Marked genetic differentiation of the UCSD juncos
Our results confirmed the striking differentiation of the urban population of Oregon juncos inhabiting the UCSD campus. Even when time of divergence was allowed to vary between 20 and 60 generations, parameter estimates, and bootstrap procedures recovered a time of divergence highly consistent with the time at which the juncos where first observed to breed at UCSD in the 1980s, with a narrow 95% CI of 20 to 32 generations, lending support to the recent founder event hypothesis. Recovered levels of divergence for the UCSD residents that were comparable to those of long-established, geographically isolated forms like townsendi and pontilis from Baja California sky islands, and in sharp contrast with the low genetic structure found among more northern forms. Both the PCA and the STRUCTURE analyses based on selectively neutral or near-neutral SNPs, revealed that the UCSD juncos represent a genetically differentiated population, although some degree of recent admixture with northern forms was detected. Juncos have become established in other suburban and natural areas of southern California in recent years (including the campus at UCLA in 2007, pers. obs.), and the degree to which the UCSD population is genetically isolated from these will require more extensive sampling in the region. At first sight, the pattern of divergence of UCSD juncos is comparable to that of other dark-eyed junco forms, known to have originated several thousand years ago (Friis et al. 2016). However, demographic inference analyses yielded results that are consistent with a very recent origin of the urban population. The inferred number of founders was as low as two to three individuals, just below the seven to 70 individuals previously estimated using microsatellite markers (Rasner et al. 2004). The analysis also favored a scenario of isolation with migration over a strict isolation model.
The migration rate from pinosus to UCSD was estimated at 6.02% with fastSIMCOAL2. This estimate is consistent with results from Yeh and Price (2004), who reported an immigration rate of 7% (95% CI: 4%-10%) based on observed ratios of banded to unbanded birds over a four-year period, although immigrants from other populations besidespinosus may be included in that estimate (see below). In any case, our results suggest that some degree of gene flow occurred during the differentiation of the urban population of juncos. In contrast, the inferred migration rate from UCSD to pinosus was negligible.
Our analyses also recovered a strong correlation between δD in feathers and population assignment likelihoods in UCSD individuals sampled during the non-breeding season, confirming the presence of both local residents and genetically distinct individuals from northern latitudes that do not represent recent immigrants but rather wintering visitors. These visitors were also genetically distinct from the juncos from Laguna Mountains and more similar to northern Oregon juncos.