j RaptorRes. 33(2):123-127 ¸ 1999 The Raptor ResearchFoundation, Inc.

COOPERATIVE HUNTING OF JACKDAWS BY THE LANNER FALCON (FALCO BIARMICUS)

GIOVANNI LEONARDI 1 Avian Scienceand ConservationCentre, Macdonald Campus of McGill University,21,111 LakeshoreRoad, Ste.Anne de Bellevue,Quebec; H9X 3V9 Canada

ABST•CT.--Cooperative hunting has been recorded for severalsubspecies of Lanner Falcon (Falcobiar- micus).On average,the successrate for pairs is higher than for singlebirds. During 1988-90, I collected data on the successof five Lanner Falcon pairs that cooperativelyhunted Jackdaws(Corvus monedula) in western Sicily. Fifty-three percent of attackswere aimed at larger groups of Jackdaws.Males made most of the initial attacks(74%) but prey captureswere mainly made by females(87%). Pairstended not to share prey and used visualcontact to coordinate chases.Most attackswere by partial surprise (60.8%), followed by nonsurprise(21.6%), and surpriseattacks (17.6%). Surpriseattacks tended to involvesmall flocks of Jackdaws,whereas partial surprisetended to involvelarge flocks. KEYWOrlDS: LannerFalcon; Falco biarmicus; cooperative hunting;, Sicily; Jackdaw; prey group size.

Caza cooperativade Corvusmonedula por Falcobiarmicus RESUMEN.--Lacaza cooperativade Falcobiarmicus ha sido registradapara variassubespecies a lo largo de su distribuci0n. En promedio la tasade •xito por pareja es mas alta que la individual. Durante 1988- 90 recolect• datos sobre el •xito de la caza cooperativade Corvusmonedula de cinco parejas en el oestede Sicilia. Cincuenta y tres porciento de los ataquesfueron dirigidos a grandesgrupos de Corvus monedula.Los machosefectuaron los ataquesiniciales (74%) pero la captura de presasfue efectuada por las hembras(87%). Lasparejas tend•an a no compartirla presay utilizaroncontactos visuales para coordinar las persecusiones(60.8%), seguidasde ataquessin sorpresa(21.6%), y ataquessorpresivos (17.6%). Los ataquespor sorpresainvolucraron pequefias parvadas de Corvusmonedula, mientras que los parcialmente sorpresivosinvolucraron parvadasmas grandes. [Traducci6n de C•sar M•rquez]

Cooperativehunting is a socialforaging behav- Hector 1986). Southern Mediterranean Peregrine ior where predatorscoordinate their movementsto Falcons (F. peregrinusbrookei) hunt cooperativelyin increase efficiency of capture (Ellis et al. 1993). areaswhere prey densityis low (Thiollay 1988). Prey may be sharedamong members according to Cooperativehunting in Lanner Falconshas been social organization, prey size and individual func- recorded for several subspeciesthroughout the tional role (Bednarz 1988, Ellis et al. 1993). Pair species'geographic range (Cramp and Simmons hunting is cooperativewhen participantsperform 1980, Tarboton and Allan 1984, Leonardi et al. separateroles. In certain speciesand under certain 1992). Lanner Falconpairs pursue swift flying prey circumstances,cooperative hunting is more suc- (e.g., swifts[Apus spp.]) along parallel paths (Mir- cessfulthan solitaryforaging (Hector 1986, Thiol- abelli 1982, Bijlsma1990). They hunt flocksof gre- lay 1988, Yosef 1991, Ellis et al. 1993). garioussmall birds (e.g., swallows[Hirundo spp.]) Cooperativehunting in the genusFalco seems to working together with repeated stoopsupon the be restrictedto bird-eatingspecies, such as Lanner same individual (Mirabelli 1982). In contrast, for Falcons (Falcobiarmicus), Aplomado Falcons (F.fe- larger perched prey (e.g., shorebirdsand pigeons moralis), and Red-headed Falcons (F. chicquera), [ Columbasp.] ), one falcon flushesthe quarry while which inhabit semi-opensavannas and desert and it is taken by the mate (Mebs 1959, Massa et al. Mediterranean scrub (Mebs 1959, Osborne 1981, 1991, Yosef 1988). Partners have distinct roles. Malesusually attack and direct prey towardfemales Presentaddress: Via SantageloFulci, 28, 1-95127 Ca- (Yosef1991) and femalestend to pursuelarge prey tania, Italy. (Brossett 1961, Tree 1963, Kemp 1993). Success

123 124 LEOtqAR• VOL. 33, NO. 2 rates when hunting in pairs (20-25%) are higher tempted to flush prey (Mebs 1959, Massa et al. 1991, than that of single birds hunting alone (15-40%) Kemp 1993, Jenkins 1995). I observedLanner Falconscooperatively hunting both (Bijlsma 1990, Yosef1991, Kemp 1993). Rock Pigeons (Columbalivia) and Jackdaw flocks near This paper describesmy observationsof coop- their nests.Both prey speciesnest on cliffs 100-300 m erative hunting in Lanner Falconsnesting in west- from Lanner Falcon nests (Sodhi et al. 1990, Suhonen et ern Sicily.In this region, pairsfrequently attack co- al. 1994). For evaluating the importance of cooperative hunting, I only investigatedhunts of Jackdaws.Prelimi- lonial nesting Jackdaws (Corvus monedula).This nary observations indicated that single female Lanner provided an opportunity to compare successrates Falconsinitiated nearly all pursuitsof pigeons.Also, <5% among attacks on different sized flocks, as it relat- of the total attempts on pigeons (N = 32) were per- ed to sex of pursuersand attack strategiesutilized. formed by males. Cooperativehunting, and necessarily, participation by males, was more common in hunts of Jackdaws.In addition,Jackdaws consistently comprised a I studied Lanner Falconson the island of Sicily in the large percentageof dietary biomassfor lanners in Sicily (Massa et al. 1991, Leonardi et al. 1992, Leonardi 1994) central Mediterranean. I observed five pairs during the breeding season:two breeding pairs near the northern Finally,Jackdaws responded to attackingfalcons with in- periphery of the Sicilian distribution and three pairs in tricate forms of mobbing behavior. This provided an op- a southern area where the specieswas studied previously portunity to investigateinteractions between cooperative by Mascara (1986). hunting and antipredator defense behavior (Kenward 1978, Caraco et al. 1980, Turner and Pitcher 1986, Cres- The climate of the northern study area is temperate- swell 1994). wet with 600-800 mm of rainfall and an average annual temperature of 12-14øC.The southernstudy area has a The number of Jackdawspresent was estimateddaily subarid climate (<600 mm of rainfall and temperature by counting the maximum number of birds seen simul- >16øC)(Instituto Geografico De Agostini1987). taneously.Jackdaw colonies typically contained 20-70 in- Land use in the studyareas was predominately farming dividuals. During an attack, I estimated the size of each and pasture. Cereal farming and pasturelandscovered by flock attacked by assumingthe members to be all birds olive ( Olea europaea)and prickly pear ( Opuntia ficus- within 25 m of each other (Cresswell 1994, 1996). At •ndica) cultivation dominated northern open spaces.The times entire colonies behaved as a single flock. Under these circumstances, I counted the number of individuals southern study area was largely in a wheat monoculture with interposingspots of xeric Mediterranean vegetation in the group first attacked (Kenward 1978). For statistical and smallEucalyptus plantations. Within both studyareas, comparisons,I placed Jackdaw flocks into three size cat- lanners nested on clay-sandand calcareouscliffs with egories according to previous studies of predation on heights of 50-1150 m (Massaet al. 1991). prey groups(Kenward 1978, Cresswell1994, 1996): 2-10, 11-30 and 31-50 individuals.I assessedthe validityof the METHODS aboveflock sizeclasses for this studythrough preliminary observationsof flocking reactions measured for single I visited breeding sites 21 times during two prerepro- and paired Lanner Falcons (Leonardi 1991, Leonardi un- ductive periods (November-January 1988-90). Each publ. data). breeding site was visited 10 times for 55 total H. I I compared F-frequencies of hunting strategiesand watched Lanner Falconshunting in pairs from 200-600 successrates among different flock sizes and strategies m with 8 X 40 and 10 X 40 binoculars.Age and sex of using chi-squared tests and C-tests (Zar 1984). I used obsex'vedfalcons was recorded for each sighting accord- Cochran's corrected chi-squaretest for differencesbe- ing to criteria in Cramp and Simmons (1980) and Porter tween malesand femalesusing a 2 X 2 contingencytable et al. (1981). (Zar 1984). Attackswere defined asvery rapid flights or stoopsto- ward one or more clearly observedprey (an individual RESULTS or group of specificprey species)(Cresswell 1994, 1996). First attackswere defined as the first, fast approach by In 52 cooperativehunts, I detected no vocaliza- falcons toward potential prey. During each attack, I re- corded the following data: position and sex of each fal- tions which might have functioned to coordinate con at the start of the attack, size of the prey flock, and pursuits.Females alone ate 70% of prey captured type of attack strategy.I placed attackstrategies into three in cooperativehunts (capturesN = 10). In only 2 categories:surprise attacks,partial surprise attacks, and of 16 cases(12%), malesfed on prey capturedin nonsurprise attacks. In surprise attacks,Lanner Falcons cooperativehunts after the departure of females. first approached close to Jackdawsfrom behind rock cliffs. In partial surpriseattacks, one of the two attacking Although Lanner Falcons preferred to attack larg- falconswas visible to prey while the other falcon attacked er flocks (Table 1; X2 = 12.33, df = 2; P < 0.001), by surprise. In partial surpriseattacks, two perched fal- hunting successwas inverselyproportional to flock cons would depart at different times (Yosef 1991, Kemp size (G = 10.7, df = 2; P < 0.005). 1993). In nonsurprise attacks,both falcons were visible Female Lanner Falcons initiated attacks less of- at the onsetof attacks,then they tried to encircleJackdaw flocks (Cresswell1994, 1996). In nonsurpriseattacks, one ten than did males (26% vs. 74%). Althoughmales falcon stooped on prey after soaringwhile the other at- preferred to pursue larger prey (87% of 52 pur- JUNE 1999 COOPERATIVEHUNTtNG BY LANNER FALCONS 125

Table 1. Distribution of Lanner Falcon attacks on flocks Table 2. Capture success(%) from first attacksof Lan- of Jackdawby cooperativehunting in Sicily. net Falconscooperatively hunting flocks of Jackdawsin Sicily. JACKDAWFLOCK SIZE CLASS JACKDAWFLOCK SIZE CLAss 2--10 11--30 31--50 2-10 11-30 31-50 Total Attempts 10 14 27 lidIls 2 5 9 Males Total 12 19 36 First attack 2 12 24 38 Captures (%) 20 35.7 33.3 Kills 0 1 1 2 Captures (%) 0 0.08 0.04 Females suits) and larger flocks more often than did fe- First attack 8 2 3 13 males (Table 2; G = 13.9; df = 2; P < 0.001), male Kills 2 5 7 14 hunting successrates tended to be lower than Captures (%) 0.25 2.5 2.3 those of females (Table 2; X2 = 2.86, df = 2; P < 0.10). Lanners attacked Jackdawsby partial surprise ence. In partial surpriseattacks, flying Lanner Fal- (60.8%) much more frequently than they did by con males from outside the colony area would nonsurprise(21.6%) and surpriseattacks (17.6%) suddenlystoop on Jackdaws. (N = 52). Although degree of surprise is one of Prey capture percentage of this studywas lower the most important factors in improving the suc- (31%) than that observedfor other lanner subspe- cessof raptor attacks,lanners used this technique cies(50%; Yosef1991, Kemp 1993) and Aplomado in only nine of 52 attempts.Also, open attacksgive Falcons(45%; Hector 1986). Sicilianlanners pur- time for antipredatorybehavior by prey. Neverthe- sued small- and medium-sizedprey with solitary less,partial surprisewas used significantlymore of- hunting strategiesand used cooperativehunting ten (X2 = 17.40, df = 2; P < 0.001). In addition, for large-sizeprey like Jackdaws.Nevertheless, this lanner pairs capturedmore prey usingnonsurprise low percentagemay have been due to Jackdawan- attacks (Table 3; Xs = 11.90, df = 2; P < 0.01). tipredator behavior.Large Jackdaw flocks frequent- Cooperative hunting techniques were not uni- ly used mobbing (43%, N = 58) againstlanners. formly distributed among prey flock classes.Hunt- This active defense, combined with the dilution ef- ing successin relation to prey flock sizewas signif- fect of individualsin a flock, can improve predator icant for partial surpriseon larger groups(22%; avoidanceby prey. The dilution effect is an advan- Table 3; P < 0.01) and nonsurpriseattacks on me- tage becauseindividuals are lesslikely to be taken dium flocks (27%; Xs = 14.40, df = 2; P < 0.01). by predatorswhen in a flock (Turner and Pitcher 1986). Morgan and Godin (1985) reported that DISCUSSION the rate of predator attack per individual prey is Evidence of coordinative signaling among hunt- inverselyproportional to group size. ing predatorsis indicativethat hunts are coopera- Although examples of role reversal are known tive (Hector 1986, Ellis et al. 1993). Male Aplo- (Mebs 1959, Mirabelli 1982, Massa et al. 1991), the mado Falcons initiate attacks and then vocalize a "chip" call (Keddy-Hectorpers. comm.). Although Table 3. Percent hunting successes(kills/attempts; total I detected no vocalizationsamong hunting lan- of 52 attempts, 16 kills) by Lanner Falconshunting co- ners, Thomsett (1987) reported that pairs of lan- operativelyon flocks of Jackdawsin Sicily. ners hunting bats gave chupping calls. Mebs (1959), however,failed to mention any calls given JACKDAWFLOCK SIZE CLASS by cooperative hunting lanners in Sicily. Partici- ATTACKTYPE 2--10 11--30 31--50 pants in hunts, however,can coordinate pursuits without vocal signals.Massa et al. (1991) suggested Partial 0 6 22 that partners monitor their movementsby visual surprise contact. Predators should avoid vocalizations dur- Surprise 11 11 0 ing surpriseattacks, which would reveal their pres- Nonsurprise 9 27 9 126 LEONARDI VOI•. 33, NO. 2 male success rate of <1% was irrelevant in com- cocolumbarius) hunting successwhen pursuingsmall parison to the 50% reported. This was probably flocks of birds (Cresswell1996). In this study,sur- becauseof the strong reversedsexual dimorphism prise attacks were less successfulthan were other (RSD) of this speciesand its tendency not to share strategies(17.6%) and were employedmostly for prey. In other words,females physically dominated attacksof pigeons.In SouthAfrica, surpriseattacks males during hunts and feedings. RSD may also from fast, low coursing flight were principally account for divergences in hunting and prey aimed at small birds and doves (Streptopeliaspp.; choice. Males of F. b. feldeggiweigh 69% that of Kemp 1993). In my study in Sicily,surprise attacks females and capture prey which average45% the on Jackdaws caused intense confusion inside size of the female's prey (Leonardi et at. 1992). It flocks.This confusion,and the dilution effect,pro- is likely that RSD favorscooperative hunting, since duced an abatement effect; Lanner Falcons had it allowsthe hunting of a wide range of prey and difficulty attacking the group repeatedly, decreas- also the use of different hunting strategies. ing capture chances (Leonardi 1991, Turner and Data on flock size choice showed that tanners Pitcher 1986, Krause and Godin 1995). prefer to attack larger groups.In previousstudies ACKNOWLEDGMENTS of flocking behavior and hunting, hunting success has been shown to be inversely proportional to Special thanks are due to Alan Kemp, C.M. White, and flock size (Kenward 1978, Turner and Pitcher 1986, R. Yoseffor their invaluablesuggestions and critical read- ing of the manuscript.Dean P. Keddy-Hector,Steve Shet- Cresswell 1994, Krause and Godin 1995). Krause rod, David Ellis, and Daniel Varland made extensive com~ and Godin (1995) suggestedthat flock conspicu- ments that improved the paper. ousness,rather than flock size per se, influenced LITERATURE CITED predator choice. Flock conspicuousnesslends to repeated attacksin a singlechase, thereby increas- BEDN•d•Z,J.C. 1988. Cooperative hunting in Harris' ing success(Krause and Godin 1995). In Jackdaws, Hawks ( Parabuteounicinctus) . Science239:1525-1527. antipredator defenseis basedon the group's con- BIJLSM&R.G. 1990. Predation by large falconson winter- ing waders in the Banc d'Arguin, Mauritania. Ardea spicuousness(which determines the encounter 78:75-82. rate) and on the total number of individuals in the BROSSETT,A. 1961. l•cologie des oiseaux du Marocori- group (dilution effect; Turner and Pitcher 1986). ental, Faucon Lanier. Tray. Inst. Sci. Chbrif Sbr.Zool In my study,the effect of group conspicuousness 22:32-33. on rates of encounter with falconsmay have been C•*•CO, T., S. M•d•TIND^LE )•ND H.R. PULLIAM. 1980. Axa- immaterial becauseJackdaws lived so close to nest- an flocking in the presence of a predator. Nature285 ing tanners (Pitcher 1986, Krause and Godin 400-401. 1995). CRAMP. S. •ND K.E.L. SIMMONS, [EDS.] 1980. The birds of Asin my study,partial surprise was the strategy the western Palearctic, Vol. 2. University Press, Ox- most commonly used by cooperativehunting Lan- ford, UK. ner Falcons in South Africa (Kemp 1993, Jenkins Cm•ssw•LL,W. 1994. Flocking is an effective antipredator strategyin Redshanks,Tringa totanus.Anim. Behav.47 1995). Sicilian lanners frequently use this strategy 433-442. (60.8%) in capturingJackdaws only. In South Af- 1996. Surprise as a winter hunting strategyin rica, nonsurpriseattacks were aimed at small prey SparrowhawksAccipiter nisus, Peregrines Falcoperegr•- (Kemp 1993,Jenkins 1995). PreviousSicilian stud- nus and Merlins E columbarius. Ibis 138:684-692. ies described nonsurprise attacks as frequent co- ELLIS,D.H., J.C. BEDNARZ,D.C. SMITHAND S.P. FLEMMING. operative techniques used against larger prey 1993. Social foraging classesin raptorial birds. Bio- (Mebs 1959, Massaet at. 1991). My data indicated Science 43:14-20. a subordinate use of this strategy in comparison HECTOR,D.P. 1986. Cooperative hunting and its relation- with the partial surpriseattack. Inversely,tanners ship to foraging successand prey size in an avian using nonsurprise attacks had good hunting suc- predator.Ethology 73:247-2,57. INSTITUTOGEOGRAFICO DE AGOSTINI.1987. Atlante ge- cessrates. This technique was probably used be- nerale metodico. IGDA, Novara, Italy. causeit involved energeticallyinexpensive soaring JENVdNS,A.C. 1995. Morphoxnetricsand flight perfor- and resulted in relatively high hunting success mance of southern African Peregrine and Lanner Fal- (Jenkins 1995). cons.J. Arian Biol. 26:49-,58. The surpriseattack was reported as the most im- KEMP,A.C. 1993. Breeding biology of Lanner Falcons portant factor in PeregrineFalcon and Merlin (Fal- near Pretoria, South Africa. Ostrich 64:26-31. JUNE1999 COOPERATIVEHUNTING BY LANNER FALCONS 127

KENWARD,R.E. 1978. Hawks and doves:factors affecting PITCHER,TJ. 1986. Functions of shoaling behavior in tel- successand selection in Goshawkattacks on pigeons. eosts.Pages 294-337 in TJ. Pitcher [ED.], The behav- J. Anita. Ecol.47:449-460. ior of teleostfishes. Chapman and Hall, London, U K KRAUSE,J. ANDj.-Gj. GODIN.1995. Predator preferences PORTER, R.F., I. WILLIS, S. CHRISTENSEN AND N.B. NIELSEN for attacking particular prey group sizes:consequenc- 1981. Flight identification of European raptors. T & es for predator hunting successand predation risk. A.D. Poyser,Calton, U.K. Anita. Behav. 50:465-473. SODHI, N.S., A. DIDIUK AND L.W. OLIPHANT. 1990. Differ- LEONARDI,G. 1991. Osservazionipreliminari sull'eco-eto- ences in bird abundance in relation to proximity of logia del Lanario Falco biarmicusj•ldeggi in Sicilia. Merlin nests. Can.J. Zool.68:852-854. Suppl.Ric. Biol. Selvaggina 17:147-149. SUHONEN,J., m. NORTDAHLAND E. KORPIM'AKI.1994. Avlan 1994. The home range of the lanner Falcobiar- predation risk modifies breeding bird community on micus:influences of territory composition.Pages 153- a farmland area. Ecology75:1626-1634. 155 in B.-U. Meyburg and R.D. Chancellor [EDS.], TARBOTON, W. AND D. ALLAN. 1984. The status and con- Raptor conservationtoday. WWGBP and Pica Press, servation of birds of prey of Transvaal. TransvaalMus Berlin, Germany. Monogn:No. 3, Pretoria, South Africa. , A. LONGO AND G. CORPINA.1992. Ecology and THIOLLAY,J.-M. 1988. Prey availabilitylimiting an island behavior of the Lanner Falcon. GEE Publications, Ca- population of Peregrine Falcons in Tunisia. Pages tania, Italy. 701-710 in TJ. Cade,J.H. Enderson, C.G. Thelander MASCARA,R. 1986. Consistenzae note sulla biologia ri- and C.M. White, [EDS.], Peregrine Falcon popula- produttiva del Lanario Falco biarmicus,nella Sicilia tions: their management and recovery. Peregrine meridionale. Riv. Ital. Ornitol. 56:203-212. Fund, Boise, ID U.S.A. MASSA, B., E Lo VALVO, M. SlRACUSAAND A. CIACCIO. THOMSETT,S. 1987. Bat hunting by Lanner Falconsin 1991. I1 Lanario (Falcobiarmicus feldeggi Schlegel) in Kenya. Gabar2:7-8. TREE,AJ. 1963. Grey Hornbill Tockusnasutus as prey of Italia: status,biologia e tassonomia.Naturalista Sicili- the Lanner Falcon Falco biarmicus. Ostrich 34:179. ano 15:27-63. TURNER,G.F. ANDT.J. PITCHER.1986. Attack abatement. MEBS,T. 1959. Beitrag zur Biologie des Feldeggsfalken a model for group protection by combined avoidance (Falcobiarmicus j•Ideggi). Vb•elwelt 80:142-149. and dilution. Am. Nat. 128:228-240. MIRABELLI,P. 1982. Biologia del Lanario Palcobiarmicus YOSEF, R. 1988. Observations on Lanner Falcons Falco in Calabria: confronti con la biologia del Falco Pel- biarmicusin the Sede Boqer area. Torgos7:68-73 (in legrino Falcoperegrinus. Atti Conv.Ital. Ornitol.1:149- Hebrew). 154. •. 1991. Foraging habits,hunting and breeding suc- MORGAN,JJ. ANDj.-G.j. GODIN. 1985. Antipredator ben- cessof Lanner Falcon (Falcobiarmicus) in Israel.J. Rap- efits of schoolingbehavior in a cyprinodontidfish, the tor Res. 25:77-81. barred killfish (Fundulusdiaphanus). Z. Tierpsychol.70: ZAR,J.H. 1984. Biostatistical analysis.Prentice-Hall Inc., 247-264. EnglewoodCliffs, NJ U.S.A. OS}•ORNE,T.O. 1981. Ecology of the Red-necked Falcon Falcochicquera in Zambia. Ibis 123:289-297. Received6 August1997; accepted30 January 1999