Skip to main content
Fig. 2 | BMC Evolutionary Biology

Fig. 2

From: Wild cricket social networks show stability across generations

Fig. 2

Plots of the predictive distance for the between year comparisons against the difference in time between each pair of compared years (a & b) and the difference in population size between each pair of compared networks (c & d). Plots (a & c) show this relationship for predicted path lengths, (b & d) for predicted clustering coefficient. Plotted are the means of the 100 predictive distances for each comparison: the difference between the mean of the simulated values and the real value. Distance in years did not affect the ability of models to predict other networks (no correlation: Mantel test, rho = −0.169 & -0.107), but were worse at predicting the path lengths of other networks when they were initially parametrised on networks with different population size (increased predictive distance with increased difference in population size: Mantel test, rho = 0.481 & 0.488) (see results). The lines are from simple regressions of the variable on the x axis on the predictive distance, which are not informed by the Mantel tests but help visualise the result

Back to article page