Results from the ANOVA after model simplification (Table 3) identified Stage, Stage:Sex, Stage:Trial (not further discussed) and Stage:Food as significant factors (Pr (>F) < 0.05) for mean swimming distance. Food as a single factor was non-significant for swimming distance. Fish-Length in mm as a fixed covariate did not have a significant effect on swimming distance. Generally, there was an increase in swimming distance across stages with a slight decrease during the adult stage. This development was mostly significant for the changes in swimming distance from the juvenile to the subadult stage. Only low-food females constantly increased swimming distances across all stages as the graph shows (Figure 25). Furthermore, low-food fish tended to swim greater distances than high-food fish. Especially the differences comparing high-food adults and low-food adults become in Figure more apparent. Summarizing these results, a peak in swimming distance was present for most groups during their subadult stage, except low-food females who peaked upon adulthood.
- Exploration 2 (Moving time) -
Results from the ANOVA after model simplification (Table 5) identified Stage, Stage:Food and Stage:Trial (not further discussed), as significant factors (Pr (>F) < 0.05) for moving time. Fish-Length in mm as a fixed covariate did not have a significant effect on time spent moving. Considering the differences in lsmeans for the Sex:Food interaction, females exposed to the high-food treatment allocated more time into swimming than males of the same treatment group. This difference was not observed for fish of the low-food treatment. However, comparing males of both groups shows that low-food males spent significantly more time swimming than high-food males. The increase in swimming time with stage becomes apparent as well as the mentioned differences between low and high-food males. Finally, there was a steady increase in time spent swimming for low-food females even upon maturation, as opposed to the other groups reducing swimming time for a non-significant amount (Figure 26).
- Shyness -
Results from the ANOVA after model simplification (Table 7) identified Stage, Sex, Stage:Sex, Stage:Food and Stage:Trial (not further discussed), as significant factors (Pr (>F) < 0.005) for time spent in outer area. Fish-Length in mm as a fixed covariate did not have a significant effect on time spent in outer area. Starting with the single factors, stage in this case, adult and juvenile specimen spent less time in the outer area than subadults, or conversely adults and juveniles spent significantly more time in the centre of the arena than subadult fish (Figure 27, Table 8). Overall, fish spent significantly more time in the outer area of the arena when reaching their subadult stage. This time allocation was reversed for male adults upon maturation while adult females mostly preferred the outer area. Figure also shows that high-food fish spent more time in the outer area than low-food fish though this difference was non-significant. The significant differences for food were confined in the Stage:Food interaction. The main trend here was that the high-food treatment, especially for subadults, made them stick longer to the outer area than low-food subadults or compared to previous and later stages.
- Activity -
Results from the ANOVA after model simplification (Table 9) identified Stage, Stage:Sex, Trial and Stage:Trial (both not discussed) as significant factors (Pr (>F) < 0.05) for mean actual velocity. Food as a single factor or interaction was non-significant for actual velocity. Fish-Length in mm as a fixed covariate did not have a significant effect on actual velocity. The bar graph in Figure, generally indicates an increase in actual velocity with advancing stage and a slight decrease for the adult phase. Just considering actual velocity over stages, both adults and subadults had a significant higher actual velocity than juveniles (0.3 (log cm/s) higher) as the log-transformed data shows (Table 10). The more detailed Stage:Sex interaction confirms the implications made by stage as a single factor. Adult females and males both swam faster than juvenile males and females (P < 2e-16). Interestingly adult female fish, had a significantly higher actual velocity, especially due to the low-food females (Figure 28, Table 10) when compared to adult males. Adult male fish featured a reduced actual velocity compared to the previous stage, while females maintained a higher speed throughout life. Values for actual velocity peaked during the subadult life-stage for male fish.
Repeatability and variance components - Exploration 1 (Swimming distance) -
Swimming distance was repeatable over all life-stages, food treatments and sexes (R = 0.287, 95% CI (0.205, 0.366)). Swimming distance was also repeatable for each single Stage:Sex:Food combination (all R > 0.347, Figure 29). While the among-individual variance statistically significantly increased with age (Table 11, Figure 29 B), repeatability and within-individual variance showed no relationship with age (Table 11, Figure 29 A and C). Neither food nor sex had an influence on repeatability or any of the variance components as seen in Table 11.
- Exploration 2 (Moving time) -
Moving time was moderately repeatable over all life-stages, food treatments and sexes (R = 0.237, 95% CI (0.163, 0.318). Moving time was also repeatable for each single Stage:Sex:Food combination (all R > 0.077, Figure 30 ). The CI for all low-food male stages included 0. While the among-individual variance and repeatability were statistically significantly higher in the high-food treatment (Table 12, Figure 30 A and B), within-individual variance showed no relationship for food (Table 12, Figure 30 C). Neither stage or sex had an influence on repeatability nor any of the two variance components (Table 12).
- Shyness-Boldness -
Time spent in outer area was scantly repeatable over all life- stages, food treatments and sexes (R = 0.158, 95% CI (0.097, 0.227). Time spent in outer area was also repeatable for each single Stage:Sex:Food combination (all R > 0.151, Figure 31). While the among-individual variance statistically significantly increased with age (Table 13, Figure 31 B), repeatability and within-individual variance showed no relationship with age (Table 13, Fig. 31 A and C). Neither food or sex had an influence on repeatability nor any of the two variance components (Table 13).
- Activity -
Actual velocity distance was moderately repeatable over all life- stages, food treatments and sexes (R = 0.274, 95% CI (0.196, 0.353). Actual velocity was also repeatable for each single Stage: Sex:Food combination (all R > 0.251, Figure 32). While the among-individual variance statistically significantly increased with age (Table 14 , Figure 32 B), repeatability and within-individual variance showed no relationship with age (Table 14, Figure 32 A and C). Neither food or sex had an influence on repeatability nor any of the two variance components (Table 14).