Late Pleistocene colonization of South Georgia by yellow-billed pintails pre-dates the Last Glacial Maximum

Aim: Glacial cycles had a major influence on the distribution of high-latitude species, and while genetic consequences of these cycles have been well studied in the Circumpolar North, notably fewer studies have been undertaken in the Antarctic. Here, we use multilocus genetic data from the South Georgia pintail (Anas georgica georgica) as a proxy to study the presence and timing of ice-free refugia on South Georgia, a glaciated subantarctic island in the South Atlantic Ocean that has been the subject of intense geomorphological and palaeoenvironmental study. Location: South Georgia. Methods: Multilocus DNA sequence data from five nuclear loci and the mitochondrial DNA control region were analysed for South Georgia pintails (n = 60) and the neighbouring population of yellow-billed pintails (A. georgica spinicauda) in Argentina (n = 64). Population genetic structure and gene flow were examined using F, assignment tests, and multilocus coalescent analyses. Results: Isolation-with-migration (IM) analysis revealed that the South Georgia pintail population was founded by pintails dispersing from South America. Although the confidence intervals on divergence dates inferred from genetic data are generally wide and there may be time-dependency in rate calibrations, our analysis suggests that this founding event probably occurred c. 34,000 years ago, prior to the Last Glacial Maximum (LGM). Our findings further suggest that South Georgia pintails might have experienced a bottleneck coinciding with complete replacement of mitochondrial DNA prior to 8700 years ago following the final advance of glaciers. Main conclusions: These findings suggest that ice-free refugia existed earlier in the chronology of deglaciation in contrast with earlier studies, but in agreement with observations that the plant community was also established on South Georgia prior to the end of the Pleistocene. Like other recent studies that have utilized genetic data to date dispersal and vicariance events in the Antarctic, our results provide a constraint on the extent of ice sheets, suggesting that past ice coverage on South Georgia through the LGM was overestimated.