A central goal in ecology and evolutionary biology is to understand the genotype-phenotype relationship in natural populations. Seasonal timing of diapause induction is a key trait for survival in many insects. However, while many studies find associations between genetic variation and phenotypic variation, functional validation of these candidate loci is rare. Additionally, studies investigating whether the genotype-phenotype map is consistent across populations are scarce.

In this thesis, I used novel and established approaches to investigate the genotype-phenotype map of photoperiodic timing of diapause induction in the butterfly Pararge aegeria. The aims were to investigate: 1) whether the genotype-phenotype map is consistent across populations, 2) what the functional effects of candidate loci on the phenotype are, and 3) whether these loci have pleiotropic effects on other seasonal timing traits. I focused specifically on two candidate circadian clock genes, period and timeless, that have been associated with phenotypic variation in photoperiodic timing of diapause induction in this species.

In Chapter I, I investigated whether the genotype-phenotype map is consistent across populations. This was done by investigating whether similar phenotypic variation in repeated colonization events in the south of Sweden is caused by genetic variation in the same genes. Taking advantage of the fact that period is positioned on the Z chromosome, I created a reciprocal hybrid population cross and compared the results to those of a similar study with a different population pair. The results showed Z-linkage of diapause induction when crossing two populations from separate colonizations of Sweden. However, there was no Z-linkage of diapause induction when crossing two populations from the same colonization event. This indicates that similar phenotypic variation has evolved by selection on different genomic regions and that the genotype-phenotype map for seasonal timing of diapause induction can vary across populations. In Chapter II, I created a highly efficient protocol for generating germline mutations in P. aegeria, which enables functional validation of candidate genes underlying non-visible physiological traits and can aid in functional validation of candidate loci in future research. In Chapter III, I introgressed a naturally occurring fixed deletion in timeless from one population into another via backcrossing and selective breeding. I created two lines, one with and one without the deletion, and found that the genomic region including the deletion had a large effect on photoperiodic timing of diapause induction. In addition, it had an effect on diapause entry timing, revealing pleiotropic effects on a seasonal timing trait within the same diapause phase. In Chapter IV, I used siblings of the two introgressed lines and found that the genomic region does not exhibit pleiotropic effects on seasonal timing of diapause termination, a different diapause phase.

The results suggest that the genotype-phenotype map for photoperiodic timing of diapause induction in the Gotland population includes a genomic region containing timeless with a deletion that can have a large effect on the phenotype. In addition, of the two seasonal timing traits, this region has pleiotropic effects only on the trait that may benefit from correlative evolution. These characteristics may have facilitated rapid adaptation during the initial colonization of the island of Gotland.

Two findings challenge current expectations: first, the fact that the genomic region containing timeless may have a large effect on the timing of diapause induction. This challenges the expectation that complex traits are primarily regulated by many loci of small effects and raises questions about the evolutionary origin and maintenance of large-effect loci. Second, it is often assumed that the genotype-phenotype map is consistent across populations that diverged within a short evolutionary time. However, that is not the case for photoperiodic timing of diapause induction across populations of P. aegeria. This suggests that for complex traits it cannot be assumed that the genotype-phenotype map is constant, especially in the case of repeated colonization events. Functional validation of candidate loci across natural populations for seasonal timing and other complex traits represents a promising approach for further investigation into the consistency of the genotype-phenotype map.