J. Field Ornithol.,63(3):324-330

ANTI-PREDATOR BEHAVIOR OF BREEDING EURASIAN

ODD W. JAGOBSENAND MORTENUGELVIK Departmentof Ecology Museumof Zoology Universityof Bergen N-5007 Bergen,Norway

Abstract.--Anti-predator behaviorwas observedduring a study(1983-1985) of the feeding ecologyand social behavior of breeding Eurasian Wigeon (Ariaspenelope L.) in western Norway. When (Vulpesvulpes) appeared, wigeon and other dabblingducks con- gregatedin a group, becamealert, quackedloudly and followedthe predator'smovement. In contrast,as avian predators appeared, wigeon becamealert, retreated into emergent vegetation,or laid down on the shorewith their headsstretched out, and remainedmotionless until thepredator passed. Wigeon responded to nestpredators, e.g., Common Raven (Corvus corax) and Hooded Crow (C. coronecornix) by assumingalert posturesin 84% and 60% of the occasionsin terrestrialand aquatichabitats, respectively. The observeddifference between habitatswas probablydue to the ' reducedability to detectpredators and escapeattacks in terrestrial environments.Foraging wigeon respondedto the presenceof avian nestpred- ators more by assumingalert posturesduring the egg-layingand incubationperiods than during the prelaying period. Wigeon becamealert in 78% (terrestrial habitats) and 66% (aquatic habitats) of the occasionswhen warning calls were given by four other species. They respondedmore to warning calls during the egg-layingand incubationperiods than during the prelaying period. It is concludedthat wigeonhave evolveddifferent strategiesto escapeattack from avian and ground predators.Wigeon may also benefit from warning calls given by other species,because warning calls may allow enhancedpredator detection and increasedforaging efficiency.

CONDUCTA ANTI-PREDATORA POR PARTE DE INDIVIDUOS REPRODUCTIVOS DE ANAS PENELOPE Sinopsis.--Durante un estudioque se 11ev6a caboentre 1983-1985 en el oestede Noruega, sobrela ecologiaalimentaria y la conductasocial de individuosreproductivos de la Europea (Ariaspenelope), se observ6conducta anti-predatora pot parte del ave. Cuando un zorro (Vulpesvulpes) aparecla, las Marecas, junto a otros patos se congregaban,asumlan posturasde alerta, vocalizabanaltamente y seguianlos movimientosdel depredador.En contrastecuando aparecian aves predatoras,las Marecas asumian posturasde alerta, se retiraban a la vegetaci6nemergente o se acostabanen la playa con su cuelloestirado en lo que pasaba el predator. Las Marecas respondierona depredadoresde nidos (Ej. Corvus corax,C. coronecornix) asumiendo posturas de alerta en el 84% y 60% de las vecesen habitats terrestresy acufiticos,respectivamente. Las diferenciasobservadas entre habitats se debieron probablementea la reducida habilidad en las aves para detectar predatoresy escaparde •stosen habitatsterrestres. Individuos que forrajeabanrespondieron, con mayor frecuencia, a la presenciade depredadoresde nidosasumiendo posturas de alerta duranteel perlodo de incubaci6nque previo a la puestade huevos.Las Marecas asumieronposturas de alerta en el 78% de los casosen habitatsterrestres yen el 66% de los casosen habitatsacufiticos, cuandootras cuatro especies de patosdieton 11amadasde alerta. Estosrespondieron mils a estas11amadas de alerta durante el periodode puestade huevose incubaci6nque durante el periodode pre-puesta.Se concluyeque Ariaspenelope ha desarrolladoa travis de la evoluci6ndiferentes estrategias para escaparal ataquede predatoresa•reos y terrestres.Las avespoddan tambien beneficiarse de 11amadasde alerta de otrasespecies, debido a que estas 11amadaspodrlan facilitar la detecci6nde predatorese incrementarla eficienciade forrajeo.

324 Vol.63, No. 3 Anti-predatorBehavior of Wigeon [325

Predationis a severeselection pressure on many ,and the ability to recognizepredators and avoid predation are critical elementsin the life history strategyof most birds. Consequently,they have evolveddif- ferent behavioraladaptations to avoid predation when primary defenses (e.g., crypsis)fail (Curio 1976). Anti-predator behaviorsare broadly dividedinto a) escape,b) defense and c) alarm signals(Curio 1976). To gain protectionagainst predators, somebirds associatewith other speciesof birds that are more vigilant (Byrkjedal 1987, Byrkjedal and Killils 1983), or larger and more ag- gressiveagainst predatorsthan themselves(Dyrcz et al. 1981, Fuchs 1977). Birds avoid predation through the use of auditory signals.These signalsindicate knowledgeof a predator'spresence and may be intended for the predatoritself or as a warning to others(Buchanan 1989). Waterfowl, such as White-fronted Goose (Anser albifrons) (Lazarus 1978) andGreen-winged Teal (Ariascrecca) (Pi3ys//1987, Tamisier 1974), often are disturbedby potential predatorswhile feeding.To avoid pre- dation, individuals scan their surroundingsbetween feeding bouts and oftendepart when a potentialpredator approaches (Pi3ys5 1987). Surveys on nonhuntingmortality of waterfowl in North America have shownthat 70% of predationwas by mammalsand the remainder by birds, and that the vast majority of depredatedbirds were dabbling ducks,with only a few geeseand swansbeing taken (Stoudt and Cornwell 1976). We collecteddata on anti-predatorbehavior of breedingEurasian Wi- geon (Ariaspenelope) during a 3-yr studyon feedingecology and social behavior of this dabbling duck in western Norway. Common predators on adult wigeonand other dabblingducks in our studyarea were Goshawk (Accipitergentilis)and Red Fox (Vulpesvulpes) (J. Stenersen,pers. comm.). Wigeon nestswere mainly preyed upon by Red Fox (own obs.), but HoodedCrow (Corvuscorone cornix) and CommonRaven (C. corax)may also have been important predatorson nestsand young (Erikstad et al. 1982, Parker 1984, Skarph•dinssonet al. 1990). On their wintering groundsin Britain, wigeonprefer feedingsites close to water, which has been interpretedas an anti-predatorresponse (May- hew and Houston 1989). We are unaware of any studythat has examined the behavioralresponse of wigeonto arian and groundpredators, however. In this paper, we describestrategies employed by wigeon to escapearian or groundpredators, and their behaviorin the presenceof nestpredators in terrestrial and aquatic habitats during the breedingseason.

STUDY AREA AND METHODS We observedbreeding wigeon from Apr. through Jun. 1983-1985 in Sveio (59ø33'N, 5ø21'E), western Norway. The study area has a fairly flat terrain approximately 30 m abovesea level, consistingof farms and moorlandwith heather (Callunavulgaris), juniper (Juniperuscommunis), and scatteredstands of Scotspine (Pinussylvestris), Norway spruce(Picea abies), and birch (Betulapubescerts). Observations were made at three lakes: 1) Lake Bjellandsvatn(10.9 ha), a shallow eutrophiclake with 326] O. w. Jacobsenand M. Ugelvik J.Field Ornithol. Summer 1992 well-developedstands of emergentvegetation, e.g., Carex rostrata,Equi- seturnfiuviatile and Phragrnitescornrnunis, and floating-leavedvegetation, e.g.,Potarnogeton natans, P. pusillus,Nymphaea alba and Glyceriafiuitans; 2) Lake Mannavatn (19.2 ha), a shallowand oligotrophiclake with sparse standsof C. rostrata,E. fiuviatile,and P. natans;and 3) Lakefi, sevatn (13.0 ha), a shallowoligotrophic lake surroundedby agriculturalareas and plantationsof Norway spruce.Emergent vegetationis sparselyde- veloped,but standsof C. rostrataand E. fiuviatile occurin secludedbays. The breeding population of wigeon in the area varied between 30 and 40 pairs since1983 (Jacobsenand Ugelvik 1990). Data were gatheredfrom five wigeonpairs in 1983, three pairs in 1984 and four pairs in 1985. We dividedthe breedingseason in to the prelaying period, which was definedas the period of dispersalfrom small arriving flocks(3-6 pairs) in early April until the day prior to laying the first egg. The egg-layingperiod was from the day the first egg was laid until clutchcompletion. Incubation included the period from clutchcompletion until hatching. During incubation,females spent 1-2 h a day away from the nest. Due to heavy nest predation by fox during incubation, we obtained few data from this period. We have thus combineddata from the egg-layingand incubationperiods in our analyses. Although none of the pairs was individually marked, we recognized eachbird by individualplumage characteristics. Males and femalesshow great variability in characteristicson the flanks, head and back (Cramp and Simmons1977). To minimize disturbance,we observedindividual pairs at distances approaching200 m usingbinoculars (8 x 40) and a spottingscope (25- 40 x). The behaviorof wigeon was recordedevery minute and assigned to the main activitycategories based on Afton (1979). All behavioraldata were recordedon a portable tape-recorder.We recordedwhether wigeon becamealert (head-up posture) when avian and ground predatorsap- peared both in terrestrial (agricultural areas closeto water and shore) and aquatic (openwater, emergentvegetation) habitats. Furthermore, we recordedthe responsesof wigeonto warning calls given by nearby Eur- asianCurlew (Nurneniusarquata ), Oystercatcher(Harnaetopus ostralegus), Common Redshank(Tringa totanus),and Mew Gull (Larus canus).

RESULTS AND DISCUSSION One morning in May 1985, we observeda Red Fox at the shoreof Lake Mannavatn.A wigeonpair grazingon the pastureimmediately becamealert whenthe fox wasabout 100 m fromthem. They thenflew into openwater and swamparallel to the fox for approximately1 min at a distanceof 30 in as it was walking alongthe shore. On four occasionsduring the periodmid May-early Jun. 1983 we observeda fox huntingat Lake Bjellandsvatn.It systematicallychecked bushesand heather.When wigeon(and other dabblingducks) detected the fox, they congregatedin a small denseflock (up to 12 individuals), Vol.63, No. 3 Anti-predatorBehavior ofWigeon [327 swam in open water parallel to the predator at a distanceof 20-30 m from the shoreand quackedloudly. When the fox disappeared,the flock dispersedinto pairs and singlemales. Wigeon and other ducksprobably follow the predator'smovement to: 1) keep track of the fox; 2) signalto the fox that it already is detected;and 3) divert the predator'sattention, thus reducing its successat finding nests. On one occasion(early Jun. 1983), a fox jumped into a dense(Carex) bed in Lake Bjellandsvatnprobably to catchhalf-grown (Arias platyrhynchos)ducklings foraging intensively on insectson the vegetation. The ducklingsrushed out of the vegetationimmediately after the unsuc- cessfulattack, and disappearedinto reed beds. Twice we observeda female Goshawk flying over the studylakes. In the first caseat Lake Mannavatn in 1983, onewigeon pair was grazingon cultivatedcrops close to water. When the Goshawk appeared,the wigeon pair immediately raised their heads in alert position, rushed from the cultivatedcrops to the shore and laid down with their necksand heads stretchedout. The birds were motionlessand well camouflagedamong stonesand blockson the shorewhere they remained for approximately 1 min, long after the Goshawk passed.At Lake Bjellandsvatnin 1984, a wigeonpair wasforaging in shallowwater containingscattered emergent vegetation.When the Goshawk appeared,the male immediatelystopped foraging,laid his head down to the water surfaceand becamemotionless. The female first performed a wings-up display closeto her mate, after which the pair disappearedinto denseemergent vegetation. When a pair of Golden Eagles (Aquila chrysaetos)circled high over Lake Mannavatn (early May 1984), onewigeon pair immediatelyrushed from cultivatedcrops, where they grazed about 10 m from water edge, to a rocky shoreand laid down with their headsand necksstretched out. The birds were motionlessfor approximately2 min. The samebehavior wasobserved in anotherwigeon pair at Lake,•sevatn (Apr. 1983)as a Rough-leggedHawk (Buteolagopus) was detectedhigh in the air. In both cases,the ducksremained motionlessuntil the raptor disappeared. In Jun. 1984, we observedan Eagle Owl (Bubo bubo) passingover Lake Bjellandsvatnearly in the morning. Upon its detection,wigeon (one pair and one singlemale), Green-wingedTeals (Ariascrecca) (two single males and one female with about 2-wk-old ducklings),and (three males plus one female with half-grown ducklings)immediately rushedinto denseemergent vegetation and disappeared. Dabbling ducksmay dive if attackedby an avian predator. In May 1985, we observeda Goshawk pursuing a male Green-wingedTeal in the air. The Green-winged Teal dropped into the lake from a height of 10 m and immediately dived to escapethe attack. Although we have not observedpredator attackson wigeon, our time budgetdata revealedthat wigeonwere more alert in terrestrial than in aquatic habitats (28%, n = 1250 min obs.vs. 17%, n = 1770) (x 2 = 52.17, df = 1, P < 0.001). The most plausible explanation is that predation risk increaseswhen the birdsmove away from water, where it takeslonger 328] O. W. Jacobsenand M. Ugelvik J. Field Ornithol. Summer 1992

TABLE 1. Percent of time Eurasian Wigeon became alert after potential nest predators appearedin terrestrialand aquatichabitats during prelayingand egg-laying/incubation periods.n = samplesize. NS = not significant. TerrestrialFisher's Aquatic Fisher's habitats exact habitats exact proba- proba- Species Periods % n bility % n bility

19 CommonRaven Prelaying 90.9 11 NS 63.1 NS (Corvuscorax) Egg-laying/ 100.0 9 100.0 4 incubation 4O HoodedCrow Prelaying 71.8 39 0.037 47.5 0.024 (C. coronecornix) Egg-laying/ 92.3 26 83.3 12 incubation Total 83.5 85 60.0 75

to seek concealmentand escapea potential attack (see Mayhew and Houston 1989). When CommonRavens and HoodedCrows, both pred- atorson nestsor youngbut not on adults,approached, the ducksbecame alert in 83.5% (n = 85) of the occasionsin terrestrial habitats, and in 60.0% (n = 75) of the occasionsin aquatichabitats (X 2 = 11.06, df = 1, P < 0.001) (Table 1). We suggestthat the responseof assumingalert posturesto the presenceof HoodedCrows and CommonRavens is partly becausethey resemblepotential predatorson adults. Wigeon becameproportionally more alert each time potential nest predatorsapproached during the egg-layingand incubationperiods than during the prelayingperiod in both terrestrial(X 2 = 5.00, df = 1, P < 0.05) and aquatic habitats (X2 = 6.81, df = 1, P • 0.01) (Table 1). Wigeon frequentlyreacted to warning callsof Eurasian Curlew, Oys- tercatcher,Common Redshank and Mew Gull (Fig. 1). Wigeon respond- ed by assumingalert postures(78.3%, n = 115), or flying from cultivated areasto openwater where theyretreated into emergentvegetation (15.6%, n = 18) after warning calls were given in terrestrial habitats. Wigeon becamealert (66.3%, n -- 89), or swam from open water and retreated into emergentvegetation (15.7%, n = 14), after warning callswere given in aquatic habitats. The ducks tend to assume alert posturesmore in terrestrial habitatsthan in aquatic habitats (X2 = 3.65, df = 1, ? • 0.1). Studieson Green-wingedTeals in Finland alsohave confirmed that this dabbling duck respondedto warning calls given by Black-headed Gulls (Larusridibundus) (Pi3ys5 1988). Wigeon assumedalert posturessignificantly more often after alarm callswere givenduring the egg-layingand incubationperiods than during the prelayingperiods (X 2 = 5.53, df = 1, P • 0.05 [terrestrialhabitats] and X2 = 8.54, df = 1, P • 0.01 [aquatichabitats]). Wigeon were more likely to fly from terrestrial habitatsthan from aquatic habitatswhen warning callswere given(X 2 = 12.54, df = 1, ? < 0.001). Vol.63, No. 3 Anti-preclatorBehavior ofWigeon [329

NS 100 - 42 NS I-• Terrestrial 34 [71 Aquatic

80- 28 NS I.l.I

Z 13 17 60- NS

26 18

40-

Ll.I 20-

Curlew Oystercatcher Redshank Mew Gull SPECIES FIGURE1. Percent of time Eurasian Wigeon becamealert after warning calls were given by Eurasian Curlew, Oystercatcher,Common Redshankand Mew Gulls in terrestrial andaquatic habitats. Numbers above columns refer to samplesize. NS = not significant.

Staying near warning birds of other speciesmay be advantageous, becausealarm callsmay allow wigeonto detectavian and groundpredators earlier and thereby reducepredation risk. Similar argumentshave been outlined by Byrkjedal and K•lls (1983) and Metcalfe (1984). If the potentialprey is alertedin an early phaseduring an attack,the predator's chanceof successis low (Kenward 1978). Wigeon spent 30.1% of the time alert when they were alone, and 21.8% of the time alert when other specieswere presenton agriculture.Grazing boutsaveraged 15.6 s (n = 120) when they were alone and 20.0 s (n = 144) when other birds were present.Furthermore, peck rates per min averaged45.7 (n = 114) when they grazed alone and 56.1 (n = 141) when they grazed in association with other specieson agriculture.Feeding efficiency is enhancedfor sev- eral specieswhen they associatewith other species(Barnard and Stephens1983, Byrkjedal 1987, Byrkjedaland Killils 1983).

ACKNOWLEDGMENTS

John Torvund and John Aambokindly gaveus permissionto work on their agricultural areasduring the study.Our sincerethanks go to Tex A. Sordahlfor improvingthe manuscript with hiscomments and for improvingthe English.Ingvar Byrkjedal, Tore Larsen,Hermund 330] O. W.Jacobsen and M. Ugelvik j. FieldOrnithol. Summer 1992

Mjelstad and Torstein Solhoymade valuable comments on an earlier draft of this manuscript. We thank Alan D. Maccaroneand an anonymousreviewer for their helpful suggestionsto improvethe quality of the manuscript.Financial assistancewas providedby the University of Bergen and the Norwegian Directorate for Protectionand Management of Nature.

LITERATURE CITED

AFTON,A.D. 1979. Time budgetof breedingNorthern Shovelers.Wilson Bull. 91:42-49. BARNARD,C. J., AND H. STEPHENS.1983. Costsand benefitsof single and mixed species flocking in Fieldfares Turduspilaris and Redwings T. iliacus.Behaviour 84:91-123. BUCHANAN,J. B. 1989. Alarm calls, habituation and falcon predation on shorebirds. Wader Study Group Bull. 55:26-29. BYRKJED^L,I. 1987. Short-billedDowitchers associate closely with LesserGolden Plovers. Wilson Bull. 99:494-495. •, AND J. A. K•L•S. 1983. Plover's Page turns into Plover's Parasite: a look at the Dunlin/Golden Plover association.Ornis Fenn. 60:10-15. CR•MP, S., AND K. E. L. SXMMONS(eds.). 1977. Handbook of the birds of Europe, the Middle East and North Africa. Vol. 1. Oxford Univ. Press,Oxford, United Kingdo•n. 722 pp. CuRXO,E. 1976. The ethologyof predation.Springer-Verlag, Berlin, Germany. 250 pp. DYRCZ,A., J. WITKOWSKI,AND J. OKULEWICZ.1981. Nesting of "timed" wadersin the vicinity of "bold" onesas an antipredatoradaptation. Ibis 123:542-545. ERIKSTAD, K. E., R. BLOM, AND S. MYRBERGET. 1982. Territorial Hooded Crows as predatorson Willow Ptarmigan nests.J. Wildl. Manage. 46:109-114. FUCHS,E. 1977. Predationand anti-predatorbehaviour in a mixed colonyof ternsSterna sp. and Black-headedGulls Larus ridibunduswith special referenceto the Sandwich Tern Sterna sandvicensis. Ornis Scand. 8:17-32. JACOBSEN,O. W., ANDM. UGELVIK. 1990. The wigeonin westernNorway. Jakt og Fiske 119:28-29. KENWARD,R. E. 1978. Hawk and doves:factors affecting success and selectionin Gos- hawks' attackson Wood-pigeons.J. Anim. Ecol. 47:449-460. LAZARUS,J. 1978. Vigilance, flock size and domain of danger size in the White-fronted Goose. Wildfowl 29:135-145. MAYHEW,P., AND D. HOUSTON. 1989. Feeding site selectionby wigeonArias penelope in relation to water. Ibis 131:1-8. METCALFE,N.B. 1984. The effectsof mixed-speciesflocking on the vigilance of shore- birds: why do they trust? Anim. Behav. 32:986-993. PARKER,H. 1984. Effect of corvid removal on reproductionof Willow Ptarmigan and . J. Wildl. Manage. 48:1197-1205. P6YSX,H. 1987. Feeding-vigilancetrade-off in the Teal (Anas crecca):effects of feeding method and predation risk. Behaviour 103:108-122. . 1988. Do foragingTeals exploit gulls as early warners?Ornis Scand.19:70-72. SKARPHI•DINSSON,K. H., O.K. NIELSEN,S. THORISSON,S. THORSTENSEN,AND S. A. TEMPLE. 1990. Breedingbiology, movements, and persecutionof ravensin Iceland. Acta Natur. Islandica 33:1-45. STOUDT,I. J., AND G. W. CORNWELL.1976. Nonhuntingmortality of fledgedNorth American waterfowl. J. Wildl. Manage. 40:681-693. T^MISIER,A. 1974. Etho-ecologicalstudies of Teal wintering in the Camargue (Rhone Delta, France). Wildfowl 25:123-133. Received25 Jun. 1991; accepted23 Oct. 1991.