A Sexual-Selection Model for the Evolution of Imitative Learning of Song in Polygynous Birds

Publisher Website
Google Scholar
Abstract
Although normal song development in oscines depends on imitative learning, it presumably was innately determined (directly inherited, nonmodifiable) in the ancestral species. Evolution of imitative song from innate song is investigated theoretically, using a genetic model of sexual selection by female choice in a polygynous population. Since song is usually limited to sexually mature males and since females exhibited preferences with respect to various aspects of song, sexual selection may well have been involved. Specifically, it is assumed that imitative learning may give rise to a variant song that some females prefer over innate song and also that female preference is not affected by early song experience. However, the assumption of polygyny restricts the applicability of the results obtained here, since most bird species are monogamous. It is shown that the coevolution of imititative song and female preference from low initial frequencies requires a strong preference. The available data, although not of direct relevance, suggest that such strong preference may sometimes occur. When females that show preference are more common than those that mate indiscriminately, a twofold difference may still be still be necessary for imitative song to be favored. If female preference is assumed to be fixed, the effect of various modes of song transmission can be compared. It is shown that transmission from father to son is critical and must be efficient, unless compensated for by strong female preference. Moreover, initial increase of imitative song is less likely to occur if paternal song is forgotten and replaced by a new song learned from another male adult. These predictions may be difficult to reconcile with the assumption of polygyny, which may imply limited opportunity for interaction between father and son. Finally, the conditions for the evolution of imitative song are compared with those for the evolution of human speech. The latter were obtained previously from a model assuming that speech serves to communicate adaptive cultural traits from parent to child. The conditions are highly analogous in spite of the disparity in the selective forces assumed.

This publication has 0 references indexed in Scilit: