Community algorithms reveal song type themes in Adelaide’s warbler song type sequence networks

Abstract

Some New World Warblers (Family: Parulidae) sing with immediate variety during the dawn chorus and eventual variety during daytime song. We used network analysis, including a community clustering algorithm, to further characterise song type sequences during the dawn chorus and daytime song in male Adelaide’s warblers (Setophaga adelaidae). Networks had longer path lengths than expected by chance, indicating that song type transitions were constrained. Community analysis revealed the presence of ‘themes’, or groups of song types that individuals deliver in close sequential proximity. To our knowledge, this is the first report of song type themes in Parulidae. Males did not cycle through their repertoires efficiently, as would be expected if large repertoires were attractive to females. Themes might emerge from the learning process or from interactions with neighbours. Themes may function to improve vocal performance or organise song types with similar functions. Relative to dawn chorus networks, daytime song networks had longer paths and stronger community structure. We hypothesise that song type networks are more structured during daytime song because song delivery is optimised for vocal warm-up during the dawn chorus, males frequently switch among intended receivers during the dawn chorus, or females prefer extended themes during daytime song.