818 SHORT COMMUNICATIONS ashi Matsudo, David Nobel, and Charles Francis for R. GOLDIZEN.1994. Parentage analysis of multi- providing blood samples. S. C. Lougheed, C. J. Oliar- male social groups of Tasmanian Native Hens nyk and three anonymousreviewers provided helpful (Tribonyx mortierii): genetic evidence for monog- comments on the manuscript. amy and polyandry. Behav. Ecol. Sociobiol. 35: 363-371. LITERATURE CITED GYLLENSTEN,U. B., AND H. A. ERLICH. 1988. Gener- ation of single-strandedDNA by the polymerase BROOKE, M. DE L., AND N. B. DAVIES. 1987. Recent chain reaction and its application to direct se- changesin host usage by cuckoos Cuculus cano- quencing of the HLA-DQ-A locus. Proc. Natl. rus in Britain. J. Anim. Ecol. 56:873-883. Acad. Sci. 85:7652-7656. BROOKE,M. DE L., AND N. B. DAVIES. 1988. Egg mim- MORITZ,C., T E. DOWLING,AND W. M. BROWN.1987. icry by cuckoos Cuculus canorus in rel%on to Evolution of animal mitochondrial DNA: rele- discriminationbv hosts. Nature 335:630-632. vance for population biology and systematics. CABOT,E. 1990. Thi eyeball sequenceeditor, version Annu. Rev. Ecol. Syst. 18:269-292. 1.09d.Dept. Biology,Univ. Rochester,Rochester, NY MOUM, T, AND S. JOHANSEN.1992. The mitochondrial DESJARDINS,R., AND P MORAIS. 1990. Sequenceand NADH dehydrogenasesubunit 6 (ND6) gene in gene organization in the chicken mitochondrial murres: relevance to phylogenetic and population genome: a novel gene order in higher vertebrates. studiesamong birds. Genome 35:903-906. J. Mol. Biol. 212:599-634. MOUM, T., S. JOHANSEN,K. E. ERIKSTAD,AND J. E EDWARDS,S. V. 1993. Mitochondrial gene genealogy PIA~T. 1994. Phylogeny and evolution of the auks and gene flow among island and mainland popu- (subfamily Alcinae) based on mitochondrialDNA lations of a sedentarysongbird, the Grey-crowned sequences.Proc. Natl. Acad. Sci. 91:7912-7916. Babbler (Pomatostromus temporalis). Evolution QUINN,T. W., AND A. C. WILSON. 1993. Sequenceevo- 47:1118-1137. lution in and around the mitochondrial control re- EDWARDS,S. V., I? ARCTANDER,AND A. C. WILSON. gion in birds. J. Mol. Evol. 37:417-425. 1991. Mitochondrial resolution of a deep branch SANGER,E S., S. NICKLEN,AND A. R. COULSON.1977. in the genealogical tree for perching birds. Proc. Sequencing with chain terminating inhibitors. R. Sot. Lond. B. 243:99-107. Proc. Natl. Acad. Sci. 7415463-5467. GIBBS, H. L., M. DE L. BROOKE,AND N. B. DAVIES. SLATER.l?. F?SLATER. AND R. SLATER.1994. The Slat- 1996. Analysis of genetic differentiation of host er field guide to Australian birds. Lansdowne, races of the Common Cuckoo Cuculus canorus Sydney. using mitochondrial and microsatellite DNA vari- WOOD,T C., AND C. KRAJEWSKI.1996. Mitochondrial ation. Proc. R. Sot. Lond. B 263:89-96. DNA sequencevariation among the subspeciesof GIBBS,H. L., A. W. GOLDIZEN,C. BULLOUGH,AND A. SarusCrane (Grus antigone). Auk 113:655-663.

The Condor 99:818-825 0 The CooperOmlthological Society 1997

MORPHOLOGICAL AND VOCAL VARIATION AMONG SUBSPECIES OF THE BLACK-FACED SHEATHBILL ’

JOELBRIED AND PIERRE JOUVENTIN Centre National de la Recherche Scientique,’ Centre dEtudes’ Biologiques de Chiz& 79360 Beauvoir sur Niort, France, e-mail: [email protected]

Abstract. The Black-faced Sheathbill (Chionis mi- and Kerguelen-Heard) on geographic and morpholog- nor) is a sedentary and polytypic species. Four allo- ical criteria. In the eastern group (Kerguelen-Heard), patric subspeciesare known, each breeding on one ar- corresponding to higher latitudes, sheathbills were chipelago in the Southern Indian Ocean. To evaluate larger and heavier, following Bergmann’s Rule. The the degree of isolation of these four subspecies,mor- sheathbills from Iles Kerguelen also had a lower- phometricsand vocalizationsof adult birds of Iles Ker- pitched voice than those from Iles Crozet, consistent guelen and Crozet were compared with those of the with their larger body size. Moreover, the birds from other localities (Prince Edward and Heard Islands). the southernmostlocality (Heard Island) had a shorter Two groupswere distinguished(Prince Edward-Crozet culmen, consistentwith Allen’s Rule, but longer tarsi and deeper sheaths.Within the western group (Prince Edward-Crozet), and at Iles Kerguelen, there also was ’ ’ Received 8 July 1996. Accepted 13 March 1997. variability on a microgeographicalscale. Differences SHORT COMMUNICATIONS 819

Kerguoelen Crozet ;:a Prince Edward He&d

\

ANTARCTICA

FIGURE 1. Breeding range of the Black-faced Sheathbill in the SouthernIndian Ocean. Island namesindicated in bold.

between subspeciesof Black-faced Sheathbill there- may be achieved by combining morphological and fore could be due not only to environmentalcorrelates ethological studies. of latitude, but also to possible genetic drift. The four subspecies are allopatric and do not differ in their METHODS breeding scheduleor in their general behavior and diet, Field work was conductedat Iles Kerguelen (48”27’- suggestingthat differentiation may be recent and main- 5o”S, 68”37’-70”35’E) in late December 1992, and at ly due to geographicalisolation. Ile de la Possessionin the eastern part of the Crozet archipelago(45”50 ’-46”3O’S, 50”-53”21’E) in Novem- Key words: speciation, reproductive isolation, ber 1994. Adult Black-faced Sheathbills were caught morphometrics,vocalizations, Chionis minor, Black- using a shepherd’s hook. We identified them as adults faced Sheathbill. becausethey were ringed, and known to have bred in the past or to be at least 3 years old, which is the The biological speciesconcept (Mayr 1970) considers minimum age of first breeding (Burger 1980a, Jouven- speciesas groupsof naturalpopulations capable of in- tin et al. 1996). Measurements (wing length, tarsus terbreedingand separatedfrom other similar groupsby length, culmen length, culmen width, and sheath mechanisms of reproductive isolation, such as geo- depth) were taken for each individual, following Bur- graphical, morphometrical,coloration, behavioral, and ger (1980b). Black-faced Sheathbills are sexually di- phenological. However, variability within the range of morphic with males larger than females (Burger a given species also can occur. Thus, geographical 1980b). Therefore, birds were sexed from size and bill variations in body size and coloration are well docu- shape index (i), the latter defined as follows (Warham mented in seabirds,particularly in Procellariiforms or 1972, Burger 198Ob): in the White-tailed Tropicbird Phaethon lepturus (de1 i = culmen length X culmen width Hoyo et al. 1992). Geographicalreproductive isolation X sheath depth X 0.1 also seems to play a major role in the Black-faced Sheathbill Chionis minor. Four subspeciesusually are Morphometrics of birds from Iles Kerguelen were recognized from body size, leg coloration and carun- comparedwith those from Prince Edward Islands (data cles, each subspeciesbreeding on a remote archipelago from Marion Island, A. E. Burger, pers. comm.) and in the Southern Indian Ocean (Marchant and Higgins Be de la Possession,using univariate statistics(Krus- 1993): C. m. marionensisat Prince Edward Islands, C. kal-Wallis H-tests and multiple comparisonpairwise z- m. crozettensisat Iles Crozet, C. m. minor at Iles Ker- tests; Scherrer 1984). We obtained the length of the guelen, and C. m. nasicornisat Heard Island (Fig. 1). appendagesrelative to individual body size by dividing Black-faced Sheathbills are sedentary and have never tarsusand culmen length by wing length (see Aldrich been recorded on the nearestcontinent or more than a and James 1991). However, wing length varies in the few kilometers offshore, becauseof their poor flying course of the year due to molt or abrasionof the outer abilities (Verheyden and Jouventin 1991). primaries. We measuredwing length during the austral This paper provides new and more accurate infor- spring and assumedthat the abrasionstage of the outer mation about biometry of Chionis minor than previ- primaries was similar in both Iles Crozet and Kergue- ously available. New insight into the speciation of len, which allowed us to compare these two localities Chionis minor and the role of reproductive isolation only. Multivariate analyseswere conducted using the 820 SHORT COMMUNICATIONS

SAS statisticalpackage (SAS Institute 1988); variables tarsi were significantly longer on Iles Crozet (Mann- were not transformed.A principal componentanalysis Whitney U-test, z = 3.69, P < 0.001, see Table 1). on correlation matrix between the variablesused (Dig- For Iles Kerguelen birds, measurementsfrom skins by and Kempton 1987) was conducted, using the were significantly smaller than those obtained from PRINCOMP procedure. Discriminant analyses also live birds (Mann-Whitney U-test, all P < O.OOl), ex- were performed by running a stepwise analysis in or- cept for culmen length and sheathdepth (sheath depth der to determine the two most discriminantparameters was significantly higher for skins in females: Mann- (STEPDISC procedure), before subsequentuse in the Whitney u-test, z = 2.39, P < 0.02). Kerguelen spec- discriminant function analysis (nonparametricmethod imens had a longer culmen than those from Heard Is- of the DISCRIM procedure using uniform kernel and land (at least for males, Mann-Whitney U-test, U = a posteriori cross-validationtechniques for estimating 14.5, rz, = 9, n, = 10, P < 0.01). Conversely, the latter error rates in classification). tended to have longer tarsi and deeper sheathsthan the Data from skins of birds from Heard Island and Iles former, althoughdifferences were not significant,prob- Kerguelen were compared separately because of the ably due to small sample size. small sample size and the way the skins had been pre- Multivariate analyses. Both principal component pared; mean measurementsfrom skins are found in and discriminant analyses revealed different popula- Marchant and Higgins (1993), the raw data were pro- tion groups. Within each population, there tended to vided by D. James (pers. comm). The measurements be severaldifferences, but thesedifferences were much taken from skins were not suitable for a multivariate less marked than those between populations. The in- analysis because they differed from those from live dividuals from Prince Edward Islands and Iles Crozet birds (see Results). So, we used them only as comple- were closer to each other than they were to those from mentary information. Iles Kerguelen. Principal component analysisused tar Visual observationsof bare parts coloration and ca- sus, culmen length and sheath depth, and Axis 1 and runcleswere performed at close range on a larger sam- Axis 2 explained 77.0% and 13.5% of the variance, ple of adult individuals (n = 55), and were easy to respectively. The two most discriminant parameters obtain becauseBlack-faced Sheathbills are very tame. were sheathdepth and culmen width, if we considered Calls (“Bob-calls,” see Burger 1980a) of Black- the birds from Prince Edward Islands, Iles Crozet and faced Sheathbills from Iles Crozet and Iles Kerguelen Iles Kerguelen simultaneously.Seventy-two percent of were recorded with a Sennheiser-MD 421 microphone the individuals subjected to a discriminant analysis and a Uher-4000C tape recorder. Frequency parame- based on sex and locality were correctly classifiedinto ters, including modal (MOD), lowest (LOF) and high- their a priori groups. If we consideredlocalities only, est (HIF) values of the fundamental frequency, were 78.5% of individuals were correctly classified. When analyzed after a fast Fourier transformation with an data from sheathbillsof Iles Crozet and Prince Edward Amiga micro-computer (sampling rate 6,512 Hz; step Islands were grouped and compared to those from Iles size 256 points, see Richard 1991). For all tests, prob- Kerguelen, the discriminant analysisperformed on the abilities 5 0.05 were consideredsignificant, and means two most discriminant parameterswithout considering are given ? SD. sex revealed that only 8.5% of individuals were mis- classified. RESULTS

BIOMETRY EXTERNAL FEATURES AND BARE PARTS COLORATION Univariate statistics.For each locality, measurements are given in Table 1. For live birds, differences be- As at Heard Island (see photographsin Downes et al. tween Iles Kerguelen, Prince Edward Islands, and Iles 1959), but not at Prince Edward Islands (photographs Crozet were significant for tarsuslength and bill mea- in Burger 1979, and drawings in Burger 1980a and surements(Kruskal-Wallis H-test, all P < 0.001, see 1980b) and Iles Crozet, adult Black-faced Sheathbills Table l), and values tended to be highest at Iles Ker- from Iles Kerguelen seem to have a roundedhead, due guelen, and lowest at Prince Edward Islands. However, to the absence of a conspicuoushead crest (Fig. 2), most measurements did not differ significantly be- and small and not very swollen caruncles.Moreover, tween Prince Edward Islands and Iles Crozet. Iles Ker- the carunclesat the baseof the bill of minor are present guelen females tended to be larger than Iles Crozet and at the sides of the upper mandible only, whereas birds Prince Edward Islands males for culmen length and of Iles Crozet and Prince Edward Islands also have sheathdepth, but these differences were not significant prominent caruncleson the forehead at the base of the (multiple comparison pairwise z-test, all P > 0.05). culmen and, at least for those from Iles Crozet, at the Iles Crozet sheathbills tended to have the widest and base of the lower mandible and on the chin (Marchant the shortestculmens. Iles Kerguelen males had signif- and Higgins 1993; this study). The orbital ring has the icantly longer wings than IlesKerguelen females-and same color whatever the locality. At Ile de la Posses- Iles Crozet individuals (Kruskal-Wallis H-test. H, = sion, birds have pinkish gray legs, whereas those from 55.0, P < 0.001, multiple comparisonpairwise z-test, Iles Kerguelen have pink legs. Legs also are pink at all P < 0.05); wing length did not differ significantly Heard Island, but they are sometimes black at Prince between Kerauelen females and Iles Crozet males and Edward Islands (Table 1). There was no seasonalvari- females (muifiple comparisonpairwise z-test, all P > ation in bare parts coloration at Iles Crozet and Ker- 0.05). Relative to body size, culmen length did not guelen. At Prince Edward Islands, the orbital ring “ap- vary significantly between Iles Crozet and Kerguelen peared to be brighter [ ] in some birds” (Burger (Mann-Whitney U-test, z = 0.48, P > 0.05), whereas 1980b) during the breeding season. SHORT COMMUNICATIONS 821

TABLE 1. Measurements(mm; mean ? SD) and coloration of bare parts of adult Black-faced Sheathbills in different localities. Material and references: (1) live birds, this study, (2) live birds, Burger (pers. comm.), (3) skins, mean values in Marchant and Higgins (1993), raw data from D. James (pers. comm.). All sex differences significant (P < 0.01) except for sheath depth in the museum specimensfrom Heard Is.

n Iles Kerguelen (1) n Prince Edward Is. (2) n Iles Crozet. (1) n Iles Kerguelen (3) n Heard Is. (3) Wing males 30 245.8 i- 4.9 24 218.0 t 5.9 7 221.0 2 2.6 9 234.2 -c 4.5 9 233.6 ? 6.1 females 29 234.0 ? 4.4 28 207.3 i 5.2 6 211.3 ? 4.0 6 222.7 r. 2.9 9 220.2 2 5.4 Tarsus males 31 48.4 2 1.7 45 46.7 C 1.6 7 45.9 2 1.6 9 44.6 ” 1.4 10 45.0 Ifr 1.4 females 29 45.5 ? 1.5 39 44.1 * 1.1 6 43.3 2 1.6 6 41.8 i- 2.8 9 42.5 2 0.7 Culmen length males 30 33.7 ? 1.0 45 32.0 ‘- 0.9 7 29.9 ? 0.6 9 33.7 r 0.9 10 32.4 2 1.1 females 29 31.1 L 0.9 38 30.0 * 0.9 6 28.4 r+ 0.7 6 31.0 ? 0.9 9 29.7 ? 1.3 Sheath depth males 30 20.9 2 1.1 34 16.8 ? 0.8 7 16.9 ” 0.6 7 21.3 ? 1.5 3 22.1 2 2.6 females 29 17.9 2 1.0 26 14.5 5 0.7 6 15.8 2 0.7 6 18.9 2 0.7 5 20.2 2 2.3 Culmen width males 30 9.4 i 0.7 43 9.4 -c 0.4 7 10.9 t 0.4 females 29 8.6 * 0.4 36 8.8 2 0.5 6 9.7 ? 0.4 - Bill shape index males 30 660.2 ? 68.9 33 515.1 2 41.6 7 552.0 ? 34.9 females 29 480.5 ? 47.3 26 390.8 2 30.8 6 437.7 + 33.1 Tarsus/wing both sexes 58 0.185 “_ 0.043 13 0.205 + 0.007 Culmemwing both sexes 59 0.130 2 0.006 13 0.135 i: 0.005 Bare parts coloration legs flesh-pink 81.3% pale pink pinkish gray flesh-pink flesh-pink to 18.7% dark pink to white-pink pinkish white or black orbital ring pink “usually bright pinkish mauve pink pink or flesh. pink” pink

VOCALIZATIONS tend to increasein the coldest parts of a species’ range, Sonogramsare shown in Figure 3. Iles Kerguelen in- which generally are at highest latitudesand/or altitude. dividuals had a significantly lower-pitched voice than Moreover, the sheathbills from Iles Kerguelen have those from Iles Crozet: the MOD, LOF and HIF values proportionally shortertarsi than thosefrom Iles Crozet, of the fundamental frequency of Bob-calls were sig- consistentwith Allen’s Rule: appendagesand extrem- nificantly lower in the Kerguelen sheathbills than in ities tend to be shorterin the coldestparts of a species’ their Crozet conspecifics(Mann-Whitney u-test, all P range. Within the minor-nasicornisgroup, individuals < 0.001). The differences were so large that no overlap from Heard Island have the shortestculmens (and per- occurred between the two populations (Table 2 and haps the shortest wings if the abrasion stage of the Fig. 4). outer primaries is similar to that of the specimensfrom Iles Kerguelen), also consistentwith Allen’s Rule. By DISCUSSION contrast, their sheaths seem the most prominent, the Our analyseshave confirmed the distinction by Peters nonsignificanceof the difference being possibly due to (in Marchant and Higgins 1993) of two main groups our small sample size. But on the whole, the differ- of Black-faced Sheathbills:marionensis-crozettensis in ences in morphometricssuggest an adaptive response the Western Indian Ocean (Fig. 1) and minor-nasicor- to higher latitudes and a colder climate, especially at nis in the Eastern Indian Ocean; the specimensfrom Heard Island where the mean annual temperature is Heard Island were not significantly different from 1.2”C (Woehler 1991) vs. 4.5”C at Kerguelen (Wei- those from Iles Kerguelen. Individuals belongingto the merskirch et al. 1989). Within the marionensis-croz- minor-nasicomis group live at higher latitudes (48”27’ ettensisgroup, morphometricdifferences cannotbe ex- to 50”s for Iles Kerguelen and 53”05’S for Heard Is- plained by Allen’s Rule, becauselatitudes (see above) land vs. 45”50’ to 46”3O’S for Iles Crozet and 46”54’S and mean annualtemperatures (5.3”C at Prince Edward for Prince Edward Islands) and have a larger body size Islands [Derenne et al. 19761vs. 4.8”C at Iles Crozet than those from the marionensis-crozettensisgroup; [Bost et al. 19921) are similar in both localities. Bare this agrees with Bergmann’s Rule: body size and mass parts, caruncle size, head shape, and coloration of legs 822 SHORT COMMUNICATIONS

FIGURE 2. Adult Black-faced Sheathbills. Top: C. m. minor. Bottom: C. m. crozerrensis standing behind a King Penguin (Aprenodytes pafugonicus). SHORT COMMUNICATIONS 823

KERGUELEN a

6

700 3 I

600

0” s g 500 !?! 0.5 ’ u CROZET 400

3 300 t 6,

200 4 iij456+6 Individuals FIGURE 4. Mean values of the modal, lowest and highest componentsof the fundamental frequency of Bob-calls recorded from Iles Crozet (n = 6) and Ker- guelen (n = 8). Vertical bars:range; boxes: SD; figures above plots = number of calls per individual.

ical isolation: Prince Edward Islands and Iles Crozet FIGURE 3. Sonogramsshowing Bob-calls of Chion- are 1,000 km apart, the distance between Iles Crozet is minor minor from Iles Kerguelen (top) and C. m. and Kerguelen is 1,500 km, and between Iles Kergue- crozerrensis from Iles Crozet (bottom). len and Heard Island is about 520 km (Fig. I). Fur- thermore, geographical isolation seems to have oc- curred not only between the four populationsof Black- also differ between the two main groups(Table I). But faced Sheathbillswhich are all allopatric and sedentary contrary to caruncle size and head shape, leg colora- (Marchant and Higgins 1993), but also between pop- tion also varies on a microgeographical scale. This ulationsof the same subspecies:movements have nev- variability is found at Prince Edward Islands (Table I), er been recorded between Marion and Prince Edward but also at Iles Crozet. In this locality, legs are grayish Islands, only about 20 km apart (Burger 1979). nor pink in 100% of adult sheathbillsfrom Ile de la Pos- between the western and the easternpart of the Crozet sessionin the eastern part of the archipelagovs. 4.8% archipelago, 170 km apart (P Jouventin,pers. observ.). of adults from Iles des Pingouins in the western part BecauseBlack-faced Sheathbills are poorly suited for of the archipelago(unpubl. data). long-distanceflights (Verheyden and Jouventin 1991). Our results suggest that the differences observed it is likely that they colonized the archipelagosof the within the Chionis minor speciesare due to geograph- SouthernIndian Ocean with the help of the strong pre-

TABLE 2. Mean modal (MOD), lowest (LOF) and highest(HIF) valuesof the fundamentalfrequency of the Bob- calls recordedfrom Kerguelenand Crozet Black-facedSheathbills.

Locality MOD (Hz) LOF (Hz) HIF (Hz) Iles Kerguelen (n = 8) 322.8 ? 28.3 258.7 -c 28.9 381.5 ? 36.7 Iles Crozet (n = 6) 679.7 ? 83.4 545.4 f 31.5 739.9 5 87.5 824 SHORT COMMUNICATIONS vailing westerly winds. Colonization probably oc- 1992). For Black-faced Sheathbills, which are terres- curred from the West to the East, perhapsstarting from trial (Verheyden 1988) the presence of an extensive Patagonia, South Georgia or the Antarctic Peninsula, intertidal zone at Iles Kerguelen gives additional op- where the Pale-faced Sheathbill Uzionis alba lives portunities for feeding and maintaining territories on (Jones 1963). However, it is impossible to know ac- shores free of penguin and cormorant colonies. How- curately where the differentiation processbetween C. ever, the intertidal zone is almost absent from other alba and C. minor started, as the fossil remains of the localities (Jouventin et al. 1996). Thus, the greater ancestor of sheathbills have not yet been discovered number of habitats at Iles Kerguelen can explain the (Lowe 1916). level of heterozygosity in the sheathbill population, Evolution of Black-faced Sheathbills in the which is three times higher at Iles Kerguelen than at Southern Indian Ocean could have proceeded as fol- Iles Crozet (Viot et al. 1993). Population size lows. In the most southernand coldest localities (Ker- (3,000-5,000 pairs at Iles Kerguelen vs. 2,000-3,000 guelen, and more especially Heard Islands), a greater pairs at Iles Crozet, Jouventin et al. 1988) might also body size evolved, following Bergmann’s Rule. At the be another factor of intrapopulational variability. same time, the head crest disappeared,the caruncles Therefore, genetic variability at Iles Crozet may be became less prominent and the culmens at Heard Is- insufficient to explain differences between islands. land shorter, in agreement with Allen’s Rule. The Coloration of legs in sheathbillsfrom Ile de la Posses- voice, at least at Iles Kerguelen, became lower-pitched sion might be explained by a founder event; this island than in the northern part of the breeding range (partic- was probably colonized from the western part of the ularly at Iles Crozet), consistentwith larger body size Crozet archipelago. as suggestedby previous studies (Guillotin and Jou- However, Black-faced Sheathbills breed at similar ventin 1980, Wallschllger 1980). dates (Jouventinet al. 1996) and seem to perform the Other climatic factors such as humidity, and trophic same agonistic and pair displays irrespectiveof breed- factors also have been invoked to accountfor morpho- ing locality (for Prince Edward Islands see Burger logical differences between bird populations (Aldrich 1980a; for Iles Crozet and Kerguelen, C. Verheyden and James 1991, Moen 1991). Thus, it was necessary and J. Bried, pers. observ.), suggestingthat the speci- to take into accountwet-bulb temperature,evaporative ation processis recent (0.4 to 1 million years ago be- power of the air and foraging constraintsto explain the tween Crozet and Kerguelen subspecies,according to variations in the American Robin Turdus migratorius, Viot et al. 1993). This latter hypothesis also is sup- a partially migratory species,the different populations ported by the known agesof the islands:about 45 mil- of which breed in different habitats and under very lion years for Kerguelen and Heard Islands, 10 million contrastingconditions (Aldrich and James 1991). Yet, years for the eastern part of the Crozet archipelago, the climate of the subantarctic archipelagos in the less than 500,000 years for the western part- (Gi%et SouthernIndian Ocean is cold-temperateand oceanic. 1987). and about 300.000 Yearsfor Prince Edward Is- Moreover, these archipelagos are very small land lands (McDougall 1971). * massesdispersed among large ocean areas.The largest More data are still needed: morphometrical data of them is Kerguelen (7,200 km*, almost as large as from a large sample of live individuals are lacking for Corsica); Heard, Prince Edward and Crozet Islands, Heard Island. A genetic study, including electropho- each covering less than 500 kmr. Temperaturesand retie and mitochondrial DNA analysis, would allow a precipitation at these localities do not vary very much comparison with similar data collected at Iles Crozet and Kerguelen (Viot et al. 1993) and provide infor- in the course of the year, due to the presence of the mation about within-group differentiation and geo- ocean which acts as a buffer. Whatever the locality, graphicalisolation processes.But other mechanismsof Black-faced Sheathbills live on the coast where the reproductiveisolation, suchas vocalizationsand visual buffer effect of the ocean is maximized. They also signals (head patterns and bare parts), are useful in have a wide dietary spectrum,being opportunisticfor- avian systematics(e.g., Lanyon 1969, Jouventin 1982, agers (Marchant and Higgins 1993, Jouventin et al. Pierotti 1987). Acoustic and visual species-specific 1996). Consequently, we can consider mean annual recognition could be investigatedto determine the ex- temperatureand latitude as the only environmentalpa- tent of behavioral differences between the subspecies rameters that vary significantly throughout the range of the Black-faced Sheathbill, perhaps by recording of this sedentary species. Hence, some of the differ- calls from Heard and Prince Edward Islands and con- ences observed among the different subspecies of ducting playback experiments, and by performing ex- Chionis minor (head pattern variations,culmen length, perimental modifications of caruncle size, head crest and wing length) probably representa responseto lat- and leg coloration. itude and temperature, whereas the discrepancy with Allen’s Rule concerningthe long tarsi and deep sheath The Administrations of the Institut Fran9aispour la of Heard Island sheathbills compared to those from Recherche et la TechnologiePolaires, and Terres Aus- Iles Kerguelen suggeststhat other morphometrics(tat- trales et Antarctiques Francaises provided logistical sus length and sheath depth at Heard Island), and leg and financial supportto this work, respectively.We are coloration might be the consequenceof genetic drift greatly indebited to A. E. Burger, who kindly provided or a founder event. us with his data from Marion Island. D. Jamesprovid- There also can be variability within the same local- ed us with measurementsfrom museum specimens ity. Thus, the influence of particular ecological con- from Heard Island and Iles Kerguelen. Thanks also to ditions at Iles Kerguelen on body size and mass also E Mougeot, D. Haubreux and P Lys for field assis- has been shown for two seabird species (Bost et al. tance, to C. Verheyden, V. Bretagnolle and E Mougeot SHORT COMMUNICATIONS 825 for helpful suggestionsand statisticaladvice. A. Mar- Australes et AntarciquesFranGaises. Livre Rouge iani helped with acoustics,and C. Mauget with com- des oiseaux menacCs des rCgions franqaises puter processing. L. Jouventin-Ruchonhrew the fig- d’outremer. C.I.P.O., Int. Council Bird Preserv. 5: ures. W. D. Koenig. D. J. Gubernick. S. J. G. Hall. A. 225-25 1. E. Burger, and t\;o anonymous reviewers provihed LANYON,W. E. 1969. Vocal charactersand avian sys- useful criticism and improved the English. tematics, p. 291-310. In R. A. Hinde [ed.], Bird vocalizations.Cambridge Univ. Press,Cambridge. LITERATURE CITED LOWE,P. R. 1916. Studies on the CharadriiformesIII. Notes in relation to the systematicposition of the ALDRICH,J. W., AND E C. JAMES.1991. Ecogeographic Sheath-bills (Chionidida~). Ibis lO:i22-155. variation in the American Robin (Z’urdus migra- MARCHANT.S.. AND P. J. HIGGINS.1993. Handbook of torius). Auk 108:230-249. Australian, New Zealand and Antarctic birds. Vol. BOST,C. A., l? JOUVENTIN,AND N. PINCSONDU SEL. 2. Oxford Univ. Press, Melbourne. 1992. Morphometric variability on a microgeo- MAYR, E. 1970. Populations, species and evolution. graphical scale in two inshore feeders. J. Zool. Belknap Press, Cambridge, MA. 226: 135-149. MCDOUGALL,I. 1971. Geoch;onology,p. 72-77. In E. BURGER,A. E. 1979. Breeding biology, moult and sur- M. Van Zinderen Bakker. St. J. M. Winterbottom. vival of Lesser Sheathbills Chionis minor at Mar- and R. A. Dyer [eds.], Mar&n and Prince Edward ion Island. Ardea 67:1-14. Islands. A. A. Balkema, Cape Town, SouthAfrica. BURGER,A. E. 1980a. An analysis of the displays of MOEN,S. M. 1991. Morphologic and genetic variation Lesser Sheathbills Chionis minor. Z. Tierpsychol. among breeding colonies of the Atlantic Puffin 52:381-386. (Fratercula arctica). Auk 108:755-763. BURGER,A. E. 1980b. Sexual size dimorphism and PIEROTTI,R. 1987. Isolating mechanismsin seabirds. aging charactersin the Lesser Sheathbill at Mar- Evolution 41:559-570. ion Island. Ostrich 51:39-43. RICHARD,J. P 1991. Soundanalysis and synthesisusing DEL HOYO, J., A. ELLIOTT,AND J. SARGATAL.1992. an Amiga micro-computer.l%oacousti& 3:45-60. - Handbook of the birds of the World. Vol. 1. Lynx SAS INSTITUTEINC. 1988. SAS/STAT user’s guide. Editions, Barcelona. Version 6.03. SAS Institute, Inc., Cary, NC_ DERENNE,PH., J. L. MOUGIN,C. STEINBERG,AND J. E SCHERRER,B. 1984. Biostatistique. Ga&an Morin, VOISIN. 1976. Les oiseaux de 1’Ile aux Cochons, Boccherville, QuCbec,Canada. Archipel Crozet (46”06’S, 50”14’E). Corn. Nat. VERHEYDEN,C. 1988. La formation de couples hiv- Franc. Rech. Antarct. 40:17-148. ernaux chez le Petit Bet-en-fourreau (Chionis mi- DIGBY, F?G. N., AND R. A. KEMFTON. 1987. Multi- nor) et sa signification biologique. Rev. Ecol. variate analysis of ecological communities.Chap- (Terre & Vie) 43:47-59. man and Hall, London. VERHEYDEN,C., AND F?JOUVENTIN. 1991. Over-win- DOWNES,M. C., E. H. M. EALEY,A. M. GWYNN,AND tering strategiesof the Lesser Sheathbill Chionis P S. YOUNG. 1959. The birds of Heard Island. minor in an impoverished and insular environ- Aust. Nat. Antarct. Res. Exped. Rep., SeriesB (I). ment. Oecologia 86:132-139. GIRET,A. 1987. GCologie des Terres Au&ales Fran- VIOT, C. R., P JOUVENTIN,AND J. BRIED. 1993. Pop- Faises.Corn. Nat. Franc. Rech. Antarct. 58:17-41. ulation genetics of southernseabirds. Marine Or- GUILLOTIN,M., AND E?JOUVENTIN. 1980. Le P&e1 des nithol. Tl:l-25. neiges B Pointe GCologie.Gerfaut 7051-72. WALLSCHLAGER.D. 1980. Correlation of song fre- JONES.N. V. 1963. The Sheathbill. Chionis nlba quency and body weight in passerine birds. Ex- (bmelin) at Signy Island, South &kney Islands. perientia 36:412. Br. Antarct. Surv. Bull. 2:53-71. WARHAM,J. 1972. Breeding seasonsand sexual di- JOUVENTIN,P 1982. Visual and vocal signals in pen- morphism in Rockhopper Penguins. Auk 89:86- guins, their evolutive and adaptive characters. 105. Adv. Ethol. 24:1-149. WEIMERSKIRCH,H., R. ZOTIER, AND F? JOUVENTIN. JOUVENTIN,P., J. BRIED,AND E. AUSILIO. 1996. Life- 1989. The avifauna of the Kerguelen Islands. history variationsof the Lesser Sheathbill Chionis Emu 89:15-29. minor in contrastinghabitats. Ibis 138:732-741. WOEHLER,E. J. 1991. Status and conservationof the JOUVENTIN,P, J. C. STAHL, AND H. WEIMERSKIRCH. seabirds of Heard Island and the McDonald Is- 1988. La conservation des oiseaux des Terres lands. Int. Council Bird Preserv. 11:263-277.