The field observations in this study revealed that wild cordon-bleus exhibit tap-dance-like displays during courtship communication. As predicted, wild cordon-bleus performed rapid dance displays in the same manner as captive birds (Table 1; Video 1 in Online Resource 1). Another similarity observed between wild and captive individuals was that they both performed dance displays on branches but never on the ground (Ota et al. 2015; Video 1 in Online Resource 1).
As my data is descriptive it is too early to conclude the functions and efficacies of multimodal signals from this study. However, my field observations strongly imply that cordon-bleu dance displays can produce multimodal and multicomponent signals including non-vocal sounds and vibrations under wild conditions. I observed that some signal receivers perched on the same substrate with signal senders, and in those cases, they seemed quite possible to perceive the non-vocal sounds and vibrations. I also found that the dance displays resulted in the shaking of several branches of the tree in which the bird performed (Video 1 in Online Resource 1). Since the movements were clearly different from those caused by the wind, shaking a branch by dancing might also function as a visual signal in addition to the dance movements and items held during the display (Fig. S2 in Online Resource 2). This possibility has been overlooked in studies under laboratory conditions.
Presumably, the multimodal signals produced during cordon-bleu courtship displays can have several functions. The multimodal signals comprised of visual elements with sounds and vibrations which could function as backup signals to ensure appropriate signal transmission under fluctuating environmental conditions. It is also possible that multimodal signals may encode multiple messages directed toward multiple signal receivers (Johnstone 1996). Since the species difference in the dance performance was not detected in the current study (Fig. S1; Table S1 in Online Resource 2), the dance display might function as conveying information such as physical abilities and motivations rather than the species. More detailed analyses are required to elucidate the information encoded in the multimodal courtship displays of cordon-bleus.
An important next step following this study is to quantitatively test how multimodal and multicomponent courtship signals affect the responses of signal receivers in cordon-bleus as well as their future breeding success. My previous (Ota et al. 2015) and current studies emphasize that both field and laboratory approaches are useful for understanding multimodal signaling behavior in cordon-bleus. While laboratory experiments would enable us to conduct detailed behavioral experiments under well-controlled conditions, the signal efficacy (e.g., the effective range of sound/vibrational amplitude) and actual mating processes via courtship display should be examined in the field.
Another question that will deserve further attention is why female courtship displays were hardly observed in the wild, considering that captive female cordon-bleus exhibit courtship displays like males (Goodwin 1982; Geberzahn and Gahr 2011; Ota et al. 2015). A likely explanation is that my observations occurred during the middle or end of the cordon-bleu breeding season. Potential sex differences in the timings of courtship displays due to the sex differences in reproductive constraints and extra-pair mating strategies (Tobias et al. 2012) may, therefore, have affected the number of female displays observed in this study.
In conclusion, I have shown for the first time in wild species that cordon-bleus perform rapid “tap-dance”-like display. Though not conclusive due to the lack of quantitative evaluation of signal efficacy, the results point to the potential of multimodal signaling behavior overlooked in previous songbird communication studies.