On the importance of scale in evolutionary quantitative genetics

The informed use of scales and units in evolutionary quantitative genetics is often neglected, and naïve standardizations can cause misinterpretations of empirical results. A potentially influential example of such neglect can be found in the recent book by Stevan J. Arnold (2023. Evolutionary Quantitative Genetics Oxford University Press). There, Arnold championed the use of heritability over mean-scaled genetic variance as a measure of evolutionary potential arguing that mean-scaled genetic variances are correlated with trait means while heritabilities are not. Here, we show that Arnold's empirical result is an artifact of ignoring the units in which traits are measured. More importantly, Arnold’s argument mistakenly assumes that the goal of mean scaling is to remove the relationship between mean and variance. In our view, mean scaling is useful because it puts traits with different units on a common scale that makes evolutionary changes, or their potential, readily interpretable and comparable in terms of proportions of the mean.

Methods
Data are part of two previously published studies, Houle 1992 Genetics and Holstad et al. 2024 Science. Details about the methods to collect the data can be found in the original papers.