Quick viewing(Text Mode)

Diets of Breeding Peregrine and Lanner Falcons in South Africa

Diets of Breeding Peregrine and Lanner Falcons in South Africa

j RaptorRes. 33(3):190-206 ¸ 1999 The Raptor ResearchFoundation, Inc.

DIETS OF BREEDING PEREGRINE AND LANNER IN SOUTH

ANDREWR. JENKINS PercyFitzPatrick Institute of African Ornithology, University of CapeTown, Rondebosch 7701, SouthAfrica

GRAHAM M. AVERY SouthAfrican Museum, P.O. Box 61, CapeTown 8000, SouthAfrica

ABSTRACT.-•Thediets of breedingPeregrine (Falcoperegrinus) and Lanner (F. biarmicus)Falcons in were determined from the analysisof prey remains collected at nest sites and through direct observationsto determinethe regionalvariation in PeregrineFalcon prey and to measurediet overlap, and the potential for competition between Peregrine Falconsand sympatriccongeners. Direct obser- vationssuggested that remainsunder-sampled small prey by about 10% and over-sampledlarge prey by about 8%. Peregrineand Lanner Falconspreyed mostlyon . Pigeonsand Streptopeliadoves com- prised the bulk (38-66% by frequency;68-85% by mass)of the PeregrineFalcon prey in each of three study areas. Columbidswere supplementedby starlings(mostly European Starling [Sturnusvulgaris]) on the Cape Peninsula, ( spp.) and swifts(Apus spp.) on the Orange River, and mousebirds(Colius spp.) in the Soutpansberg.Cape PeninsulaPeregrine Falcons had the leastdiverse diet, the narrowestfeeding niche and they took the largestproportion of juvenile birds. Peregrine Falconson the Orange River had the broadestfeeding niche and preyedmainly on 'commuter' rather than sedentaryresidents. Lanner Falconsin the Soutpansbergtook mainlyterrestrial or cursorial species,particularly young (Gallus gallus, 40%; 37%) and charadriids,but columbidswere also important.The dietsof sympatricPeregrine and Lanner Falconsoverlapped by about 35%. Peregrines Falconsconcentrated their foraging on woodland and cliff-dwellingprey, while Lanner Falconstook mainly open-countryspecies. Close-neighboring pairs of congenersdid not obviouslyaffect the food- niche parametersof either speciessuggesting that they were not activelycompeting for food. KEYWORDS: PeregrineFalcon; Falco peregrinus; LannerFalcon; Falco biarmicus; diet;,niche dimensions; competition.

Dieta de Falcoperegrinus y Falco biarmicus en reproduccitn en Surifrica RESUMEN.--Lasdietas de Falcoperegrinus y Falcobiarmicus en reproduccitn en Surfifricafueron deter- minadasa partir del anilisis de restosde presasrecolectados en los sitiosde anidacitn y a travts de observacionesdirectas para determinar la variacitn regional en las presasde halctn peregrino,medir su coincidenciay la competenciapotencial entre halconesperegrinos y suscongtneres simpitricos. Las observacionesdirectas sugirieron que las presaspequefias fueron subvaloradasen un 10% y que las presasgrandes fueron sobrevaloradasen un 8%. Loshalcones peregrinos y lanariosse alimentaron mas que todo de aves.Palomas y otrasaves del gtnero Streptopeliarepresentaron la mayoria(38-66% de la frecuencia; 68-85% de la biomasa) de las presasde los halconesperegrinos en cada una de las tres ireas de estudio.Las Columbiformesfueron suplementadaspor Sturnusvulga•is en la Peninsuladel Cabo, Pteroclesspp. y Apusspp. en el Rio Orange y Coliusspp. en el Soutpansberg.Los halconespere- grinos de la Peninsuladel Cabo tuvieron la dieta menosdiversa y el nicho alimenticiomis restringido cazando avesjuveniles. Los halconesperegrinos del R/o Orange tuvieron el nicho alimenticio mils amplio y depredaron en "especiespasajeras" contrario alas sedentarias.Los halconeslanarios en el Soutpansbergse alimentaron de especiesterrestres particularmente de pollos. (Gallusgallus, 40%; 37%) y Charadriformes,las Columbiformes fueron tambitn importantes.Las dietas de loshalcones peregrinos y lanariossimpitricos coincidi6 en un 35%. Los halconesperegrinos concetraron su forrajeo en presas de bosquesy riscosmientras que los halconeslanarios simpitricos coincidi6 en un 35%. Loshalcones peregrinosconcentraron su forrajeo en presasde bosquesy riscos,mientras que los halconeslanarios en presasde espaciosabiertos. La cercania de parejas de congtneres obviamente no afect6 los pari- metrosdel nicho alimenticiode ninguna de las dos especieslo que sugiereque estasdos especiesno estan compitiendo por cornida.

190 SEPTEMBER 1999 PEREGRINE AND LANNER DIETS IN SOUTH AFRICA 191

[Traducci6n de Cfsar M/trquez]

Food habits of the Peregrine (Falcopere- northwest to southeast, and temperatures are mild to warm. grinus)have been describedfor arctic and temper- Prey remains were collected from falcon nest ledges ate regions (Mearns 1983, Hunter et al. 1988, Rat- and from below roost sitesand feeding perches.Uneaten cliffe 1993, Rosenfield et al. 1995), and for remains, regurgitated pellets and plucked feathers were southwesternAustralia (Pruett-Joneset al. 1981, used in combination to minimize bias in diet analyses Marchant and Higgins 1993). However,in the trop- (Simmons et al. 1991, Oro and Tella 1995). Collections were made from just after egg laying to soon after fledg- ics, the diet is poorly known and data for African ing. The frequency of collectionsvaried between sites, populationsare limited (Mendelsohn1988). The years,and areas (Table 1). Care was taken to remove all only studiesof (E biarmicus)diets material on each collection to prevent duplication in sub- have been in the northern periphery of its range sequent samples. Pellets were broken up with tweezers and individual bone remnants were separated from the in southern Europe, eastern Sahara, and Israel feather matrix. (Massa et al. 1991, Yosef 1991, Goodman and All avian osteological material was identified using Haynes 1992). Although Peregrine Falcons share comparativeskeletons in the South African Museum and many parts of their distribution with other large additional material loaned from the Transvaal Museum, falconshaving similar resourceneeds (Cade 1982), Pretoria and the National Museum, Bloemfontein. Indi- vidual body parts were separatedaccording to taxon and few studieshave comparedPeregrine Falcon diets recorded as the number of identified specimens (NISP). with thoseof sympatriccongeners (Cade 1960, Por- The minimum number of individuals (MNI) was calcu- ter and White 1973). This paperdescribes the diets lated from the most common body part among the iden- of breeding Peregrine and Lanner Falcons in tified specimens,after accounting for paired elements South Africa, the and diversityof prey taken (Klein and Cruz-Uribe 1984). Incompletely ossified boneswere consideredto be thoseof juveniles. Mammal by Peregrine Falcons in different environments, remains were identified according to cranio-dental char- defines the feeding niches of the two speciesin an acteristics. area of sympatry and assessesthe potential for and mammal remains were identified to the low- competitionbetween these two congenersin terms est possibletaxonomic level. All prey were assignedto a of diet overlap. size classbased on bone size for unidentified prey and on body massdata from the literature (Brownet al. 1982, METHODS Smithers1983, Maclean 1993) when prey were identified to at least the family level. Size classeswere small (up to We conductedour studyat sitesin three areasof South sparrowsize, averageabout 20 g), small to medium (star- Africa. Sixteen pairs of Peregrine Falconswere studied ling size, averageabout 60 g), medium (dove size, aver- on the Cape Peninsula,Western Cape Province(34ø10'S, age about 130 g), medium-large (large dove size,average 18ø25'E) from 1989-95. This area had both inland and about 220 g), large (pigeon sizeand larger, averageabout coastalnest sitesdispersed in a mosaicof urban, forest, 350 g), and very large (francolin size and larger, average and heathland habitats. Altitude varied from 0-1100 m about 600 g). In biomasscalculations, unidentified prey and the climatewas temperate with locallyvariable winter were given these averagemass values. Mean body mass rainfall (400-2000 mm per year). Seven pairs of Pere- valuesfrom the literature were used for prey identified grine Falconswere also monitored annually from 1989- to species.Prey identified only to higher taxonomiclevels 95 on the lower Orange River, Northern Cape Province were assignedmass estimates based on published weights (28ø30'S,17ø00'-20ø40'E). This is an arid, hilly to moun- of similar or related forms. tainous area, sparselyvegetated except for narrow strips Arthropod, amphibian and reptile remains were iden- of riparian bush or irrigated croplandsalong the river tified at a grosslevel only. A small number of nonavian banks. Altitude varied from 30-700 m and the climate remains were consideredto be unlikely prey of falcons, washot and dry (mean annualrainfall 60-130 mm). Prey particularlyPeregrine Falcons, on the basisof sizeor hab- remains were collected at seven nest it. For example, observationssuggest that African Pere- sites in the Soutpansberg range, Northern Province grine Falconshunt only flying prey (Hustler 1983, Tar- (23ø00'S, 29ø40'E) from 1988-95, and the diet of these boton 1984) and, therefore, are not likely to take birds was compared with that of a sympatricpopulation terrestrial mammals and reptiles. These taxa (Appendix of nine pairs of Lanner Falconsover a 3-yTperiod from 1) were excludedfrom the analyses.Some relatively com- 1991-93. Evergreen forest and moist woodland occurred plete, easilyidentifiable prey were identified in the field along the eastern foot of the Soutpansbergescarpment, and discarded. These were not included in the data and, with dry thornveldin the westand grasslandwith patches to avoid duplication, only contributed to the total num- of scrub forest and protea woodland along the summit ber of identified prey where fewer individualsof the rel- of the mountains (Tarboton 1990). Altitude ranged from evant taxon were subsequentlyidentified from other re- 900-1700 m. There was low to moderate summer rainfall mains in the corresponding collection. (400-1000 mm per year), increasingon a gradient from Pluckings were identified using study skins in the 192 JE•i•i•s •4I) AwRY VoL. 33, No. 3

Table 1. Samplesof prey remainscollected each year at falcon nestsites in three areasof SouthAfrica. NISP denotes the numberof identifiedspecimens in the sampleand MNI denotesthe numberof prey individualsidentified. The MNI totalsare the sum of prey individualsidentified from skeletalremains, pluckings and remainsidentified in situ and discarded.Remains collected from consecutiveyears were pooled where samplesizes were small.

NUMBER NUMBER MNI YEAR COLLECTIONS SITES NISP SKELETAL PLUCKINGS DISCARDED TOTAL

Cape Peninsula Peregrine Falcons 1989 8 I 318 37 3 22 62 1990 11 4 608 83 6 8 97 1991 7 2 187 33 8 0 41 1992 11 5 412 65 6 1 72 1993 10 5 46 63 10 1 74 1994 22 7 417 88 30 0 118 1995 7 4 161 44 4 1 49 Overall 76 10 2649 413 67 33 513

OrangeRiver PeregrineFalcons 1989-90 7 5 680 102 10 0 112 1991-92 4 4 410 72 8 I 81 1993-95 4 2 333 48 7 0 55 Overall 15 7 1423 222 25 I 248 SoutpansbergPeregrine Falcons 1988-90 5 4 245 36 6 4 46 1991 13 6 719 113 8 0 121 1992 13 5 378 89 5 0 94 1993 14 4 585 93 10 0 103 1994-95 4 3 225 41 2 0 43 Overall 49 7 2152 372 31 4 407

SoutpansbergLannerFalcons 1991 14 9 178 65 5 0 7O 1992 15 7 302 62 3 0 65 1993 10 7 188 38 0 I 39 Overall 39 9 668 165 8 I 174

South African Museum. No attempt was made to deter- x 40 binocularsor a 20-60x spotting scopefrom dis- mine the number of individualsrepresented by the sum tances of 200-400 m. Whenever a falcon was seen with of pluckingsof a particularspecies collected in each sam- food, an effort wasmade to identify or at least estimate ple. To prevent duplication in these samples, species the size of its prey. Size classesused were the same as identified from feather remains contributed one individ- those used for prey remains. Only largely intact prey ual to the total for a collection,but only when the rele- could be identified or size classified. vant taxon was not recorded in the skeletal remains.Ju- Indices of diet breadth and overlap were calculated veniles were identified from plumage characteristics basedon the relativefrequency of taxa identifiedin prey where thesediffered from adult birds,and from the pre- remains at the specieslevel wherever possibleto refine dominance of sheathed or incompletely grown tail or the quality of these estimates (Greene and Jaksi• 1983, flight t•athers, which indicated that nestlingor recently Sherry 1990). Diet breadth (BA)was calculated using Lev- fledged individuals had been taken. ins' (1968) standardized formula: Some dietary data were obtained from field observa- BA = B- l/n- 1, nons. These were used to determine biasin the analyses of prey remains (Collopy 1983, Rosenbergand Cooper where B = 1/E p•2and p• is the proportionof the diet 1990). Nearly 1000 hr of observationswere made at Per- contributed by the ith taxon. Valuesof Ba range from 0- egrine Falcon nest siteson the Cape Peninsula,and over 1, with larger valuesindicating a broader diet. Also, the 200 hr were obtained each at nest siteson the Orange number of frequently used taxa (thosecomprising 3% or River and in the Soutpansbergand at Lanner Falconsites more of the total number of identified prey;Krebs 1989) in the Soutpansberg.Observations were made using 10 was tallied for each falcon population as an additional SEPTEMBER 1999 PEREGRINE AND LANNER DIETS IN SOUTH AFRICA 193 estimate of diet breadth. Diet overlap was measured us- low, open grasslandwith patches of short deciduous for- ing Morisita's (1959) index of similarity: est. Deciduous woodland comprised short, moderately closed woodland and occurred on the northern backslo- pes and on the middle and upper slopesat the western C= 2 E P•jP•k/E pv[(N• 1 +•Pik [(N• l) end of the range. The lower scree slopesof the escarp- ment were thickly vegetated with low, closed woodland where p, and Pi• are the proportionsthat taxon i makes or moist thornscrub in the east and short, moderately up of the dietsof speciesj and k, respectively,n•j and n• are the number of individuals of taxon i in the diets of closedwoodland or dry thornscrub further west. One or two siteswere selected as typical of each habitat type and speciesj and k, respectively,and N• and N• are the total at least three 1 km line transects were walked at these number of individuals in the diets of speciesj and k, respectively.Values of C range from 0-1, with larger val- siteseach breeding seasonfrom 1991-93. Transect sites ues indicating a greater dietary overlap. This index is were located between 150-3000 m from the main escarp- considered the least prone to biasesassociated with sam- ment. Transectswere completed at various times of day, ple size and the number of resourcesused (Smith and but mostlyin the morning or in late afternoon. Transects Zaret 1982). Multivariate cluster analyseswere conducted were walked briskly so only the birds which were con- to examine qualitativedifferences in the diet of sympatric spicuouslyactive in the area, and hence most likely to Peregrine and Lanner Falcons,using the PRIMER soft- provide falconswith hunting opportunities,were record- ware package (Plymouth Marine Laboratory,U.K.). Diet ed. High, overflying birds unlikely to be resident in the composition data were compared using the Bray-Curtis habitat being sampled were not counted, and species similarity coefficient (using group average linking) to consideredtoo large for falconsto catch and subdueun- generate a dendrogram of hierarchical clustersand a der normal circumstances()800 g in weight) were also processof nonmetric multidimensional scaling to gen- excluded. The species, number of individuals and ap- erate an ordination plot. proximate perpendicular distancefrom the transectline To examine the possiblecompetitive influence of near- were recorded for each sighting (Bibby et al. 1992). by Lanner Falcon pairs on the diet of SoutpansbergPer- egrine Falconsand viceversa, the food niche parameters RESULTS of falcon pairs with congenersas particularly close neigh- bors (pairs (1 km apart, Peregrine FalconsN = 3, av- A minimum of 1168 individuals of at least 82 erage distanceto nearestLanner Falcon pair = 0.6 km, speciesfrom 34 families were identified as prey Lanner Falcons N = 4, average distance to nearest Per- from the remains collected at Peregrine Falcon egrine Falconpair = 0.7 km) were comparedwith those of relatively isolated pairs ()2 km apart, Peregrine Fal- nest sites, and a minimum of 174 individuals of at cons N = 4, averagedistance to nearest Lanner Falcon least 24 speciesfrom 15 families were identified pair = 6.5 km, Lanner FalconsN = 5, averagedistance from Lanner Falcon prey remains (Table 1, Ap- to nearestPeregrine Falcon pair = 4.7 km). The mean pendix 1). Samples of remains collected at Pere- distancebetween conspecific pairs was 9.7 km for Pere- grine and 5.2 km for Lanner Falcons. grine Falcon sitesin the Soutpansbergwere more Two sampling techniqueswere used to estimate the substantialthan those from sympatricLanner Fal- abundance and distribution of potential avian prey in the con sites,yielding more specimensper collection vicinityof falcon nest sitesin an attemptto make quali- (Peregrine Falcon i = 45.6, range = 2-211, N = tative assessmentsof prey selectionand habitat use. An estimate of local prey availabilitywas made at nest cliffs 47 collections; Lanner Falcon i = 18.6, range = by countingthe number of potentialprey per hour which 1-60, N = 36 collections;Mann-Whitney Z = 3.01, flew acrossa samplingarea prescribedby a 1 mr frame. P = 0.003) and more identified individuals per col- This frame waspositioned at the top of the cliff, looking lection (Peregrine Falcon • = 8.0, range = 1-31, out and down, or on the screeslope below the face look- ing up at the crest. An observer sat 2 m behind the N = 49 collections;Lanner Falcon i = 4.5, range frame, looking through it, to make the count. Prey = 1-10, N = 39 collections; Z = 2.77, P = 0.006). countswere made at different times of the day at a cross- Peregrine Falcons were observed catching or section of nest sites on the Cape Peninsula and in the feeding on 296 prey individuals, about half of Soutpansbergonly, mostlyfrom 1991-93. Line-transect counts were walked in the Soutpansberg which could be identified at least to family. Fifteen to estimate bird densities in seven habitats identified in speciesfrom eight families were identified. Only terms of a broad-scale classification of the character and one species,Black-eyed Bulbul (Pycnonotusbarba- structure of vegetation present (sensuEdwards 1983). tus)was observed as Peregrine Falcon prey and not Plains to the south of the mountain range comprised ei- ther short, semi-opensavanna woodland or short, sparse recorded in prey remains.Only 36 prey individuals denuded woodland, often found adjacentto rural settle- were identified during observationsat Lanner Fal- ments and heavily impacted by clearing and cultivation. con sites in the Soutpansberg.These comprised Forest and plantation included tall, closedstands of both one locust,one 1-3 day-oldchicken (Gallusgallus), indigenousevergreen and exotic eucalyptusor pine for- est. Areas of montane grasslandand scrub forest were one Streptopeliadove, one unidentified murid ro- typicalof the upper slopesof the mountainsand featured dent, seven unidentified small birds, nine small- 194 JENKINSAND AVERY VOL.33, NO. 3 medium birds, 12 medium birds, and four medi- grine Falconson the Orange River. The relative um-large birds. frequencyof juvenilesin the diet of Cape Penin- Becauserelatively few prey items were identi- sula Peregrine Falconsincreased through the fied in the field, and the sampleof observedprey breedingseason from 1.7% by frequencyin late at Lanner Falconsites was so small,the qualityof October to 6.2% in early November,6.8% in late the diet information providedby the analysisof November,and 8.1% in early December.A high prey remains could only properly be assessedin proportion of the remains recovered from Lanner terms of the size of prey observedat Peregrine Falconnests were young chickens,and juvenile Falconsites in eacharea. These comparisons sug- birds comprised39.7% of identified prey from gestedthat remainscollected on the Cape Pen- these sites. insula provided the least accurate diet estimate, Columbidsmade up the majorityof preytaken while thosefrom sitesin the Soutpansbergwere by PeregrineFalcons in all three studyareas (Table the most accurate (Fig. 1). On average,prey re- 2). Other consistentlyimportant taxa were species mains underrepresentedsmaller prey by about in the familiesApodidae, Sturnidae and Ploceidae. 10% and overrepresentedlarger prey by about SoutpansbergLanner Falcons took mainly phasian- 8%. Taxonomicallythe sampleswere similar with ids,charadriids and columbids.The averagesize of birds, and particularlycolumbids, predominating prey taken by Peregrine Falconsin each of the in all. three studyareas varied (Kruskal-WallisH = 26.8, The relativeimportance of key speciesin the di- P < 0.001). CapePeninsula Peregrine Falcons took eu of both falconswas fairly consistentbetween significantlylarger prey (averagemass = 144.5g, years,so data on prey remainsfor eachyear were range = 15-390 g, N = 513 individuals) than pooled.Dominant species in the diet of Peregrine Orange River PeregrineFalcons (average mass = Falconson the Cape Peninsulawere Laughing 128.3 g, range = 3-347 g, N = 248 individuals; Dove (Streptopeliasenegalensis, 28% by frequency Mann-WhitneyZ = 4.33,P < 0.001) and Peregrine and 20% by biomass),European Starling (Sturnus Falconsin the Soutpansberg(average mass = 123.3 vulgaris,14 and 7%, respectively),and CapeTurtle g, range = 3-600 g, N = 407 individuals;Z = 4.11, Dove (Streptopeliacapicola). Peregrine Falconson P • 0.001). On average,Soutpansberg Peregrine the Orange River took Rock Pigeons ( Falconstook significantlysmaller prey (average guinea,15 and 41%, respectively),Namaqua Sand- mass= 116.9g, range = 3-600 g, N = 318 individ- grouse(Pterocles namaqua, 10 and 13%, respective- uals)than syrnpatricLanner Falcons (average mass ly), and LaughingDove (8 and 6%, respectively). = 123.7 g, range = 3-500 g, N = 174 individuals; Peregrine Falcons in the Soutpansbergtook Z = 2.41, P = 0.02). PeregrineFalcon diet includ- LaughingDoves (19 and 16%, respectively),Red- ed a greaterproportion of small and large prey faced Mousebirds (Urocoliusindicus, 10 and 4%, re- individuals, while Lanner Falcons tended to con- spectively),and Red-eyedDoves (Streptopelia semi- centrate on medium-sizedprey. torquata,8 and 16%,respectively). During 1991-93, Of the three Peregrine populations studied, Peregrine Falconsin the Soutpansbergpreyed those on the Cape Peninsulahad the leastdiverse mainly on doves(35 and 45%, respectively)and diet (30 specieswere identified from preyremains; mousebirds(15 and 7%, respectively),while syrn- Appendix 1) and the narrowestdiet breadth (BA patric Lanner Falcons took chickens (40 and 30%, = 0.17). PeregrineFalcons in the Soutpansberg respectively),doves (20 and 25%, respectively), fed on a wide diversityof prey (--•50 species)but and Crowned Plovers (Vanellus coronatus,9 and concentrated on a few taxa and had a moderate 12%, respectively;Appendix 1). diet breadth (BA = 0.21). On the Orange River, The relativeimportance of juvenile birds in the only 38 specieswere taken but there was lessem- dietsof PeregrineFalcons varied significantly (X 2 phasison particular families so the diet was rela- = 19.5, P < 0.001) amongthe three studyareas. tively broad-based(BA = 0.34). Most of the juveniles recorded were columbids Overall,Peregrine Falcons in the Soutpansberg (Appendix1) andjuveniles were more frequent in had a broaderfeeding niche than syrnpatricLan- the diet on the Cape Peninsula(32 of 513 individ- ner Falconsand there wassome overlap in the di- ualsor 6.2%) than in the Soutpansberg(3 of 407 ets of the two species(C = 0.34). On a per site or 0.7%; X2• with Yates' correction = 15.3, P < basis(excluding one peregrine site and one lanner 0.001), and comprised 3.2% of the diet of Pere- site with insufficientsamples), the difference in SEPTEMBER 1999 PEREGRINE AND LANNER DIETS IN SOUTH AFRICA 195

5O Cape Peninsula 4O

t- 3O

2O

o small small-medium medium medium-large large 60

5o Orange River 40

t-

30

20

10

o small small-medium medium medium-large large 40

Soutpansberg 3o

t-

2o

lO

small small-medium medium medium-large large prey size classes

prey remains I observations

Figure 1. The relative frequencyof prey of different size classesin the diets of Peregrine Falconsin three areasof South Africa determined from the analysisof prey remainsand direct observations.Prey remainsdiffered significantly from observationson the Cape Peninsula (remains N = 513, observationsN = 159, X'•4= 42.0, P • 0.001) and on the Orange River (remainsN = 248, observationsN -- 81, X24= 14.0, P = 0.007) but not in the Soutpansberg (remains N = 407, observationsN = 56, X24= 7.3, P = 0.12). 196 JENKINSAND AVERY VOL. 33, No. 3

Table 2. Summary of falcon diets from the analysisof prey remains collectedat nest sites (see Appendix I for raw data). The importanceof each taxon is expressedin terms of its relativefrequency in the sampleof individuals(%F) and in termsof its contributionto the total biomassof prey in eachsample (%M). Familiescomprising 1% or more of the diet in each sampleare listed.Peregrine Falcons (all years):Cape PeninsulaN = 513 individuals,74 107 g; Orange River N = 248 individuals,31 811 g; and SoutpansbergN = 407 individuals,50 193 g. Soutpansberg(1991- 93): PeregrineFalcons N = 318 individuals,37 188 g; Lanner FalconsN = 174 individuals,21 532 g.

PEREGRINEFALCONS (1989--95) SOUTPANSBERG(1991--93)

CAPE ORANGE PENINSULA RIVER SOUTPANSBERG PEREGRINES LANNERS

FAMILY %F %m %F %M %F %m %F %m %F %m

Phasianidae .... 1.5 3.6 1.3 3.6 40.2 37.4 Charadriidae 1.0 ------3.9 5.3 3.5 4.9 9.2 12.4 Recurvirostridae ...... 1.2 1.6 Glareolidae .... 1.0 ..... Laridae -- -- 2.4 2.1 ...... Pteroclididae -- -- 9.7 14.0 1.0 1.9 -- 1.3 -- -- 66.1 85.1 37.5 69.9 48.2 68.4 46.5 67.2 28.2 39.1 Ps•ttacidae 1.0 ...... Apodidae 6.2 1.9 10.1 4.1 6.1 2.9 5.4 2.7 1.2 -- Coliidae 1.0 -- 1.2 -- 13.5 6.1 14.8 7.0 -- -- Alaudidae -- -- 7.7 1.0 1.0 -- 1.3 ------Hlrundinidae -- -- 6.1 ...... Sturnidae 15.0 8.2 1.6 1.4 3.2 2.3 3.1 2.1 -- -- Ploceidae 4.5 1.1 4.4 -- 4.2 1.0 4.4 1.2 4.6 1.2 -- -- 1.2 ...... Frlngillidae -- -- 1.6 -- 1.0 ------1.2 -- Molossidae -- -- 6.5 ...... Pteropodidae .... 1.5 1.6 1.3 1.4 -- -- Muridae ...... 1.2 --

diet breadth was statisticallysignificant (Peregrine rates in terms of numbers of individuals counted Falcon averageBA = 0.14, range = 0.11-0.18, N = (Cape Peninsula average = 116.3 birds per hour, 6; Lanner Falcon averageBA = 0.04, range = 0.01- range = 2-542 birds per hour, N = 64 counts; 0.14, N = 8; Mann-Whitney Z = 2.52, P = 0.01). Soutpansberg average = 119.0 birds per hour, Cluster analyseson diet composition data clearly range = 0-929 birds per hour, N -- 114 counts).A separatedPeregrine and Lanner Falconsites (Fig. significantlygreater biomassof birds per hour was 2). Food-nichewidths of falcon pairs with pairs of recorded in counts on the Cape Peninsula (aver- congeners as close neighbors did not differ signif- age = 8567 g per hour, range = 405-33 479 g per icantly from those of more isolated pairs (Pere- hour, N = 64 counts) than in the Soutpansberg grine Falcon pairs with close neighboring Lanner (average= 7104 g per hour, range -- 0-56 669 g Falcon pairs average BA = 0.14, N = 3, isolated per hour, N = 114 counts;Mann-Whitney Z = 3.09, pairs averageBA = 0.14, N = 3, Z = 1.00, P = 1.00; P = 0.002). Mostly cliff-dwelling specieswere in- Lanner Falcon pairs with close neighboring Pere- cluded in these counts (e.g., Rock Pigeons,Alpine grine Falcon pairs average BA ----0.08, N = 4, iso- Swifts [Apusmelba] and Red-wingedStarlings [On- lated pairs averageBA = 0.05, N = 4, Z = 0.72, P ychognathusmorio]). Larger species were more = 0.47). However, collectively,interspecific diet abundant at cliffs on the Cape Peninsula than in overlap was greatest between pairs with close the Soutpansberg(e.g., Rock Pigeonswere record- neighboring congeners (close pairs C = 0.44, dis- ed in 72 vs. 18% of counts at cliffs on the Cape tant pairs C = 0.26). Peninsula and in Soutpansberg,respectively, with Counts at cliffs on the Cape Peninsula and in an average of 3.9 vs. 0.6 individuals per count), the Soutpansberghad similar averageprey passage whereas aerial insectivores were more abundant in SE]'TE•aBER 1999 PEP.•Cea•E AND LANNER DIETS IN SOUTH AFRICA 197

LCl (14)

LC2 (16)

LC3 (35) *

LC4 (15) *

LC5 (26)

LC6 (32)

LC7 (20) *

LC8 (12) *

PC2 (15) *

PC1 (92)

Pc3 (110) *

Pc4 (34)

Pcs (so)

PC6 (17) * i i i i i i i i i 20 30 40 •0 •0 70 •0 •0 too Bmy-C-rti5 similarity

Figure 2. Resultsof cluster analyseson diet compositiondata for Peregrine (P) and Lanner (L) Falcon sitesin the Soutpansberg.The dendrogramgroups sites according to the Bray-Curtissimilarity coefficient (with the number of preyindividuals identified from eachsite in parentheses).Sites with closeneighboring pairs of congenersare marked with an asterisk.The ordination plot (inset) illustrates the relatedness of each site to the others in terms of a multidimensional scalingprocedure. the Soutpansberg(e.g., Rock Martins [Hirundofu- na and denuded woodland to only 189 individu- ligula] were recordedin 19 vs. 52%, respectively, als/kin 2 in moist thornscrub. with an average of 0.8 vs. 10.2 individuals per Some specieswhich were recorded regularly in count). the diet of falcons in the Soutpansbergwere not Of the habitatsidentified in the Soutpansberg, common in the environment. In particular, Red- the woodlandson the plain below the escarpment facedMousebirds made up about 10% of the Per- supported the highest diversityand the greatest egrine Falcon diet by frequency but were not en- numberand biomassof birds (average-- 6-7 spe- countered on any of the transect counts. Three cies, 20-30 individuals and 800-1300 g per tran- species (Chestnut-backed Finchlark [Eremopterix sect, respectively),while the dense thornscrub on leucotis],Melba Finch [Pytilia melba]and Blue Wax- the scree slopes of the mountains featured the bill [ Uraeginthusangolensis]) were common on the most depauperateavifauna (average= <2 species, plains at the foot of the escarpment(i = 13.6, 1.5, 2-6 individualsand 40-200 g per transect).Tran- and 8.5 individuals, respectively,of each species sect data extrapolated to avian density estimates were recorded per transectthrough savannaor de- ranging from 1441-1690 individuals/kin9 in savan- nuded woodland) but were infrequent or did not 198 JENKINSAND AVERY VOL. 33, No. 3

Table 3. Foraging habitat use by breeding Peregrine and Lanner Falconsin the Soutpansbergas inferred from the habitat affinities of prey speciesidentified in food remains. (A) only indicatesprey specieswhich were recorded in passage-rateor line-transectcounts. (B) combinesthe speciesin (A) with a number of important prey specieswhich were not sighted on sample counts (e.g., domestic , Namaqua Dove and Red-facedMousebird). Habitat preferencesin speciesin (A) were determined from count data. The additional speciesin (B) were assignedto habitatson the basisof incidental observationsin the area. Both comparativedistributions were significantlydifferent: (A) X26= 12.6, P = 0.049; (B) X26= 79.7, P • 0.001.

% FREQUENCYIN % FREQUENCYIN PEREGRINE DIET LANNER DIET FAVORED HABITAT OF PREYSPECIES (A) (B) (A) (B)

Savanna woodland 26.7 39.0 19.0 19.5 Denuded woodland 7.9 12.3 12.1 49.4 Forest and plantation 0.3 0.3 -- -- Grassland and scrub forest 5.4 5.4 1.4 1.4 Moist thornscrub .... Deciduous woodland 0.3 0.3 1.4 1.4 Dry thornscrub 0.3 0.3 • -- Cliffs 9.1 9.1 5.2 5.2

occur in falcon diets, perhaps becausethese small DISCUSSION birds were underrepresentedin prey remains.Sim- Columbids and particularly Streptopeliadoves ap- ilarly, Black-eyedBulbuls were ubiquitous in the pear to be the staple food of Peregrine Falconsin area (0.5-1.6 individualsper transectin all habitats southern Africa (Hustler 1983, Tarboton 1984, except denuded woodland), but were recorded as Mendelsohn 1988, this study).A similarpreference falcon prey only once. There was a significantdif- for columbiformshas been recorded in many oth- ference in the foraging habitatsused by Peregrine er areas of the Peregrine Falcon's distribution and Lanner Falcons in the Soutpansberg,as in- where pigeonsand dovesare availableprey (Cade ferred from the habitat preferences of their prey 1982, Ratcliffe 1993). Small, aerially dexterous spe- (Table 3). While both speciesfavored the wooded cies such as swifts and appear more consis- plains below the mountain range, Peregrine Fal- tently in the diet of African PeregrineFalcons than cons took more woodland speciesfrom relatively has generally been reported for other populations pristine habitats whereas Lanner Falcons took (Hustler 1983, Brown 1988, Mendelsohn 1988). open-countryspecies from denudedwoodland and Conversely,relatively large, terrestrialspecies such free-range chickens from around human settle- as anatids,galliforms, and charadriiformsare less merits.Also, PeregrineFalcons preyed more heavi- frequently taken in Africa than in other areas ly than Lanner Falcons on cliff-dwelling species (Cade 1960, Porter and White 1973, Pruett-Jones (Table 3). SoutpansbergPeregrine Falcons took et al. 1981, Mearns 1983, Ratcliffe 1993). These significantlymore 'commuter species' (aerial in- differencesmay partly reflect differencesin prey sectivores, migrants or species which regularly availability,but may also be due to a greater ten- commute between distant resources, e.g., sand- dency for African PeregrineFalcons (and perhaps grouse, large columbids) than sympatricLanner other small,tropical forms of the species)to 'catch Falcons (76 of 318 or 23.9% vs. 19 of 174 or 10.9% and carry' prey rather than strike it to the ground commuters,respectively; X21 = 11.3, P • 0.001), and retrieve it (Cade 1982). but fewer overall than Cape Peninsula Peregrine Three Streptopeliadoves comprised nearly half Falcons(27.0 vs. 31.6% commuters,respectively) (by frequencyand mass)of the diet of Peregrine and significantlyfewer than Peregrine Falconson Falconson the Cape Peninsula.British Peregrine the Orange River (110 of 407 or 27.0% vs. 114 of Falconsspecialize to a similardegree on larger col- 248 or 46.0% commuters,respectively; X21 = 23.7, urnbids (Mearns 1983, Ratcliffe 1993). Typical of P • 0.001). peri-urban peregrine populations around the SEPTEMBER 1999 PEREGRINE AND LANNER DIETS IN SOUTH AFRICA 199 world (Cade and Bird 1990), Cape Peninsulapairs the cliff-line to the screeslopes below, as suggested also took substantial numbers of commuting by Hustler (1983). or domesticpigeons (C01umbalivia) and European Diet breadth indices calculated for South Afri- Starlings,and occasionallypreyed on escapedavi- can Peregrine Falcons are comparable with equiv- ary birds. Inexperienced, newly-fledgedbirds are alent data for populations in other areas (Table 4). particularly vulnerable to by raptors These figures suggestthat peregrines in temperate (Newton and Marquiss 1982, Rosenfield et al. areas (e.g., the Cape Peninsula [this study], south- 1995), and young dovesand starlingswere taken ern Scotland [Mearns 1983], Victoria, Australia quite frequentlyby Cape PeninsulaPeregrine Fal- [Pruett-Joneset al. 1981] ) are relativelyspecialized cons. As observed in other temperate falcon pop- feeders. In contrast, Peregrine Falcons appear to ulations (Newton et al. 1984, Parr 1985), juvenile become more generalized in extreme environ- birds became more prevalent in the diet as the ments such as deserts (e.g., this study), taiga and breeding cycle progressed.This supportsthe con- tundra (Cade 1960, Cade et al. 1968). tention that breeding by Peregrine Falconson the Relatively small samples of identifiable prey re- Cap•Peninsula may in partbe timed to exploitthe mains were recovered from Lanner Falcon nest period of maximum productivityof their principal sites, perhaps because much of the prey they con- prey species(Jenkins 1991, 1998) sumed was completely digestible. Few data are Streptopeliadoves featured least prominently in available on the diet of southern African Lanner the diet of Peregrine Falconson the Orange River, Falcons. The only quantitative studies, based on the most generalized feeders of the three Pere- small samples, suggestthat phasianids (including grine Falcon populations studied. These falcons domestic fowl) and columbids are important prey usuallyforaged in the river valley (Jenkins 1995), groups (Barbour 1971, Tarboton and Allan 1984, which was the focus of bird movements in the area. Kemp 1993). Lanner Falcons in the Soutpansberg Aerial insectivores (swifts, hirundines, and micro- were relatively specialized feeders, with young chiropteran bats) and obligate drinkers (sand- chickenscomprising nearly 40% of the prey re- grouse, columbids, and granivorous ) corded in remains. Given the greater susceptibility were the groups most frequently taken and were of juvenile bones to damage and acid erosion, probably exposed to attack crossingrocky gorges chickens may have been underrepresented in or open stretchesof the river, or travelingover arid these samples.Columbids were alsoimportant, and flats adjacent to the floodplain. contributed more than phasianids to the biomass Soutpansberg Peregrine Falcons preyed on a of identified prey. Sourcesof free-ranging wide variety of species,most of which were secured were a considerable distance from most Lanner in bird-rich woodlands at least 2 km from the base Falcon nest sites, and their frequency in prey re- of the escarpment. Relatively fewer transient or mains suggeststhat Lanner Falcon hunting ranges commuter species,and more sedentary residents, extended well beyond those of nearby Peregrine were taken by these pairs than by Peregrine Fal- Falcon pairs. One radio-tracked male Lanner Fal- cons in the other two areas. However, the wood- con foraged over rural settlement areas >10 km land speciesrecorded as prey were generally those from the main escarpment.Terrestrial or cursorial likely to fly furthest from, and highest above, the speciespredominated in the diet of Soutpansberg protective canopy of trees (e.g., Red-faced Mouse- Lanner Falcons,and most prey were probably tak- birds, Lamprotornisstarlings). Mousebirds were en on or close to the ground. more common Peregrine Falcon prey in the Sout- Southern African Peregrine and Lanner Falcons pansberg than has been found in studiesat other are of approximately equal body mass and com- woodland sites,whereas rollers and woodpeckers parisons of key food handling and flight perfor- were not taken (Hustler 1983, Tarboton 1984), mance measurementsyield ratios consistentlylow- perhapsbecause they were lesscommon in the en- er than values traditionally considered minimum vironment. Francolins are among the largest spe- for noncompetitive coexistence (Jenkins 1998). cies regularly taken by African Peregrine Falcons, Despite these structural similarities, diet overlap and those recorded as prey of SoutpansbergPere- between SoutpansbergPeregrine and Lanner Fal- grine Falconswere probably all caught by females. cons was moderate relative to those calculated us- These essentially cursorial birds may have been ing data from other studiesof Peregrine Falcons caught in flight as they descendedfrom the top of and sympatric congeners, and fell well below the 200 JENraNSA•r• AVERY Vo•.. 33, No. 3

Table 4. Indices of Peregrine Falcon diet breadth for a number of locations around the world, and measuresof diet overlap betweenPeregrine Falconsand sympatriccongeners.

NUMBER OF NUMBER OF DIET BREADTH STUDYPOPULATION TAXA USED TAXA >3% (BA)

Southern Scotland I 90 5 0.04 V•ctoria, Australia 2 77 6 0.12 Cape Peninsula,South Africa • 39 6 0.17 Soutpansberg,South Africa 4 63 9 0.21 West Greenland 5 11 6 0.24 NW Territories, Canada 6 19 7 0.34 OrangeRiver, South Africa 7 45 10 0.34 Talga zone, Alaska8 61 11 0.81 SympatricPeregrine Falconsand (F. rusticolus),Colville River,Alaska '• Peregrine Falcon 55 12 0.33 55 2 0.004 Diet overlap: C = 0.07 SympatricPeregrine and Lanner Falcons, Soutpansberg,South Africa 1ø Peregrine Falcon 67 10 0.20 Lanner Falcon 67 6 0.07 Diet overlap: C = 0.34 SympatricPeregrine and Prairie Falcons(F. mexicanus),Wasatch Mountains, Utah TM Peregrine Falcon 33 9 0.31 33 6 0.26 Diet overlap: C = 0.58 Food niche parameterswere calculatedusing data from: 1 Mearns 1983, 9 Pruett-Joneset al. 1981, • this study,4 this study(all years), 5 Rosenfieldet al. 1995,6 Bradleyand Oliphant1991, 7 thisstudy, 8 Cadeet al. 1968,9 Cadeand White 1971,l0 thisstudy (1991-93), li Porter and White 1973.

suggestedcritical value for competing speciesof prey remains collected at a Peregrine Falcon nest site in about 0.62 (Bosakowski and Smith 1992). Also, the Soutpansbergin 1988. South African National Parks allowed us to conduct researchin the AugrabiesFalls Na- while competition theory predicts narrower food tional Park. Tom Cade, Rob Davies,James Enderson, Phil niches and reduced diet overlap between close Hockey, Rob Simmons, and Reuven Yosef made useful pairs of competing speciesvs. distantpairs (Nilsson commentson earlier drafts of this manuscript.This re- 1984, KorpimSki1987), the diet of falcon pairs in searchwas partly funded by the Foundationfor Research this studywas not obviouslyaffected by the pres- Development. ence of nearby pairs of congeners. Therefore, there was little evidence suggestingactive compe- LITERATURE CITED tition for food between Peregrine and Lanner Fal- BARBOUR,D.Y. 1971. Notes on the breedingof the lanner. cons in the Soutpansberg.As concluded in previ- Bokmakierie 23:2-5. ous studies of Peregrine Falcons and sympatric BIBBY,CJ., N.D. BURGESSAND D.A. HILL. 1992. Bird cen- congeners (e.g., Cade 1960, Porter and White sus techniques.Academic Press,London, U.K. 1973), relativelysubtle differences in morphology, BOSAKOWSm,T. AND D.G. SMrr•. 1992. Comparativediets flying performance, and hunting techniques(Jen- of sympatricnesting raptors in the easterndeciduous kins 1995) are evidentlysufficient to segregatefor- forest biome. Can.J. Zool.70:984-992. aging habitat and diet of Peregrine and Lanner BRADLEY,M. AND L.W. OLIPHANT. 1991. The diet of Per- Falcons in South Africa. egrine Falcons in Rankin Inlet, North-West Territo- ACKNOWLEDGMENTS ries: an unusually high proportion of mammalian Dave Allan, Zelda Bate, Tim Wagner, and Anthonyvm• prey. Condor93:193--197. Zyl assistedin the field. Dave Allan kindly contributed BROWN,CJ. 1988. Diet of Booted Eaglesm•d Peregrine SEPTEMBER 1999 PEREGRINE AND LANNER DIETS IN SOUTH AFRICA 201

Falcons in the Waterberg Plateau Park, Namibia. Ga- John Voelcker Bird Book Fund, Cape Town, South bar 3:26-28. Africa. BROWN, L.H., M.K. Um•AN AND K. NEWMAN. 1982. The MARCHANT,S. AND P.J. HIGGINS [EDS]. 1993. Handbook birds of Africa. Vol. 1. Academic Press, London, U.K. of Australian, New Zealand and Antarctic birds. Vol CADE,T.J. 1960. Ecologyof the peregrine and Gyrfalcon 2. Oxford Univ. Press, Auckland, New Zealand. populationsin Alaska. Univ. CaliforniaPubl. Zool.63: MASSA,B., F. Lo V•VO, M. SIRACUSA,M. AND A. CIACCIO. 151-290. 1991. II Lanario Falco biarmicusfeldeggi (Schlegel) •n --. 1982. The falcons of the world. Collins, London, Italia: status,biologia e tassonomia.Naturalista Sicil.S. U.K. 15:27-63. •AND C.M. WHITE.1971. Cliff-nestingraptors and MEARNS,R. 1983. The diet of the Peregrine Falcon Falco ravensalong the Colville River in ArcticAlaska. Living peregrinusin south Scodandduring the breeding sea- Bird 10:107-150. son. Bird Study30:81-90. --AND D.M. BIRD. 1990. Peregrine Falcons,Falcoper- MENDELSOHN,J.M. 1988. The statusand biology of the egrinus,nesting in an urban environment: a review. peregrine in the AfrotropicalRegion. Pages297-306 Can. Field-Nat. 104:209-218. in T.J. Cade, J.H. Enderson, C.G. Thelander and C.M. CADE,TJ., C.M. WHITE ANDJ.R. HAUGH.1968. Peregrines White [EDS.], Peregrine Falcon populations: their and pesticidesin Alaska.Condor 70:170-178. management and recovery.The Peregrine Fund, Boi- se, ID U.S.A. COLLOPY,M.W. 1983. A comparisonof direct observation and collectionsof prey remains in determining the MORISITA,m. 1959. Measuringof interspecificassociation diet of Golden Eagles.J. wildl. Manage.47:360-368. and similarity between communities. Mem. Fac. Scz. EDWARDS,D. 1983. A broad-scale structural classification KyushuUniv. Sez.E (Biol.) 3:65-80. of vegetationfor practicalpurposes. Bothalia 14:705- NEWTON,I. ANDM. MARQuISS.1982. Food, predation and 712. breeding seasonin Sparrowhawks(Accipiter nisus). J Zool., Lond. 197:221-240. GOODMAN, S.M. AND C.V. HAYNES. 1992. The diet of the •, E.R. MEEKAND B. LITTLE. 1984. Breeding season lanner (Falcobiarmicus) in a hyper-arid region of the foods of Merlins Falco columbarius in Northumbria. eastern region of the Sahara.J. Arid Env. 22:93-98. Bird Study31:49-56. GREENE,H.W. AND F.M. JAKSI•. 1983. Food niche rela- NILSSON,I.N. 1984. Prey weight, food overlap, and repro- tionshipsamong sympatricpredators: effects of level ductive output of potentially competing Long-eared of prey identification. Oikos40:151-154. and TawnyOwls. OrnisScand. 15:176-182. HUNTER,R.E., J.A. CRAWFOPmAND R.E. AMBROSE.1988. ORO,D. ANDJ.L.TELIA. 1995.A comparisonof two meth- Prey selectionby Peregrine Falconsduring the nest- ods for studyingthe diet of the Peregrine Falcon.J. ling stage.J. Wildl. Manage.52:730-736. RaptorRes. 29:207-210. HUSTLER,K. 1983. Breeding biologyof the PeregrineFal- PARR,SJ. 1985. The breeding ecology and diet of the con in Zimbabwe. Ostrich 54:161-171. Hobby Falcosubbuteo in southern . Ibis127:60- JENKINS,A.R. 1991. Latitudinalprey productivityand po- 73. tential densityin the PeregrineFalcon. Gabar6:20-24. PORTER,R.D. ANDC.M. WHITE.1973. The PeregrineFal- . 1995. Morphometricsand flight performance of con in Utah: emphasizingecology and competition southern African Peregrine and Lanner Falcons.J. with the Prairie Falcon. BrighamYoung Univ. Sci.Bull. Avian Biol. 26:49-58. 18:1-74. --. 1998. Behaviouralecology of Peregrineand Lan- PRUETT-JoNES,S.G., C.M. WHITE ANDW.R. DEVINE.1981. ner Falcons in South Africa. Ph.D. dissertation, Univ. Breeding of the Peregrine Falcon in Victoria. Emu 80: Cape Town, Cape Town, South Africa. 253-269. KEMP,A.C. 1993. Breeding biology of Lanner Falcons RATCLIFFE,D.A. 1993. The Peregrine Falcon. T. & A.D. near Pretoria, South Africa. Ostrich 64:26-31. Poyser,London, U.K. KLEIN,R.G. ANDK. CRUZ-UPdBE.1984. The analysisof an- ROSENBERG,K.V. ANDRJ. COOPER.1990. Approachesto imal bones from archaeologicalsites. Univ. Chicago avian diet analysis.Stud. Avian Biol. 13:80-90. Press,Chicago, IL U.S.A. ROSENFIELD,R.N., J.W. SCHNEIDER,J.M. PAPPAND W.S. Ko•a'I 'MAI•,E. 1987. Dietary shifts,niche relationships SEEGAR.1995. Prey of Peregrine Falconsbreeding in and reproductive output of coexisting kestrelsand west Greenland. Condor 97:763-770. Long-earedOwls. Oecologia74:277-285. SHERRY,T.W. 1990. When are birds dietarily specialized? KREBS,CJ. 1989. Ecologicalmethodology. Harper Col- Distinguishing ecological from evolutionary ap- lins, New York, NY U.S.A. proaches.Stud. Avian Biol. 13:337-352. LEVINS,R. 1968. Evolution in changing environments: SIMMONS,R.E., D.M. AVERYAND G. AVERY.1991. Biases in some theoreticalexplorations. Princeton Univ. Press, diets determined from pellets and remains: correc- Princeton, NJ U.S.A. tion factors for a mammal and bird-eating raptor J MACLEAN, G.L. 1993. Roberts' birds of southern Africa. RaptorRes. 25:63-67. 202 JENKINSAND AVERY VOL. 33, No. 3

SMITH, E.P. •d•I) T.M. ZARET. 1982. Bias in estimating --AND D.G. ALLAN. 1984. The status and conserva- niche overlap.Ecology 63:1428-1253. tion of birds of prey in the Transvaal. TransvaalMus. SMITHERS,R.H.N. 1983. The mammals of the southern Monogz3:1-115. African subregion.Univ. Pretoria, Pretoria, South Af- YOSEF,R. 1991. Foraging habits, hunting and breeding rica. success of Lanner Falcons (Falco biarmicus) in Israel. TAV,BOTON, W.R. 1984. Behaviour of the African pere- J. RaptorRes. 25:77-81. grine during incubation. RaptorRes. 18:131-136. 1990. The Soutpansberg.Fauna & Flora47:1-6. Received5 December 1998; accepted 26 April 1999

Appendix 1. Invertebrateand vertebrate taxa identified from remains collectedat Peregrine and Lanner Falcon nest sitesin SouthAfrica. Remainswere collectedat 10 PeregrineFalcon sites on the Cape Peninsula,seven Peregrine Falcon siteson the Orange River,and sevenPeregrine Falconsites and nine Lanner Falcon sitesin the Soutpansberg. Data provided are the averagemass of each taxon and the number of individualsidentified, with the number of nest sitesat which each taxon was collectedin parentheses.Items marked with an asteriskwere not consideredas falcon prey in the diet analyses.

SOUTPANSBERG AVERAGE CAPE ORANGE PREYTAXON MASS(g) PENINSULA RIVER PEREGRINES LANNERS

Insects Unidentified orthopterans 3 i (1) •2 (2) 3 (2) Unidentified coleopterans 3 i (1) 2 (1) -- Amphibians Unidentified frogs 20 *2 (1) B•rds Little Sparrowhawk--male 80 1 (1) Accipiterminullus Peregrine Falcon--nestling 250 *2 (2) -- Lanner Falcon--nestling 250 -- '1 (1) Coqui Francolin 230 1 (1) -- Francolinuscoqui Crested Francolin 340 2 (1) 1 (1) E sephaena Grey-wingFrancolin 390 1 (1) E africanus Cape Francolin--pullus 150 1 (1) E capensis Natal Francolin--female 425 2 (2) E natalensis Natal Francolin--unsexed 520 1 (1) Unidentified adult francolin 500 2 (2) Unidentified pullus francolin 150 i (1) Domesticchicken--<3days old 60 41 (8) Gallusgallus Domesticchicken--+7 daysold Domestic chicken--_+ 14 daysold 150 23 (3) Harlequin Quail 250 i (1) C. delegorguei White-fronted Plover Charadriusmarginatus 49 2 (2) Kittlitz Plover C. pecuarius 43 2 (1) SEPTEMBER1999 PEREGRINEAND LANNER DIETS IN SOUTH AFRICA 203

Appendix 1. Continued.

SOUTPANSBERG AVERAGE CAPE ORANGE PREY TAXON MASS(g) PENINSULA RIVER PEREGRINES LANNERS Crowned Plover--adult 167 I (1) -- 16 (5) 14 (6) Vanellus coronatus Crowned Plover--immature 167 -- -- 2 (2) Greenshank 191 -- 1 (1) • (•) Tringa nebularia CurlewSandpiper 57 3 (1) -- Calidrisferruginea Little Stint 24 -- I (1) C. minuta Black-wingedStilt 175 -- -- 2 (1) 2 (2) Himantopushimantopus Temminck's Courser 74 -- -- 4 (3) Cursorius temminckii Common Tern 124 2 (1) 3 (1) -- Sterna hitundo White-wingedTern 57 -- I (1) -- -- Chlidoniasleucopterus Unidentified Tern 100 w 2 (2) -- -- NamaquaSandgrouse 185 -- 24 (6) -- -- Pteroclesnamaqua Burchell'sSandgrouse 235 -- -- 1 (1) P. burchelli Double-bandedSandgrouse 235 -- -- • (•) -- P. bicinctus Feralor domesticpigeon--adult 320 47 (7) 8 (4) 17 (3) -- Columba livia Feral or domesticpigeon--young 320 1 (1) ------RockPigeon--adult 347 28 (7) 35 (6) 20 (•) • (2) C. guinea RockPigeon---fledgling 347 8 (3) 2 (2) • (•) • (1) Rameron Pigeon 415 -- -- 2 (2) -- C. arquatrix Unidentified pigeon--adult 330 1 (1) ------Unidentified pigeon--young 330 2 (2) -- Red-eyedDove--adult 235 40 (7) 16 (5) 34 (6) 10 (4) Streptopeliasemitorquata Red-eyedDove--fledgling 235 1 (1) -- -- CapeTurtle Dove 153 61 (7) 5 (4) 25 (5) 8 (•) S. capicola LaughingDove--adult 102 140 (8) 20 (5) 79 (7) 21 (7) S. senegalensis LaughingDove---fledgling 102 4 (2) w 2 (2) Unidentified dove--fledgling 130 3 (3) -- -- NamaquaDove 40 3 (3) 7 (5) 15 (5) -- Oena capensis Green-spottedDove 65 -- -- I (1) I (1) Turtur chalcospilos Small unidentified columbid 50 -- -- Medium unidentified columbid 160 -- -- 1 (1) Cockatiel 90 I (1) w Nymphicushollandicus 204 JV.NraNS AND AVERY VOL. 33, No. 3

Appendix 1. Continued.

SOUTPANSBERG AVERAGE CAPE ORANGE PREY TAXON MASS(g) PENINSULA RIVER PEREGRINES LANNERS Budgerigar 28 4 (1) -- 1 (1) -- Melopsittacusundulatus Grey Loerie 269 -- -- 1 (1) -- Corythaixoidesconcolor Freckled Nightjar 60 -- 2 (1) -- -- Caprimulgustristigma Unidentified nightjar 60 -- -- I (1) -- Black Swift adult 45 13 (5) -- 12 (3) -- Apu,s barbatus Black Swift--fledgling 45 I (1) ------Bradfield's Swift 45 • 18 (6) -- -- A. bradfieldi Little Swift 26 12 (3) -- 2 (2) -- A. affinis Alpine Swift 77 6 (4) 6 (5) 11 (3) 2 (2) A. melba Unidentified swift 35 -- I (1) -- -- Speckled Mousebird 53 -- -- 14 (6) 1 (1) Colius striatus White-backed Mousebird--adult 41 1 (1) 2 (2) -- -- C. colius White-backedMousebird--fledgling 41 -- 1 (1) -- -- Red-faced Mousebird 56 4 (3) • 41 (7) -- Urocolius indicus Narina Trogon 65 -- -- 1 (1) -- Apalodermanarina Pied Kingfisher 82 -- 1 (1) -- -- Cerylerudis European Bee-eater 55 -- -- 1 (1) -- Meropsapiaster Hoopoe 57 -- I (1) 1 (1) -- Upupaepops Scimitar-billedWoodhoopoe 35 -- 1 (1) -- -- Phoeniculuscyanomelas Red-billed Hornbill 130 ------1 (1) Tockuserythrorhynchus Black-collaredBarbet 57 -- -- 2 (2) -- L ybiustorquatus Red-cappedLark 26 -- • I (1) -- Calandrella cinerea Chestnut-backedFinchlark 14 -- -- I (1) 1 (1) Eremopterixleucotis Grey-backedFinchlark 17 -- 19 ( 1) -- -- E. verticalis Unidentified lark 20 -- • 2 (1) -- European Swallow 18 I (1) ------Hirundo rustica Rock Martin 22 I (1) 4 (4) -- -- H. fuligula Brown-throated Martin 13 -- 11 (3) -- -- Riparia paludicola SEPTEM}•ER1999 PEP.•GRINEAND LANNER DIETS IN SOUTH AFRICA 205

Appendix 1. Continued.

SOUTPANSBERG AVERAGE CAPE ORANGE PREY TAXON MASS(g) PENINSULA RIVER PEREGRINES LANNERS Black-headed Oriole 69 -- -- 1 (1) Oriolus larvatus Unidentifiedsylviid 20 -- 2 (2) Fiscal Shrike 41 -- -- I (1) Lanius collaris Unidentified shrike 60 -- I (1) EuropeanStarling--adult 76 68 (9) -- Sturnusvulgaris EuropeanStarling--fledgling 76 5 (3) -- Pied Starling 107 1 (1) -- Spreobicolor Plum-coloredStarling 46 -- -- 1 (1) Cinnyricinclusleucogaster Cape GlossyStarling 84 -- -- 6 (3) Lamprotornisnitens GreaterBlue-eared Glossy Starling 76 -- -- 1 (1) L. chalybaeus Unidentified starling 80 -- -- 2 (1) Red-wingedStarling--adult 135 3 (3) -- I (1) Onychognathusmorio Red-wingedStarling--fledgling 135 -- -- 1 (1) Pale-wingedStarling 107 -- 4 (2) -- O. nabouroup Unidentified starling 80 -- -- I (1) 1 (1) White-browedSparrow-weaver 48 -- -- 3 (3) I (1) Plocepassermahali Sociable Weaver 27 -- 1 (1) -- Philetairus socius HouseSparrow 24 2 (1) -- -- Passer domesticus CapeSparrow 26 10 (3) 4 (3) 4 (3) 2 (2) P. melanurus CapeWeaver 45 7 (4) -- 3 (2) Ploceuscapensis Masked Weaver 27 1 (1) 1 (1) -- 1 (1) P. velatus Unidentified weaver 35 -- -- 1 (1) Red-billedQuelea 19 -- 2 (2) 6 (3) I (1) Queleaquelea Red Bishop 23 -- 2 (2) Euplectesorix Yellow-rumpedWidow 45 2 (1) -- E. capensis Unidentifiedploceid 40 1 (1) 1 (1) 2 (2) Common Waxbill 8 -- 3 (3) -- astrild Red-headed Finch 23 -- -- 3 (3) Amadina erythrocephala Yellow-eyedCanary 13 -- -- I (1) Serinus mozambicus 206 J•.NK•NSAND AVER-'/ VOL. 33, No. 3

Appendix 1. Continued.

SOUTPANSBERG AVENGE C•E O•GF. PREY TAXON M_•S (g) PwN•NSUI• Rear PEREGRINES LANNERS Forest Canary 16 l (1) S. scotops Bully Canary 26 1 (1) S. sulphuratus YellowCanary 17 1 (1) 1 (1) S. fiaviventris White-throated Canary 27 2 (1) S. albogularis Unidentified canary 15 (1) -- 1 (1) -- Unidentified fringillid 15 -- 1 (1) 1 (1) I (1) Unidentified small birds--adults 2O (3) 2 (2) 9 (3) 2 (2) Unidentified small birds--young 2O (•) -- _ _ Unidentified small-medium birds 6O (4) -- 11 (4) 8 (6) Unidentified medium birds 130 (3) 4 (4) 2 (2) 4 (2) Unidentified very large birds 6OO _ _ • (•) -- Mammals Unidentified shrew 10 -- • O) Unidentified fruit 130 6 (2) -- Flat-headed free-tailed bat 14 8 (•) Sauromyspetrophilus Egyptianfree-tailed bat 15 8 (•) Tadaridaaegyptiaca Hildebrandt's horseshoe bat 30 1 (1) Rhinolophushildebrantii Unidentified insectivorous bats 15 1 (1) 6 (4) s (s) Multimammate mouse 65 -- -- '1 (1) Mastomysnatalensis Namaqua rock mouse 45 ß • (•) Aethomysnamaquensis Small unidentified murid 25 ------• (D Medium unidentified murid 45 *2 (2) *3 (3) *2 (2) 1 (1) Large unidentified murid 120 -- -- *2 (2) -- Unidentified squirrel 160 -- -- '1 (1) -- Small unidentified bovid 5000 _ _ ,• (•) -- TOTAL 515 (10) 252 (7) 418 (7) 175 (10)

Top View