Metabolic rate shapes phenotypic covariance among physiological, behavioral, and life-history traits in honeybees

Abstract

Metabolic rate is often cited as the fundamental rate which determines the rate of all biological processes by shaping energetic availability for the various physiological, behavioral, and life-history traits that contribute to performance. It has therefore been suggested that metabolic rate drives the widely observed covariance among these different levels of phenotypic traits. However, much of the work on this topic has relied on pairwise correlational analysis on a handful of traits at a time, leaving an important gap in our understanding regarding the functional links that shape this phenotypic covariance, often referred to as pace of life. Using honeybees as a model, we measured a large number of behavioral, life-history, and physiological traits in individual bees and used a path analysis to demonstrate that variation in metabolic rate plays a fundamental proximate role in driving the covariance among these traits. We combined this with a factor analysis in a structural equation model framework to characterize the overall phenotypic covariance or the pace-of-life axis in honeybees. We discuss the importance of these findings in the context of how interindividual variation in terms of slow–fast phenotypes may drive the phenotype of a group and the functional role metabolic rate might play in shaping division of labor and social evolution. The study provides empirical support for the theoretical idea that metabolic rate acts as a proximate driver of phenotypic covariance among several physiological, behavioral, and life-history traits at the individual level and that behavior acts as a mediator for how metabolic rate affects life history. In addition, using honeybees as an experimental model for this study establishes a framework for asking questions regarding how these individual-level phenotypic covariance patterns lead to observed phenotypic covariance patterns at the colony level, an idea which has functional consequences for division of labor and social evolution. The results of this study therefore contribute toward a better understanding of the rules of life that shape processes across different levels of biological organization. Our use of different structural equation modeling approaches for inferring heuristics and proximate causal relationships among multiple phenotypic traits also informs future research efforts on this topic.