Fig. 4

Relative frequency and movement quantity of explorer individuals over the course of population spread. Each row corresponds to an experimental modality, in the same order as in Fig. 2. a–d Dots represent the relative proportion of each component, as inferred from the G-mixture (2) model fit (see Fig. 3), every 15 min, over time windows of 10 min. Solid lines delineate the corresponding 95% bootstrap confidence intervals. Estimates before about 1 h (shaded parts) may not be reliable as they correspond to a phase in which the population distribution was still close to Gaussian (Fig. 3); the presumed initial state is 100% residents. e–h Each dot represents the inferred MSD for the explorer component, using the same methodology as in a-d. Best supported linear or piecewise linear regressions are represented, together with their slopes (diffusion coefficients) and associated standard errors. Significantly different values harbour different superscript letters