The Condor 92444-360 0 The Cooper Ornithological Society 1990

POPULATION DYNAMICS OF THE WOOD IN SOUTHERN VERACRUZ, MEXICO ’

Bell Museum of Natural History, Department of Ecology and BehavioralBiology. Universityof Minnesota, 10 ChurchStreet, SE, Minneapolis, MN 55455

JOHNH. RAPPOLE Conservationand ResearchCenter, National ZoologicalPark, Front Royal, VA 22630

MAFUOA.RAMOS World Wildrif Fund, US, 1250 24th Street NW, WashingtonDC 20037

Abstract. The Wood Thrush Hylocichla mustelina(Muscicapidae: Turdinae) was studied on wintering grounds in the rainforest of southern Veracruz, Mexico, using mist netting, radiotelemetry, and field observation. Data were collected during three boreal winters (pri- marily November-March), and 46 individuals received transmitters. All sex and age classes were present, but difficulties in aging and sexing did not allow accurate determination of proportions. Telemetry showed two types of spatial use: “sedentary” (n = 18) and “wan- dering” (n = 21). Wanderers moved greater relative distances,had higher subcutaneousfat reserves, and seemed to incur greater mortality then sedentary . Analysis of eight specimens suggeststhat these two strategiesare practiced by both sexes. Six birds that receivedtransmitters remained unclassifieddue to insufficientdata. One isolatedbird showed intermediate behavior. Netting revealed a lack of sedentarinessin part of the population throughoutthe winter. Movements were correlatedwith weather changes:cold, wet weather systemsbrought an influx of transientsinto the lowland forest site. Vocalization and playback data imply that during these weather systems aggressionsubsided. Home-range size of sedentarybirds averaged 0.44 ha (0.12-1.03 ha). Some wanderers seemed to make move- ments of >2 km. Wanderers were estimated to constitute approximately 50% of the pop- ulation. Becausewanderers apparently occupy which undergo intolerable fluctua- tions in suitability and appear to sufferhigher mortality,we tentatively acceptBrown ’s (1969) critical density level 3 as a valid descriptor of wintering Wood Thrush densities in this region. Key words: Nearctic migrant; neotropics;rainforest; territoriality; nonbreedingecology; radiotelemetry.

INTRODUCTION nearctic wintering in the neotropics. The rate of conversion of natural tropical habi- One reason for this deficiency may be the short- tats in the neotropics to agriculture is alarming age of studies in which individuals have been (Myers 1980a, 1980b, 198 1; Caufield 1984), and marked and fo11owed. concern has been expressedover what effectsthis The importance of studying individuals when loss may have upon both resident and investigating population-level phenomena has migrant speciesthat utilize these resources recently been emphasized(Lomnicki 1978,1988; (Vogt 1970, Terborgh 1980). Discussion over the Hassell 1986), and we concur with this advice. potential effectsof tropical habitat loss on nearc- The study of individuals is easier when move- tic migrants that use these resourcesduring the ments are relatively restricted, making speciesin boreal winter (e.g., Morse 1980, Holmes and which individuals hold winter territories attrac- Sherry 1988) has been hampered by a lack of tive subjects for study. Nonbreeding, insectivo- relevant data from nonbreeding populations. rous, passerinemigrants are frequently territorial Thus far we lack even the demonstration of a (see Rappole et al. 1983, p. 22-25 for a review; mechanism for the winter-limitation of any also Elgood et al. 1966, Price 1981) and terri- toriality has been suggestedas the most common social systemamong wintering nearctic migrants. I Received 15 September 1989. Final acceptance15 (Greenberg 1986). December 1989. The hypothesis that tropical wintering habitat

[4441 WINTERING WOOD THRUSHES IN VERACRUZ 445

loss may have a negative effect upon wintering ber through March. These systems,called nortes, populations of nearctic avian migrants would usually last 2-6 days. Severalnormally occureach postulate that these migrants are experiencing month. The rest of the year is divided into the conditions of high density relative to the carrying dry season, late March through May, and the capacity of the most favorable part of their win- rainy season,early June to early November. tering environments. When examined in the con- Annual temperature means range from 18- text of Brown’s (1969) critical density model of 26°C depending mostly on altitude. The highest the role of territoriality in population regulation, peaks in the range, Volcan San Martin and Vol- this postulation results in two predictions about can Santa Martha, are more than 1,500 m above a population of territorial wintering migrants: (1) sealevel. Temperatures at the lowland study site habitats of decreasinglevels of suitability should (elevation 150 m) during the 1984-1985 field be occupied by individuals of that population, seasonaveraged 2 1°C from 11 December through and (2) a class of nonterritorials should exist. 20 March, with a low of 12°C (12 December This study examines these predictions in a win- 1984, 7 January 1985) and a high of 31°C (31 tering population of Wood Thrushes (Hylocichla January 1985). For a more detailed summary of mustelina), through exploration of movements, the climatology and geography of Los Tuxtlas, and, indirectly, behavior at the individual and see Andrle (1967), Lot-Helgueras (1976) and population levels. Soto-Esparza (1976). The Wood Thrush breeds primarily in the The original vegetation of much of Los Tuxtlas United Stateseast of Texas, Oklahoma, Kansas, was the “selva alta perennifolia” (tall, evergreen, Nebraska, and the Dakotas, and winters largely tropical forest)of Miranda and Hemandez (1963) on the Caribbean slope from eastern and south- and Pennington and Sarukhan (1968). Los Tux- em Mexico through Central America to Panama tlas marks the northernmost extent of such hab- and northwestern Colombia (AOU 1983). Dur- itat in the Western Hemisphere. Other foresttypes ing the breeding seasonit inhabits mesic decid- existedor are still presentin the region (seeAndrle uous forests. It winters primarily in moist low- 1967, p. 168), but our work was conducted in land forests.This specieswas chosen for several selva, or rainforest. Deforestation has caused reasons:(1) it is a near&c migrant whose entire profound habitat changein Los Tuxtlas, and most winter range is undergoing extreme habitat loss; of the area is now grasslands, pastures, citrus (2) individuals are territorial on the wintering groves, and weedy fields, broken by tree rows grounds(Willis 1966, Rappole and Warner 1980); and small forest remnants. Andrle (1966) esti- (3) its preferred habitat is rainforest (Rappole mated that one-half of the Sierra de Los Tuxtlas and Warner 1980; Rappole and Morton 1985; was still forested in 1962. Rappole and Warner D. W. Warner, unpubl. data); (4) it is a common (1980) estimated that less than one-third re- and easily captured speciesin southern Veracruz mained forested in 1975. Using recent Landsat (Rappole and Warner 1980, Ramos 1989); and imagery and air photos, we estimated that less (5) it is large enough to bear a 2-g radio trans- than 15% was forested in February 1986. Very mitter. little lowland forest remains in this region. The La Peninsula study site lies on an edge of STUDY AREA the largest remaining tract of selva left in Los The Sierra de Los Tuxtlas, about 42,000 km* in Tuxtlas (<60,000 ha), in the steep valley of the size, is a rugged,mountainous region of volcanic Rio de Coxcoapan, a fast-flowing, rocky moun- origin situated on the Gulf of Mexico near the tain river, near the small village of Ejido La Pen- Isthmus of Tehuantepec, 90 km southeast of la insula de Moreno (Fig. 1). Selva at this site is Ciudad de Veracruz (Fig. 1). Average rainfall characteristic of the plant community at an el- varies annually from 2,500 mm-4;500 mm, de- evation of 150 m, having a canopy height of 20- pending on location. The orographiceffect of Los 30 m (with emergents to >40 m). Topography Tuxtlas on air masses moving inland from the is varied (minor valleys and ridges),with streams gulf causeshigher rainfall on the immediate gulf and treefall gapscommon. Additional netting was slope than on the surrounding lowlands (Gold- conducted in March 1985 on the Volcan Santa man 195 1, p. 269). This effect is noticeable dur- Martha in virgin forest beside the Rio de Su- ing the periodic cold, wet weather systemsthat chiapan, at an altitude of about 470 m. Fieldwork move into the area from the north from Novem- was conducted during the periods of November 446 K. WINKER, J. H. RAPPOLE AM) M. A. RAMOS

A study site ------road/trail - FIGURE 1. Location of study sites. Site A is the La Peninsula site; site B is the Santa Martha site. Enclosed area around the sites indicates current extent of relatively undisturbed forest. An ejido is a small agricultural community. WINTERING WOOD THRUSHES IN VERACRUZ 441

1982, January-March 1983, October 1983-Jan- ers. Two transmitters fell off after 2-3 days due uary 1984, late March-early April 1984, and No- to failure of year-old (i.e., not fresh) glue. vember 1984-April 1985. Transmitters were attached and birds released as quickly as possible at the point of capture; MATERIALS AND METHODS usual elapsedtime was l-l .5 hr. Only birds that Study sites were mapped and gridded, and num- appeared to be in good condition were selected bered flags were placed at 25-m intervals to fa- to bear transmitters. Radio-tagged birds were lo- cilitate plotting ofprecise locations both for radio cated at least twice a day (morning and evening) locations and resightings.Nylon mist nets (12 x when possible, using a four-element Yagi anten- 2.6 m) were set at alternate grid points (50 m na and a 12-channel receiver (Custom Teleme- apart). Net spacingwas chosen based on the es- try). Tracking was done on foot. Both triangu- timated OS-ha averageterritory size of the Wood lation and approaching a bird until it moved Thrush found by Rappole and Warner (1980, p. were usedto pinpoint locations.Home-range sizes 363). On the La Peninsula site, 350 m x 250 m were estimated from these locations using the (8.75 ha), 48 nets were set, oriented perpendic- modified minimum area method of Harvey and ular to the valley (approximately 50”). Nets were Barbour (1965). The few points excluded by dis- openedwhen weatherpermitted, and over 3 1,000 tance from inclusion in home-range size esti- net hr were accumulated. The Santa Martha site mation were classified as extra-home-range for- (2.5 ha) was netted from 9-19 March 1985 with ays, or “sallies outside the area” (seeHarvey and 16 nets (830 net hr). Barbour 1965, p. 400; Burt 1943). Each Wood Thrush captured received a unique During the 1984-l 985 season,the date, time, colored leg-band combination and an aluminum type, and duration of every vocalization was not- U.S. Fish and Wildlife Service band. Date, time, ed. Additional observations were made by arti- net number, age(when possible),molt condition, ficially stimulating presumably territorial indi- and subcutaneous lipid deposits were recorded viduals using a taped playback of calls used in for all birds. Lipid levels were estimated using agonisticinteractions. Cassettetape sessionswere the method of Helms and Drury (1960). Age was conductedat various times of day, both with and determined by skull ossification through mid- without the presentation of a mounted conspe- December, after which it became unreliable due cific model. The tape contained a repeated mod- to shrinking of the unossihed regions. Exami- erate to high intensity aggressivevocal confron- nation of 85 museum specimens of known win- tation between two individuals and was run for tering birds taken from the area during the win- 12-15 min during each session, for a total of ters of 1973-1974 and 1974-1975 (see Ramos approximately 13 hr. 1983) showed that aging by the presence or ab- sence of wing bars or wing spots is inaccurate RESULTS (Appendix I). Birds were releasedat the banding Despite being reported as uncommon in Vera- laboratory (on site) unless fitted with a trans- cruz (Loetscher 1955, p. 36), and rare in Mexico mitter, in which case they were released at the (Edwards 1972, p. 190) the Wood Thrush was point of capture. the dominant understory forest species in this During the study, 46 Wood Thrushes were fit- winter avian community, as indicated by mist- ted with radio transmitters (average weight 2.0 net captures(unpubl. data). During the study 2 17 g, Custom Telemetry, 185 Long View Drive, Wood Thrushes were caught on the La Peninsula Athens, GA 30605). Expectedtransmitter life was site. Many of these birds were recaptured or re- 14 days. Transmitters were attached between 14 sighted periodically throughout the winter sea- December and 15 February (all years). The at- son, and several from one year to the next (Table tachment technique used was essentially that of 1). A disparity between captures and recaptures Raim (1978), which involves the gluing of a pri- per 1,000 net hr for the 1982-1983 field season mary cloth with artificial eyelash adhesive to the and the 1984-1985 and 1985-1986 seasonswas back of the bird, then the gluing of a secondary due to an intensive November 1982 netting ef- cloth, with transmitter attached, to the primary fort, during which migrants were still passing cloth. Attachment usually failed after approxi- through. Wood Thrushes returned to Los Tuxtlas mately 40 days due to the regrowth of back feath- from mid-October to mid-November, and de- 448 K. WINKER, J. H. RAPPOLE AND M. A. RAMOS

TABLE 1. Wood Thrush nettingresults from La Peninsula.

Previous season captur& ,000 Recap~m;l,OOO Season Captures Recaptures recaptures 1982-1983 81 80 0 10.54 10.42 1983-1984 40 14 3 5.76 2.02 1984-1985 96 42 3 6.59 2.88

parted around mid-April. Early and late dates slowly, although occasionally in heavy rain wan- were 16 October 1983 and 10 May 1985, al- derers behaved similarly. though Ramos (1983, p. 20) recordedboth slight- Sedentaries.The movements of sedentarybirds ly earlier and later dates in the region. were much more restricted, both in rate and dis- Because accurate determinations of age and tance, than those of wanderers (Table 3). These sex were not possible on live birds throughout restricted movements were reflected in a 78% the winter, we determined the age and sex struc- recapturerate among sedentariesreceiving trans- ture of the Wood Thrush population wintering mitters. The mean length of within-season cap- in Los Tuxtlas by examining museum specimens ture history for recaptured sedentarieswas 38.6 taken from the area during the winters of 1973- days (first to last capture, Table 2). Since loca- 1974 and 1974-l 975 (mid-December-late tions were only taken twice per day for most March). Differences in age and sex proportions birds, the movement rates in Tables 2 and 3 were between these years were not significant, so the not accurate per-day figures; they were used as samples from both winters were pooled (age: z an index of movement based upon known lo- = 0.599; P = 0.27; sex: z = 0.212; P = 0.29; cations. Wyatt and Bridges’ [ 19671 test for differences The behaviors of well-known sedentaries(e.g., between population proportions). Based on this 213,466[‘85], 1645,1747,1832,1841,and 1645 sample, the population during those years was and 1778 when in residence) and the behaviors composedof 72% males and 28% females. Using ofwell-known nonterritorials, or wanderers(e.g., only specimens collected earlier than mid-Feb- 1235,1773,1779,1780,1793,1807, and 1856), ruary (to eliminate the possibility of including were used as “templates” against which to com- fully ossifiedfirst-year birds), a more reliable age pare other birds with fewer data. Not all birds ratio was obtained-60% adults: 40% juveniles tracked were classifiable as either sedentary or (Appendix I). wanderer. Of the 46 birds which received trans- mitters, 18 were classedas sedentary, 2 1 as wan- INDIVIDUAL MOVEMENTS derer, and six were unclassifieddue to insufficient Radio-tracking results demonstrated two gen- data (Table 2). The single bird from the Santa eral, but distinct, types of movement among in- Martha site was also unclassified. dividual Wood Thrushes in Los Tuxtlas. We will Figure 2 depictsthe movements of a sedentary use “sedentary” and “wanderer” as labels for individual. Number 466 was present on the site thesetwo behavioral categories.Movement alone in all three winters of the study, and was fitted was not always an accurate predictor of a bird’s with a transmitter in both the first and last years. disposition, however, and several criteria were The chronological sequence of locations shows used to categorize each individual. Site fidelity that daily movements encompasseda large por- and movement distances,patterns, and rateswere tion of what was eventually determined to be the found to be good separators,but capture records, home range. There was a great degreeof overlap agonistic behavior, and behavior when ap- in the two areas of movement, but the third-year proached and flushed were also used. Table 2 home-range center was 15 m away from the first- shows the individualistic nature of birds fitted year center, suggestingthat movement patterns with transmitters and the various criteria used had changed,and that perhaps home-range and/ in categorizing each bird as “sedentary,” “wan- or territory boundaries are not entirely stable derer,” or unknown. Sedentary individuals gen- between years. erally allowed a much closer approach by the Home-range sizes for sedentary birds with tracker, and when flushed tended to move off transmitters varied from 0.12 to 1.03 ha, and WINTERING WOOD THRUSHES IN VERACRUZ 449 averaged 0.44 ha. The bird holding the 0.12-ha upriver coincided with still another. The length home range (no. 42 1) was found out of this area of time no. 1793 stayed in the riverine disturbed only once. For comparison, no. 1743 had a home area and the small spaceit used while there were range of 0.4 ha, yet was found out of this area not typical. It is possible that the bird was de- four times. This may reflect differencesin home- fending this space as a territory, but no change range quality: much of no. 1743’s home range was in behavior was observed, and, perhaps most rocky stream-edgeand steepbank. Only two oth- suggestive,its flushing disposition did not change. er extra-home-range forays were recordedamong These movements along the river may represent all birds with transmitters. an attempt by a nonterritorial bird to “fit” itself Two sedentary birds showed movements of a into the territorial mosaic of the area by estab- kind rarely describedin wintering territorial sys- lishing residence on an edge. The marginality of tems. Each left its home range for several days the site it utilized along the river was twofold: it (5 and IO), then returned again. Number 1645 was young second growth (not forest), and was took up residencein its home range upon return; frequently flooded. The bird’s eventual depar- no. 1778 (adult male) remained on its former ture coincided with the norte-causedflooding of range for 4 days, then departed again. It returned part of the small area it had been utilizing. again 15 days later and was collected. Shortly Wanderers included four individuals that dis- after this, another bird (or perhapstwo) took over appeared almost immediately after transmitter the home range. The whereabouts of nos. 1645 attachment. These birds were similar to other and 1778 while gone was unknown, but they were birds subsequently classed as wanderers: they at least 2 km away. The initial return of these were newly captured birds, and their captures two birds to the site coincided with the beginning occurredat the end of nortes.Transmitter failure of a norte, the same one that presumably caused on all of these birds was considered improbable, a shift out of the riverine disturbed habitat by for in each casethe transmitter was activated on wanderer no. 1793 (see below). the day of attachment and was functioning per- Wanderers. Within the wanderer group, pat- fectly before release. We interpreted these signal terns were not consistent. Two very general types disappearancesas large-scalemovements out of of movement emerged. In one type (Type 1, the Coxcoapan valley. Including these four, a Movers; nos. 555, 1773, 1779, 1780) repre- total of nine disappearanceswas recordedamong sented by Figure 3 (no. 1773) movements were the wanderers; three of these birds (nos. 1773, largely random when near the study site, but those 1780, 1856) returned to the valley following dis- birds for which a final, distant signal was received appearance, but none returned to the area uti- (prior to total disappearance) were moving lized prior to departure. Home-range sizes were up the valley toward higher elevations and the not estimated for wanderers becauseof their ap- interior of the forest tract. parent transience. The second general type of movement found The higher movement rate of wanderers re- among the wanderers(Type 2, Drifters; nos. 527, sulted in a lower averagenumber of tracking days 1182, 1235, 1793, 1807, 1810, 1856) is repre- per individual than among sedentary birds (Ta- sentedby Figure 4 (no. 1793). These birds’ move- ble 3). These high rates of movement were prob- ments were relatively localized comparedwitb the ably also responsible for a relatively low recap- wide-ranging movements of the mover group. ture rate (24%) and a short mean length of capture Movement from the La Peninsula site to a nearby history (6 days) among transmitter-bearing wan- (ca. 150 m) disturbed riverine habitat and sub- derers which were recaptured (first to last cap- sequent departure from the area was shown by ture, Table 2). These figuresreflect the tendency six of the Type 2 wanderers. of wanderers to leave our study site, and often, Movements of all wanderers appeared to be apparently, to leave the Coxcoapan valley. Once correlated with adverse weather. Number 1793 out of the valley, signals were never received, was initially captured at the beginning of a norte, although checkswere made daily from the high- and its recapture 4 days later (transmitter at- est available points, and periodically from the tached) coincided with the end of this norte. Its valley rim. These birds had to move at least 2 subsequentmovement from the small area it uti- km to cause signal disappearance due to topog- lized in the riverine habitat coincided with the raphy (apparatus was known to be suitable for onset of the next norte, and its disappearance at least 3-4 km line-of-sight), and were probably 450 K. WINKER, J. H. RAPPOLE AND M. A. RAMOS

TABLE 2. Vital statisticson Wood Thrushesreceiving transmitters during the study.Transmitters were attached between 14 December and 16 February (all years). Groupingswere based upon movement pattern (not shown), averagedistance moved per day, capture records, and behavioral disposition (see text for details).

Approachesto a25 m First to Home- Aqres- Submis- Dayst last AWL?ge range swe OCCW- sincefirst capture Bird number Trackingdays m/day size (ha) display” displayb‘lve rences Attempts capture (days) A. Sedentaries,La Peninsula 165 (1983) 9.5 ca. 80.0 - - 3 82 733 213 12.5 49.6 0,; - - 7 ii 84 84 390 9.0* 21.1 0.20 - - 6 6 69 75 421 11.5 34.8 0.12 - - 5 5 9 9 466 (1983) 6.5* -d 0.20 - - 0 737 466 (1985) 21.5 40.5 0.20 2 - 24 24 737 737 548 * 26.7 0.30 - - 25 1107 :::* 33.1 0.35 - - : : 3: 38 1645e 32.5 (23) 54.0 0.50 - - 24 26 21 57 1694 6.5 49.2 0.25 - - 10 10 3 1736 6.0 63.8 0.40 1 - 10 12 : 1743’ 13.5 99.3 0.80 - - 3 7 : 0 1747 34.5 41.9 0.58 - - 7 8 0 177ti 42.5 (24) 52.7 1.03 ? - 15 17 : 43 1794 8.5 64.7 0.35 - - 7 9 0 0 1832 24.5 45.2 0.45 1 - 12 14 12 12 1841 24.5 72.4 0.65 2? - 17 20 0 4 1870 2.5d 85.6 0.20 - - 6 6 22 22 B. Wanderers, La Peninsula. 460 1.0 388.0 - - - 0 0 526 1.5 104.0 - - - 1 527 :.: * 62.9 - - - :, 1 555 - - - 0 1182 6:0* 51:; - - - : 1235 13.5* 41.8 - - - :, 7 1727’ 0.0 - - - - - 1728’ 1.5 - - - - ; 1 1741j 1.5 - - - - 0 2 1773* 5.5 (4.5) 421.9 - - - 1774 0.5 - - - - :, zl 103.5 - - - 2 17791780k ::: (5) 276.6 - - - : 1 1793 19.5 164.3 - - - 1 1807 4.5 235.6 - - 1 : 1810 2.5 116.0 - - - : 2 1849 176.0 - - - 1 1 1850’ ------1856k 16.0 (11) 178.7 - - - 1 9 1872’ 0.0 ------1890’ 0.0 ------C. Unclassified birds, including single Santa Martha bird. 077’ 5.0* 25.0 ? - - 3 3 81 81 165 (1984)m 1.0 50.0 ? - - 2 2 402 733 113@ 3.0* 15.7 ? - - 2 2 29 490 12040 ------0 1219~ a0 ------: 1238q 4:0* 126.3 - - - : SM52‘ 4.5 131.0 - - - : : 0 8 ;t TFsmitter attachedx daysafter bird’s initial capture. Bird locatedonly onceper day. sGiven either in a natural encounteror in responseto tapedplayback. b Immediate retreatin responseto aggression. r Readingsmixed up with a Habia gutturalis bearing a transmitter,making movementestimate approximate and accuratehome-range size estimate tmposslble. WINTERING WOOD THRUSHES IN VERACRUZ 451

TABLE 3. Comparisonsbetween transmitter-bearing Wood Thrush sedentariesand wanderers.

Sedentaries Wanderers

Number of individuals 18 Total tracking days 280.0 GA.0 Mean tracking days (SD) per bird 15.6 (11.6) 4.7 (5.7) Mean distance (SD) moved per bird per day (my 53.8 (21.5) 178.5 (122.3) Deaths 0 2 *An index of movementbased on a maximum of two radio locationsper day. Sedentaries:n = 17; Wanderers:n = 13. more distant, since valley-rim checks were un- (1 male: no. 1778, 1 female: no. 1130), and two successful.No wanderers were recaptured from were shot when responding aggressivelyto play- one winter to the next. back of aggressive vocalizations (2 females). During the study, two wanderersbearing trans- Wanderers in this sample were no. 1856 (male), mitters were taken by predators, one by an un- and one probable wanderer, an unbanded female known predator and the other by a raptor. Be- collected after a norte. cause the sample periods within the two groups Subcutaneouslipids, Subcutaneous fat depos- (tracking days) were disparate, the effective “win- its upon initial capture were higher in wanderers dow” into the lives of sedentarieswas much long- than in sedentaries (z = 5.49, P < 0.001, Wil- er than for individual wanderers; we had nearly coxon’s rank sum test; Fig. 5). three times the sample period in the sedentary SantaMartha site.Only one bird was captured group, yet not one sedentarybird died while being on the Santa Martha site, and this individual tracked.This trend was substantiatedduring three received a transmitter. Its movements were dis- subsequent years of data collection (Rappole et similar from those of any other bird tracked. It al. 1989; summarized in Discussion). moved about on the valley floor loosely, gener- Although differencesin movement rates made ally using the flatter terrain along the river. The any quantitative assessmentdifficult, and there- area it covered was long (260 m) and narrow, fore crude, we estimated that wanderers consti- unlike the home range of any sedentary mapped, tuted approximately 50% of the Wood Thrush but its flushing disposition matched the seden- population at La Peninsula. tary type, rather than that of wanderers (Table Collectedbirds. Six birds successfullycollected 2~). Without conspecificsin the area its move- to determine sex (two additional birds had go- ments were probably not constrained. It is note- nads destroyedby shot), showedthat sedentaries worthy that one Wood Thrush was seen on this comprised birds of each sex (3 females, 1 male), site the day before a nortecame. Netting com- and wanderers included one male and probably menced during this norte,and no Wood Thrush- one female. Sedentariesin this sample consisted es were seen, heard, or captured for the duration of two birds with radio and recapture histories of this weather system (3 days). On the day fol- t dTransmitter problems:broken antenna (466); experimental0.7-g transmitter(1870); home-rangesizes estimated from locations,resighting, and recapturedata. eBird left area,later returned.Tracking daysin parenthesesand averagedistance moved are values for residentperiod only. Seetext. ‘Newcomer which establisheda home rangenear studysite. ’ ’ Vocalizationheard from vicinit of radio-taggedbird. 8Only detectionswere 0.45 km fymm ate. Movement betweenlocations inestimable due to terrain. hFirst transmitterfell off after 3 days;new transmitterattached upon recapture,3 dayslater signaldisappeared. ’ ’ Signal (and thus bird) disappearedshortly after transmitterattachment. Functmnality of transmittercertain upon attachment. JBird ’s only movementswere due to long distanceflushes (by tracker)prior to disappearance. ’ Bird let?valley (signaldisappearance), later returned.Tracking days in parenthesesand averagedistance moved are for residentperiod only. ’ ’ Movementslocalized, but locatedon only 2 daysin a 5-&y period. mSedentary by recapturerecord; problems with shortmgout transmitterin heavy rain. Tracking recordalone considered insufficient evidence for categorization.Before transmitter failure, no. 165 wasfound out of its home ran e, whereit showedwanderer-type movement (moving rapidly and keepingwell away from tracker)until returningto its home mage,when it settled down againin its movements(moving slowly and circularlywhen pushed). L1Sedentary by recapturerecord, transmitter fell off, however. DTransmitter found fallen off after 2 days,its position making the previousday ’s locationuncertain. pFunctionality of transmitterupon releaseuncertain. 0 Signal behavedas thoughantenna had broken. ’ ’ Individual from SantaMartha. Although the other birds in Table 2C are unclassifiabledue to insufficienttelemetry data, this individual simply doesnot fit into either category. 452 K. WINKER, J. H. RAPPOLE ANLIM. A. RAMOS

......

FIGURE 2. Movements of Wood Thrush sedentary FIGURE 4. Movements of Wood Thrush wanderer individual no. 466 over 21 days (1985). 8 indicates no. 1793 over 19 days (example of a drifter). 8 indi- capture location; open circles are nets. Black dots are cates capture locations. Black dots are connected in connected in chronological order and show locations chronologicalorder and show locations (maximum of (maximum of two per day). Note relatively even use two per day); black dots not connected by lines are of spaceand narrow scope of movement pattern. nets. Note relatively localized wanderings. lowing this norte, this single Wood Thrush was level through netting and vocalization data. The captured and fitted with a transmitter. During 1984-1985 field season was the most thoroughly the next norte no signal was found. netted; in the following analyses the 3 days dur- ing this seasonon which nets were open for less POPULATION-WIDE MOVEMENTS than 4 hr were excluded. These days gave mis- The apparent correlation between individual be- leading capture rates, since most captures oc- havior and weather is evident at the population curred in the morning (when nets were open on

12 ,

0 Sedentaries

FIGURE 3. Movements of Wood Thrush wanderer 1 2 3 no. 1773 over 4.5 days (example of a mover). 8 in- dicates capture location. Black dots are connected in Fat Levels chronologicalorder and show locations (maximum of two per day); black dots not connected by lines are FIGURE 5. Comparison of subcutaneouslipid de- nets. Note long distancesbetween locations and broad posits between transmitter-bearing Wood Thrush se- scope of movements. dentaries and wanderers. WINTERING WOOD THRUSHES IN VERACRUZ 453

A

rn a 1 .km B ’ Feb. Nov. I DC?C. I Jan. 1 FIGURE 6. Captures(A) and recaptures(B)ofWood Thrushesfor the 1984-l 985 field season,by day, expressed as birds per 1,000 net hr. Fine horizontal lines denote seasonmeans. Closed circles represent rainy days and open circles represent sunny days. these days). Netting results confirmed the rela- > 0.0 1; G-test using Yates’ correction [Sokal and tionship between movements and weather re- Rohlf 198 11; Table 5). A separate analysis of ported for Los Tuxtlas by Ramos (1983). Sig- vocalization data corroborated this apparent nificantly more Wood Thrushes were captured weather-related change in aggression.Vocaliza- during days associatedwith nortes (rainy days: tions consistedof calls (breeding ground songwas n = 22, K = 10.12, SD = 7.62), than on days rarely heard in winter), and appearedto be given with good weather (sunny days: n = 27, K = 4.40, primarily during agonistic encounters with con- SD = 3.80; captures and recaptures included, z specifics (unpubl. data). Observed intraspecific = 2.83, P = 0.005, Mann-Whitney U-test). No interactions were always accompanied by vocal- correlations were found between capture rates izations, and spontaneous vocalizations were and either minimum or maximum daily tem- confined mainly to periods of low light: usually peratures.Figure 6 showsthe distribution of cap- dawn and dusk (unpubl. data). Only vocaliza- tures, recaptures,and weather events during the tions believed to be given in agonistic intraspe- 1984-1985 field season. The recapture rate of cific interactions were included in this analysis. Wood Thrushes was also higher on rainy days Vocalization data for the 1984-l 985 field season (excluding days upon which no birds were re- were analyzed by frequency under different captured: z = 2.49, P = 0.012, Wilcoxon’s rank sum test). Capture rates of new birds and total captures TABLE 4. Correlationcoefficients (r) for Wood Thrush (capturesand recaptures)decreased significantly captures,recaptures, and totals (captures+ recaptures) during the course of the 1984-l 985 field season, during (A) the entire course of the 1984-1985 field while recapturesincreased significantly (Table 4, season,and (B) during a period after which all residents case A). Individuals of the wanderer class were should have been captured. captured throughout the winter season. NO. Vocalizationsand agonisticencounters. Taped &YS New captures Recaptures Totals playbacks were less likely to elicit an aggressive A II -0.52** 0.27* -0.32* response from presumed territory holders on B 56 -0.31* 0.21 -0.11 rainy days (days associatedwith nortes) than on *P < 0.01. sunny and nonrainy days (G = 6.47, 0.25 > P **P < 0.001 454 K. WINKER, J. H. RAPPOLE AND M. A. RAMOS

TABLE 5. Weather-related occurrencesof aggressive TABLE 7. Wilcoxon’s rank sum test resultsof Wood responsesto playback of recordedaggressive vocaliza- Thrush vocalizations per day under different weather tions in the Wood Thrush. conditions.

NO Comparison Z P RespLTM response Rainy vs. sunny - 19.48 CO.001 Rainy days 0 10 Sunny vs. cloudy -1.40 0.082 Sunny and nonrainy days 23 30 Rainy vs. cloudy -5.66 CO.001 (G = 6.47, 0.25 > P > 0.01; corrected G-test). Rainy vs. all others -5.06 -Co.001 weather conditions (Tables 6, 7), showing that DISCUSSION significantly fewer vocalizations were given on days when it rained (days associated with nortes) Our two categories,sedentary and wanderer, fit than on days when there was no rain. a general conception of “territorial” and “float- In examining vocalization data, all low-light er”; however, in keeping with the caution urged vocalizations that were not verified aggressive by authors such as Burt (1943) and Emlen (1956) encounterswith conspecificswere excluded. This we refrain from using those particular labels in weather-related difference in vocalization rates describing individuals. Although many individ- was not an artifact of reduced hearing ability uals in this Wood Thrush population were ter- during rain; during heavy rain we were usually ritorial, the descriptor “territorial” cannot be not at the study site (such days were excluded confidently attached to all sedentary individuals from the analysis), and the rain on those days given a lack of corroborant behavioral data (i.e., that vocalizations were noted was usually re- aggressivebehavior), which were collected only strictedto a light, steadydrip, or to brief showers, incidentally. Nevertheless, the preceding results neither of which affectedour ability to hear other are considered in the context of a territorial sys- vocalizing forest residents. Hearing was notice- tem, reflecting our knowledge of the underlying ably reducedonly during l-3 day periodsof strong social system, if not of the exact status of all southerly winds (surudus),which were included individuals considered. in the “sunny” days sample. With spontaneous WEATHER-RELATED BEHAVIOR vocalizations removed from the sample we be- lieve these vocalization data are a good index of Wanderers.During nortes,our lowland study site intraspecific agonistic encounters. became host to large numbers of newly arrived, These analysessuggest that during nortesthere unbanded birds of many species(unpubl. data, is a decline both in the likelihood of an aggressive Ramos 1983). Ramos (1983, part 2, p. 4) found responseto a conspecific stimulus and in actual a significant difference in capture rates between aggressiveencounters. This is opposite to what rainy and “sunny” days for many speciesin this might be expected,given that intrusion rateswere region, as well as for overall (all species)capture higher at these times. rates. Our netting results show that these tem- porarily resident individuals and speciesdepart- ed either immediately or soon after the norte. TABLE 6. Number of Wood Thrush vocalizations Telemetry data revealed that the movements of heardper day under different weatherconditions (1984- transient Wood Thrushes during nortes usually 1985). On 15 rainy days no calls were heard, while on 5 sunny days two calls were heard, etc. encompassed an area much larger than the La Peninsula site. Unbanded birds, presumably

Vocalizations/day Rainy days Sunny days Cloudy days transients, were often seenduring nortesforaging along trails and small, dry streambeds,either in- 0 15 6 3 dividually or in loose aggregationsof two to three 1 7 12 10 2 1 2 individuals. 3 ; 2 Becausebirds resident on the site usually do : 1 1 not depart, new captures during nortes presum- 2 0 1 1 ably came from habitats other than the lowland Total types in which we were netting. Rappole and vocalizations 9 32 30 Warner (1980) and Ramos (1983, p. 9) conclud- WINTERING WOOD THRUSHES IN VERACRUZ 455 ed that since these new birds were accompanied some habitats outside of our lowland study site by species normally associated with the local undergo intolerable fluctuations in suitability, highlands (e.g., Myioborus miniatus, making lowland selva a desirable temporary ref- mexicanus), the newcomers came from higher uge. A large portion of the Wood Thrush pop- elevations. Reduced food availability in the high- ulation (ca. 50%) appearsto be subjecteddirectly lands was hypothesized as the reason for these to these fluctuations. movements. Sedentaries. Through the presence of higher Although we have no measurements of food numbers of conspecificsduring nortes, sedentar- availability, our observations seem to support ies seem to be indirectly affected by suitability this hypothesis. The highlands are exposed to fluctuations in habitat(s) outside the lowland sel- strongerwinds, heavier rains (see also Goldman va in which they reside. Whereas most sedentary 195 1, p. 283) and cooler temperatures than the birds remained during nortes, two left the area, lowlands (unpubl. data), suggestingthat insect later returning to their home ranges. These two food might be depleted (Buskirk and Buskirk departures (nos. 1645, 1778) suggestthat the 1976) and that foraging efficiency may be re- suitability of the La Peninsula site underwent duced. We have observed feeding arboreal in- fluctuations that were intolerable to at least some sectivorous birds slowly descend steep valley residents. Although it is probable that the ulti- walls on Volcan Santa Martha as the cloud cover mate cause of these two temporary departures of a mild norte descended( 16 March 198 5). The was weather, the proximate causemay have been movements of the single Wood Thrush captured intruder pressure,lowered resourceavailability, and followed on the slopesof Volcan Santa Mar- or both. tha also suggestthat norte conditions in the high- There are few documented casesof wintering lands might be the causeof lowland influxes. The territorial birds being forced by climatological mover categoryofwanderer included individuals changesto leave their areas.Schwartz (1964) not- which fit Ramos’ (1983) hypothesisof temporary ed that in the Northern Water-thrush (Seiurus visitors from the highlands during inclement noveboracensis)in Venezuela, individuals oc- weather (e.g., nos. 555, 1773, 1779, and 1780). cupying territories on higher ridges abandoned Number 1773 (Fig. 3) was capturedduring a norte, these sites as the winter progressedand these and its final disappearanceupriver corresponded areas became dry. Snow (1956) described terri- with another norte. Although this general move- torial Eurasian Blackbirds (Turdus merula) in ment pattern supportsRamos ’ “lowland retreat” Britain leaving their winter territories during se- hypothesis, it should be recalled that there was vere weather due to reduced food availability almost no forest further downriver (i.e., at lower and returning to their territories when conditions elevations) for these birds to use. improved. These Eurasian Blackbird territories Weather-related altitudinal movements by were defended for breeding season purposes, birds in the tropics have been reported by Wet- however, and not strictly as winter feeding ter- more (1926) and “hard-weather movements” ritories. Davies and Houston (198 1) noted that were noted by Lack (1968) and summarized for territorial Pied Wagtails (Motacilfa alba) fre- palearctic thrushes by Simms (1978). quently left their territories to feed on super- The reason norte-resident transients leave fol- abundant food sources,but it appeared that the lowing the nortesthat brought them in is a mys- territories were being defended as insurance tery. Perhaps it is a combination of the “return” against the time that the concentrated sources of suitable habitat elsewhere (a place to go) and were not available. Myers et al.( 1979) found that an increase in aggressionon the part of the local individual Sanderlings (Calidris alba) alternated territorial birds. An increasein aggressionduring between territorial and nonterritorial behavior, a norte could have little effect if there is nowhere presumably due to fluctuations in prey density. for expulsed intruders to go; with such a “sink” Vocalizations and agonistic encounters. The available after a norte, however, a resurgenceof decreasein vocalizations during nortes implies aggressionmight effectively lower densities (see that fewer aggressiveinteractions between Wood also Lack 1968). Thrushes occurredduring theseweather systems. The large, weather-related shifts in movement Given that this change in behavior occurswhen patterns evident through telemetry (wanderers) population densities are at their highest, we pos- and netting (norte-resident transients) imply that tulate that territorial defense becomes uneco- 456 K. WINKER, J. H. RAPPOLE AND M. A. RAMOS

nomical for most of the La Peninsula residents during 6 years of study: seven killed were of a during nortes,due to increased costs associated group of 27 wanderers (moved > 150 m from with foul weather and increased intruder pres- point of capture); two killed were of a group of sure. Recher and Recher (1969, p. 146) described 34 sedentary birds (moved 150 m or less from decreasedlevels of aggressionat high population initial capture point). These two proportions are levels among some shorebird species,but they significantly different (chi-square test, P < 0.05), hypothesizedthat theseaggregations were formed and these data suggestthat wandering birds ex- becauseof locally abundant food organisms, and perience a higher mortality rate than sedentary that this, in combination with high population individuals. Mortality rates in birds per tracking density, causedthe reduced frequency of aggres- day were estimated as 0.026 for wanderers and sive interactions observed. Myers et al. (1979) 0.004 for sedentary individuals (Rappole et al. described curtailed aggression when the non- 1989). Note that the criteria for categorization breedingterritories of individual Sanderlingswere in Rappole et al. (1989) are simpler than those suddenly invaded by nonterritorial flocks of con- used here, causing some incongruity in the ca- specifics.Given a sufficient number of intruders, tegorization of individuals between thesepapers. a similar responsemight occur within an entire This difference has no effect upon the outcome local population. However, there are two poten- of this comparison, however. tial costs to territorial Wood Thrushes during Although avian evidence for differential mor- nortes-increased intruder pressure and bad tality between sedentary individuals and wan- weather. derers is scarce (but see Tompa 1962; Carrick 1963; Jenkins et al. 1963, 1964, 1967; cf. Cat- terall et al. 1989), mammalian nomads are known DIFFERENCES BETWEEN SEDENTARIES to sufferhigh mortality (seeErrington 1943, 1963; AND WANDERERS Beer and Meyer 1951; Hawkins et al. 1971; Three hypothesescould explain the existence of Gaines and McClenaghan 1980, p. 184-185; two behavioral classesamong wintering Wood Packard and Mech 1980, Andelt 1985). Nomad- Thrushes. (1) A genetic polymorphism exists in ic individuals lack intimate knowledgeof an area, the population, causingsome individuals to wan- a fact that may be disadvantageous(Brown 1963). der and others to be sedentary. (2) The two strat- In addition to lacking stable food resources,no- egiesexist in an evolutionary stable mixture, each mads can also be more vulnerable to predation. having equal fitness (Davies 1980, p. 72). (3) In experimental studies, Metzgar (1967) and Territorial aggressionby sedentary individuals Ambrose (1972) found that in rodents similar prevents some individuals from settling. If either otherwise, those unfamiliar with their environ- of the first two hypothesesapplied to the Wood ment (experimentally induced transients) sus- Thrush in Los Tuxtlas, we would expect to see tained higher predation rates by than sed- no differential mortality between sedentariesand entary individuals. Physically degradedtransients wanderers(because both hypothesesimply equal might show a higher mortality rate than seden- fitness). Also, in regard to the first hypotheses, tary conspecificsin good condition (Gaines and we would expect little or no behavior switching McClenaghan 1980, p. 185), but this is not the (e.g., Rappole et al. 1989, p. 407,408). Although only reason for observed differential mortality; we did not examine interactions between seden- nomads are not always physically inferior to sed- taries and wanderers, aggressivereaction to con- entary individuals (Metzgar 1967, Ambrose specificswas demonstrated through playback ex- 1972). Environmental unfamiliarity may have periments, suggestingthat the latter hypothesis been the reason for the different flushing behav- merits consideration. Removal experiments iors observedwhen tracking Wood Thrushes.The needed to test this hypothesis were not per- benefits of familiar surroundings may be a factor formed, however (discussedbelow). We have only in the season-to-seasonsite fidelity often seen in one observation to suggestthat wanderers might sedentary (in many casesterritorial) birds, both be prevented from settling becauseof aggression on the breeding and wintering grounds. from sedentary conspecifics (unpubl. data). As A legitimate question, given this apparent Greenberg (1986) noted, these interactions are mortality imbalance, is whether wanderers de- difficult to observe in dense tropical vegetation. clined over the season. If we suppose that all Dtjizrential mortality. Rappole et al. (1989) residents (i.e., sedentaries) should be captured found that nine birds were killed by predators after 2 1 days of netting (an arbitrary figure) and WINTERING WOOD THRUSHES IN VERACRUZ 451

that new captures after that are mostly nonter- Migrants defending territories on their win- ritorials, or wanderers, it appears that this cat- tering grounds have been considered under the egory continues to decline significantly (Table 4, light cast by territorial theory developed pri- case B). What this means is uncertain. It could marily from temperate breeding ground research mean: (1) that wanderers are dying; (2) that wan- (e.g., Rappole and Warner 1980, Price 1981). derersare leaving the area (facultative migration, Although studies of true nonbreeding territorial sensu Ten-ill and Ohmart 1984); (3) wanderers systemsare few (as opposedto systemsin which are finding suitable vacant spaceand are becom- territories are held during the nonbreeding sea- ing sedentary, or (4) that nortesare not as severe son on the breeding grounds), evidence suggests later in the season, causing fewer birds to seek that territorial theory has wide applicability in refugein the lowlands. None ofthese possibilities nonbreeding systems(e.g., Schwartz 1964; Rap- need be occurring exclusively. Although wan- pole and Warner 1976,198O; Kodric-Brown and dering Wood Thrushes appear to be disadvan- Brown 1978; Myers et al. 1979; Davies and taged, practicing this strategyis not immediately Houston 1981; Price 1981; Morton et al. 1987). detrimental: three were tracked for 2 weeks, and Basic similarities probably exist among popu- individuals assignableto the wanderer classwere lations in which individuals defend resourcesand present throughout the season. the spacethese resourcesoccupy (e.g., the unify- Subcutaneouslipids. The physical status of ing concept of economic defendability; Brown wandering Wood Thrushes seemed very similar 1964). These similarities apparently constitute a to that of sedentary birds. In fact, wanderers had sufficient basis for considering nonbreeding ter- higher subcutaneousfat deposits than sedentar- ritorial systemsunder the rubric of a generalbody ies, which is commonly regardedas a “superior” of territorial theory. condition (Rappole and Warner 1976, Real and Brown (1969, p. 294) proposed that popula- Caraco 1986). We hypothesize that the ultimate tions on their breeding grounds could be classi- reason for the larger fat deposits of wanderers fied into three critical density levels, based on was their uncertain food supply. Survival when how territoriality affectsindividual distribution. dependent upon an insecure food source prob- As population size increases, individuals are ably necessitatesthe practice of storing energy forced to utilize less suitable habitats, until, at when food is available. A similar reasoning was level 3, all habitats suitable for breeding are taken followed by Rappole and Warner (1980, p. 380- by territorial individuals, and a nonbreeding sur- 384) in comparing the fat deposits of a variety plus is left over for which space is unavailable of speciesto the social systemsof those species. (breeding-ground floaters). Brown’s (1969) crit- They saw fat deposition as an adaptation to an- ical density model, together with the more rig- ticipate energy stress.They made the prediction orous but similar Fretwell-Lucas ideal despotic and presented evidence that fat levels should be model (Fretwell and Lucas 1970, Fretwell 1972) higher in speciessubsisting on resourcesclumped seems applicable to populations that are terri- in spaceand time than in specieswhose resources torial during the nonbreeding season. Both of are evenly distributed. The Wood Thrush data these models predict that habitats of lower suit- suggestthat this prediction also holds within a ability will be occupied when population de- single population. mands exceed the amount of available space/ resourcesin preferred habitats. Further, Brown (1969) predicts the occurrenceof a nonterritorial THE FUNCTION OF WINTER TERRITORIALITY class at peak, or level 3 densities. Unlike territories held for breeding purposes, Because many Wood Thrushes appear to (1) where reproduction is the primary concern, the use habitats that undergo intolerable fluctuations wintering territories of nearctic migrant passer- in suitability, and (2) often use a nonterritorial ines are probably held for survival value. Hold- strategy,we accept these models as a valid con- ing a suitable nonbreeding territory has in some ceptual paradigm in which to consider this non- casesbeen demonstrated to confer a survival ad- breeding territorial population. In addition, we vantage (e.g., Jenkins et al. 1964, Dhondt 197 1). tentatively acceptBrown ’s (1969) critical density In the territorial systems of wintering migrants level 3 as a valid descriptor of wintering Wood it appears that the resource being defended is Thrush densities in Los Tuxtlas. food, but familiarity with an area might also in- Nomadic individuals in this population ap- creasethe probability of survival. pear to be at a disadvantage when compared to 458 K. WINKER, J. H. RAPPOLE AND M. A. RAMOS sedentary conspecifics.Our data (cf. Rappole et to provide a mechanism for winter population al. 1989) are among the first for territorial sys- regulation, we have not shown that this mech- tems of nonbreeding migrants, and further re- anism occurs as a result of territorial spacing, search is needed to ascertain the degree of this only that it occurs in a territorial system. Prox- disadvantage. If further work supports the ap- imately, it seemsthat this mechanism is not star- plication of the above models to nonbreeding vation, as Price (198 1) implied it should be in a migrants, however, our data imply a mechanism system in which food is defended, but predation. for winter population regulation in territorial Individual strategiesof resourceexploitation fre- nonbreeding migrants. It is clear that further ex- quently reflect predation factors, however (Pul- amination of the applicability of these models to ham and Caraco 1984). Ekman (1984) found that populations of territorial wintering migrants is although the proximate factor contributing to necessary.In particular, it must be demonstrated wintering Willow Tit (Parus montanus) mortal- that nonterritorials adopt the wandering strategy ity was predation, this predation was ultimately as a forced choice due to the presence of terri- a consequenceof intense competition for food. torial conspecifics.Removal experiments will be crucial to this demonstration. ACKNOWLEDGMENTS Through removals of territorial individuals, The World Wildlife Fund (US) financed the project presumable nonterritorials have been shown to through a grant to JHR and MAR. The senior author occupy vacanciesrapidly and become territorial. received additional assistancefrom the Dayton Nat- Numerous studies have demonstrated the pres- ural History Fund of the Bell Museum of Natural His- tory, University of Minnesota. The Instituto National ence of thesenonterritorials (critical density level de InvestigacionesSobre RecursosBioticos (INIREB), 3) on the breeding grounds (reviewed by Klomp Mexico, provided greatly appreciated help with logis- 1972, Stutchbury and Robertson 1985). Thus far, tics and in-country cooperation. For assistancein the however, comparable evidence from the winter- field we thank S. Barrios, A. Hartlage, C. Nemerov, and especiallyJ. L. Alcantara, M. Neri F., and S. Stuck- ing grounds of migrants is scarce. Rappole and er. We would also like to thank B. C. Eliason,R. Green- Warner (1980) reported seven removal experi- berg, F. McKinney, J. R. Tester, D. W. Warner, and ments in four speciesresulting in the occupancy three anonymousreviewers for their comments on ear- of the vacated area by presumed floaters. Morton lier drafts ofthe typescript. Lastly, we thank the people et al. (1987) performed removal experiments in of La Peninsula, particularly Agostin Mozo and Pas- cual Velasco, for their generosityand friendship. wintering Hooded Warblers ( Wilsoniacitrina), but were uncertain of the behavioral status of LITERATURE CITED territorial replacements.We found (unpubl. data) AMBROSE,H. W., III. 1972. Effect of habitat famil- three similar instances in the Wood Thrush after iarity and toe clipping on the rate of predation accidental removal (death) of territory holders in Microtus pennsylvanicus.J. Mammal. 53:909- (only one of these birds, no. 1778, had a trans- 912. mitter; the other two were deemed territorial AMERICANORNITHOLOGISTS UNION.’ 1983. Check- based upon capture, recapture, and/or resighting list of North American birds. 6th ed. American Ornithologists’ Union, Washington, DC. records).In each case,the vacancy was occupied, ANDELT,W. F. 1985. Behavioral ecology of coyotes but it was uncertain whether the occupantswere in south Texas. Wildl. Monogr. 94:145. neighbors expanding their home ranges or for- ANDRLE,R. F. 1966. North American migrants in mer wanderers. the Sierrade Tuxtla of southernVeracruz, Mexico. Condor 68:177-184. Winter population regulation? It is apparent ANDRLE,R. F. 1967. Birds of the Sierra de Tuxtla of from this study that two wintering strategiesexist southernVeracruz, Mexico. Wilson Bull. 163-187. in the wintering population of Wood Thrushes BEER,J. R., ANDR. K. MIXER. 1951. Seasonalchanges that we studied: sedentary (most of which are in the endocrine organsand behavior patterns of probably territorial) and wanderer (most of which the muskrat. J. Mammal. 32: 173-l 9 1. BROWN,J. L. 1963. Aggressiveness,dominance and are probably not territorial). The wanderer class social organization in the Steller Jay. Condor 65: may be disadvantaged through a higher rate of 460484. predation, but this study has demonstrated nei- BROWN,J. L. 1964. The evolution of diversity in ther the degreeof this disadvantage nor the fac- avian territorial systems.Wilson Bull. 76:160-169. BROWN,J. L. 1969. _Territorial behavior and popu- tor(s) causing wanderers to adopt this strategy. lation regulation in birds: a review and re-evalu- In addition, although these data, together with ation. Wilson Bull. 81:293-329. those presented in Rappole et al. (1989) appear BURT, W. H. 1943. Territoriality and home range WINTERING WOOD THRUSHES IN VERACRUZ 459

conceptsas applied to mammals. J. Mammal. 24: 1971. Dispersal of deer from Crab Orchard Na- 346-352. tional Wildlife Refuge. J. Wildl. Manage. 35:216- BUSKIRK,R. E., ANDW. H. BUSKIRK. 1976. Changes 220. in arthropod abundancein a highland Costa Rica HELMS,C. W., ANDW. H. DRURY. 1960. Winter and forest. Am. Midl. Nat. 95288-398. migratoryweight and fat field studiesof someNorth CARRICK,R. 1963. Ecologicalsignificance of territory American buntings. Bird-Banding 3 1:l-40. in the Australian Magpie, Gymnorhina tibicen. HOLMES.R. T.. AND T. W. SHERRY. 1988. Assessing Proc. XII Int. Omithol. Congr. (1962):740-753. population trends of New Hampshire forest bird; CATTERALL, C. P., J. KIKKAWA, AND C. GRAY. 1989. local vs. regional patterns. Auk 105:756-768. Inter-relatedage-dependent patterns of ecologyand JENKINS, D., A. WATSON,AND G. R. MILLER. 1963. behaviour in a populationof silvereyes(Aves: Zos- Population studies on Red Grouse, Lagopus la- teropidae). J. Anim. Ecol. 58:557-570. gopus scoticus(Lath.) in northeast Scotland. J. CAU~~EI& C. ’ 1984. In the rainforest. A. A. Knopf, Anim. Ecol. 32:3 17-377. New York. Jnrnu~s, D., A. WATSON,AND G. R. MILLER. 1964. DAVIES,N. B. 1980. The economicsof territorial be- Predation on Red Grouse populations. J. Appl. havior in birds. Ardea 68:63-74. Ecol. 1:183-195. DAVIES,N. B., AND A. I. HOUSTON. 1981. Owners JENKINS,D., A. WATSON,AND G. R. MILLER. 1967. and satellites:the economics of territory defense Population fluctuations in the Red Grouse, La- in the Pied Wagtail (Motacilla alba). J. Anim. Ecol. gopuslagopus scoticus. J. Anim. Ecol. 36:97-122. 50:157-180. KLOMP,H. 1972. Regulation of the size of bird pop- DHONDT,A. A. 1971. Some factors influencing ter- ulations by means of territorial behavior. Neth. J. ritory in the Great Tit, Parus major, L. Gerfaut Zool. 22:456488. 61:125-135. KODRIC-BROWN,A., AND J. H. BROWN. 1978. Influ- EDWARDS,E. P. 1972. A field guide to the birds of ence of economics,interspecific competition, and Mexico. E. P. Edwards, Sweet Briar, VA. sexual dimorphism on territoriality of migrant EKMAN,J. 1984. Density-dependent seasonal mor- Rufous Hummingbirds. Ecology 59:285-296. tality and population fluctuationsof the temperate LACK,D. 1968. and natural selection. zone Willow Tit (Parus montanus).J. Anim. Ecol. Oikos 19:1-9. 53:119-134. LOETSCHER,F. W. 1955. North American migrants ELGOOD,J. H., R. E. SHARLAND, AND P. E. WARD. in the state of Veracruz, Mexico: a summary. Auk 1966. Palearctic migrants in Nigeria. Ibis 108: 72~14-54. 84-l 16. LOMNICKI,A. 1978. Individual differences between EMLEN,J. T. 1956. Defended area?-a critique of the and the natural regulation of their num- territory conceptand of conventionalthinking. Ibis bers. J. Anim. Ecol. 47:461475. 99~352. LOMNICKI, A. 1988. Population ecology of individ- ERRJNGTON,P. L. 1943. An analysis of mink pre- uals. Princeton Univ. Press, Princeton, NJ. dation upon muskrats in north-central United LOT-HELGUERAS,A. 1976. I.a Estacion de Bioloaia States. Iowa Agric. Exp. Stn. Res. Bull. 320:797- tropical de 10sTuxtlas: pasado,presente y future, 924. p. 3 l-5 1. In A. Gomez-Pomna led.1. Investiga- ERRINGTON,P. L. 1963. Muskrat populations. Iowa &ones sobre regeneration de Selvas Altas en Ve- State Univ. Press, Ames, IA. racruz, Mexico. Cecsa, Mexico. FRETWELL,S. D. 1972. Populationsin a seasonalen- METZGAR,L. H. 1967. An experimental comparison vironment. Princeton Univ. Press.Princeton. NJ. of ScreechOwl predation on residentand transient FRETWELL,S. D., AND H. L. LUCAS,JR. 1970.’ On White-footed Mice (Peromyscus leucopus). J. territorial behavior and other factors influencing Mammal. 48:387-39 1. habitat distribution in birds. I. Theoretical devel- MIRANDA,F., AND E. HERNANDEZXOLOCOTZI. 1963. opment. Acta Biotheor. 19:16-36. Los tipos de vegetation de Mexico y su clasifica- GAINES,M. S., ANDL. R. MCCLENAGHAN.1980. Dis- cion. Bol. Sot. Bot. Mex. 28:29-179. persal in small mammals. Annu. Rev. Ecol. Syst. MORSE,D. H. 1980. Population limitation; breeding 11:163-196. or wintering grounds?,p. 505-516. In A. Keast GOLDMAN,E. A. 1951. Biological investigations in and E. S. Morton [eds.], Migrant birds in the neo- Mexico. Smithsonian Institution Press, Washing- tropics. Smithsonian Institution Press, Washing- ton, DC. ton, DC. GREENBERG,R. 1986. Competition in migrant birds MORTON,E. S., J. F. LYNCH,K. YOUNG,AND P. MELHOP. in the nonbreeding season,p. 281-307. In R. F. 1987. Do male Hooded Warblers excludefemales Johnston [ed.], Current ornithology. Vol. 3. Ple- from nonbreedingterritories in tropical forest?Auk num Press, New York. 104:133-135. HARVEY,M. J., AND R. W. BARBOUR. 1965. Home MYERS, J. P., P. G. CONNORS,AND F. A. PITELKA. range of Micro&s ochrogasteras determined by a 1979. Territory size in wintering Sanderlings:the modified minimum area method. J. Mammal. 46: effects of prey abundance and intruder density. 398-402. Auk 96:551-561. HASSELL,M. P. 1986. Detecting density dependence. MYERS,N. 1980a. The presentstatus and future pros- Trends Ecol. Evol. 1:90-93. pects of tropical moist forests. Environ. Conserv. HAWKINS,R. E., W. D. KLIMSTON,AND D. C. AUTRY. 7:101-l 14. 460 K. WINKER, J. H. RAPPOLE AND M. A. RAMOS

MYERS,N. 1980b. The conversionof tropical forests. SOKAL,R. R., AND J. H. ROHLF. 1981. Biometry. Environment 22:6-13. W.H. Freeman and Co., San Francisco. MYERS,N. 1981. The hamburger connection: how SOTO-ESPARZA,M. 1976. Algunos aspectosclimati- Central America’s forests become North Ameri- cos de la regi6n de 10sTuxtlas, Veracruz, Mexico. ca’s hamburgers.Ambio 10:3-g. Cecsa, Mexico. PACKARD,J. M., ANDL. D. MECH. 1980. Population STUTCHBURY,B. J., AND R. J. ROBERTSON.1985. regulationin wolves, p. 132-l 50. In M. N. Cohen, Floatingpopulations of female Tree Swallows.Auk R. S. Malpass, and H. G. Klein [eds.], Biosocial 102:65l-654. mechanismsof population regulation. Yale Univ. TERBORGH,J. W. 1980. The conservation status of Press, New Haven, CT. neotropical migrants: present and future, p. 21- PENNINGTON,T. D., ANDJ. SARUKHAN.1968. Arboles 30. In A. Keast and E. S. Morton [eds.], Migrant tropicales de Mexico. Instituto National de In- birds in the neotropics. Smithsonian Institution vestigacionesForestales, Mexico. Press, Washington, DC. PRICE,T. 1981. The ecologyof the Greenish Warbler TERRILL,S. B., AND R. D. OHMART. 1984. Facultative Phylloscopus trochiloides in its winter quarters.Ibis extension of fall migration by Yellow-rumped 123:131-144. Warblers (Dendroica coronata). Auk 101:427-l38. PULL=, H. R., AND T. CARACO. 1984. Living in TOMPA,F. S. 1962. Territorial behavior: the main groups: is there an optimal group size?, p. 122- controlling factor of a local Song Sparrow popu- 147. In J. R. Krebs and N. B. Davies [eds.], Be- lation. Auk 79:687-697. havioural ecology: an evolutionary approach. VOGT, W. 1970. The avifauna in a changingecosys- Blackwell Scientific Publications, Oxford. tem. Smithson. Contrib. Zool. 26:8-16. RAIM, A. 1978. A radio transmitter for small pas- WETMORE,A. 1926. The migrations of birds. Har- serines.Bird-Banding 491326-332. vard Univ. Press, Cambridge, MA. Amos, M. A. 1983. Seasonal movements of bird WILLIS, E. 0. 1966. The role of migrant birds at populationsat a neotropical study site in southern swarms of army . Living Bird 5: 187-23 1. Veracruz, Mexico. Ph.D.diss., Univ. of Minne- WYATT, W. W., AND C. M. BRIDGES,JR. 1967. Sta- sota, Minneapolis. tistics for the behavioral sciences.D.C. Heath and RAMOS, M. A. 1988. Eco-evolutionaryaspects of bird Co., Boston, MA. movements in the northern neotropical region. Proc. XIX Int. Omithol. Congr. (1986):250-293. RAPPOLE,J. H., AND D. W. WARNER. 1976. Rela- tionships between behavior, physiology, and APPENDIX. Data from 85 specimens collected as weatherin avian transientsat a migratory stopover winter residentsin Los Tuxtlas during 1973-1974 and site. Oecologia 26: 193-2 12. 1974-1975. Skins are housed in the Bell Museum of RAFTOLE,J. H., AND D. W. WARNER. 1980. Ecological Natural History, Minneapolis, Minnesota. Numbers in aspects of migrant bird behavior in Veracmz, parenthesesare the number of specimens for which Mexico, p. 353-393. In A. Keast and E. S. Morton, data were available. [eds.], Migrant birds in the neotropics. Smithson- ian Institution Press, Washington, DC. Presence(P) or absence(A) of wingbars (70): RAPPOLE,J. H., E. S. MORTON,T. J. Lowou, III, AND HY AHY J. L. Ruos. 1983. Nearctic avian migrants in the P ? A P ? A neotronics.U.S. Fish and Wildlife Service. Wash- ington; DC. 11 8 5 6 8 32 RAFTOLE.J. H.. AND E. S. MORTON. 1985. Effects of habitat alteration on a tropical avian forest com- munity, p. 1013-1021. In P. A. Buckley, M. S. Age”: Foster, E. S. Morton, R. S. Ridgely, and F. G. HY AHY Buckley [eds.], Neotropical ornithology. Omithol. Monogr. No. 36. American Ornithologists’ Union, All winter (70): 24 (34%) 46 (66%) Washington, DC. Pre-25 February (57): 23 (40%) 34 (60%) RAPPOLE,J. H., M. A. &OS, ANDK. WINKER. 1989. Movements and mortality in wintering Wood Sex (82): Thrushes. Auk 106:4024 10. Rw, L., AND T. CARACO. 1986. Risk and foraging Male Female in stochasticenvironments. Annu. Rev. Ecol. Syst. 17:371-390. 59 (72%) 23 (28%) RECHER,H. F., ANDJ. A. RECHER.1969. Some aspects of migrant shorebirds.II Aggression.Wilson Bull. Stomach contents (26): 81:140-154. Seed@ Fruit Arthrouods SCHWARTZ,P. 1964. The Northern Waterthrush in Venezuela. Living Bird 3: 169-l 84. 11 (42%) 5 (19%) 16 (62%) SIMMS,E. 1978. British thrushes.Collins, London. aDue to progressingossification, a more reliable age ratio is obtained SNOW,D. W. 1956. Territory in the Blackbird Turdus by usingthe 57 specimenscollected before 25 February. merula. Ibis 98:438-447. bProbably fruit seeds.