Seasonal Decline in Clutch Size of Te Barnacle Goose in Svalbard

The Condor98:42-$7 0 The CooperOrnithological Society 1996

SEASONAL DECLINE IN CLUTCH SIZE OF THE BARNACLE GOOSE IN SVALBARD ’

LAILA DALHAUG, INGUNN M. TOMBRE AND KJELL EINAR ERIKSTAD Foundationfor Nature Researchand Cultural Heritage Research,Department for Arctic Ecology, c/o Tromse Museum, N-9006 Troms0, Norway and Universityof Tromso, Dept of Ecology/Zoology,Institute for Biology and Geology, Universityof Troms0, N-9037 Tromsp, Norway

Abstract. We examinedseasonal decline in clutchsize in the higharctic Barnacle Goose Branta leucopsisin Svalbard.Females that arrivedfirst at the breedingground nested first, had the longestprelaying period, produced the largestclutches and also had the longest incubationperiod. Nevertheless, their brood sizeat hatchingwas larger than late nesters. Theseresults do not supportthe nutrient reallocation hypothesis which suggests that seasonal declinein clutchsize in arcticnesting geese results from the mobilizationof nutrientreserves duringthe periodbefore egg laying. Instead, we proposethat geesewhich arrive late at the breedingarea spendrelatively less of their reserveson eggs.They do this presumablyto achievesynchronous hatching with early nestingbirds and/or to preventa late breeding seasonwhich may reducegosling survival, adult survivaland/or futurefecundity. Kev words: Barnacleneese;Branta leuconsis; clutch size; breeding date;incubation period; bodyWcondition. -

INTRODUCTION resulting from degeneration of follicles) in At- When arctic-nesting geese depart from staging lantic Brent Geese Branta bernicla was more areas in spring, nesting conditions may still be common for individuals with small clutchesand unfavorable on their breeding grounds and may for those which failed to breed compared to suc- also vary considerably among years (Barry 1962, cessful females which produced large clutches Ely and Raveling 1984, Petersen 1992, Prop and (Barry 1962). Barry (1962) suggestedthat follic- de Vries 1993). The timing of breeding may affect ular atresia was induced by prolongedbad weath- individual fitness, and a number of proximate er which covered the nesting areas with snow. factorshave been proposedto explain the timing Consequently, resorption of nutrients for main- of breeding in geese; weather conditions (e.g., tenance while waiting for nesting conditions to Barry 1962, Mickelson 1975), age, physical con- be favorable becomes necessary. dition of the laying female and available food The argument for small clutches in late years resources (reviewed by Johnson et al. 1992). may also apply to the decline in clutch sizeswith- When spring thaw is late, the average clutch size in seasons(Ryder 1970). According to the NRH, in geese is reduced (Ryder 1970). Clutch sizes it is assumed that all geese arrive at the same have been found to vary within, as well as be- time and in similar body condition. Since bad tween, seasonsand a negative relationship be- weather and snow cover may induce competition tween nesting date and clutch size has been re- for available nest sites, some females delay egg corded for all speciesstudied in detail (reviewed laying longer than others. During this delay, the by Rohwer 1992). Several hypotheseshave been females utilize stored nutrients which could oth- put forward to explain the seasonal decline in erwise have been used for egg production (Ha- clutch size in geese(see Sjijberg 1994 for a recent mann et al. 1986). Therefore, according to the review). The most acceptedhypothesis for arctic- NRH, females laying late in the season have a nesting geeseis the nutrient reallocation hypoth- longer prelaying period and produce smaller esis(NRH) (Barry 1962). This hypothesisis based clutches compared to early layers. on the observation that follicular atresia (vesicles Several ultimate hypotheses have also been proposed to explain seasonal decline in clutch size. The growth rate and survival of young, and LReceived 10 March 1995. Accepted20 October thereby recruitment of young to the population, 1995. decreaseas the breeding seasonproceeds (Sedin-

1421 SEASONAL DECLINE IN CLUTCH SIZE 43

ger and Raveling 1986, Larsson and Forslund date of egg laying (determined by extrapolation 199 1, Lindholm et al. 1994). Time could be saved using the number of eggs found in a nest and by producing smaller clutcheslate in the season, assuming that one eggis laid per day), length of the cost-of-delay hypothesis (Drent and Daan the prelaying period (number of days between 1980). The balance between current reproduc- arrival and egg laying), clutch size, length of the tive effort and future breeding could also change incubation period (defined as the number of days as the seasonproceeds; for example, the parents’ between the last egglaid and the first egghatched) reproductive effort should be lower late in the and the female body condition at hatching. The seasonsince late breeding is costly for the parents females were caught at the nest at the end of (Toft et al. 1986, reviewed in Rohwer 1992). incubation (one to three days before hatching) So far, there are few critical studies of the sea- using a noose on the end of a fishing rod. The sonal decline in clutch size in arctic-nestinggeese. geesewere weighed and measurements of skull No data document the relationship between the (head including bill) and wing length were taken. date of arrival, length of the prelaying period and Then body size was calculated using a principal clutch size, which would make it possible to ex- component analysis (PCA). The first principal amine the NRH in any detail (for review see component (PCl) explained 68.8% of the stan- Rohwer 1992). dardized variance in the two characters. The In this study we examined the seasonaldecline female’s body condition was defined as the re- in clutch size in Barnacle GeeseBranta leucopsis siduals from a body mass vs. PC1 regression. breeding in Svalbard. The main purpose of the The total explained variance in the female body study was to describe the relationship between mass due to size was 35% (r* = 0.35, df = 16, P date of arrival at the breeding area, the length of = 0.001). the prelaying period and their effects on clutch The statistical analyseswere carried out using size. We also present data on the length of the the SAS statistical package(SAS Institute 1990). incubation period, egg predation during incu- All values are given as means & SE. The signif- bation and female body condition at hatching. icance level is set at P < 0.05. We discuss these data in relation to the NRH and the ultimate hypothesesdescribed to explain seasonaldecline in clutch size for Barnacle Geese. RESULTS We also outline an alternative explanation for DATE OF ARRIVAL AND START OF how female geeseallocate their body reservesto EGG LAYING eggsand incubation in relation to the time of the There was a considerablevariation in arrival dates seasonwhen they start nesting. among females to the breeding ground. Early arriving females started egglaying first; the first egg METHOD AND STUDY AREA laying date was 4 June (Fig. la). These females The study was carried out in 1993 in a colony also had the longest prelaying period (Fig. lb). near Ny-Alesund, in the Kongsfjord area on the Females arriving at the breeding ground when NW coast of Svalbard. The colony (ca. 240 other females already had startedegg laying wait- breeding pairs) studied was on an island (Stor- ed approximately three days before they laid their holmen 78”55’N, 12”15’E) of 30 ha. Almost 70% first egg (Fig. lb). For these females the laying of the adults in the study area have color rings date was determined by the date of arrival. with individual letter codes, and some are of known breeding age. CLUTCH SIZE AND LAYING DATE Birds were identified daily as they arrived in Clutch size decreased significantly with laying the Kongsfjorden area. The colony was visited date (Fig. 2). Egg loss to predators (based on loss every day during the egg laying and hatching of single eggs) was very low (7.0%, n = 397). periods. Birds which were observed for the first Accordingly, mean brood size at hatching (3.32 time during egg laying were excluded from the + 0.08, it = 110) was only slightly lessthan mean prelaying period analysis. Females laying more clutch size produced (3.56 f 0.06, IZ = 123), but than once (renesters), and all nests containing differed significantly (t = 2.39, df = 231, P < fewer than three eggs, were excluded from all 0.01). Therefore, females laying large clutches analyses. The following data were collected for early in the seasonare also those which produce each pair: date of arrival to the breeding area, the largest number of goslingsat hatching. 44 L. DALHAUG, I. M. TOMBRE AND K. E. ERIKSTAD

A early nesters (Fig. 3). There was no correlation 15 between the length of the incubation period and r . clutch size (r* = 0.005, df = 68, P = 0.6), and there was a positive trend for late-laying females . to be in slightly better body condition at hatching l (9 = 0.17, df = 16, P = 0.11). / . . . DISCUSSION The Barnacle Geese which arrived first at the breeding ground startedegg laying first. They had the longestprelaying period, but produced larger clutches than late arrivers. Females which pro-

I,,,, I I I I , duced the largestclutches early in the seasonalso 29 30 1 3 6 7 9 11 13 B DATE OF ARRIVAL (May 6 June) had a longer incubation period. These, however, hatched a larger number of goslings. These results do not support the NRH. According to this hypothesis,females with a long prelaying period should use more body reservesfor maintenance . while waiting to nest and therefore have less reserves to spend on egg production than those with a shorter prelaying period (Barry 1962, Ry- der 1970, Rohwer 1992). In this study, however, . females with the longest prelaying period pro- .%. duced larger clutches than females with a short . prelaying period. Late arriving females waited approximately 29 39 1 3 5 7 9 11 13 DATE OF ARRIVAL (May 6 June) three days before they laid their first egg, pre- FIGURE 1. The relationship between A) date of egg sumably a minimum of time needed before egg laying and the date of arrival at the breeding grounds, laying could start. Closely related speciesto the and B) length of the prelaying period and the date of Barnacle Goose have a rapid follicle growth arrival at the breeding grounds in a Barnacle Goose (RFG) of approximately lo-13 days (reviewed populationin Kongsfjorden,Svalbard (a. Early arriving geese:R2=0.001,df= 14,P=O.91,y=8.30-0.04x, in Alisauskas and Ankney 1992). Therefore, the Late arriving geese:R2 = 0.70, df = 9, P = 0.003, y = Barnacle Geese in Ny-Alesund have probably 0.79x - 19.63, b. Early arriving geese:RZ = 0, 55, df already started the RFG before they arrive at = 14, P = 0.002, y = 46.47x - 1.25, Late arriving their breeding grounds since most females have geese,R2=0.14,df=9,P= 0.29, y= 11.38 -0.21x. a prelaying period shorter than 12 days. A number of studieson geesehave shown that LAYING DATE AND CLUTCH SIZE IN young females producesmaller clutchesthan old- RELATION TO FEMALE AGE 6 . There was no indication from this study area that the age of the breeding females had any signifi- I cant effect on laying date and clutch size. First- 5t l * l breeders (2 years old) had the same mean laying date (9.69 * 0.8 1 in June) as females 5 years old or more (9.41 2 0.57 in June) (t = 0.29, df = 28, P = 0.77). Furthermore, there were no dif- ferencesin the mean clutch sizes of the two age groups (young geese: 3.23 f 0.25, n = 13, old geese:3.64 f 0.20, n = 17, t = -1.26, df = 28, I P = 0.21). 4 9 9 10 12 14 15 CLUTCH SIZE AND INCUBATION PERIOD DATE OF EGG LAVING (June) FIGURE 2. Clutch size in relation to the date of first In addition to smaller clutches, late nesting fe- egglaying(y=4.49-0.10x,R2=0.17,df=60,P= males also had a shorter incubation period than 0.001). SEASONAL DECLINE IN CLUTCH SIZE 45

also Barnacle Geese from Svalbard (Prop and de Vries 1993) have shown that early nesters are f 28 observed with more young than late nesters in i their wintering area. It is obvious from this study that the seasonal decline in clutch size in this 1:: j+-+& Barancle Goose colony is not consistent with the NRH as schematically shown in Fig. 4a. Females $22$- . . . do not produce small clutches late in the season due to longer prelaying periods. Fig. 4b sum- marizes the results from the present study where geesearrive at the breedingarea at different times, 4 6 2 10 12 14 16 and where late arrivers use relatively lessof their DATE OF EGG LAYING (June) reservesin the production of eggs,more on in- FIGURE 3. The length of the incubation period in cubation and have better body condition at relation to date of egglaying in a BarnacleGoose population breeding in Kongsfjorden, Spitsbergen1993 0, hatching. = 25.76 - 0.18x, R2 = 0.21, df = 68, P = 0.0001). There are at least two mutually non-exclusive explanations why geeseallocate their resources differently in relation to the time of breeding er geese(Brakhage 1965, Cooper 1978, Finney season.Firstly, the breeding season in the high and Cooke 1978, Owen 1980, Raveling 198 1). Arctic is short, and those which arrive late may Such an effect has also been documented in Bar- shortentheir seasonby producing smaller clutch- nacle Geese (Forslund and Larsson 1992) where es and use more reserveson incubation to short- individuals of two and three years produced en the incubation period. Using more reserves smaller clutchesand young compared to birds of on incubation also gives females an opportunity four or more years of age. However, there is no to increase nest attentiveness. This may reduce evidence in this study suggestingthat female age the incubation period and decreasethe chances had any impact on clutch size or laying date. that the eggsare taken by predators(Aldrich and In this study, clutch size decreasedwith laying Raveling 1983). Prop and de Vries (1993) found date. However, clutch size was not related to that the probability of hatching eggs increased extra costscaused by an increased prelaying pe- progressively for Svalbard Barnacle Geese who riod. So far we know nothing about how the length started egg laying later in the season. Such an of the prelaying period affectsthe body reserves investment may also synchronize hatching with ofbreeding females. Most likely, females arriving those nesting earlier. To synchronize hatching first must use more reserves for maintenance will be adaptive becausea large number of gos- compared to those arriving later (e.g., Ryder lings may satiate local predators (Findlay and 1970). It is even more likely that geesearriving Cooke 1982). late are in better body condition at the start of Secondly, if the chances that offspring fledge egg laying since they have a short prelaying pe- decreasewith date of egg laying (Cooke et al. riod. Furthermore, a significant positive corre- 1984, Sedinger and Raveling 1986, Prop and de lation has been recorded between female body Vries 1993, Lindholm et al. 1994), females may condition upon arrival and date of arrival for the vary the relative allocation of resourcesto pro- Barnacle Geese in Ny-Alesund (I. M. Tombre et duction of young and their own survival in real., unpubl. data). Accordingly, early nesters lation to the time of the season. Those which which produce the largest clutchesuse relatively arrive early spend relatively more reserves on more of their reserveson eggsthan those arriving eggsat the expenseof body condition at hatching later. This argument is also consistent with the but they presumably produce more fledglings observation that females producing clutcheslate (Prop and de Vries 1993). This suggestsa trade- in the seasonhave similar, or slightly better, body off between investment in young and survival condition at hatching compared to early nesting (e.g., Williams 1966, Chamov and Krebs 1974). females. The length of the incubation period was A number of bird studies (Linden and Moller not related to clutch size, suggestingno differ- 1989) indicate that there exists a negative rela- encesin costsof incubating clutches of different tionship between body mass at the end of the sizes. breeding season and the chance to survive and Studiesof Snow Geese (Cooke et al. 1984) and produce in the future. This also has been re- 46 L. DALHAUG, I. M. TOMBRE ANDK. E. ERIKSTAD

egg-laying date and clutch size, a pattern found in several arctic nesting geese(Rohwer 1992).

ACKNOWLEDGMENTS We are gratefulto G. W. Gabrielsen,K.-B. Strann,I. Alsos,M. Loonenand H. Skarstjordfor valuableas- sistancein the field. We would also like to thank J. Black, Wetland and Wildfowl Trust, for giving us permission to participate in, and use information from, their ringing program on Svalbard barnacle geese.M. Loonen kindly gave us permission to use age data on ringed geese.I. Byrkjedal gave valuable comments on an earlier draft on this paper, R. Barrett correctedthe English,and two anonymousreferees gave constructive comments. We thank the staff on the researchstation of The Norwegian Polar Institute in Ny-Alesund, and also the Governor of Svalbard for permission to work in the nature reservesin Kongsljorden. This study is a part of a terrestrial ecologicalprogram in Svalbard financed by The ResearchCouncil of Norway. Finan- cial supportwas also received from the Foundation for Nature Researchand Cultural Heritage Research. FIGURE 4. a) A graphicalmodel of thenutrient real- locationhypothesis for seasonaldecline in clutchsize (redrawnfrom Rohwer 1992).A femalethat delayed LITERATURE CITED nestingmight lay only 4 eggsbecause stored nutrients havebeen used for maintenance.In the simplestform ALDRICH,T. W., ANDD. G. RAVELING. 1983. Effects of this model,all individualsarrive togetherand have of experience and body weight on incubation be- the samelevel of nutrientreserves at arrival,at the end haviour of Canada Geese. Auk 100:670-679. of layingand at hatching.b) A graphicalmodel based ALISAUSKAS,R. T., AND C. D. ANKNEY. 1992. The on the resultsfrom this study.Geese with the longest cost of egg laying and its relationship to nutrient delaylay the largestclutches, are in poorerbody con- reserves& waterfowl, p. 30-6 1. In B.D.J. Batt, A. dition after laying and hatching,and have longerin- D. Afton. M. G. Anderson. C. D. Anknev. D. H. cubationperiods. Geese with shortprelaying periods Johnson,‘J.A. Kadlec, and G. L. Krapu [eds.], allocatefewer reserves for eggs,start incubation with Ecologyand managementof breedingwaterfowl. more reservesand alsohave betterbody conditionat Univ. of MinnesotaPress, Minneapolis and Lon- hatchingcompared to earlynesters. They thereforeal- don. locatemore reserves for incubation,survival and future BARRY,T. W. 1962. Effectof late seasonson Atlantic fecundity. Brantreproduction. J. Wildl. Manage.26: 19-26. BRAKUGE, G. K. 1965. Biologyand behaviourof tub-nestingCanada Geese. J. Anim. Manage.29: 751-771. cordedin long-lived species(Jacobsen et al. 1995). CHARNOV,E. L., AM) J. R. KREBS. 1974. On the The benefit of producing small clutches may be clutch size and fitness. Ibis 116:217-2 19. COOKE.F.. C. F. FINDLAY.AND R. F. ROCKWELL.1984. an adaptation to shortenthe breeding seasonand/ Recruitment and the timing of breeding in Lesser or to synchronize hatching. However, it may also Snow Geese(Chen caerulescens caerulescens). Auk suggesta possible trade-off between investment 101:451-458. in eggsand own survival, related to the time of COOPER,J. A. 1978. The history and breedingbiology of the Canada Geese of Marshy Point, Manitoba. breeding and the chancesthat the young survive Wildl. Monogr. 61. and recruit to the population (Toft et al. 1986, DRENT,R., ANDS. DAAN. 1980. The prudentparent: Hamann and Cooke 1989). energeticadjustments in avian breeding.Ardea 68: Differences in costsand benefits to rear young 225-252. during the breeding seasonmay as well be used ELY, C. R., AND D. R. RAVELMG. 1984. Breeding biology of Pacific White-fronted Geese. J. Wildl. to explain yearly variation in clutch size. In years Manage. 48:823-837. of late snow melt and egglaying, the chancesthat FINDLAY,C. S., AND F. COOKE. 1982. Synchrony in young would survive may decrease.In suchyears the LesserSnow Goose (Ansercaerulescens). Evo- females may profit from spending less reserves lution 36~786-799. F~NNEY,G., ANDF. COOKE. 1978. Reproductive hab- on eggsand more on self-maintenance and sur- its in the Snow Goose:the influenceof female age. vival. Such an allocation of reservesresults in a Condor 80:147-158. negative relationship between yearly variation in FORSLUND,P., AND K. LARSSON.1992. Age-related SEASONAL DECLINE IN CLUTCH SIZE 47

reproductivesuccess in the barnaclegoose. J. Anim. PETERSEN,M. R. 1992. Reproductive ecologyof Em- Ecol. 6 1:195-204. peror Geese:survival ofadult females.Condor 94: HAMANN, J., AND F. COOKE. 1989. Intra-seasonal 398-406. decline in clutch size in lesser snow geese. Oec- PROP,J., ANDJ. DE VRIES. 1993. Impact of snow and ologia 7983-90. food conditions on the reproductive performance HAMANN,J., B. ANDREWS,AND F. COOKE. 1986. The of Barnacle Geese Branta leucopsis.Omis Stand. role offollicular atresiain inter- and intra-seasonal 24:110-121. clutch size variation in LesserSnow Geese (Anser RAVELING, D. G. 1981. Survival experienceand age caerulescenscaerulescens). J. Anim. Ecol. 55:48l- in relation of breeding successof Canada Geese. 489. J. Wildl. Manage. 45:8 17-829. JACOBSEN,K. O., K. E. ERIKSTAD,AND B.-E. SAETHER. ROFF, D. A. 1992. The evolution of life histories. 1995. An experimental study of the costs of re- Chapmann and Hall, New York. production in the kittiwake Rissa tridactyla. Ecol- ROHWER,F. C. 1992. The evolution of reproductive ogy 76~1636-1642. patternsin waterfowl, p. 486-539. In B.D.J. Batt, JOHNSON,D. H., J. D. NICHOLS,AND M. D. SCHWARTZ. A. D. Afton, M. G. Anderson, C. D. Ankney, D. 1992. Population dynamics of breeding water- H. Johnson,J. K. Kadlec, and G. L. Krapu [eds.], fowl, p. 446-485. In B.D.J. Batt, A. D. Afton, M. Ecologyand management of breeding waterfowl. G. Anderson, C. D. Ankney, D. H. Johnson,J. A. Univ. of Minnesota Press, Minneapolis and Lon- Kadlec, and G. L. Krapu [eds.], Ecologyand man- don. agement of breeding waterfowl. Univ. of Minne- RYDER,J. P. 1970. A possiblefactor in the evolution sota Press, Minneapolis and London. of clutch size in Ross’ Goose. Wilson Bull. 82:5- LARSSON,K., ANDFORSLUND, P. 1991. Environmen- 13. tally induced morphological variation in the Bar- SAS INSTITUTE. 1990. SASSTAT user’s guide: sta- nacleGoose, Branta leucopsis.J. Evol. Biol. 4:6 19- tistics, release 6.04. Cary, NC. 636. SEDINGER,J. S., ANDD. G. RAVELING. 1986. Timing L~NDBN,M., AND A. P. MILLER. 1989. Cost of re- of nestingby Canada Geesein relation to the phe- production and covariation of life history traits in nology and availability of their food plants. J. birds. Trends Ecol. Evol. 4:367-37 1. Anim. Ecol. 55:1083-l 102. LINDHOLM,A. G., G. GAUTHIER,AND A. DESROCHERS. SJBBERG,G. 1994. Early breeding leadsto intra-sea- 1994. Effects of hatch date and food supply on sonalclutch size declinein Canada Geese.J. Avian gosling growth in arctic-nesting Greater Snow Biol. 25:112-118. Geese. Condor 96:898-908. Tom, C. A., D. L. TRAUGER,AND H. W. MURDY. MICKELSON,P. G. 1975. Breedingbiology of Cackling 1986. Seasonaldecline in brood sizes of sym- Geese and associatedspecies on the Yukon-Kus- panic waterfowl (has and Aythya, Anatidae) and kokwim Delta, Alaska. Wildl. Monogr. 45. a proposed evolutionary explanation. J. Anim. MINEAU, P., AND F. COOKE. 1979. Territoriality in Ecol. 53:75-92. Snow Geese or the protection of parenthood. WILLIAMS,G. C. 1966. Natural selection,the cost of Wildfowl 30:16-19. reproduction,and a refinement of Lack’s principle. OWEN,M. 1980. Wild geeseof the world. B. T. Bats- Am. Nat. 100:687-690. ford, London.