We live in an era of significant environmental and climatic change and it has even been suggested that the world is entering a new epoch, the ‘Anthropocene’. To understand better how species might cope under different future climate scenarios, studies are now frequently looking to explore how they responded to rapid environmental change in the past. Whilst census data can capture contemporary trends, genetic approaches can infer population trends stretching tens, or even thousands, of years back in time. In this thesis, I first used skyline plots to infer historical demographic trends from genetic data of a well-studied system, humans. Using this gold standard, my work revealed detailed demographic profiles, but also identified issues relating to the way key methodological assumptions are contravened. In Chapter 2 I present a discussion about the risk of misinterpretation or overinterpretation in the context of Bayesian skyline plot (BSP) analysis. Understanding that any single profile can be problematic, when moving to non-model species, I chose to work as many species as possible. This approach exploits the recent boom in sequencing projects that has generated a huge volume of publicly available data. By building large, novel, multi-species datasets it becomes possible to construct profiles averaged over many species with similar properties, such as habitat preference. The expectation is that average profiles will prove better at capturing broad trends for the species they contain. Collating and processing public domain data is not a trivial task. I therefore developed a pipeline, now an R package, to access and compile sequence data for over 100 species of bird, focusing on mitochondrial DNA (mtDNA). I found differences in the mean time of population expansion after the ice age between bird species associated with different habitats. However, notably, the demographic trends drawn from BSPs did not reveal a close match with the amount of available habitat indicated by species distribution models. BSPs frequently indicated population increases even though species’ habitat ranges were decreasing. These results further emphasise the level of care needed when interpreting BSPs. If genetic methods for demographic reconstruction are to be used extensively in the future, it is important that we understand what confounding factors commonly exist in real world populations so as to prevent misleading or inaccurate interpretations. To explore the impact of historic range dynamics on BSPs I created a realistic spatial demographic model for a small North American passerine, the yellow warbler (Setophaga petechia). From this I simulated mtDNA sequences for a number of populations across the modern species’ range. With these data I’d hoped to investigate how BSP profiles varied depending on local population history. However, true demographic signals proved hard to capture and further work will be required to explore my original question more fully.