Detection of change in population size with Msvar. The Bayes factors (BF) are given for each of the following demographic scenarios: a stable population (with *N*_{0} = *N*_{1} = 464, *T*_{a} = 500), a declining population (with *N*_{0} = 100, *N*_{1} =10,000, and *T*_{a} = 500), and an expanding population (with *N*_{0} = 10,000, *N*_{1} =100, and *T*_{a} = 500). We considered three different mutation models, which differ from each other by the value of *p*, the frequency of multistep mutation changes: *p* = 0.00 (stepwise mutation model, SMM), *p* = 0.22 (moderate generalized stepwise model, GSM_{1}), and *p* = 0.74 (strong generalized stepwise model, GSM_{2}). For the stable population scenario, the lower triangle provides the Bayes factor for a population decline (*i.e*., the ratio of the posterior probability of a population decline divided by the posterior probability of a population expansion), and the upper triangle provides the Bayes factor for a population expansion (*i.e*., the ratio of the posterior probability of a population expansion divided by the posterior probability of a population decline). Following Jeffreys (1961), BF ≥ 10 indicate strong support, and BF ranging from 3 to 10 indicate substantial support. BF ranging from 0.33 to 3 indicate no support and values <0.33 indicate false detection of contraction or expansion. Nonconverged (NC) analyses are also indicated.

## PubMed Commons