Can the diverse styles of song development in songbirds be understood in an
evolutionary context? Are song imitation and song improvisation strategies
that evolved in identifiable ecological circumstances? Differences among C
istothorus wrens suggested that song imitation was used in stable, resident
populations by Marsh Wrens (Cistothorus palustris), but that song improvis
ation evolved in the more nomadic populations of North American Sedge Wrens
(C. platensis). Toward understanding this seemingly unique strategy of imp
rovisation by North American Sedge Wrens, we reexamined song development in
the laboratory and singing behavior and population movements among free-ra
nging males. Nestling Sedge Wrens were collected in North Dakota and during
their first year of life tutored with 10 Sedge Wren song types; throughout
the experiment, males were in adjacent cages and could both hear and see e
ach other. Songs of the laboratory birds were not close imitations of songs
from the training tape or immediate neighbors; rather, songs were either i
mprovised (different from but most likely derived from training songs) or i
nvented (no similarity to other songs in their environment). In nature, mal
es at a Nebraska site also had unique song repertoires, a pattern that is c
onsistent with the improvisational mode of song development. Our field surv
eys also verified that Sedge Wren populations are highly mobile, arriving a
t or departing from breeding sites at seemingly odd times of the summer bre
eding season. These data, together with evidence of song imitation among se
dentary populations of Sedge Wrens in Central and South America, reinforce
the idea that song improvisation among North American Sedge Wrens is a deve
lopmental strategy. Because songs are improvised, each male is unique, but
songs do not vary geographically; hence, it seems likely that males and fem
ales can communicate with one another no matter where they find themselves
in the geographic range of the species.