J. Field Ornithol.

, 70(2):230-235

HABITAT COMPOSITION OF MAURITIUS KESTREL

HOME RANGES

Jo ELLENCARTER
1ANDMICHAELHUNTJONES
2,3
Departmentof Biology
BoiseState University
Boise, Idaho 83725 USA

Abstract.--Lossof its original native forest habitat has been identified as one causeof the
population decline of the Mauritius Kestrel (Falcopunctatus).During the 1988-1989 breed-
ing seasonwe examined the size and habitat compositionof the home rangesof two pairs
of nesting Mauritius Kestrels,one located in a traditional location and the other in a new
nest cavityat the edge of agriculturalfields and other non-nativevegetation.Kestrelhome
ranges were determined by radio tracking. Habitat typeswere identified using overstory
canopiesonly.Nativevegetationcomposed12-41% of kestrelhome range area,with exotic
woodland,agriculturalland, and scrubcomposing15-40%. Kestrelswere observedtaking
prey from non-nativehabitats,and both pairsfledgedyoung.TheseMauritiusKestrelsappear
to have adjustedsuccessfully
to non-nativehabitats.

COMPOSICION
DELHABITATQUEOCUPAFALCOPUNCTATUS
Sinopsis.--Una de las causasde la reducci6n poblacional del Falc6n de Mauricio (Falco
punctatus)ha sido la pfirdida de habitat pot la destrucci6nde susbosquesnativos.Durante
la fipocareproductivade 1988-89 examinamosel tamafioy composici6ndel habitatde fireas
utilizadas(home ranges)de dos paresde falcones.Una de estaslocalidadesse encontraba
en un /trea tradicionaly la segundaen el borde de un campo agrfcolacon vegetaci6nno-
nativa.Las fireasutilizadaspor las avesfueron determinadasutilizando radio localizaci6n.Se
identificaronlos habitatsutilizandotan solola vegetaci6ndel docel del bosque.De un 12-
41% del /trea de utilizaci6nde las avesestuvocompuestade vegetaci6nnativay de un 15-
40% de firbolesex6ticos,tierra agricolay vegetaci6narbustiva.Ambasparejascriaron en los
dos tipos de habitatsy se observarona los falconescazandoen el /trea con vegataci6nno-
nativa. E1Falc6n de Mauricio parece haberseajustadoadecuadamentea habitatsno-nativos
en su lugar de origen.

The Mauritius Kestrel (Falco punctatus), one of the rarest falcons in

the word, is endemic to the island nation of Mauritius in the southwestern
Indian Ocean (Temple 1977, Cade 1982,Jonesand Owadally1988). Be-
causeof its preferencefor arboreal geckosand method of still-hunting,
which requiresan open canopy,the Mauritius Kestrelwasthought to be
dependentupon nativeforest(Temple1974, 1977, 1987;Jonesand Owa-
daily 1988). Most of the kestrel'soriginal native forest habitat has been
lost to cultivation or commercialdevelopment (Cade and Jones 1993).
The remainingnativeforestis severelyalteredby invasionof exoticplant
species,which changethe composition,structure,and densityof the for-
est canopyand the understoryvegetation,in turn affectingthe availability
of potentialprey species.By 1974, MauritiusKestrelshad declinedto a
1 Currentaddress:
Departmentof MedicalMicrobiology,
The Ohio StateUniversity,Columbus,
Ohio 43210 USA.
9 Currentaddress:
Departmentof Evolution,Ecology,and OrganismalBiology,The Ohio State
University,Columbus,Ohio 43210 USA.
•Author to whomreprintrequests
shouldbesent.

230
Vol.70,No.2 MauritiusKestrel
Habitats [231

low of four individualsin the wild (Temple1977).Intensiveconservation

efforts, including releaseof captivelybred birds and thosereared from
harvestedeggsinto areasof native vegetationand other habitats (Jones
et al. 1991), have led to recoveryof this species.
Here we describethe size and habitat compositionof home rangesof
two pairsof nestingMauritiusKestrelsoccupyingheterogeneoushabitats
during the late 1980s.One pair occupied a traditional nest cavityin the
region of the greatestremaining native forest,and the other occupieda
new nest cavityat the edge of degraded forest and agriculturalfields.
METHODS

Studyarea.--Mauritiusis an oceanic,subtropicalisland (1860 km2) in

the southwesternIndian Ocean, 800 km east of Madagascar.Since colo-
nizationbeganin the 17th century,mostof the islandhasbeen cleared
for cultivation and development.Inhabited by more than one million
people,most of the islandremainsrural. Forestsare limited primarilyto
steepridgesand gorgesand someriparian areas.At the onsetof thisstudy,
kestrelswere found only in the vicinity of the Black River Gorges (55
km2) on the southwestcorner of the island, and only 10 nesting pairs
were known.This area wasthe only region of the islandnot systematically
treatedwith pesticides(Cheke1987).Nativetreesstilloccupymostof the
gorges,but manyareasare dominatedby one or more exoticspecies,and
other areasare a mosaicof nativeand non-nativespecies.The gorgesare
surroundedon all sidesby cultivatedlands,providingavenuesfor invasion
of non-nativeplants.
Breedingpairs.--During the 1988-1989 breeding season,two nesting
pairsof Mauritius Kestrelswere studiedduring three nestingstages(Table
1). One pair used a traditional nest site on Mr. Zaco in the Black River
Gorges.Both birds were wild (i.e., not captivelyreared) but had been
banded severalyearspreviouslyand had successfully fledged young the
two seasonsprior to the startof the presentstudy.The secondpair nested
on the Trois Mamellesmassifat the fringe of the gorges.The male was
captivelybred and hackedapproximately3 km away,whereasthe female
wasa wild bird. The male wasobservedcourtingthe femalein September
1988. The pair searchedfor cliff cavitiesfacing forestedand unforested
aspects,and eventuallychosea cavityfacing agriculturalland. This site
was unique becauseit was the first nest cavitylocated outside forested
areas.

Both pairswere providedwith approximately30 g of white mice directly

belowtheir respectivenestsitesdaily.Supplementalfeedingwasbelieved
necessaryby the local kestrel conservationorganization,the Mauritius
Wildlife Appeal Fund, to facilitatereproductivesuccessand conservation
goals.The behaviorof both male kestrelswasapparentlymodifiedby this
supplement becausethey remained near the feeding site for approxi-
mately 1-2 h prior to the feeding, and remained in the vicinity as long
as the workersdeliveringthe food remained at the site.
Habitat classification.--Nativeforest and many exotic tree and shrub
232] J. E. Carter
andM. H. Jones j. Field
Ornithol.
Spring 1999

TABLE1. Mauritius Kestrelhome range sizeand habitat compositionfor the Mt. Zaco and
Trois Mamelles nest sites in 1988-1989. The bottom portion of the table showsthe
percentageof habitat type within each home range, followedby the percentageof sta-
tisticallyindependent radio trackingobservationsmade in each habitat.

Site

Mt. Mt. Trois Trois Trois

Zaco Zaco Mamelles Mamelles Mamelles

Sex male male male male female

NestingStage incubation post-fledging nestling post-fiedging post-fiedging
Home Range (kin2) 0.80 0.80 1.54 1.61 0.79
Habitat Type (%)/Observations(%)
Native forest 12/12 13/6 29/12 34/32 41/11
Mixed exotic wood-
land 40/54 40/37 1/0 0/0 0/0
Agricultural 0/0 0/0 39/61 37/47 29/38
Monotypicexotica 9/0 9/2 2/0 2/0 2/2
Open ground 7/4 9/25 3/0 1/1 1/9
Rock 3/4 2/0 10/15 6/5 10/9
Scrub 25/15 24/27 15/7 18/6 16/16
River Reserves 0/0 0/0 3/5 4/9 3/16
Acacia 4/12 4/2 0/0 0/0 0/0

-•For Mt. Zaco, this categorycombinesthe tecoma (Tabetruiapallida) and terrainalia (Ter-
minaha sp.) habitat types;for Trois Marnelies,this typewasprimarily Liane de Cerf (Hiprage
benghalensis)
.

specieson Mauritiushave distinctivecanopiesthat are readilydistinguish-

able from one another when viewedfrom vantagepoints availablein the
Black River Gorgesarea. Native forest is heterogeneousin terms of species
composition and canopy structure, and native trees are taller than most
exoticvegetation.In contrast,many exoticsform monotypicstands,either
through deliberate planting or invasion of native vegetation. Further-
more, many exotic specieshave canopystructuresthat are too densefor
kestrelflight, do not supportkestrelprey,or both (Jones1987). Kestrels
hunt primarily in the inner portion of a tree's canopyby still-hunting
(i.e., perching to watch prey and then pursuing by hopping or flying
amongthe branches; Jones1987). The structureof the overstorycanopy
is therefore important to kestrel ecology.By identifying habitat types
basedsolelyon overstorycanopies,we producedhabitat categoriesmean-
ingful for kestrel research and conservation,and avoided the need to
identify vegetation communities using more rigorous floristic methods
requiring extensivesamplingand ground truthing.
We used previousdescriptionsof native forest by Vaughan and Wiehe
(1937,1941) to identifynativehabitattypesavailableto the breedingpairs
included in this study (Table 1). Becausenative forest is now extensively
invadedby exotic plants (Lorrence and Sussman1988), we establisheda
thresholdof 25% canopycoverby native tree speciesfor a habitat to be
considerednative forest.Forestedareaswith lessthan 25% canopycover
were defined as mixed exotic woodland. We also usedVaughan and Wie-
Vol.70,No.2 MauritiusKestrel
Habitats [233

he's (1937, 1941) descriptionof riparian vegetation,calledriver reserves.

Outside of the Black River Gorges,river reservesare narrow bands of
exotic trees and other vegetation,with native trees present in some lo-
cations.Agricultural lands usuallyabut these corridorsand are typically
planted in sugar cane. We identified an additional habitat type called
monotypic exotic that included a single exotic specieswithin any one
mapped habitat type. Habitat patcheswere mapped if they were a mini-
mum of 25 m in one dimension.
Radio tracking.--Kestrelswere captured using bal-chatri traps. Radio
transmitters(Terry Roundy, Salt Lake City, Utah, U.S.A.) weighing ap-
proximately5 g (3.1-3.5% of kestrelbodymass)were clippedto the two
central retrices,as describedby Kenward(1978, 1980). Batterieswere
field-replaceable,facilitating continuoustracking and minimizing han-
dling. Three-channelreceivers(RB4 "Falconer,"CustomElectronics,Ur-
bana, Illinois, U.S.A.) with three-element hand-held Yagi antennas and
earphoneswere used to locate the birds. Signaldirection wasdetermined
using the null-averagemethod.
Kestrelswere observedduring three nestingstages,including incuba-
tion, nestling,and post-fledgingover a 3-mo period from October 1988-
January 1989. Locationswere triangulatedevery15 min during each ob-
servationperiod using at least two receivers.Samplingwas conducted
0600-1800 h, at 3-10 h intervals.Each hour of the day from 0600-1800
h was sampled an equivalentnumber of times. We tested for indepen-
dence of data points following Schoener (1981) and Swihartand Slade
(1985). Upper and lower critical t2/r 9 valueswere calculatedfor each
bird. If data were autocorrelated for the 15-min observation intervals, we
used longer time intervals (i.e., multiples of 15) until data achievedin-
dependence.
Homerangesizeand habitatusedetermination.--Homerange sizeswere
determinedusingthe minimum convexpolygonmethod (Macdonaldet
al. 1980, Schoener 1981). Total size of home range and habitat type areas
within home-rangepolygonswere calculatedusinga digitizer.From these
we calculatedthe percentagesof habitat typesavailableto each individual
kestrel during each nesting stage.
RESULTS AND DISCUSSION

Mauritius Kestrelsmaintained home rangesthat encompassedmosaics

of nativeand non-nativehabitat types(Table 1). For the Trois Mamelles
pair, nativeforestand agriculturalland each accountedfor roughly33%
of the birds' home ranges,with scrub and rock making up most of the
rest. Some 47-61% of the observations of the male and 38% of the female
at Trois Mamelles were made in sugar cane fields. Both birds were fre-
quentlyobservedtaking Calontes versicolor,
a 50-gexoticterrestrialagamid
lizard that was common only in sugarcane fields. This was the first re-
corded instance of agricultural land being used for hunting within a
Mauritius Kestrel home range. Kestrelsfrom Trois Mamelles were ob-
served hunting on the wing, hovering, and hopping or sitting on the
234] J. E. Carter
andM. H. Jones J.Field
Ornithol.
Spring 1999

ground, unusual behavior for this species.The home range of the Trois
Mamellesmale was•1.5 km2,roughlytwicethe sizeof the Mt. Zacomale.
However,this male wasa yearling, and this washis first nestingattempt.
The Trois Mamelles pair successfully fledged young the year this study
was conducted.
Most of the Mt. Zaco male home range was made up of mixed exotic
woodland(40%) and scrub(25%), whileonly 12% wasnativeforest(Ta-
ble 1). The relativeproportionsof thesehabitat typesvariedlittle between
nesting stages,but the proportion of times the Mt. Zaco male was ob-
servedin each habitat type changedsubstantially for most habitat types
from the incubationto post-fledgingstage(Table 1). One of the predom-
inant non-nativetreesin severalhabitatsin the Mt. Zaco home range was
Terminaliasp.This tree typicallyhaslarge diameterbranches,a relatively
open canopystructure,and peelingbark, characteristics that are believed
favorablefor Phelsuma geckos,the kestrels'primarypreyspecies(Gardner
1984,Jones 1985). Carter (1991) found that severalnon-nativetree spe-
cies in Mauritius Kestrel home ranges,including Terminalia,supported
Phelsumapopulationsgreater than thosein native forest, but that other
exotic speciessupportedfew geckos.The Mt. Zaco pair also successfully
fledged young the year of this study.
Prior to this study,it was believed that even though home ranges in-
cluded non-nativehabitat types,kestrelswould neverthelessrely heavily
on native forest for prey. Although the number of kestrelsexamined in
this studywasinsufficientto draw conclusionsregardinghabitat use,our
data show that Mauritius Kestrels selected home ranges that included
relativelysmall proportionsof native forest, and that kestrelsdid not ap-
pear to spend a disproportionateamount of time in native forest relative
to other habitats.Supplementalfeeding undoubtedlyhad an effect on
kestrel behavior and the way in which the data can be interpreted, but
the importanceof supplementalfeeding to MauritiusKestrelconservation
at the time of this study clearly outweighedshort-term researchgoals.
Since this studywas completed,kestrelshave been observedto nest in
non-native habitatson many parts of the island not known to be previ-
ouslyoccupied.By 1991 there were asmanyas 145 non-captiveindividuals
on the island (Cade andJones1993,Joneset al. 1991), and by 1996 some
100 breeding pairs in the wild.
ACKNOWLEDGMENTS

This work was conductedin collaborationwith and supportedby the Mauritius Wildlife
Appeal Fund, ForestryQuarters,Black River,Mauritius.Principalfinancialsupportcame
from the Raptor ResearchCenter at BoiseStateUniversityand the World Center for Birds
of Preyin Boise,Idaho, with additionallogisticalsupportprovidedbyAir Mauritius.We thank
C. Jones, R. Lewis, L. Duverge, L. Jenkins, R.-M. Cretien, K. Swinnerton,and K. Evansfor
kestrel researchassistance in Mauritius. Botanicalassistance wasprovidedby D. Lorrence,
W. Strahm, and the staff at the Mauritius SugarResearchInstitute herbarium. Thanks to J.
Munger, M. Bechard,and T. Cade for their many forms of guidance,and to E. Atkinson,K.
Bildstein,and C. R. Chandler for additionalhelpful comments.
Vol.70,No.2 MauritiusKestrel
Habitats [235

LITERATURE CITED

CADE,T.J. 1982. The falconsof the world. Cornell UniversityPress,Ithaca,New York. 188 pp.
--, AND C. G. JONES.1993. Progressin restorationof the Mauritius Kestrel. Conserv.
Biol. 7:169-175.
C^RTER, J. E. 1991. Home range, habitatutilization,and prey deliverypatternsof the Maul:
itius Kestrel (Falcopunctatus).M.Sc. thesis.BoiseState University,Boise,Idaho, U.S.A.
132 pp.
CHar,E, A. S. 1987.Aa•ecologicalhistoryof the Mascarenelslands,with particularreference
to extinction and introductionsof land vertebrates.Pp. 5-89, in A. W. Diamond, ed.
Studiesof Mascarenelsland birds. Cambridge UniversityPress,Cambridge,United King-
dom. 458 pp.
G•X•r)N•R,A. S. 1984. The evolutionaryecologyand population systematics of day geckos
(Phelsuma)in the Seychelles.Ph.D. dissertation.Universityof Aberdeen, Scotland,Unit-
ed Kingdom,387 pp.
JoNEs,C. G. 1985. The biology of the criticallyendangeredbirds of Mauritius. M.Sc. thesis.
UniversityCollegeof Swansea,United Kingdom.384 pp.
--. 1987.The larger land-birdsof Mauritius.Pp. 208-300, inA. W. Diamond,ed. Studies
of Mascarenelsland birds. CambridgeUniversityPress,Cambridge,United Kingdom.
458 pp.
--, AND h. W. OW.M)ALLY. 1988. The life histories and conservation of the Mauritius
KestrelFalcopunctatus,pink pigeon Columbamayeri,and echo parakeetPsittaculaeques.
Proceedingsof the Royal Societyof Arts and Science5:81-130.
--, W. H•c•:, R. E. L•wm, Y. ML•N(mOO,^ND T. J. CADE.1991. A summary of the con-
servationmanagementof the MauritiusKestrelFalcopunctatus1973-1991. Dodo 27:
81-99.
KENWOOD,R. E. 1978. Radio transmitters tail-mounted on hawks. Ornis Scand. 9:220-223.
ß 1980. Radio monitoring birds of prey. Pp. 97-104, in C. J. Amlaner and D. W.
McDonald, eds.A handbook on biotelemetry and radio tracking. PergamonPress,Ox-
ford, United Kingdom. 804 pp.
LORENCE, D. H., ANDR. W. SUSSMAN. 1988ßDiversity,density,and invasionin a Mauritian
wet forest.Monogr.Syst.Bot. 25:187-204.
MACDONALD,D. W., F. G. BALL,AND$. G. HOUGH. 1980. The evaluationof home range size
and configurationusing radio-trackingdata. Pp. 405-424, in C. J. Amlaner and D. W.
MacDonald, eds.A handbook on biotelemetry and radiotracking.PergamonPress,Ox-
ford, United Kingdom. 804 pp.
SCHOENER, T. W. 1981.gal empiricallybasedestimateof home range.Theor. Popul.Biol. 20:
281-325.
SWIHART,R. K., .•ND N. A. SLADE.1985. Testing for independence of observationsin animal
movements.Ecology66:1176-1184.
TEMPLE,S. A. 1974. Wildlife in Mauritiustoday.Oryx 13:584-590.
ß1977. Statusand conservationof endemickestrelson Indian Ocean islands.Pp. 74-
81, in R. D. ChancelIra; ed. Report of Proceedingsof World Conference on Birds of
Prey,Vienna, 1-3 Oct. 1975. ICBP,London, United Kingdom.442 pp.
ß 1987. Foragingecologyof the endangeredMauritiusKestrel (Falcopunctatus).Bio-
tropica 19:1-6.
VAUGHAN, R. E., ANDP.O. •A[IEHE.1937. Studieson the vegetationof Mauritius.I. A. prelim-
inary surveyof the plant communities.J. Ecol. 25:289-343.
, AND--'. 1941. Studieson the vegetationof Mauritius. lII. The structure and
developmentof the upland climaxforest.J. Ecol. 29:127-160.
Received6 Jan. 1998; accepted23 Jul. 1998.