4.1 Intraspecific independent evolution of spur length among mountain regions
The spur length of A. buergeriana var. buergeriana was correlated with the average flower visitor size (only bumblebees) of the population; spur length varied greatly with the average visitor size even among spatially close populations in the same mountain region (Figures 2 and 4a). The PCA results for the three floral traits (spur length, corolla width, and petal width), showed that the floral traits tended to be differentiated even among populations within the same mountain region (supplementary material, Figure S2). The genetic results obtained by PCA and the STRUCTURE analysis of genome-wide SNPs suggest that populations within each mountain region are more closely related to each other than to populations in other mountain regions (Figures 3 and 4; supplementary material, Figure S4). Genetic isolation was proportional to geographical distance and did not reflect trait differences (Table 2). These results suggest that spur length ofA. buergeriana var. buergeriana evolved independently in each mountain region. The Ontake and Norikura regions are part of a series of volcanic massifs called the Norikura volcanic chain (Kimura & Yoshida, 1999; Sekiguchi & Yamagishi, 2013), and the colonization history of the two regions seems to be very closely related. The close genetic relationship detected between the populations of these two mountain regions may be related to the related origin of the massifs. In the Iizuna region, populations at different altitudes seem to belong to different genetic clusters (Figure 4d); further, flowers in lower altitude populations were visited by B. diversus and those in higher altitude populations were visited by B. consobrinus (Table S1). These results suggest that genetic differentiation may occur between higher and lower altitude populations because of a lack of pollinator sharing. Further studies are needed to determine whether gene flow by pollination is hindered between populations at higher and lower altitudes in the Iizuna region.
Hodges et al. (2002) have reported a genetic basis for spur length in two Aquilegia species (A. formosa and A. pubescens ), and they have performed quantitative trait locus mapping for spur length variation. In addition, the functions of some of the quantitative genes that cause spur length variation in A. coerulea have been elucidated (Zhang et al., 2020). Therefore, we think it highly likely that spur length has a genetic basis in A. buergeriana var. buergeriana , and that the evolution of spur length is facilitated by flower visitors.
In anole lizards, leg length has evolved independently on different islands to suit local habitats (Losos, 2010), and in stickleback fishes, the evolution of marine to freshwater forms (sticklebacks that move between rivers and the sea) occurred independently in different marine and freshwater locations in various regions of the world (Jones et al., 2012). We propose that plant species distributed across a wide geographic range with site-specific, different-sized pollinators constitute another model suitable for testing independent adaptive radiation. We have demonstrated that spur length in an Aquilegiaspecies may have evolved independently among mountain regions by using a population genetics approach to compare traits among mountain regions. Independent evolution in different mountain regions has recently been examined in various model systems: for example, the independent evolution of upland and short-winged forms of scorpionflyPanorpodes (Panorpodidae) (Suzuki et al., 2019), the independent evolution of Potentilla matsumurae (Rosaceae) in fellfield and snowbed environments on different mountains in Japan (Hirao et al., 2019), and the independent evolution of alpine morphology inAntirrhinum species (Antirrhineae) (DurĂ¡n-Castillo et al., 2021). Further, we recently presented a case in which we used microsatellite markers to show the independent adaptation of floral tube size inLamium album var. barbatum (Lamiaceae), associated with flower visitor size, in the Utsukushigahara and Norikura regions of the Japanese Alps (Toji et al., 2021). These examples show that comparisons between mountain regions can be used to study independent trait evolution in various organisms, and similar patterns might be found in many places around the world.