale-eaters relative to generalists within this subspace, and estimating a survival spline for hybrid phenotypes on each arbitrary multivariate axis (Fig. 6). We did not measure all traits potentially affecting fitness, this procedure creates a random set of discriminant morphospaces to visualize all fitness pathways across our set of thirty morphological traits. This procedure aligns these multivariate linear axes in the same direction from generalist to scale-eater phenotype enabling com
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
