152 Sno}•r COMMUNICATIONS VOL. 38, NO. 2

nus). During one hunt, eight falconschased a king- A wild hatch-year(HY) male falcon wasfbund eating for about 1 rain until it sought refuge in a mes- prey while perched on the rafters underneath one of quite. The falconsfollowed it, with five waiting in the the hack boxes.Attendants reported that this falcon treetop while three others ran and hopped through "generouslyshared" his kill, possiblya swallow,with a the lower branchesuntil the kingbird flushed and was female hacked falcon. captured, a sequenceoften exhibited by mated pairs A wild-hatched adult female arrived at a hack site and (Keddy-Hector 2000). led the first successfulgroup hunt of the yr. She cap- 3. In 1993, a group of sevenfalcons chasedand caught tured a meadowlarkafter chasingit together with two a Ladder-backedWoodpecker (Picoidesscalaris). Sev- HY hacked falcons, all three stooping in turn. At a eral falcons fed on it simultaneously,while the others different hack site, a previouslyhacked adult female settled on perches nearby. When a Northern Harrier regularly visited from 1999-2002. This falcon occa- (Circuscyaneus) approached the kill site, two of the sionallyfed from the tower,joined in hunts and tower non-feedingfalcons left the group and drove the har- defense, and tolerated food-beggingfrom the HY fal- rier awaywhile the others continued their meal un- cons. Attendants described her behavior as "mentor- disturbed (C. Perez pers. comm.). ing."

j RaptorRes. 38(2):152-157 ¸ 2004 The Raptor ResearchFoundation, Inc.

SUMMER ROADSIDE RAPTOR SURVEYS IN THE WESTERN PAMPAS OF

MICHAELI. GOLDSTEIN1 AND TOBYJ. HIBBITTS 9 TexasA&M University,Department of Wildlifeand Fishe•esSciences, 2258 TAMU, CollegeStation, TX 77843 U.S.A.

KEyWOP, DS: ChimangoCaracara; chimango;Crest- to Zapala, Neuqu•n (Travaini et al. 1995); we add to this ed ;Caracara plancus; agriculture,,mesquiW,, Argen- body of knowledge and report results obtained from tzna; survey. roadside raptor surveyscarried out during December 1998 and January 1999 in the provincesof La Pampa, Roadsidesurveys are useful for assessinghabitat pref- C6rdoba, and San Lugs. erences of diurnal raptors. Although the limitations and STUDY AREA AND METHODS biases inherent in roadside counts are well known (Fuller and Mosher 1987), roadsidesurveys serve as a practical Survey routes extended from Huanchilla, C6rdoba in means for rapidly assessingraptor distribution and abun- the north and Intendente Alvear, La Pampa in the east dance over large areas (Ellis et al. 1990). Roadside sur- to the western border of La Pampa Province,approach- veys have been used to compare speciesrichness and ing the Rio Negro near the city of Neuqu•n in the prov- abundancebetween broad regions and to assessimpacts ince of Neuqu6n, Argentina (ca. 35øS,64øW; Fig. 1). The climate becomes more arid from the eastern coast (Buen- of anthropogenic-habitat transformations on raptors. os Aires) to the mountains of western Argentina, with These typesof surveyshave been carried out in Europe vegetationchanging l?om agriculturalgrasslands to mes- (Meyburg 1973), Al'rica (Cade 1969), North America quite (Prosopisspp.) to desert-scrubgrasslands. We chose (Woffinden and Murphy 1977), Latin America (Ellis et four primary landscapedivisions based on characteristics al 1990), Patagonia (Don•zar et al. 1993), and a grass- of the predominant vegetation type: agriculture, mixed land-agricultural ecosystemin Argentina (Leveau and agriculture/mesquite, mesquite, and desert-scrubgrass- Leveau 2002). The distribution of raptors acrosscentral lands. Argentina was surveyed east to west from Buenos Aires The agriculture category consistedof a mix of cattle ranching and row-crop agriculture,with dominant sum- ruer crops of alfalfa, sunflower, sorghum, and corn. In the agricultural region, forestsand shrubs exist intermit- • Presentaddress: USDA ForestService, Chugach Nation- tently, generally planted as shade areas for cattle, for al Forest, 3301 C Street, Suite 300, Anchorage,AK 99503 wind breaks between fields, and as entrance corridors to U.S.A.; e-mail address:[email protected] estatehouses. These forestsmost frequently consistedof 9 Present address: School of , Plant and Environ- groves of introduced eucalyptus (Eucalyptusspp.) trees. mental Sciences,University of the Witwatersrand,Private The mixed agriculture/mesquitecategory contained 25- Bag 3, Wits 2050, South Al'rica. 75% mesquite,while the mesquitecategory contained JUNE2004 SHORTCOMMUNICATIONS 153

N

Cordoba

Figure 1. Map of roadsidesurvey locations in the westernpampas and central Argentina. 154 SHORT COMMUNICATIONS VOL. 38, NO. 2

Table 1. Number of raptorsobserved in landscapessurveyed in the westernpampas and centralArgentina, Decem- ber 1998-January1999.

HABITAT TYPES

SCIENTIFIC AGRICULTURE/ DESERT SPECIES NAMES AGRICULTURE MESQUITE MESQUITE SCRUB TOTAL AmericanKestrel Falcosparverius 2 10 7 1 20 AplomadoFalcon Falcofemoralis 2 0 2 1 5 White-tailed Kite Elanus leucurus 0 1 0 0 1 ChimangoCaracara Milvagochimango 292 138 38 22 490 CrestedCaracara Caracaraplancus 42 48 63 19 172 Swainson's Hawk Buteoswainsoni 36 1 0 0 37 Red-backedHawk Buteopolyosoma I 1 1 3 6 White-tailed Hawk Buteo albicaudatus 0 3 0 3 6 BurrowingOwl Athenecunicularia 12 1 7 0 20 Short-earedOwl Asioflammeus 1 1 0 0 2 TurkeyVulture Cathartesaura 0 0 0 12 12 BlackVulture Coragypsatratus 0 4 0 1 5 Richness 8 10 6 8 Abundance 388 208 118 62 776 Richness/100 km 1.5 2.6 2.8 6.0 Abundance/100 km 74.9 54.6 54.6 46.6

>75% mesquite with small grassy patches scattered the greatestabundance (•--20 individuals)using a repli- throughout. The desert-scrubgrassland category con- cated goodness-of-fittest (Sokal and Rohlf 1995). We tained <10% tree cover and generally consistedof nat- used replicatedgoodness-of-fit tests to determinewheth- ural . er raptors were distributedin proportion to available We traveledsurvey routes between 0600-1100 H and habitat, or whether individual raptor speciesdeviated 1630-2030 H (local time; on 1 January 1999 sunrisewas from the expectedproportions in the samefashion (i.e., at 0545 H and sunsetat 2010 H). Surveyswere completed whether G for the pooled data, G• and G for the hetero- on 32 routes over 16 d, with a minimum of 50 km and a geneity,GH, were significant.We did not analyzespecies maximum of 248 km per route. Each route wassurveyed with low abundance (<20 observations). once to ensurebird sightingswould be independent.The weather on surveydays was partly cloudyto sunnywith RESULTS either no wind or a slightbreeze. We recorded each oc- currenceof birdsof preyand New World vulturesin each We traveled 518 km through agriculturalhabitat, 381 habitat.We traveledon pavedhighways at approximately km through the mixed agriculture/mesquitehabitat, 216 80-90 km/hr, slightlyfaster than recommended (Fuller km through pure mesquitehabitat, and 133 km through and Mosher 1987), but still at speedsat which we were desert-scrubgrasslands. We counted 12 speciesof raptors able to detect speciesin different habitats,particularly and vulturestotaling 776 individuals(Table 1). Agricul- the five common speciesanalyzed for habitat selection. tural lands had the lowest relative richness and highest To minimize differences in detectability among sur- relative abundance(1.5 species/100km, 74.9 individu- veys,we standardizedtime of d, weather,driving speed, als/100 kin), while desert scrub had the highest relative and number of observers (Fuller and Mosher 1987). In a few cases,we needed to stop the vehicle for positive richnessand lowestrelative abundance (6.0 species/100 identification;during thesetimes we did not includenew km, 46.6 individuals/100 km; Table 1). The raptor ob- observations. Because of time and distance constraints, servationsfor the four habitat typeswere determined to time in each habitat was not uniform. have the followingspecies accumulation curves (Fig. 2): We determined speciesrichness and abundanceby the MMF model (Morgan et al. 1975) for agriculture(y habitat type. We used curve-fitting software (Curve- = [ab + cxd/] [b q- xd]; a = -8.78; b = 0.90; c = 11.03; Expert¸, 1995-2001;Daniel Hyams,Version 1.37) to d = 0.29; SE = 0.45; r = 0.99); the logisticmodel for demonstrate how richness increased with increasing mixed argriculture/mesquite(y= a/(1 + b X e-CX);a = number of km surveyed.CurveExpert¸ usesdouble-pre- 9.51; b = 4.07; c = 0.038; SE -- 0.79; r= 0.97); the Power cisionfloating-point numbers to calculateand rank best- fit curves.In this manner,we describethe rate of species Fit for mesquite(y = axb;a = 0.14; b = 0.68; SE = 0.92; accumulationsin each habitat type and include the cor- r = 0.94); and the rational function model for desert- relation coefficient, r. scrubgrasslands (y = (a + bx)/(1 + cx + dx9);a -- 0.23; We analyzedhabitat preference for the five specieswith b --- 0.83; c -- 0.11; d = -0.0001; SE = 0.26; r = 0.99). JUNE 2004 SHORTCOMMUNICATIONS 155

Agriculture Mixed Agriculture-Mesquite

,oo ,;o" 2/0 0 25 60 76 100 •125 t60 Distance Traveled (kin) Distance Traveled (kin)

0 60 •oo i60 20o 26o •00 o 26 6o 76 100 i2• 16• Dista.ce Traveled(kin) DistanceTraveled (kin)

Figure 2. Speciesaccumulation curves for raptors encountered during roadside surveysof four habitat types in Argentina.

Of the 12 speciesobserved, we recorded 7 infrequent- A replicated goodness-of-fittest on the five most com- ly: Aplomado Falcon (Falcofemoralis), White-tailed Kite mon speciesindicated that raptors were not distributed (Elanus leucurus),Red-backed Hawk (Buteopolyosoma), in proportion to available habitat. The pooled goodness- White-tailed Hawk (B. albicaudatus), Short-eared Owl of-fit test (G e = 43.8; df = 2; P < 0.001) indicated that (Asio flammeus), Turkey Vulture (Cathartesaura), and the raptor community as a whole wasobserved in habitats Black Vulture (Coragypsatratus). Chimango Caracaras in proportions different from those available. In partic- ( Milvago chimango),Crested Caracaras( Caracaraplancus), ular, raptor abundance in agricultural lands exceeded American Kestrels (E sparverius),migratory Swainson's the proportion of available habitat. The heterogeneity Hawks (B. swainsonz),and Burrowing Owls (Athenecuni- goodness-of-fittest (GH = 173.1; df = 12; P < 0.001) cularia) were most common. Chimango Caracaras (N = indicated that individual raptor speciesdid not all deviate 490) were most frequently found in agriculture or mixed from the expected proportions in the same fashion. All agriculture/mesquitem•d CrestedCaracaras (N = 172) individual goodness-of-fittests also were significant (Ta- were most frequently found in mesquite. These two spe- ble 3). American Kestrelsused mesquiteand agriculture/ cies occurred in all habitats and accounted for 85% of mesquite habitats, Crested Caracarasused mesquite, and the total number of individuals sighted. American Kes- Burrowing Owls used agriculture or mesquite. Both Chi- trels (N = 20) were recorded infrequently, but also oc- mango Caracarasand Swainson'sHawks were observed curred in all surveyed habitats (Table 1). Swainsoh's largely on agricultural lands. Hawks and Burrowing Owls were found most commonly DISCUSSION in agriculture. We often fbund Chimango Caracaras, Crested Caracaras,and Swainson'sHawks in groups (Ta- Our resultsshow that the five most commonly encoun- ble 2). tered specieschose habitats differently and not in pro- 156 SHORT COMMUNICATIONS VOL. 38, NO. 2

Table 2. Group sizesof speciesobserved on roadside surveys.

No. INDMDUALS 1 2 3 4 5 6--10 >10

Species American Kestrel 16 2 Aplomado Falcon 1 2 Chimango Caracara 200 61 17 7 7 4 2 Crested Garacara 81 22 7 4 1 Swainson's Hawk 1 1 1 2 Red-backed Hawk 6 White-tailed Hawk 2 2 Burrowing Owl 6 2 Turkey Vulture 8 2 Black Vulture 2

poruon to their availability. Chimango Caracaraswere rent findings,associate Swainson's Hawks with agricultur- more abundant than other specieswe encountered, sim- al lands in Argentina, where they forage opportunistically fiat to findings from other roadsidesurveys in Argentina. on swarmsof grasshoppers(Jaramillo 1993, Goldstein et ChimangoCaracaras accounted for 54% of all observa- al. 1999). Opportunisticfeeders that forage in groupsof- uons in Patagonia(Donftzar et al. 1993) and 74% of all ten are found in associationwith agricultural fields and raptors surveyed in central Argentina (Travaini et al. rangelands(Ellis et al. 1990, Eakle 1994). 1995). Ghimango Garacaras,a poorly studied yet com- Smaller thlcons may be more difficult to detect than mon species,use agricultural areas extensively.This spe- larger , particularly while completing roadsidesur- cies is often found near trees, foraging on snakes, ro- veysat fast speeds.Similarly, our ability to detect falcons dents, birds, and insects,depredating nestlingsand eggs, may change with habitat type. Detectability of small eaung roadkills, other carrion, and refuse near houses perching falcons may be reduced with dense habitat (M. Goldstein unpubl. data). Previous studies, like cur- structure (Fuller and Mosher 1987). Aplomado Falcons

Table 3. Replicated goodness-of-fittest of habitat selectionfor the five most common raptor speciesencountered on road surveysin the western pampas and central Argentina. Speciesnumbers are followed by percent observedin parentheses.

AGRICULTURE/ DESERT SPECIES AGRICULTURE MESQUITE MESQUITE SCRUB TOTAL km 518 381 216 133 1248 (41.5%) (30.5%) (17.3%) (10.7%)

American Kestrel 2 10 7 1 20 12.5' (10%) (50%) (35%) (5%) Burrowing Owl 12 1 7 0 20 15.1' (60%) (5%) (35%) Gh•mangoGaracara 292 138 38 22 490 90.1' (60%) (28%) (8%) (4%) Crested Caracara 42 48 63 19 172 42.7* (24%) (28%) (37%) (11%) Swmnson's Hawk 36 1 0 0 37 56.5* (97%) (3%) Total 384 198 115 42 739 (52.0%) (26.8%) (15.6%) (5.7%) GH = 173.1' Gp = 43.8* Gr • = 216.9' a Gr= Gu + G• * S•gnificantat et < 0.001. JUNE2004 S•40R'rCOMMUNICATIONS 157 and American Kestrels perched on fence posts and anonymous reviewer provided critical input for earher barbed wires may stand out more than when perched on versions of the manuscript. trees in forested habitats, which may have led to under- LITERATURE CITED counting in the latter. However, electrical and phone wires did not exist acrossthe entire sample area. There CADE,T.J. 1969. The status of the peregrine and other were no wires in the desert scrub habitat and wires were in Africa. Pages289-321 inJ.J. H•ckey intermittent acrossother regions.In the two habitatswith [ED.], Peregrine Falcon populations: their biology greater tree structure,mixed argriculture/mesquiteand and decline. Univ. Wisconsin Press, Madison, WI mesquite,we detected more kestrels.In a 35-km section U.S.A. of road with electrical wires, through mixed agricultural DONXZAR,J.A., O. CEBALLOS,A. TRAVAINI,AND F. HIRAt- and mesquitehabitat, we detectedsix American Kestrels DO. 1993. Roadside raptor surveysin the Argentinean on wires. Although we noted no other landscapediffer- Patagonia.J. RaptorRes. 27:106-110. ences (e.g., ridges or valleys),we do not know whether EAKLE,W.L. 1994. A raptor roadside survey in western American Kestrels were more visible on this section of Turkey and Eastern Greece.J. RaptorRes. 28:186-191 road, wires influenced their visibility,or they simply had ELLIS,D.H., R.L. GEINSKI,AND D.G. SMITH. 1990. Raptor greater abundance in this area. road surveysin South America. J. RaptorRes. 24'98- Relative richness was inversely correlated with the 106. number of km traveled. Although we observedthe great- FULLER,M.R. ANDJ.A. MOSHER.1987. Raptor surveytech- est relative richness in desert-scrub habitat and the lowest niques.Pages 37-65 in B. Giron-Pendleton,B. Millsap, relative richnessin agriculture, this dichotomy may have K. Cline, and D. Bird [EDS.], Raptor management been due to unit effbrt (Heyer et al. 1994). In other techniquesmanual. Natl. Wildl. Fed., Washington,DC words, if the rapid species accumulationswe found in U.S.A. these two habitatswere equivocal,then relative richness GOLDSTEIN,M.I., T.E. LACHER,JR., B. WOOD•RIDGE,MJ. was a function of sampling effort and we over-estimated BECHARD, S.B. CANAVEELI, M.E. ZACCAGNINI, G P its value. Nonetheless, our results indicate that common COBB,E.J. SCOLLON,R. TRIBOLET,AND M.J. HOOPER raptors exhibited distinct landscapepretbrences, and for 1999. Monocrotophos-induced mass mortality of the two most common species,Chimango and Crested Swainson'sHawks in Argentina, 1995-96. Ecotox.8. caracaras,it was likely a result of greater foraging oppor- 201-214. tunities in disturbed landscapes. HEYER,W.R., M.A. DONNELLY,R.W. MCDIARMID, L.C. HAY- EK, AND M.S. FOSTER.1994. Measuring and morntot- RESUMEN.--Contamos776 rapacesy buitresa lo largo de ing biologicaldiversity: standard methods for amphib- estudiosal horde de carretera que totalizaron 1248 km ians. Smithsonian Institution Press, Washington, DC durante diciembre de 1998 y Enero de 1999. Viajamos U.S.A. 518 km a trav6sde habitatsagricolas, 381 km a trav6sde JARAMILLO,A.P. 1993. Wintering Swainson'sHawks in Ar- un habitat arbustivo mixto de cultivos agricolasy mes- gentina: food and age segregation.Condor 95:475- 479. quite (Prosopisspp.), 216 km a trav6sde habitat de puro mesquite,y 133 km a travis de desiertoarbustivo en las LEWAU, L.M. ANDC.M. LEVEAU.2002. Uso de h•tbitat pot pampasoccidentales y el centro de Argentina. De las 10 aves rapaces en un agroecosistemapampeano. Hot- nero 17:9-15. especiesobservadas, los caracaraschimango (Milvagochi- mango)y los Caracarascrestados (Caracara plancus) ocur- MEYgUR•, B.-U. 1973. Observations sur l'abondance rel- tieton a lo largo de todas las rutas estudiadasen todos ative des rapaces (Falconiformes) dans le nord ct los habitatsy fueron los m'•s comunes.Las tierras agri- l'ouest de l'Espagne. Ardeola19:129-150. colas tuvieron la m•s baja riqueza relativa y la m'•s alta MORGAN,P.H., L.P. MERCER,AND N.W. FLOmN. 1975. Gen- abundancia relativa, mientras que el desierto arbustivo eral model for nutritional responsesof higher organ- tuvo la m'•s alta riqueza relativay la m'•sbaja abundancia isms. Proc. Natl. Acad. Sci. 72:4327-4331. relativa. Un test replicado de bondad de ajuste para las SOKaL,R.R. ANDF.J. ROHLF. 1995. Biometry, 3rd Ed. W H cinco especiesmas comunes indic6 que las especiesno Freeman, NY U.S.A. estuvieron distribuidas en proporci6n al habitat disponi- TP-AVA•NLA., A. RODRigUEZ,O. CEt•ALLOS,J.A. DONXZA}•, ble (Gp = 43.8; P < 0.001) y diferentes especiesmostra- ANDF. HIRALDO.1995. Roadside raptor surveysin cen- ton preferencia pot difbrentes h•bitats. tral Argentina. Hornero14:64-66. [Traducci6n de C6sar M•rquez] WOnqNDEN,N.D. ANt)J.R. MURPHY.1977. A roadsiderap- tor censusin the eastern Great Basin 1973-74. Raptor ACKNOWLEDGMENTS Res. 11:62-66. We thank A. Lanusseand M. Bechardfor logisticalsup- port and field assistancewhile in La Pampa. T. Lacher, Received 3 February 2003; accepted 30 November 2003 M. Cotson, K. Kosciuch,J.J. Negro, M. Carfete, and an AssociateEditor: Juan jos6 Negro