j. RaptorRes. 37(1):31-36 ¸ 2003 The Raptor Research Foundation, Inc.

DIFFERENTIAL SPRING MIGRATION OF ADULT AND JUXNILE LEVANT SPARROWHAWKS ( A CCIPITER BREVIPES) THROUGH EILAT, ISRAEL

REUVEN YOSEF 1 InternationalBirding and ResearchCentre in Eilat, Departmentof Life Sciences,Ben-Gurion University of theNegev, P.O. Box 774, Eilat 88000, Israel

LORENZO FORNASARI Departmentof EnvironmentalSciences, University of Milano Bicocca, Piazza della Sdenza 1, 1-20126,Italy

PIOTR TR•ANOWSKI Departmentof Avian Biologyand Ecology, Adam Mickiewi•z University, Fredry 10, PL-61-701Pozna•i, Poland

MARCJ. BECHARDAND GREGORYS. KALTENECKER IdahoBird Observatory,Department of Biology,Boise State University, 1910 UniversityDrive, Boise, 1D 83725 U.S.A

KEITH BILDSTEIN Hawk MountainSanctuary, 1700 Hawk MountainRoad, Kempton, PA 19529 U.S.A.

ABSTRACT.--Asmany as 50 000 LevantSparrowhawks ( brevipes) are countedduring migrationat the northern end of the Gulf of Aqaba each spring.We presentdata from 1819 migrantsthat were capturedand ringed at Eilat, Israel:459 from 1984--88,21 from 1989-95, and 1345 capturedfi:om 1996- 2000. Of these,396 (22%) were adult females,631 (35%) were adult males,359 (20%) werejuvenile fkmales,and 433 (24%) werejuvenile males. We comparemigration timing and bodysizes in .juvenile (i.e., first-timespring migrants) and adult migrants,and in malesand females.Wing chord length and body massin malesand femaleschanged significantly with date of arrival. Further, a significantcorre- lationwas found for both sexesbetween wing chord length and bodymass in spring.Within age classes, both wing chord and body massdeclined significantly with date of ringing. Body masswas also signifi- cantlyrelated to sizeobtained from PCA analyses(PC1), both in malesand females.We computedalso standardizedresiduals of body masson PC1. Date of passagewas also significantlycorrelated to the standardizedresiduals, both in malesand females.This suggested,testing for allometryvs. isometry, that in better than expected'condition' migrated earlier. Moreover,results from analysisof vari- ancerevealed that bodymass and agewere significantlyrelated to the date of passage.The mediandate of passagefor adultspreceded that of juvenilesby 2.5 days.We believejuveniles on their first spring passagemigrate slowerthan adultsand that they are more likely to be later and in poorer body con- dition.

KEYWORDS: Levant5)barrowhawk; Accipiter brevipes;ag• Eilat, Israel;sex,' spring migration.

MIGRACI6NDIFERENCIAL DE PRIMAVERAENTRE ADULTOS JUVENILES DEL AZOR DEL MED- ITE•EO ORIENTAL(ACCIPITER BREVIPES) A TRAVES DE EILAT,ISRAEL P•SUMEN.--Tantoscomo 50 000 azoresdel mediterr•tneooriental (Accipiterbrevipes) son contadosdur- ante su migraci0nen el limite del golfo de Aqabacada primavera. Presentamos datos de 1819 emigrantes que fueron capturadosy anilladosen Eilat, Israel:459 de 1984--88,21 de 1989-95, y 1345 capturados entre 1996-2000. De estos,396 (22%) fueron hembras adultas, 631 (35%) eran machos adultos, 359 (20%) hembrasjuveniles, y 433 (24%) machosjuveniles. Comparamos el tiempode migraci6ny los tamafiosdel cuerpo en emigrantesjuveniles (v.gr., azoresmigrantes que lo hacian por primera vez en primavera)y adultos,yen machosy hembras.La longitudde la cuerdadel ala y la masadel cuerpoen

E-mailaddress: ryosef@eilatcity. co.il

31 32 YOSEFET AL. VOL. 37, NO. 1

machosy hembras cambi6 significativamentecon la fecha de arribo. Por otro lado, se encontr6 una correlaci6n significativapara ambos sexosentre la 1ongitud de la cuerda del ala y la masa corporal en primavera.Dentro de las clasesde edad, tanto la longitud de la cuerda del ala como la masadel cuerpo declin6 significativamentecon la fecha del anillado. Las masascorporales adem•tsestuvieron relacion- adassignificativamente al tamafio obtenido mediante analisisPCA (PC1), tanto en machoscomo hem- bras. Calculamosadem•ts residuos estandarizadosde masa corporal en PC1. La fecha de paso estuvo tambi6n correlacionada significativamentecon los residuos estandarizados,tanto en machos como en hembras. Esto sugiere, haciendo pruebas de alometrfa vs. isometrfa, que las avesen mejor "condici6n" que la esperada,migraban m•tstemprano. Por otra parte, los resultadosdel analisisde varianzarevelan que la masa del cuerpo y la edad estuvieronrelacionadas significativamente con la fecha de paso. La fecha promedio para los adultos precedia a la de losjuveniles en 2.5 dias. Creemosque losjuveniles en su primera pasadade primavcra migran mas lcnto que los adultosy que son m•tspropensos a estar retrasadosyen condicionescorporales mas pobres. [Traducci0n de C6sar Marquez]

The Levant Sparrowhawk (Accipiter brevipes) mist nets, dho-gazas, box traps, and Bal-chatri traps breedsprincipally within the westernPalearctic Re- (Glark and Yosef 1997). All data were pooled because 63% of Levant Sparrowhawkswere caught in 7-m-high gion, in southeastEurope, locally through Turkey mist nets in the early mornings when the flocks started to northern , and is widespreadin southwest the day's migration. We assumedthat these birds repre- Russia and Kazakhastan (Snow and Perrins 1998). sent a sample of the general population becausethe ma- It winters in the east Sahel of sub-Saharan Africa, jority was not trapped with food as bait (Gorney et al 1999). All captured individuals were aged, sexed, mea- and sporadic reports are received also from the sured (including unfiattened wing chord), and weighed Ethiopian highlands and the southern Arabian Aging wasbased on plumage, molt, and iris color (Glark Peninsula (Snow and Perrins 1998, Shirihai et al. and Yosef 1998). In addition, birds divided into two cat- 2000). Principal migration routes lie entirely with- egories:juveniles--in second calendar year (SY) and in the Middle East (Frumkin et al. 1995, Shirihai adults--after second calendar year (ASY). Data presented in this paper were collected during 1984-88 (Gorney and et al. 2000) with especiallylarge concentrations Yom Toy 1994, Gorney et al. 1999), 1989-95 (Yosefand found at Eilat during spring (Shirihai and Christie Fornasari 2000), and 1996-2000 (Glark and Yosef 1997, 1992, Yosef 1995). Shirihai et al. 2000). We excluded from the analysesfive In Israel, the Levant Sparrowhawkis an abun- birds (two identified as female and three as male) be- cause according to the biometrics we consider them to dant nilgrant in both spring and autumn and have been sexed incorrectly. about 90% of the world population of the species Owing to differences in weather and other local con- passesthrough Israel within a short period of a ditions which influence the phenologyof migration (me- fortnight (Shirihai et al. 2000), and this is the only dian test, X2 = 294.5, df = 13, P < 0.0001), data were raptor known to nilgrate at night in the region standardized between years. By computing the median date of passagefor each year, data were collected. Then (Stark and Liechti 1993). Visible migration surveys the dates of all captured hawkswere transformed as vab conductedsince 1977 suggestthat Eilat is an im- ues before (minus) or after (plus) the median. portant stopoversite for the speciesin spring.Eilat All the basic statisticswere performed according to So- is at the northern edge of the Sahara and Sinai kal and Rohlf (1995). We computedcorrelation between birds' measurementsand timing of migration (in Julian deserts,and in spring many northbound migrants dates). However, becausein accipiterswing chord is an stop there to rest and feed (Safriel 1968, Yosef indicator of body size (e.g., Mueller and Meyer 1985, Wyl- 1996a). Levant Sparrowhawkswere trapped arid lie and Newton 1994), we established a condition index. ringed at and around Eilat during spring (mid- We used a total body-sizemeasurement, which was ob- April through early May) from 1984-2000 (Clark tained from the principal component analysiswith VAR- IMAX rotation of four (wing chord, culmen, hallux, and and Yosef 1997, Yosef and Fornasari 2000). In this tail length) log-transformed, external measurements paper, we presentmorphometric data collectedaf- (Piersma and Davidson 1991) computed separatelyfor ter capture to compare age- and sex-relateddiftbr- the sexes.All morphometric variableshad positiveand a encesin body sizeand migration tinting in this spe- similar magnitude of loading on the first component cies. (0.318-0.742, eigen value = 1.811, 45.3% of total vari- ance explained for females and 0.234-0.636; eigen value STUDY AREA AND METHODS = 1.577, 39.4% of total variance explained for males) We analyzedstructural size (PG1)/massrelationships sep- Levant Sparrowhawkswere captured and ringed at and arately within four age/sex classesto test for allometry around Eilat, Israel (29ø33'N, 34ø57'E), using bow-nets, or isometry. MARCH 2003 SPRING MIGRATION OF LEVANT SPARROWHAWI•S 33

(A) After-second-year females (A) After-second-year males ß ß y = -0.6244x + 208.23 ß y = -0.3255x + 167.85 ß ' P = 0.058 ß .. . P = 0.029 .....i,.: :,;:.: ß ß ß :.: '.h,fiL ,, ... -_._!,h,,1111,l!. ,!. ß ":;!ill!11:..' ß ß ß .;11-;i.- ß

ß ß

ß ß ß ß ß ß ß

i -25 -15 -5 5 15 25 -25 -15 -5 5 15 25 $tandarJzed ¾ea• Valne (Median = 0) Sta•dadzedYear Value (Media• = 0)

26O

240 (B) Second-year males 240

220 220

200 200180 180 •6o ]60 140 ]40 120

120 100

100 -25 -15 -5 S 15 25 -25 -15 -5 5 15 •andar•edYearValue(Median=O) S•andarizedYear Value (Median: O) Figure 2. Body massof (A) after:second-year(ASY) and Figure 1. Body massof (A) after-second-year(ASY) and (B) second-year(SY) female Levant Sparrowhawkin re- (B) second-year(SY) male Levant Sparrowhawkin rela- lation to date of capture and passageat Eilat. Regression tion to date of capture and passageat Eilat. Regression line is non-significantfor SY females. line is non-significantfor SY males.

more ASY individuals of both sex migrate earlier than SY birds. RESULTS Wing chord length in males (r•0s0= -0.207, P A total of 1819 migrant Levant Sparrowhawks < 0.0001) and females (r•44 = -0.127, P = 0.001) was captured: 459 in 1984-88, 21 in 1989-95, and changed significantlywith date of arrival, as did 1345 in 1996-2000. Of these, 396 (22%) were adult body massin males (r•022= -0.141, P < 0.0001; females, 631 (35%) were adult males, 359 (20%) Fig. 1) and females (rv40= -0.137, P < 0.0001' werejuvenile females,and 433 (24%) werejuvenile Fig. 2) changed significantly with date of arrival. males. Sex ratio amongjuvenile birds differed sig- Further, a significant correlation was found for nificantlyfrom 1:1 (Xa = 27.4,P < 0.0001),but was both sexesbetween wing chord length and body not significantlydifferent in adults(X 9 = 3.6, P = mass in spring (males y = 0.140 (+__0.011)x + 0.06). Hence, sex ratio between the two age cate- 0.367, t = 12.61, P < 0.0001 versusfemales y = goriesdiffered significantly(X 9 = 8.4, P = 0.004). 0.128 (+0.011) x + 0.380, t = 11.17, P < 0.0001). Each spring, transientswere trapped in Eilat be- In addition, we found a significant correlation be- tween early March and early May. The two sex and tween mass (log transformed before analyses)and two age groups differed significantlyin their me- wing chord length for juveniles (r1006= 0.625, P < dian time of migration (SY femalesmedian date = 0.000l) and adults (r7s3 = 0.665, P < 0.000l). How- 27 April, range = 10 March-15 May; ASY females ever, there was no statistical difference between the median date = 24 April, 18 March-16 May; SY correlation values for the two-age classes(P = males,27 April, 18 March-21 May; ASY males,25 0.152). April, 28 March-12 April; Kruskal-Wallistest, X9 = Body masswas also significantly correlated with 142.36, df = 3, P < 0.0001). A post-hocDunn's Q- size obtained from PCA analyses(PC1), both in test (P < 0.05) demonstrated that significantly males: •'597: 0.368, P < 0.0001 and females: r740= 34 YOSEFET AL. VOL. 37, NO. 1

Table 1. Analysis of Variance analysesof biDmetric val- "time selected" migrants, whereasjuveniles were ues of Levant Sparrowhawksrelated to migration passage "energy-selected"migrants. Becausemost juvenile at Eilat, Israel. birds do not breed, they would not need to reach their "breeding grounds" as early in spring as SOURCE OF SUM OF adults. We offer another explanation: that adults VARIATION SQUARES df F P precedejuveniles becausethey are better prepared Covariates 166.525 2 3.978 0.019 and more efficient at migration, therefore, are Body mass 134.968 1 6.448 0.011 more capable and faster migrants en route. Little Wing chord 4.626 1 0.221 0.638 is known about the Levant Sparrowhawkson their Main effects 139.636 2 51.107 <0.001 breeding grounds in Eurasia, and the specieshas Age 1755.369 1 83.856 <0.001 yet to be studied in detail on its wintering grounds Sex 0.147 1 0.007 0.933 in Africa (Shirihai et al. 2000). The fact that adults 2-way interactions arrive in Eilat only a few daysearlier than juveniles Age*Sex 30.493 1 1.457 0.228 and because adults are heavier than juveniles, in our view,suggests that adults are more capablemi- Explained 3301.656 5 31.545 <0.0001 Residual 37 218.904 1778 20.933 grants, rather than that the age classesare using Total 40 520.560 1783 22.726 different migration strategies.However, it is also possiblethat age classeswinter in separateregions, at different distancesfrom Eilat, resulting in dis- 0.330, P < 0.0001. We computed also standardized crepancy in phenology between the two classes. residualsof body masson PC1. Date of passagewas However, the latter cannot be verified at present significantlycorrelated to the standardizedresidu- owing to the lack of data and observationsfor the als, both in males: rs98= -0.154, P < 0.0001 and speciesfrom the wintering grounds. Another pos- females: r349= -0.110, P = 0.040. This suggested, sible explanation is that the different age classes testing for allometry versusisometry, that birds in may have different migration strategiesbecause better than expected 'condition' migrated earlier. adults have longer wings and tails in the spring Moreover, results from Analysis of Variance re- (Yosefand Fornasari 2000). Proportionallylonger vealed that only body mass and age were signifi- wing and tail length allow for a greater proportion cantly related to the date of passage(Table 1). of time spent in soaringflight, which is in contrast to juveniles who have shorter wing and tail which DISCUSSION requires comparatively more flapping flight that In many raptors, adults migrate earlier in spring requires greater fat reservesand better body con- than do juveniles (Newton 1979, Gorney and Yom- dition. Tov 1994, Yosef et al. 2002). With an overall 10-yr Soaring migrants such as Levant Sparrowhawks median trapping date of 25 April for adults versus (Spaar 1997, Spaar et al. 1998) typicallytravel in 27 April juveniles,our results,which extend an ear- large flocks, presumably so that individuals can lier 5-yr study of Gorney et al. (1999), confirm that more quicklylocate thermals needed to assisttheir Levant Sparrowhawksin Israel, too, exhibit age-re- long-distancemovements (Kerlinger 1989). Obser- lated differences in the timing of migration. Al- vationsof Broad-wingedHawks (Buteoplatypterus) though age differencesin raptor migration are not in North America suggestthat juveniles are more completely understood, previous work suggests likely to be wind drifted and blown off coursethan that such differences occur becausebreeding pres- are adults (Hagar 1988, Hoffman and Darrow sures on adults select for earlier arrival on the 1992). It is thought that coastalraptor migrations breedinggrounds (Newton 1979, Gorney and Yom- consistprimarily of juveniles of all speciesbecause Tov 1994), juveniles require more time either to of this fact (Kerlinger 1989). We propose that ju- initiate or complete their journeys, or both (Gor- venile Levant Sparrowhawkspass through Eilat lat- ney and Yom-Tov 1994, Gorney et al. 1999) or al- er than adults because they are less efficient mi- ternatively,juveniles may over-winterfarther from grants as shown by their lower body weights and their breeding grounds than do adults (Kr61 1983). lower 'condition' (Clark and Yosef1997, Gorneyet Gorney and Yom-Tov (1994) argued that earlier al. 1999). passageof adult steppe Common Buzzards (Buteo Although Gorney et al. (1999) found no signif- buteovulpinus) at Eilat suggestedthat adults were icant associations between condition indices of MARCH 2003 SPRING MIGRATION OF LEVANT SPARROWHAWKS 35

adults and juveniles, we believe that large numbers tre through the years, W.S. Clark, Dayton Baker of the of Levant Sparrowhawks,particularly iramatures, National Aviary in Pittsburgh,World Wildlife Fund-Inter- national, Sir and Lady Kaye, Mr. and Mrs. Speers, Dr reach Eilat in poor body condition. Many raptors Gerold Dobler and SwarovskiOptics, Earthwatch InsU- ringed at Eilat have tar and oil residueson their tute, and Raft Saar and Montal of Kibbutz Eilot. We thank feathersand feet (Clark and Gorney 1987). Per- Mick Marquiss and three anonymousreferees for com- hapsthey mistakenlyland in oil fields of the Sahara ments on an earlier version of the manuscript. and Sinai deserts in search of fresh water during LITERATURE CITED migration flights (Clark and Gorney 1987). Data BILDSTEIN, K.L., W.S. CL•m•:, D.L. EVANS, F. iVIA•sHALL, L collected from 1996-2000 indicate that 81% (N = SOUCY,aND E. HENCKEL.1983. Sex and age differenc- 43) of oil-contaminatedLevant Sparrowhawkswere es in fall migration of Northern Harriers. J. Field Or- juveniles. We assume that juveniles are more nithol. 55:143-150. stressedduring migration, and therefore, more Ck4vm, W.S. AND E. GORNE¾.1987. Oil contamination of likely to seek drinking water. raptors migrating along the Red Sea. Environ.Pollul Gorney et al. (1999) found no significantasso- 46:307-313. ciationswithin age and sex groupsbetween date of --AND R. YOSEF.1997. Migrant Levant Sparrow- migration and physicalcondition. In contrast, our hawks (Accipiterbrevipes) at Eliat, Israel: measurements and timing.J. RaptorRes. 31:317-320. studyshows that birds with longer wing chordsand --AND ---. 1998. In-hand identification guide to greater body masspassed through the area earlier palearctic raptors. International BirdwatchingCentre than smaller individuals of the same sex and age in Eilat. Tech. Publ. 7, Eilat, Israel. class.In this study, we find that wing chord was FRUMKIN, R., B. PINSHOW, AND S. KLEINHAUS. 1995. A re- significantlycorrelated with body mass,suggesting view of migration over Israel.J. Ornithol.136:127- that while on migration, larger birds are heavier. 147. Gorney and Yom-Tov (1994) suggestthat the large C.ORNEY, m. ANDY. YOM-TOV.1994. Fat, hydration condi- proportion of iramaturesringed at Eilat may have tion, and moult of steppebuzzards Buteo buteo vulpinus resulted from age differences in migration routes on spring migration. Ibis 136:185-192. , W.S. CLARK, AND Y. YOM-TOV. 1999. A test of the as has been reported for other raptor species(Bild- condition-bias hypothesisyields different results for stein et al. 1983, Yosef 1996b, Yosef and Alon two speciesof sparrowhawks(Accipiter). Wilson Bull 1997). No evidence suggeststhat adult Levant 111:181-187. Sparrowhawksfollow a different migration route HaG?m,J.A. 1988. Broad-wingedHawk: migration. Pages than juveniles (Shirihai and Christie 1992, Shirihai 1-25 in R.S. Palmer [ED.], Handbook of North Amer- 1996, Shirihai et al. 2000). Rather, geography of ican birds. Vol. 5. Yale Univ. Press, New Haven, CT the region (Shirihai et al. 2000, Zalles and Bild- U.S.A. stein 2000) suggeststhat northern end of the Gulf HOFFMAN,W. ANDH. DAm•OW.1992. Migration of diurnal of Eilat servesas a concentration point for many raptors from the Florida Keys into the West Indies. HMANA Hawk MigrationStud. 17:7-14. western Palearcticmigrating raptor speciesduring KERLINGER,P. 1989. Flight strategiesof migrating hawks spring,regardless of age or sex (Spaaret al. 1998). Univ. of Chicago Press,Chicago, IL U.S.A. We think that the limited sample size included by limoL,W. 1983. Bird ringing resultsin Poland. Migration Gorney and Yom-Tov (1994) and Gorney et al. of the buzzards Buteo buteo buteo. Acta Ornithol. 19:137- (1999) in their studymay have led them to conclu- 151. sionsnot supportedby our data set,which includes MUELLER, H.C. AND K. MEYER. 1985. The evolution of re- more than double in the number of birds involved versed sexual dimorphism in size--a comparative in the previousanalyses. analysisof the Falconiformesof the western palearc- In conclusion, the fact that within sex and age tic. Curt. Ornithol. 2:65-101. classes, heavier and better 'condition' individuals NEWTON,I. 1979. Populationecology of raptors.Buteo Books, Vetmillion, SD U.S.A. are trapped early in the seasonsuggests that in PIERSMA,T. AND N.C. DAVIDSON. 1991. Confusion of mass both juveniles and adults, the more efficient mi- and size. Auk 108:441-444. grantspass earlier than lessefficient migrants,and SAFP,IEL, U. 1968. at Eilat, Israel. Ibis'110: that adults are more efficient than juveniles. 283-320. SHImHAt, H. 1996. The birds of Israel. Academic Press, ACKNOWLEDGMENTS London, U.K. We thank the many organizationsand individualswho --AND D. CHInSTIE.1992. Raptor migration at Eilat. have helped the International Birding and ResearchCen- B•: Birds 85:141-186. 36 YOSEFET AL. VOL. 37, NO. 1

, R. YOSEF, D. At,ON, G. KIRWAN, AND R. SPAAR. --. 1996a. Raptors tbeding on migration at Eilat, Is- 2000. Raptor migration in Israel and the Middle rael: opportunisticbehavior or migration strategy?J East--a sunimary of 30 years of field research. Inter- RaptorRes. 30:242-245. national Birding & Research Centre in Eilat, Eilat, Is- --. 1996b. Sex and age classesof migratingraptors rael. during the spring of 1994 at Eilat, Israel.J. RaptorRes SNOW, D.W. •,ND C.M. PERRINS.[EDS.] 1998. The birds of 30:160-164. the western palearctic. Vol. 1. Non-passerines.Oxford --. AND D. ALON. 1997. Do immature palearctic Univ. Press, Oxford, U.K. Egyptian Vultures Neoph•vnpercnopterus remain in Af- SouAt,, R.R. AND F. j. ROHLF. 1995. Biometry, 3rd Ed. rica during the northern summer? Vb•elwelt118:285- Freeman Press, New York, NY U.S.A. 289. SP^^R,R. 1997. Flight strategiesof migratingraptors: a --. AND L. FORNASAm. 2000. Biometric differences comparative study of interspecific variation in flight between age and sex classesof the Levant Sparrow- characteristics. Ibis 139:523-535. hawk Accipiterbrevipes on migration at Eilat, Israel. Is- , n. STARK,AND E LIEGHTI.1998. Migratory flight raelJ. Zool.46:207-214. strategiesof Levant Sparrowhawks:time or energy --, P. TRYJANOWSKI,AND K.L. gILDSTEIN.2002. Spring minimization? Anim. Behav. 56:1185-1197. migration of adult and immature Buzzards(Buteo bu- teo)through Eilat, Israel:timing and bodysize. J. Rap- ST)din,H. 2kNDE Lmcn'n. 1993. Do Levant Sparrowhawks tor Res. 36:115-120. Accipiterbrevipes also migrate at night? Ibis 135:233- ZALLES,jd. ^NI) K.L. BILDSTE1N.(EDS.) 2000. Raptor 236. watch: a global directory of raptor migration sites. WY•Lm, I. ^ND I. N•WTON. 1994. Latitudinal variation in BirdLife Conservation Series No. 9. BirdLife Inter- the body-sizeof SparrowhawksAccipiter nisus within national, Cambridge, U.K. and Hawk Mountain Sanc- Britain. Ibis 136: 434-440. tuary. Kempton, PA U.S.A. YO$•F, R. 1995. Spring 1994 raptor migration at Eilat, Israel.J. Raptw'Res. 29:127-134. Received19 December 2001; accepted9 September2002