Artificial mass loading disrupts stable social order in pigeon dominance hierarchies

1. Introduction

Many animals live and travel in groups [1,2]. The benefits of group living can include enhanced vigilance and predator detection [1,2], energetic saving through positive aero- or hydro-dynamic interactions [3,4] and increased foraging efficiency [5,6]. Within a group, however, individual characteristics in personality, morphology and physiology can lead to conflict. A product of such conflicts can be the emergence of dominance hierarchies [7], and these dominance relationships are a frequently documented characteristic of group living [8].

A dominance hierarchy within a group can confer benefits to all its members by decreasing the severity and incidence of physical conflicts [9]. By reducing the time devoted to agonistic encounters, time can be invested in other important behaviours such as maintenance, vigilance and foraging [10]. Dominance hierarchies within animal societies are frequently arranged in a linear fashion; higher ranked individuals dominate all individuals of lower rank [8]. Linear hierarchies have often been linked to parameters such as body mass/size [11] and have shown to be either stable [8] and unstable [12] over time. The degree to which there is apparent temporal variation in dominance hierarchies appears linked to certain life-history traits, with animal groups either confined to a limited area or living together for prolonged periods of time favouring stable hierarchies [13].

How dominance and body mass interact both within and between seasons is not fully understood (although see [14]). Given that body mass can vary substantially throughout the annual cycle in response to key life-history events such as breeding, moult and migration, how these changes in body mass are reflected in the stability of group hierarchies and individual positions therein is likely to have significant consequences on overall group dynamics and levels of aggression. Therefore, a better understanding of how responsive—and the rapidity of response—dominance hierarchies are to smaller scale instantaneous changes in body mass has the potential to offer insight into both collective and individual energy expenditure.

Using a captive flock of homing pigeons (Columba livia), we tested whether (i) group dominance hierarchies were stable over successive years in a closed population, (ii) whether any hierarchical structure was directly related to body mass and (iii) if linear hierarchies were stable and correlated with body mass, whether they could be disrupted by artificial instantaneous manipulations of body mass.

2. Material and methods

3. Results

All dominance hierarchies for non-weighted sampling sessions (N = 7) over the 3 years were significantly linear (table 1) and strongly correlated with body mass (figure 1a; electronic supplementary material, tables S1 and S2). Taking the mean rank and body mass for each individual bird for the seven sampling periods, body mass was significantly correlated with rank position (figure 1b). David's score across the seven (not mass loaded) trials was significantly repeatable (R = 0.78 ± 0.07 (s.e.m.), 95% CI: 0.60–0.86, p < 0.001), as was body mass (R = 0.96 ± 0.02 (s.e.m.), 95% CI: 0.91–0.98, p < 0.001). Spearman's rho comparisons supported the repeatability of David's score across non-weighted sessions, with comparisons between each unweighted sampling session being significantly correlated (ρ range 0.78–0.99; table 2), indicating that rank in one (unweighted) session was a good predictor of rank the following session.

• Download figure
• Open in new tab
• Download PowerPoint

Upon application of the artificial mass, the dominance hierarchy changed significantly (tables 1 and 2; figure 1c,d) but remained linear; the dominance hierarchy observed when nine birds were artificially weighted was significantly different to all seven non-weighted hierarchies (Spearman's rho, ρ, range 0.04–0.09; table 2). Artificial mass loading significantly increased an individual's dominance score (LMM: d.f. = 118, t = 4.52, p < 0.001) by 42.85 ± 9.47 (s.e.m.) (David's score)). The nine individuals who were artificially mass manipulated significantly, on average, increased their aggression (figure 1c,d), resulting in a significant increase in their rank (figure 1c,d). On average, individuals that were artificially mass loaded increased their number of aggressive behaviours by 134 ± 261.6% (s.d.). Not all birds increased their aggressive behaviours (figure 1c,d); the maximum decrease in aggressive behaviours observed by artificially mass loaded birds was 38.33%. Of those birds which did increase aggressive behaviours, the maximum and minimum increases were 750% and 11.3%, respectively (electronic supplementary material).

There was a significant interaction between sex and mass loading (d.f. = 117, t = 3.72, p < 0.001). This relationship was driven by the males increasing their dominance score when mass loaded (Tukey post hoc test for lme in ‘emmeans' package: non-mass loaded males versus mass loaded males, estimate = 66.88 ± 11.1 (s.e.m.), t.ratio = 6.043, p < 0.001), whereas mass loaded females showed no difference to their non-mass loaded behaviour and subsequent dominance rank (estimate = −4.56 ± 15.6 (s.e.m.), d.f. = 117, t.ratio = 0.292, p = 0.991).

The application of artificial mass resulted in an overall increase in aggression and aggressive encounters in the flock (table 1); the total number of aggressive interactions recorded during the artificial mass loading sampling session was nearly double (N = 2580) that of the nearest number of aggressive interactions recorded during an unweighted session (N = 1321, session 7; table 1; electronic supplementary material, figure S1). During the artificial mass-manipulated session, the dominance hierarchy remained linear (table 1). All eight dominance hierarchies—both weighted and unweighted—had significant steepness (p < 0.001; table 1), indicating that the agonistic relationships of the pigeons were organized in a steep and linear fashion (the size of the absolute differences between adjacently ranked individuals in David's score is large). The composition of aggressive behaviours remained significantly stable between sessions, and between the non-weighted and weighted trials (electronic supplementary material), with none of the key five behaviours measured (pecking, chasing, beak grab, neck pull and wing slap) changing significantly in terms of relative contribution to overall aggressive behaviours recorded (chi square, all weighted to non-weighted comparisons, p = 0.99; electronic supplementary material).

4. Discussion

Over a 31-month period, the dominance hierarchy of the pigeon group did not significantly change, with individuals retaining their position within the hierarchy throughout the experimental period. Previously, it has been demonstrated that in animal groupings of less than approximately 10 individuals, stable hierarchies are more commonly observed than in larger groups [27]. The linear dominance hierarchy in the pigeons was significantly related to body mass. There is no clear pattern yet determined as to why body mass is such a strong determinant of dominance in some species but not others [11,28]. It is possible that body mass is significantly correlated with dominance in species where secondary-sexual ornamentations are less pronounced, and as a result, signalling is less clear. In such cases, body mass may become more of an important indicator of fitness. The hierarchy returning to its stable structure upon the removal of the additional mass load suggests that no carry-over or ‘memory' effects of mass loading persist and implies an instantaneous neurological feedback mechanism regarding changes in body mass (e.g. [29]).

It is possible that the addition of the extra mass to the backs of the subordinate birds aggravated or stressed the birds, causing them to exhibit higher levels of aggression. During the addition of the artificial mass, the subordinate birds did not show any obvious signs of aggravation at the lead weights attached to them, nor did they try to peck or remove them, either on themselves or on conspecifics (S.J.P. personal observation), suggesting this is an unlikely explanation for their increased aggression. Similarly, the composition of aggressive behaviours did not change between weighted and unweighted sessions, suggesting behaviours were not more focused on the back, where the weights were attached. An alternative explanation, however, is that the addition of artificial mass—although only for a short period—increased the energetic requirements of the weighted birds, thus requiring them to be more aggressive to ensure adequate access to food [30,31]. Such a theory is akin to ‘lead according to need', an idea which has previously linked to motivation and leadership in group behaviour [32].

Only males responded to the artificial mass loading by significantly increasing their aggressive behaviour, while females did not seemingly respond, suggesting that increasing aggression in response to artificial mass loading is sex specific. Previously, it has been demonstrated that injections of testosterone into male pigeons did not make male pigeons more aggressive or dominant, [33], yet a perceived possible increase in physiological condition through the addition of mass in the present study did elicit a response. This sex-specific response may be linked to competition for females, with female pigeons preferentially selecting males for partnering who hold dominant positions within a hierarchy [34]. An avenue worthy of further investigation is the impact that the pairing status of an individual has on their respective rank, as it has been previously demonstrated in birds that being paired increases your rank within a hierarchy [35,36].

The present study demonstrates the plasticity of aggressive traits, and the rapidity with which they can be modified based on physiological condition. Fruitful future investigations would be to ascertain the attributes that lead to greater body masses in wild-type scenarios and in turn greater dominance. The ‘prior attributes' hypothesis [8,27,28], for example, suggests hierarchies are predetermined by personality or physiological differences in dominance ability. This in turn may be linked to leadership during flocking and associated energy expenditure [37–40]. How natural seasonal variations in body mass [41–43] manifest in terms of dominance and general social behaviour would further explore the interactions between individual physiology, energetics and social behaviour. Moreover, experiments that supplementary feed specific individuals over a longer period of time to increase body mass may yield different results with respect to the changes in their respective ranks. Our study focused on only one flock of birds, and to determine the full nature of these instantaneous changes in body mass, further studies are needed with larger sample sizes, both in terms of number of flocks and sampling sessions where mass was added, and ideally additional species.

Ethics

All experiment protocols were approved by the R.V.C. Ethics and Welfare Committee.

Data accessibility

All data are available in the electronic supplementary material.

Authors' contributions

S.J.P. and A.M.W. were involved in conceptualisation; S.J.P. was involved in methodology; J.R.U. and A.M.W. were involved in resources; S.J.P. was involved in data collection; S.J.P, C.R.W. and D.W.E.S. were involved in formal analysis; S.J.P. was involved in writing the original draft; S.J.P., J.R.U., C.R.W., D.W.E.S. and A.M.W. were involved in writing, reviewing and editing. All authors are in agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Competing interests

We declare we have no competing interests.

Funding

Funding was provided by an EPSRC grant to A.M.W. and J.R.U. (EP/H013016/1) and a Wellcome Trust Fellowship (095061/Z/10/Z) to J.R.U.