Table 3.
Results of bootstrap analyses. On the left, the CV, Max/Min, R%, and mean method scores for the fossil footprints and comparative samples are presented.
| Sample | n | Mean method quotient | CV | Max/Min | R% | US Army percentiles | Daasanach percentiles | Gorilla percentiles | |||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Mean method quotient | CV | Mean method quotient | CV | Mean method quotient | CV | ||||||
| Ileret (FwJj14E UFL) | 16 | 1.115 | 8.1 | 1.488 | 0.391 | 0.427 | 0.918 | 0.655 | 1.000 | 0.009 | 0.024 |
| Ileret (pooled) | 25 | 1.142 | 9.0 | 1.488 | 0.392 | 0.909 | 0.999 | 0.991 | >1.000 | 0.010 | 0.036 |
| Laetoli | 6 | 1.194 | 11.6 | 1.364 | 0.314 | 0.970 | 0.994 | >1.000 | >1.000 | 0.508 | 0.595 |
| Engare Sero | 21 | 1.146 | 8.2 | 1.296 | 0.270 | 0.925 | 0.956 | 0.992 | 1.000 | 0.018 | 0.014 |
| Walvis Bay | 21 | 1.009 | 0.5 | 1.009 | 0.009 | 0.008 | 0.008 | 0.013 | 0.013 | 0.011 | 0.011 |
| Willandra Lakes | 7 | 1.069 | 4.6 | 1.111 | 0.106 | 0.097 | 0.140 | 0.152 | 0.243 | 0.017 | 0.027 |
| P. troglodytes | 37 | 1.108 | 6.3 | 1.090 | 0.249 | 0.232 | 0.257 | 0.480 | 0.719 | <0.001 | <0.001 |
| P. paniscus | 8 | 1.057 | 3.7 | 1.121 | 0.111 | 0.022 | 0.024 | 0.036 | 0.053 | 0.003 | 0.004 |
| Gorilla | 31 | 1.206 | 11.1 | 1.298 | 0.412 | >1.000 | >1.000 | >1.000 | >1.000 | — | — |
| Daasanach | 29 | 1.124 | 6.9 | 1.219 | 0.249 | 0.596 | 0.546 | — | — | <0.001 | <0.001 |
| US Army | 2000 | 1.121 | 6.8 | 1.507 | 0.407 | — | — | >1.000 | >1.000 | <0.001 | <0.001 |
On the right are the results of bootstrapping from the US Army, Daasanach, and Gorilla comparative samples. For each fossil footprint locality, a ‘fossil sample’ of equal size was drawn from each comparative sample 1000 times. We present the relative percentile of each fossil mean method quotient and CV within the bootstrapped comparative sample, which corresponds to the probability of drawing that quotient or CV from within the given bootstrapped sample. Hypothesis tests were one-tailed, so when a fossil sample falls at a percentile greater than 0.95 within the modern human distributions, that sample has a significantly higher level of dimorphism. Likewise, when fossil samples fall below 0.05 in the Gorilla distribution, this indicates a significantly lower level of size dimorphism.
Note: the sample sizes (column n) are the numbers used in the analysis, after the smallest, presumably juvenile, individuals had been removed (see text for explanation of methods).