Figures and data
Experimental design.
(A) Experiment 1: Triadic cohabitation and formation of affiliative relationships. Step 1: three male crows with already established relative dominance relationships through each dyadic encounter but no affiliative relationships were introduced into an aviary for two weeks. These three crows are referred to as high (orange), middle (green), and low (blue) based on their relative dominance. Step 2: pre-test dyadic cohabitation for 3 days to ensure any affiliative relationships noted were not due to prior dyadic interactions. Step 3: triadic cohabitation for two weeks to observe the formation of affiliative relationships, characterised by reciprocal allopreening between two specific crows in each triad. Step 4: post-test dyadic cohabitation for 3 days to confirm the persistence of these relationships. If no affiliative relationship was formed in any dyad, that triad would not proceed to Experiment 2, the behavioural pharmacological experiment.
(B) Experiment 2: Effects of V1a receptor antagonism on affiliative relationships between dominant and subordinate individuals. Steps 5 and 6: V1aR antagonists and control vehicle solutions were administered peripherally to one of the two individuals with established affiliative or unaffiliative relationships from Experiment 1, prior to their dyadic interaction. In each trial, the drug to either the dominant or subordinate individual, and free interactions between the two were compared pre- and post-administration. After completing all vehicle trials, the antagonist was administered following the same protocol. Allopreening, aggressive behaviours and submissive vocalisations, were coded for analysis to determine the influence of the V1aR antagonist on the maintenance of affiliative behaviours and the emergence of aggression or submission.
Determination of relative dominance in dyads.
The numbers of aggressive (A), submissive (B) and allopreening behaviours (C) across the three trials of dyadic social encounters. The numbers of aggressive behaviours (jab, peck, aggressive vocalisation and displacing approach) by the dominant (thick solid black line) and submissive behaviours (submissive begging vocalisation and avoidance) by the subordinates (thick dotted black line) decreased through the three encounters, suggesting the stable formation of dyadic dominance relationships. Allopreening behaviours were not observed in any dyads. Grey-coloured solid and dotted lines indicate dominant and subordinate individual data, respectively. Data are represented with mean ± s.e.
Individuals used in Experiments 1 and 2.
Birds OW, YA, VV, BB, #28, and KR are included in two different triads. Triad #6 (RM, YA, VA) did not receive Experiment 2 (see text for the details). n.a. indicates data is not available.
Triadic cohabitation facilitated the formation of an affiliative relationship between two particular males.
(A) Total number of allopreening behaviours in each dyad across the experimental periods: pre-tests, 2-week triadic cohabitation (early, middle, and last), and post-tests. Data plots are presented as mean ± 1 s.e. Panels indicate triads (#1−#8). Bright-coloured plots (pink, green, purple, blue, orange and red) in the #1−#6 triads represent dyads that showed the formation of affiliated relationships. Darker plots denote the dyads that did not form affiliative relationships.
(B) Scatter plots of allopreening behaviours initiated by dominants (x-axis) versus subordinates (y-axis) of dyads in the pre-tests (opened) and the post-tests (shaded). Coloured plots represent dyads wherein affiliative relationships were formed. Allopreening occurred exclusively in the coloured dyads after the cohabitation (i.e., post-test). The colours correspond to the dyad in Figure 3A. Data are presented as the mean with natural log-transformed values.
Outputs of discriminant analysis.
Effects of V1a receptor antagonist (V1aRA) on the number of allopreening behaviours in affiliated and unaffiliated dyads.
In affiliated dyads, allopreening behaviours initiated by both dominant (DOM, left panel) and subordinate (SUB, right panel) crows exhibited a significant decrease in both the high-dose and low-dose V1aRA conditions post-administration, compared to pre-administration. Conversely, in the vehicle condition, no significant difference was observed in allopreening frequency between pre- and post-administration phases. In contrast, within unaffiliated dyads, V1aRA administration exerted no effect on allopreening for either the dominant (DOM, left panel) or subordinate (SUB, right panel) individuals.
The plots illustrate individual data for each trial. The upper, middle, and lower bounds of the boxes represent the 75th, 50th, and 25th percentiles, respectively, with whiskers extending to the maximum and minimum values. Numbers in the brackets below the boxes denote the number of samples (no. of birds × no. of trials). Asterisks denote statistically significant differences (Wald χ2statistics, *** p < 0.001). The colours correspond to the affiliated dyads shown in Figure 3.
Outputs of GLMMs for allopreening behaviour in affiliated dyads.
Post-hoc pairwise comparisons of pre- and post-administration on allopreening behaviour by drug conditions and subject dominance status.
Post-hoc pairwise comparisons of drug conditions on allopreening behaviour by subject dominance status in the post-administration phase.
Effects of V1aRA on the number of aggressive behaviours in affiliated and unaffiliated dyads.
In affiliated dyads, aggressive behaviour initiated by the dominant individual (DOM, left panel) exhibited a significant increase following administration of the high-dose V1aRA, as compared to pre-administration phase. In contrast, no such increase was observed in the low-dose or vehicle conditions. Additionally, this elevation in aggression was not observed in any condition for the subordinate individual (SUB, right panel). Furthermore, in unaffiliated dyads, no significant effects were detected in either the dominant or subordinate individuals under any of the administered conditions.
The plots illustrate individual data for each trial. The upper, middle, and lower bounds of the boxes represent the 75th, 50th, and 25th percentiles, respectively, with whiskers extending to the maximum and minimum values. Numbers in the brackets below the boxes denote the number of samples (no. of birds × no. of trials). Asterisks denote statistically significant differences (Wald χ2statistics, *** p < 0.001). The colours correspond to the affiliated dyads shown in Figure 3.
Outputs of GLMMs for aggressive behaviour in affiliated dyads.
Post-hoc pairwise comparisons of pre- and post-administration on aggressive behaviour by drug conditions.
Post-hoc pairwise comparisons of drug conditions on aggressive behaviour in the post-administration phase.
Outputs of GLMMs for aggressive behaviour in unaffiliated dyads.
Effects of V1aRA on the number of submissive begging vocalizations in affiliated and unaffiliated dyads.
In affiliated dyads, submissive vocalization initiated by the subordinate individual (SUB, right panel) exhibited a significant increase following administration of the high-dose V1aRA, as compared to pre-administration phase. In contrast, no such increase was observed in the low-dose or vehicle conditions. Additionally, this elevation in submissive vocalization was not observed in any condition for the dominant individual (DOM, left panel). Furthermore, in unaffiliated dyads, no significant effects were detected in either the dominant or subordinate individuals under any of the administered conditions.
The plots illustrate individual data for each trial. The upper, middle, and lower bounds of the boxes represent the 75th, 50th, and 25th percentiles, respectively, with whiskers extending to the maximum and minimum values. Numbers in the brackets below the boxes denote the number of samples (no. of birds × no. of trials). Asterisks denote statistically significant differences (Wald χ2statistics, *** p < 0.001). The colours correspond to the affiliated dyads shown in Figure 3.
Outputs of GLMMs for submissive vocalization in affiliated dyads.
Post-hoc pairwise comparisons of pre- and post-administration on submissive vocalization by drug conditions.
Post-hoc pairwise comparisons of drug conditions on submissive vocalization in the post-administration phase.
Outputs of GLMMs for submissive vocalizations in unaffiliated dyads.
