NESTING OF THE GREATER KESTREL FALCO R UPICOLOIDES IN ZAMBIA by TimothyO. Osborne AlaskaDepartment of Fish and Game Box 155 Galena, Alaska 99741 and ]. F. R. Colebrook-Robjent Musumanene Box 303 Choma Zambia

Abstract Breedingdata were obtainedfrom 10 pairs of Greater Kestrelsat Minyanya Plain, Zambia,during September 1975. Black Crow nestswere the only nestingplatforms uti- lized,and we concludedthat they were acquiredaggressively by the kestrel.Sixty % of the nestskestrels used had beenbuilt by crowsduring the 1975 seasonindicating a fre- quentoccurrence of nestpiracy. Incubationwas undertakenby the female which was very shyon the nest.The distancebetween nests averaged 2.3 km, slightlymore than the distancebetween all crownests. The breedingdistribution and nestingdensity of the GreaterKestrel in Zambiais directly influencedby the distributionand densityof the Black Crow.

Introduction Bensonet al. (1971)summarized the distributionof the GreaterKestrel (Falco rupico- loides)in Zambiaand notedthat there were no breedingrecords. Aspinwall (1979) re- cordedthe firstnest of the speciesin Zambiawhich contained 4 fresheggs on 30 August 1974 at Mitashi Plain (13ø35'S.,22ø50'E.), western Zambezi (formerly Balovale)Dis- trict. The eggswere laid in an old nestof a BlackCrow (Corvuscapensis). The statusof the GreaterKestrel in Zambiahas been discussedbriefly in a previouspaper (Osborne and Colebrook-Robjent1980). This paperpresents breeding data on an undisturbedpop- ulationof GreaterKestrels at MinyanyaPlain, Zambia.

StudyArea and Methods MinyanyaPlain (13ø09'S.,22ø23'E.) is a watershedLoudetia grassland lying between the North and South KashijiRivers in ZambeziDistrict, North-Western Province, Zambia. The plain,approximately 1150 m above sealevel, is borderedon the northby the North Kashijifloodplain; to the southand westby brokenstands of Diplorhynchuswoodland and to the eastby degradedKalahari (Baikiaea) woodland. The soilsare Barotse sandswith an unevensurface due to numerousCubitermes mounds. Scattered over the plainare solitarytrees or sparseclumps of treesrarely greater than 6 m high. From 10-15 September1975 we locatedall raptorand crownests within a 5 km radiusof our campsite in a smallisolated stand of Syzigiumjust northof an east-westtrack whichbisected the plain (Fig. 1). The open natureof the plain and the low statureof the treesenabled us to find all the nests.On 15 Septemberwe drovea transectSSE 60 km towardsSouth Kashiji School and examinedevery nest we saw.

71 RaptorResearch 16(3):71-76 72 RAPTOR RESEARCH Vol. 16, No. 3

Results Ten pairsof kestrelwere locatedin the studyarea (Fig. 1). In addition,3 singlekes- trelswere observedon the edgeof the studyarea. All 10 pairswere associatedwith ei- ther new or old nestsconstructed by BlackCrows. The crow nestswere within the up- per canopyof the tree and were composedof a platformof sticksusually lined with animal hair. Of 19 crow nestsin the studyarea, 4 were old and unoccupied,5 containedkestrel eggs,2 containedBlack Crow eggs,2 containedBlack Crow nestlings,2 were under constructionby crows,2 were new nestswith kestrelsin attendance,and 2 were old nestswith kestrelsin attendance.We attributedone nestwhich containedbroken eggs andappeared deserted to a kestrelpair whichwas perched 500 m away. We observedincubation only by females.They allowedus to approachto within250 to 400 m of the nestbefore flying unobtrusivelyaway, iust above ground level, out of sight.One femalekestrel vigorously defended her empty nestevery time we examined it, but neitherthis bird nor any otherkestrel ever uttereda call.

NORTHFLOODPLAINKASHIJI

Figure 1.--Locationof GreaterKestrel nest sites at MinyanyaPlain, ZambeziDistrict, Zambia,September 1975.The plain is bisectedby a motortrack; the Diplorhynchuswoodland is showncrosshatched; the Baikiaeawoodland stippled, and the Syzigiumstand is black. Fall 1982 Osborneand Colebrook-Robjent--Greater Kestrel 73

The averagedistance between kestrel nest sites was 2.3 km; the densitywas one nest- ing pair per 415 ha. The nestingactivities of the kestrelpairs are summarizedin Table 1. Clutchsize of the kestrelwas three C/4 and two C/2. Egg layingstarted at the end of August(Aspinwall 1979) and continuedpast mid-September.The averageweight of •ven unincubatedeggs was 24.2 g (23.1-27.0 g). Measurementsof 26 eggsaveraged 41.5 by 32.7 mm (range39.4-43.3 by 31.4-34.6 mm). TheseZambian Greater Kestrel egg measurements,which includedAspinwall's (1979), are smallerthan the averages givenin McLachlanand Liversidge (1978). On our transecttowards South Kashiji School, we found 13 Black Crow nests,4 of whichwere occupied by kestrels. The BlackCrows at Minyanyabred over an extendedseason. We foundnests contain- ing largeyoung (eggs laid in earlyAugust), and nestsstill in the early stagesof construc- tion. Like Aspinwall(1979) and White (1946) we found the crowsto be very shy,and

Table1. Summaryof GreaterKestrel nests at MinyanyaPlain, Zambia, September 1975

Pair Nest site Contents Notes on the kestrels 1 New crownest 4.5 m highin leafy nil Femaleunusually aggressive Parinarimobola, crow recently whennest inspected 12 Sept. dispossessed Still empty 1000h 15 Sept.

2 Old crow nest2.4 m high in small C/4 Three eggs10 Sept.fourth defoliated tree egglaid before0700 h 13 Sept.

3 Old crownest 2.1 m high in small nil Pair at nest 10 to 15 Sept. defoliated tree

4 New unlinedcrow nest 4.2 m high nil Pair perchednearby 11 Sept. in medium-sized tree. Nest still unlined14 Sept.

New fidly linedcrow nest 4.8 m 2 eggs No eggs11 Sept.first egglaid highin solitary'Mukutanlonga' before 1000 h 12 Sept.,second tree,partly defoliated egglaid before 1400 h 14 Sept.

6A Old crow nest4.5 m high in 1 or more Desertednest with one egg solitarymedium-sized tree eggs and remainsof others12 Sept.

6B Incompletenew crow nest nil Pair perched500 m north 12 Sept. in solitarytree probablypair from failed nest6A

New crow nest5.4 m high in partly C/3 12 Sept. defoliated tree

New crow nest in medium-sized nil Pair 100 m from nest 12 Sept. tree

Old crownest 5.4 m high in partly C/4 Freshlyplucked button quail defoliated tree Turnixsylvatica remains on groundnear nestsite 13 Sept.

Old anddelapidated crow nest nil Kestrelperched nearby 10 Sept. 3 m highpartly tippedover 74 RAPTOR RESEARCH Vol. 16, No. 3 we werenot ableto observeany pairs at theirnests. The densityof the crowsaveraged 381 ha per nestingpair, considerablylarger than the 60 ha averagefound in SouthAf- rica (Skead1952).

Discussion The GreaterKestrel uses the nestsof crowsor hawksin SouthAfrica (McLachlanand Liversidge1978); however, the Minyanyapopulation of GreaterKestrels in Zambiahas developeda restrictivebreeding strategy in whichBlack Crow nestsare exclusivelyuti- lized as the nest site. We concluded that when a suitable nest site was not available within the kestrel's homerange, the pair waiteduntil a BlackCrow built a nestand thenpirated it. Of the 10 pairsof kestrelswe observed,2 were quietlyperched 100-200 m awayfrom crow nestsunder construction. Although we do not havedirect observations of nestpiracy, it wasfrequent because 60% of the kestrelswere using nests built duringthe 1975nesting season. The averagedistance between crow nests built in the 1975 seasonirrespective of the presentocccupant (either crow or kestrel)was 2.3 km. When this is comparedwith the 2.3 km we foundbetween kestrel nests, it is apparentthat the densityof the kestrelnests is directlyrelated to the densityof crow nests.Once the minimumrequirement of a nest siteis satisfied,other factors, such as food, must affect density since the total crow nests outnumberedthe nestsoccupied by kestrels. A reviewof the genusFalco in Zambiaindicated that perhapsbreeding density is in- flueneedby "host"nest site density in severalother species. In partsof the rangeof the Bed-neckedFalcon (F. chicquera),a Pied Crow (Corvusalbus) nest is the nestplatform (Colebrook-Robjentand Osborne1974). On the KafueFlats a relationshipbetween fal- conand crow density is suspected(pets. obs.). In largeareas of Zambiaand Zimbabwe, whererock outcrops do not existto providecliff faces,the LannerFalcon (F. biarmicus) nestin newor oldBateleur (Terathopius ecaudatus) nests (pers. obs., Steyn 1980). The distributionof the GreaterKestrel in Zambiaas stated by Bensonet al. (1971)im- pliesthat the bird is widespreadin suitablehabitat throughout the southand west of the country.Their recordsmainly covered the periodof Julyto October,and, since no nests hadbeen discovered, they suggestedthat mostkestrels found in Zambiawere derived from Botswananbirds driven north during the cold dry season.However, with our knowledgeof kestrelnest site preferences, a distinctionmust be madenot in the date of the observationbut whetheror not the record occurswithin the known range of the BlackCrow. Kestrelrecords from elsewherein Zambia (Fig. 2), regardlessof the date, canbe assumedto be wanderingbirds. It is not possibleto discussthe breedingdistribution of the GreaterKestrel in Zambia without reviewingthe distributionof its nest building "host,"the Black Crow. The rangeof the crow in Zambia(Benson et al. 1971)parallels the distributionof the water- shedgrassland vegetation type foundwest of the ZambeziRiver and north of the Kalabo River (Fig. 2). There are furthercrow recordsfrom KonkanoPlain (13ø15'S.,24ø00'E.), alsoa watershedgrassland, and from Livingstone,but the latter recordis thoughtto be an escape(Benson et al. 1971).It is not knownwhat elementin the watershedgrassland determinesits suitabilityto the crow. The distributionof the kestrelshows a concentra- tion of recordsfrom thesesame grasslands. We expectthe kestrelwill be found as a breedingspecies on the KonkanoPlain. Fall 1982 Osborneand Colebrook-Robient--GreaterKestrel 75

ZAMBIA _10øS

-12

Mi.nyanya

-14

K afue R.

16

ezi R.

vingstone -18

212 214 216 2,8 30I 32øE

Figure2.--Distribution of GreaterKestrel in Zambia.Watershed and floodplain grasslands of the westernhalf of thecountry represented by stippledarea. Locations where Greater Kestrels have been collected or sightre- cordsare represented by the symbol ( • ). Locationswhere Black Crows have been collected or sightre- cordsare represented by the symbol( • ). FromNational Museums of ZambiaProvisional Bird Atlas species numbers112 and 384 (courtesyR. J. Dowsett). The Black-shoulderedKite (Elanuscaerulus) was common on the NorthKashi]i flood- plain,the northernborder of our studyarea, but we did not observeany on Minyanya Plain.Likewise Britton (1970) considered it wasprobably absent from watershed plains in the District.Conversely, we failedto find any kestrelson the floodplain.Since both speciesprey on similarfood items (mice and insects)(McLachlan and Liversidge 1978), exclusiveterritories were not expected.Possible reasons for the exclusioncould be inter- specificaggression, competitive exclusion or a differenthunting technique. The hov- eringhunting style of the kite maygive it an advantageon the floodplainwhere grasses are tallerthan on the plain. 76 RAPTOR RESEARCH Vol. 16, No. 3

The differencesin the averageegg size between Minyanya and SouthAfrican kestrels suggestthat the Zambiankestrels may be smaller.We measuredan additional46 eggsof the kestrelfrom Transvaal;their average42.5 by 34.2 mm (range 40.0-44.6 by 32.3-35.7mm) fallsbetween the Minyanyaand SouthAfrican eggssuggesting a size cline. However,a comparisonof wing lengthsdoes not showa significantdifference. The wingsof four femalescollected in Zambiaaveraged 280 mm (range273-289 mm) (Bensonand Irwin 1967,Aspinwall 1979) and 18 southernAfrican females average 282 mm (range272-290 mm) (McLachlanand Liversidge1978). More data are neededto clarifythis issue.

Summary The breedingdistribution of the GreaterKestrel in Zambiais very restricted.The breedingstrategy of the kestrelrestricts it to regionswhere its nestbuilding "host," the BlackCrow, nests. The densityof kestrelnests on MinyanyaPlain is relatedto the den- sityof crownests. The breedingseason commenced in August,and somepairs had not laid eggsby mid-September.The averageegg size was smaller than SouthAfrican birds suggestinga difference in the sizebetween the birds.

Acknowledgments We thankthe followingwho renderedfield assistance,R. A. Conant,L. Mbewe, and L. Y. Osborne.D. R. Aspinwalland R. J. Dowsettmade useful comments on an earlier draft and H. Walter is thankedfor his help on the final draft. We thankthe Ministerof Lands,Natural Resources and Tourism S. Kalulufor the opportunityto visit Minyanya Plain.

Literature Cited Aspinwall,D. R. 1979. Bird notesfrom Zambezi District, North-WesternProvince. ZambianOrnithol. Soc. Occas. Papers 2:1-60. Benson,C. W., R. K. Brooke,R. J. Dowsett,and M. P.S. Irwin. 1971.The Birds of Zam- b/a. Collins,London. 414 pp. Benson,C. W. and M. P.S. Irwin. 1967.A contributionto the ornithologyof Zambia. ZambiaMuseum Papers 1:1-139. Britton, P. L. 1970. Birds of the Balovale District of Zambia. Ostrich 41:145-190. Colebrook-Robjent,J. F. R. andT. O. Osborne.1974. High densitybreeding of the Red- neckedFalcon Falco chicquera in Zambia.Bull. Br. Ornithol.Club 94:172-176. McLachlan,G. R. and R. Liversidge.1978. Roberts Birds of SouthAfrica. FourthEd. JohnVoelcker Bird BookFund, Cape Town. 660 pp. Osborne,T. O. and J. F. R. Colebrook-Robjent.1980. The statusof the genusFalco in Zambia.Proc. IV Pan-Afr. Orn. Congr.301-306. Skead,C. J. 1952.A studyof the BlackCrow Corvuscapensis. Ibis 94:434-451. Steyn,P. 1980. Breedingand food of the Bateleurin Zimbabwe(Rhodesia). Ostrich 51:168-178. White, C. M. N. 1946.The ornithologyof the Kaonde-LundaProvince, Northern Rho- desia. Ib/s 88:68-103, 206-224, 502-512.