Falkland Island Sedge Wrens (Cistothorus platensis) imitate rather than improvise large song repertoires

ABSTRACT.-Among songbirds, does reduced fidelity to a breeding site lead to vocal improvisation? Data for Cistothorus wrens suggest it does, because North American Sedge Wrens (C. platensis) have low breeding-site fidelity and improvise their large song repertoires, but sedentary or site-faithful populations of this and other Cistothorus species in the Neotropics and North America all imitate. We attempted to falsify this hypothesis by studying extreme south-- temperate zone populations of Sedge Wrens in the Falkland Islands. We banded and recorded males on Kidney Island and Carcass Island, and then compared song matching among males both within and between islands. Birds on those islands were highly site-faithful from one breeding season to the next. Song repertoires were large, up to 400 in one bird, and songs of birds within an island were more similar to each other than to songs on the other island, showing that these birds do imitate. These results further support the idea that site fidelity promotes imitation of neighbors, and continue to highlight the unique correlation between reduced site-fidelity and song improvisation in the North American Sedge Wren.

RESUMEN.-?Lleva la baja fidelidad al sitio de cria a la improvisacion vocal entre las aves canoras? Datos para el genero Cistothorus sugieren que si, porque poblaciones de C. platensis que presentan baja fidelidad al sitio de cria improvisan sus repertorios de largos cantos, mientras que poblaciones de esta especie y de otros Cistothorus del Neotropico y de America del Norte que son sedentarias o fieles a sus sitios imitan. Intentamos falsear esta hip6tesis estudiando poblaciones de C. platensis en el extremo sur de America del Sur, en las Islas Falkland (Malvinas). Anillamos y registramos machos en las Islas Kidney y Carcass, y luego comparamos la similitud de los cantos entre machos de una misma isla y de islas diferentes. Las aves de estas islas fueron muy fieles a sus sitios entre una estacion reproductiva y la siguiente. Los repertorios de cantos fueron grandes, llegando a 400 cantos en un ave. Los cantos fueron mas similares entre si dentro de una isla que entre islas, mostrando que estas aves sf imitan. Estos resultados tambien apoyan la idea que la fidelidad al sitio de cria promueve la imitation de los vecinos, y continuan destacando la correlacion entre baja fidelidad al sitio e improvisacion de cantos en C. platensis.

Patterns in geographic song variation help reveal the processes by which those variations have evolved. Among Chestnut-sided Warblers (Dendroica pensylvanica), for example, songs used to attract females vary geographically far less than do songs used with territorial male neighbors; intersexual and intrasexual forces have thus produced highly contrasting geographic patterns (Byers 1996a, b). Also, males in resident populations of several species seem to share more songs with immediate neighbors than do males in migratory populations, suggesting that strategies for song development (such as how accurately songs are learned) might be related to patterns of population movements (e.g. Ewert and Kroodsma 1994). A similar relationship seems especially pronounced among Cistothorus wrens, with fidelity to the breeding site promoting song imitation and reduced site fidelity song improvisation (Kroodsma 1999).

Support for this relationship among the Cistothorus wrens comes largely from the uniqueness of the Sedge Wren (C. platensis) in North America (Kroodsma et al. 1999a). Breeding site fidelity there is low, and males improvise their large song repertoires, as if a repertoire of generalized rather than site-specific songs were important to communicate with whatever other male and female wrens were encountered within and between seasons. In contrast, tropical, sedentary populations of that same species imitate (Kroodsma et al. 1999b), as do resident or site-faithful Marsh Wrens (C. palustris) (review in Kroodsma and Verner 1997), Merida Wrens (C. meridae, Kroodsma et al. 2001), and Apolinar's Wrens (C. apolinari, P. Caycedo unpubl. data).

Given the apparent uniqueness of the north temperate zone populations of the Sedge Wren, it is important to try to falsify this hypothesis relating breeding-site fidelity and styles of song development. To do so, we studied the Falkland Island populations, which live in extreme environments of the south temperate zone. We verified that those island populations of Sedge Wrens are highly site-faithful from one breeding season to the next (and most likely permanently resident), and discovered that males there have large repertoires of 300-400 song types that are largely shared with immediate neighbors. Like other resident or highly site-faithful Cistothorus wrens in the tropics and north temperate zone, these Sedge Wrens also imitate, again emphasizing the seemingly unique relationship between low fidelity to a breeding site and song improvisation among North American Sedge Wrens.

Methods.-We studied populations on two islands in the Falkland Islands, Kidney Island and Carcass Island. From 19 to 26 October 1995, we visited Kidney Island, just outside the port of Stanley on East Falkland, color-banding 17 males; we counted 25 territories in our study area of ~20 ha, for a density of 1.25 territories ha^sup -1^ in the lush tussac grass (Poa flabellata) covering the island. From 4 to 13 November 1995, we visited Carcass Island, ~200 km away off the coast of West Falkland, color-banding 12 males there. Suitable habitat on Carcass Island was largely in natural remnants of tussac grass, but because of past grazing the habitat was less dense, which probably contributed to the much larger territories there. Each October or November from 1996 to 2000, R.W.W. returned to Kidney Island to census the birds there; dates were 21-28 November 1996, 15-19 October 1997, 15-19 October 1998, 31 October to 6 November 1999, and 13-20 October 2000. Since 1959, R.WW. has visited Kidney Island on numerous occasions, confirming that the wrens were present on each visit from September through April; no visits were made during the midwinter months of May through August, but we believe the wrens are resident on their territories year round.