Model | gof(U^{2}) | P | μ_{1} (degrees) | s (degrees) | p_{1} | μ_{2} (degrees) | s (degrees) | p_{2} |
---|---|---|---|---|---|---|---|---|
1VM | 0.644 | <0.01 | 1.26±178.37 | 74.51 | 1.00 | - | - | - |
2VM | 0.098 | 0.16 | −2.91±6.55 | 28.30 | 0.58 | 177.72±8.83 | 32.45 | 0.42 |
For each model, a goodness of fit was calculated and significance tested against a parametric bootstrap as described in the Materials and methods. The mean direction of each mode (μ±95%CI) as well as the angular standard deviation (s) and proportion of data fit by that mode (p_{n}) are shown. The fitting of mixtures was stopped when the probability exceeded 10%. P represents the probability that the data are better fit by a model containing the corresponding number of modes than by a model containing additional modes.