J. Raptor Res. 25(1):9-17 ¸ 1991 The Raptor ResearchFoundation, Inc. DEVELOPMENT OF FORAGING BEHAVIOR IN THE AMERICAN KESTREL DANIEL E. VARLAND AND F•RWIN E. KLAAS U.S. Fish and WildlifeService, Iowa CooperativeFish and Wildlife ResearchUnit, Iowa State University,Ames, IA 50011 THOMAS M. LOUGHIN AgriculturalExperiment Station, Department of Statistics,Iowa State University,Ames, IA 5001 ABSTRACT.--Weobserved the developmentof foragingbehavior after nestdeparture in 12 siblinggroups of American Kestrels (Falcosparverius). Perch resting decreasedwhereas perch hunting, eating self- capturedprey, and flying increasedover the 5-wk periodthat youngwere observed.Kestrels used perch hunting more than other typesof hunting and fed exclusivelyon invertebrates,primarily grasshoppers. Perch hunting success(captures/pounces) increased significantly 3 wk after fledging.After this period there was no significantchange. Significant increasesin capture rate (captures/hour) occurred4 and 5 wk after fledging due to increasedpounce rates. We observedsocial hunting among siblings,families, and alsoamong unrelated kestrels. Social hunting occurred during both perchhunting and groundhunting. Socialforaging in thesekestrels was imitative rather than cooperative. Desarrollo de los h•tbitos de caceria en los Halcones Cernlcalos EXTe,•CTO.--Hemosobservado el desarrollode losh•tbitos de bfisquedade alimentoen HalconesCernlcalos (Falcosparverius) con 12 gruposde hermanos.Despu•s de un perlodo de 5 semanasde observaci6n, notamosque estosj6venes halcones disminuyeron la frecuenciadel posarsepara descansar,mientras que aumentaronla frecuenciaen el posarsepara cazar, el comersu presacapturada, y el volar. La cacerla desdeuna perchaocurri6 m•ts que otrostipos de caceria.Se alimentaronexclusivamente de invertebrados, principalmentede saltamonteso longostas.E1 •xito de la cacerladesde una posici6nperchada (captura/ embestida)creci6 significativamente despu•s de 3 semanasde haber dejadoel nido. Desputsde estono hubo cambiosignificativo. Un notableincremento en la proporci6nde capturas(captura/hora) ocurri6 entre 4 y 5 semanasdespu•s de dejar el nido, y se debi6 al incrementoen la proporci6nde embestidas. Se observ6cacerias en gruposentre hermanos,familias, y tambi•n entre individuossin parentesco.Las caceriasen grupo se realizaron o bien desdeel sueIv o desdeuna percha. La provisi6nde comidaen grupoen estoscernicalos rue m•tsbien imitativa que cooperativa. [Traducci6n de Eudoxio Paredes-Ruiz] The post-fledgingperiod, here definedas the pe- In 1988 we begana study of American Kestrels riod of parental dependencyfor foodin youngbirds nestingin nest boxesattached to the backsof high- after leaving the nest (see van Tyne and Berger way signsalong InterstateHighway 35 (I-35) in 1966), hasreceived relatively little attentionin avian Central Iowa. In this paper we describethe devel- research.This is partly becauseof the difficultiesin opmentof foragingbehavior in youngkestrels during observingthe behaviorof youngonce they leavethe the post-fledgingphase and during the period of nest (e.g., Brown and Amadon 1968, Newton 1979, recent independencefrom parents. Alonso et al. 1987). STUDY AREA AND METHODS The post-fledgingperiod and the subsequentpe- Severalyears prior to the initiation of this study,kestrel riod of recent independencefrom parents are im- nest boxeswere attachedto the backs of highway signs portantlife historystages, when youngdevelop for- along1-35 at approximately2-km intervals,from northern aging skills essentialto survival (Weathers and Polk County to northern Worth County in northcentral Sullivan 1989). High mortality ratesof recentlyin- Iowa. The studyarea was a corridorapproximately 2 km wide on either side of 1-35 from 18 km south to 99 km dependentjuveniles and othersduring their first year north of Ames. Land bordering1-35 was farmed inten- of life reflectsthe critical nature of this time (e.g., sively with row crops. Lack 1954, Henny 1972, Sullivan 1989). We banded 97 fledglingsobserved in 1988 and 1989 10 VARLAND •.T AL. VOL. 25, NO. 1 with U.S. Fish and Wildlife Serviceleg bands,and indi- appearedfrom view. We did not use data if the bird left vidually marked them with coloredvinyl leg jessesbefore in < 5 min. We analyzeddata for 93 observationsessions they fledged.We captured76 percent(35/46) of the adult (mean length = 57.5 min, SD -- 32.0). kestrels in the nest box or with bal-chatri noose traps A metronometiming device(Wiens et al. 1970), set at (Berger and Mueller 1959). We bandedand individually 20 sec intervals, cued spot observationsof behavior and marked adults with coloredvinyl leg jesses. social activity. At each sound of the tone, we recorded To locatefledged young for behavioralstudies we used behavior and social activities of the focal kestrel. We re- the signalsfrom back-mountedradio transmitters (Holohil cordedfour main classesof activity:general behavior, so- Systems,Ltd., Woodlawn,Ontario, Canada). We attached cial behavior, hunting behavior, and allopreening and transmittersto birds severaldays before fiedging. In 1988 beaking.We recognizednine subclassesof generalbehav- we attached radio transmittersto 12 nestlingsin 10 nest ior and five of social behavior. boxes.Survival of radio-markedkestrels was high (11 of General Behavior. "Perch resting"describes a kestrel 12 survivedthe post-fledgingperiod) and siblingsgenerally perchedand not engagedin any other observedbehavior. maintained closecontact for 4 to 5 wk after fiedging.This Rudolf (1982) and Toland (1987) distinguished"perch confirmed the technique'susefulness and feasibility for hunting" from other perching activity by alert posture, monitoring family group activity. We made observations erectbody or bodyleaning slightly forward, frequent star- in 1988 to gain insightinto AmericanKestrel post-fiedging ing at ground (Fig. 1), and head bobs. Becauseyoung behaviorand to developan efficientdata recordingsystem. kestrelsthat have never huntedmay exhibit someof these These data are not part of the presentanalysis. behaviorswithout attemptingprey captures,behavior was We testedthe transmittersused in 1989 alongthe high- not recordedas perch hunting until at least one pounce way right-of-way at a heightof I m. Signalrange averaged was observed.Flights to/and from the ground and flights 2.3 km (N = 13, SD = 0.60, range = 1.1-3.5 km). In betweenperches during perchhunting bouts were included 1989, we radio-taggedone randomly selected nestling from in perchhunting behavior.We defined"ground hunting" eachof 13 nests.Young observedin 1989 (50 individuals as a bird on the ground searchingfor prey for >20 sec. from 13 nests)fledged between 27 and 31 d after hatching Searchesof shorterduration involvingflight from a perch (mean = 29.2, SD -- 1.4), from 13 June to 3 July. were consideredperch hunting. "Flight" was any non- One radio-taggednestling died 7 d after fledgingbefore huntingflight. We usethe term "eating" only for kestrels we could collect behavioral data. We lost signals from 3 eating self-capturedprey. "Maintenance activity" includ- of the remaining 12 transmitterswithin 5 d after the tagged ed preening,plumage rousals (shaking), and stretching. birds fledged. For two of these sibling groups,we were "Lying on belly" describesa postureyoung kestrels often unable to determine whether the transmitters failed or if assumedon fenceposts,utility poles,and largetree branch- the individuals left the area. For the third, transmitter es. "Begging"was solicitationof food from parents."Out failure became evident when we observed the radio-marked of sight" refersto a focal kestrelconcealed by vegetation kestrel with another siblinggroup in the study 37 d after or other objects.A sessionwas discontinuedwhen a bird fiedging.Despite the early lossof signalsfrom thesethree was out of sight >5 min. "Other" was usedto categorize transmitters, we were able to collect data on behavior of behaviorsobserved relatively infrequently,and included individuals in these broods. walking,hover hunting, aggressive interactions among sib- We observedfledglings between 0600 and 1300 H at a lings, parent-to-youngprey transfers,and eating prey distance of 70-100 m with a 20x or 20-60x spotting caught by parents. During observationsessions, one or scope.We did not use a blind becausebirds under obser- both adults frequently vocalizedaggressively at us. We vation frequently changedlocations. We monitoredfledg- therefore suspectthat the interactionswith parents oc- ling groupson a rotational basisat 1-3 d intervalsuntil curredless frequently than they would havein the absence we lost contact with the brood. When we could not find of observers. a brood,we searchedby vehiclean area of approximately Social Behavior. Lett and Bird (1987) defined social 6 km 2 around their last known location. behaviorfor American Kestrel fledglingsas any behavior We adoptedWyllie's (1985) definitionof dispersal,which which occurredwithin 2 m of one or more siblings.We is movementof a fledgedbird farther than I km from its adoptedthis operationaldefinition with two modifications. nestwithout return. We determinedtime of dispersalonly We extendedthe distanceto 3 m and includednon-sibling for kestrelswith transmittersknown to be functioning1 kestrelsin socialinteractions (adults late in the post-fledg- wk after fledging (N = 9). ing periodwhich no longerfeed their youngand kestrels At the beginningof each observationsession, we ran- from outsidethe parent/sibling family unit). "Associa- domly selectedone fledgling, which was not