No change in genetic variance
The effect of selection on genetic variation is a key topic in evolutionary biology. Strong stabilizing or directional selection is expected to erode genetic variance (Barton and Turelli, 1989; Estes and Arnold, 2007). However, when many loci are contributing to variation in the trait under selection, selection might leave no detectable signature on the genetic variance associated with that trait (Bulmer, 1971; Johnson and Barton, 2005). Although our truncated selection increased and decreased the phenotypic mean by three-quarters of a standard deviation in the high and low line, respectively, we observed no significant decrease in the coefficient of additive genetic variance. There are several explanations for this observation. First, several genetic factors may be involved. Previously, genetic mapping studies revealed several QTL underlying difference in sex pheromone composition in H. subflexa (Groot et al., 2013). Since variation in pheromone composition in H. subflexa is also explained by fitness variation (Blankers et al., 2021), many different genetic factors likely contribute small additional fractions to the total variance. Another explanation comes from population genetics: re-orientation of genetic covariances in response to bottlenecks may free up additive genetic variance, thereby paradoxically increasing levels of genetic variance (Carson, 1990; Templeton, 2008). Since truncation selection is effectively a non-random bottleneck of the population, this may also explain the maintenance of genetic variance that we observed. Lastly, to avoid inbreeding depression, we maintained large populations and avoided mating first and second-degree relatives. Such a mating scheme likely counteracts some loss in genetic variability due to selection (Du et al., 2021).