876 SHORT COMMUNICATIONS od, temperature and diet quality. Physiol. Zool. Calow [eds.], Physiologicalecology: an evolution- 66537-560. ary approach to resourceuse. Blackwell, Oxford. NOVOA, F. F. 1993. Ecofisiologiade Zonotrichia ca- STEEL, R. G. D., ANDJ. H. TORRIE. 1985. Bioestad- pensis:cambios estacionalesen el gasto y la ad- istica: principios y procedimientos.McGraw-Hill, quisici6n de energia. Ph.D. diss., Universidad Bogota. Chile, Santiago. WALSBERG, G. E., AND C. W. THOMPSON. 1990. An- SIBLEY, R. M. 1981. Strategiesin digestionand defe- nual changesin gizzard size and function in a fru- cation, p. 109-139. In C. R. Townsend and P. givorous bird. Condor 92~794-795. The Condor 98:876-879 0 The CooperOrnithological Society 1996 WINTERING SWAINSONS’ HAWKS IN CALIFORNIAS’ SACRAMENTO-SAN JOAQUIN RIVER DELTA ’ SEBASTIANK. HERXKY Department of Wildlife, Fish, and ConservationBiology, Universityof California at Davis, Davis, CA 95616 Key words: Swainsons’ Hawk; Buteo swainsoni; served foraging flocks of up to 16 Swainson’s Hawks natural history; winter distribution;Sacramento-San on five Delta islands previna on small rodents in close JoaquinRiver Delta. associationwith farming operations. The objective of the present study was to systematically monitor the The Swainson’s Hawk (Buteoswainsoni) is considered population to determine wintering seasonchronology, a summer resident breeder on lame Darts of North size, composition, distribution, and foraging ecology. America, leaving in winter to migrate south to South STUDY AREA AND METHODS America (AOU 1983, Palmer 1988). Wintering Swain- son’s Hawks in North America had been confirmed The study area was in the Sacramento-SanJoaauin only from southernFlorida, where somejuveniles were River Delta in the Central ValIey ofCalifornia (38005’N, found during most winters, and from the southwestern 121”35’w) (Fia. 1). Historicallv. the Delta was an ex- United Statesas occasionalmigration dropouts(Palm- tensive freshw%er marsh. Naturally deposited levees er 1988). had been forming regularlyflooded islandsdominated Swainson’s Hawks in California breed chiefly in the by tule (Scirpusspp.), cattail (Typha spp.), and reeds Central Valley and on the Northeastern Plateau (Cal- (Phragmitesspp.) (Her-boldand Moyle 1989). Sincethe ifornia Department of Fish and Game 1993), and the 185Os,> 80% of the Delta’s wetlands were converted speciesis listed as threatened in the state. Until 1990, to mostly agriculturallands (California Department of few credible winter recordshad been reported from the Water Resources 1987). The area’s main crops were state (e.g., Browning 1974, McCaskie 1985, Morlan et corn, grain, and hay. Accesswas limited becausemost al. 1987, McCaskie 1989), most of which were consid- Deltaislandswere privately owned and few public roads ered to represent very late or early miarants. Several led through the study area. Boats could not be used reports from California’s Sacramento-San Joaquin becausehigh leveesand elevationsbelow sealevel lim- River Delta (LeValley and Rosenberg 1984, Campbell ited visibility of islands. et al. 1986, Morlan et al. 1987, Campbell et al. 1988, Swainson’s Hawk population surveys were con- Yee et al. 1989, Erickson et al. 1990) suggestedsome ducted by car on 23 islands from 25 September 1993 Swainson’s Hawks were winterina locallv. Durine the to 11 March 1994. I first surveyedareas where hawks 1990/1991 winter, 28 hawks were cont%med ig the were seen previously (Holt and Yee, unpubl. data). Sacramento-SanJoaquin River Delta, mostly consist- Surveyswere extended to other accessibleislands and ing of adults (Yee et al. 1991). A population of com- conducted 4-6 times a week by mid-October and re- parable size was reconfirmed in the two subsequent duced to twice a week by late February. Survey fre- winters (W. Holt and D. G. Yee, unpubl. data). quency was determined by logistic factors such as ac- Holt and Yee (unpubl. data) observedthe population cessibilityand remoteness.I conductedfrom 5 1 to 92 irregularly during all three winters. Thev discovered a surveysper island on Andrus, Brannan, Bouldin, and communal roost on Andrus Island (Fig. 1) and ob- Terminous, from 31 to 50 surveysper island on Em- pire, King, Staten,and Twitchell, from 11 to 30 surveys per island on Sherman and Venice, and from one to ten surveysper island on Bacon, Bethel, Byron, Hol- I Received 19 June 1995. Accepted 31 July 1996. land. Jersev, Lower Jones. Medford. Palm. Roberts. z Current address:Foundation for Tropical Research Upper Jones,Veale, Victoria, and Webb. Potentialroost and Exploration, P.O. Box 74431, Davis, CA 95617, sites were located during initial surveys and visited e-mail: [email protected] regularlyfrom mid-November to early February in late SHORT COMMUNICATIONS 877 A Roost Site FIGURE 1. Study area and location of the Swainson’s Hawk roost site in the central Sacramento-SanJoaquin River Delta, California. afternoon. One roost used by Swainson’s Hawks in November. Previously, I observed 5 30 possiblytran- previous years on Andrus Island (Fig. 1) was inten- sient Swainson’s Hawks per day from 25 September sively monitored from late November to late February. to 17 October, whereasonly zz 5 were located from 18 Numbers, locations,and behavior were recorded, as October to 18 November, with no birds present at the well as color morphs and age classes(juvenile, im- Andrus Island roost. Thereafter, numbersobserved per mature, adult). Whenever hawks were observed for- day varied from zero to a maximum of 29 at the Andrus aging, I also recorded foraging techniques,prey items, Island roost on 28 January. The population consisted numbers of other buteos present, interspecific inter- mostly of adults colored light to dark rufous, with only actionswith other buteos,field conditions,and farming one immature and two juveniles observed. The last operations. group of wintering Swainson’s Hawks was sighted on Prey populationswere sampledfor presence/absence 20 February on Andrus Island. Afterwards, only one on Terminous Tract from 8-l 2 Februarv 1994. I cao- wintering bird was seen, a juvenile at the roost on 22 tured rodents using 36 Sherman live-traps set 15 m February. The first spring migrants, which differed in apart on a squaregrid. They were checkedat dawn and plumagecolor from all wintering birds, were observed dusk.Two gridswere placedrandomly on two adjacent on 3 March on Terminous Tract. wheat fields flooded after harvest and drained in early As in previousyears, Swainson ’sHawks useda com- winter. The first grid was placed on a field that was munal roost at the southeasterntip of Andrus Island, tilled four days previously, the secondon the adjacent, a 240 m row of approximately 18-20 m tall eucalyptus untilled field. Swainson’s Hawks were foraging exten- (Eucalyptussp.) on a breakwater surroundinga boat sively over both fields during the till. marina. Even though the number of birds observed at the roost varied from a few up to 29, and despite the RESULTS presence of tall stands of eucalyptus on other Delta Swainson’s Hawks wintered in the Delta from at least islands, I found no roosting Swainson’s Hawks else- 19 November to 20 February. The arrival of the bulk where. They were seen at the roost at all times of day, of the winter population was marked by the sighting but mid-day sightingsmostly coincided with fog. De- of 13 Swainson’s Hawks on Bouldin Island on 19 No- parture and arrival times from and to the roost were vember and of 8-9 at the Andrus Island roost on 21 variable and seemed to be affected by weather condi- 878 SHORT COMMUNICATIONS tions. Hawks usually arrived singly, in pairs, or rarely DISCUSSION in groupsof 5 5 from mid-afternoon to sunset.Mom- A population of about 30 Swainson’s Hawks regularly ing departures were more irregular. Occasionally, I winters in the Sacramento-SanJoaquin River Delta. found no birds during early morning visits (about 08: It probably does not migrate becausecorn and wheat 00), suggestingthey had already left the roost. For un- farming present foraging opportunities almost contin- known reasons,29 Swainson’sHawks departedin mass uouslyfrom fall to spring.Fallow fieldsof corn or wheat from 08:22-08:25 on 28 January, and 20 departed stubbleprovide suitable habitat for small rodent pop- within one minute at 11:32 on 30 January. The roost ulations. Interestingly, the foraging behavior of win- was shared by several Red-tailed Hawks (Buteo ja- tering Swainson’s Hawks closely resembled that of maicensis)and 5 120 Great Egrets- (Cmmerodius al- breedingconsoecifics in California’s Central VaIlev_. (ea., _. bus). ’ Estep 1989). Among other similarities, breeding birds Swainson’s Hawks were observed 52 times from 19 regularly formed foraging groupsand spent > 50% of November to 20 Februarv on nine islands outside the their foragingtime hunting in responseto farming ac- immediate vicinity of the roost. Of these, 36 (69%) tivities such as tilling and harvesting (Estep 1989). were of foraging birds. Thirty (83%) foraging obser- Swainson’s Hawks varied in number at the Andrus vations of Swainson’s Hawks were of mixed species Island roost, suggestingsome birds regularly spent the flocks including Red-tailed Hawks and Rough-legged night elsewhere.Alternate roost sites might have been Hawks (Buteo lagopus),which varied in size from 4 to usedon islands I was unable to access,or hawks might z 60 buteosand typically consistedof about 27 birds. have roostedon the ground. There also were time gaps Swainson’s Hawks varied from 1 to 16, Red-tailed as long as 20 days in December and January, during Hawks from 3 to about 50, and Rough-leggedHawks, which I wasunable to locateforaging Swainson ’sHawks. which occurred in 18 of the 30 flocks. from 1 to 4 A period of frequent, dense, low-lying fog from mid- individuals per flock. Nonforaging and foragingSwain- December to mid-January and occasionalfoggy days son’s Hawks not associatedwith other Buteo species until the end of Januarywere partially responsiblefor occurredeither singly or in groupsof 5 5 birds.