Despite generations of rearing in a laboratory environment, zebrafish (Danio rerio) retain innate preferences for environmental features that mimic their natural habitat (e.g. plants and gravel), but the effects of such preferred housing conditions on agonism are poorly understood. In this study, we worked with > 450 fish in 41 tanks to test two competing hypotheses about body size variation within shoals (a possible proxy for dominance hierarchy strength) and agonistic behaviour: A) the provision of preferred resources increases agonism (relative to levels in empty, barren tanks), possibly due to resource guarding behaviour, and B) barren housing conditions increase agonism (relative to levels in tanks containing preferred resources), possibly because subordinates are less able to hide from dominant fish. In Experiment 1, groups of 11–13 fish were housed for ∼ 1 year in 7 tanks in each of 3 treatments: empty lab-typical tanks, larger empty plastic tanks, and larger “well-resourced” plastic tanks containing preferred resources (gravel and artificial plants). Similarly, in Experiment 2, groups were housed in 5 tanks in each of 4 cross-factored treatments: empty glass aquaria, empty plastic tanks, well-resourced glass aquaria, and well-resourced plastic tanks. Live observations of agonistic behaviours and measures of body size variation (coefficient of variation for standard length and condition factor) were compared between treatments (using GLMs). In Experiment 1, empty lab-typical tanks exhibited significantly more agonistic behaviours than the other two treatment groups, which were not significantly different from each other. In Experiment 2, we found a significant interactive effect of housing treatment and tank material on agonism, where agonistic behaviours occurred most often in empty glass aquaria, possibly due to mirror-like reflections on glass walls. We therefore suggest that reflective surfaces in zebrafish housing systems should be minimized or visually disruptive resources like plants should be added to reduce agonism. Together both experiments also indicate that we need not be hesitant about adding preferred resources to captive zebrafishes’ housing environments, since they likely do not exacerbate agonism (and indeed may even reduce it). We also found two unexpected results: in Experiment 1, shoals exhibiting increased agonism had significantly lower coefficients of variation in standard length and in Experiment 2, fish in barren tanks grew significantly longer than those housed with preferred resources. Possible growth-related hypotheses are discussed and we suggest that high coefficients of variation in body size should not be used to infer high levels of agonism in zebrafish shoals.
Read full abstract