Patterns of geographic variation in the songs of a neotropical lowland bird
I studied song variation throughout the range of a Neotropical lowland bird species, the Buff-throated Woodcreeper (Xiphorhynchus guttatus ; Aves, Dendrocolaptidae), to examine support for the use of vocal characters in defining species-limits in suboscines. This species is common across most of the lowlands of Amazonia and the Atlantic forest, its populations are separated by geographic barriers well-known in regional biogeography, and vocalizations likely used in reproductive isolation presumably are innate. I present a detailed overview of woodcreeper behavior in general and of the systematics, distribution, identification, voice, behavior, and conservation of the Xiphorhynchus guttatus superspecies in particular. Analyses of my field data revealed that this species' song varies more than songs of most suboscines and that song variation within individuals is best quantified using the relative length of three song parts: the introduction, middle, and conclusion. Within each song part, the structure of component notes and the rate at which these notes are repeated vary little. I could best distinguish subspecies groups by differences in note structure and repetition rate. Using multivariate analysis of spectrogram parameters and the location of breakpoints separating song parts, I could place a song within a subspecies group with a high degree of certainty. Thus, song types correspond well with both subspecies groups and known centers of endemism in the region. Not only do subspecies groups have strikingly different songs, but songs differ among and within subspecies. Even different individuals within populations can be diagnosed on the basis of their songs. Early researchers reported that suboscine song does not vary geographically, but more recent quantitative studies have revealed geographic variation in the songs of at least four suboscine species. Even so, vocal differences among suboscine populations are more subtle than those in oscines, so rigorous sampling of large numbers of songs and of individuals is necessary to detect these differences. Unlike a recent study of antshrike vocalizations, geographic variation in voice is not correlated with geographic variation in genetics in the Xiphorhynchus guttatus complex.