The Coalescent in a Continuous, Finite, Linear Population

In this article we present a model for analyzing patterns of genetic diversity in a continuous, finite, linear habitat with restricted gene flow. The distribution of coalescent times and locations is derived for a pair of sequences sampled from arbitrary locations along the habitat. The results for mean time to coalescence are compared to simulated data. As expected, mean time to common ancestry increases with the distance separating the two sequences. Additionally, this mean time is greater near the center of the habitat than near the ends. In the distant past, lineages that have not undergone coalescence are more likely to have been at opposite ends of the population range, whereas coalescent events in the distant past are biased toward the center. All of these effects are more pronounced when gene flow is more limited. The pattern of pairwise nucleotide differences predicted by the model is compared to data collected from sardine populations. The sardine data are used to illustrate how demographic parameters can be estimated using the model.