The Condor 94:383-397 0 The Cooper Ornithological Society 1992

REPRODUCTIVE ECOLOGY OF EMPEROR GEESE: ANNUAL AND INDIVIDUAL VARIATION IN NESTING ’

MARGARET R. PETERSEN Alaska Fish and Wildlife ResearchCenter, U.S. Fish and Wildlife Service, IO1 1 E. Tudor Road, Anchorage,AK 995032 and Department of Wildlife and FisheriesBiology, Universityof California, Davis Campus, Davis, CA 95616

Abstract. I studied the nesting ecologyof adult female Emperor Geese(Chen canagicus) from 1982-1986 on the Yukon-Kuskokwim Delta (YKD), Alaska. I evaluated variation among and within yearsin date of migration, nest initiation date, clutch size, nestingsuccess, and nesting frequency of 90 individually marked females. Geese arrived on their nesting areas at the same relative date of the seasoneach year and initiated nests five days after their arrival on the study area. Individuals initiated nests at the same relative time of the seasoneach year. There was no significantrepeatability for clutch size, nor did mean clutch sizes vary among years. This may be related to geesehaving ready accessto foods during spring migration and a relative short distance (600 km) between migration staging areas and nestinggrounds. When data were pooled, clutch sizesdeclined as the seasonprogressed, however, this decline was not significant within individuals among years. The proportion of adult females that nested annually varied from 38.S52.0%; the probability of nesting was independent of clutch size or nesting successthe previous season. This low nesting frequency may reflect high annual mortality resulting in a high proportion of newly paired birds each year (thus less likely to nest). Emperor Geese remain at food-rich, spring staging areasuntil nest sitesare available, initiate nestsearly, lay large clutches,and frequently forgo nesting. Key words: Reproductiveecology: annual variation; individual variation; nesting;Chen canagicus;Alaska.

INTRODUCTION skimies 1957, MacInnes and Dunn 1988, Gau- Studies of breeding biology of birds frequently thier 1989, Lessells et al. 1989). Some of these deal with average population statistics such as traits are heritable (Cooke 1987, van Noordwijk clutch size or hatching success. Selection, how- and van Balen 1988), although population av- ever, acts on individuals. Understanding varia- erages may change with annual variation in tion in reproductive success within and among weather and other proximate factors. individuals over the lifetime of the individual To better understand how reproductive suc- can lead to a more complete understanding of cess may vary among individuals within a pop- population demography, adaptation, and selec- ulation, I studied the nesting ecology of female tion (e.g., Clutton-Brock 1988, Newton 1989). Emperor Geese (Chen canugicus). The Emperor Studies of waterfowl suggest that some indi- Goose is a maritime species that nests primarily viduals may consistently produce more offspring along the coastal fringe of the Yukon-Kusko- than others (e.g., Raveling 198 1, Owen and Black kwim Delta (YKD), Alaska and winters through- 1989, Bacon and Andersen-Harild 1989). Al- out the Aleutian and Komandorski islands (Ga- though variation inevitably exists within popu- brielson and Lincoln 1959). Emperor Geese are lations, individual birds are often consistent, mi- monogamous and nest either dispersed or semi- grating to nesting areas about the same time each colonially, primarily along elevated shorelines of year (Rees 1989) laying eggs on similar dates, ponds and sloughs (Eisenhauer and Kirkpatrick and laying similar clutch sizes each year (Ko- 1977, Portenko 1981). In this paper I evaluate the variation within and among years in arrival, nest initiation, clutch size, nesting success, and I Received 16 August 1991. Accepted 8 November frequency of nest initiation for individual adult 1991. female Emperor Geese. I also compare this vari- 2Corresponding address. ation with other goose species nesting on the YKD

W31 384 MARGARET R. PETERSEN to gain a more complete understanding of how hatch dates (Eisenhauer 1976, Krechmar and these speciesmay vary in similar environmental Kondratiev 1982). For nestscontaining eggslaid conditions. by more than one female (see below) I excluded obvious parasitic eggsbefore calculating initia- METHODS tion dates. Parasitic eggsare all eggsadded to a nest after incubation begins, extra eggs added STUDY AREA during egg laying (i.e., more than one egg laid/ I conducted this study on a 760-ha area near day), eggslaid outside a nest and then found in Kokechik Bay, YKD, Alaska (63”39’N, the nest, and eggswith viable embryos that were 165”5l’W) from 1982-1986. This area included not completely developedafter most eggshatched the entire study site used by Eisenhauer and and the brood had abandoned the nest. Clutch Kirkpatrick (1977) in 197 l-1973, plus adjacent size includes all but obvious parasitic eggs of area (Petersen 1990). The area contains a mix- Emperor Geese incubated in the nest. ture of lowland, intermediate, and upland tundra Evidence of hatching included pipped and similar to that described by Ely and Raveling hatched eggs,goslings in the nest, and egg shells (1984). Lowland tundra is generally lessthan 0.5 with the inner membrane intact but separated m elevation above mean high tide and domi- from the shell. For eggs that failed to hatch, I nated by grassesand sedges.Upland tundra is determined the type of predator from direct ob- ~0.5 m of elevation and characterized by pros- servation of predation events or from physical trate willows (S&x spp.), dwarf birch (B&&Z evidence at the nest. Evidence of arctic fox (Alo- nana) and Labrador tea (Ledum palustre), and pex lagopus)predation included tooth marks on contains pingos similar to those described by eggs eaten near the nest, scat in the nest, foot Bums (1964). Intermediate tundra contains plants prints to the nest, or remains of an adult killed characteristic of both upland and lowland tun- near the nest. Foxes normally removed and dra. The flora, fauna and physical features are cached the entire clutch (Stickney 1989, pers. described by Holmes and Black (1973), Eisen- observ.). Evidence ofavian predation (gulls [Lar- hauer and Kirkpatrick (1977) and Jackson(198 1). us spp.] and jaegers [Stercorariusspp.]) included Nesting Emperor Geese use similar habitats on eggswith holes pecked in them or large fragments the northern Chukotka Poluostrov, Siberia with peck marks. Gulls and jaegers ate eggsin (Kistchinski 197 1, Portenko 198 1) and else- the nest or removed a single egg if disturbed by where on the YKD (Spencer et al. 195 1). the goose(Strang 1976, pers. observ.).

NESTS MARKED GEESE I determined clutch sizes, nesting success(pro- I marked 90 nesting female geesewith individ- portion of nests in which one or more eggs ually-coded yellow neck collars and standard hatched),and hatching success(proportion of eggs USFWS leg bands. Geese were trapped on their that hatched in successfulnests) of 746 Emperor nests when hatch began (68 birds) or after 10 Goose nests. Clutch size data were also collected days of incubation (22 birds) using modified for Greater White-fronted Goose (Anser albi- Weller (Weller 1957) traps (85 birds) or by hand from), Cackling Canada Goose (Brunta cana- (3, plus 2 with a dip net). Daily observations of densisminima), and Black Brant (B. bernicla ni- geesefrom blinds allowed me to determine ar- gricans) nests found on the study area. Nest rival dates of these marked females during sub- searchingmethods were similar to those used by sequent springs. Nesting frequency is defined as Raveling (1978) and Ely and Raveling (1984). I the number of known nesting attempts of geese determined some nest initiation dates by direct returning to the nesting area as determined by observations of birds building nests and laying direct observations of marked geeseduring egg eggs.For nests found during egg laying, I deter- laying or incubation. mined initiation dates by backdating from the date the nest was found, using an egg-laying rate WEATHER AND SNOW CONDITIONS of 1.2 days/egg(Eisenhauer 1976, Krechmar and Temperatures on the wintering area and spring Kondratiev 1982). For nests found after incu- staging areas are from National Oceanographic bation began, I assumedthe same egg-laying rate and Atmospheric Administration monthly sum- and a 24-day incubation period to backdate from mary sheetsof local climatological data. Consis- EMPEROR GOOSE NESTING ECOLOGY 385 tent yearly winter weather data are available only of the first migrants and continued to appear well from Cold Bay, Alaska, a location on the north- into the nesting period (Fig. 1). Marked geese easterly portion of the primary winter range of arrived on similar relative days of the migration Emperor Geese (Gabrielson and Lincoln 1959, period in different years (Table 2). Based on the Dement’ev and Gladkov 1952). Data on early timing of snow melt, I considered 1982, 1985 spring weather are from Cold Bay, and data on and 1986 to be late seasonsand 1983 and 1984 weather during spring migration are from King to be early seasons(Petersen 1990). There was a Salmon and Port Heiden, Alaska. Port Heiden tendency for marked geeseto arrive more syn- is a major spring staging area (Petersen and Gill chronously in late seasonsthan in early seasons 1982) approximately 650 km SE of Kokechik (Fig. 1). Conditions on the nesting grounds were Bay; however, long term weather data for the not indicative of conditions the preceding winter spring staging area are available only from King on the wintering area (Table 3). General condi- Salmon. 250 km NE of Port Heiden. tions on the spring staging areas, however, were similar to conditions in spring on the nesting STATISTICAL ANALYSIS areas. I expressed annual variation in arrival, clutch initiation, and hatch dates in relation to the first NEST INITIATION date of the event in a given year. I determined Median dates of nest initiation varied from 20 repeatability estimates of arrival date, nest ini- May to 3 June (Table l), and only in late years tiation date, clutch size, and hatch date following (1982, 1985 and 1986) coincided with snow melt Lessellsand Boag (1987) and standard errors of and drainage of water from nesting areas (Peter- the repeatability estimates following Becker sen 1990). Marked geeseinitiated nests five days (1984). Statisticaltests were conductedwith SPSS” after they were first seen on the study area (Table (SPSS 1986). I used bootstrapping to determine 4). There was no significantdifference in the length the standard errors of coefficientsof variation in of time marked geesewere present before initi- clutch sizes among other goosespecies (50 sam- ating nests either among years or between early ples/species).Preliminary assessmentof the data and late seasons.Nest initiation dates of indi- suggestedthat nest mortality rates were constant viduals were sinnificantlv_ reDeatable_ (Table 2): throughout the nesting period in some years and the timing of nest initiation for an individual was variable in others, and in some years detect- similar each year. ability of destroyed nests was high whereas in CLUTCH SIZES others it was low (due to high predation rates during egg laying). Thus, neither the Apparent Mean clutch sizeswere similar among years (Ta- or the Mayfield procedures were useful for com- ble l), and did not vary significantly between paring nest success(Johnson and Shaffer 1990). early and late seasons(ANOVA F,,d,, = 0.7 1, P Successof nests were therefore calculated using = 0.4 1). Clutch sizes of marked geesewere sim- the Kaplan-Meier procedure (Kaplan and Meier ilar to clutch sizes of unmarked geese(ANOVA 1958) as modified by Pollock et al. (1989) to F ,,47,,= 0.42, P = 0.52). Clutch sizes of marked permit comparisons of data throughout the nest- geese were also similar in both early and late ing period and to include nests found at various seasons(ANOVA F,,,,, = 1.OO, P = 0.32). The stagesof the egg-laying and incubation cycle. I most common clutches were 4-6 eggs (Fig. 2) used the log-rank test modified for staggereden- and the median clutch size was five eggs. try design (Pollock et al. 1989) to compare sur- Repeatability of clutch size was not significant vival rates of nests at five-day intervals among (Table 2); individual geesedid not consistently years. lay the same number of eggseach year. Differ- ences in clutch size for marked geeseaveraged RESULTS lessthan one eggbetween years, although several individuals had wide variations in clutch size. ARRIVAL No individual consistentlylaid large(> 6) or small The first Emperor Geese arrived at the nesting (~4) clutches. area in early to mid-May, with large influxes of There was no consistenttrend for the 18 marked geeseoccurring two to 16 days later (Table 1). individuals from which two or more years data Marked geesearrived within two to seven days are available to lay fewer eggswhen nesting later 386 MARGARET R. PETERSEN

10 - a- 1983,early season 6- . 4 2 ?._L ?i 1984,early season ii8 26

1985,late season

Za 1986,late season 6 4

1 M$ June & Major influx of geese. FIGURE 1. First observations of individually marked geeseand major influxes of unmarked geesein spring. in relation to the median nest initiation date than NESTING FREQUENCY in other years. Only for one individual was this When data from all individuals which survived trend significant over four years (rZ = 0.9259, n to the next year are combined, 5 1.5% of 68 fe- = 4, P = 0.04). Six individuals showed no or males that nested one year nested the next year, only a slight changein clutch size with later nest- while 72.0% of 25 females that failed to initiate ing, one an increase in clutch size, and eight a a nest one year failed to nest the following year decreasein clutch size. Two individuals laid the (Table 5). Nesting frequency was independent of same number of eggsbeginning on the same date clutch size the previous year (Median test x2 = each year for three years and two years, respec- 0.02, df = 1,P = 0.90), although not independent tively. Clutch size of all Emperor Geese declined con- of nesting status the previous year (x2 = 4.07, df tinuously throughout the nesting season(Fig. 3) = 1, P = 0.04) (Table 5). The frequency with at similar (ANCOVA F,,d,, = 0.73, P = 0.57) which females nested ranged from five times in rates each year. Clutch sizes of marked geese five years to once in five years (Table 6). Fifteen exhibited a similar trend at a slightly higher, al- of 45 geese (33.3%) nested each year and, on though not significantly different (ANCOVA F,,d,, average, a female nested 69.8% of the years she = 2.43, P = 0.12), rate of daily decline. was known to be alive. The overall proportion EMPEROR GOOSE NESTING ECOLOGY 387

TABLE 1. Arrival, nesting chronology, and clutch sizes of Emperor Geese.

First Major First nest Median nest Clutch size E&clutch hatched Year arrival influxes initiation initiation (n) f + SE (n) 1 + SE (n)

1982 All geese’ 29 May 3 June (54) 4.9 f 0.2 (74) 4.6 + 0.2* (65) Marked geese 29 May 3 June (34) 5.0 f 0.3 (36) 4.6 + 0.3 (34) 1983 All geese 7 May 23 May 12 May 20 May (105) 5.1 * 0.2 (111) 4.7 5 0.2* (102) Marked geese 13 May 17 May 15 May 21 May (28) 5.0 -+ 0.2 (29) 4.7 * 0.2 (29) 1984 All geese 8 May 13, 22-23 17 May 23 May (73) 4.8 + 0.2 (88) 3.8 k 0.2 (85) May Marked geese 15 May 18 May 19 May 24 May (22) 5.3 -t 0.3 (21) 4.4 2 0.2 (20) 1985 All geese 17 May 19 May 25 May 1 June (110) 4.8 f 0.1 (109) 4.5 + 0.2* (109) Marked geese 19 May 26 May 27 May 31 May (30) 4.9 + 0.2 (30) 4.6 + 0.2 (30) 1986 All geese 5 May 7, 16 May 21 May 28 May (55) 5.0 f 0.2 (72) 4.3 ? 0.2* (56) Marked geese 12 May 23 May 23 May 28 May (16) 4.8 + 0.3 (16) 4.4 t 0.3 (14) Total All geese 4.9 ? O.l* (472) 4.4 * 0.14 (417) Marked geese 5.0? O.l’(l32) 4.6 f O.ls (127)

I Includesmarked and unmarkedgeese nesting on the studyarea. z F,,,, = 0.941, P = 0.44. Clutch size amongyears. ’ F,,,,, = 0.533, P = 0.7 I. Clutch size amongyears. : p,, = 5.016, P = 0.0005. Number of e&q hatchedper successfulnest among years. = 0.522, P = 0.72. Number of eggshatched per successfulnest among years. * S’i9lar clutch sizes;LSD test, P > 0.05.

of geesenesting each year was similar (x2 = 1.87, proportion for adult females on the Kokechik df = 3, P = 0.60), and ranged from 38.5% in Bay study area. 1986 (n = 26) to 52.0% in 1984 (n = 25) with intermediate values for 1983 (44.4% of 18) and NESTING SUCCESS 1985 (50.0% of 24). Marked geese successfully The proportion of nests having one or more eggs nested at least once (the year they were nest hatch varied markedly among years (90.6% in trapped), and geesethat did not hatch eggsin at 1982 to 0.1% in 1986) (Table 7). Nesting success least one year were not sampled. The sample of of marked geese was similar to unmarked geese. marked individuals is probably biased toward In years when few nests were destroyed (1982- older, more successfulfemales. Thus the overall 1983), nests were lost to avian and mammalian nesting frequency of 69.8% based on the average predators in similar proportions (x2 = 0.15, df of individual geesethat were known to be alive = 1, P = 0.70). In contrast, during years when is probably a maximum and 42.6% based on predation rates were high (1984-l 986) the rel- average annual nesting frequency a minimum ative number of nests destroyed by mammals

TABLE 2. Repeatability of arrival date, nest initiation date, clutch size, and hatch date for individual female Emperor Geese.

F ratio df P Repeatability SE

Arrival date 2.46 17,33 0.013 0.341 0.164 Nest initiation date 2.23 18, 34 0.021 0.274 0.144 Clutch size 0.83 19,34 0.660 -0.179 0.142 Hatch date 1.89 16, 30 0.065 0.240 0.175 388 MARGARET R. PETERSEN

TABLE 3. General weather conditions experiencedby Emperor Geeseduring winter, late-winter, springstaging and summer.

Winter Late-winter springstaging Summer Aleutian Islands’ C”;gyY King Salmon Port Heiden’ KokechikBay ’ Dec.-Mar. AY;;~Y A,;~Y M$-$ - 1.8%’ 0.6%

1981-1982 conditions5 Warm Cold Very cold Cold 1982-1983 conditions Warm Very warm Very warm Very warm Very warm 1983-1984 conditions Warm Cold Average Average Average 1984-1985 conditions Very warm Very cold Very cold Cold Very cold 1985-l 986 conditions Average Average Cold Average Cold ’ ’ Aleutian Island weatherbased on Cold Bay reportedmonthly summaries. 2Port Heiden weatherbased on averagemin.-ma. monthly temperatures,no information on deviation of temperaturefrom long term normal available. General weatherconditions based on ditTerencesin temperaturesamong years. ’ ’ Kokecbik Bay weatherbased on averagemin.-max. temperaturesfrom 10 May-10 June taken at the field camp. General weatherconditions basedon differencesin temperaturesamong years. 4 Long teml mean. 5 General weathercondition categories based on deviation from long tam meantemperatures: Very cold < - Z.OT, Cold -2.(P to -0.4’c, Average -0.5” to 0.6=C,Warm 0.7” to Z.O=C,Very warm >2.O=C.

was high while the number destroyed by birds = 4.1 f 0.3, n = 34) was similar among years remained low (Table 8). The arctic fox was the (ANOVA F34.398= 1.5 1, P = 0.08). Smaller primary mammalian predator on the study area clutches were less likely to be successfulthan and two pairs were present each year. larger clutches (x2 = 17.84, df = 4, P = 0.0013) Predators frequently destroyed nests early in when all years were pooled. Small clutches (I 3) the nesting season before females began incu- were combined becauseof small sample sizes,as bating (Fig. 4). Only in 1985 were any nests de- stroyed by predators after the first 10 days of incubation. When few nestswere sampled before incubation began (1982 and 1983), the proba- bility of a nest surviving to hatch appeared to be very high. However, too few nests were visited early in those years to accurately determine pre- incubation lossrates. With the exceptionsof 1984 and 1986, rates of nest loss were similar between years (log rank tests,P > 0.05) (Fig. 4). Predation rates in 1984 were similar only to 1985 (log rank tests, P > 0.05). The predation rate in 1986 was different from all other years (log rank tests, P < 0.005). The mean number of eggsin successful(X = 5.0 * 0.1, n = 398) and unsuccessfulclutches (X

TABLE 4. Length of time between date of first ob- servation and nest initiation for individually marked geese.

Days present YeaI Seasontype n Median Range

1983 Early 4 5 3-6 1984 Early 8 4 o-7 1985 Late 11 6 2-10 1986 Late 4 4 l-8 2 4 6 8 10 Total 27 51,2 O-10 Clutch size laid FIGURE 2. Clutch sizedistribution of Emperor Geese ’ ’ Among years,Krwkal-Wallis x’ = 3.74, df = 3, P = 0.29. 2 Among seasons,Krwkal-Wallis x2 = 1.32, df = 1, P = 0.24. by year. EMPEROR GOOSE NESTING ECOLOGY 389

TABLE 6. Number of nestingattempts of 45 marked . . a. geeseknown alive. . . . . I YCU3 Number of yearsnested observed 5 4 3 2 I 5 1 2 0 1 2 4 3 3 1 2 3 4 3 5 2 7 11

2 b. F 9.422 = 70.505, P < O.OOOl),but not among years o A . (ANGVA F+,22= 2.288, P= 0.06) (Fig. 5). Among successfulnests, the number of eggshatching per nest increased with increasing clutch size. This pattern was consistent for each clutch size in each year (year x clutch size interaction, ANOVA F18 = 1.21, P = 0.06). The most common clutches (4-6 eggs/clutch)frequently lost one or more eggs (x2 = 22.21, df = 5, P = 0.0005) (Table 9). is 3b 4 8 14 lb May June DISCUSSION FIGURE 3. Clutch size as related to date of first egg. NESTING CHRONOLOGY a. Y* = 0.32, n = 366, y = 37.75 + (-0.21)x; b. r2 = 0.36, n = 18 birds, 49 clutches,y = 45.44 + (-0.26)x. Weather influences nesting chronology in several speciesofgeese (Coach 1958, Barry 1962, Hudec and Kux 197 1, Ryder 1972, Newton 1977 and were large clutches (~7). Sample sizes are too citations therein, Raveling and Lumsden 1977), small for statistical comparison of nesting suc- including Emperor Geese. Emperor Geese ar- cess by clutch size by year. However, among rived at later dates and initiated nests at later completed clutches, small clutches were de- dates when freezing temperatures and snow and stroyed in significantly higher proportions than melt-water covered the nesting area (Eisenhauer larger clutches in late years when fox predation and Kirkpatrick 1977, Mickelson 1975, Petersen was severe (1985 and 1986; x2 = 22.01, df = 4, 1990). Delayed nesting seasons,however, did not P = 0.0002) (Fig. 5). result in changes of relative arrival patterns of individual Emperor Geese. Individuals that ar- HATCHING SUCCESS rived before the median arrival date in mild years The mean number of eggshatched per nest var- arrived early in cold years, and birds that arrived ied significantly among clutch sizes (ANOVA after the median arrival date in mild years ar- rived late in cold years. This consistency in ar- rival patterns suggeststhat birds responded sim- TABLE 5. Status of adult female geesein relation to ilarly to the same environmental cues for clutch size and reproductive history the previous year. initiation of migration each year. Weather con- ditions at spring staging areas along the Alaska Statusor clutch Number (%) of geesereturning in year t + I size in year t Nested Not nested Peninsula reflected conditions on the nesting grounds on the YKD and may provide a reliable Not nested 7 (28.0%) 18 (72.0%) cue regarding the availability of nest sites. In Nested 35 (51.5%) 33 (48.5%) other speciesof geese,older birds initiated nests Clutch size earlier in the season than younger, less experi- 53 4 (57.1%) 3 (42.9%) enced geese (Brakhage 1965), and adult geese 4 7 (58.3%) 5 (41.7%) 5 6 (42.9%) 8 (57.1%) initiated nests about the same time (as modified 6 10 (41.7%) 14 (58.3%) by weather) each year (Findlay and Cooke 1982, 7 4 (66.7%) 2 (33.3%) but see MacInnes and Dunn 1988). No trend of 28 2 (100%) 0 (0%) earlier nest initiation dates in subsequentseasons Unknown 2 (66.7%) 1 (33.3%) was apparent for female Emperor Geese trapped 390 MARGARET R. PETERSEN

TABLE 7. Survival of Emperor Goose nests until hatch.

Apparent method Pollwk et;; ;(:;9) method Uncollared geese Collared geese’ n % surv. Var. 95% Cl. n % SUrv.‘.~ n % surv.’

1982 63 90.6 0.18 82.2-99.0 80 96.3 1983 118 78.6 0.23 69.2-87.9 102 94.1 7 100 1984 105 37.6 0.15 29.9-45.3 129 67.4 14 78.6 1985 134 48.5 0.22 39.3-57.7 135 71.8 17 88.2 1986 171 0.1 0.00 0.0-O. 1 151 33.8 10 40.0

I Includes only eesx marked in previous years. * Nests found a Fier egg laying had begun were included in analysis. This is the probability of a nest surviving to hatch. ’ ’ Hatching success calculated using the Apparent method (n hatching/n found) (Johnson and Shaffer 1990). 4 Includes only geese not collared in previous years. at hatch, suggestingthat the femalessampled were late years than in early years (Mickelson 1975, experienced, older females. No data are avail- Raveling 1978, Dau and Mickelson 1979, Ely able, however, from known-aged individuals. and Raveling 1984, pers. observ.). Unlike the other speciesnesting in the area, Emperor Geese CLUTCH SIZE exhibited no significant variation in clutch size Reduced clutch sizesin some years for arctic and between early and late seasonsfrom 1982-l 986. sub-arctic nesting geesehave been attributed to The variation in clutch size among these years insufficient food (or insufficient quality) available in the same area for the other goosespecies was at spring staging areas and resultant poorer con- significantly greater than for Emperor Geese (Z dition of females on their arrival to nesting areas = -2.30, P = 0.01). This suggeststhat for Em- (Newton 1977, Ebbinge et al. 1982, Davies and peror Geese the nutrient reserve levels needed Cooke 1983, Coach et al. 1989). In years with for eggproduction remained high in late seasons delayed snow melt on the nesting areas, reduced during this period. This may be becauseEmperor clutch sizes were attributed to use of energy re- Geese have ready accessto foods during spring servesby females for maintenance during a pro- migration, or because they have a shorter mi- longed pre-nesting period and subsequent re- gration between stagingareas and nestinggrounds duction in reserves available for egg laying than do other speciesof geese. (Ankney and MacInnes 1978, Raveling 1979a). Similar weather conditions at stagingand nest- Seasons severely delayed because of prolonged ing areas could clue Emperor Geese to remain snow and ice melt can result in non-breeding by on stagingareas where abundant food is available significant proportions of arctic goose popula- until nesting areasare accessible.Emperor Geese tions (Barry 1962, Raveling 1978, Cooke et al. feed extensively on blue mussels(Mytilus edulis) 1981, Prop et al. 1984). and the bivalve Macoma balthica in intertidal Spring weather currently influences Emperor regions of lagoons along the north side of the Goose clutch size lessthan for other geese.Black Alaska Peninsula during spring migration (Pe- Brant, Cackling Canada Geese, and White-front- tersen 1983). The delay in migration in late years ed Geese nesting on the YKD laid fewer eggsin would result in geeseremaining in thesefood rich

TABLE 8. Proportion of nestsdestroyed by mammalian and avian predators.Other lossesinclude nestsflooded due to storm tides, deserted nests, and unknown losses.

Type of loss Mammalian predation Avian%py ;lation Other Total nests Total destroyed % W) n % (n)

1982 102 8 37.5% (3) 62.5% (5) 1983 168 13 46.2% (6) 53.8% (7) 1984 141 53 79.2% (42) 15.1% (8) 5.7% (3) 1985 159 43 81.4% (35) 11.6% (5) 7.0% (3) 1986 176 114 89.5% (102) 8.8% (10) 1.8% (2) EMPEROR GOOSE NESTING ECOLOGY 391

Egg laying Incubation Hatch TABLE 9. Clutchsizes in relation to total and partial failure, and number of eggshatched by Emperor Geese in 1982-1986.

% % % total martial Clutch total nest _ nest size n nests failure failure x k SE at hatch ’

1 3 0.6 33.3 0.0 1.0 f o.o* 2 21 4.5 28.6 20.0 1.8 f O.l* 3 52 11.2 19.2 19.0 2.8 + 0.1* I 4 113 24.3 15.9 30.5 3.6 + 0.1 5 104 22.4 8.6 34.4 4.5 * 0.1 1986 : 6 102 22.0 9.8 44.4 5.2 k 0.1 0 2 4 6 a 10 7 55 11.8 12.7 52.1 6.0 + 0.2** Stage of nesting cycle 9 1.9 11.1 25.0 7.4 +- 0.5** FIGURE 4. Survival probabilities of Emperor Goose ; 5 1.0 40.0 33.3 7.7 f 1.3** nests at stagesof the nesting cycle. Distance between 10 1 0.2 0.0 0.0 10.00** each stage is five days. Stages: 1-first nest initiated, *, ** Denotes pain of groups.hat are similar. Scheffe procedure P < 0.05. a-median nest initiation, 3-median initiation of in- ’ ANOVA F,,,, = 65.79, P = 0.0000. cubation, 4 to 7-incubation at 5 day intervals, 8- median hatch, 9 to lo-hatch at 5 day intervals. that reported by Rohwer and Eisenhauer (1989) for Emperor Geese on the same study site in areas until nesting areas became available. This 1973. This seasonal decline is found in many would result in reduced clutch size variability speciesof birds that exhibit variation in clutch between early and late years because geese would size (e.g., Klomp 1970, von Haartman 197 1, Per- not spend prolonged periods on the nesting area rins and McCleery 1989, but see Eldridge and and use energy reserves needed for egg formation Krapu 1988). Toft et al. (1984) and Murphy and incubation. (1986) summarized explanations of reduced Reduction of clutch sizes in late seasons could clutches within a season.Smaller average clutch occur because of local depletion of mussel and sizes later in the seasonhave been attributed to clam beds by foraging Emperor Geese. Mickel- renesting individuals laying smaller clutches lat- son (1975) and Eisenhauerand Kirkpatrick (1977) er in the season, young or less-experiencedin- reported lower clutch sizes in Emperor Geese in dividuals laying smaller clutcheslater in the sea- late seasonswhen populations were much larger son, and depletion of nutrient reserves while (140,000-l 60,000; Eisenhauer and Kirkpatrick waiting to nest. Toft et al. (1984) suggestedthat 1977). During the years of this study the spring different optimum clutch sizes may exist for in- Emperor Goose population declined from dividuals that nest at different times during the 101,000 in 1982 to 42,000 in 1986 (R. King, season.This is supported by studies confirming U.S. Fish and Wildlife Service, pers. comm.). A the repeatability of clutch size and laying date density dependent effect on clutch size may have found in some species (Batt and Prince 1979; occurred during previous studies, but could have van Noordwijk et al. 1980, 1981a, 1981b; Find- been absent from 1982-1986. Emperor Geese lay and Cooke 1982,1983; van Noordwijk 1987; arrive at their nesting area with heavy accumu- Gauthier 1989). lations of body fat (Portenko 198 1). Geese could Although Emperor Geesearrived and initiated accumulate additional energy storesafter arrival nests at the same time each year, they did not on the nesting grounds, thus reduce the density always lay the same size clutch. This low re- dependent effect on clutch size. Some forage peatability of clutch size is consistent with other plants are available during the pre-laying period speciesof geese(Canada Geese [MacInnes and (Raveling 1979b, pers. observ.) and Emperor Dunn 19881 Lesser Snow Geese [Lessells and Geese feed at this time; however adult female fat Boag 19871) and does not support Toft et al. and protein levels did not increase significantly (1984) in relation to clutch size in geese. This after arrival to the nesting area (K. Laing and D. lack of strong repeatability may reflect variation G. Raveling, pers. comm.). in physical ability (i.e., energy reserveson arriv- Clutch sizesof Emperor Geese declined within al) of females to lay their maximum clutch each each seasonfrom1982-1986 at a rate similar to year (Ankney and MacInnes 1978, Raveling 392 MARGARET R. PETERSEN

lo- 1982 - 8. - + 8, + -c 4, - 2 -- l (1) (12) (15) (13) (14) VI (2) (0) (1)

100

- 50 $ Ti 2 100 -;; i!Y 50 cii 4 +- % IT (9) (19) (20) (16) (9) 0) (1) .__25

I” (131 129) 127) 1291 1101 111 12) I-

;]‘ 1988 /---_ 110

Cl)/= (61 (14) (12) (15) (61 (1) I If _ . , I 2 4 8 8 10 Clutch size laid FIGURE 5. Clutchsize (X f SE) at hatchand percentof nestshaving one or more eggshatch by clutchsize laid and year. Samplesize (n) includedall nestswith known fate and clutchsize laid.

1979a, Findlay and Cooke 1983, Hamann et al. 1983). It is commonly assumed that after her 1986, Eldridge and Krapu 1988). first nesting season a female goose nests each successiveyear (Cooper 1978, Cooke and Rock- NESTING FREQUENCY well 1988). This was not the case for Emperor In normal conditions most adult geeseattempt Geese. Only 42.6-69.8% of adult females nested to nest each season (Craighead and Stockstad each year. Nesting frequency was independent of 1964, Brakhage 1965, MacInnes et al. 1974, Prop the previous year’s nest fate, clutch size, nest et al. 1984) although non-nesting occursin some initiation date, and arrival date and the current individuals (MacInneset al. 1974, Cooper 1978). year’s arrival date or timing of habitat avail- In some years a large segment of the population ability. An estimated 21% of females nest par- may not nest because of weather conditions in- asitically (Petersen 199 1). Although laying eggs, fluencing habitat availability (Barry 1962, Prop most parasitically laying females were not re- et al. 1984) or food shortages(Davies and Cooke corded as nesting birds since few could be pos- EMPEROR GOOSE NESTING ECOLOGY 393 itively identified with nests. Becauseparasitic fe- dents, which are often alternate prey of primary males were recorded as non-nesting individuals, nest predators (Pehrsson 1986, Summers 1986, their nesting frequency may be higher than in- Summers and Underhill 1987, Mason 1988, but dicated. seeOwen 1987). At Kokechik Bay in spring 1986, In monogamous birds the death of a mate and when predation rates were high on Emperor subsequent re-pairing may result in non-nesting Goose nests,alternative prey (Microtus spp.) were or failed nesting in the following breeding season. as abundant as in 1985 (R. M. Anthony, pers. Re-pairing in Snow Geese (Chen caerulescens) comm.). High predation apparently occurred in resulted in no significant reduction of reproduc- 1986 becausefoxes did not den successfullythat tive potential (Cooke et al. 1981). However, in spring and had more time available to searchfor Owen et al. (1988), reproductive potential ofBar- and cache gooseeggs (Stickney 1989). nacle Geese (Branta leucopsis)fell following re- Most destruction of Emperor Goose nests oc- pairing, and it is suggestedthat re-pairing soon curred during laying when geesewere seldom at after the previous breeding season reduces the their nests. As with other larger arctic and sub- negative reproductive effectsof re-pairing. Thus, arctic nesting geese (e.g., Snow Geese, Harvey the reproductive potential of geesere-pairing in 1970; Pink-footed Geese [Anser brachyrhyn- early fall should be similar to geesewho remain thus], Inglis 1977; Canada Geese, Raveling and paired into the next breeding season. For Em- Lumsden 1977) incubating Emperor Geese vig- peror Geese, hunting mortality occurs predom- orously defended their nests from foxes and re- inantly in spring during native subsistencehar- duced lossesto avian predators by almost con- vest activities (Klein 1966, Pamplin 1986). This tinuous incubation (Thompson and Raveling spring kill coupled with a high annual adult mor- 1987). Nests are susceptibleto egg lossesduring tality rate from all causes(Petersen 1992) would laying (Harvey 197 1, Inglis 1977). During this result in a large proportion of adult geese that period arctic foxes were most successfulin taking are newly paired in spring and lesslikely to nest. eggsfrom Emperor Goose nests. Geese that pair in spring may be physiologically There was no single “best” clutch size for Em- unprepared to nest (Akessonand Raveling 1984). peror Geese. Most clutches contained 4-6 eggs, Females who have lost their mates may have with a few 13 and 27 egg clutches each year. insufficient energy reservesfor successfulnesting, Clutches of five and six eggsexperienced the least since one role of the male is to allow the female total failure, and clutches of two and three eggs sufficient uninterrupted time to feed (Mc- the least partial failure. Clutches of seven, eight, Landress and Raveling 198 1, Raveling 1988). and nine eggsproduced the most hatched eggs Delayed maturity is believed to be an adap- per successfulclutch, and also had the highest tation, exhibited by many species,that results in variance. Females laying six or more eggs,how- more viable offspring produced over the lifetime ever, survived at higher rates than birds laying of the individual (Bell 1980). A similar argument five or fewer eggs(Petersen 1992). Females that could be made for frequent non-nesting. Non- laid smaller clutches (34 eggs)lost more entire nesting is the apparent strategyused by Emperor clutches, hatched fewer eggs, and survived at Geese when the costs of nesting (higher adult lower rates (Petersen 1992) than geesethat laid mortality) are high and survival of non-nesting larger clutches (5-7 eggs).Over half of the geese individuals is high (Petersen 1992). An indi- that nested in any one year, however, laid five vidual should nest only when it is likely to pro- or fewer eggs.One explanation for this prepon- duce its maximum potential number of surviving derance of small clutches would be that the offspring. Older geesehave larger clutches(Brak- smallest clutches are laid by younger, inexperi- hage 1965, Cooper 1978, Finney and Cooke enced geesenot yet laying to their full potential 1978) higher hatch rates (Brakhage 1965, Al- (Kossack 1950, Finney and Cooke 1978, Rock- drich and Raveling 1983) and higher fledging well et al. 1983, Aldrich and Raveling 1983). rates (Raveling 198 1). More young female Emperor Geese may have nested in the two early seasons(1983 and 1984) NESTING SUCCESS when a shift toward four and five egg clutches Many species of birds suffer high rates of nest was apparent. This disparity between the mean predation. Poor reproduction of several species and optimum clutch size could also reflect dif- of arctic waterfowl with wide geographicdistri- ferences in female condition due to environ- butions is correlated with lows in cycles of ro- mental variability (Rockwell et al. 1987, Ankney 394 MARGARET R. PETERSEN and Afton 1988) and be reflected in among year reducing their preincubation maintenance costs; differences of individual Emperor Geese. they initiate nests early which subsequently re- The largest number of eggslaid, eggshatched, sults in increased gosling size and survival to or young fledged is not necessarilythe most pro- fledging; they lay a large clutch which results in ductive clutch based on return rates and subse- more eggshatching per female; and adult females quent nesting of offspring (van Noordwijk et al. frequently forgo nesting thereby increasing their 1980). Similarly, the most productive clutch one survival. year may not be the most productive clutch the next year (Lack 1966, van Noordwijk et al. 1980). ACKNOWLEDGMENTS The increased survival of adults laying six or This study was funded by the U.S. Fish and Wildlife more eggs coupled with clutch sizes of six or Service, Region 7, Anchorage, Alaska, and was con- more eggs producing the most young suggests ducted on lands belongingto the Chevak, Paimiut, and that, on average, females laying six or more eggs SeaLion corporations. The staff of the Yukon Delta will produce more offspring than other females. National Wildlife Refuge- provided invaluable logistic support throughout the study. I particularly thank K. However, the proportion of those offspring sur- G. Becker. W. I. Butler. Jr.. E. Peltola. Jr.. and M. B. viving to nesting age is unknown. Rearden for mail, supplies’and moral support. P. L. Flint, K. Laing, D. Lake, B. Murran and S. C. Thomp- INDIVIDUAL VARIATION son cheerfully assistedwith field work. I thank T. W. Emperor Geese exhibited both intra- and inter- Arnold, D. W. Derksen, C. R. Ely, R. E. Gill, Jr., D. Keppie, and D. F. Lott for helpful comments on var- individual variation in clutch size. Although the ious versionsof the manuscript. I thank D. G. Raveling population as a whole exhibited a seasonal de- for support throughout this project, helpful comments cline in clutch size (a trait common to other arctic on this manuscript, and for sharinghis insight ofgeese. and temperature waterfowl), individual birds did not exhibit that pattern. Emperor Geese that ar- LITERATURE CITED rived and initiated nestsearly in one year tended AKESSON, T. R., AND D. G. RAVELING. 1984. Endo- to do so in other years, despite wide variation of crine and behavioural correlatesof nestingin Can- weather conditions and timing of the season. ada Geese. Can. J. Zool. 62845-850. Clutch sizes of individuals, however, were not ALDRICH, T. W., AND D. G. RAVELING. 1983. Effects of experience and body weight on incubation be- correlated with nest initiation date. The intra- havior of Canada Geese. Auk 100:670-679. individual variation in clutch size suggeststhat ANKNEY,C. D., AND A. D. AFTON. 1988. Bioener- individuals are not consistently in peak physio- getics of breeding Northern Shovelers: diet, nu- logical condition each year. trient reserves,clutch size, and incubation. Con- dor 90459-472. Individuals that lay eggs late have smaller ANKNEY,C. D., ANDC. D. MACINNES. 1978. Nutrient clutch sizes, on average, than those that lay eggs reservesand reproductive performance of female early. Perhaps individuals nesting early in the Lesser Snow Geese. Auk 95:45947 1. season have a larger potential clutch size than BACON,P. J., ANDP. ANDERSEN-HARILD.1989. Mute individuals that normally nest later. Inter-indi- Swan, p. 363-386. In I. Newton [ed.], Lifetime reproductionin birds. Academic Press,New York. vidual variability in maximum potential clutch, BARRY, T. W. 1962. Effect of late seasonson Atlantic however, is unknown. Small clutch sizes in later Brant reproduction. J. Wild]. Manage. 26: 19-26. initiated nests result in a shorter pre-incubation BAG, B.D.J., AND H. H. PFUNCE. 1979. Laying dates, period such that smaller clutches hatch earlier clutch size and egg weight of captive Mallards. within a season than would large clutches laid Condor 81:3541. BECKER,W. A. 1984. Manual ofquantitativegenetics, by the same individuals. Early goslingscan take 4th ed. Academic Enterprises, Pullman, WA. advantage of the early availability of high quality BELL,G. 1980. The costs of reproduction and their foods (Sedinger and Raveling 1986). These gos- consequences.Am. Nat. 116:45-76. lings should experience increased survival BRAKHAGE,G. K. 1965. Biology and behavior oftub- nesting Canada Geese.J. Wildl. Manage. 29:75 l- through fall migration since larger goslings sur- 771. vive at higher rates than smaller goslings(Owen BURNS,J. J. 1964. Pingos in the Yukon-Kuskokwim and Black 1989). By hatching early, geese also Delta Alaska: their plant successionand use by avoid being unable to leave before freeze-up mink. Arctic 17:203-2 10. CLU~ON-BREAK, T. H. [ED.] 1988. Reproductive (Coach 1958). success.Univ. ChicagoPress, Chicago. Emperor Geese remain at food-rich, spring COOCH, E. G., D. B. LANK, R. F. ROCKWELL, AND F. stagingareas until nest sitesare available thereby COOKE. 1989. Long-term decline in fecundity in EMPEROR GOOSE NESTING ECOLOGY 395

a Snow Goose vovulation: evidence for densitv GABRIELSON,I. N., AND F. C. LINCOLN. 1959. The dependence?J. knim. Ecol. 58:77 l-726. . birds of Alaska. StackpoleCo., Harrisburg, PA. Cooc~, F. G. 1958. The breeding biology and man- GAUTHIER,G. 1989. The effect of experience and agement of the Blue Goose Chen caerulescens. timing on reproductive performance in Buffle- Ph.D.diss. Cornell Univ., Ithaca, NY. heads. Auk 106:568-576. COOKE.F. 1987. Lesser Snow Goose: a lone-term HAMANN,J., B. ANDREWS,AND F. COOKE. 1986. The population study, p. 407432. In F. Cooke [ed.], role of follicular atresiain inter- and intra-seasonal Avian genetics.Academic Press, New York. clutch size variation in Lesser Snow Geese (Anser COOKE,F., M. A. BOUSF~ELD,AND A. SADURA. 1981. caerulescenscaerulescens). J. Anim. Ecol. 55:48 l- Mate changeand reproductive successin the Less- 489. er Snow Goose. Condor 831322-327. HARVEY,J. M. 1970. Factorsaffecting nesting success COOKE,F., ANDR. F. ROCKWELL.1988. Reproductive in the Blue Goose. Ph.D.diss. Cornell Univ., Ith- successin a Lesser Snow Goose population, p. aca, NY. 237-250. In T. H. Clutton-Brock [ed.], Repro- HARVEY,J. M. 1971. Factors affecting Blue Goose ductive success.Univ. Chicago Press, Chicago. nesting success.Can. J. Zool. 491223-234. COOPER,J. A. 1978. The history and breedingbiology HOLMES,R. T., AND C. R. BLACK. 1973. Ecological of the Canada Geese of Marshy Point, Manitoba. distribution of birds in the Kolomak River-As- Wildlife Mono. No. 6 1. kinuk Mountain region, Yukon-Kuskokwim Del- CRAIGHEAD, J. J., AND D. S. STOCKSTAD.1964. Breed- ta, Alaska. Condor 75: 150-163. ing age of Canada Geese. J. Wildl. Manage. 28: HUDEC,K., ANDZ. Kux. 1971. The clutch size of the 57-64. Greylag Goose (Anser anser) in Czechoslovakia. DAU, C. P., AND P. G. MICKELSON. 1979. Relation Zool. Listy 20:365-376. of weather to spring migration and nesting of INGLIS. I. R. 1977. The breedine behaviour of the Cackling Geese on the Yukon-Kuskokwim Delta, Pink-footed Goose:behavioural correlatesof nest- Alaska, p. 94-104. In R. L. Jarvis and J. C. Bar- ing success.Anim. Behav. 25~747-764. tonek [eds.], Management and biology of Pacific JACKSON,M. T. 1981. Vegetation patternsof an Em- Flyway Geese. Oregon State Univ. Book Stores, peror Goose nesting area near Kokechik Bay, Corvallis, OR. Western Alaska. National GeographicSociety Re- DAVIES,J. C., AND COOKE, F. 1983. Annual nesting searchReports 13:287-296. productivity in Snow Geese: prairie droughtsand JOHNSON,D. H., AND T. L. SHAFFER. 1990. Estimat- arctic springs.J. Wildl. Manage. 47:291-296. ing nest success:when Mayfield wins. Auk 107: DEMENT’EV,G. P., AND N. A. GLADKOV[EDS.] 1952. 595-600. Birds of the Soviet Union, Vol. IV. Israel Program KAPLAN,E. L., AND P. MEIER. 1958. Nonparametric for Scientific Translations, Jerusalem.U.S. Dept. estimation from incomplete observations.J. Am. Interior and National ScienceFoundation, Wash- Stat. Assoc. 53:457-481. ington, DC. IOSTCHINSKI,A. A. 1971. Biologicalnotes on the Em- EBBINGE,B., A. ST. JOSEPH,AND B. SPAANS. 1982. peror Goose in north-east Siberia. Wildfowl 22: XXVII. The importance of springstaging areas for 29-34. arctic-breeding geese, wintering in western Eu- KLEIN, D. R. 1966. Waterfowl in the economy of the rope. Aquila 89:249-258. Eskimos on the Yukon-Kuskokwim Delta, Alas- EISENHAUER,D. I. 1976. Biology and behavior of the ka. Arctic 19:319-336. Emperor Goose (Ansercanagicus Sewastianov) in KLOMP, H. 1970. The determination of clutch-size Alaska. M.S.thesis. Purdue Univ., W. Lafayette, in birds: a review. Ardea 58: l-l 24. IN. KOSKIMIES, J. 1957. Variations in size and shape of EISENHAUER,D. I., AND C. M. KIRKPATRICK. 1977. eggsof the Velvet Scoter, Melanitta fusca (L.). Ar- Ecology of the Emperor Goose in Alaska. Wildl. chivum 12:58-69. Mono. 57. KOSSACK,C. W. 1950. Breeding habits of Canada ELDRIDGE,J. L., AND G. L. KRAPU. 1988. The influ- Geese under refuge conditions. Amer. Midland ence of diet quality on clutch size and laying pat- Nat. 43~627-649. tern in Mallards. Auk 105:102-l 10. KRECHMAR,A. V., AND A. YA. KONDRATIEV. 1982. ELY, C. R., AND D. G. RAVELING. 1984. Breeding Ecology of nesting of Philucte cunagicuson the biolonv of Pacific White-fronted Geese. J. Wildl. north part of Chukotka Peninsula. J. of Zoology Manage. 48:823-837. 6 11254-264 (in Russian). FINDLAY,C. S., AND F. COOKE. 1982. Breeding syn- chrony in the Lesser Snow Goose (Anser caeru- LACK, D. 1966. Population studies of birds. Clar- endon Press, Oxford, England. lescenscaerulescens). I. Genetic and environmen- tal componentsof hatch date variability and their LESSELLS,C. M., ANDP. T. BOAG. 1987. Unrepeatable effectsof hatch synchrony.Evolution 36:342-35 1. repeatabilities:a common mistake. Auk 104:116- FINDLAY,C. S., AND F. COOKE. 1983. Genetic and 121. environmental componentsofclutch size variance LESSELLS,C. M., F. COOKE,AND R. F. ROCKWELL.1989. in a wild population of LesserSnow Geese (Anser Is there a trade-off between egg weight and clutch caerulescenscaerulescens). Evolution 37~724-734. size in wild Lesser Snow Geese (Anser c. caeru- FINNEY.G.. ANDF. COOKE. 1978. Reoroductive hab- lescens)?J. Evol. Biol. 2t457-472. its in the Snow Goose: the influence of female age. MACINNES,C. D., R. A. DAVIS, R. N. JONES,B. C. Condor 80:147-158. LIEFF,AND A. PAKULAK. 1974. Reproductive ef- 396 MARGARET R. PETERSEN

ficiency of McConnell River small Canada Geese. tion and National Science Foundation, Washing- J. Wildl. Manage. 38:686-707. ton, DC and Amerind Publishing, New Delhi. MACINNES,C. D., AND E. H. DUNN. 1988. Compo- PROP,J., M. R. VANEERDEN, AND R. H. DRENT. 1984. nentsofclutch size variation in arctic-nestingCan- Reproductive successof the BarnacleGoose Brun- ada Geese. Condor 90:83-89. ta leucopsisin relation to food exploitation on the MASON,C. F. 1988. Parallel changesin numbers of breeding grounds, western Spitsbergen.Nor. Po- waders and geese.Bird Study 36:80-82. larinst. Skr. 181:87-117. MCLANDRESS,M. R., AND D. G. RAVELING. 1981. RAVELING,D. G. 1978. The timing of egg laying by Changesin diet and body composition of Canada northern geese.Auk 95294-303. Geese before spring migration. Auk 98:65-79. RAVELING,D. G. 1979a. The annual cycle of body MICKEL~ON,P. G. 1975. Breedingbiology of Cackling composition of Canada Geese with special refer- Geese and associatedspecies-on the-Yukon-Kus- ence-to control of reproduction. Auk 96:234-252. kokwim Delta, Alaska. Wild]. Monog. 45. RAVELING,D. G. 1979b. The annual energy cycle of MURPHY, M. T. 1986. Temporal components of re- the Cackling Canada Goose, p. 81-93. 1r1R. L. productive variability in Eastern Kingbirds (Ty- Jarvis and J. C. Bartonek [eds.], Management and runnus tyunnus). Ecology 67:1483-1492. biology of Pacific Flyway Geese. Oregon State NEWTON,I. 1977. Timing and successof breeding in Univ. Book Stores, Corvallis, OR. tundra-nesting geese, p. 113-126. In B. Stone- RAVELING,D. G. 1981. Survival, experience,and age houseand C. M. Perrins [eds.], Evolutionary ecol- in relation to breeding successof Canada Geese. ogy. Univ. Park Press, Baltimore, MD. J. Wildl. Manage. 45:s 17-829. NEWTON,I. [ED.] 1989. Lifetime reproductionin birds. RAVELING,D. G. 1988. Mate retention in Giant Can- Academic Press, New York. ada Geese. Can. J. Zool. 66:2766-2768. OVEN, M. 1987. Brent Goose Branta b. berniclu RAVELING,D. G., ANDH. G. LUMSDEN. 1977. Nesting breeding and lemmings-a re-examination. Bird ecologyof Canada Geese in the Hudson Bay low- Study 34: 147-149. lands of Ontario: evolution and population regu- OWEN,M., ANDJ. M. BLACK. 1989. Factors affecting lation. Ontario Ministry of Natural Resources.Fish the survival of Barnacle Geese on migration from and Wildl. Res. Report No. 98. the breedinggrounds. J. Anim. Ecol. 58:603-6 17. Ram. E. C. 1989. Consistencv in the timing of mi- OWEN, M., J. M. BLACK,AND H. LIBER. 1988. Pair gration for individual Bewick’s Swans.Amm. Be- bond duration and timing of its formation in Bar- hav. 38:384-393. nacle Geese (Branta leucopsis),p. 23-38. In M. ROCKWELL,R. F., S. C. FINDLAY,AND F. COOKE. 1983. W. Weller [ed.], Waterfowl in winter. Univ. Min- Life history studiesof the LesserSnow Goose: the nesota Press, Minneapolis. influence of age and time on fecundity. Oecologia PAMPLIN,W. L., JR. 1986. Cooperative efforts to halt 56:318-322. population declines of geese nesting on Alaska’s ROCKWELL,R. F., S. C. FINDLAY,AND F. COOKE. 1987. Yukon-Kuskokwim Delta. Trans. N. Am. Wildl. Is there an optimal clutch size in Snow Geese? Nat. Resour. Conf. 5 1:487-506. Am. Nat. 130:839-863. PEHRSSON,0. 1986. Duckling production of the Old- ROHWER,F. C., AND D. I. EISENHAUER.1989. Ees, squaw in relation to spring weather and small- massand clutch size relationshipsin geese,eiders, rodent fluctuations. Can. J. Zool. 64: 1835-l 84 1. and swans. Omis Stand. 20:43-48. PERRINS,C. M., ANDR. H. MCCLEERY. 1989. Laying RYDER,J. P. 1972. Biology of nesting Ross’s Geese. datesand clutch size in the Great Tit. Wilson Bull. Ardea 60:185-215. 101:236-253. SFDINGER,J. S., ANDD. G. RAVELING. 1986. Timing PETERSEN,M. R. 1983. Observations of Emperor of nesting by Canada Geese in relation to the phe- Geese feeding at Nelson Lagoon, Alaska. Condor nology and availability of their food plants. J. 85:367-368. Anim. Ecol. 55:1083-l 102. P~ERSEN,M. R. 1990. Nest site selectionof Emperor SPENCER,D. L., U. C. NELSON,AND W. A. ELKINS. Geese and Cackling Canada Geese. Wilson Bull. 1951. America’s greatestgoose-Brant nesting area. 102:413426. Trans. N. Am. Wildl. Conf. Nat. Resour. 16:290- PETERSEN,M. R. 199 1. Reproductive ecologyof Em- 295. neror Geese. Ph.D.diss. Univ. of California. Da- SPSB. 1986. SPS4 user’s guide, 2nd ed. McGraw- vis, CA. Hill Book Company, New York, NY. PETERSEN,M. R. 1992. Reproductive ecologyof Em- STICKNEY,A. A. 1989. The foragingbehavior, habitat peror Geese: survival of adult females. Condor use, and diet of arctic foxes (Alopex lagopus)in a 94:398-406. goose nesting area near Kokechik Bay, Alaska. PETERSEN,M. R., AND R. E. GILL, JR. 1982. Popu- M.S.thesis. Univ. Alaska, Fairbanks, AK. lation and statusof Emperor Geesealong the north STRANG,C. A. 1976. Feeding behavior and ecology side of the Alaska Peninsula. Wildfowl 33:31-38. of Glaucous Gulls in western Alaska. Ph.D.diss. POLLOCK,K. H., S. R. WINTERSTEIN,C. M. BUNCK, Purdue University, W. Lafayette, IN. AND P. D. CURTIS. 1989. Survival analvsis in SUMMERS,R. W. 1986. Breedingproduction of Dark- telemetry studies: the staggeredentry design. J. bellied Brent Geese Brunta bernicla bernicla in Wildl. Manage. 53:7-15. relation to lemming cycles. Bird Study 33:105- POR~NKO, L. A. 1981. Birds of the Chukchi Pen- 108. insula and Wrangel Island. Smithsonian Institu- SUMMERS,R. W., AND L. G. UNDERHILL. 1987. Fac- EMPEROR GOOSE NESTING ECOLOGY 397

tom related to breeding production of Brent Geese Clutton-Brock [ed.], Reproductive success.Univ. Branta b. bernicla and waders (Charadrii) on the Chicago Press, Chicago. Taimyr Peninsula. Bird Study 34: 16l-l 7 1. VAN NOORDWUK,A. J., J. H. VAN BALEN,AND W. THOMPSON.S. C.. AND D. G. RAVELING. 1987. In- SC-. 1980. Heritability ofecologically im- cubatibn behavior of Emperor Geese compared portant traits in the Great Tit. Ardea 68: 193-203. with other geese:interactions of predation, body VAN NOORDWUK,A. J., J. H. VAN BALEN,AND W. size, and energetics.Auk 104:707-7 16. Sc~~tuoo. 1981 a. Genetic and environmental Tonr, C., D. L. TRAUGER,AND H. W. MURDY. 1984. variation in clutch size of the Great Tit. Neth. J. Seasonaldecline in brood sizes of sympatric wa- Zool. 31:342-372. terfowl (Anas and Aythya, Anatidae) and a pro- VAN NOORDWUK,A. J., J. H. VAN BALEN, AND W. posedevolutionaryexplanation.J. Anim. Ecol. 53: SCHAFUOO.198 1 b. Genetic variation in the tim- 75-92. ing of reproductionin the Great Tit. Oecologia49: VANNOORDWUK, A. J. 1987. Quantitative ecological 158-166. genetics of Great Tits, p. 363-380. In F. Cooke VON HAARTMAN,L. 1971. Population dynamics, p. and P. A. Buckley [eds.],Avian genetics.Academ- 392459. In D. S. Famer and J. R. Ring [eds.], ic Press, London. Avian biology, Vol. 1. Academic Press,New York. VAN Noc~a~wt.t~,A. J., AND J. H. VAN BALEN. 1988. WELLER,M. 1957. An automatic nest trap for wa- The Great Tit, Pam major, p. 119-l 35. In T. H. terfowl. J. Wildl. Manage. 21:456458.