The Condor9 1:515-544 0 The CooperOrnithological Society 1989

DISTRIBUTION OF OVERWINTERING NEARCTIC MIGRANTS IN THE YUCATAN PENINSULA, I: GENERAL PATTERNS OF OCCURRENCE ’

JAMES F. LYNCH SmithsonianEnvironmental Research Center, P.O. Box 28, Edgewater,MD 21037

Abstract. Point countsand mist-net surveyswere employed to studythe winter distri- bution of nearcticmigratory landbirds in Mexico’s YucatanPeninsula. Overwintering mi- grantscomprised 42 regularlyoccurring species, and accountedfor 30-58% (mean = 41%) of the individual birds encounteredin surveysof a wide rangeof natural and disturbed habitats.Some migratory species were encountered most frequently in pastures,agricultural fields,or brushysecond growth, but the main habitatfor many migrantswas tropical forest. Migratoryand residentspecies showed similar degreesof specializationwith respectto the successionalmaturity of their habitats,but residentswere more likely to specializeon particulartypes of matureforest. After mist-netcapture data werestandardized by rarefac- tion, therewere no statisticallysignificant differences in the ratio of migrantsto permanent residentsin habitatsthat ranged from pasture,through brushy old fields,to maturesemiever- greenforest. For any given number of captures,species richness of both migrantsand residentswas substantiallyhigher in mist-net samplesfrom pasturesand brushyold fields than in thosefrom maturesemievergreen forest. With few exceptions,nearctic migratory species that breed in mature temperate-zone forestoccurred both in forestand in brushysecond growth during winter, althoughsome specieswere substantiallyless frequentin the latter habitat. In contrast,overwintering migrantsthat breedin edgeor field habitatstended to avoid tropicalforest in the Yucatan. Key words: Parulinae; migratory birds;distribution; habitat use; Yucatan Peninsula.

INTRODUCTION reflect species-widereductions that are causally In recent years, biologists and conservationists related to the destruction of overwintering hab- have expressedincreasing concern over the sta- itat in the neotropical region (Briggsand Criswell tus of nearctic migratory birds, speciesthat breed 1979, Morse 1980, Terborgh 1980, Rappole et in arctic or temperate North America, but spend al. 1983, Wilcove and Terborgh 1984, Powell the nonbreeding season south of the Tropic of and Rappole 1986). Certainly, there can be no Cancer (e.g., Keast and Morton 1980, Rappole doubt as to the rapid pace of deforestation in et al. 1983, Wilcove and Whitcomb 1983, Powell Middle America and the Caribbean islands and Rappole 1986). Since the 1950s breeding (Myers 1980), areas where the majority of mi- populations of some nearctic migrants that breed gratory speciesand individuals overwinter (Keast in forest have undergone alarming local declines, and Morton 1980). particularly in regions where remaining wood- While a link between tropical deforestationand lands are highly fragmented (e.g., Lynch and declining populations of migratory landbirds in Whitcomb 1978, Morse 1980). Studiesin eastern North America is both plausible and a cause for North America have shown that the area, iso- great concern, there is surprisingly little direct lation, physiognomy, and floristic composition evidence for such a connection. Indeed, the ex- of remnant forest patches all influence the oc- istence of a pattern of global decline in migrant currence and density of breeding birds, partic- populations, as opposed to purely local pertur- ularly nearctic migrants (e.g., Robbins 1980, bations, has yet to be unequivocally demonstrat- Whitcomb et al. 198 1, Lynch and Whigham 1984, ed. The only published continent-wide data set Askins et al. 1987). For these long-distance mi- on population trends, the Breeding Bird Survey grants, it has also been suggestedthat population (BBS),indicates that breedingpopulations of most declines in local portions of the breeding grounds forest-associatedmigratory specieswere stable, or actually increased, between 1966 and 1979, the only period for which data have been pub- I Received12 July 1988. Final acceptance16 Feb- lished (Robbins et al. 1986). This latter finding, ruary 1989. which has struck some biologists as counter-in-

[5151 516 JAMES F. LYNCH tuitive, seems to harken back to an earlier, and Maya Mountains of Belize, essentially the entire largely discredited, view (e.g., Slud 1960, Willis peninsula is a low-lying shelf of Tertiary to 1966, Leek 1972) that moderate levels of tropical Quaternary marine limestone. The present-day forest disturbance might actually benefit the con- Yucatan gradually emerged from the sea floor, siderable number of nearctic migrant speciesthat beginning at the southern end during the Oli- utilize edge habitats and successionalvegetation gocene-Miocene epoch, and continuing north- on their tropical wintering grounds. ward into Pleistocene and Recent times (Flores Analysis of more recent BBS data indicates 1952, Hatt et al. 1953, West 1964). that many ofthe increasesobserved between 1966 The northern half of the peninsula lacks rivers and 1979 have levelled off, or even have been and streams, and surface water is restricted to reversed (Robbins et al., in press),but most mi- scatteredshallow seasonalpools and a few lakes. grant populations still appear to be as high or Rainfall, which is stronglyconcentrated in a May- higher than they were in the late 1960s according November wet season,increases markedly from to BBS data. In large part, current uncertainties north to south, and, in the northern half of the as to the specific threats posed to nearctic mi- region, from west to east (Garcia 1965). At Pro- grants by tropical deforestation reflect our ig- greso,on the northwesterncoast of Yucatan state, norance of the winter ecology of even the com- the annual rainfall is 450 mm, and the native moner migrant species(Keast and Morton 1980, vegetation is low thorn scrub. Merida, only 30 Rappole et al. 1983). km inland, receives about twice as much precip- Within the northern neotropical region, the itation, and the native vegetation there is low 240,000-km* Yucatan Peninsula stands out as deciduous forest, termed “selva baja caducifol- an important overwintering center for nearctic ia” (Miranda 1958). Between Merida and the migratory landbirds. Previous studies (e.g., Caribbean coast,some 300 km to the east,annual Paynter 1955, Tramer 1974, Waide 1980, Waide precipitation increasesto about 1,200 mm, and et al. 1980) had shown that overwintering mi- low deciduousforest gives way to moderately tall grants are both diverse and abundant in the semideciduous forest (“selva mediana subcadu- northern Yucatan Peninsula, but a comprehen- cifolia”), then to moderately tall semievergreen sive, quantitative survey of migrant numbers and forest (“selva mediana subperennifolia”) in habitat associationshas not previously been at- northern and central Quintana Roo. According tempted. In 1982 I undertook such a survey in to the Holdridge systemof life zone classification order to addressthe following questions:(1) What (Holdridge et al. 197 l), this west-east gradient are the relative densities of individual migratory corresponds to a transition from thorn wood- species,and of migrants as a group, in the most land, through arid tropical forest, to dry tropical widespread categories of natural and disturbed forest (Thien et al. 1982). vegetation? (2) Are migrants as a group more In southernmost Quintana Roo and Cam- tolerant of human-related habitat disturbance peche, rainfall increases to 1,500 mm, streams than are resident species?(3) What will be the and rivers are common, and taller, floristically likely impact of future land-use changeson mi- richer semievergreen forest (“selva alta subpe- gratory landbirds in the Yucatan? rennifolia”) appears. This forest, which is inter- mediate between “tropical dry forest” and “trop- STUDY AREA AND METHODS ical moist forest” according to the Holdridge Fieldwork was concentratedin the Mexican states system, is inhabited by a number of resident bird of Quintana Roo and Yucatan, but migrants were speciesthat do not range farther northward into also surveyed in the states of Campeche and the drier portions of the peninsula (Paynter 1955). Chiapas, and in neighboring Belize and Guate- The wettest climate in the Yucatan Peninsula mala (Fig. 1). The vegetation, climate, and ge- occurs along the Caribbean-facing slope of the ology of the Yucatan Peninsula have been well Maya Mountains in southern Belize. Here, an- described from a zoogeographicalviewpoint by nual rainfall approaches 4,000 mm (Russell Lee (1980), and only a brief account is presented 1964), and the climax vegetation is “tropical wet here. forest” according to the Holdridge classification. A number of distinctive natural vegetation CLIMATE AND NATIVE VEGETATION types (e.g., mangrove swamp, sawgrass-palmetto The Yucatan Peninsula is fairly uniform in ge- savanna, dune scrub) occur along the Yucatan ology and topography. With the exception of the coast(see Lynch et al. 1985: fig. 2). For migratory OVERWINTERING MIGRANTS 517

I I 9o” 1 89O 88O 87’

YUCATAN 0

!OO

CAMPECHE 190

IS’

GUATEMALA BELIZE i

7O 5,o 100 kilometers 170, I 900 +89O I I 88O 87O F IGURE 1. Map of the Yucatan Peninsula showing sites where point counts were made 1982-1987. Large ? fmbols represent > 12 counts;small symbols represent4-l 2 counts.Several localities in northern Chiapas (just SCmthwest of the area mapped) are not indicated. 518 JAMES F. LYNCH

FOREST

WOTt!,KEW, BWWA A

<=>A YEWA INBU RBGRyT A PABU,. _ / ACAHUAL - FIELD FIGURE 2. Diagramindicating relative rates of occurrencein threemajor habitatsfor 2 1 commonmigratory species,based on point-countdata. Contours delimit zonesof equalhabitat diversity, based on exp(H’). Species that plot near the centerof the diagram,where exp(H ’) = 3.0, are habitat generalists(i.e., they use all three successionalstages in proportionto their availability);species that plot neara givencomer are extreme specialists on that successionalstage. Size of eachsymbol is proportionalto overallrate of occurrencein all habitats.Species representedare LeastFlycatcher (LEFC), Wood Thrush (WOTH), White-eyedVireo (WEVI), Gray Catbird (GRCA),American Redstart (AMRE), Black-throatedGreen Warbler (BTGW), Black-and-whiteWarbler (BAWW), Blue-wingedWarbler (BWWA), CommonYellowthroat (COYE), HoodedWarbler (HOWA), KentuckyWarbler (KEWA), Magnolia Warbler (MAWA), Northern Parula Warbler (NOPA), Ovenbird (OVEN), Worm-eating Warbler (WEWA), Yellow Warbler (YEWA), Yellow-throatedWarbler (YTWA), SummerTanager (SUTA), Indigo Bunting(INBU), PaintedBunting (PABU), and Rose-breastedGrosbeak (RBGR). landbirds, the most important of these forma- primary forest has been eliminated from the state tions probably is dune scrub, remnants of which of Yucatan and from northern Campeche (Be- form a narrow, discontinuous fringe along the quaert 1935), although there are sizeable areas shores of both the Caribbean and the Gulf of of secondary forest. To the east, the transition Mexico (Moreno-Casasola and Espejel 1986). from semideciduousto semievergreenforest cor- This vegetation exhibits strong floristic affinities responds approximately to the political border with the islands of the northern Caribbean, and between the statesof Yucatan and Quintana Roo, is the sole or major mainland habitat for several and between a landscapethat has been profound- resident and migratory bird speciesthat are oth- ly altered by human activities and one that re- erwise restricted to the Antilles (Lopez-Omat et tains extensive tracts of old-growth forest (Lynch al. 1989). et al. 1985: fig. 3). Even in forested areas, the presenceof literally hundreds of Maya ruins tes- HUMAN MODIFICATION OF tifies to the fact that virtually the entire peninsula NATIVE VEGETATION has been subjected to human disturbance in the For thousands of years natural vegetation gra- past. dients in the Yucatan Peninsula have been al- At least since the collapse of the Classic phase tered, sometimes virtually beyond recognition, of Maya culture (ca. 900 A.D.), human settlement by human land-use. In the semiarid northwest, has been concentrated in the dry northwestern a century and a half of intensive cultivation of portion ofthe Yucatan (the present Yucatan state henequen (the source of sisal fiber) has com- and adjacent northern Campeche). The remain- pounded the effects of more than two millenia der of the peninsula had retained extensive areas of slash-and-bum maize agriculture. As a result, of native forest until recent decades,when con- OVERWINTERING MIGRANTS 519

Dry, Medium Ht.

Moist, Medium Ht. Moist, Toll FIGURE 3. Diagram indicating occurrence rates in three widely distributed forest types by frequently en- countered migrants (large closed symbols) and residents(open symbols). See legend to Figure 2 for explanation of contoursand mnemonic abbreviationsfor migratory species.Infrequently encounteredmigrants (small closed symbols) that are not representedin Figure 2 are Ruby-throated Hummingbird (rthu), Northern Waterthrush (nowa), Cape May Warbler (cmwa), Palm Warbler (pmwa), Swainson’s Warbler (swwa), Yellow-mmped Warbler (yrwa), Blue Grosbeak (blgr), Northern Oriole (noor), and Orchard Oriole (oror). struction of new roads began to allow large-scale 1970). However, the area that is presently for- agricultural and ranching development in central ested may actually exceedthe extent of forest at and southernCampeche, southern Quintana Roo, the height of the ClassicMaya civilization, 1,OOO- and northern Belize. The largestremaining blocks 1,500 years ago. Clearly, migratory birds, like of essentially intact tropical forest occur in the other elements of the Yucatan’s biota, have had northern two-thirds of Quintana Roo, and in an to accommodate a long history of human dis- irregular area that extends eastward from the turbance. northern Peten of Guatemala and southern Cam- peche,through southernmost Quintana Roo, and HABITAT CLASSIFICATION into Belize. The remainder of the peninsula is For purposes of analysis nine major classes of now mostly a patchwork of pastures,permanent vegetation were recognized: (1) active pasture fields and orchards,shifting agricultural plots (lo- (potrero) or cornfield (milpa), (2) brushy old field cally termed “milpas”), brushy old fields (“aca- (acahual), (3) native coastal scrub, (4) semide- huales”), and secondary forest. ciduous forest, (5) semievergreen forest, (6) ev- There is less mature forest in the Yucatan to- ergreen forest, (7) mangrove swamp forest, (8) day than at any time since the 16th century Span- riverine floodplain forest, and (9) seasonally in- ish conquest, a period of catastrophic depopu- undated low woodland (tintal). Some forest types lation for the area’s Maya inhabitants (Thompson (categories4-9) were further subdivided into two 520 JAMES F. LYNCH

TABLE 1. Number of point counts performed in 19 habitats in the Yucatan Peninsula, 1982-1987.

Stature of vegetation LOW Medium Tall Very tall Major habitat group (53 In) (410 In) (I I-20 m) (>20 m) Total counts

Open field 92 92 Native coastal scrub 60 60 Savanna 4 4 Brushy old field 72 72 Semideciduousforest 60 116 8 184 Semievergreenforest 20 196 208 424 Moist forest 4 12 36 12 64 Tall floodplain forest 16 16 Mangrove forest 20 28 Seasonallyinundated forest 24 32 Grand total 916

to four height intervals, such that the total num- olds et al. 1980, DeSante 198 1). However, com- ber of defined habitat categorieswas 19 (Appen- parisons with the results of intensive spot-map- dix I). Each bird survey point was assigned to ping indicate that the methods based on point one of these 19 habitat types (Table 1). Details counts may yield misleading absolute (or even of vegetation structure and floristic composition relative) densities for most of the speciesin some were documented at several hundred 0.04-ha bird communities, particularly in forested eco- plots located within 20 m of survey points. The systems(DeSante 1986, Hutto et al. 1986). We results of the latter study will be presented else- therefore interpret our point counts as indices of where (J. F. Lynch, D. F. Whigham, and E. S. relative abundance (Verner 1985) that are best Morton, unpubl.). suited for comparisons of the prevalence of a given species across habitats or geographic re- POINT SURVEYS gions. Migrant and resident birds were surveyed in the Within a given tract, each of the four survey Yucatan Peninsula during the winters of 1982- points were located at least 150 m apart, and 1987 (Fig. 1). Our point-count technique com- were no closerthan 100 m to habitat boundaries, bined aspects of the I.P.A. (“Index Ponctuel except where the linear form of the habitat (e.g., d’Abondance”) and E.F.P. (“Echantillonnage mangrove fringe) made this impossible. In prac- Frequentiel Progressif‘ ) methods developed by tice, we were able to achieve the desired spacing Blonde&Ferry and coworkers(Blonde1 et al. 1970, between four survey points in most stands that 1981; Ferry 1974; Blonde1 1975, 1977), and pre- were larger than ca. 15 ha. Where single tracts viously applied to North American breeding bird of this extent were unavailable (e.g., small slash- communities (e.g., Whitcomb et al. 198 1, Lynch and-bum agricultural plots), we divided the re- and Whigham 1984). In each tract of vegetation quired four points between two nearby tracts of we counted all birds that were seen or heard (re- a given vegetation type. Stands of less than 2 ha gardlessof distance) on a single visit to each of were not included in our study. four survey points. We used a lo- to 12-min The procedure was for two observers (occa- observation period per point, rather than the 20 sionally more) to walk 100-l 50 m into a desig- min specified for both the I.P.A. and the E.F.P. nated tract, wait a few moments for bird activity methods. However, counts of 1O-l 2 min should to equilibrate, and begin the count. Surveys were give acceptablecoverage, according to the results continued for 10-l 2 min, the shorter interval of comparative studies of count efficiency (Scott being used only if no new birds had been detected and Ramsey 198 1, Pyke and Recher 1985, Ver- between min 8 and min 10. Becausethe primary ner 1985, Hutto et al. 1986). focus of this study was the status of nearctic mi- Variants of the unlimited point-count method grant landbirds, many of which respond to gen- (e.g., the variable circular-plot method, the fixed- eralized distresscalls and other stimuli, we em- radius point count method) have been used to ployed several techniques to increase the estimate absolute densities of birds (e.g., Reyn- likelihood of detecting migrants at close range OVERWINTERING MIGRANTS 521 within the count period. First, territorial chip Standard 12-m x 2-m nylon mist nets (mesh notes of two common overwintering migratory size 3.2 cm and 3.8 cm) were operated inter- species (Hooded Warbler and Kentucky War- mittently between 1984 and 1987 at 15 locations bler) were played back for 3-5 min, using a Uher in northern and central Quintana Roo (Fig. 1). 4000 Report Monitor tape recorder. Imitations Each site was an essentially homogeneous stand of the distress squeak call of a small passerine of a major natural or secondaryvegetation type, and thepoip-poip-poipwhistle ofthe Ferruginous ranging from pastures and milpas, through old Pygmy-Owl were then used to elicit responses. fields of varying ages, to mature semievergreen To quantify the effectsof these stimuli, a series forest (Table 2). At each site 20-36 nets were set of paired comparisonswere made between counts either in a single line or in two or three parallel using aural stimuli counts where no active at- lines of eight to 12 nets each. Net lines were tempt to elicit responseswas made. The results separated by at least 50 m. In one intensively of these comparisons, which will be presented studied tract of semievergreen forest, two mist elsewhere(J. Lynch, unpubl.), indicated that mi- nets were placed in each of 12 40-m x 40-m gratory warblers and vireos showed minor, but plots that extended over an area of about 5.5 ha. a statistically significant, positive responseto our A netting bout consistedof operating all the nets aural stimuli, but resident speciesas a group were at a given site for 24 (usually 3) consecutive unaffected. The main benefit to this study of us- days. The number of bouts per site ranged from ing aural stimuli was that they tended to cause one to six, with most sites being sampled on two migrants to approach the observer more closely, to four occasions(Table 1). Nets were opened at thereby allowing more accurate sex/age deter- first light, and were closed at 11:30-12:30, by mination, reading of color bands, etc. which time capture rates usually had declined Aerial foragers that did not make use of the and rising temperatures tended to stressnetted vegetation were not counted, as they could not birds, especially in the more open habitats. Cap- be unambiguously assigned to particular habi- tured birds were identified to species,sexed and tats. Thus, many common species(e.g., vultures, aged (where possible), weighed, and color-band- soaring hawks, swifts, swallows) did not appear ed for future recognition. “Captures” refer to the on our survey lists unless they were observed number of different individuals netted per net perched or otherwise associatedwith terrestrial hour (or net morning) in a given bout. Thus, the vegetation. same bird caught in two different years (or at two different seasonsof the same year) was counted MIST NETTING as two captures, but an individual captured re- Because many previous studies of nearctic mi- peatedly within the same bout was counted only grants in the Yucatan and elsewherein the neo- once. tropics have employed mist netting to estimate patternsof relative abundance(e.g., Tramer 1974; RESULTS Waide 1980; Waide et al. 1980; Karr 1981a, 198 1b), and becausemark-recapture data arising DIVERSITY OF THE AVIFAUNA from netting studies can provide valuable infor- Fieldwork between the months ofNovember and mation on home-range size and survivorship in March revealed a total of 2 13 speciesof landbirds some species(e.g., Karr 198 la), we supplement- (excluding aerial foragers), of which 44 (21%) ed our point counts with mist netting. We rec- were overwintering nearctic migrants (Appendix ognize that the latter technique is strongly biased II). Limited observations in October and April against speciesthat are mainly active in the forest revealed the presenceof several additional tran- canopy, or that are too large or too small to be sient migratory species(Appendix II), but these efficiently captured by nets of a given mesh size. are not included in our analysis of winter bird Nevertheless, mist-net capture rates can be communities. Of the overwintering migratory meaningful indices of abundance for speciesof species,one (Cedar Waxwing) was casually ob- appropriate size that are active mainly in the served on a single occasion, but was not en- ground and shrub strata. In addition, we believed countered during either the point counts or mist- it was important to compare the results of mist- netting surveys.The other 43 winter visitors were netting studies directly with those of the point- either recordedonly in point surveys(10 species), count method. only by mist netting (5 species),or by both sur- 522 JAMES F. LYNCH

TABLE 2. Localitiesand habitatswhere mist nettingwas conducted in the Mexicanstate of QuintanaRoo in 1984-1987.Only data for the time of yearwhen migrants were present (late Augustto late April) are presented. Localitieslabelled “ Sian Ka’an” are within the Sian Ka’an BiosphereReserve, east of Felipe Carillo Puerto,in centralQuintana Roo; “Puerto Morelos” refersto a clusterof localitieswithin 15 km of the village of Puerto Morelos,on the northeasterncoast of QuintanaRoo.

Major habitat type Locality Netting bouts Net hours Net mornings Active field or pasture Sian Ka’an 4 673 164 PuertoMorelos 1 310 69 Subtotal 5 983 173 Early acahual PuertoMorelos 1 144 32 (l-3 years) Sian Ka’an 1 528 96 PuertoMorelos 1 459 34 Subtotal 3 1,131 162 Mid-stageacahual Sian Ka’an 1 264 48 (4-7 years) PuertoMorelos 3 1,356 280 Subtotal 4 1,620 328 Semievergreenforest PuertoMorelos 6 2,826 576 (medium stature) PuertoMorelos 4 1,920 336 Sian Ka’an 1 240 48 Sian Ka’an 1 384 72 PuertoMorelos 2 598 120 Subtotal 14 5,968 1,152 Semievergreenforest PuertoMorelos 2 858 141 (tall) Native coastalscrub Sian Ka’an 2 322 56 Sian Ka’an 2 644 116 Subtotal 4 1,824 172 Grand total 32 11,562 2,128

vey methods (28 species).A majority (28/43 = patterns (see below-Results). The remaining 65%) of the overwintering speciesof migratory seven habitat types were each sampled 60-208 landbirds belonged to the emberizid subfamily times. Even for the commonest species of mi- Parulinae (Appendix II). grants, the mean number of individuals detected in point counts where the species was present POINT SURVEYS was only slightly greater than unity (X = 1.11; A total of 976 point counts was made during range = 1.07-l .26; n = 12 species).“ Frequency” the winters of 1982-1987. The following 12 hab- and “abundance” are assumed to be related, but itat types were not censusedfrequently enough statistical analyses are based on presence/ab- (< 50 counts) to permit detailed statistical anal- sence data, not on absolute numbers. Absolute ysis (Table 1): tall semideciduousforest (8 points), densities were not determined in this study, ex- low semievergreen forest (20) low (4) medium cept for Hooded Warblers (Green et al. 1988). (12) tall (36), and very tall (12) moist forest, Successionalpatterns. Major successional savanna (4) low mangrove forest (20) medium trends were summarized by comparing point sur- mangrove forest (8) tall floodplain forest (16) vey data for three common habitat classes:(1) low inundated deciduous forest (24), and me- fields and pasturesthat were in active use or had dium inundated deciduous forest (8). Results of been abandoned for less than a year, (2) brushy surveysin these 12 habitat types are not included old fields (acahuales) 4-7 years into abandon- in the following statistical analysis, but some ment, and (3) mature medium-stature semiev- qualitative results are noted in our discussions ergreen forest (selva mediana subperennifolia), of general habitat associations and geographic an extensive vegetation type in the eastern half OVERWINTERING MIGRANTS 523 of the peninsula (Fig. 2). Most of our mist netting in acahualeswere Magnolia Warbler, Common was conducted in this forest type and its seral Yellowthroat, Hooded Warbler, White-eyed Vir- stages. eo, Least Flycatcher, Gray Catbird, Indigo Bunt- Active and very recently abandoned fields ing, and American Redstart. Four of these eight (milpas) and pastures(potreros), which in Quin- speciesare also associatedwith brushy and edge tana Roo almost always contain at least a few habitats on their breeding grounds, and all except living and dead trees and clumps of brush, ap- the Hooded Warbler and American Redstart were peared to support a rich bird community (Table frequent in field-brush communities as well as 3), although it was not known whether or not acahuales during winter (see above). The most birds observed in cleared areas also made use of frequently detected resident speciesin acahuales adjacent forest. Nearctic migrants accounted for (Scrub Vireo, Melodious Blackbird, Tropical 17% (17/101) of the speciesencountered in 92 Kingbird, Golden-fronted Woodpecker, Hooded point counts within this habitat, and migrants Oriole) were also commonly encounteredin fields were among the most frequently detectedspecies: and pastures (Table 3). 10 of the 34 most commonly encountered field- The 404 point surveys conducted in late- associated species (29%) were winter visitors, successionaland mature medium-stature semi- which also accounted for 27% of all individual evergreenforest (selva mediana subperennifolia) birds that were tallied. In order of decreasing revealed the presenceof 99 bird species,of which frequency of occurrence, the main migratory 18 (18%) were nearctic migrants (Table 3). Five speciesin potreros and milpas (hereafter termed of the 10 most frequently encountered speciesin “field/brush” communities) were Common Yel- semievergreen forest were migrants, which also lowthroat, Least Flycatcher, Magnolia Warbler, accounted for 40% of the total individuals en- Indigo Bunting, Gray Catbird, White-eyed Vir- countered. In decreasing order, the most fre- eo, Northern Parula Warbler, Yellow-throated quently detected migrants in semievergreen for- Warbler, Rose-breasted Grosbeak, and Yellow est were the Hooded Warbler, Magnolia Warbler, Warbler (Table 3). American Redstart, White-eyed Vireo, Black- Four of the six most frequently detected per- and-white Warbler, Wood Thrush, Black-throat- manent residents in the field-brush community ed Green Warbler, Kentucky Warbler, Gray Cat- (Golden-fronted Woodpecker, Tropical Mock- bird, and Least Flycatcher. Whereas most of the ingbird, Tropical Kingbird, Scrub Vireo) are eco- migratory speciesthat occurred in acahualesalso logically similar to congeners that inhabit dis- were found, at least occasionally, in heavily dis- turbed habitats in temperate North America. turbed open fields, only two typical acahual Thus, the winter bird community associatedwith species (Magnolia Warbler, White-eyed Vireo) field-brush vegetation appearsto be numerically were also commonly encountered in mature for- dominated by nearctic migrants and by resident est. When the degreeof specialization of migrant speciesthat have closely related counterparts in specieson actively disturbed (Field), midsucces- the North Temperate Zone. sional (Acahual), and late successional(6-20 m The 72 point surveys conducted in acahuales height) semievergreen forest (Forest) is sum- (successionalcommunities dominated by shrubs marized in a triangular diagram (Fig. 2) most and young trees) revealed the presence of 80 speciescluster along a line passing through the species(Table 3). Although fewer acahualesthan “Forest” apex, and bisecting the Field-Acahual field-brush sites were sampled, both the total axis. number of migratory species(19) and the pro- Forest stature appeared to be a relatively un- portion of migratory species(24%) that was de- important determinant of migrant occurrence. tected were somewhat higher in acahuales. As There were virtually no statistically significant was also true in field-brush communities, mi- differencesbetween the rates of occurrenceof any grants were disproportionately represented migrant speciesin medium-stature (6-10 m) vs. among the most frequently encountered species: taller (1 l-20 m) selva mediana subperennifolia seven of the 10 most commonly detected species (Table 4). in acahuales were migrants (Table 3) and mi- Of the 10 most frequent resident species in grants accounted for 33% of all individual birds agricultural fields and pastures,only one (Brown that were encountered. In decreasingorder, the Jay) also ranked among the 10 most frequent most frequently encountered migratory species speciesin mature forest. Similarly, only one of

OVERWINTERING MIGRANTS 525

the 10 most frequent resident species in aca- green) were classed as forest generalists (i.e., huales (Fawn-breasted Hummingbird) also exp[H’] > 2.7) only 19 of the 61 (31%) most ranked within the 10 most frequently detected frequently detected forest-dwelling resident species in forest. No resident species was fre- specieswere so classed. At the other extreme, quent in all three seral stages,although a number none of the 12 most frequent forest-dwelling mi- of speciesoccurred at least occasionally across grants was classifiedas highly specialized on one the entire successionalcontinuum (Table 3). At forest type (i.e., exp[H’] < 1.6), but 7 of 6 1 (12%) the level of the three successionalstages that were of the permanent residentsfell into this category. compared,the 26 most frequent migratory species A tendency for residents to specialize on tall, and the 52 most frequent residents did not differ moist forest is evident in a triangular plot of the significantly in their degree of specialization on three forest types (Fig. 3). one successionalhabitat (Mann-Whitney U = Occurrenceof migrantsin otherhabitats. Con- 778; P > 0.14).As was true for migrants, resident sistent patterns of occurrenceof some migratory speciesoccurred with similar frequency in me- species in some of the 12 less-studied habitat dium-stature and tall selva mediana subperen- types (seeabove) suggestedrestricted habitat use nifolia (Table 4). in some instances (Tables 5 and 6). Although Comparisonsbetween semievergreenand sample sizesare too small to permit reliable sta- semideciduousforest. Although the semideci- tistical analysis in all species,moist tropical for- duous woodland that occurs in the heavily pop- est (including selva alta subperennifolia), which ulated northwestern portion of the Yucatan is was surveyed in southernmost Quintana Roo, entirely secondgrowth, it commonly attains the Chiapas, Belize, and northern Guatemala, was same stature as mature semievergreenforest that the forest type with the highest occurrencerates is found to the east and south. With few excep- for severalmigratory species,including the Wood tions, the same migratory species occurred in Thrush, Gray Catbird, Kentucky Warbler, and both the drier and more humid forest types, al- Worm-eating Warbler. Conversely, the White- though the relative frequency of most species eyed Vireo, Northern Parula Warbler, Black- differed from one forest type to the other (Table throated Green Warbler, Hooded Warbler, 4). Black-throated Green Warblers and White- American Redstart, and Black-and-white War- eyed Vireos were significantly more frequent in bler were less frequent in the lush forests of the semideciduous forest, while the Wood Thrush, southern peninsula than in dry forests to the Gray Catbird, Hooded Warbler, Kentucky War- north. Yellow-throated Vireo, Magnolia War- bler, Magnolia Warbler, and American Redstart bler, and Ovenbird showed no consistent dif- were more frequent in semievergreenforest. Least ferencesin frequency of occurrencebetween wet Flycatcher, Ovenbird, Black-and-White War- and dry forest types. bler, and Northern Parula Warbler occurredwith The Cape May Warbler, Palm Warbler, Yel- similar frequency in both types of forest. The low-throated Warbler, Yellow-rumped Warbler, only two migratory speciesthat were relatively and Blue Grosbeak all occurred with highest fre- frequent in one forest type but virtually absent quency in coastal scrub, although none of these from the other were the Wood Thrush and Ken- except the Blue Grosbeak was encountered at a tucky Warbler, which occurred, respectively, in sufficient number of survey points to allow the 11% and 13% of 408 point counts in semiever- statistical significance of this pattern to be as- green forest, but each of which occurred in only sessed.The primary tropical wintering areas for l/ 10 1 ( 1%) of the counts in semideciduousforest these speciesare the Bahamas and the Greater (P < 0.01; binomial test). Antilles (Emlen 1977, Rappole et al. 1983) and Resident speciesshowed a significantly greater except for the Cape May Warbler, these species tendency than migrants to occur preferentially in also occurredregularly in pasturesand milpas in either semideciduousor semievergreenforest, al- the northern Yucatan, sometimes many kilo- though most residents were found at least oc- meters from the coast. Other migratory species casionally in both forest types (Table 4). Whereas were relatively scarcein coastalscrub exceptdur- 8 of 12 (67%) migratory species that occurred ing fall and spring, when large numbers of tran- commonly in one or more of the three wide- sient TennesseeWarblers, Red-eyed Vireos, and spread forest types (short semideciduous, short Eastern Kingbirds were present (A. Lopez, pers. semievergreen, and medium-stature semiever- comm.; pers.observ.). A few resident species(e.g., 526 JAMES F. LYNCH

TABLE 4. Degree of specializationof the 88 most frequently encounteredforest-dwelling specieswith respect to three types of forest, based on point-count data. Speciesmarked with an asterisk (*) were more frequently found in nonforest habitats, but also occurred in forest. Numbered codes for resident speciesare identified in Appendix II. Degree of specializationis based on the exponential of the Shannon-Weiner diversity statistic (H’) calculated from the relative rates of occurrencein three forest types: (1) semideciduous,(2) medium height semievergreen,and (3) tall semievergreen.Exp(H ’), which can be interpreted as the number of equally used habitats, can vary from a minimum of 1.0 (complete specialization on one forest type) to a maximum of 3.0 (equal use of all three forest types). Ignoring the 14 specieswhose main habitats were not forest, migrants show significantly less tendency toward habitat specialization than do permanent residents (x2 = 37.3; df = 7; P < 0.005).

Degree of specialization Migrant species (n = 18) Resident species (n = 70)

Generalists Least Flycatcher,* White-eyed Vireo, 14, 38, 42, 45, 63, 73, 77, 84, 86, 96, (exp[H’] > 2.7) Black-throated Green Warbler, 102, 108, 109, 113, 115, 120, 138, Magnolia Warbler, American Red- 150, 151, 155* start (20 species= 29%) (9 species= 50%) Moderate specialists (1.6 < exp[H’] < 2.7) Semideciduousforest Northern Parula Warbler, Common 1, 28, 41, 53, 79, 82, 88, 89, 97,* 99, Yellowthroat,* Rose-breasted 112, 127,* 129, 148,* 165 Grosbeak* (15 species= 2 1%) (3 species= 17%) Medium-stature semi- Summer Tanager 121,* 125, 128 evergreenforest (1 species= 6%) (3 species= 4%) Tall semievergreen Wood Thrush, Gray Catbird,* Yel- 32, 33,* 36, 39, 44, 49, 52, 60, 62, 74, forest low-throated Vireo, Kentucky 92, 107, 119, 139, 140, 141, 143 Warbler (17 species= 24%) (4 species= 22%) Equal occurrencerates (none = 0%) 7, 67, 85, 94 in both semiever- (4 species= 6%) green types Extreme specialists (exp[H’] < 1.6) Semideciduousforest Indigo Bunting* 116, 155, 164* (1 species= 6%) (3 species= 4%) Medium height semi- (none = 0%) 162 evergreenforest (1 species= 1%) Tall semievergreen (none = 0%) 17, 19,* 58, 59, 75, 104, 161* forest (7 species= 10%)

Black Catbird, Tropical Mockingbird, Banana- resident races),Yellow-throated Warbler, Black- quit) reached extraordinarily high densities (as and-white Warbler, American Redstart, and measuredby mist-net capturerates) in some areas Worm-eating Warbler. ofcoastal scrub (A. Lopez and J. Lynch, unpubl.). Degreesof habitat specialization. The method Coastal mangrove forest was inhabited by a of contingency analysis described by Hobbs and wide variety of migrants, and point surveys in Bowden (1982) helps to clarify the pattern of this habitat showed a higher percentage of mi- differential use of natural and disturbed habitats grant individuals (58%) than any other habitat by overwintering migrants. Contingency analysis that was sampled (Table 5). Nevertheless, only of the 23 speciesthat were abundant enough for one migratory species (Northern Waterthrush) meaningful statistical analysis reveals distinctive had its maximum rate of occurrencein mangrove patterns in the use of the nine most frequently forest (Table 6). Other migrants that were fre- sampled habitat categories(Table 6). Six species quently detected in mangrove forest were the (Yellow Warbler, Yellow-throated Warbler, Yel- Common Yellowthroat, Northern Parula War- low-rumped Warbler, Common Yellowthroat, bler, Yellow Warbler (both migratory and Indigo Bunting, Rose-breasted Grosbeak) were OVERWINTERING MIGRANTS 527

TABLE 5. Absolute and relative occurrencerates of nearctic migrants in major vegetationtypes in the Yucatan Peninsula. Tabled entries are mean numbers of species(S) and individuals (I) detectedper cluster of four point counts within each contiguoustract of vegetation that was surveyed.

All birds Migrants Percentmigrants Habitat No. tracts s I s I s I

Field/pasture 22 19.2 49.0 5.6 16.3 29.2 33.3 Brushy old field 16 20.6 40.0 6.2 12.4 30.1 30.1 Secondarydrier forest 31 17.0 28.6 6.2 12.1 36.5 42.3 Inundated drier forest 7 13.7 23.3 6.7 11.6 48.9 49.8 Mature drier forest 9 18.9 28.4 7.4 13.3 39.2 46.8 Low moist forest 10 16.0 30.5 5.6 11.7 35.0 38.4 Secondarymoist forest 48 19.1 37.0 7.0 15.1 36.6 40.8 Mature moist forest 58 21.1 39.2 6.9 15.1 32.7 38.5 Evergreen forest (all) 20 18.2 27.3 6.1 10.7 33.5 39.2 Coastal mangrove forest 7 13.1 27.3 7.1 15.9 54.2 58.2 Coastal scrub 14 19.2 49.0 3.1 10.4 29.2 33.3

TABLE 6. Patternsof habitat specializationin the 23 most common speciesof overwintering nearctic migrants in the Yucatan Peninsula, based on point counts. Overall x2 values and significancelevels (** = P < 0.0 1, *** = P < 0.001, **** = P < 0.0001, ns = P > 0.05) are given for the degreeof heterogeneityshown by each species in its use of nine major habitats. Tabled entries indicate whether occurrencerate in each individual habitat was statisticallydifferent from random at P < 0.1 (Hobbs and Bowden 1982); H = higher than expectedoccurrence rate, L = lower; a minus sign (-) indicatesno significantdifference from random; habitats:Fi = field and pasture; AC = acahual;LD = low, semideciduousforest; MD = medium-height, semideciduousforest; MM = medium- height semievergreenforest; TM = tall semievergreenforest; MO = moist tropical forest; Fl = mangrove and seasonallyflooded forest; CS = native coastal scrub.

Habitat Speci.3 Overall x’ Fi AC LD MD MM TM MO Fl CS

Early successionalspecialists Yellow Warbler 47.9*** - _ _ L L L L _ _ Yellow-throated Warbler 37.4*** H _ L L L L L - _ Yellow-rumped Warbler 23.3*** _ L _ L _ L L _ _ Common Yellowthroat 100.1*** H - _ - L L L _ _ Indigo Bunting 62.8*** H H _ _ L L _ L _ Rose-breastedGrosbeak 24.4*** H _ - _ _ L L L L Generalists Least Flycatcher 15.4 (ns) ______Gray Catbird 25.2*** _ _ - L _ _ _ _ _ White-eyed Vireo 31.0**** _ - _ - _ _ _ _ _ Northern Parula Warbler 8.7 (ns) - ______Magnolia Warbler 107.8**** - _ L _ H H _ L L Forest generalists Black-throated Green Warbler 48.1**** _ L _ H _ _ _ _ L American Redstart L ______L Ovenbird ;;:;I::: L ______L Forest specialists Wood Thrush 65.0**** L L L L _ H H L L Yellow-throated Vireo 22.6*** L L L _ - H L L L Blue-winged Warbler 52.2*** L L L _ _ _ - _ L Black-and-white Warbler 63.2**** L L - _ _ H _ L L Kentucky Warbler 52.6**** L L L L _ H H L L Hooded Warbler 228.3**** L _ _ - H H L L L Extreme habitat specialists Wilson’s Warbler 73.6**** L L L L L L H L L Northern Waterthrush 86.8**** _ _ _ _ L - L H _ Blue Grosbeak 20.6** L L L _ L L L L H 528 JAMES F. LYNCH significantly more frequently encounteredin fields northern and central Quintana Roo, where all of or pastures, or in early old fields, than in any of our mist netting was conducted. several types of forest. Another group of five Active pastures and recently abandoned agri- species(Least Flycatcher, Gray Catbird, White- culturalfields. These brushy field habitats, which eyed Vireo, Northern Parula Warbler, Magnolia were sampled at five locations (Table 2) yielded Warbler) occurred across essentially the entire 72 species,of which 20 (27.8%) were migrants. successionalspectrum, and were found in semi- The total capture rate (CR = 100 x number of deciduous, semievergreen, and evergreen forest. individuals captured/net/morning) was high in Although the habitat distribution of three of the pasturesand fields (CR = 247.7), but the sample five specieswith the weakest tendency to spe- was numerically dominated by only a few abun- cialize in their habitat use is not completely ran- dant species.The two most frequently captured dom, occurrencesin any one habitat type are not species,the Indigo Bunting (a migrant) and White- substantially higher or lower than would be ex- collared Seedeater(a resident), together account- pected by chance. These five migratory species ed for 30.4% of all captures in open habitats may be termed “habitat generalists.” (Table 3a). Both speciesoccur in flocks during Three species (American Redstart, Black- winter. As was true for point counts, migrants throated Green Warbler, Ovenbird) tended not made up a disproportionate number (9 of 24 = to occur in the highly disturbed vegetation of 37.5%) of the speciesthat were most frequently milpas and pastures,but were found with similar netted in open habitats. In order of decreasing frequency acrossthe entire remaining spectrum abundance the most frequently captured mi- ofwooded habitats, regardlessof stature or mois- grants in open fields were Indigo Bunting, Com- ture regime. Such speciesmay be termed “forest mon Yellowthroat, Gray Catbird, White-eyed generalists.” Vireo, Least Flycatcher, Painted Bunting, Mag- A sizeablegroup of migrants (six species:Wood nolia Warbler, and Yellow-breasted Chat. Thrush, Yellow-throated Vireo, Blue-winged Brushy oldjields (acahuales).This structurally Warbler, Black-and-white Warbler, Kentucky and floristically heterogeneoushabitat was sam- Warbler, Hooded Warbler) can be categorizedas pled at two sites (Table 2) that were 5-7 years “forest specialists.” These specieswere detected into abandonment when netted. At this stage, with significantly higher frequency than expected regenerating pastures and old fields in central in mature forest,and with significantly lower than Quintana Roo typically contain patchy stands of expectedfrequency in early successionalhabitats Cecropia and other fast-growing trees, areas of (Table 6). All six species occurred more fre- shrubs and tall bracken fern (Pteridium), and quently in semievergreenand/or evergreenforest remnant patchesof the grassesand weedstypical than in semideciduous forest. of earlier successionalstages. A total of 328 net Finally, three species(Northern Waterthrush, mornings in acahuales yielded 87 bird species, Blue Grosbeak, Wilson’s Warbler) were closely ofwhich 22 (25.3%) were nearctic migrants. The associatedwith a single habitat type (inundated total capture rate was high (CR = 113.3), though woodland, coastal scrub, and moist evergreen lower than in open fields. Numerical dominance forest, respectively), within the portion of the was also relatively high, and the two commonest Yucatan that was studied. Such species,which species(the migratory Gray Catbird and the res- may be termed “extreme habitat specialists,” ident Blue Bunting) accounted for 23.0% of all rarely or never were encountered in any other individuals captured (Table 3b). Ten of the 27 habitat. most frequently captured species (35.2%) in brushy old fields were migrants. In order of de- MIST-NETTING RESULTS creasing abundance, the most frequently netted Mist nets were operated at 15 sites for a total of migratory specieswere the Gray Catbird, White- 2,128 net mornings (11,526 net hr). In all, 144 eyed Vireo, Wood Thrush, Common Yellow- bird specieswere captured, of which 38 (26.4%) throat, Hooded Warbler, Least Flycatcher, Mag- were nearctic migrants. The following discussion nolia Warbler, and Indigo Bunting. focusseson the use by migrants and resident Medium-stature semievergreen forest. This speciesof the successionalcontinuum of habitats widely distributed foresttype was sampled at five typical of the zone of semievergreen forest in locations (Table 1). Interviews with local resi- OVERWINTERING MIGRANTS 529

dents and comparisons with other stands of maturity of habitats. This analysis was restricted known age indicated that none of the forestswe to three habitat categories: pastures and fields studied was lessthan 30 years old, and most were lessthan 1 year into abandonment (473 captures; more than 50 years old. A netting effort of 1,096 67 species),acahuales 5-7 years into abandon- net mornings in semievergreenforest yielded rel- ment (556 captures; 74 species), and medium- atively few captures (CR = 40.5) and revealed stature semievergreen forest (662 captures; 54 the presenceof only 54 species.To a degree,the species).Rarefaction of these data (Table 7; Fig. low capture rate in this habitat reflected the fact 4) revealed that any given number of captures some siteswere netted monthly or bimonthly for would be expectedto include essentiallythe same part ofthe study period. Net avoidance may have number of species in pastures as in acahuales. reduced the successrate per net per morning in However, for mature forest, the predicted num- such instances (Karr 198 1a), but even if data for ber of speciesfor a given number of captureswas frequently sampled plots were disregarded, the approximately 30% lower than in earlier succes- capture rate for maturing forest was only a frac- sional stages. This result holds even for fairly tion of that in earlier successionalstages. large hypothetical samples (300-500 captures). Migrants made up a relatively high percentage As noted earlier, mist-net samplesin forest inev- (25.4%) of the species netted in semievergreen itably under-sample speciesthat are active main- forest, and comprised an even higher proportion ly in the canopy, and the total number of species of the most frequently netted species: seven of in a tropical forest might be substantially greater the 18 commonest speciesin our samples(38.9%) than in successionalhabitats. Nevertheless, the were migrants. Dominance wasunexpectedly high mist-netting data reinforce the point made ear- in this habitat, and the two most frequently net- lier on the basis of point survey results: Second- ted species, both residents (Red-throated Ant- ary vegetation in the Yucatan is used by a diverse Tanager and Ruddy Woodcreeper), contributed assemblage of both resident and migratory 44.2% of all captures (Table 3~). The most fre- species. quently netted nearctic migrants in semiever- green forest were the Hooded Warbler, Oven- DISCUSSION bird, Wood Thrush, Kentucky Warbler, DIVERSITY AND ABUNDANCE OF Black-and-white Warbler, White-eyed Vireo, and NEARCTIC MIGRANTS IN THE Magnolia Warbler. YUCATAN PENINSULA Efect oftotal captureson apparentspecies rich- Data from point counts and mist-net surveys ness. Because different numbers of individuals agreed in showing that overwintering nearctic were captured in each habitat, observed be- migrants were both diverse and common in most tween-habitat differences in the total species habitatsthroughout the Yucatan. Despite the very richness of the nettable portion of the commu- different sampling biasesof the two methods, the nity, and in the relative abundance of migrant proportion of migrant species and individuals species,could be interpreted as artifacts of dif- was similar in point counts and mist-net samples ferent sample size (Sanders 1968; Simberloff of a given habitat type. However, the rank order 1972, 1978). To test this possibility, the statis- of individual species abundances varied sub- tical technique of rarefaction was applied to the stantially between the two sampling methods mist-net capture data, using the program SIM (Table 3). Results from both survey methods sug- (Simberloff 1978). When mist-net capture data gest that in most habitats a disproportionately for open fields, early acahuales, mid-stage aca- large fraction of the most frequently detected (or huales, and semievergreenforest were rarefied to netted) specieswas migrants. a common number of capturesper site (n = 100) Theseresults differ in some major respectsfrom the predicted number of migratory species per those reported by Tramer (1974) who surveyed sample of captures did not differ significantly (P migrants and residentsin nine plots in the north- > 0.05) among the four successionalstages (X = westernYucatan. Although the proportion of mi- 13.5; range = 11.2-14.2). grant speciesrecorded by Tramer (7-40%; me- A secondrarefaction analysis helped to clarify dian = 28%) was similar to the values determined the relationship between total speciesrichness in in the present study, he reported that migrants the ground/shrub stratum and the successional made up only 3-36% (median = 14%) of the 530 JAMES F. LYNCH

Brushy 01Id-field \

Mature Forest

4 100 200 300 400 500 600 NUMBER OF CAPTURES FJGURE 4. Rarefaction curves for mist-net data, showing the rate at which speciesrichness increasesas a function of the number of captures. The vertical lines at n = 450 capturesrepresent 2 standard deviations of the projected mean speciesrichness. individuals in the communities he surveyed.That The results of the present study are in closer is, Tramer concluded that the average migrant accord with those reported by Waide (1980), specieswas less abundant than the average res- based on his mist netting of six plots in south- ident. This disparity between Tramer’s results easternCampeche. There, the median percentage and ours is difficult to reconcile, even allowing of new capturescontributed by migrants was 50% for the fact that our use of playbacks slightly (range = 32-57%). When Waide’s data for three increasedthe proportion of migrants to residents early successionalsites are pooled, migrants made in our point surveys (see above). Tramer’s con- up 20/65 (30.8%) of the specieshe captured, a clusions were based on censuses of a limited proportion that is similar to the 20/72 (27.8%) number of plots, most of which were too small we found for similar disturbed habitats in Quin- (< 10 ha) to justify general conclusions on rela- tana Roo. Waide’s (1980) data indicate that in tive abundance (Waide et al. 1980). Apparently, southeasternCampeche migratory speciesmade numerous identification problems also con- up 14/43 (32.6%) of the species netted in two founded some of Tramer’s (1974) results (Austin areas of maturing semievergreen forest, as com- et al. 1981). pared with 14/55 (25.4%) for the same type of forest in Quintana Roo. The slightly lower pro- portion of migrant speciesin the two Quintana TABLE 7. Rarefaction analysis of mist-net capture Roo habitats can reasonably be attributed to our data for three major successionalhabitats in Quintana greater sampling intensity in the latter state (189 Roo. Entries are predicted means (and standard de- net days vs. Waide’s 14 1 net days in early succes- viations) of speciesas a function of the total number of captures(n). For any given subsamplesize, species sional fields; 1,096 net days vs. Waide’s 136 net richnessis significantlylower in mature forest than in days in maturing semievergreenforest). With in- fields or acahuales. creasing netting effort, the number of resident speciescontinues to increase as more and more Predictedspecies richness rare speciesare added (Fig. 4) whereasthe num- Individuals Field ACahlK%l Forest in subsample (n = 473) (n = 556) (n = 662) ber of migrant speciescaptured in a given habitat quickly approachesan asymptote. 50 25.9 (2.6) 27.8 (2.6) 21.6 (2.3) In maturing semievergreen forest, the propor- 100 38.4 (2.9) 40.1 (3.0) 29.6 (2.6) 200 52.2 (2.7) 54.0 (3.0) 38.2 (2.5) tion of captured individuals that belonged to mi- 300 59.7 (2.2) 62.3 (2.6) 43.4 (2.4) gratory species was similar in Quintana Roo 400 64.5 (1.4) 67.9 (2.1) 47.2 (2.1) (32.9%) and Campeche (36.7%), but in early 500 - 72.1 (1.3) 50.3 (1.7) successionalhabitats the proportion of migrants 600 - 51.6 (1.4) was much higher in Campeche(5 5.1% vs. 38.8%). OVERWINTERING MIGRANTS 531

This difference is almost entirely due to the ex- currence rates in samples from these disturbed traordinary abundance of one migratory species, habitats. However, many migratory specieswere the Indigo Bunting, in the three old fields sam- more frequent in one or another classof tropical pled by Waide in Campeche. This patchily dis- forest (or in forest generally) than in pasturesand tributed (Rappole et al. 1983; pers. observ.), early old fields (Tables 3, 6). Despite this fact, flocking speciesaccounted for 60.5% of the mi- no migratory specieswas restrictedto old-growth grants Waide (1980) captured at his three early forest. successionalsites. Although the Indigo Bunting In agreementwith previous studiesof breeding was also the single most abundant migrant we bird communities in eastern North America netted in open fields in Quintana Roo, there the (Willson 1974, May 1982, Mehlhop and Lynch speciesaccounted for only 124/287 (43.1%) of 1986) overwintering migrants appear to show a the total migrants captured. Migrants other than discontinuity in habitat use between the shrub Indigo Buntings made up similar proportions of stageof succession(when the tree canopy has not the total capturesin the field community in Cam- yet closed) and closed-canopy forest, regardless peche (32.6%) and Quintana Roo (30.1%). Ex- of its stature. As has also been documented dur- treme seasonaldominance of tropical bird com- ing the breeding season&richer 1973, Mehlhop munities by a single migrant speciesis unusual, and Lynch 1986) some migrants whose main but not unique. As an example, Emlen (1977) winter habitat is forest (e.g., Black-and-white reported that Palm Warblers made up 77% of Warbler, Black-throated Green Warbler, Oven- the total winter bird community in open pine bird, Worm-eating Warbler, Magnolia Warbler) woodlands in the Bahamas. were also regularly encountered in early succes- Summarizing, the results of the present study sional habitats. Although some of these species are in basic agreementwith thoseof Waide (1980) (e.g., Black-throated Green Warbler) tended to indicating that overwintering birds make up about be restricted to hedge rows and remnant trees one-third of the individuals from netted samples within areas of early secondaryvegetation, most in Yucatan forests, and up to one-half or more were also observedto foragein low, densegrasses of the individuals netted in secondary habitats. and weedy vegetation. Whether some individu- The migrants captured in secondaryhabitats are als of such speciesactually restrict their winter dominated by one species,the Indigo Bunting. activity to early successionalvegetation is an im- portant question whose answer will require de- HABITAT ASSOCIATIONS OF tailed studies of known individuals through the MIGRANTS AND RESIDENTS winter. In the Yucatan Peninsula a few overwintering In the Yucatan, overwintering forest-associ- migrants (e.g., White-eyed Vireo, Magnolia War- ated migrant speciesalso commonly made use bler) occurred acrossessentially the full range of of early successional habitats, but migratory successionalhabitats, although they were not en- speciesthat were most frequent in old fields tend- countered with equal frequency in all vegetation ed not to utilize closed-canopy forest. Of nine types. Other migrants occurred at much higher migrant specieswhose main winter habitat in the than expected frequencies in open fields and Yucatan was either natural scrub or early succes- brushy successionalhabitats (e.g., Common Yel- sional vegetation, six (Yellow-throated Warbler, lowthroat, Yellow-breasted Chat, Indigo Bunt- Yellow Warbler, Common Yellowthroat, Indigo ing, Rose-breasted Grosbeak), or in closed-can- Bunting, Painted Bunting, Rose-breasted Gros- opy forest(e.g., Wood Thrush, Kentucky Warbler, beak) were rarely or never observed in mature Black-and-white Warbler, American Redstart). forest (Table 3, Fig. 2). As a group, migrants were no more strongly as- Residents and migrants did not differ signifi- sociated with disturbed habitats than were resi- cantly in their tendency to specialize on a par- dents, but this was not becausemigrants avoided ticular successionalstage of vegetation, but for- secondaryvegetation. Instead, both migrants and est-dwellingresident species were more likely than residentsoccurred at relatively high frequency in forest-dwelling migrants to occur at higher than early successionalvegetation (cf, Willis 1980). expectedfrequency in a particular foresttype (Fig. The co-occurrence of forest-adapted and field- 3). Half of the 18 migratory speciesthat occurred adapted species in secondary vegetation pro- frequently in at least one of three widespread duced very high local species richness and oc- forest types (medium-stature semievergreen, tall 532 JAMES F. LYNCH

TABLE 8. Summary of correlation analysis of mist- COMPARISON OF POINT COUNTS AND net caoture rates vs. occurrencerates in uoint counts. MIST-NET SURVEYS Tabled values are Spearman’s rank correlation coeffi- cients and the total number of speciesrecorded in each There was no significant correlation between the habitat. Significancelevels: ** = P < 0.01; * = P < rank order of occurrencerates of individual res- 0.05; ns = P > 0.05. ident species as estimated by point counts vs. mist-net surveys in any of the three successional Mierants Residents habitats where comparisons could be made (Ta- Fields and pastures 0.49* 0.20 ns ble 8). For migrants, Spearman’s rank correlation (n = 21) (n = 79) coefficientsfor point-count data vs. mist-net sur- Brushy old fields 0.49** -0.02 ns vey resultswere statisticallysignificant (rs= 0.48- (n = 26) (n = 83) 0.49; P -c 0.05)but the coefficients of determi- Medium semievergreen 0.48* 0.21 ns nation (0.23-0.24) were relatively small. For sur- forest (n = 20) (n = 88) veys of entire bird communities, point counts appear to be superior to mist-netting surveys, both becauseof the latter method’s biasesagainst speciesthat are large, or that forage high in the semievergreen, medium semideciduous) were canopy, and because mist netting is relatively classedas forest generalists,i.e., they occurred at costly in time and labor. As a concrete example, approximately expected frequencies in all three a 3-day bout of operating 20-30 mist nets in forest types. Eight additional migratory species Quintana Roo, plus the additional time needed showed a moderate degree of concentration in to cut lanes, and to set up and take down the one forest type, but only one species (Indigo nets, typically produced a strongly biased sample Bunting) was categorizedas highly specializedon of 35-100 individual birds. By comparison, four one of the three forest types. The latter species mornings of point counts would be expected to is, in fact, mainly associated with brushy old yield 32-64 individual surveys, depending on fields (Table 3), but on a few occasions it was logisticsand weather. This translatesto 280-560 observed in semideciduous woodlands. individual bird contacts per 4-day period. Forest-dwelling residents, on the other hand, Even at the level of individual species,the re- tended to exhibit a well-defined tendency to oc- sults of point counts and mist-net surveys are cur at higher than expected frequency in medi- not generally related by a constant proportion- um-height or tall semievergreenforest. More than ality factor acrossdifferent habitats. For exam- two-thirds (50 of 70) ofthe most common forest- ple, the Magnolia Warbler, a generalist species associatedresident species were classifiedas either with respectto the successionalmaturity of vege- moderately or markedly specialized on one of tation, occured in 28% of our point counts in the three forest types. The proportion of forest fields and pastures, 36% of the counts in aca- specialistswould have been higher if there had huales,and 57% ofthe points within forest (Table been sufficient censusdata to include the humid 3). The actual number of individual Magnolia forests of the southern Yucatan in the formal Warblers per point count where the specieswas analysis. This region harbors a number of res- observedwas 1.26 (n = 197 occurrencesfor 1982 ident speciesthat do not range north of south- and 1983) a figure which did not vary signifi- ernmost Quintana Roo and Campeche (Paynter cantly with habitat type. This suggeststhat the 1955, Lopez-Omat et al. 1989). The ranges of frequency of occurrence should be a good index most suchspecies are centeredin the wet lowland to the actual abundance ofthis species.Assuming forests of Middle America, and even semiever- this to be the case,Magnolia Warblers were about green forest is probably suboptimal habitat for twice as common in forest as in open fields. How- them; semideciduous forest is altogether avoid- ed. Several nearctic migrants (Wood Thrush, ever, becausethe Magnolia Warbler tends to for- Gray Catbird, Magnolia Warbler, Kentucky age in the canopy of whatever type of vegetation Warbler) occurred with higher frequency in the it inhabits, the species is much less subject to humid evergreenforests of the southern Yucatan capture by nets in tall forest than in low scrub, than in the semievergreenor semideciduousfor- relative capture rates in fields, acahuales, and eststo the north, although larger sample sizesare forest were 5, 5, and 1, respectively (Table 3). In needed to confirm these trends. the absenceof other information, thesemist-net- OVERWINTERING MIGRANTS 533 ting data might causeone to conclude that Mag- munities. Among the few nearctic migrants in nolia Warbler strongly “prefers” disturbed hab- this categorywere the Gray Catbird and Yellow- itats to forest, which is precisely the reverse of breasted Chat. the pattern revealed by point counts. Point counts are also subject to bias, but few An even more extreme example of the inap- specieswere revealed by mist-netting that were propriatenessof mist-netting data as an index of not also detected in point counts. The reverse abundance for canopy speciesis provided by the was not true, and many large speciesand canopy American Redstart. This insect-hawking canopy specialistswere rarely or never netted. Interspe- specialist was detected at 45% of 408 survey cific differencesin detectability exist (e.g., Whit- points (mean number of individuals per occur- comb et al. 198 l), and point counts do not nec- rence = 1.17) in semievergreen forest. However, essarily provide a reliable basis for estimating redstarts very rarely forage close to the forest absolute (as opposed to relative) densities, nor floor, and they were never netted in nearly 6,000 for comparing different communities whose con- net hours of mist netting in closed-canopy forest stituent species differ greatly in detectability (Table 3). On the other hand, the American Red- (DeSante 1981, Vemer 1985, Hutto et al. 1986). start occurs uncommonly in brushy old fields (8% of point counts), and even in milpas (2% of WITHIN-SPECIES HABITAT DIFFERENCES point counts), and individuals that forage in the Although each species has been analyzed as a “canopy” of these low-stature habitats are oc- single entity in the foregoing assessmentof hab- casionally captured by standard 2-m-high mist itat use, at least some migratory species show nets. Capture rates for redstartsin acahuales(CR important age- and sex-related differences in = 2) and in fields (CR = l), though quite low, habitat occurrence during the nonbreeding sea- were higher than in forest (CR = 0), where the son. In perhapsthe best-documented caseof this speciesis one of the most frequently observed sort, Lynch et al. (1985) showed that in the Yu- migrants. catan, overwintering male and female Hooded One category of birds for which mist netting Warblers defend nonoverlapping feeding terri- might be predicted a priori to yield better indices tories that differ dramatically in vegetation struc- of abundance than point counts consistsof small, ture. Males were found to be most common in flocking, ground-foraging species. Among the closed-canopyforest with relatively open under- nearctic migrants in the Yucatan, the most ob- story; females tended to occur in lower, but more vious example is the Indigo Bunting, which had dense,woody vegetation,conditions that are most a capture rate almost four times as high as the often seenin acahualesand young secondaryfor- next most frequently netted migrant in fields and est. Subsequentfield studieshave confirmed this pastures, but which ranked only fourth in the pattern for the Hooded Warbler, and have re- frequency of detection in point counts (Table 3). vealed similar, if less striking, modes of habitat A reasonable interpretation of this result is that separationbased on ageor sex, or both, in several mist netting is more effective than point counts other speciesof small, insectivorous migrants in at tallying absolute numbers of flocking species the Yucatan region (R. Greenberg, A. Lopez- within a designatedarea. However, as R. Green- Omat, and J. F. Lynch, unpubl.). Whether such berg has suggested(pers. comm.), the density of local differentiation might in turn produce re- wide-ranging, nonterritorial, flocking species(e.g., gional gradients in the distribution of sex and age Indigo Bunting) may be substantially overesti- classeshas not yet been determined, although mated by mist netting, because the number of preliminary results suggestthat the sex ratio of flock members that eventually pass through a Hooded Warbler may be biased toward males in given small area (e.g., a grid of mist nets) will be the heavily forested eastern portion of the Yu- higher than the number of equivalent territorial catan (Lynch et al. 1985). individuals that could permanently reside within the same area. In the present study, the only FUTURE PROSPECTS FOR MIGRATORY speciesfor which mist-net data clearly appeared BIRDS IN THE YUCATAN PENINSULA to provide better indices of abundance than point Over the past two millennia, the Yucatan Pen- counts were fairly small, secretive, nonllocking insula has undergone several cycles of wide- taxa that inhabited dense low-stature plant com- spread human disturbance, followed by exten- 534 JAMES F. LYNCH sive regeneration of native vegetation. Although acahuales might reduce the regional population Mexico’s presentrapid pace of population growth of a forest-dependent species such as the Ken- and economic development has thus far spared tucky Warbler by as much as 80-lOO%, depend- some parts of the peninsula, ever-increasing ing mainly on how much the fallow cycle is short- pressuresfor higher production of lumber, live- ened. Once secondary successionhas produced stock, and food crops are threatening remaining a sapling-stageforest, which in the Yucatan re- extensive tracts of tropical forest. Which of the quires 10-15 years (pers. observ.), most forest- many possible scenarios for economic develop- dependent migrants are able to invade a tract ment actually come to pass in the Yucatan will and maintain reasonably high to very high den- have profound consequencesfor migratory birds. sities. However, as the human population in a If, for example, future disturbance of the tropical region increases,there is a tendency for the fallow forest resulting in a mosaic of remnant wood- period to become shortened, sometimes to less lands, small and temporary agricultural plots, than 10 years. The latter pattern has already oc- and regenerating old fields, as has been the case curred in heavily settled portions of Yucatan and historically in much of the area, prospects are Campeche states. relatively good for those migratory speciesthat For most overwintering migratory passerines, can make use of low-stature disturbed or native the single most critical feature of the Yucatan’s vegetation. Examples include species that ac- future landscapewill be the extent to which some tually appear to prefer such habitats (e.g., Com- form of woody vegetation, even if it is succes- mon Yellowthroat, Yellow-breasted Chat, In- sional scrub or young woodland, persists or is digo Bunting, Painted Bunting, Rose-breasted allowed to regenerate. The present study has Grosbeak), as well as ecological generalists that demonstrated that secondary habitats are rou- occur with similar frequencies in both forest and tinely used by many overwintering migrants, in- shrub-field habitats (e.g., Gray Catbird, White- cluding some species (e.g., Least Flycatcher, eyed Vireo, Magnolia Warbler, Northern Parula Hooded Warbler, Magnolia Warbler, Black-and- Warbler). However, this is not to say that some white Warbler, American Redstart) normally re- speciesin the latter category might not undergo stricted as breeders to sizeable tracts of mature substantial density reductions if most mature forest. Despite the fact that winter densities of forest were to be replaced by. brushy second some of these speciesappear to be substantially growth. As an example, if one assumesthat oc- lower in successionalscrub than in mature trop- currence rates in point surveys are proportional ical forest (basedon point-count data), secondary to local densities for a given species, our data vegetation constitutes an important reservoir of suggestthat a landscapeconsisting of milpas and winter habitat for the majority of nearctic mi- acahuales might support only one-half to two- grants in the Yucatan and, one suspects,else- thirds the number of Magnolia Warblers as a where in Middle America. The ability of most pristine forested landscape(Table 3). Still, fairly migratory speciesto utilize other vegetation types generalized migrants such as the Magnolia War- in addition to mature forest may help to explain bler probably would maintain a wintering pop- the otherwise anomalous observation that most ulation in the Yucatan even if the area devoted forest-associatedmigrants do not appear to have to traditional slash-and-burn agriculture were to sufferedglobal declines in abundance during the expand considerably. 1960s and 1970s (Robbins et al. 1986), even in At the other extreme, migrants that rely on (or the face of unprecedenteddestruction of primary at least occur most frequently in) moist, closed- forestin the northern neotropics. However, more canopy forest (e.g., Wood Thrush, Yellow- recent continent-wide population data (Robbins throated Vireo, Kentucky Warbler) would be et al., in press)suggest that these same migratory expected to suffer severe setbacks if wholesale specieshave tended to decline over the last de- destruction or major disturbance of the penin- cade (1978-l 987). Population trends in neotrop- sula’s remaining forestswere to take place, even ical migrants, particularly speciesthat overwin- if some landscape diversity were to be main- ter in tropical forest, should be carefully tained in the form of hedgerowsand successional monitored for evidence of continued decline. woodlots. Again, basing an estimate on point The fact that many forest-associatedmigrants survey data, the change from a forested land- in the Yucatan region also occur in woody second scapeto one covered by milpas and early-stage growth does not, of course, mean that they are OVERWINTERING MIGRANTS 535 immune to the negative effects of tropical de- LITERATURE CITED forestation. Where forest is permanently con- AMERICANORNITHOLOGISTS UNION.’ 1983. Check- verted to overgrazed pasture or intensively list of North American birds. 6th ed. American farmed fields. as is now the casein much of Mid- Ornithologists’ Union, Washington, DC. dle America ‘(e.g., Sader and Joyce 1988) one ASKINS,R. A.,M. PHILBRICK,AND 6. &GENO. 1987. Relationship between the regional abundance of would predict that populations of most species forest and the composition of forest bird com- of migrants will experience local dislocations or munities.Biol. Conserv. 39: 129-l 52. overall population reductions, or both (cf., Rap- AUSTIN. G. T., E. R. BLAKE.P. BRODKORB.M. R. pole and Morton 1985). To the extent that the BR&xN&, W. E. GODF&, J. P. HUBB&D, G. future economic development of Middle Amer- MCCASKIE,J. T. MARSHALL,G. MONSON,S. L. OLSON.H. OUELLET.R. S. PALMER.A. R. PHILLIPS. ica and the northern Caribbean entails the com- W. M.PULICH, M. A. RAMOS,A. M. REA, ANDD: plete suppressionof native woody vegetation in A. ZIMMERMAN. 1981. Ornithology as science. favor of cattle pasture and food crops, species- Auk 98:636-637. wide declines are expected to occur in breeding BEQUAERT,J. C. 1935. Botanicalnotes from Yucatan, p. 505-523. In G. C. Shattuck[ed.], The peninsula populations of migratory birds, particularly in of Yucatan. Carnegie Inst. Wash. Publ. No. 43 1. speciesthat are (1) restricted (or nearly so) to BLONDEL,J. 1975. L’analyse des peuplements d’oi- forest during the nonbreeding season, and (2) seaux; element d’un diagnostic tchantillonnages limited in their winter distributions to Middle frecptentielsprogressifs (E.F.P.). Terre Vie 29:533- America and the Caribbean islands, where de- 589. forestation is proceeding at the most rapid rate. BLONDEL,J. 1977. The diagnosisofbird communities by means of frequential sampling (E.F.P.). Pol. Should complete deforestation occur throughout Ecol. Stud. 3: 19-26. the Yucatan (admittedly an unlikely prospect), BLONDEL,J., C. FERRY,AND B. FROCHOT. 1970. La only those few migratory speciesthat under pris- method des indicesponctuels d ’abundance(I.P.A.) tine conditions were associated with sparsely on des releves d’avifaune par “stations d’ecoute.” vegetatedbeach strandsor savanna (e.g., Yellow- Alauda 38:55-71. BLONDEL,J., C. FERRY,AND B. FROCHOT.198 1. Point rumped Warbler, Palm Warbler, Blue Grosbeak, countswith unlimited distance. Stud. Avian Biol. Indigo Bunting) would be expected to thrive. A 6:414-420. combination of quantitative studies of bird den- BRIGGS,S. A., AND J. H. CRISWELL.1979. Gradual sities in a range of natural and disturbed habitats, silencing of spring in Washington. Atl. Nat. 32: together with a regional assessmentof vegeta- 19-26. DESANTE,D. F. 1981. A field test of the variable tional cover, will be necessaryif we are to track circular-plot censusingtechnique in a California and project the impact of habitat change on in- coastalscrub breeding bird community.Stud. Avi- dividual migratory speciesin sufficientdetail (e.g., an Biol. 6:177-185. Green et al. 1988). DESANTE,D. F. 1986. A field test of the variable circular-plot censusingmethod in a Sierran sub- alpine foresthabitat. Condor 88: 129-142. ACKNOWLEDGMENTS EMLEN,J. T. 1977. Land bird communities of Grand I extend special thanks to Gene Morton, who gener- Bahama Island: The structureand dynamics of an ously sharedmuch of his profound knowledgeof trop- avifauna. Omithol. Monogr. No. 24. American ical birds during the early stagesof this project. I also Ornithologists’ Union, Washington, DC. thank the following people who ably assisted in the FERRY,C. 1974. Comparisonsbetween breedingbird fieldwork in Mexico: Ed Balinsky, Mauro Berlanga, communities in an oak forest and a beech forest, Edgar Cabrera, Gilbert0 Chavez, Carin Chitterling, censused in The IPA method. Acta Omithol. Laurie Greenberg,Russ Greenberg, Jamie Harms, Burt (Warsaw) (English Trans.) 14:302-309. Jones, Elizabeth Ley, Arturo Lopez, Bill Mayher, Jay Ftoans, G. 1952. Geology of northern British Hon- O’Neill, Felipe Sanchez,Martita Van der Voort, Den- duras. Am. Assoc. Pet. Geol. Bull. 36:404409. nis Whigham, and Patricia Zugasty. Alfred0 Careaga GARCIA,E. 1965. Distribuci6n de la precipitation en and Enrique Carillo, former directors of the Centro de la Republica Mexicana. Publicacionesde Geogra- Investigacionesde Quintana Roo (CIQRO), encour- fia de la Universidad National de Mexico, Vol. 1. aaed the researchand facilitated our fieldwork in the GREEN,K. M., J. F. LYNCH, J. SIRCAR,AND L. Z. Syan Ka’an reserve and elsewhere in Quintana Roo. GREENBERG.1988. Use of Landsat remote sens- The manuscriptprofited from critical readingsby Gene ing to assesshabitat for migratory birds in the Morton, Russ Greenberg, and Dick Hutto, and an Yucatan Peninsula. Vida Silvestre Neotropical 1: anonymousreviewer, none of whom necessarilyagree 27-38. with all of my interpretations.Margaret McWethy pre- HATT, R. T., H. I. FISHER,D. A. LANGEBARTEL,AND paredthe illustrations.Generous financial support was G. W. BRAINERD. 1953. Fauna1and archaeolog- provided by the World Wildlife Fund and the Smith- ical researchesin Yucatan caves. Cranbrook Inst. sonian Institution. Sci. Bull. 33. 536 JAMES F. LYNCH

HOBBS,N. T., AND N. BOWDEN. 1982. Confidence servation. Smithsonian Institution Press, Wash- intervals on food preference indices. J. Wildl. ington, DC. Manage. 46:505-507. MYERS,N. 1980. The present statusand future pros- HOLDRIDGE,L. R., W. C. GRENKE,W. H. HATHEWAY, pects of tropical moist forests. Environ. Conserv. T. LIANG.AND J. A. TOSI. 1971. Forest environ- 7:101-l 14. ments in tropical life zones. PergamonPress, New PAYNTER,R. A., JR. 1955. The omithogeographyof York. the Yucatan Peninsula. Peabody Mus. Nat. Hist. HUTTO, R. L., S. M. PLETSCHET,AND P. HENDRICKS. Yale Univ. Bull. 9:1-347. 1986. A fixed-radiuspoint count method for non- POWELL,G.V.N., AND J. H. RAPPOLE. 1986. The breeding and breeding seasonuse. Auk 103:593- Hooded Warbler. Audubon Wildlife Renort 1986: 602. 827-853. KARR, J. R. 1981a. Surveying birds with mist-nets. PYKE, G. H., AND H. F. RECHER. 1985. Estimated Stud. Avian Biol. 6162-67: forest bird densities by variable distance point KARR, J. R. 1981 b. Surveying birds in the tropics. counts. Aust. Wildl. Res. 12:307-319. Stud. Avian Biol. 6:598-553. RAPPOLE,J. H., AND E. S. MORTON. 1985. Effects of KEAST,A., ANDE. S. MORTON[EDS.]. 1980. Migrant habitat alteration on a tropical avian forest com- birds in the neotropics:ecology, behavior, distri- munity, p. 1013-1021. In P. A. Buckley, M. S. bution and conservation. Smithsonian Institution Foster, E. S. Morton, R. S. Ridgely, and F. G. Press, Washington. Buckley [eds.], Neotropical ornithology. Omithol. K~CHER, J. C. 1973. Summer bird diversity in re- Monogr. No. 36. American Ornithologists’ Union, lation to secondarysuccession on the New Jersey Washington, DC. Piedmont. Am. Midl. Nat. 89:121-137. RAPPOLE.J. H.. E. S. MORTON.T. E. Lovmo~. ANDJ. LECK.C. F. 1972. The imuact of some North Amer- L. I&OS. ‘1983. Nearctic avian migrants in the ican migrants at fruiting trees in Panama. Auk 89: neotropics. Fish and Wildlife Service, U.S. Dept. 842-850. Interior, Washington, DC. LEE, J. C. 1980. An ecogeographicanalysis of the REYNOLDS,R. T., J. M. SCOTT,AND R. A. NUSSBAUM. herpetofauna of the Yucatan Peninsula. Univ. 1980. A variable circular-plot method for esti- Kansas Mus. Nat. Hist. Misc. Publ. 67: l-75. mating bird numbers. Condor 82:309-3 13. LOPEZ-ORNAT,A., J. F. LYNCH,AND B. MACK~NNON ROBBINS,C. S. 1980. Effects of forest fragmentation DE MONTES. In press. New and noteworthy rec- on breeding bird populations in the Piedmont of ords of birds from the easternYucatan Peninsula. the Middle-Atlantic region. Atl. Nat. 33:31-36. Wilson Bull. ROBBINS,C. S., D. BYSTRAK,AND P. H. GEISSLER. LYNCH,J. F., E. S. MORTON,AND M. E. VAN DERVOORT. 1986. The breeding bird survey: its first fifteen 1985. Habitat segregationbetween the sexes of years, 1965-1979. U.S. Fish and Wildl. Serv. Res. winteringHooded Warblers( Wilsonia citrina). Auk Publ. 157:1-196. 102:714-721. ROBBINS,C. S., J. R. SAUER,R. S. GREENBERG,AND S. LYNCH,J. F., AND D. F. WHIGHAM. 1984. Effects of DROEGE.In press. Recent declinesin populations forest fragmentation on breeding bird communi- of North American birds that migrate to the trop- ties in Maryland, U.S.A. Biol. Conserv. 28:287- ics. Proc. Natl. Acad. Sci. 324. RUSSELL,S. M. 1964. A distributional study of the LYNCH,J. F., AND R. F. WHITCOMB. 1978. Effects of birds of British Honduras. Omithol. Monogr. No. the insularization of the eastern deciduous forest 1. American Ornithologists’ Union, Washington, on avifaunal diversity and turnover, p. 46 l-489. DC. SADER,S. A., AND A. T. JOYCE. 1988. Deforestation In A. Marmelstein [ed.], Classification,inventory, and analvsisof fish and wildlife habitat. U.S. Fish ratesand trends in Costa Rica. 1940 to 1983. Bio- and Wildlife Serv. OBS 78/76. Washington, DC. tropica 20: 1 l-l 9. SANDERS, H. L. 1968. Marine benthic diversity: a MAY, P. G. 1982. Secondarysuccession and breeding comnarative studv. Am. Nat. 102:243-282. bird community structure: patterns of resource SCOTT, J: M., AND F.-L. RAMSEY. 1981. Length of utilization. Oecologia 5 5:208-2 16. count period as a possible source of bias in esti- MEHLHOP,P., AND J. F. LYNCH. 1986. Bird/habitat mating bird densities. Stud. Avian Biol. 6:409- relationships along a successionalgradient in the 413. Maryland coastalplain. Am. Midl. Nat. 116:225- SIMBERLOFF,D. S. 1972. Propertiesof the rarefaction 239. diversity measurement. Am. Nat. 106:414-418. MIRANDA,F. 1958. Vegetation de la peninsula Yu- SIMBERL&D. S. 1978. Use of rarefaction and re- cateca,p. 215-227. In E. Beltran [ed.], Los recur- lated methods in ecolosv.v. 150-165. In J. Cairns. SOSnaturales de1 sureste y su aprovechamiento. R. J. Livingston, and K. L. Dickson [eds.], Bio: Vol. II. Inst. M&x. Rec. Nat. Ren., Mexico City. logical data in water pollution assessment:quan- MORENO-CASASOLA,P., AND I. ESPEJEL.1986. Clas- titative and statisticalanalyses. Am. Sot. Testing sification and ordination of coastal sand dune and Materials Spec. Tech. Publ. No. 652. vegetation alona the Gulf and Caribbean Sea of SLUD, P. R. 1960. -The birds of Finca “La Selva,” Mexico. VegetaGo 66:147-182. CostaRica: a trovical wet forest localitv. Bull. Am. MORSE,D. H. 1980. Population limitation: breeding Mus. Nat. Hist.~121:49-148. . or wintering grounds?,p. 505-516. In A. Keast TERBORGH,J. W. 1980. The conservation status of and E. S. Morton [eds.], Migrant birds in the neo- neotropical migrants: present and future, p. 21- tropics: ecology, behavior, distribution, and con- 30. In A. Keast and E. S. Morton [eds.], Migrant OVERWINTERING MIGRANTS 537

birds in the neotropics:ecology, behavior, distri- APPENDIX I. Continued. bution, and conservation.Smithsonian Institution Press, Washington, DC. but which almost always contained abundant na- THIEN, L. B., A. S. BRADBURN,AND A. L. WELDEN. tive grasses,herbs, shrubs,and regeneratingtrees. 1982. The woody vegetation of Dzibilchaltun, a In most areasof the Yucatan that were studied, at Maya archeological site in northwest Yucatan, least a few living and dead canopy treeswere pres- Mexico. Mid. Am. Res. Inst. (Tulane Univ.) Oc- ent in open fieldsas scatteredemergents. One phase cas. Pap. No. 5: l-24. was distinguished:Open field (CHT O-2 m). THOMPSON,J.E.S. 1970. Maya history and religion. 2. Brushy old-jield(“acahual~.’ Densesecondary scrub Univ. of Oklahoma Press, Norman. that developswithin 2-3 yearsof abandonment of TRAMER,E. 1974. Proportions of wintering North pastureor cropland. The canopytended to be high- American birds in disturbed and undisturbed dry ly patchy and irregular in stature, but its height tropical habitats. Condor 76460-464. was generally less than 3 m. Scattered emergent VERNER,J. 1985. Assessmentofcounting techniques, clumpsof fast-growingpioneer trees(e.g., Cecropia p. 247-302. In R. F. Johnston [ed.], Current or- obtusifolia) or stump sprouts of previous canopy nithology. Vol. 2. Plenum Press, New York. trees were typically present. Acahualestaller than WAIDE, R. B. 1980. Resource partitioning between about 3 m graded into the young serial stagesof migrant and resident birds: the use of irregular various forest types, depending on the regional resources,p. 337-352. In A. Keast and E. S. Mor- climate and on local edaphicconditions. One phase ton [eds.],Migrant birds in the neotropics:ecology, was distinguished:Brushy old field (CHT l-3 m). behavior, distribution, and conservation. Smith- 3. Native coastal scrub. A heterogeneousnatural as- sonian Institution Press, Washington. semblagethat was restricted to a narrow, discon- WAIDE, R. B., J. T. EMLEN,AND E. J. TRAMER. 1980. tinuous strip along the coastof the peninsula. Ma- Distribution of migrant birds in the Yucatan Pen- jor components were salt-tolerant grasses,herbs, insula: a survey, p. 165-171. In A. Keast and E. and shrubs, together with emergent palms (e.g., S. Morton [eds.], Migrant birds in the neotropics: Coca nuctfera, Thrinax radiata) and stunted in- ecology,behavior, distribution, and conservation. dividuals of certain speciesof treesfrom the neigh- Smithsonian Institution Press, Washington. boring forest communities (e.g.,Bursera simaruba, WEST,R. C. 1964. Surface configuration and asso- Metopium brownet). Substratewas typically loose ciated geology of Middle Am&ica, p. 33-83. In calcareous sand, occasionally with outcrops of R. C. West led.1. Handbook of Middle American limestone. The canopy was low and uneven, and Indians. Vol. 1: ‘Univ. of Texas Press, Austin. was commonly interrupted by areas of grassand WHITCOMB,R. F., C. S. ROBBINS,J. F. LYNCH, B. L. herbs, or by areasof exposed substrate.Along the WHITCOMB,M. K. KLIMKIEWICZ,AND D. BYSTRAK. dry northern coast of the peninsula, the native 1981. Effectsof forest fragmentation on avifauna scrubfeatured emergentcacti (Cactaceaespp.) and of the easterndeciduous forest, p. 125-2 15. In R. agaves(Agave spp.). One phasewas distinguished: L. Burgessand D. M. Sharpe [eds.], Forest island Native coastal scrub (CHT l-3 m). dynamics in man-dominated landscapes.- Sprina- -_ 4. Savanna. A frequently inundated grassland that er-Verlag, New York. typically featured combinations of sawgrass(Cla- WILCOVE.D. S.. AND J. W. TERBORGH.1984. Patterns dium jamaicense) and other grasses,cattail (Typha of population decline in birds. Am. Birds 38:10- dominguensis), and sedges,with scatteredclumps 13. or individualsof emergentshrubs, palms, and trees. WILCOVE,D. S., AND R. F. WHITCOMB. 1983. Gone Extensive savannasflanked freshwaterwetlands in with the trees. Nat. Hist. 92:82-9 1. east-centralQuintana Roo. One phasewas distin- WILLIS, E. 0. 1966. The role of migrant birds at guished:Savanna (CHT l-3 m). swarms of army ants. Living Bird 5: 187-23 1. 5. Seasonally inundatedforest. This low, patchily dis- WILLIS,E. 0. 1980. Ecologicalroles of migratory and tributed foresttype typically occurredat the fringes resident birds on Bar&Colorado Island, Panama, of savannasand within shallow, seasonallyflooded v. 205-225. In A. Keast and E. S. Morton leds.1. depressions.Surface conditions varied from flood- Migrant birds in the neotropics:ecology, behavior; ed to extremely dry, dependingon seasonand year. distribution, and conservation. Smithsonian In- The canopy treestended to be deciduous,and were stitution Press, Washington. dominated by Leguminosae.Characteristic species WILLSON,M. F. 1974. Avian community organiza- included Bucida buceras, Haematoxylon campe- tion and habitat structure.Ecology 55: 1017-1029. chianum, Dalbergia glabra, and Pithecellobium al- bicans. Two phaseswere distinguished:Low (2-3 APPENDIX I. Characteristicsof vegetationtypes that m), Medium-stature (4-10 m). were surveyedin the Yucatan Peninsula.“Phases” are 6. Mangroveforest. A specializedassociation that oc- subcategoriesof a given vegetationtype, based on can- curred in permanently inundated or mucky soils opy height (CHT). Categories are listed in order of in a narrow, discontinuousfringe along the Gulf increasingcanopy height, with drier formations before and Caribbeancoasts. Canopy height and tree den- inundated formations. sity varied markedly in responseto local condi- tions, especially in the most widespread species, red mangrove (Rhizophora mangle). Other char- 1. Agriculturaljields and pastures. Highly disturbed, acteristic trees included black mangrove (Avicen- low-stature communities that were often domi- nia germinans) and buttonwood (Conocarpus erec- nated by exotic cultivated cropsand foragegrasses, 538 JAMES F. LYNCH

APPENDIX I. Continued. APPENDIX I. Continued.

tus). Two phaseswere recognized: Low (l-3 m), 9. Moist forest. Includes the tallest, floristically rich- Medium-stature (4-10 m). est phase of the“Selva Perennifolia” (Miranda I. Semideciduousforest. Included various forms of 1958), which occurred in southern Quintana Roo the “Selva Subcaducifolia” of Miranda (1958). and adjacent Campeche, as well as the tall, essen- Characterized by the presence of 50-75% decid- tially evergreenforests of central Belize, northern uous species, this diverse forest type was wide- Guatemala, and northern Chiapas. Dominant tree spread in areas of Yucatan, Campeche, and west- specieswere similar to those in semievergreenfor- ern Quintana Roo that received 700-1,100 mm est, but they attained greater stature in moist for- of annual rainfall. Characteristic trees included est. Rainfall in areassupporting moist forestranged Bursera simaruba, Lonchocarpusspp., Lysiloma from about 1,200-2,000 mm. Four phases were latisiliqua, Metopium brownei, Senna racemosa, distinguished:Low (CHT 2-3 m), Medium-stature and Vitex gaumeri. Three phases were distin- (CHT 4-10 m), Tall (CHT 1 l-20 m), Very tall guished:Low (CHT 2-3 m), Medium-stature(CHT (CHT 21-30 m). 4-10 m), Tall (CHT 1 l-20 m). 10. Floodplainforest. This associationoccurred in wet 8. Semievergreenforest. Included all but the tallest alluvial soils alongpermanent river systemsin Be- phasesof the “Selva Subperennifolia” of Miranda lize and northern Chiapas. The understory was (1958). In this diverse association,which occurred typically species-poorand was structurally modi- in areas of Quintana Roo and Campeche where fied due to frequent scouringby flood waters, but annual precipitation was 1,100-l ,400 mm, about canopy trees attained considerable stature. Al- 50-75% of the plant specieswere evergreen.Char- though various height classesof floodplain forest acteristic tree speciesincluded Brosimum alicas- occurred in the study region, only one phase was trum, Manilkara zapota, Metopium brownei,Psi- censused:Tall floodplain forest (CHT 1 l-20 m). dium sartorianum, Swietenia macrophylla, and Vitex gaumeri. Three phaseswere distinguished: Low (CHT 2-3 m), Medium-stature (CHT 4-10 m), Tall (CHT 1l-20 m).

APPENDIX II. List of migratory and resident landbirds encounteredin point counts(PC) and mist-net surveys (MN) on the Yucatan Peninsula 1982-1987. Scientific and common names, and the order of families follow the AOU (1983) check-list. Within families, genera are arranged alphabetically, as are common names within genera. List does not include aquatic species,aerial foragersnot seen usingvegetation, speciescasually observed outside of formal point counts, or speciesobserved only during summer. Migratory speciesmarked with an asterisk (*) were transients observed only during fall and/or spring migration.

A. Nearcticmigrants How detected Species PC MN

Falconidae 1. American Kestrel (Falco sparverius) X Trochilidae 2. Ruby-throated Hummingbird (Archilochuscolubris) X Picidae 3. Yellow-bellied Sapsucker(Sphyrapicus varius) X Tyrannidae Tyranninae 4. Least Flycatcher (Empidonax minimus) X X 5. Empidonax sp. X 6. Great Crested Flycatcher (Myiarchus crinitus) X X Muscicapidae Turdinae *7. Gray-cheeked Thrush (Catharus minimus) X 8. Swainson’s Thrush (C. ustulatus) X *9. Veery (C. fuscescens) X 10. Wood Thrush (Hylocichla mustelina) X X OVERWINTERING MIGRANTS 539

APPENDIX II. Continued.

A. Nearcticmigrants How detected Species PC MN

Mimidae 11. Gray Catbird (Dumatella carolinensis) X X Vireonidae * 12. Red-eyed Vireo (Vireo olivaceus) X 13. White-eyed Vireo (V. griseus) X X 14. Yellow-throated Vireo (V. _fZuvifons) X X Emberizidae Parulinae 15. Black-throated Blue Warbler (Dendroicacaerulesct Ins) X 16. Black-throated Green Warbler (D. virens) X 17. Cape May Warbler (D. tigrina) X * 18. Chestnut-sided Warbler (0. pensylvunicu) X 19. Magnolia Warbler (D. magnolia) X 20. Palm Warbler (D. palmarum) X 2 1. Prairie Warbler (D. discolor) 22. Yellow Warbler (D. petechiu)-migratory races X 23. Yellow-throated Warbler (D. dominica) X 24. Yellow-rumped Warbler (D. coronata) X 25. Common Yellowthroat (Geothylypistrichas) X 26. Worm-eating Warbler (Helmitheros vermivorus) X 27. Yellow-breasted Chat (Zcteria virens) X 28. Swainson’s Warbler (Limnothlypisswainsonit) X X 29. Black-and-white Warbler (Mniotilta varia) X X 30. Kentucky Warbler (Oporornisformosus) X X 3 1. Northern Panda Warbler (PuruZuamericana) X X *32. Prothonotary Warbler (Protonotaria citrea) X 33. Louisiana Waterthrush (Seirus motacilla) X 34. Northern Waterthrush (S. noveborucensis) X X 35. Ovenbird (S. aurocapillus) X X 36. American Redstart (Setophaguruticillu) X X 37. Blue-winged Warbler (Vermivora pinus) X X 38. Golden-winged Warbler (I’. chrysopteru) X 39. Nashville Warbler (V. rujicupillu) X *40. TennesseeWarbler (V. peregrina) X 4 1. Hooded Warbler ( Wilsonia citrma) X X 42. Wilson’s Warbler ( W. pusilla) X Thraupinae 43. Summer Tanager (Pirungu rubru) X X Cardinalinae 44. Blue Grosbeak (Guiraca caerulea) X X 45. Indigo Bunting (Pusserinacyaneu) X X 46. Painted Bunting (P. ciris) X X 47. Rose-breastedGrosbeak (Pheucticusludovicianus) X X Icterinae 48. Northern Oriole (Zcterusgulbulu) X 49. Orchard Oriole (I. spurius) X X

Summary: Overwintering species:43 species Detected only in point surveys: 10 species(23.2%) Detected only by mist-netting: 5 species(11.6%) Detected by both methods: 28 species(65.1%) Fall/spring transients: 6 species Grand total: 49 nearctic migratory species 540 JAMES F. LYNCH

APPENDIX II. Continued.

B. Permanent residents How detected

Species PC MN

Tinamidae 1. Thicket Tinamou (Crypturelluscinnamomeus) 2. Great Tinamou (Tinamus major) Ardeidae 3. Cattle Egret (Bubulcusibis) X 4. Green-backed Heron (Butoroidesstriatus) X Acciptridae 5. Bicolored Hawk (Accipiterbicolor) 6. Gray Hawk (Buteo nitidus) 7. Roadside Hawk (B. mugnirostris) 8. Zone-tailed Hawk (B. albonotatus) 9. Common Black Hawk (Buteogullusanthrucinus) 10. Black-shoulderedRite (Elanus leucurns) 11. Crane Hawk (Gerunospizucuerulescens) Falconidae 12. Bat Falcon (F&o rujiguluris) X 13. Collared Forest-Falcon (Micrustur semitorquatus) X X Cracidae 14. Plain Chachalaca(Ortalis vetulu) X 15. Crested Guan (Penelopepurpuruscens) X Phasianidae 16. Black-throated Bobwhite (Colinus nigroguluris) X

17. Blue Ground Dove (C/uravispretiosu) X X 18. Pale-vented Pigeon (Columbu cuyennensis) X 19. Red-billed Pigeon (C. flavirostris) 20. ScaledPigeon (C. speciosu) : 21. Short-billed Pigeon (C. nigrirostris) 22. White-crowned Pigeon (C. leucocephala) Common Ground-Dove (Columbina passerina) ;:. Ruddy Ground-Dove (C. tupacott) 25: Ruddy Quail-Dove (Geotrygontulpucott) 26. Caribbean Dove (Leptotila jumuicensis) 27. Gray-chested Dove (L. cussinit) 28. White-tipped Dove (L. verruuxt) 29. White-winged Dove (Zenaidu asiatica) Psittacidae 30. Red-lored Parrot (Amazona autumnalis) X 3 1. White-fronted Parrot (A. ulbifons) 32. Yellow-lored Parrot (A. xuntholoru) :: X 33. Olive-throated Parakeet (Arutingu nanu) X X Cuculidae 34. Groove-billed Ani (Crotophagasulcirostris) X X 35. Lesser Roadrunner (Geococcyxvelox) X 36. Squirrel Cuckoo (Piuyu cuyunu) X X Strigidae 37. Mottled Owl (Ciccubavirgutu) X 38. FerruginousPygmy-Owl (Glaucidium brusiliunum) X X Trochilidae 39. White-bellied Emerald (Amazifiu cundidu) X X 40. Rufous-tailed Hummingbird (A. tzucutl) X OVERWINTERING MIGRANTS 541

APPENDIX II. Continued.

B. Permanentresidents How detected Species PC MN

4 1. Cinnamon Hummingbird (A. r&la) X X 42. Buff-bellied Hummingbird (A. yucutanensis) X X 43. Green-breasted Mango (Anthracothoraxprevostil] X X 44. Wedge-tailed Sabrewing(Campylopterus curvipennis) X X 45. Fork-tailed Emerald (Chlorostilbon cunivetz] X X 46. Mexican Sheartail (Doricha eliza) X 47. Little Hermit (Phaethornis longuemareus) X 48. Long-tailed Hermit (P. superciliosus) X Trogonidae 49. Citreoline Trogon (Trogon citreolus) X X 50. Collared Trogon (T. collaris) X X 5 1. Slaty-tailed Trogon (T. masse@ X 52. Violaceous Trogon (T. violaceus) X X Momotidae 53. Turquoise-browed Motmot (Eumomotus superciliosa) X 54. Blue-crowned Motmot (Momotus momota) X X Alcedinidae 55. Green Kingfisher (Chlorceryle americana) X X 56. Pygmy Kingfisher (C. aeneu) X X Bucconidae 57. White-necked Pullbird (Bucco macrorhynchus) X Ramphastidae 58. Collared Aracari (Pteroglossus torquatus) X X 59. Keel-billed Toucan (Ramphastos sulfuratus) X X Picidae 60. Pale-billed Woodpecker (Campephilus guatemalensis) X 61. Chestnut-coloredWoodpecker (Celeuscastaneus) X 62. Lineated Woodpecker (Dryocopus lineatus) X 63. Golden-fronted Woodpecker (Melanerpes aurifrons) X X 64. Yucatan Woodpecker (M. pygmaeus) X X Ladder-backed Woodvecker (Picoides scalaris) X X Golden-olive Woodpecker (Pkulus rubiginosis) X X Smokey-brown Woodpecker (Veniliornis jiimigatus) X X Furnariidae 68. Rufous-breastedSpinetail (Synallaxis erythrothorux) X 69. Plain Xenops (Xenops minutus) X X Dendrocolaptidae 70. Ruddy Woodcreeper (Dendrocincla homochroa) X X 7 1. Tawny-winged Woodcreeper (0. anabatina) X X 72. Barred Woodcreeper (Dendroclaptes certhia) X X 73. Olivaceous Woodcreeper (Sittasomus griseicapillus) X X 74. Ivory-billed Woodcreeper (Xiphorhynchusflavigaster) X X Formicariidae 75. Black-faced Antthrush (Formicarius analis) X X 76. Barred Antshrike (Thamnophilus doliatus) X X Tyrannidae Tyranninae 77. Bright-rumped Attila (Attilu spadaceus) X 78. Northern Beardless-Tyrannulet(Camptostoma imberbe) X :: 79. Tropical Pewee (Contopus cinereus) X X 80. Caribbean Elaenia (Elaenia martinica) X X 8 1. Yellow-bellied Elaenia (E. flavogaster) X X 542 JAMES F. LYNCH

APPENDIX II. Continued.

B. Pemtanent residents How detected Soecies PC MN

82. Boat-billed Flycatcher (Megurhynchuspitang@ X X 83. Ochre-bellied Flycatcher (Mionectes oleugineus) X X 84. Dusky-capped Flycatcher (Myiurchus tubercullifer) X X 85. Wied’s Crested Flycatcher (M. tyrunnulus) X X 86. Yucatan Flycatcher (M. yucutunensis) X X 87. Sulfur-rumped Flycatcher (Myiobius sulphureipygius) X 88. Greenish Elaenia (Myiopugis viridicutu) X 89. Social Flycatcher (Myiozetetes similis) X 90. Royal Flycatcher (Onychorhynchus coronutus) X 91. Great Kiskadee (Pita&us sulphuratus) X 92. White-throated Spadebill (Plutvrinchus mvstuceus) X 93. Eye-ringed Flatbill (Rhynchocyclus brevirustris) X 94. Northern Bentbill (Oncostomu cinereigulure) X 95. Common Tody-Flycatcher (Todirostrum cinereum) X 96. Yellow-olive Flycatcher (Tolmomyius sulphurescens) X 97. Tropical Kingbird (Tyrunnus meluncholicus) X Tityrinae 98. Gray-collared Becard (Puchyrumphus major) X 99. Rose-throated Becard (P. ugluiue) X X 100. Rufous Mourner (Rhytipternu holerythru) X 101. Black-crowned Tityra (Tityru inquisitor) X 102. Masked Tityra (T. semifasciutu) X X Pipridae 103. White-collared Manakin (Munacus candet) X 104. Red-capped Manakin (Pipru mentalis) X X 105. Thrushlike Manakin (SchzJornis turdinus) X X Hirundidae 106. Gray-breasted Martin (Progne chulybeu) X Corvidae 107. Brown Jay (Cyunocorux morio) X 108. Green Jay (C. yncus) X 109. Yucatan Jay (C. yucutunicus) X Troglodytidae 110. Band-backed Wren (Cumpylorhynchus zonutus) X 111. White-breasted Wood-Wren (Henicorhinu leucostictu) X 112. Carolina Wren (Thryothorus 1udovicianu.s) X X 113. Spot-breastedWren (T. mucuhpectus) X X 114. House Wren (Troglodytes uedon) X X 115. White-bellied Wren (Uropsilu leucogustru) X X Muscicapidae Sylviinae 116. Blue-gray Gnatcatcher (Polioptilu cueruleu) X X 117. Tropical Gnatcatcher (P. plumbeu) X 118. White-lored Gnatcatcher (P. ulbiloris) X 119. Long-billed Gnatwren (Rumphocuenius melunurus) X X Turdinae - 120. Clay-colored Robin (Turdus gruyiz) X X Mimidae 121. Black Catbird (Melunoptilu glubirostris) X X 122. Tropical Mockingbird (Mimus gilvns) X X 123. Cozumel Thrasher (Toxostomu guttutum) X OVERWINTERING MIGRANTS 543

APPENDIX II. Continued.

B. Permanent residents How detected Species PC MN

Vireonidae Vireoninae 124. Lesser Greenlet (Hylophilus decurtatus) X X 125. Tawny-crowned Greenlet (H. ochraceiceps) X X 126. Cozumel Vireo (Vireo bairdi) X 127. Mangrove Vireo (V. pallens) X X 128. Yucatan Vireo (V. magister) X X \ -I Cyclarhinae 129. Rufous-browed Peppershrike(Cyclarhis gujanensis) X X Emberizidae Parulinae 130. Golden-browed Warbler (Busileuterus bellz) X 13 1. Yellow Warbler (Dendroica netechial-resident race X 132. Gray-crowned Ykllowthroat(Geothl$pis poliocephala) X 133. Gray-throated Chat (Grunatellus salluez) X Coeribinae 134. Bananaquit (Coerebu jhzveola) X Thraupinae 135. Red-legged Honeycreeper (Cyunerpes cyuneus) 136. Gray-crowned Tanager (Eucometis penicilluta) X 137. Olive-backed Euphonia (Euphoniu gouldt) 138. Scrub Euphonia (E. afinis) X 139. Yellow-throated Euphonia (E. hirundinacea) X 140. Red-crowned Ant-Tanager (Hubiu rubicu) X 141. Red-throated Ant-Tanager (H. fuscicauda) X 142. Black-headedShrike-Tanager (Lanio uuruntius) 143. Rose-throated Tanager (Pirangu roseogularis) X 144. Crimson-collared Tanager (Rumphocelus sunuginolentus) 145. Scarlet-rumped Tanager (R. pusseriniz) 146. Stripe-headed Tanager (Spindulis zena) X 147. Yellow-winged Tanager (Thraupis ubbus) . Cardinalmae 148. Northern Cardinal (Cardinalis cardinalis) X X 149. Black-faced Grosbeak (Curyothruustes pohogaster) X 150. Blue Bunting (Cyanocumpsu parellinu) X X 151. Black-headed Saltator (Saltutor atriceps) X X 152. Buff-throated Saltator (S. maximus) X 153. Grayish Saltator (S. coerulescens) X Emberizinae 154. Orange-billed Sparrow (Arremon aurantiirostris) X 155. Green-backed Sparrow (Arremonops chloronotus) X X 156. Olive Sparrow (A. rufivirgatus) X X 157. White-collared Seedeater(SporovhiZa toraueola) X X 158. Yellow-faced Grassquit (T&-is dhvucea) _ ’ X X 159. Blue-black Grassquit ( VoZatiniajucurina) X X Icterinae 160. Yellow-billed Cacique (Amblycercus holosericeus) X X 161. Melodious Blackbird (Dives dives) X X 162. Altamira Oriole (Zcterus gularis) X X 163. Black-cowled Oriole (I. dominicensis) X X 164. Hooded Oriole (I. cucullatus) X X 165. Orange Oriole (I. aura&s) X X 166. Yellow-backed Oriole (I. chrysater) X X 167. Yellow-tailed Oriole (I. mesomelas) X X 168. Montezuma Oropendula (Psarocoilius montezuma) X 169. Great-tailed Grackle (Quiscalus mexicunus) X 544 JAMES F. LYNCH

APPENDIX II. Continued.

Summary: Detected only in point surveys:6 1 species(note: 2 l/6 1 do not occur in region of central Quintana Roo that was netted; 40 specieswere seen in central Quintana Roo, but not netted there) Detected only by mist netting: 2 species Detected by both methods: 106 species Grand total: 169 resident species

MIST NETS NET POLES CALIPERS SPRING SCALES

We are supplyingmist nets and other bird banding equipmentby mail order.

For information,write: Avinet, Inc. 12021Wilshire Blvd. #6&l Los Angeles,California 9002.5 (213)396-6387-FAX (213) 3964697