Stasis in Pleistocene Owls: Tables

TABLE 1. Basic statistics of La Brea TMTs of Athene cunicularia. SD: Standard Deviation; CV: Coefficient of Variation; N: sample size; TMT: tarsometatarsus.

TABLE 2. Basic statistics of La Brea TMTs of Asio otus. SD: Standard Deviation; CV: Coefficient of Variation; N: sample size; TMT: tarsometatarsus.

TABLE 3. Mann-Whitney U tests for each dimension of TMTs for Athene cunicularia, comparing measurements from each time interval to the rest of the pooled samples; jackknife method was used instead of straight pooled variation due to highly uneven sample sizes. Bold face indicates results that are significantly different at p < 0.05 with Bonferroni correction applied. The significant differences given by this test are consistent with the significance levels reported by the Kruskal-Wallis test (df = 4). Corrected significance level: p < 0.01. Depth and robustness measurements were significantly different from pooled samples for the 9 ka and modern samples.

TABLE 4. Mann-Whitney U tests for each dimension of TMTs for Asio otus, comparing measurements from each time interval to the rest of the pooled samples; jackknife method was used instead of straight pooled variation due to highly uneven sample sizes. Corrected significance level: p < 0.00625. No differences were significant by either test.

TABLE 5. Time-series analysis, given in Akaike weights, for Athene cunicularia. GRW: general random walk; URW: unidirectional random walk; for “strict stasis” definition, see Hunt et al. (2015). The favored model for each time series is shown in bold face. Strict stasis was preferred for length and width measurements. Depth measurements fit best as ordinary stasis (i.e., true mean not constrained) due to the larger measurements in the 0 ka group. Robustness, derived from the combination of the three measurements, favored an unbiased random walk model.

TABLE 6. Time-series analysis, given in Akaike weights, for Asio otus. GRW: general random walk; URW: unidirectional random walk; for “strict stasis” definition, see Hunt et al. (2015). The favored model for each time series is shown in bold face. All time series best fit a strict stasis model except width, which favored ordinary stasis due to the large difference in the 12 ka sample.