Table 1.

Responses to simulated territorial intrusions with noise-adjusted songs in control environments results in atypical responses for several behaviours.

	Number of songs	Number of calls	Number of attacks	Number of wing flicks	Min approach distance	Ln time to min approach
Song type (adjusted vs. unadjusted)	−1.16 ± 0.93 (0.214)	8.27 ± 3.45 (0.017)	−1.44 ± 0.45 (0.001)	−0.94 ± 0.42 (0.025)	2.76 ± 2.8 (0.351)	−0.04 ± 0.2 (0.833)
CORT	−0.41 ± 1.22 (0.739)	16.16 ± 7.74 (0.037)	0.66 ± 0.61 (0.277)	0.05 ± 0.73 (0.943)	0.69 ± 2.98 (0.819)	−1.15 ± 0.37 (0.009)
Song type x CORT	3.86 ± 1.47 (0.008)	−15.58 ± 7.81 (0.046)	0.93 ± 0.61 (0.126)	1.24 ± 0.69 (0.071)	−3.79 ± 3.72 (0.335)	0.12 ± 0.26 (0.669)

Results are given as coefficient ±s.e. (P), with P ≤ 0.05 in bold. Categorical variables are shown with 1 vs. 0 (i.e. adjusted = 1 and unadjusted = 0), such that a positive β indicates an increase with category 1.We fit generalised linear mixed models with song played, adrenocortical responsiveness and two-way interactions as fixed effects, with male ID as a random effect, using only data from control sites. Playbacks (n = 19) to individuals (n = 11) were divided by song type: unadjusted songs at control sites (n = 10) and adjusted songs at control sites (n = 9).