Corella,2OO7, 31 (1\: 1 0-12

SEXUALDIMORPHISM IN THE PROVIDENCEPETREL Fterodroma solandri USING DNA ANALYSIS

ADAM J. BESTER''', N. I. KLOMP' and D. PRIDDEL'

'CharlesSturt University,PO Box 789, Albury, NSW 2640, Australia. rDepartmentof Environment and Conservation (NSW), PO Box 1967, Hurstville, NSW 2220, Australia 'Current address:Arid Recovery,PO Box 150 Roxby Downs, SA, 5725, Australia.

Received:25 October 2005

The ProvidencePetrel Pterodromasolandri was studied on Lord Howe lsland during the 2000 breedingseason. Molecular sexing was conducted on feather samples taken from 18 pairs of breeding . Morphological measurementstaken from these sexed birds indicatethat maleswere significantlyheavier than females(mean mass of males507.2 + 41.5 9; females476.6 t 44.7 9) and had a significantlylonger culmen (males 36.5 t 1.1 mm; females35.5 r 1.3 mm). This study confirms earlierevidence, based on small sample sizes,that ProvtdencePetrels are indeed sexuallydimorphic. Sexual dimorphism has only been positivelyidentified in two other Pterodroma.The abilityto sex individualProvidence Petrels using this DNA analysis provides a means to investigatesex-specific differences in foragingand breedingbehaviour.

INTRODUCTION Pennoyer1952). Despite males having slightly larger wings, culmenand tarsi,Murphy andPennoyer (1952) found that the and a small island off differencesbetween sexes were not significant enough to (Philip lsland) supportthe only known breedingpopulations of concludesexual dimorphism in this species. the Providence Petrel Pterodroma solandri in the world (Fullagaret al. 1974);consequently the speciesis listed as The ability to sexindividual Providence Petrels provides a vulnerableunder both the 2000 IUCN Red List of Threatened meansto investigatesex-specific differences in foraging and (Criteria D2) and the New South Wales Threatened breedingbehaviour. Unfortunately many speciesare very SpeciesConservation Act 1995, althoughit is not listed under difficult to sex in the hand (Robertson €/ al. 2000), commonwealth legislation (Envircnment Protection and particularly seabirds.Based on the information available Biodiversity ConservationAct 1999).On Lord Howe Island, (Murphy and Pennoyer1952; Marchant and Higgins 1990),it ProvidencePetrels breed predominantly on Mt Lidgbird andMt appearsthat male ProvidencePetrels are slightly heavierand Gower, with smaller populationsoccurring at lower elevations havelonger wings, culmenand tarsi.Some studies have used on the southernsection of the island(Fig. 1). a discriminantanalysis function to sexbirds (Anderson1975; Fullagarand Disney1981; Johnstone and Niven 1989),which Adult body mass among the Pterodroma is highly is able to determineparticular morphological characteristics variable, ranging from 163 grams for Pycroft's Petrel P that distinguishbetween sexes. Unfortunately this technique pycrofti to 525 grams for the Great-winged Petrel P canbe quiterestrictive (Johnstone and Niven 1989)and is also macroptera macroptela (Warham 1990). The Providence subjectto high error.Johnstone and Niven (1989) suggested Petrelis at the heavierend of the scale,although prior to this that the inclusionof masswould restrictthis procedureto the study, morphometricdata were scant and based on small startof the incubationperiod, given changesin massover the samplesizes. Data providedin Marchantand Higgins (1990) season.Nowadays most studiesare using DNA analysisas a suggestthat the averagemass of males(517.6 x.64.32 g, tool for sexingindividuals. Molecular analysis is currentlythe 443-600g, n = 3) is greaterthan that of females(423.5 t 9.40 most feasibleand accurate(99.997o) method for determining g, 414433 g, n = 2). sex (Griffiths et al. 1998).Another advantageof molecular Wing lengthhas been measured by Murphy and Pennoyer analysisis that it can be conductedthroughout the year. (1952)at 306.1t 5.45 millimetres(296-317 mm, n = 25) for malesand 302.1x.7.59 millimetres (284-316 rnrn,n = 58) for Much morphological information regarding the females. The method used to sex these birds was not ProvidencePetrel is still lacking andit is unclearwhether this mentioned by the authors, although it was probably by speciesis sexually dimorphic based on previous research. dissectionand internal examination. In ProvidencePetrels. the Small samplesizes provided by Marchantand Higgins( 1990) bill is laterally compressedand considerablyhooked (Bailey and lack of infbrmationregarding the methodused for sexing et al. 1989).The culmen measures34.6 + 1.05 millimetres by Murphy and Pennoyer(1952) necessitatea more rigorous 32.7-37 .l mm. ,?= 25) in malesand 34.3 + I .18 millimetres study to supporttheir findings.This study will thereforeuse (30.1-36.2 mm, r? = 58) in females(Murphy and Pennoyer molecular analysis to accurately sex individuals and 1952). Tarsi are rounded anteriorly with some slight determine if this species is sexually dimorphic. Such compression,and have been measuredat 42.6 + 1.09 information will be used in other studiesof this speciesto millimetres(40.944.1 rlm, n = 25)for malesand 41.8 t 1.04 determine any sex-specificdifferences in foraging and millimetres(39.144.6 mm, n = 58) for females(Murphy and breedinsbehaviour.

10 March2007 A.J.Bester, N.l. Klomp and D. Priddel: Sexual Dimorphism inthe Providence Petrel Using DNA Analysis 11

METHODS The mass and morphometric measurements of each DNA- sexed bird were compared using paired t-tests (to avoid Lord Howe Island (31'33'5, 159"05'E)is locatedin the pseudoreplicationof data, following Gray and Hamer 2001) to South Pacific Ocean,770 kilometres northeastof Sydney and determine if, based on these particular morphological features, 800 kilometresnorthwest of New Zealand(Hutton 1991).Birds sexual dimorphism existed. werestudied on the summitof Mt Gower(Fig. l). RESULTS In total, 36 adultswere weighed and measuredover a one- week period during chick rearingin 2000.Adults were weighed The DNA analysisconfirmed the sex of 18 male/female to + 5gramsin calico bagsusing a 1000-gramPesola spring breedingpairs. The meanmass of thesemales (507.2 x. 41.5g, balance. Weights were only taken after adults had finished n = 18)and females (476.6 x.44.7 g, n = l8) werefound to be feeding their chick. Culmen lengths and tarsus lengths were significantlydifferent (Paired t-test:tlJ = 3.6, P = 0.002). measuredto + 0.05 millimetresusing dial callipers,and the Similarly, culmen measurementswere significantlydifferent maximum flattenedchord (wing) of eachbird was measuredto betweenmales and females (36.5 t 1.1mm and35.5 + 1.3mm, + 1.0millimetre using a butt-stoppedruler. Adults were banded respectively,n = 181'Paired t-test: tll = 2.8, P = 0.012).No with individuallynumbered, stainless steel bands provided by significant differencesbetween sexes were found for tarsus the AustralianBird andBat BandingScheme. measurements(Paired t-test tl,7 = l.l, P = 0.311),or wing chord (Pairedt-test: /15 = 0.7, P = 0.526).Table I providesa Each bird was sexedby Dr. David Groth from molecular comparisonof measurementsand weightstaken from male and analysisof DNA taken from feathersamples. Contour feathers femalebirds. werecollected from the mantleusing tweezers and stored in lock- tight plastic bags.Only featherscontaining the node were kept. DISCUSSION Featherswere sexedin the laboratoryby extractingthe DNA material from skin cells inside the feather shaft to detect the The meanmass of adult ProvidencePetrels in this studywas presenceof a particularsex-linked gene (followin g Giffrths et al. 507 gramsfor malesand 471 gramsfor females(n = 36). These I 998).This testgives an accuracyof 9999% and is considereda massfigures are somewhatdifferent to those in Marchant and better alternativethan sexing adults from cloacal examination Higgins (1990),although this may be a function of their smaller (Serventy1956) or from vocalisation(Warham 1990). samplesize (n = 5). Murphy and Pennoyer(1952) found slight

D

0 Admiralty .K. lslands o

Mount Eliza Nt I

TransitHill Mutton Bird o lsland

Study site

0 2tn

- Road - - Walkingtrack : Studyarea A Mountainsummil MountGower

e Gowerlsland

Figure l. The location of the study area on Lord Howe Island. 12 A.J.Bester, N.l. Klomp and D. Priddel: Sexual Dimorphism inthe Providence Petrel Using DNAAnalysis Corella31(1): 10-12

TABLE1 0.01Vo enor). Even though birds were found to be sexual dimorphic in this study it is suggestedthat other researchers Massand lengths of culmen, tarsus and maximum shouldcollect feathersand conductDNA analysisto accurately individuals.DNA analysis conductedat any time flattenedchord (MFC) of DNAsexed Providence Petrels sex can alsobe of the season,although there is lag time with the analysis. in2000 The ability to sexindividuals enables sex-specific differences Males Females in behaviourto be identified(Merton et al. 1984:Gonzdlez-Solis et al. 2000). This DNA analysis has been essentialto other Samplesize 18 18 ProvidencePetrel studiesinvestigating differences in foraging Meanmass (g) 507 477 and breedingbehaviour between sexes (Bester 2003). sD(g) 41.5 44.7 ACKNOWLEDGMENTS

Range(g) 420440 405-625 The Lord Howe IslandBoard providedlogistical support for this study.Nicholas Carlile, Lisa O'Neill, MichaelJarman, Robert Wheeler Samplesize '18 18 and numerousother volunteersprovided assistancein the field. Dr. David Groth undertookthe molecularanalysis of DNA extractedfrom Meanculmen (mm) ?46 35.5 feathersamples. The maunscriptwas improvedby commentsfiom five anonymousreferees. SD(mm) 1.1 1.3 Range(mm) 34.9-381 329-37.8 REFERENCES Anderson,A. (1975).A methodof sexingmoorhens. Wildfowl 26: Samplesize 18 47 71-82. Bailey,S. F., Pyle,P. and Spear,L.B. (1989).Dark Pterodromapetrels Meantarsus (mm) 45.3 44.7 in the North Pacific: identification,status, and North American 11 {a occurrence.American Birds 43: 400-415. SD(mm) t.l Bester,A. J. (2003).'The breeding,foraging ecology and conservation of Range(mm) 42.5-46.8 42.7-47.6 the ProvidencePetrel Pterodroma solandri breeding on Lord Howe Island,Australia.' Ph.D. thesis,Charles Sturt University,Albury. Fullagar,P J. andDisney, H. J. de S. (1981).Discriminant functions fbr Samplesize 17 16 sexingwoodhens. Corella 5: 106-109. MeanMFC (mm) 306 305 Fullagar,P. J., McKean,J. L. and VanTets, G. F. (1974).Appendix F, report on the birds. In 'Environmental Survey of Lord Howe SD(mm) 5.8 6 Island'. (Eds H. F. Recherand S. S. Clark) Pp.55-72. (AGP: Sydney.) Range(mm) 296-320 297-318 Gonziilez-Solis,J., Croxall, J. P. and Wood, A.G. (2000). Sexual dimorphism and sexual segregationin foraging strategies of northerngiant petrels, Macronectes halli, during incubation. Oikos differenceswith wing length, culmen and tarsi betweensexes, 90: 390-398. althoughresults were not statisticallysignificant. Results from Gray,C. M. andHamer, K. C. (2001).Food-provisioning behaviour of this studysupport Murphy and Pennoyer's (1952) findings, with male and female Manx shearwaters,Puffinus puffinus. the inclusionof significantdifferences found with culmenlength Behaviour62: 117-121, betweensexes. Griffiths,R., Double,M. C., On, K. and Dawson,R. J. G. (1998).A DNA testto sexmost birds. Molecular Ecology 7: l0'71-1075. Thus,this studyhas shown that ProvidencePetrels are indeed Hindwood,K. A. and Serventy,D. L. (1941).The Gould petrel of sexuallydimorphic based on greatermass and culmenlengths of CabbageTree Island. Emu 4l: 1-20. 'Birds males.Sexual dimorphism has only been recordeddefinitively Hutton,I. (1991). of Lord Howe Island:Past and Present.' (Ian for two otherPterodroma species: Trindade Petrel Pterodroma Hutton: Coffs HarbourPlaza.) arminjoniana(Murphy andPennoyer 1952),Grey-faced Petrel P Imber, M. J. (1971). Filoplumesof petrelsand shearwaters.New m. gouLdi(Imber 1971;Johnstone and Niven 1989)and possibly ZealandJournal of Marineand Freshwater Research 5: 396-403. Johnstone,R. M. andNiven, B. E. ( 1989).Sexing grey-faced petrels by in Gould's PetrelPterodroma leucoptera leucoptera (Hindwood discriminantanalysis of measurements.Notornis 36: 261-265. 'The and Serventy 1941). All studiesexcept for Hindwood and Marchant,S. and Higgins,P. J. (1990). Handbookof Australian, Serventy(1941) found that males were significantly heavier than New Zealandand Antarctic Birds, Vol. l.' (Oxford University females,while studieson the Grey-facedPetrel also recorded Press:Melbourne.) longerand deeperbills in malebirds. Male Gould's Petrelswere Merton, D. V., Monis, R. B. and Atkinson, I. A. E. (198.1).Lek found to havelonger culmen lengths (Hindwood and Serventy behaviourin a parrot: the kakapo Strigopshabroptilus of New 1941),however, sample sizes were too smallto positivelysupport Zealand.Ibis I 26: 227-283. sexualdimorphism in this species(n = 5 males;n = 3 females). Murphy,R. C. andPennoyer, J. M. ( 1952).Larger petrels of the genus Johnstoneand Niven (1989)used a discriminantfunction to sex Pterodroma.American Museum Novitates 1580: 1-43. Grey-facedPetrels with eight per cent error. A discriminant Robertson,B. C., Millar, C. D., Minot, E. O., Merton, D. V. and Lambert,D. M. (2000).Sexing the critically endangeredkakapo analysis function was also applied to the ProvidencePetrel, Strigopshabroptilus. Emu 100:336-339. per greater althoughit was subjectto I I centerror based on mass Serventy,D. L. (1956). A method of sexing petrels in field and larger culmen length of males (Bester 2003). Thus, observations.Emu 56:213-214. ProvidencePetrels should not be sexedin the hand using this Warham,J. (1990). 'The Petrels.Their Ecology and Breeding techniqueas DNA analysisproves to be more reliable (only Systems.'(Academic Press: London.)