Fig. 2

Results of the diving test used to assess group risk‐taking behaviour in the zebrafish selection lines in two consecutive trials at 230 and 240 dpf. Data are presented as cumulative time (bin in 30‐s period) spent by zebrafish shoals at the surface of the water (7 cm; see Figure S2; Video S1) once moved to a novel tank. Food has been added at 03:00 and the predator approached at 03:30. We first model the data with a non‐orthogonal quadratic polynomial mixed model by splitting the data into two sections: (i) before the approach of the simulated predator; (ii) after the approach of the simulated predator (a). Moreover, to assess differences among selection lines in risk‐taking behaviour as specific time points of the experimental assay, we implemented a linear mixed model for each 30‐s time bins (b). Finally, we calculated adjusted repeatability (i.e. after controlling for fixed effects of selection lines) of risk‐taking behaviour for each 30‐s time bin across time (230–240 dpf) by using the same mixed effects models implemented before (a). Repeatability scores (right axis in a) are presented as empty circles (when they were not significantly different from zero) and solid circles (when they were significantly different from zero; see also Tables S2–S4 and Figure S4 for further details and partitioning of variance components). The different colours represent the selection lines (N = 12 for each selection line) together with confidence intervals (dashed coloured lines in a), median (points in b) and lower and upper quartiles (vertical lines in b). Bird pic source: https://www.piqsels.com. Significant difference is indicated by black horizontal lines (†p < 0.10; *p < 0.05; **p < 0.01)