9During the Quaternary, large climate oscillations had profound impacts on the distribution, 10 demography and diversity of species globally. Birds offer a special opportunity for studying these 11 impacts because surveys of geographical distributions, publicly-available genetic sequence data, and 12 the existence of species with adaptations to life in structurally different habitats, permit large-scale 13 comparative analyses. We use Bayesian Skyline Plot (BSP) analysis of mitochondrial DNA to 14 reconstruct profiles depicting how effective population size (Ne) may have changed over time, 15 focussing on variation in the effect of the last deglaciation among 102 Holarctic species. Only 3 16 species showed a decline in Ne since the Last Glacial Maximum (LGM) and 7 showed no sizeable 17 change, whilst 92 profiles revealed an increase in Ne. Using bioclimatic Species Distribution Models 18 (SDMs), we also estimated changes in species potential range extent since the LGM. Whilst most 19 modelled ranges also increased, we found no correlation across species between the magnitude of 20 change in range size and change in Ne. The lack of correlation between SDM and BSP 21 reconstructions could not be reconciled even when range shifts were considered. We suggest the 22 lack of agreement between these measures might be linked to changes in population densities 23 which can be independent of range changes. We caution that interpreting either SDM or BSPs 24 independently is problematic and potentially misleading. Additionally, we found that Ne of wetland 25 species tended to increase later than species from terrestrial habitats, possibly reflecting a delayed 26 increase in the extent of this habitat type after the LGM. 27
28The Quaternary period has been characterised by extensive cycles of glaciation and de-glaciation. 29 The legacy of these ancient large-scale climate alterations is evident today in everything from 30 species' genetic diversity to population structure (Twitchett 2006;Svenning et al. 2015). Despite the 31 profound impact that past climate changes have had on both flora and fauna, there is limited 32 quantitative evidence on which factors determined how different species fared during these cycles, 33 or how species responded to subsequent post-glacial climate amelioration. 34One of the most widely used genetic methods for inferring demographic history is the so-called 35 skyline plot, a family of graphical, non-parametric methods first introduced by Pybus et al. (2000). 36Grounded in the principles of Kingman's coalescent theory (Kingman 1982), the 'skyline framework' 37 aims to use DNA sequence data to reconstruct a gene tree. The rate of coalescent events within the 38 gene tree can then be used to infer how the population changed in size over time: in essence, 39 periods of low coalescent rates imply a large population while a high density of coalescent events 40 implies a small population. Although skyline plots have been used to reconstruct demographic 41 histories for many species, both extant and...