It remains a central problem in population genetics to infer the past action of natural selection, and these inferences pose a challenge because demographic events will also substantially affect patterns of polymorphism and divergence. Thus it is imperative to explicitly model the underlying demographic history of the population whenever making inferences about natural selection. In light of the considerable interest in adaptation in African populations of Drosophila melanogaster, which are considered ancestral to the species, we generated a large polymorphism data set representing 2.1 Mb from each of 20 individuals from a Ugandan population of D. melanogaster. In contrast to previous inferences of a simple population expansion in eastern Africa, our demographic modeling of this ancestral population reveals a strong signature of a population bottleneck followed by population expansion, which has significant implications for future demographic modeling of derived populations of this species. Taking this more complex underlying demographic history into account, we also estimate a mean X-linked region-wide rate of adaptation of 6 · 10 211 /site/generation and a mean selection coefficient of beneficial mutations of 0.0009. These inferences regarding the rate and strength of selection are largely consistent with most other estimates from D. melanogaster and indicate a relatively high rate of adaptation driven by weakly beneficial mutations.T HERE is considerable interest in understanding the nature and extent of adaptation in natural populations. Drosophila melanogaster has been the focus of many such studies and a variety of approaches to address this fundamental question have been developed with this system. These include divergence-based methods (e.g., Sattath et al. 2011), polymorphism-based approaches (e.g., Kim and Stephan 2002;Sabeti et al. 2002;Kim and Nielsen 2004;Li and Stephan 2006b;Jensen et al. 2008a), and approaches that use both polymorphism and divergence data (e.g., Fay et al. 2002;Welch 2006;Andolfatto 2007;Macpherson et al. 2007;Shapiro et al. 2007). In some cases, these methods aim to localize putative targets of selection (e.g., Harr et al. 2002;Glinka et al. 2003;Jensen et al. 2007a), while in others the ultimate goal is to generally characterize the average rate and strength of adaptation (Wiehe and Stephan 1993;Li and Stephan 2006a;Andolfatto 2007;Macpherson et al. 2007;Jensen et al. 2008a). These approaches have enjoyed much success, and it is becoming clear that the signatures of both recent and recurrent selection abound in the D. melanogaster genome (for review see Sella et al. 2009).However, demographic history will also leave its signature in the genome, which has the potential to confound inferences of natural selection. Indeed, to effectively investigate models of natural selection, which are of great general interest in evolutionary biology, one must take into account the underlying demographic history of a population. Consequently, understanding demographic history is integr...