Chestnut mannikin (Lonchura malacca), an Estrildine finch recently introduced in Puerto Rico, is a widespread Southeast Asian species. In its native habitat it frequently nests in vegetation over water, but in Puerto Rico it usually nests in dense sugar cane. In a marsh ecosystem in Humacao, mannikins nest in clumps of Cyperus spp. growing on top of stumps in a lagoon (depth = 50 cm). Their nests are in sedge clumps that are in shallower water and on higher stumps, and that are taller, wider and denser than clumps without mannikin nests. Nests were usually placed on the south side of the clump, with the entrance facing open water. The predator protection offered by nesting over water may account for nests being placed in relatively open sites. SPECIES THAT EXPAND THEIR BREEDING RANGE, with or without human assistance, into new geographical areas may expand their nesting sites either by nesting in entirely new habitats (i.e., moving from forests to fields), by nesting in different sites within preferred habitats, or both. Presumably constraints from predation pressures, nest site competition, and environmental factors shape nest site choices in the newly-colonized breeding range. In this paper we examine nest site selection in a population of chestnut mannikin (or munias Lonchura malacca) in a marsh in Puerto Rico. Factors affecting nest site selection include social and physiognomic features of the environment (Burger 1985). Social factors include attraction to conspecifics that provide antipredator or information transfer advantages (Fuchs 1977, Nuechterlein 1981, Burger 1981), or avoidance of other species because of predation (Greig-Smith 1982). Physiognomic aspects of nest site selection include factors such as structural stability (Coon et al. 1981), foliage density (Furrer 1980), vegetation height (Cody 1985), orientation relative to the sun (Reynolds & Knapton 1984), and perch site availability (Gochfeld 1979). Some species appear to have very strict nest site requirements or preferences, whereas others tolerate a wide variety of sites. The chestnut mannikin, munia, or black-headed nun is one of several recently introduced Estrildine finches which has become common in Puerto Rico (Raffaele 1983). It naturally ranges from India through southern China and Southeast Asia to the Philippines and Indonesia (Delacour & Mayr 1946, Smythies 1960, Long 1981, Ali & Ripley 1983). The recentness of its establishment in Puerto Rico is reflected by the fact that Long (1981) did not include it in his compendium of feral exotic bird species. The genus Lonchura is a popular cagebird, and eight species within the genus have successfully established feral populations in regions remote from its natural range (Long 1981). This suggests that the genus is either adaptable with respect to habitat selection, or cosmopolitan with respect to habitat requirement. We wished to examine qualitatively whether chestnut mannikin in Puerto Rico used habitats similar in physiognomy to those in its native habitat, and examine quantitatively whether its choice of nest sites differed from the available habitat. Our general approach to nest site selection is to contrast nest sites with sites that are available but not used (Burger & Gochfeld 1985). We interpret these findings in the light of Skutch's (1976) concept that nest sites should be invisible, inaccessible, impregnable. STUDY AREA AND METHODS We studied chestnut mannikin in Frontera Palmas Lagoon of the Humacao Rufuge (operated by the Puerto Rico Departamento de Recursos Naturales) in June 1987 and February 1988. The Humacao Refuge contains 600 ha of brackish coastal marshes created by hurricanes in the late 1970's, and presently maintained by dikes and management. The palmas lagoon once supported a coconut palm plantation (Cocos nuciferos), but as the trees have died since the flooding, the rotting stumps support a growth of grasses (e.g., Paspalum spp.), Cyperus sedges, and ferns. Small stumps often had no vegetation, whereas wide stumps usually had one dominant plant. ' Received I I April 1988, revision accepted 24 August 1988. 364 BIOTROPICA 21(4): 364-368 1989 This content downloaded from 157.55.39.46 on Mon, 13 Jun 2016 05:24:06 UTC All use subject to http://about.jstor.org/terms