A functional view reveals substantial predictability of pollinator-mediated selection

Øystein Opedal

doi:10.26786/1920-7603(2021)673

Abstract

A predictive understanding of adaptation to changing environments hinges on a mechanistic understanding of the extent and causes of variation in natural selection. Estimating variation in selection is difficult due to the complex relationships between phenotypic traits and fitness, and the uncertainty associated with individual selection estimates. Plant-pollinator interactions provide ideal systems for understanding variation in selection and its predictability, because both the selective agents (pollinators) and the process linking phenotypes to fitness (pollination) are generally known. Through examples from the pollination literature, I discuss how explicit consideration of the functional mechanisms underlying trait-performance relationships can clarify the relationship between traits and fitness, and how variation in the ecological context that generates selection can help disentangle biologically important variation in selection from sampling variation. I then evaluate the predictability of variation in pollinator-mediated selection through a survey, reanalysis, and synthesis of results from the literature. The synthesis demonstrates that pollinator-mediated selection often varies substantially among trait functional groups, as well as in time and space. Covariance between patterns of selection and ecological variables provides additional support for the biological importance of observed selection, but the detection of such covariance depends on careful choice of relevant predictor variables as well as consideration of quantitative measurements and their meaning, an aspect often neglected in selection studies.

Highlights

Natural selection is the primary mechanism by which populations adapt to their environments
I will argue that the predictability of variation in selection can be improved by leveraging a combination of (i) functional knowledge of trait-performance-fitness relationships, (ii) quantitative interpretation of selection estimates, and (iii) explicit consideration of spatiotemporal variation in ecological variables relevant to the focal component of selection
Through my discussion of pollinator-mediated selection I have argued that a predictive understanding of natural selection requires a combination of (i) functional knowledge of traitperformance-fitness relationships obtained from experiments and detailed observations, (ii) quantitative interpretation of selection estimates, and (iii) explicit consideration of spatiotemporal variation in ecological variables relevant to the focal component of selection

Summary

Introduction

Natural selection is the primary mechanism by which populations adapt to their environments. I will argue that the predictability of variation in selection can be improved by leveraging a combination of (i) functional knowledge of trait-performance-fitness relationships, (ii) quantitative interpretation of selection estimates, and (iii) explicit consideration of spatiotemporal variation in ecological variables relevant to the focal component of selection.

Results

Conclusion