We determined habitat preferences for males and females of nine species of dichromatic warblers in the Sian Ka'an Reserve, Quintana Roo, Mexico. We found significant habitat segregation in the American Redstart (Setophaga ruticilla), Common Yellowthroat (Geothlypis trichas), Magnolia Warbler (Dendroica magnolia), and Northern Parula (Parula americana), and we confirmed previously reported habitat segregation in the Hooded Warbler (Wilsonia citrina). Possible differences in habitat distribution were found in the Black-and-white Warbler (Mniotilta varia), Prothonotary Warbler (Protonotaria citrea), and Black-throated Green Warbler (Dendroica virens). Only migratory Yellow Warblers (Dendroica petechia) showed no sign of sexual segregation. In all five cases where there is a significant difference between successional stages, males occupied the more mature habitat. The overall frequency and spatial pattern of recaptures were similar between sexes. Received 29 March 1989, accepted 26 January 1990.

Estacion Biologica Donana, Consejo Superior de Investigaciones, Avenue Maria Luisa s/n, Seville, 41113 Spain, and 2National Zoological Park, Washington, D.C. 20008 USA HABITAT segregation between the sexes was first described for American Kestrels (Falco sparverius, Mills 1976) and Eastern Great Reed Warblers (Acrocephalus orientalis, Nisbet and Medway 1972). More recently the first species of Neotropical migrant land bird was found to display intersexual segregation. Male and fe- male Hooded Warblers (Wilsonia citrina) defend territories in different habitats (Lynch et al. 1985, Morton et al. 1987). Males were common in ma- ture forest, and females occurred in a variety of scrub habitats. However, the overall frequency of habitat segregation among Neotropical mi- grants has not been established. Sexual habitat segregation has important implications for un- derstanding the life history and demography of particular species and for assessing the pos- sible effects of habitat alteration. We report the results of a survey of the most common dichro- matic migratory wood warblers found in the Northeastern Yucatan Peninsula. The results summarize two different surveys; each author conducted one survey. We used two techniques (mist-netting and visual-auditory surveys), which produced complementary data sets. Where the data can be compared, the results are generally in agreement. The data sets together provide information on all of the common di- 3 To whom reprint requests should be sent. 539 chromatic warblers in the Sian Ka'an Reserve, Quintana Roo, Mexico. METHODS Study area.--The fieldwork was conducted in the Sian Ka'an Biosphere Reserve and surrounding ejidos (rural land cooperatives) in the northeast coastal re- gion of the Yucatan Peninsula. The reserve receives 1,200 mm/yr of rainfall, primarily during a rainy sea- son from June to November. The area supports a high abundance and diversity of Nearctic migrants during the nonbreeding season (Waide 1980, Lynch 1989). We conducted surveys in coastal dune scrub, man- grove scrub, seasonally flooded scrub, low forest, me- dium-height subdeciduous forest, medium-height subperennial forest, pasture, and postclearing succes- sional scrub. These habitats represent the major vege- tation types of the eastern Yucatan Peninsula (Miran- da 1958, Olmsted et al. 1983). Within each habitat where we mist-netted, the veg- etative cover was categorized and estimated for an imaginary circle centered at each of 24 net sites. The categories include (a) herb, grass, or fern; (b) scrub (1-4 m); and (c) trees (4 m+). Cover was estimated to one of five classes, each representing 20% (e.g. cover class 1 = 1-19%, class 2 = 20-39%). A quantitative description of the habitats is presented (Table 1). Me- dium-height subperennial forest was not sampled in an equivalent fashion, but it has greater stature (mean canopy height = 13-15 m), fewer deciduous canopy trees, and a moister, shadier understory than subde- ciduous forest. To analyze the bird survey data, we pooled the TABLE 1. plot. LOPEZ ORNAT AND GREENBERG [Auk, Vol. 107 Habitat structure measurements of study plots. Numbers are means (+SD) based on 24 points per Plot size Canopy ht. Vegetative cover Mist-net (ha) (m) Trees Shrubs Herbs hours Second growth 1-2 yr 1.5 1.1 + 0.3 0 2.6 + 0.9 2.6 + 1.6 1,105 3-4 yr 1.5 1.6 + 0.6 0.4 + 0.6 3.1 + 1.0 2.0 + 1.0 1,230 7-8 yr 3.0 3.7 + 1.1 2.6 + 1.0 1.5 + 0.7 1.2 + 1.0 1,250 Mangrove scrub 1.5 2.1 + 1.2 1.6 + 0.8 4.0 + 0.8 0.4 + 0.6 576 "Sabana" 1.5 2.1 + 1.5 1.2 + 1.6 2.0 + 1.3 2.7 + 1.0 1,656 Low forest 1.5 5.5 + 1.6 3.6 + 0.5 1.6 + 0.9 0.6 + 0.6 1,728 Medium forest 2.5 11.0 + 1.6 4.1 + 0.6 0.9 + 0.3 0.2 + 0.1 1,404 habitats in two ways. The mist-netting data (Table 2) are compared for open habitats (second growth 3-4 yr of age or younger and sabana, a marshlike habitat), scrub (7-8-yr-old second growth and mangrove scrub), low forest (seasonally flooded low forest), and me- dium forest. To analyze the broad-scale visual survey, we contrasted scrub habitats (from pasture to 7-8-yr- old second growth), subdeciduous, and subperennial forests. Mist-netting.--In each of the vegetation types (ex- cept subperennial forest), we established grids that consisted of 3 parallel lines with 8 mist nets (2 x 12 m, 30 mm mesh) on each line. Nets were spaced 15- 25 m apart along the lines, and 50-70 m between the lines. The total areas netted varied between habitats from 1.5 to 3.0 ha. Nets were opened at dawn (ca. 0600) and closed 6 h later. Nets were opened for at least three consecutive days in each habitat during the autumn, winter, and spring of 1985/1986 and of 1986/1987 for a total of 9,372 net hours. All of the habitats were sampled each year in midwinter. Ad- ditional netting was done in September-October and April in some habitats. Because only ground-level mist nets were used, the relative numbers of individ- uals captured do not reflect differences in species' abundance between different habitats (Lynch 1989). Visual surveys.--The Sian Ka'an Reserve and nearby areas were surveyed during the winter of 1987/1988 and of 1988/1989. The surveys were designed only to determine sex ratios, not to estimate relative abun- dance. A particular habitat patch or trail was selected for each survey, and the sex of each dichromatic war- bler was recorded. Effort was made to select sites that were away from edge situations and could be un- ambiguously assigned to one of three major habitats: scrub, subdeciduous, or subperennial forests. Each site was visited only once each winter. Criteria used to distinguish sexes.--Sex was deter- mined for all species by plumage characteristics. The criteria were developed from the USFWS banding manual (1984) and supplements from Pyle et al. (1987). There is a likelihood of some error for subtle dis- tinctions (such as between first-year male and female American Redstarts, and male and female Magnolia Warblers. For participants in the visual surveys, in- terobserver consistency was tested both with mist- netted and observed birds, and it was found to be high for both. For Magnolia Warblers, observer clas- siftcation from plumage characters was checked in- dependently against wing chord (using a cutoff of 58 mm or greater for males), and they were consistent 97% of the time. Resident Yellow Warbler females were distinguished from migratory females by their grayer upperparts and duller yellow underparts. TABLE 2. Ratio of males to females in mist-netted samples of warblers. Data were pooled to conduct Chi- square contingency test for heterogeneity in sex ratio. a Medium Open Scrub Low forest forest P Black-and-white Warbler 8:7 0:3 5:15 1:1 NS Prothonotary Warbler 15:6 1:2 4:2 1:2 NS Yellow Warbler 4:6 6:7 0:0 0:0 NS American Redstart 2:19 6:3 4:10 4:3 <0.025 Hooded Warbler 3:11 1:20 9:10 8:1 < 0.01 Common Yellowthroat 25:27 b 22:12 8:0 0:0 <0.025 ' Data were pooled to make the following comparisons: Black-and-white Warbler, American Redstart, and Hooded Warbler open+scrub habitat vs. forest; Prothonotary Warbler and Common Yellowthroat open habitat vs. scrub+ forest; and Yellow Warbler open habitat vs. scrub.  Significant heterogeneity exists between the recently cleared fields (11:24) and the sabana or marsh-like habitat (14:3). T^BLE 3. The percentage of males that comprises sightings of dimorphic warblers in scrub and forest habitats (sample size in parentheses). Species Scrub Subdeciduous Subperennial Black-and-white Warbler 58 (12) 55 (20) 48 (62) Northern Parula a 37 (83) 59 (27) 67 (48) Black-throated Green Warbler 47 (30) 50 (40) 63 (72) Magnolia Warbler a 36 (99) 55 (64) 60 (101) American Redstart  35 (113) 74 (90) 80 (139) (adult males) a 16 48 47 Hooded Warbler a 11 (72) 67 (83) 89 (97) Significant (P < 0.05) heterogeneity based on Chi-square contingency test pooled forest vs. scrub data. Statistical tests.--Testing for differences in the pro- portion of males in different habitats was by a Chi- square contingency test. For the visual data, tests com- pared forest and scrub habitats. Because of variable and sometimes small sample sizes, the mist-net data were also grouped. In all cases, adjacent habitats along the successional gradient were pooled. However, be- cause of the differences in distribution across the successional gradient, habitats were grouped in dif- ferent ways. RESULTS Prothonotary WarbIer.--The Prothonotary Warbler (Protonotaria citrea) was a common tran- sient (mainly in autumn). The few mist-net cap- tures in scrub and forest habitat showed no skew in sex ratio. However, most birds captured in open habitats were males. The difference in sex ratio between open and other habitats was not significant. Black-and-white WarbIer.--Black-and-white Warblers (MniotiIta varia) were found primarily in forests. The mist-nets yielded a high pro- portion of females in the low forest, but sample sizes for other habitats were generally small and the difference between pooled forest and open- scrub samples was not significant. Visual sur- veys indicated an even sex ratio for scrub and forest habitats. Northern ParuIa.--Northern Parulas (ParuIa americana) were found in most terrestrial hab- itats in Sian Ka'an, but were most common in older secondary-scrub with a substantial num- ber in the canopy of the subperennial forest. Few were mist-netted. Based on the visual sur- veys, Northern Parulas had a significantly fe- male-biased sex ratio in scrub and male-biased sex ratio in the taller forests (Table 3). Yellow Warbler.--During winter most of the Yellow Warblers (Dendroica petechia) in the Sian Ka'an area were migrants. Yellow Warblers were common, but had a very restricted habitat dis- tribution. They were most common in the nat- ural scrub vegetation (mangrove and coastal scrub), and uncommon in postclearing succes- sion. In the open and scrub habitats considered here, Yellow Warblers displayed an overall even sex ratio with no indication of interhabitat vari- ation. When additional data from coastal scrub were included, the overall sex ratio was also equal (30:32). Magnolia Warbler.--Magnolia Warblers (D. magnolia) were common in every habitat except recently cleared fields, although they were more common in forest than scrub. They were cate- gorized only by sex in the visual surveys (Table 3). There was a small but significant bias to- wards females in the scrub and towards males in the forest habitats. Black-throated Green WarbIer.--Black-throated Green Warblers (D. virens) were most common in forest, and were uncommon in secondary scrub with high densities of Acacia and other leguminous trees. Based on the visual surveys (Table 3), we found a nonsignificant trend to- wards an increased proportion of males in taller forest habitat. Hooded Warbler.--Hooded Warblers were in every habitat except mangrove scrub. Based on mist-netting (Table 2), we found 8 times as many males as females in medium forest, but only 13% as many males as females in open and scrub habitats (P < 0.001). The structurally interme- diate low forest had an intermediate sex ratio. Similar results were obtained in the visual sur- vey (Table 3) with the proportion of males rang- ing from 11% in scrub to 89% in tall forest hab- itat (overall Chi square, P < 0.001). American Redstart.--American Redstarts (Se- tophaga ruticiIIa) were numerous in all habitats with at least some woody vegetation. Mist-net- ting data (Table 2) show that almost all birds in the open habitat were females, whereas males made up 29-67% of the small samples captured in scrub to subdeciduous forest. There was no consistent trend within scrub and forest habi- tats. Perhaps, because of a larger sample size (particularly for forest habitats), the difference in sex ratio was clearer with the visual survey data. Males made up 74-80% of the forest ob- servations, and only 35% of the scrub obser- vations (P < 0.001). The proportion of adult males, where no problem in accurate determi- nation by plumage should exist, also increased dramatically between scrub and forest habitat. Common Yellowthroat.--The Common Yellow- throat (Geothlypis trichas) was the most common warbler in open habitats, but it was absent from all forest types except seasonally flooded low forest. Mist-net captures were significantly skewed towards females in the open pastures and fields (11:24), and male-biased for the older second-growth scrub habitats (30:12). In the natural open habitat, "sabana," males were more common (14:3). Recaptures.--Of the individual warblers from four species (Table 4), 21% (n = 223) were re- captured after at least 30 days from their first capture. Although there was considerable in- terspecific variation, there was no significant difference in site tenacity between the sexes. Females were recaptured at a slightly and non- significantly higher rate (males, 16%; females, 24%). The mean distance between captures and recaptures (activity radius, Table 4) was similar for both sexes in all species, with the possible exception of the American Redstart. DISCUSSION Some degree of habitat segregation by sex is the rule for dimorphic warblers that migrate through or winter in the northeastern Yucatan Peninsula. In all five species where there is a significant pattern, males occupy the more ma- ture habitat along the successional gradient. In addition, the Black-throated Green Warbler showed a nonsignificant trend in the same di- rection. The one exception was in migratory populations of the Prothonotary Warbler. This exception is noteworthy because Prothonotary Warblers have been found commonly in male- female pairs during winter in Panama (Morton 1980, pers. obs.). However, although the sex ratio is female-biased in open habitats, it is not significantly different from the samples from other habitats. Furthermore, sex ratio was based on a small sample of transient individuals. In the Hooded Warbler, Morton (1990) has argued that the proximate mechanism for sex- ual habitat segregation is an innate difference in the response to simple cues. Support for this hypothesis comes from two experiments. In one, a removal experiment, females failed to occupy mature forest when the resident males were removed (Morton et al. 1987). In a laboratory experiment, naive hand-reared males and fe- males showed distinct preferences for plant height and habitat structure (Morton 1990). In- nate differences in habitat preference should lead to relatively invariant differences in hab- itat occupied. Habitat segregation in this species is, in fact, nearly complete, particularly when fine-scaled habitat measures are analyzed with multivariate techniques (Lynch et al. 1985). However, few of the species show habitat seg- regation that is as clearly defined as in the Hooded Warbler. In the case of Magnolia War- biers and Northern Parulas, the magnitude of the segregation was small, although statistically significant. In these situations, the role of dom- inance interactions between the sexes may be more important. Experiments similar to those conducted on Hooded Warblers would make an interesting comparison. We found no support for the idea that females behave as "floaters" in the nonforest habitat. The overall frequency and pattern of recaptures was similar between the sexes. Perhaps a more facultative basis for habitat segregation explains the apparent differences between studies for the same species. Holmes et al. (1989) reported no consistent habitat seg- regation between the sexes of American Red- start in four Jamaica study areas. Another ex- planation is that the same degree of habitat segregation occurs in Jamaica and Mexico, but the study in Jamaica focused on a more narrow and structurally intermediate range of habitats. Two possibilities have been presented for the evolution of habitat segregation in Hooded Warblers (Lynch et al. 1985). One explanation holds that differences in sex roles during the breeding season may lead to asymmetrical dom- inance relationships in the winter. Males, the larger and more aggressive sex, are able to ex- clude females from the better habitat. An alter- native hypothesis is that the two habitats rep- resent different adaptive peaks. Males and females are best adapted for feeding and avoid- TABLE 4. Activity radius (mean [___SD] intercapture distance) and recaptures for four species of warblers. The sex ratio of recaptures is not significantly different from the sex ratio of captures in any species (Chi-square contingency test); sample sizes are in parentheses. Species Sex Activity radius (m) Captures Recaptures Black-and-white Warbler M 87 ___ 44 (33) 14 3 F 65 ___ 32 (39) 26 9 American Redstart M 37 + 21 (16) 16 3 F 93 ___ 35 (47) 35 10 Hooded Warbler M 100 ___ 81 (33) 21 6 F 82 ___ 40 (42) 42 9 Common Yellowthroat M 80 + 37 (43) 41 3 F 57 ___ 38 (33) 36 4 ing predators in their preferred habitats. We cannot help distinguish these two hypotheses. However, these observations and others (for Black-throated Blue Warblers, Dendroica caeru- lescens; Wunderlie pers. comm.) suggest that habitat preference is not an arbitrary feature of a particular sex for a particular species. Rather, there seems to be a repeated pattern: male war- biers are associated with forest and females with more open habitats. ACKNOWLEDGMENTS We thank James F. Lynch for comments on earlier drafts of the manuscript. We also thank Thomas W. Sherry and two anonymous referees for their careful reviews, and Mauro Berlanga, Oscar Barrera, James Lynch, Daniel Niven, Dana Bradley, David and Jill Heath, and Rosa Maria Vidal for assistance in field- work. We thank Barbara McKinnon de Montes and Amigos de Sian Ka'an, A.C., for support during field- work. We thank SEDUE for use of the facilities at the Sian Ka'an Biosphere Reserve. Research was funded by grants from the Direcci6n General de Investi- gaci6n Cientifica y Tecnica, Ministerio de'Educaci6n of Spain, and the National Ecology Research Center of the U.S. Fish and Wildlife Service. LITERATURE CITED HOLMES, R. T., T. W. SHERRY, & L. REITSMA. 1989. Population structure, territoriality and overwin- ter survival of two migrant warbler species in Jamaica. Condor 91: 545-561. LYNCH, J. F. 1989. Distribution of overwintering Nearctic migrants in the Yucatan Peninsula: I. General patterns of occurrence. Condor 91: 515- 545. ß , E. S. MORTON, & M. E. VAN DER VOORT. 1985. Habitat segregation between the sexes of over- wintering Hooded Warblers (Wilsonia citrina). Auk 102: 714-721. MILLS, G. S. 1976. American Kestrel sex ratios and habitat separation. Auk 93: 740-748. MIRANDA, F. 1958. Estudios acerca de la vegetacion. Pp. 215-271 in Los Recursos Naturales del Sureste y su aprovechamiento, parte II: Estudios partic- ulates. Mexico City, Inst. Mexicano Recursos Nat. Renovables (IMERNAR). MORTON, E.S. 1980. Adaptation to seasonal changes by migrant landbirds in the Panama Canal Zone. Pp. 437-457 in Migrant birds in the Neotropics (A. Keast and E.S. Morton, Eds.). Washington, D.C., Smithsonian Inst. Press. 1990. The evolution of habitat segregation by sex in the Hooded Warbler: experiments on proximate causation. Am. Nat. 135: 319-333. , J. F. LYNCH, K. YOUNG, & P. MELHOP. 1987. Do male Hooded Warblers form nonbreeding ter- ritories in tropical forest? Auk 104: 133-135. NISBET, I. C. T., & L. MEDWAY. 1972. Dispersion, pop- ulation ecology, and migration of Eastern Great Reed Warblers A crocephalus orientalis wintering in Malaysia. Ibis 114: 451-494. OLMSTED, I. C., A. LOPEZ ORNAT, & g. DUP,q. 1983. Vegetacion. Pp. 41-64 in Sian Ka'an, estudios pre- liminares. Puerto Morelos, Q. Roo, Mexico, CIQRO-SEDUE. PYLE, P. D., S. N. G. HOWELL, R. P. YUNICK, & D. F. DESANTE. 1987. Identification guide to North American passerines. Bolinas, California, Slate Creek Press. U.S. FISH AND WILDLIFE SERVICE. 1984. Bird banding manual. Washington, D.C. WAIDE, g. g. 1980. Resource partitioning between migrant and resident birds: the use of irregular resources. Pp. 331-352 in Migrant birds in the Neotropics (A. Keast and E. S. Morton, Eds.). Washington, D.C., Smithsonian Inst. Press.