Delayed colonisation of Acacia by thrips and the timing of host-conservatism and behavioural specialisation

Table 2 Bayes factor comparisons of timing models

Hypothesis	Stepping-stone sampling	BF for H₁₄	Path sampling	BF for H₁₄
ln(H₁₄)	−15934.63	NA	−15933.89	NA
ln(H₂₀)	−15934.94	0.6	−15934.45	1.2
ln(H₂₅)	−15936.31	3.4	−15935.68	3.6
ln(H₅₀)	−15952.21	35.2	−15952.00	36.2

Path and stepping-stone sampling was used to estimate the marginal likelihoods of divergence models inferred with BEAST. Bayes factor tests were made between the 14 Mya origin of Acacia thrips (best model = H₁₄) and alternative timings based on penalised likelihood and Bayesian inferences. A BF indicates the strength of support for the best over the worst (ln) model.

ISSN: 2730-7182