Isolation associated aggression--a consequence of recovery from defeat in a territorial animal

Abstract

Population density has profound influences on the physiology and behaviour of many animal species. Social isolation is generally reported to lead to increased aggressiveness, while grouping lowers it. We evaluated the effects of varying degrees of isolation and grouping on aggression in a territorial insect, the Mediterranean field cricket, Gryllus bimaculatus. Substantiating early observations, we show that dyadic contests between weight-matched, adult male crickets taken from groups rarely escalate beyond threat displays, whereas interactions between pairs of previously isolated crickets typically escalate to physical fights lasting several seconds. No significant differences were found between 1, 2 and 6-day isolates, or between individuals grouped for a few hours or lifelong. Unexpectedly, crickets grouped in immediate proximity within individual mesh cages that precluded fighting while permitting visual, olfactory and mechanical, antennal contact, were as aggressive as free isolates. This suggests that reduced aggression of grouped animals may be an acquired result of fighting. Supporting this notion, isolated crickets initially engage in vigorous fights when first grouped, but fighting intensity and duration rapidly decline to the level of life-long grouped crickets within only 10 min. Furthermore, grouped crickets become as aggressive as life-long isolates after only 3 hours of isolation, and on the same time course required for crickets to regain their aggressiveness after social defeat. We conclude that the reduced aggressiveness of grouped crickets is a manifestation of the loser effect resulting from social subjugation, while isolation allows recovery to a state of heightened aggressiveness, which in crickets can be considered as the default condition. Given the widespread occurrence of the loser effect in the Animal Kingdom, many effects generally attributed to social isolation are likely to be a consequence of recovery from social subjugation.

Figure 1. Effects of crowding and isolation on cricket aggression.

(A, B) Bar graphs giving the level and duration respectively (circles: medians, bars: interquartile range) of aggression for encounters between pairs of male, weight matched crickets taken from the same test groups (from left to right): crowded colony (dark blue), grouped (light blue: groups of 30, 20, and 10 individuals as indicated), caged-grouped (light blue, hatched), caged-isolated (red, hatched), isolated (red: for 1, 3 and 6 days as indicated). Numbers in parentheses above the x-axis in A give the number of pairs or crickets for each group. Significant differences between groups are indicated (Mann–Whitney U-test, * p < 0.05, ** p < 0.01, *** p < 0.001, ns not significant).

Figure 2. The influence of females.

(A, B) Bar graphs comparing the level of aggression and fight duration respectively (circles: medians) exhibited by pairs of weight matched grouped crickets (light blue bars, compare also with Figure 1) compared to pairs of weight matched crickets taken from groups of 20 males together with 10 mature, adult, virgin females (stippled light blue bars). (C, D) Bar graphs comparing the level and duration of aggression respectively (circles: medians) exhibited by pairs of weight matched, isolated male crickets (light red bars, compare also with Figure 1) compared to pairs of weight matched isolated males that were each housed with an individual mature, adult, virgin female (stippled light red bars). The third bar in series (light red, darkly stippled) depicts fights between isolated males housed with females, as previous, whereby the females were also present in the fighting arena. Significant differences between groups are indicated (Mann–Whitney U-test, * p < 0.05, ** p < 0.01, *** p < 0.001, ns not significant). Differences that do no survive Bonferroni correction for alpha are placed in parentheses (*).

Figure 3. Acquisition of the crowding effect.

(A, B) Bar graphs giving the level and duration respectively (circles: medians, bars: interquartile range) of aggression for encounters between pairs of male, weight matched, initially isolated crickets after being re-grouped for various lengths of time (pale blue bars). Each bar gives data accumulated from 5 separate observations, each with 10 different re-grouped crickets, for which all interactions occurring within the observation period were evaluated (n, given above the x-axis in A). Data for isolated crickets (red bar) and crickets taken from our crowded colony (blue bar) are included for comparison. Significant differences between groups are indicated (Mann–Whitney U-test, * p < 0.05, ** p < 0.01, *** p < 0.001, ns not significant). ). Differences that do no survive Bonferroni correction for alpha are placed in parentheses (*).

Figure 4. Recovery from crowding and losing.

(A, B) Bar graphs comparing the level and duration respectively (circles: medians, bars: interquartile range) of aggression for encounters between pairs of male, weight matched crickets that were previously grouped (blue bars) or had lost a previous fight (losers; yellow bars). Each individual bar gives data from a different set of crickets each observed at different times after either isolation or losing; the number of pairs tested for each is given in parentheses under the x-axis in A. The differences between grouped and loser crickets for each time period are not statistically significant (Mann–Whitney U-test).