Carlos Bustamante and colleagues demonstrate the potential of using local ancestry analysis of the genome sequences of admixed individuals for improved demographic inference (Am. J. Hum. Genet. 91, 660–671, 2012). The authors examined the whole-genome sequences of 50 individuals from 11 diverse populations. Their initial comparisons showed expected patterns, including an average of 18% greater genetic variation in genomes from African individuals and more rare variants than expected under a standard neutral model. They further examined the genome sequences of 12 individuals with recent admixed ancestry. Using the PCAdmix algorithm to analyze each chromosome and partition genomic regions by inferred European, African or Native American ancestry, they were able to assign inferred local ancestry to 79–95% of each genome with an accuracy of 80–99%. They modeled the distribution of ancestry-tract lengths and found departures from predictions of a pulse admixture model. A continuous or repeated migration model provided improved fit to genetic and historic data, including continuous gene flow from Europe. The authors estimate that 70% of the European ancestry in modern African-Americans dates to European gene flow 7–8 generations ago. They also highlight the need for new methods for analyzing admixed genomes.