sign in sign up
<<
Home , Black noddy, Frigatebird, Great frigatebird

The Condor96987-993 Q The Cooperomithological society 1994

GREAT : SEX-SPECIFIC HOST CHOICE AND AGE-RELATED PROFICIENCY ’

JAMESD. GILARDI Section of Evolution and Ecology, Universityof California, Davis, CA 95616

Abstract. Avian kleptoparasitismis a specializedforaging behavior requiring rapid flight speedsand extreme agility. As such, it is expected that proficiency will correlate with ex- perience and that the size of the kleptoparasite will influence its choice of host .I investigated age- and sex-differencesin the kleptoparasiticbehavior of Great (Fregata minor) on Midway Atoll. The results indicate that chasesby more than one fri- gatebird are more successfulthan solitary ones, and that successrates vary with the host pursued. In contrast with previous work, successimproved with age and sub-adultswere more likely to chaseas part of a groupthan were adults.Adults showedstriking sex-differences in their selection of host species,females preferring large speciessuch as and trop- icbirds while males preferred the smaller noddies and . The role of these sex-differences in the evolution of sexual size dimorphism is discussed.Finally, since marked breeding individuals were not observedin kleptoparasiticevents near the colony, this foraging mode is apparently not important to breeding in this population. However, in some cases the frigatebird kleptoparasitismmay directly influence the behavioral patterns and the sur- vival of host individuals. Key words: : Fregata minor; kleptoparasitism;age effects;sex-d$er- ences;.

INTRODUCTION both the parasite and host. As frigatebirds show The acquisition of previously caught prey, or delayed maturation and are sexually di- kleptoparasitism, is particularly common in the morphic in size and plumage (Fregata minor: Fregatidae, , and Stercorariidae (Brock- Females 1,800 g, males 1,000 g, Schreiber and mann and Barnard 1979, Fumess 1987). Near Schreiber 1988), predictions of age- and size/sex- many tropical colonies, frigatebirds pur- differencesare readily testable. sue a variety of species(hosts), frequently catch- In studies that have looked for age-differences ing them by the wing or tail, encouraging them in the successof kleptoparasitism by Magnificent to regurgitate prey items. In this specialized for- Frigatebirds (Fregata magnijicens), results were aging mode, many factors potentially effect the either equivocal or no differencesin successwere successof such attempts: i.e., the frigatebird’s detected (Gochfeld and Burger 198 1, Gibbs and assessmentof a host’s alertness,fullness of crop, Gibbs 1987, Osomo et al. 1992). These findings agility, location relative to land, as well as its contradict a wealth of evidence on avian foraging own capabilities of speed and agility, and the behavior demonstrating that proficiency im- presenceof potentially competing frigatebirds. proves with age (Recher and Recher 1969, Ori- In species with such highly skilled foraging ans 1969, Morrison et al. 1978, Carroll and Cra- techniques, performance typically improves with mer 1985, Weathers and Sullivan 1989). age and experience (Orians 1969, Carroll and Sex-differences in kleptoparasitic behavior Cramer 1985). Moreover, aerodynamic theory have been observed in several fiigatebird species predicts that physical parameters such as mass (Vemer 1965, Diamond 1975, Gochfeld and wing-loading constrain both agility and peak dive Burger 198 1, Gibbs and Gibbs 1987, Osomo et velocity in birds (Andersson and Norberg 198 1). al. 1992). In general, one sex performs virtually For frigatebirds then, it is expected that success all of the kleptoparasitic behavior although this of kleptoparasitic events should improve with varies with the colony and speciesof frigatebird. age and that successwill vary with the size of For instance, in Magnificent Frigatebirds females were more likely to chaseboobies in Belize (Ver- ner 1965), while in the Galapagos males were the more common kleptoparasites(Gibbs and Gibbs I Received 14 January 1994. Accepted 27 July 1994. 1987).

[9871 988 JAMES D. GILARDI

TABLE 1. Breeding populations, frequency of attack, and mass of potential host specieson Midway Atoll, 1989-1991.

Number Massg + SpXkS of indiv. 22~~” (ranee)* rare (Naous tenurostris) 3,000* 108 (85-133) frequent (Anousstolidus) 5,000* 205 (153-275) Grey-backed infrequent (Sterna Zunata) 1,ooot 146 (115-177) very rare (Diomedea immutabilis) 500,000* 2,400 (1,900-3,075) Red-footed frequent (Ma sula) 1,ooo* 1,110(905-1,400) Red-tailed very freq. (Phaethon rubricauda) 30,ooot 624 (540-750) very freq. (Sternafuscata) 100,000~ 198 (153-320) Wedge-tailed rare (Pu&n45 pacijicus) 400t 388 (320-5 10) rare (Gygis alba) 20,ooot 111 (92-139)

m&cator of how often this type of chase is observed. l+AroY Data om Harmon ’ et al. (1983). t Personal observation, 1989-1991.

In contrast, on Midway Atoll, all sexes and based on plumage characters (Harrison 1985). ages of Great Frigatebird commonly kleptopar- For each kleptoparasitic event the following data asitize a great diversity of host speciesvarying were collected: number of individuals involved, in size from Grey-backed Terns (100 g) to Laysan their sexes and ages, individual marks (see be- Albatross (2,400 g; Table 1). This study focuses low), successof the event, and the host species. on the four most commonly chased species: As the plumage of young frigatebirds is highly Brown Noddies, Sooty Terns, Red-tailed Trop- variable, I disregardeddistinctions between sub- icbirds, and Red-footed Boobies. During the frig- adult, immature, and juvenile birds, considering atebirds’ breeding seasonsof 1989-l 99 1, I eval- them all ‘sub-adults’ for the purposesof this study. uated age and sex-related differences in Kleptoparasitic attempts lasting longer than one kleptoparasitic behavior and explored the im- second were deemed chase events and I consid- portance of this behavior to breeding individu- ered events successfulwhen the frigatebird re- als. trieved prey either in flight or from the water. For group events, only frigatebirds that clearly STUDY AREA AND METHODS acquired prey were considered successful. Midway Atoll (28” N, 178” W) is near the north- In 1990, to determine whether breeding adults west end of the Hawaiian leeward chain and is were responsible for a significant portion of the comprised of a fringing reef 8- 10 km in diameter kleptoparasitism, 53 breeders were individually surrounding a shallow lagoon and three small marked with coded notchesin the outer rectrices. islands. The Great Frigatebird population num- These marks were observablewhen the birds were bers 500-900 non-breeders and a colony of 50- in flight, and allowed determination of both 80 breeding pairs (pers. observ. 1989-1991). breeding status and individual recognition. Al- Midway Atoll is also the breeding site for a large though at times these birds represented = 5% of and diverse community of potential kleptopar- both roosting and breeding frigatebirds, the lluc- asitic hosts (Table 1). tuating size of the roosting population and ac- I observed kleptoparasitic events from several cumulation of the marked group over time pre- locations on the three islands, from June-August vent knowing the exact proportion of marked 1989, May-July 1990, and June-August 1991, individuals. These observations therefore were using 9 x binoculars to age and sex individuals not treated statistically. GREAT FRIGATEBIRD KLEPTOPARASITISM 989

TABLE 2. Successof Great Frigatebird kleptoparasiticevents by all agesand sexes,success/total (o/o).

Brown Noddies SootyTerns Red-tailed Tropicbiids Red-footedBoobies

Females Solitary 4/14 (28.6) 16/95 (16.8) 2/10 (20.0) Group : 17/25 (68.0) 47/92 (51.1) 4/8 (50.0) Males

SolitaryGroup 17/192/12 (89.5)(16.7) 24/3915132 (61.5)(46.9) 8/602/13 (13.3)(15.4) : Sub-adults Solitary 2/10 (20.0) o/21 (0.0) o/2 (0.0) Group 5/6 :83.3) 9/28 (32.1) 7/24 (29.2) t t Not observed.

As the frigatebirds involved in kleptoparasitic of a group. Females clearly preferred the larger events were from a large pool ofindividuals com- host species, whereas males preferred smaller muting to or from the colony, I assume that few hosts. events involved the same individual (parasite or In the general effectsmodel, forward selection host). All variables used in the analysis were cat- indicated that all two-way interactions were egorical in nature: [A] age/adult-sex (sub-adult, significant and backward deletion included an male, female), [G] association with other frigate- additional three-way interaction [SHG] (Table birds in a chase(solitary, group), [H] host species 3A). This suggeststhat the successof kleptopara- (the four most common were analyzed, BRNO, sitic events was affectedby both the host species SOTE, RTRR, RFBO, seeTable l), and [S] suc- pursued and whether the fiigatebird chased sol- cessof the events (+ or -). itarily or with others (Fig. 1). Group events were I analyzed the data with log-linear models us- more successfulthan solitary ones, although as ing the BMDP 4F procedure (Dixon 1990) and the three way interaction suggests,this varied as selected the best model with a combination of a function of host species.As both ageand gender forward selectionand backward deletion (Agresti effects were significant ([SA] [HA] [GA]), they 1990). Similar to other stepwise methods, the were explored directly in the two subsequent forward selection procedure adds terms to the models. model starting with the most statistically signif- In the age model both selection procedures icant and stopping at a selectedlevel of signifi- yielded the same result: all but one two-way in- cance (P = 0.05). The first model includes all the teraction were significant ([AS] [AG] [AH] [SH] data to consider general effects. To look specif- [SG], Table 3B). Sub-adults differed from adults ically at age and sex effects,I ran two subsequent in three ways: the successof their attempts, the models, an age model comparing sub-adults and frequency with which they chased as part of a adults, and a gender model which explores sex- group, and the host speciesthey selected(Table differences comparing only males and females. 2). Sub-adults were less successfulthan adults in For each model, the significance of each term both solitary and group events and were also and the selected model is presented, and to fa- more likely to chase as part of a group. In com- cilitate interpretation, I also present some of the parison with adults, they also pursued tropic- data in graphical form, collapsing acrossone or birds and noddies somewhat more often than the more variables. other species. In the gender analysis, both selection proce- RESULTS dures converged on the same model which in- Frigatebirds pursued a number of host species, cluded all but one two-way interaction: [SH] [SG] the ages and sexesshowing considerable varia- [HG] [HA] [GA] (Table 3C). Female and male tion in proficiency, selection of host species,and Great Frigatebirds showed dramatic differences whether to chaseas part of a group or solitarily in the speciesthey chose to parasitize (Fig. 2), (Tables 1,2). In general sub-adults were lesssuc- females preferring the larger boobies and trop- cessfulthan adults and chasedmore often as part icbirds while males selected terns and noddies 990 JAMES D. GILARDI

TABLE 3. Analysis of log-linear models of Great all frigatebirds and 30-50% of the breeding birds Frigatebirdkleptoparasitism on Midway Atoll. [S] = on Midway. While the marked breeders were success, [H] = host species, [G] = group vs. solitary, frequently observed, they were never seen in [A] = age or adult gender. kleptoparasitic events.

A. GBNERAL MODEL DISCUSSION Partial i3%wciation El&t df X1 P These results provide insight into the types of

3 25.16 0.0000 decisions made by the various agesand sexesof 1 47.84 0.0000 Great Frigatebirds participating in kleptoparas- [SAI 2 14.23 0.0008 itic behavior. Independent of the age and sex of [HGI 14.55 0.0022 the frigatebirds, the successof kleptoparasitic events was affectedby both the host speciespur- tz; :2 121.6925.73 0.0000 suedand whether the frigatebird chasedsolitarily 2 4.637.04 0.07050.5920 or with other conspecifics.In general group chas- [sGAj 2 2.21 0.3318 es were more successfulthan solitary ones and lHGA1 6 6.05 0.4176 smaller hostswere more easily kleptoparasitized ~SHGA] 6 5.31 0.5041 than larger ones. The frequency of the less suc- forward selection [SH] [SG] [SA] [HG] [HA] [GA] cessfulchase events (i.e., solitary attempts or large backward deletion [SHG] [SA] [HA] [GA] hosts) may be maintained by trade-offs in the B. AGE MODEL benefits of a given chase type. That is, chasing Partial association

Effect df X2 P as part of a group yields a partial reward on av- erage, and second, higher probability of success 3 25.39 0.0000 on smaller hosts may be countered with larger 1 52.66 0.0000 PAI 1 12.86 0.0003 rewards from large hosts. As both host avail- ability and potential competitors/conspiratorsare t::i 3 5.898.72 0.11690.0332 spatially and temporally variable, the most ef- [GA1 : 11.22 0.0008 fective pursuit strategy is likely to vary accord- :FIz; 3 7.371.08 0.06090.7828 ingly. tSGA1 : 0.02 0.8949 Adults were more often successfulthan sub- [HGAI 2.29 0.5137 adults in both group and solitary chase events, [SHGA] 3 2.01 0.5702 supporting the hypothesisof age-relatedforaging model selected by both methods: [SH] [SG] [SA] [HA] proficiency. Given the delayed breeding and an [GA1 unpredictable food supply in this species(Nelson C. GENDER MODBL 1975), this age difference is predicted by both Partial association theoretical and empirical work (Recher and EffWl df P X ’ Recher 1969, Orians 1969, Morrison et al. 1978, 3 20.29 0.000 1 Gochfeld and Burger 198 1, Carroll and Cramer 1 41.62 0.0000 1985, Hesp and Barnard 1989, Tuljapurkar PAI : 1.01 0.3160 1990). It is surprisingthat other studiesof frigate- :::! 3 111.5410.31 0.01610.0000 kleptoparasitism have not found such dif- PAI : 13.20 0.0003 ferences(Gochfeld and Burger 198 1, Gibbs and Gibbs 1987, Osomo et al. 1992), and it remains [%Z! 3 7.723.97 0.05210.2653 to be seenwhether more extensive work on Mag- ISGAl : 3.29 0.0698 [HGAI 4.94 0.1765 nificent Frigatebirds will concur with the present [SHGA] 3 2.04 0.5644 study. p;i;l selected by both methods: [SH] [SG] [HG] [HA] The age x group interaction suggeststhat by chasing more often in groups, sub-adults may compensate for their inferior foraging ability. Many chasesare of sufficient duration that the more frequently. Also, males pursued hosts sol- sub-adults might assessthe likelihood of success, itarily more often than females (Table 2). and then join the event or not. Brown Noddies By the end of the 1990 season, individually often taggedalong on such longer chasesas ‘sec- marked birds comprised roughly five percent of ondary kleptoparasites’ with surprisingly fre- GREAT FRIGATEBIRD KLEPTOPARASITISM 991

90

80 fgj SOLITARY 70 GROUP 60

50

40

30

20

10

0 i Brown Sooty Tern Red-tailed ’ Red-footed Noddy Tropicbird Booby

HOST SPECIES FIGURE 1. The effects of host speciesand group size on the successof kleptoparasitic attempts by Great Frigatebirds at Midway Atoll. quent success.Further studies on the age varia- smaller hostsand the larger females selectedcor- tion in the sequence of frigatebirds joining the respondingly larger hostsis consistent with other group might shed light on this question. work on kleptoparasitism (Kushlan 1978, Ter- The striking sex-differencein host choice sug- shey and Breese 1990, but see Dully 1980). It is gestsa correlation between the size of the fiigate- tempting to consider this trend in host size se- bird and the size of the preferred host species. lection as a contributing factor to the reversed- The fact that the smaller males selected the size sexual dimorphism in Fregatidae, the origin

100 18

90 2 80 ?Z & 70 .s 60 13 cd 50 g % 40 BZ 30

@ 20 10

0 I I -r 1 Brown Sooty Tern Red-tailed Red-footed Noddy Tropicbird Booby HOST SPECIES FIGURE 2. Sex-differencesin hostchoice by adult Great Frigatebirdsat Midway Atoll. 992 JAMES D. GILARDI of which remains an open question (Pennycuick though the loss of an occasional meal seemsun- 1983, Schreiber and Schreiber 1988, Gilardi likely to reduce reproductive success,it has been 1992). Since frigatebirds interact with hosts of suggestedthat kleptoparasitism by fiigatebirds is varying body-sizes both at sea and near the col- responsiblefor the noctumality of Swallow-tailed ony, the sexescould potentially partition the host (Creugrusfurcutus, Hailman 1964, Snow community, each specializing on a different size and Snow 1968). Indeed, Sooty Terns and Brown range thereby reducing intrapair competition. Noddies which are preferred hosts on Midway However, the colonial nature of both fiigatebirds also feed their chicks predominantly at night and their victims undermine the benefits of this (Brown 197 5). Although suggestive,this pattern type of niche partitioning which has been ap- would require extensive investigation before plied, albeit inconclusively, to terrestrial raptors drawing conclusions on causal links. Finally, I which are generally territorial (cf. Andersson and encountered numerous Sooty Terns and a trop- Norberg 198 1). icbird with broken or severedwings which could Rather than causingthe size-dimorphism, sex- only have been caused by frigatebirds. Clearly, differencesin host choice may themselves be the kleptoparasite-induced mortality would have result of a dimorphism resulting from other caus- major fitnessconsequences for hosts,particularly es. Possibly more important to the evolution of in tropical where adult survival is an dimorphism are the unusual sex-differences in important life history character. chick provisioning shown by Great and Magnif- icent Frigatebirds and that males may be annual ACKNOWLEDGMENTS breeders and females biennial (Diamond 1972, I thank the Northwestern Hawaiian Islands N.W.R. pers. observ.). for the opportunity to conduct research on Midway; K. McDermond, C. Roland, D. Williamson and D. Hu Finally, the relationship between host size were particularly helpful. I also thank J. R. Gilardi and preference and the size of the frigatebird is not K.V.K. Gilardi for assistanceand patience in the field perfect.Despite the similarity in massof the Sooty and J. Dykes for extensive statistical advice. N. War- Terns and Brown Noddies, females frequently neck, S. Carroll, K. Gilardi, and two anonymous re- chasedthe former but never the latter. This may viewers provided extensive constructivecomments on previous drafts. Finally, I am indebted to W. B. Tyler result from differencesin wing-loading asthe terns for the introduction to Midway and for help in the soar almost exclusively whereas noddies rarely early stagesof this study. do so (pers.obsver.). As wing-loading affectsboth agility and dive velocity (Andersson and Norberg LITERATURE CITED 198 l), analysis of this character in both frigate- AGRESTI,A. 1990. Categorical data analysis. John birds and their hostsmay yield additional under- Wiley and Sons, New York. standing of the observed sex differences in host ANDERSSON,M., AND R. A. NORBERG. 1981. Evo- choice. lution of reversedsexual size dimorphism and role Observations of marked individuals suggest partitioning among predatory birds, with a size that kleptoparasitism near the colony is not an scalingof flight performance.Biol. J. Linnean Sot., Lond. 15:105-130. important source of nutrition for breeding fiig- BR~,H. J. ANDC. J., BARN-. 1979. Klep- atebirds at Midway Atoll. Although frigatebird toparastttsmin birds. Anim. Behav. 27:487-5 14. kleptoparasitism behavior is not often observed BROWN,W. Y. 1975. Parental feeding of youngSooty at sea (pers. observ.), these data do not address Terns (Sterna.fuscata)and Brown Noddies ( stolid& in . J. Anim. Ecol. 44~7311742. its importance away from the colony. Also, klep- CARROLLS. P.. AND K. L. CRAMER. 1985. Aee dif- toparasitism may be important to a small num- ferences in kleptoparasitism by Laughing-Gulls ber of specialist individuals, breeding or other- (Lmus atricilla) on Brown (Pelicanus oc- wise, which this study might very well have cidentalis).Anim. Behav. 33:201-205. missed. In fact, Vickery and Brooke (1994) sug- D~OND, A. W. 1972. Sexualdimorphism inbreed- ing cycles and unequal sex ratio-in Magnificent gest that kleptoparasitism in this speciescould Frieatebirds. Ibis 114:395-398. theoretically account for as much as 40% of the DIAMO&, A. W. 1975. Biology and behaviour of daily energy expenditure of some individuals. frigatebirdsFregata spp. on atoll. Ibis 117: From the hosts’ perspective however, klepto- 302-323. DIXON, W. J. 1990. BMDP statisticalsoftware man- parasitism may cause the loss of a prey load, ual. Univ. of California Press, Berkeley. changesin diurnal behavior patterns (cf. Emms DUFFY, D. C. 1980. Patterns of piracy by Peruvian and Verbeek 1991) and even fatal injuries. Al- seabirds:a depth hypothesis. Ibis 122:521-525. GREAT FRIGATEBIRD KLEPTOPARASITISM 993

EMMS,S. K., AND N.A.M. VERBEEK. 1991. Klepto- Brown (Pelicanusoccidentalis). Anim. Be- parasite avoidance tactics of pigeon guillemots, hav. 17:316-319. Cepphw columba. Anim. Behav. 41:907-909. OSORNO,J. L., R. Toaans, AND C. MACIAS-GARCIA. Fumnsss,R. W. 1987. Kleptoparasitism in seabirds, 1992. Kleptoparasitic behavior of the Magnifi- p. 7 l-100. In Croxall, J. P. [ed.], Seabirds:feeding cent Frigatebird: sex bias and success.Condor 94: ecologyand role in marine ecosystems.Cambridge 692-698. University Press, Cambridge. PENNycutc~,C. J. 1983. Thermal soaringcompared GIBBS,H. L., ANIJJ. P. GIBBS. 1987. Prey robbery by in three dissimilar tropical bird species,Fregata non-breeding Magnificent Frigatebirds (Fregata magnijicens,Pelicanus-occidental& and ?orabps magnijcens). Wilson Bull. 99:101-104. - stratus. J. Exp. Biol. 1021307-325. GILARDI.J. D. 1992. Sex-snecific foraaina distribu- Rncm~, H. F., AND J. A. RECHER. 1969. Compara- tions of Brown Boobiesin the easterntropical Pa- tive foraging efficiency in adult and immature cific. Colonial Waterbirds 15:148-l 5 1. Little Blue . Anim. Behav. 17:320-322. GOCHFELD,M., AND J. BURGER.198 1. Age-related SCHREIBER, E. A., ANDR. W. SCHREIBER. 1988. Great differencesin piracy of frigatebirds from laughing- _ fiigatebird size dimorphism on two central Pacific gulls. Condor83:79-82. - atolls. Condor 90:90-99. HAILMAN.J. P. 1964. The GalaoaeosSwallow-tailed SNOW,B. K., AND D. W. SNOW. 1968. Behavior of is nocturnal. Wilson Bull.-76:347-254. the Swallow-tailed Gull of the Galapagos.Condor HARRISON,C. S., T. S. HIDA, AND M. P. SEKI. 1983. 70~252-264. Hawaiian seabirdfeeding ecology.Wildl. Monogr. TERSHEY,B. R., ANDBREESE, D. 1990. The influence 85:1-71. of sexualdimorphism on kleptoparasitismof Blue- -SON, P. 1985. Seabirds,an identification guide. footed Boobies by Brown Boobies. Can. J. Zool. Houghton Mifflin, Boston. 68:197-199. Hasp, L. S., AND C. J. BARNARD. 1989. Gulls and TUtJAPUlUCAa,S. 1990. Delayed reproduction and plovers: age-related differences in kleptoparasi- fitnessin variable environments. Proc. Natl. Acad. tism among Black-headed Gulls (Larus ridibun- Sci. USA 87:1139-1143. dus). Behav. Ecol. Sociobiol. 24: 197-304. VERNER,J. 1965. Flight behavior of the Red-footed KUSHLAN,J. A. 1978. Nonrigorous foraging by rob- Booby. Wilson Bull. 77:229-234. bing egrets.Ecology 59:649-653. VI-Y. J. A.. AND M. DE L. BROOKE.1994. The MORRISON,M. L., R. D. SLACK,AND E. SHANLEY,JR. kleptoparasitic interactions between Great Frig- 1978. Age and foraging ability relationships of atebirds and Masked Boobies on Henderson Is- OlivaceousCormorants. Wilson Bull. 90:41@22. land. south Pacific. Condor 96:331-340. NELSON,J. B. 1975. The breeding biology of frigate- WEA&, W. W., AND K. A. SULLIVAN. 1989. Ju- birds-a comparative review. Living Bird 14:113- venile foragingproficiency, parental effort and avi- 155. an reproductive success.Ecol. Monogr. 59:223- ORIANS,G. H. 1969. Age and hunting successin the 246.