Skip to main content

Table 3 Marginal likelihood estimates from path-sampling and stepping-stone analyses and 2logeBF test statistics comparing evidence for different analyses

From: The first known fossil Uma: ecological evolution and the origins of North American fringe-toed lizards

One-rate analyses
 Model Path Sampling loge (Marg. Lik.) Stepping Stone loge (Marg. Lik.) Path Sampling 2logeBF Stepping Stone 2logeBF
 All fossil calibrations −111,014.69 −111,014.42 3.24 (181.58) 3.58 (181.60)
aNo sand lizard calibration −111,013.07 − 111,012.63 (178.34) (178.02)
 All fossil calibrations
 + Pleistocene calibration
−111,023.13 − 111,022.44 20.12 (197.64) 19.62 (198.46)
Two-rate analyses
 Model Path Sampling loge (Marg. Lik.) Stepping Stone loge (Marg. Lik.) Path Sampling 2logeBF Stepping Stone 2logeBF
aAll fossil calibrations −110,923.90 −110,923.62
 No sand lizard calibration −110,924.78 − 110,924.17 1.76 1.10
 All fossil calibrations
 + Pleistocene calibration
−110,932.31 −110,931.95 16.82 16.66
  1. The two-rate analysis with all fossil calibrations (in bold) is the best supported analysis overall. The favored analysis for each clock rate model is denoted by a a. The 2logeBF values are the favored model compared to the given model. For the one-rate analyses, the 2logeBF outside of parentheses are in comparison to the one-rate analysis without the sand lizard calibration, and 2logeBF in parentheses are in comparison to the overall favored analysis (two-rate analysis with all fossil calibrations)