Table 3.

The models used to simulate pseudo-replicate datasets for assessing the power of the models in Table2

	Priors
Model series	t	τ	θ
$${\mathcal{ℳ}}_{\mathit{\text{msBayes}}}$$	|τ|=22	τ∼U(0,0.2 [ 0.5 MGA])	θA∼U(0,0.05)	$${\overline{\theta }}_D\sim U(0,0.05)$$
	|τ|=22	τ∼U(0,0.4 [ 1.0 MGA])	θA∼U(0,0.05)	$${\overline{\theta }}_D\sim U(0,0.05)$$
	|τ|=22	τ∼U(0,0.6 [ 1.5 MGA])	θA∼U(0,0.05)	$${\overline{\theta }}_D\sim U(0,0.05)$$
	|τ|=22	τ∼U(0,0.8 [ 2.0 MGA])	θA∼U(0,0.05)	$${\overline{\theta }}_D\sim U(0,0.05)$$
	|τ|=22	τ∼U(0,1.0 [ 2.5 MGA])	θA∼U(0,0.05)	$${\overline{\theta }}_D\sim U(0,0.05)$$
	|τ|=22	τ∼U(0,2.0 [ 5.0 MGA])	θA∼U(0,0.05)	$${\overline{\theta }}_D\sim U(0,0.05)$$
$${\mathcal{ℳ}}_{\mathit{\text{Uniform}}}$$	|τ|=22	τ∼U(0,0.2 [ 0.5 MGA])	θA∼θD1∼θD2∼Exp(mean=0.025)
	|τ|=22	τ∼U(0,0.4 [ 1.0 MGA])	θA∼θD1∼θD2∼Exp(mean=0.025)
	|τ|=22	τ∼U(0,0.6 [ 1.5 MGA])	θA∼θD1∼θD2∼Exp(mean=0.025)
	|τ|=22	τ∼U(0,0.8 [ 2.0 MGA])	θA∼θD1∼θD2∼Exp(mean=0.025)
	|τ|=22	τ∼U(0,1.0 [ 2.5 MGA])	θA∼θD1∼θD2∼Exp(mean=0.025)
	|τ|=22	τ∼U(0,2.0 [ 5.0 MGA])	θA∼θD1∼θD2∼Exp(mean=0.025)
$${\mathcal{ℳ}}_{\mathit{\text{Exp}}}$$	|τ|=22	τ∼Exp(mean=0.058 [ 0.14 MGA])	θA∼θD1∼θD2∼Exp(mean=0.025)
	|τ|=22	τ∼Exp(mean=0.115 [ 0.29 MGA])	θA∼θD1∼θD2∼Exp(mean=0.025)
	|τ|=22	τ∼Exp(mean=0.173 [ 0.43 MGA])	θA∼θD1∼θD2∼Exp(mean=0.025)
	|τ|=22	τ∼Exp(mean=0.231 [ 0.58 MGA])	θA∼θD1∼θD2∼Exp(mean=0.025)
	|τ|=22	τ∼Exp(mean=0.289 [ 0.72 MGA])	θA∼θD1∼θD2∼Exp(mean=0.025)
	|τ|=22	τ∼Exp(mean=0.577 [ 1.44 MGA])	θA∼θD1∼θD2∼Exp(mean=0.025)

The distributions of divergence times are given in units of 4N_C generations followed in brackets by units of millions of generations ago (MGA), with the former converted to the latter assuming a per-site rate of 1 × 10⁻⁸ mutations per generation. For all of the ${\mathcal{ℳ}}_{\mathit{\text{msBayes}}}$ models, the priors for theta parameters are θ_A∼U(0, 0.05) and θ_D1,θ_D2∼Beta(1, 1)×2×U(0, 0.05. The later is summarized as ${\overline{\theta }}_{\mathit{\text{D}}}\sim \mathit{\text{U}}\left(\text{0, 0.05}\right)$. For the ${\mathcal{ℳ}}_{\mathit{\text{Uniform}}}$ and ${\mathcal{ℳ}}_{\mathit{\text{Exp}}}$ models, θ_A,θ_D1, and θ_D2 are independently and exponentially distributed with a mean of 0.025.