Consistency of fish‐shoal social network structure under laboratory conditions

Abstract

We investigated the consistency of association network structure for groups of sticklebacks Gasterosteus aculeatus. Each group was observed twice and we varied the duration between observations and the size of the experimental arena that they were observed in. At the dyad level, we found positive correlations between dyad interaction frequencies across observations. At the group level we found variation in four network metrics between observations, but only in treatments where the duration between observations was short. Specifically, fish formed more and smaller groups in the second observation in this treatment. Fish were also organized into more subunits in the larger arenas. Finally, we saw positive correlations between some group network metrics across observations suggesting relative consistency at the group level. There are several processes that might drive these interaction patterns. Our findings have implications for experimental design and the comparison and integration of findings of experiments from different studies carried out under different conditions.

Comment on “Consistency of fish‐shoal social network structure under laboratory conditions (Gaffney & Webster, 2018)”

“…Another interesting aspect of the study by Gaffney and Webster () is that they document, albeit tentatively, that shoals differed repeatedly from one another in group‐level network measures, particularly the size of their largest subunit. Although the study was not properly set up for investigating variability among groups, as the authors acknowledge, this finding suggests that groups, even small communities, may differ consistently from one another.…”

“…While much work on the topic has focused on understanding the emergence of group‐level properties, relatively little is known about the variability among groups and about collective dynamics and structure at the community level. In a recent issue of this journal, Gaffney and Webster () present some simple yet neat experiments with shoals of three‐spined stickleback Gasterosteus aculeatus L. 1758 that, although perhaps a bit limited in terms of the size and number of groups tested, present some promising insights into social network characteristics at the dyadic and community level.…”

“…Gaffney and Webster () looked at fission‐fusion characteristics by recording the number and distribution of smaller groups when shoals split. They found that, at the dyadic level, pairwise association strengths were positively correlated for most shoals, suggesting that even while shoals split and merged, individual fish tended to associate with the same individuals over time.…”

Species‐specific schooling behaviour of fish in the freshwater pelagic habitat: an observational study

“…Understanding of schooling behaviour in the pelagic zone of freshwater ecosystems is rather sparse (Milne et al ., 2005) compared with marine ecosystems where their high economic importance has resulted in increased attention ( e.g., Sund 1935). Recent studies were focused mostly on laboratory experiments ( e.g., Gaffney & Webster, 2018; Krause et al ., 2000a) and theoretical models (Couzin et al ., 2002; Hensor et al ., 2005) which might not bring as accurate results as field observations could, but those are also still lacking (reviewed in Krause et al ., 2000a).…”

Reply to Jolles (2018): Comment on “Consistency of fish‐shoal social network structure under laboratory conditions (Gaffney & Webster, 2018)”

“…In a paper this year (Gaffney & Webster, ) we presented a series of experiments in which we investigated the effects of arena size and inter‐observation duration upon basic social network structure metrics and related measures for shoals of threespine sticklebacks Gasterosteus aculeatus L. 1758. We were motivated to perform this experiment by the question of whether such measurements can meaningfully be compared between experiments run by different researchers and under different conditions.…”