4.2 Flower color does not contribute to genetic isolation
Although red-flowered individuals were observed in some populations, genetic analyses (STRUCTURE and PCA) based on neutral genes did not differentiate red- and yellow-flowered individuals in those populations. These results suggest that red flower color is maintained in each population merely as a flower color polymorphism. Throughout the diversification history of Aquilegia , flower color changes have been shown to be associated with pollinator shifts (Whittall & Hodges, 2007). Another well-known example is the Mimulus aurantiacusspecies complex, in which flower color influences pollinator preference, which in turn leads to genetic isolation. Within the M. aurantiacus species complex, there are two ecotypes, one with red flowers, which are preferred by hummingbirds, and the other with yellow flowers, which are preferred by hawkmoths. Although these two ecotypes are very closely related, cluster analysis by RAD-seq (restriction site-associated DNA sequencing) based on genome-wide SNP data has shown that they are genetically distinct (Sobel & Streisfeld, 2015). In the hybrid zone between the two ecotypes, the MaMyb2 gene, which is involved in the synthesis of flower pigments, is geographically maintained despite neutral gene flow occurred (Streisfeld & Kohn, 2005; Sobel & Streisfeld, 2015; Stankowski & Streisfeld, 2015). Gene flow between yellow and red flower M. aurantiacus ecotypes in the early stages of speciation seems to be limited mainly by differences in pollinator preference (Sobel & Streisfeld, 2015). Similarly, gene flow between two closely related Aquilegia species: hummingbird-pollinated, red-flowered A. formosa and hawkmoth-pollinated, yellow-flowered A. pubescens is also limited by pollinator isolation when the two species are distributed parapatrically (Fulton & Hodges, 1999; Noutsos et al., 2014).
Why have yellow- and red-flowered individuals in A. buergerianavar. buergeriana not become genetically isolated? In the central Nagano region, where this study was conducted, bumblebees appear to be abundant and many flowers depend on bumblebees for pollination (e.g. Egawa & Itino, 2020), whereas potential pollinators such as birds and butterflies that prefer red flowers are scarce. In another Japanese mountain region (the Taisetsu mountains), flowers are dominantly visited by bees and flies at the community level (Mizunaga & Kudo, 2017). A recent review has reported that Lepidoptera account for less than 10% of insect visitors to flowers in many parts of Asia, whereas bees and flowers account for more than half (Funamoto, 2019). It is possible that in the central Japanese Alps, because only the locally abundant bumblebees contribute to pollination of A. buergeriana var.buergeriana irrespective of the flower color, pollinator shifts to other taxa such as birds have not triggered the evolution of extreme traits. Whether bumblebees cause selection or act neutrally with respect to flower color requires further investigation, but the maintenance of small numbers of red-flowered individuals in some populations suggests that the frequency of red flowers may be determined by genetic drift. The maintenance of this small number of different flower-color polymorphisms in some populations might become a driving force for a pollinator shift should the plants be faced with a new pollinator environment.