Experiment 1 – Diet manipulation
The average amount of food that birds consumed did not vary by diet treatment (χ2 = 1.75, P = 0.416). Body mass decreased over the course of the experiment (day: χ2 = 49.49, P < 0.0001), but did not vary by treatment (Figure S1; χ2 = 0.36, P = 0.836) or the interaction between treatment and time (χ2 = 0.38,P = 0.829). We observed a non-significant trend where fat score also decreased over the course of the experiment (Figure S2; day: χ2 = 3.61, P = 0.058). Fat score did not vary by treatment (χ2 = 0.47, P = 0.792) or the interaction between treatment and time (χ2 = 1.29P = 0.526). Corticosterone concentrations significantly increased after 30 min, being higher in stress-induced samples (Figure S3; sample type: χ2 = 150.99, P < 0.0001), but were not affected by diet treatment (P = 0.124) or time sampled (P = 0.060). Complement activity was also unaffected by treatment, day, and the interaction between treatment and day (Figure S4; all P
Bacterial richness and bacterial diversity of the gut microbiota did not vary across diet treatments, over time, or in response to the interaction between treatment and time (Table S1; Figure S5; allP adj ). However, when considering the interactive effects of each physiological metric (Table S2; complement, baseline corticosterone, stress-induced corticosterone) with diet treatment and time (day 0, day 8), we found that complement activity and changes in complement activity over time significantly affected microbial richness and Shannon diversity. Higher complement activity levels in blood plasma were associated with greater richness (Figure 2; complement:P adj = 0.0280) and higher Shannon diversity (Figure 2; complement: P adj = 0.0492). This relationship was strongest prior to treatment (day 0), and appeared to be disrupted over the course of the experiment (complement*day; Richness: P adj = 0.0023, Shannon diversity:P adj = 0.0118). Shannon diversity was also influenced by a significant three way interaction between treatment, time, and stress-induced corticosterone concentrations (Figure 3; treatment*day*SIcort: P adj = 0.0276). Higher stress-induced plasma corticosterone concentrations were generally associated with higher Shannon diversity, however this positive association was not apparent in birds on a protein diet prior to treatment (day 0) or birds fed a high fat diet for 8 days. We did not find a significant effect of diet, time, or the interaction between diet and time on any beta diversity metrics (Table S3, Figure S6; allP adj d that birds on the high protein diet had a significant reduction in the abundance of the phylum Campylobacterota (Figure 4; χ2 = 6.75, P adj = 0.0469) when comparing samples collected on day 0 (pre-treatment) to samples collected on day 8 (post-treatment). Conversely, there were no significant changes in relative abundance at the phylum level in the high fat (all P adj ≥ 0.870) or equal ratio (allP adj ≥ 0.603) diet treatments when comparing pre- and post- treatment samples (Table 1). We did not detect any significant changes in any of the three diet treatment groups at the genus level when comparing pre- and post- treatment samples (Table 1, Figure S7; allP adj ≥ 0.143).