Mixed-species flocks of bird represent an ecological phenomenon resulting from a diversity of complex interactions and evolutionary pressures. In the mixed-flocks of Amazonia, species of antshrike (Thamnomanes sp.) have been shown to provide other flock members valuable information regarding threats from predation through their frequent alarm calls. Several species of flock attendees respond to the alarms of antshrikes with predator-avoidance behaviors, but no study has yet shown a reciprocal relationship between flock members and flock-leading antshrikes. Apparently distinct from mixed-flock systems of other forests, there has been little evidence to support that flock-leading species of antshrike benefit from the alarm signals of other flock members.

Male Dusky-throated Antshrike Thamnomanes ardesiacus. This species provides frequent alarm calls warning conspecifics and eavesdropping species of potential predators.

Spectrograms of alarm and control examples used during playback experiment. A) Alarm: Dusky-throated Antshrike Thamnomanes ardesiacus. B) Alarm: Bluish-slate Antshrike T. schistogynus. C) Alarm: Gray Antwren Myrmotherula menetriesii. C) Alarm: White-flanked Antwren M. axillaris. D) Song (Control): Screaming Piha Lipaugus vociferans

Using alarm signals recorded during in situ exposures of mixed-flock species to trained raptors, we conducted a playback experiment to 16 different mixed flocks to see to what degree antshrikes may be eavesdropping on the alarms of Myrmotherula antwrens. Eight Dusky-throated Antshrike Thamnomanes ardesiacus and eight Cinereous Antshrike T. schistogynus were presented with i) conspecific alarms ii) White-flanked Antwren Myrmotherula axillaris alarms iii) Grey Antwren M. menetriesii alarms and iv) a control. 

The results of our experiment support that the alarms of the two species of antwren illicit a significant response from both species of antshrike. Our model also shows the responses of antshrikes to antwren alarms are indistinguishable from their response to conspecific alarms, yet different from a control (Marginal R2=.36, Conditional R2=.37). This suggests that highly social flock-leading species such as Thamnomanes antshrikes may benefit from the information provided by flock mates regarding predators. Further, these patterns are consistent with eavesdropping interactions observed within mixed-flocks in ecologically similar but geographically distant forests.

​Summary graph showing the probability of response by T. ardesiacus (black circles) and T. schistogynus (grey circles) to alarms of their own species (conspecific), M. menetriesii, M. axillaris, and a control. Our statistical model does not distinguish a significant difference in the probability of an antshrike response between any of the alarm treatments. All of the alarm treatment responses were shown to be significantly different from control treatments. Shown are predicted means and 95% confidence intervals derived from model results.