Table 1.
Selection gradients of skin redness and darkness with standard errors, and corresponding GLM results. (For males, demographic variables were square-root-transformed and fecundity log-transformed. Estimates of selection gradients provided for males were assessed based on models using non-transformed measures of fecundity, however (see Material and methods for details). Note that regression coefficients and standard errors of the quadratic selection gradients were doubled (see Material and methods for more details). Significant selection gradient results are given in bold. AIC, Akaike information criteria.)
| linear | quadratic | correlational | |
|---|---|---|---|
| males | |||
| redness | β = 0.149 ± 0.110 | γii = 0.080 ± 0.173 | γij = −0.001 ± 0.134 |
| t = 1.329, p = 0.187 | t = 1.167, p = 0.247 | t =−1.224, p = 0.225 | |
| darkness | β = 0.162 ± 0.120 | γii =−0.229 ± 0.165 | γij = 1.229 ± 1.337 |
| t = 1.649, p = 0.103 | t =−0.092, p = 0.362 | t = 2.041, p = 0.044 | |
| age | t = 0.732, p = 0.466 | t = 0.840, p = 0.403 | t = 0.807, p = 0.422 |
| dominance ranka | t =−3.208, p = 0.002 | t =−2.961, p = 0.004 | t =− 2.531, p = 0.014 |
| group productivityb | t = 2.494, p = 0.015 | t = 2.640, p = 0.010 | t = 2.490, p = 0.015 |
| AIC | 169.82 | 171.67 | 168.88 |
| females | |||
| redness | β = 0.118 ± 0.030 | γii = 0.024 ± 0.030 | γij =−0.318 ± 0.272 |
| t = 3.897, p < 0.001 | t = 0.805, p = 0.423 | t =−1.169, p = 0.246 | |
| darkness | β = 0.013 ± 0.035 | γii =−0.000 ± 0.024 | γij =−0.407 ± 0.230 |
| t = 0.357, p = 0.722 | t =−0.002, p = 0.998 | t = 1.769, p = 0.081 | |
| age | t = 18.767, p < 0.001 | t = 18.526, p < 0.001 | t = 18.418, p < 0.001 |
| rank | t =−1.847, p = 0.069 | t =−1.702, p = 0.093 | t = 2.676, p = 0.351 |
| reproductive statusc | t = 2.122, p = 0.037 | t = 2.141, p = 0.035 | t = 2.676, p = 0.009 |
| AIC | 28.793 | 31.999 | 40.613 |
aFor males, dispersal rate is used as a proxy for dominance rank through a subject's life.
bFor males only. Group productivity is calculated based on the average number of offspring produced yearly in the groups in which each male resided.
cEstimates based on behavioural observation and date of birth during the following birth season (see Material and methods for details).