Journal Article

Measuring fitness and inferring natural selection from long-term field studies: different measures lead to nuanced conclusions

Coltman, David W, Raveh, Shirley, Tamian, Anouch, Murie, Jan O, Viblanc, Vincent A, Dobson, F. Stephen, Saraux, Claire, Criscuolo, François

2022

Overview

Measuring individual reproductive success in the wild is often achieved by counting the number of descendants produced by individuals. In seeking to understand how reproductive success can inform us about natural selection, however, we are faced with a conundrum. In terms of timing, what is the most relevant measure for examining selection? We might count the number of offspring born, surviving to the termination of parental care, surviving to adulthood, or only those surviving to themselves reproduce. Clearly, only the latter are passing on genes and traits to future generations, but this estimate may not always be available. So, are different estimates of fitness consistent? Do they provide us with similar inferences of selection on phenotypic traits? We examined these questions on a 29-year long-term study of individually monitored male and female Columbian ground squirrels (Urocitellus columbianus). We used the long-term data to calculate male and female fitness based on an annual measure of adult survival and the yearly production of offspring counted at the stages of birth, weaning, and yearling age. We then decomposed fitness into its constitutive elements including (1) adult survival to the next spring, and (2) the yearly production of offspring counted at the stages previously mentioned. We then compared fitness metrics to evaluate if they provided similar or contrasting information in the wild. Next, we used those fitness metrics to test for selection on the date of emergence from annual hibernation, a phenotypic trait previously shown to be highly variable, heritable, and associated with reproduction. Finally, we partitioned selection on emergence date into additive episodes of selection by looking at how selection changed from reproduction measured at birth, weaning, and when offspring reached yearling age. Overall, fitness metrics were well correlated, but correlations weakened the further offspring were counted from birth. We generally found directional selection for earlier emergence dates both in males and females. The strength of selection depended on which fitness metric was used. Most of the selection gradient on emergence date was explained by offspring born, and the selection differential was stronger in males than females. We evaluate how the choice of fitness metrics in life-history studies may nuance our inferences about natural selection.Significance statementThis study explores how our inferences about natural selection acting on organismal traits vary depending on our choice of fitness metrics. Focusing on the timing of emergence from hibernation in Columbian ground squirrels, we show that directional selection for earlier emergence dates occurs, but the strength of selection depends on whether fitness is evaluated from offspring counted at birth, at weaning, or later in life. These results show that the choice of timing for fitness measurements may nuance inferences about natural selection in life-history studies.

• Publisher: Springer Nature B.V
• Subject: Animal reproduction / Birth / Breeding success / Counting / Descendants / Evaluation / Females / Field study / Fitness / Genes / Ground squirrels / Hibernation / Inference / Life / Life history / Life history studies / Long term / Males / Measures / Natural selection / Offspring / Production / Reproduction / Reproductive fitness / Sex differences / Survival / Termination / Time measurement / Weaning