DOES SEA ICE CONSTRAIN THE BREEDING SCHEDULES OFHIGH ARCTIC NORTHERN FULMARS?

MARK L. MALLORY1,3

AND MARK R. FORBES2

1Canadian Wildlife Service, Box 1714, Iqaluit, NU X0A 0H0, Canada2Department of Biology, Carleton University, Ottawa, ON K1S 5B6, Canada

Abstract. The Northern Fulmar (Fulmarus glacialis) is a pelagic seabird that breedsacross 25u of latitude, from the boreal to the high Arctic oceanographic zones. Weexamined the breeding schedule of fulmars in the remote Cape Vera colony in theCanadian high Arctic, a marine region covered by sea ice much of the year, to determine ifthe timing of breeding and colony attendance patterns of birds differed from the breedingphenology of fulmars in colonies farther south. Cape Vera fulmars arrived at the colonylater in the year, spent less time at the colony before egg-laying, and took a significantlylonger prelaying exodus from the colony compared to fulmars nesting in more southerlycolonies. After egg-laying, however, patterns of colony attendance by fulmars in the highArctic were similar to patterns for fulmars in southern colonies; this part of the fulmarbreeding schedule may be inflexible. The differences in breeding schedules across thespecies’ range might reflect behavioral adaptations by arctic-nesting birds to accommodatethe physical and biological limitations imposed by extensive sea ice near arctic colonies,particularly early in the breeding season. Given that climate warming and correspondingreductions in sea ice are taking place in the Arctic, it remains to be determined whetherfulmars in the high Arctic have the behavioral flexibility in their breeding phenology tocompensate for rapidly occurring changes in their environment.

Key words: Arctic, breeding, Fulmarus glacialis, Northern Fulmar, sea ice.

¿Restringe el Hielo Marino la Fenologıa Reproductiva de Fulmarus glacialis a Altas Latitudes

en el Artico?

Resumen. Fulmarus glacialis es un ave marina pelagica que se reproduce en las zonasoceanograficas desde los 25u de latitud norte hasta altas latitudes en el Artico.Examinamos la fenologıa reproductiva de individuos de F. glacialis en la colonia remotade Cabo Vera ubicada a altas latitudes en el Artico canadiense, una region marina cubiertapor hielo durante gran parte del ano. Trabajamos allı para determinar si el inicio de lareproduccion y los patrones de presencia de aves en las colonias diferıa de la fenologıareproductiva de colonias de F. glacialis que se encuentran mas al sur. Los individuos deCabo Vera llegaron a las colonias mas tarde, permanecieron por menos tiempo en lascolonias antes de comenzar la puesta de huevos y se ausentaron antes de la postura por unperıodo significativamente mas largo que los individuos que nidificaron en colonias massurenas. Sin embargo, despues del periodo de postura de huevos, los patrones de presenciade individuos fueron similares entre las colonias del artico y las surenas; esta parte delperiodo reproductivo podrıa ser inflexible. Estas diferencias en la fenologıa reproductivaen el area de distribucion de esta especie pueden reflejar adaptaciones de comportamientopor parte de las aves que nidifican en el Artico para hacer frente a las limitaciones fısicas ybiologicas impuestas por las extensas areas de hielo en el mar cercanas a las colonias delArtico, especialmente durante las etapas tempranas de la epoca reproductiva. Debido alcalentamiento climatico y la consecuente reduccion del hielo marino que esta ocurriendoen el Artico, es aun necesario determinar si los individuos que se reproducen a altaslatitudes en el Artico tienen la flexibilidad conductual en su fenologıa reproductivanecesaria para compensar los cambios que estan ocurriendo repentinamente en suambiente.

INTRODUCTION

The timing of breeding by birds has receivedconsiderable attention. Timing of nesting might

relate to the ability of the female to gathersufficient food for egg production (Perrins1970), the parents’ abilities to gather resourcesfor self-maintenance during incubation (Brooke1978), or the parents’ abilities to find sufficientfood to provision young while maintaining theirown body condition (Lack 1954, Gaston andHipfner 1998). Recent studies have identified

Manuscript received 7 December 2006; accepted 4June 2007.

3 E-mail: mark.mallory@ec.gc.ca

The Condor 109:894–906# The Cooper Ornithological Society 2007

[894]

connections between breeding phenology andlarge-scale effects of climate change on marineecosystems (Montevecchi and Myers 1997,Ainley 2002). Among seabirds, timing of re-production typically reflects annual patterns ofmarine productivity (Cairns 1987, Abrahamand Sydeman 2004). In years when marineproduction is low, often a result of majorclimatic factors, marine birds may nest later(Schreiber 2002, Gaston et al. 2005).

For seabirds nesting in polar regions, breed-ing phenology may be particularly affected bymarine conditions (Hunt 1991), because sea-icecover creates a physical barrier to open wateraccess (Brown and Nettleship 1981) and alsoreduces light penetration, thereby reducing ordelaying productivity at lower trophic levels(Raymont 1976, Welch et al. 1992). In yearswhen sea-ice cover is extensive or persists longerinto the breeding season, marine birds shouldexperience higher foraging costs, because theyhave to travel farther to find food, as well asreduced food availability, because marine pro-duction is adversely affected. Both of thesefactors could influence reproductive decisionsby breeding birds. Indeed, several studies ofpolar seabirds have found delayed egg-layingand reduced reproductive success in years withmore severe sea-ice conditions (Ainley et al.1983, Gaston and Hipfner 1998, Gaston et al.2005).

Another approach for addressing the impor-tance of ice conditions to seabird reproductionis to compare timing and duration of keybreeding events within species across a rangeof latitudes. This reduces the contribution ofinterannual variation in other factors that mayrelate to annual variation in sea-ice coverduring the breeding season. Specifically, if sea-ice cover acts as an important constraint onseabird breeding as inferred, then we wouldexpect to find consistent differences in breedingphenology between northern (high Arctic) andsouthern populations of widespread species.Using published information, such a studyrequires comparing birds at different coloniesin different years, necessitates accounting forvarious study designs, and assumes that breed-ing metrics for each colony were derived duringtypical annual environmental conditions.

The Northern Fulmar (Fulmarus glacialis) isideal for such comparisons: it is a fulmarinepetrel with a circumpolar distribution, and it

breeds between 55u and 80uN in the NorthAtlantic Ocean (Hatch and Nettleship 1998).The Northern Fulmar is long-lived, lays one eggper clutch, and does not renest followingbreeding failure (Dunnet et al. 1963, Ollasonand Dunnet 1978). Like all petrels, this fulmarundertakes an exodus (a key breeding event)from the breeding colony prior to egg-laying,which allows the female to gather nutrients toform her single, relatively large, energy-rich egg(Warham 1990, 1996). Additionally, breeding atfulmar colonies appears to be relatively syn-chronized annually, and shows little variationin mean laying dates across years (Hatch andHatch 1990), similar to many other petrelsincluding Sooty Shearwaters (Puffinus griseus;Richdale 1963), Short-tailed Shearwaters (P.tenuirostris; Serventy 1963), Manx Shearwaters(P. puffinus; Brooke 1978), Antarctic Fulmars(Fulmarus glacialoides), Antarctic Petrels (Tha-lassoica antarctica), Cape Petrels (Daptioncapense), and Snow Petrels (Pagodroma nivea;Hodum 2002).

In his seminal work on the species, Fisher(1952) deduced that Northern Fulmarsthroughout the boreal oceanographic zonehad a similar breeding schedule, but suggestedthat this schedule might be different for arcticcolonies. In the boreal zone of the Atlantic,fulmars typically begin to attend their breedingcolony intermittently in the autumn precedingthe breeding season and lay eggs in mid- to lateMay, with eggs hatching in early to mid-July(Fisher 1952, Dunnet et al. 1963, Coulson andHorobin 1972, MacDonald 1977, 1980). Severalauthors have provided some evidence thatfulmars in the Arctic do indeed nest later thanfulmars in the boreal zone (Freuchen andSalomonsen 1958, Falk and Møller 1997, Hatchand Nettleship 1998), but no studies to datehave provided details of the first arrival of thebirds at these colonies, nor measures of theduration of the prelaying exodus. These twokey features of fulmar breeding should be mostaffected by sea ice or climatic conditions,because sea-ice extent should be greatest andpelagic marine productivity lowest early in theyear.

We studied the breeding phenology ofNorthern Fulmars in the remote Cape Veracolony in the Canadian high Arctic. At this site,fulmars have to fly over extensive sea ice inBaffin Bay and Jones Sound for over half of

HIGH ARCTIC FULMAR BREEDING SCHEDULE 895

their breeding season (first arrival at the colonyto date of hatching), a constraint experiencedby only a few fulmar colonies across the highArctic (Falk and Møller 1997, Hatch andNettleship 1998). Because of this sea-ice cover,we predicted that birds would arrive at thecolony later, would have a shorter prelayingperiod at the colony, and would take a longerprelaying exodus than fulmars nesting in theboreal oceanographic zone. We tested theseexpectations, while also comparing differencesin breeding phenology related to gender orbreeding status as other studies have done. Ourstudy thereby provides baseline data on timingand duration of key breeding events of fulmarsnesting at this colony.

METHODS

We conducted fieldwork between 14 May and 9August 2004 and from 20 April to 10 August2005 at the ‘‘Cape Vera’’ fulmar colony(76u159N, 89u159W) on northern Devon Island,beside the Hell Gate–Cardigan Strait polynya(Fig. 1). This marine area is located in the highArctic oceanographic zone (Salomonsen 1965),and nearby Jones Sound remains ice-coveredfrom October through July (except for the openwater near the polynya; Mallory and Fontaine

2004). We obtained data on the ice conditionsof Jones Sound and Baffin Bay from theEnvironment Canada – Canadian Ice Serviceclimate archives (available at ,http://ice-glaces.ec.gc.ca.). Typical conditions for maximum,circumpolar extent of arctic sea ice were takenfrom satellite imagery (National Snow and IceData Center 2007).

The Cape Vera colony is the most northerlyand remote fulmar colony in North America(Hatch and Nettleship 1998, Mallory 2006).The coastline at Cape Vera features prominentcliffs that rise 245–313 m above sea level, andapproximately 11 000 pairs of fulmars occupynest sites along 6.4 km of deeply incised cliffs(Gaston et al. 2006). Breeding fulmars do notappear to feed in the Hell Gate polynya, andinstead fly east to forage between EllesmereIsland and Greenland (MLM, unpubl. data).

We established eight long-term monitoringplots along the breeding cliffs at Cape Vera,with each plot supporting 25–300 nest sites(Mallory and Gaston 2005). Nest sites andattending birds could be easily viewed using 103 42 binoculars or a 603 spotting scope fromvantage points 30–300 m distant along the clifftop, meaning that birds were not disturbedduring counting. Cape Vera is highly suscepti-

FIGURE 1. The location of the Cape Vera study site on northern Devon Island, Nunavut, Canada.

896 MARK L. MALLORY AND MARK R. FORBES

ble to fog and high winds, and these weatherconditions meant that data on all plots werecollected on only 34 of 87 (39%) days in 2004.One plot was particularly problematic, and thisplot was replaced in 2005 (accounting for someof the total count differences between years). In2005, data on all plots were collected on 93 of107 (87%) days, and data were collected onsome of the plots on days when other plots wereobscured. In 2005, we also used a CanonH 20Ddigital camera with a 400 mm lens to producean 8 megapixel photograph of each plot eachday. These were downloaded to field compu-ters, and counts could be conducted or con-firmed in camp from these photos during poorweather. In situations where a nest could not beobserved on day x, but we identified the samebird at the nest on day x 2 1 and day x + 1, weassumed the bird was on the nest during day x.If the other pair member appeared on day x + 1,we assumed that the new bird arrived that day,and attributed the missed observation to thebird observed on day x 2 1.

To determine which member of the pair wasin attendance, we relied on several types ofinformation. Northern Fulmars range in plum-age color from very dark (DD) to intermediate(D or L) to very light (LL), which refers toincreasing amounts of white on the breast,neck, and head of birds (Fisher 1952, Hatchand Nettleship 1998). Fulmars of the CanadianArctic are mostly intermediate birds (L and D),with numerous very light morphs (LL), andrelatively few very dark birds (DD). Plumagedifferences, along with distinct markings on thebill, allow experienced observers to distinguishmembers of a pair based on these character-istics. Live fulmars cannot be sexed unlessmeasured (Mallory and Forbes 2005), or unlessthe position of birds during copulation is noted(Hatch 1990a). However, males are usually firstto arrive at the colony (as with many otherseabirds; Ainley et al. 1983), last to leave andfirst to return from the exodus, and they takethe first long incubation shift (Fisher 1952,Hatch 1990a, 1990b, 1990c).

It is not easy to determine if a fulmar is on anegg, as nonbreeders often appear to be in-cubating for periods of hours, but then leavethe nest site (Fisher 1952, Gaston et al. 2006).In preliminary work in 2003, we estimated thatapproximately one-third of the time we in-correctly assumed that observed birds at nest

sites were breeding adults and then saw them flyoff with no egg or chick at the site. To reducethis error, we determined that repeated obser-vations of the same individuals for severalhours each day were required.

On five of the study plots in 2004 and 2005,we conducted detailed, daily observations for1–4 hr of selected nest sites to determine whichpair member was attending the nest, and toidentify whether an egg or chick was in the nest.In some cases we recorded whether the birdexhibited ‘‘egg behavior’’ and confirmed thestatus of the nest at a later date when the eggwas subsequently observed. Repeated observa-tions of the same plots by the same personnelmeant that observers became familiar with theindividual birds’ behaviors. Thus, for thesewell-studied plots, ‘‘egg behavior’’ was incor-rectly ascribed to only 1% of birds for which welater confirmed that there was no egg present.We defined the following three categories ofbirds: successful breeders (fulmars at nests thatproduced a chick which was still alive by 10August); failed breeders (fulmars at nests thatproduced an egg, but either the egg or chick waslost during incubation or brood-rearing); andnonbreeders (fulmars at nests where we neverobserved an egg). Observing whether thesetypes of birds differ in attendance patterns isimportant as it might tell us whether inter-colony comparisons have to be made somewhatcarefully.

In 2005, we collected six fresh fulmar eggswithin one day of laying. We used the stainingand yolk ring counting technique of Astheimerand Grau (1985) to determine how many daysfemales spent developing eggs.

We followed the approach of Hatch (1990c)to divide the fulmar breeding season into stagesas follows: stage 1 (pre-exodus) – 40 to 21 daysbefore egg-laying, covering the period of pre-laying colony attendance and the departure forthe exodus; stage 2 (prelaying) – 20 to 1 daysbefore egg-laying, covering most of the exodusfrom the colony and the return of fulmars priorto the day the egg was laid; and stage 3(incubation) – 0 to 62 days after egg-laying,covering incubation and the posthatching guardstage where one of the parents broods the chick.We then compared the amounts of time themale and female spent at the colony in each ofthese stages for those fulmars that successfullyreached the end of stage 3. We were not at the

HIGH ARCTIC FULMAR BREEDING SCHEDULE 897

colony long enough to cover Hatch’s (1990c)stage 4 (chick-rearing), nor to observe youngfledge.

To compare breeding schedules of fulmars atCape Vera to those from colonies elsewhere, wereviewed the literature for published values ofkey phenological dates (e.g., first landing, startand end of exodus, egg-laying, date of hatch-ing). In situations where studies were conductedover several years, we took the earliest datesfrom any of the years as our measure for eachcolony-specific date. Aside from Cape Vera,only fulmar colonies in Canada, Greenland,Spitzbergen (Norway), and Franz Joseph Land(Norway) encounter extensive sea ice at somestage of the breeding season (National Snowand Ice Data Center 2007).

For calculations, all dates were converted toordinal dates, with 1 January 5 001. Insituations in which the date of first landingreported in the literature was in the autumnprior to a new breeding season, we subtractedthe number of days from 1 January (e.g., 1December 5 2031). Because Cape Vera datawere missing for many days in 2004, we referprincipally to 2005 data below, but include anddescribe 2004 data where appropriate. Depend-ing on the distribution of data, we used t-tests,ANOVAs, Kruskal-Wallis tests, or Pearson orSpearman rank correlations to compare dataamong years or colonies. We used ANOVAs forcomparisons with summary data from otherstudies, but in all cases the Bartlett’s testssuggested significant differences in standarddeviations among studies (all P , 0.01).Therefore, statistical differences among studiesshould be interpreted cautiously. Means arepresented 6 SE unless otherwise noted.

RESULTS

SEA-ICE DISTRIBUTION

During April and May, fulmars encountered.90% sea-ice cover from Disko Island, Green-land (69u309N, 54uW) to the Cape Verabreeding colony, a distance of approximately1200 km. The exception was a recurrent area ofopen water between Ellesmere Island andGreenland (the Northwater Polynya, 76uN,75uW), which varied annually in size and couldbe 350–500 km away at this time. Between 1996and 2005, the mean distance to open water atCape Vera in the second week of June was 190

6 5 km, a distance that was consistent acrossyears (range: 175–225 km, coefficient of varia-tion 9%). By the third week of July, fulmars stillhad to cross at least 100 km of sea ice in 2004,and 300 km of ice in 2005, to reach feedingareas.

PRELAYING COLONY ATTENDANCE ANDTHE EXODUS

In 2005, Northern Fulmars were not observedin the nearby polynya nor landing on the cliffsbefore 30 April (two weeks before observationswere initiated in 2004). The first birds to appearat known nesting sites in monitoring plots wereobserved one day later, and numbers increasedrapidly to a maximum count for the year on 7May, before declining to ,5 fulmars by 21 May(Fig. 2). In 2005, the colony was effectivelydeserted for 14 days: only 10 fulmars wereobserved between 17 and 30 May at 307 nestsites. At 48 nests where we could distinguishpair members, knew the sex of the birds, andtracked attendance to hatching, 35% of thepairs arrived at the colony on the same day,with the male arriving before the female in allother cases. Thirty-three percent of pairs de-parted together for the exodus, while the femaledeparted before the male in all but 6% of theremaining pairs. Collectively, males arrivedearlier, spent longer at the colony prior todeparting for the exodus, were away from thecolony for a shorter exodus, and returned

FIGURE 2. Total counts of Northern Fulmars ineight study plots at Cape Vera in 2004 and 2005,showing the early season peak in attendance, thevirtual desertion of the colony following the peak,and the fluctuating attendance during incubation andearly chick-rearing.

898 MARK L. MALLORY AND MARK R. FORBES

earlier before egg-laying than females (Table 1).At 23 days long, the male exodus was five daysshorter than the female exodus, with femalesaway from the colony for approximately onemonth (Table 1).

Colony attendance from late prelaying toearly incubation (14 May to 23 June) was highlycorrelated between counts on the same day in2004 and 2005 (rs 5 0.86, P , 0.05, n 5 14;Fig. 2), suggesting that the colony was syn-chronized across years.

EGG-LAYING, INCUBATION, DATE OFHATCHING, AND CHICK-REARING

There was no variation in counts of yolk ringsamong fulmar eggs; all six females took 23 daysto form their egg. Mean egg-laying day forCape Vera in 2005 was 157.9 6 0.2 (6 June,range: 2–16 June, n 5 103), with 80% of eggslaid between 4 and 9 June. This was similar to2004, in which the egg-laying day for 16 nestsaveraged 157.1 6 0.6, and the earliest egg wasalso observed on 2 June.

The incubation period of fulmars at CapeVera was 49.0 6 0.2 days (range: 46–52, n 5

65), and mean day of hatching was 206.1 6 0.3(25 July, range: 22–31 July, n 5 65). Like egg-laying, date of hatching was highly synchro-nized, with 80% of eggs hatching between 23and 27 July. In 2004, we observed 39 nestshatch between 21 and 31 July (mean day 5

205.8 6 0.5), and behaviors of birds suggestedthat hatching occurred at a further nine nestsduring this period, but we were unable toconfirm the presence of a chick at these nests

until 3 August. Thus, it seems probable thatoverall mean dates of hatching in both yearswere within one day of each other.

For 24 nests where the full chick-guardingstage was observed, at least one parent was withthe chick for 14.6 6 0.4 days (range: 12–18 days). At another 16 nests, an adult wasobserved still attending the chick 10–19 daysafter hatching when our observations ended(i.e., before the completion of chick-rearing).

SEASONAL PATTERN OF TIME ALLOCATIONAT THE COLONY

During stage 1 (pre-exodus) of the breedingseason, males spent 44% and females spent 28%of their time at the colony, although thesedifferences were not statistically significant(Table 1). This stage was marked by highsynchrony in attendance within pairs, withmany pairs arriving and departing on the sameday (above). Males spent only 11% of stage 2(prelaying) at the colony, but this was signifi-cantly more than females, who attended thecolony for only 2% of those days. Duringincubation and early chick-rearing (stage 3),males spent slightly more than half of their timeat the nest, which was more than females, whoattended the colony for 44% of the period(Table 1). Thus, over the entire 103-day breed-ing period that we observed, males spent 46% oftheir time at the nest and females spent 35% oftheir time at the nest, for an overall mean of40% 6 1% (n 5 38). During this period, birdswere observed together at the nest 9% 6 1% ofthe time.

TABLE 1. Comparison of key dates and time spent at the Cape Vera colony by paired male and femaleNorthern Fulmars during the 2005 breeding season (paired t-tests, all t $ 6.7, all P , 0.001). The breedingseason was divided into three stages as follows: stage 1 5 pre-exodus, 40 to 21 days before egg-laying; stage 25 prelaying, 20 to 1 day before egg-laying; and stage 3 5 incubation, 0 to 62 days after egg-laying.Calculations of the percentage of time spent at the colony used only pairs that had nests where incubation wascompleted, a chick was observed, and the identity of each member of the breeding pair was ascertained reliably.

Parameter

Male Female

Mean 6 SE Range n Mean 6 SE Range n

First arrival at colony (ordinal date) 123.8 6 1.4 121–129 48 124.9 6 1.4 122–129 48Duration of prelaying attendance (days) 8.9 6 2.0 5–15 48 5.7 6 1.7 2–10 48Duration of exodus (days) 23.1 6 2.8 15–28 48 28.1 6 3.1 22–39 48Return from exodus (ordinal date) 154.7 6 2.0 151–159 48 157.7 6 2.5 153–165 48Time spent at colony in stage 1 (%) 44 6 10 25–60 38 28 6 8 10–50 38Time spent at colony in stage 2 (%) 11 6 11 0–50 38 2 6 5 0–15 38Time spent at colony in stage 3 (%) 52 6 5 42–61 38 44 6 6 31–56 38Time at colony prelaying to early

chick-rearing (%)46 6 5 36–57 38 35 6 4 26–43 38

HIGH ARCTIC FULMAR BREEDING SCHEDULE 899

BREEDING SCHEDULES ANDBREEDING STATUS

At Cape Vera, males and females of known sex(above) appeared to have consistent patterns ofarrival and departure, so we assumed that thefirst birds to arrive and the last birds to departduring prelaying were males. Using this ap-proach, we found no significant differences for133 apparently occupied nest sites amongsuccessful, failed, or nonbreeders in any of thefollowing: the synchrony of arrival and de-parture from the colony; the amount of time thepair spent together during prelaying; theamount of time spent at the colony by malesor females during stage 1; or the duration of themale or female exodus (Table 2). Fulmarsattended their nest sites from prelaying tochick-rearing even if they were non- or failedbreeders, but over the entire 103-day observa-tion period, nonbreeders were more oftenobserved together at the colony than successfulbreeders (Table 2). In addition, nest sites wereleft unattended significantly less often during

incubation by successful breeders than by failedbreeders, and both of these groups were at thenest more often than nonbreeders (Table 2;Kruskal-Wallis tests, all P , 0.03). Thus, priorto egg-laying, fulmars exhibited similar breed-ing schedules irrespective of breeding status,but during incubation, nonbreeders and failedbreeders spent more time away from the nestsite.

COMPARISONS WITH BREEDINGSCHEDULES OF NORTHATLANTIC FULMARS

For 16 fulmar colonies located across the NorthAtlantic and Arctic Oceans, the first landing offulmars at the colony was significantly later athigher latitudes (rs 5 0.68, P 5 0.004; Table 3,Fig. 3a). This same pattern held when theanalysis was restricted to seven colonies situ-ated around Baffin Bay (rs 5 0.68, P 5 0.11),although the trend was not significant due tothe small sample size. First egg dates (rs 5 0.74,P 5 0.04, n 5 8), mean egg dates (rs 5 0.64, P

TABLE 2. Values of colony attendance parameters (mean 6 SE) for successful, failed, and nonbreedingNorthern Fulmars nesting in the Cape Vera colony in 2005. The breeding season was divided into three stagesas follows: stage 1 5 pre-exodus, 40 to 21 days before egg-laying; stage 2 5 prelaying, 20 to 1 day before egg-laying; and stage 3 5 incubation, 0 to 62 days after egg-laying. For failed breeding adults, date of hatching wascalculated as 49 days from the laying date. For nonbreeding adults, the overall mean egg-laying date, meandate of hatching, and end of observations during chick-rearing were used to calculate proportions of timespent at the colony during stage 3.

Parameter

Breeding status (n)Statistical

comparison

Successful (52) Failed (48)Nonbreeding

(33) F or K-W P

Synchrony of prelaying arrival (days)a 1.2 6 0.2 1.4 6 0.2 1.6 6 0.3 0.7 0.48Synchrony of departure for exodus (days)b 2.0 6 0.3 1.7 6 0.3 1.2 6 0.4 1.3 0.26Proportion of days paired during prelaying (%) 76.8 6 3.9 72.8 6 4.8 64.1 6 5.1 1.3 0.28Male at colony during stage 1 (days) 8.8 6 0.3 9.0 6 0.4 8.7 6 0.4 0.2 0.84Female at colony during stage 1 (days) 5.5 6 0.2 5.7 6 0.3 6.0 6 0.4 0.7 0.50Male exodus (days)c 23.2 6 0.4 23.8 6 0.4 24.6 6 0.7 2.0 0.14Female exodus (days)d 28.4 6 0.4 28.5 6 0.4 28.2 6 1.0 0.1 0.94Synchrony of return from exodus (days)a 3.2 6 0.4 2.9 6 0.4 3.2 6 0.6 0.1 0.86Nest site unattended during incubation (days) 0.5 6 0.2 3.2 6 0.7 12.5 6 1.3 64.8 ,0.001e

Days pair together during stage 3 (days) 4.7 6 0.4 8.7 6 0.6 11.6 6 1.0 40.7 ,0.001f

Days pair together from prelaying to earlychick-rearing (%)

9 6 1 13 6 1 14 6 1 20.3 ,0.001f

a Calculated as date of first arrival – date of second arrival.b Calculated as date of last to depart – date of first to depart.c Assumes male is last to depart and first to arrive back from exodus.d Assumes female is first to depart and last to arrive back from exodus.e Nonparametric Kruskal-Wallis and Dunn’s multiple comparison test conducted due to significant

difference in variation of each group; each group significantly different from the others (three groups tested).f Nonparametric Kruskal-Wallis and Dunn’s multiple comparison test conducted due to significant

difference in variation of each group; successful lower than failed or nonbreeding.

900 MARK L. MALLORY AND MARK R. FORBES

TA

BL

E3

.L

oca

tio

ns,

sea

ice

pre

sen

ce,

an

dre

po

rted

bre

edin

gp

hen

olo

gy

of

No

rth

ern

Fu

lma

rco

lon

ies

inA

lask

aa

nd

acr

oss

the

No

rth

Atl

an

tic

Oce

an

(No

rth

ern

Fu

lma

rsa

rek

no

wn

as

Fu

lma

rsin

Eu

rasi

a).

Pa

ram

eter

sre

po

rted

incl

ud

e:d

ate

of

firs

tla

nd

ing

(wh

enfu

lma

rsa

refi

rst

ob

serv

eda

tten

din

gth

eco

lon

yaft

erth

ep

rev

iou

sb

reed

ing

sea

son

);M

:Fex

od

us

(th

ele

ng

tho

fth

em

ale

an

dfe

ma

lep

rela

yin

gex

od

us

fro

mth

eco

lon

y);

da

teo

fth

eea

rlie

stre

po

rted

egg

-la

yin

g(f

irst

egg

),m

ean

da

teo

feg

g-l

ay

ing

,a

nd

mea

nd

ate

of

ha

tch

ing

,a

sw

ell

as

the

len

gth

of

the

incu

ba

tio

np

erio

d.

Fu

lma

rco

lon

yL

ati

tud

eL

on

git

ud

eS

eaic

ep

rese

nt?

aD

ate

of

firs

tla

nd

ing

M:F

exo

du

s(d

ays)

Da

teo

ffi

rst

egg

Mea

neg

g-l

ay

ing

da

te

Mea

nh

atc

hin

gd

ate

Incu

ba

tio

np

erio

d(d

ays)

Ref

eren

cec

Ala

ska

(Sem

idi

Isla

nd

s)5

6uN

15

6uW

No

11

Ap

r1

2.2

:17

.62

6M

ay

9Ju

n2

7Ju

l4

8.4

1U

nit

edK

ing

do

m(W

eyb

ou

rne)

53uN

1u6

9EN

o2

3O

ct2

Un

ited

Kin

gd

om

(Sa

nd

so

fF

orv

ie)

57u1

89N

1u5

69W

No

22

Oct

9.3

:19

.61

2M

ay

21

Ma

y3

Jul

3,

4U

nit

edK

ing

do

m(S

t.K

ild

a)

57u4

89N

8u3

69W

No

30

Oct

5M

ay

12

Ma

y2

1Ju

n5

Un

ited

Kin

gd

om

(Ork

ney

,S

het

lan

d)

59u3

09N

2uW

No

29

Oct

14

.5:1

5.9

12

Ma

y2

2M

ay

24

Jun

5,

6,

7

Fa

ero

eIs

lan

ds

62u3

09N

7u3

09W

No

30

Oct

19

Ma

y1

5Ju

l5

Icel

an

d6

5u3

09N

22u3

09W

No

20

Oct

11

Ma

y1

6M

ay

7Ju

l5

Gre

enla

nd

(Dis

ko

)6

9uN

54uW

No

8A

pr

5G

reen

lan

d(U

ma

na

q)

71uN

52uW

Yes

12

Ap

r2

8M

ay

5N

orw

ay

(Ja

nM

ay

en)

71uN

9uW

No

19

Dec

14

Julb

5C

an

ad

a(B

ail

larg

eB

ay

)7

3u2

49N

84u3

09W

Yes

23

Ap

r5

Ca

na

da

(Pri

nce

Leo

po

ldIs

lan

d)

74uN

90uW

Yes

2M

ay

3–

5Ju

n7

Jun

25

Jul

47

.78

,9

Ca

na

da

(Ca

pe

Lid

do

n)

74u3

69N

91u1

29W

Yes

26

Ap

r8

No

rwa

y(B

jorn

oy

a)

74u3

09N

19uE

No

Dec

14

Julb

5C

an

ad

a(C

ap

eV

era

)7

6u1

29N

89u1

29W

Yes

30

Ap

r2

3.2

:28

.23

Jun

8Ju

n2

5Ju

l4

9.0

10

No

rwa

y(S

pit

zber

gen

)7

7u3

09N

15uE

No

22

Jan

3Ju

n5

No

rwa

y(F

ran

zJo

sep

hL

an

d)

81uN

50uE

Yes

13

Ma

r5

Gre

enla

nd

(Ma

llem

uk

fjel

d)

80u1

29N

16u3

69W

Yes

24

Ap

r3

1M

ay

19

Jul

11

aE

xte

nsi

ve

sea

ice

typ

ica

lly

pre

sen

td

uri

ng

earl

yb

reed

ing

sea

son

(Ap

ril–

Ma

y).

bR

epo

rted

ha

tch

esa

lmo

sto

ne

mo

nth

earl

ier

seem

hig

hly

imp

lau

sib

le;

giv

end

ate

sub

sta

nti

ate

db

y.

1st

ud

y.

c1

5H

atc

ha

nd

Ha

tch

(19

90

);2

5S

late

r(1

99

0);

35

Ma

cDo

na

ld(1

97

7);

45

Ma

cDo

na

ld(1

98

0);

55

Fis

her

(19

52

);6

5D

un

net

eta

l.(1

96

3);

75

Bro

ok

e(2

00

4);

85

Ha

tch

an

dN

ettl

esh

ip(1

99

8);

95

Ga

sto

net

al.

(20

05

);1

05

this

stu

dy

;1

15

Fa

lka

nd

Mø

ller

(19

97

).

HIGH ARCTIC FULMAR BREEDING SCHEDULE 901

5 0.01, n 5 8), and first observed hatchingdates (rs 5 0.78, P 5 0.01, n 5 10; Table 3,Fig. 3b) also were later for fulmar colonies athigher latitudes.

Fulmars breeding at Cape Vera departed thecolony for a prelaying exodus that averagedmore than one week longer than fulmars atcolonies farther south (Table 3). Both males(ANOVA on summary data: F2,790 5 159.2, P, 0.001) and females (F2,779 5 113.6, P ,

0.001) spent significantly longer periods awayfrom the colony, and differences among theCape Vera, Semidi Islands, and Sands of Forviecolonies were all significant (Tukey-Kramermultiple comparisons tests, all P , 0.05).

The incubation period for fulmars at CapeVera (49.0 days, n 5 65) was slightly longer

than for fulmars in Alaska (48.4 6 0.1 days, n5 52; Hatch 1990a) or Prince Leopold Island(47.7 6 0.2 days, n 5 30; Hatch and Nettleship1998; ANOVA on summary data: F2,144 5 11.3,P , 0.001; Tukey-Kramer multiple compar-isons tests, all P , 0.05), although the range ofincubation periods was similar.

DISCUSSION

Hatch and Nettleship (1998) hypothesized thatNorthern Fulmars breeding in the high Arcticmight exhibit adaptations to the colder, ice-covered environmental conditions, includinga compressed breeding season, compared withboreal fulmars elsewhere in the Pacific andAtlantic Oceans. Results of our study ofNorthern Fulmars breeding at Cape Vera weresimilar in some respects to those of studies ofbreeding fulmars in the boreal zone, but we alsofound support for the hypothesis above. Aswith other fulmars, we found that females atCape Vera took 23 days to form their eggs(Astheimer and Grau 1985). Moreover, CapeVera fulmars completed incubation in 46–52 days (Hatch and Nettleship 1998) andguarded the chick for .14 days (Hatch1990c), despite having to fly over extensive seaice to forage throughout the prelaying andincubation periods. This result is similar topatterns observed by Hodum (2002), whereincubation, chick-guarding, and nestling peri-ods of four antarctic fulmarine petrels variedlittle for colonies separated by 21 degrees oflatitude. However, Cape Vera fulmars wereamong the last fulmars of the North Atlanticpopulation to begin breeding each year, withthe second-latest colony arrival date and thelatest mean egg-laying date. What consequencesdoes this have for their allocation of time todifferent stages of their breeding schedule?

Surprisingly, both male and female fulmarsbreeding at Cape Vera spent about the sameamount of time at the colony from arrivalthrough chick-guarding as fulmars in theSemidi Islands in Alaska (46% and 35%,respectively; Hatch 1990c), and even similaramounts of time together at the nest site.However, time at the nest appeared to beallocated differently between these colonies.Male fulmars at Cape Vera spent proportion-ally more time at the colony during pre-exodusbut less time during prelaying and incubationthan Hatch (1990c) observed in any of six years

FIGURE 3. The breeding phenology of selectedNorthern Fulmar colonies in the North AtlanticOcean, demonstrating earlier colony attendance, egg-laying, and hatching for birds breeding at lowerlatitudes. (a) First arrival by fulmars at the colony inrelation to colony latitude. (b) Mean date of egg-laying (circles) and mean date of hatching (squares)for fulmar colonies in the arctic (unfilled) and boreal(filled) oceanographic zones. In both panels, datafrom a colony in Alaska are shown by triangles.

902 MARK L. MALLORY AND MARK R. FORBES

for male fulmars in Alaska. In contrast, femalefulmars at Cape Vera spent about 6% less timeat the colony during prelaying, attributable totaking a 10-day longer exodus, but were at thecolony for similar proportions of time duringpre-exodus and incubation compared to femalefulmars in Alaska.

COMPRESSED BREEDING SEASON, THEEXODUS, AND SYNCHRONIZATION

Boreal fulmars first attend colonies six or moremonths prior to the annual breeding season(Fisher 1952, Coulson and Horobin 1972).However, arctic colonies surrounded by exten-sive sea ice are not visited by fulmars untilabout one month prior to egg-laying (except,occasionally, Pribilof Island colonies; S. Hatch,Alaska Science Center, pers. comm.). In north-ern Baffin Bay, there are no published observa-tions of fulmars before 24–26 April (Fisher1952), thus we suspect that there were nofulmar landings at Cape Vera before ourobservations on 30 April. Therefore, at CapeVera, fulmar mates typically arrive within twodays of each other and spend only six to ninedays together at the colony before departing forthe exodus. Unless pairs meet at sea (which isunknown), fulmars breeding in the high Arctichave a shorter period each year to choose orcheck nest sites, reestablish pair bonds, andcopulate than fulmars breeding in boreal or lowArctic colonies (Hatch and Nettleship 1998).

The prelaying exodus of fulmars at CapeVera is 9–14 days longer than that reported inother locations, and is characterized by a lackof intermittent, short visits to the colony whichmay be observed at southern colonies (Hatch1990c). Fulmars can travel several hundredkilometers to feed (Weimerskirch et al. 2001;MLM, unpubl. data), and surveys by McLaren(1982) suggested that fulmars may leave thehigh Arctic completely during the exodus,perhaps to feed near southwestern Greenlandor southern Davis Strait. We suspect thatmarine production is still too low early in theseason to support the energetic needs of fulmarsat this time (Welch et al. 1992, Lewis et al.1996), thus the birds appear to move to distant,more productive waters.

If the physical and biological effects of sea iceconstrain the breeding schedule of arcticfulmars (Fisher 1952, Falk and Møller 1997,Hatch and Nettleship 1998), it follows that

polar colonies should be more synchronized intheir schedules than southern colonies. Indeed,the range of egg-laying dates in the Canadianhigh Arctic (11–19 days) was shorter than therange in northeastern Greenland (24 days; Falkand Møller 1997), Alaska (18–34 days; Hatchand Hatch 1990), or the eastern North Atlantic(,30 days; Fisher 1952). Falk and Møller(1997) demonstrated how the phenology offulmar reproduction in northeastern Greenlandmatched the sea-ice cycle of the nearby polynya,a pattern also found by Hodum (2002) for fourspecies of fulmarine petrels near Antarctica.Thus, fulmars nesting in the most ice-con-strained sites appear to have higher synchroni-zation of egg-laying than fulmars breeding atcolonies surrounded by open water year-round.

COMPRESSED BREEDING SCHEDULE ANDEXTRA-PAIR COPULATIONS

While overall time spent at the colony fromarrival through chick-guarding was similarbetween fulmars at Cape Vera and Alaska,fulmars breeding in Alaskan colonies may beginattending breeding cliffs up to 47 days beforeegg-laying, compared to 40 days for fulmars atCape Vera (Hatch 1989). Pairs also may beobserved copulating at Alaskan colonies overa 30-day period, up to 10 days before egg-laying (Hatch 1987), because the prelayingexodus is shorter in Alaska. At Cape Vera, nopairs that attempted to breed were observed atthe colony after 14 May, and those same pairslaid eggs starting 5 June, meaning that nocopulations occurred ,22 days before egg-laying. This timing is similar to the length ofthe egg formation period (23 days), and thetime at which copulation frequencies begin todecline at boreal colonies (20 days before egg-laying; Hunter 1998), both of which suggestthat females at Cape Vera have already decidedwhether to breed and have secured sperm fromtheir mates potentially much earlier than atother colonies (Hunter 1998).

During the pre-exodus period, male fulmarsalways arrived at the Cape Vera colony beforeor with their mate, and males spent 8% moretime at the colony than males in Alaska (Hatch1990c). This increased attendance may reflecta behavioral adaptation of male fulmars to theshort breeding season in the high Arctic toenhance opportunities to copulate with thefemale, to secure extra-pair copulations, or to

HIGH ARCTIC FULMAR BREEDING SCHEDULE 903

guard the female from extra-pair copulationattempts by other males (Hatch 1987, Hunter1998). If this interpretation is correct, we makethe following predictions: (1) levels of extra-paircopulations by fulmars in high Arctic colonieswill be more highly correlated with proportion-al male attendance during prelaying than isfound in low Arctic or boreal colonies; (2) highArctic fulmars will have lower egg hatchability,because there will be fewer fertilizations due toabsent males during the sole opportunity tomate, and females may use older, degraded,stored sperm; and (3) male fulmars in highArctic colonies will exhibit energetic adapta-tions to breeding in ice-constrained environ-ments which require high attendance at the nestduring prelaying, and specifically that they willarrive at the colony with proportionally largerenergetic reserves than male fulmars at moresouthern colonies.

SCHEDULING DIFFERENCES BETWEENBREEDING AND NONBREEDING FULMARS

Nonbreeding fulmars attend colonies through-out the breeding season (Fisher 1952, Coulsonand Horobin 1972, Falk and Møller 1997),presumably to gain experience with potentialnest sites, timing of movements, and locationsof feeding areas (Warham 1990, 1996), whichare critical for successful reproduction (Ollasonand Dunnet 1978). At Cape Vera, nonbreedingfulmar pairs exhibited similar levels of synchro-ny in colony attendance as breeding birds, andspent similar amounts of time at the colonyprior to the exodus, unlike nonbreeding fulmarsin Alaska, which spent less time at the colonythan breeding adults. With little food andextensive ice nearby, there is probably anenergetic disadvantage for nonbreeders at CapeVera to leave the breeding cliffs early in theseason, whereas those in Alaska may be able tomake brief foraging trips at this time.

The differences in attendance patterns be-tween breeding and nonbreeding adult fulmarswere most evident during incubation. Nonbree-ders were observed either absent from the nestor with their mate at the nest site much morethan breeders, a result undoubtedly attributableto the greater energetic limitations placed onbreeding birds. At least one member ofa breeding pair must remain at the nest toincubate the egg and protect it from inclementweather or predation. In contrast, the value of

colony attendance by failed or nonbreedersdeclines as the season progresses, for tworeasons. First, it is too late in the season fornonbreeders to initiate a nest, and failedbreeders do not renest (Hatch and Nettleship1998). Second, the high energetic costs ofreplacing feathers during molt, particularlyprimary molt, are deferred by breeding birdsuntil late in chick-rearing, but nonbreedersinitiate molt shortly after mean egg-laying datesfor the colony, and failed breeders enter moltshortly after losing their egg or chick (Fisher1952, Hatch 1990c, Hatch and Nettleship 1998).

Collectively, the data from Cape Vera andstudies of other fulmar colonies affected by seaice indicate that high Arctic fulmars havea compressed and substantially adjusted breed-ing schedule compared to fulmars breeding inthe boreal oceanographic zone. These schedul-ing adaptations occur during prelaying; eggdevelopment and attendance during incubationand chick-rearing appear to be similar acrossthe species’ range.

With climate change already contributing toreductions in arctic sea ice (McBean 2004),fulmars may be less constrained in schedulingearly season breeding activities near theircolonies in the future. However, it is unclearwhether fulmars adapted to high Arctic condi-tions are flexible enough to adopt a more borealbreeding schedule, or whether they can main-tain a typical high Arctic breeding schedule butshift it earlier into the season. Barbraud andWeimerskirch (2006) showed that several spe-cies of antarctic seabirds have begun arriving atcolonies and laying eggs later over the past55 years. The authors noted that the birdsapparently have the behavioral plasticity toaccommodate the resultant shorter breedingperiod, perhaps by decreasing the time requiredfor activities in the prelaying period, similar tophenology at Cape Vera. However, somestudies have suggested that climate-mediatedearlier production of prey items has led toa mismatch of avian breeding phenology andprey abundance (Stenseth and Mysterud 2002),which could have detrimental long-term effectsif behavioral adaptations of fulmars cannotoccur as rapidly as changing food supplies.

ACKNOWLEDGMENTS

We thank the many students and collaborators whomade the Cape Vera project possible in 2003–2005,

904 MARK L. MALLORY AND MARK R. FORBES

particularly Jason Akearok, Tracy Arsenault, SarahChisholm, Darryl Edwards, Alain Fontaine, AlissaMoenting, Kieran O’Donovan, and Karen Truman,whose assistance was essential for gathering data onbreeding schedules in 2004 and 2005. Thanks also toKerrie Wilcox at Bird Studies Canada for analyzingfulmar eggs for us. Two reviewers provided insightfuland helpful comments on the manuscript, withparticular thanks to Scott Hatch for suggestionsregarding egg hatchability and fulmar mate-guard-ing. We are very grateful for the financial and logisticsupport provided by Environment Canada (NorthernConservation Division and the Northern EcosystemInitiative), Natural Resources Canada (Polar Conti-nental Shelf Project), the Nunavut Wildlife Manage-ment Board, Indian and Northern Affairs Canada(Environmental Capacity Development Initiative andNorthern Scientific Training Program), and CarletonUniversity.

LITERATURE CITED

ABRAHAM, C. L., AND W. J. SYDEMAN. 2004. Oceanclimate, euphausiids and auklet nesting: inter-annual trends and variation in phenology, dietand growth of a planktivorous seabird, Ptychor-amphus aleuticus. Marine Ecology ProgressSeries 274:235–250.

AINLEY, D. G. 2002. The Adelie Penguin: bell-weather of climate change. Columbia UniversityPress, New York.

AINLEY, D. G., R. E. LERESCHE, AND W. J. L.SLADEN. 1983. Breeding biology of the AdeliePenguin. University of California Press, LosAngeles, CA.

ASTHEIMER, L. B., AND C. R. GRAU. 1985. Thetiming and energetic consequences of egg forma-tion in Adelie Penguins. Condor 87:256–267.

BARBRAUD, C., AND H. WEIMERSKIRCH. 2006.Antarctic birds breed later in response to climatechange. Proceedings of the National Academy ofSciences USA 103:6248–6251.

BROOKE, M. 1978. Some factors affecting the layingdate, incubation and breeding success of theManx Shearwater, Puffinus puffinus. Journal ofAnimal Ecology 47:477–495.

BROOKE, M. 2004. Albatrosses and petrels across theworld. Oxford University Press, London.

BROWN, R. G. B., AND D. N. NETTLESHIP. 1981. Thebiological significance of polynyas to arcticcolonial seabirds, p. 59–65. In I. Stirling and H.Cleator [EDS.], Polynyas in the Canadian Arctic.Canadian Wildlife Service Occasional Paper 45.

CAIRNS, D. K. 1987. Seabirds as indicators of marinefood supplies. Biological Oceanography 5:261–271.

COULSON, J. C., AND J. M. HOROBIN. 1972. Theannual reoccupation of breeding sites by theFulmar. Ibis 114:30–42.

DUNNET, G. M., A. ANDERSON, AND R. M.CORMACK. 1963. A study of survival of adultFulmars with observations on the pre-layingexodus. British Birds 56:2–18.

FALK, K., AND S. MØLLER. 1997. Breeding ecologyof the Fulmar Fulmarus glacialis and the

kittiwake Rissa tridactyla in high-arctic north-eastern Greenland, 1993. Ibis 139:270–281.

FISHER, J. 1952. The Fulmar. Collins, London, UK.FREUCHEN, P., AND F. SALOMONSEN. 1958. The

Arctic year. G. P. Putnam and Sons, New York.GASTON, A. J., H. G. GILCHRIST, AND M. L.

MALLORY. 2005. Variation in ice conditionshas strong effects on the breeding of marinebirds at Prince Leopold Island, Nunavut. Eco-graphy 28:331–344.

GASTON, A. J., AND J. M. HIPFNER. 1998. The effectof ice conditions in northern Hudson Bay onbreeding by Thick-billed Murres (Uria lomvia).Canadian Journal of Zoology 76:480–492.

GASTON, A. J., M. L. MALLORY, H. G. GILCHRIST,AND K. O’DONOVAN. 2006. Status, trends andattendance patterns of the Northern FulmarFulmarus glacialis in Nunavut, Canada. Arctic59:165–178.

HATCH, S. A. 1987. Copulation and mate guarding inthe Northern Fulmar. Auk 104:450–461.

HATCH, S. A. 1989. Diurnal and seasonal patterns ofcolony attendance in the Northern Fulmar,Fulmarus glacialis, in Alaska. Canadian Field-Naturalist 103:248–260.

HATCH, S. A. 1990a. Incubation rhythm in theFulmar Fulmarus glacialis: annual variationand sex roles. Ibis 132:515–524.

HATCH, S. A. 1990b. Individual variation in behaviorand breeding success of Northern Fulmars. Auk107:750–755.

HATCH, S. A. 1990c. Time allocation by NorthernFulmars Fulmarus glacialis during the breedingseason. Ornis Scandinavica 21:89–98.

HATCH, S. A., AND M. A. HATCH. 1990. Breedingseasons of oceanic birds in a subarctic colony.Canadian Journal of Zoology 68:1664–1679.

HATCH, S. A., AND D. N. NETTLESHIP. 1998.Northern Fulmar (Fulmarus glacialis). In A.Poole and F. Gill [EDS.], The birds of NorthAmerica, No. 361. The Birds of North America,Inc., Philadelphia, PA.

HODUM, P. J. 2002. Breeding biology of high-latitudeAntarctic fulmarine petrels (Procellariidae).Journal of Zoology (London) 256:139–149.

HUNT, G. L., JR. 1991. Marine ecology of seabirds inpolar oceans. American Zoologist 31:141–142.

HUNTER, F. M. 1998. Timing of pair and extra-paircopulations in the Northern Fulmar. Condor100:141–145.

LACK, D. 1954. The natural regulation of animalnumbers. Clarendon Press, Oxford, UK.

LEWIS, E. L., D. PONTON, L. LEGENDRE, AND B.LEBLANC. 1996. Springtime sensible heat, nu-trients and phytoplankton in the Northwaterpolynya, Canadian Arctic. Continental ShelfResearch 16:1775–1792.

MACDONALD, M. A. 1977. The pre-laying exodus ofthe Fulmar Fulmarus glacialis (L.). Ornis Scan-dinavica 8:33–37.

MACDONALD, M. A. 1980. The winter attendance ofFulmars at land in NE Scotland. Ornis Scandi-navica 11:23–29.

MALLORY, M. L. 2006. The Northern Fulmar(Fulmarus glacialis) in arctic Canada: ecology,

HIGH ARCTIC FULMAR BREEDING SCHEDULE 905

threats, and what it tells us about marineenvironmental conditions. Environmental Re-views 14:187–216.

MALLORY, M. L., AND A. J. FONTAINE. 2004. Keymarine habitat sites for migratory birds inNunavut and the Northwest Territories. Cana-dian Wildlife Service Occasional Paper 109.

MALLORY, M. L., AND M. R. FORBES. 2005. Sexdiscrimination and measurement bias in North-ern Fulmars (Fulmarus glacialis) from theCanadian Arctic. Ardea 93:25–36.

MALLORY, M. L., AND A. J. GASTON. 2005.Monitoring Northern Fulmars in the CanadianArctic: plot locations and counts at selectedcolonies. Canadian Wildlife Service TechnicalReport No. 432.

MCBEAN, G. 2004. Arctic climate: past and present,p. 21–60. In C. Symon, L. Arris, and B. Heal[EDS.], Arctic climate impact assessment. Cam-bridge University Press, New York.

MCLAREN, P. L. 1982. Spring migration and habitatuse by seabirds in eastern Lancaster Sound andwestern Baffin Bay. Arctic 35:88–111.

MONTEVECCHI, W. A., AND R. A. MYERS. 1997.Centurial and decadal oceanographic influenceson changes in Northern Gannet populations anddiets in the north-west Atlantic: implications forclimate change. ICES Journal of Marine Science54:608–614.

NATIONAL SNOW AND ICE DATA CENTER [ONLINE].2007. State of the cryosphere. National Snowand Ice Data Center, University of Colorado,CO, ,http://nsidc.org/sotc/sea_ice.html. (10May 2007).

OLLASON, J. C., AND G. M. DUNNET. 1978. Age,experience and other factors affecting the breed-ing success of the Fulmar Fulmarus glacialis inOrkney. Journal of Animal Ecology 47:961–976.

PERRINS, C. M. 1970. The timing of birds’ breedingseasons. Ibis 112:242–255.

RAYMONT, J. E. G. 1976. Plankton and productivityin the oceans. Pergamon Press, Toronto, Ca-nada.

RICHDALE, L. E. 1963. Biology of the SootyShearwater Puffinus griseus. Proceedings of theZoological Society of London 141:1–117.

SALOMONSEN, F. 1965. The geographical variation ofthe Fulmar (Fulmarus glacialis) and the zones ofmarine environment in the North Atlantic. Auk82:327–355.

SCHREIBER, E. A. 2002. Climate and weather effectson seabirds, p. 179–216. In E. A. Schreiber and J.Burger [EDS.], Biology of marine birds. CRCPress, New York.

SERVENTY, D. L. 1963. Egg laying time-table of theSlender-billed Shearwater, Puffinus tenuirostris.Proceedings of the XIII International Ornitho-logical Congress 1:338–343.

SLATER, C. A. 1990. First arrival dates at two FulmarFulmarus glacialis colonies in Norfolk. BirdStudy 37:1–4.

STENSETH, N. C., AND A. MYSTERUD. 2002. Climate,changing phenology, and other life history traits:nonlinearity and match-mismatch to the envi-ronment. Proceedings of the National Academyof Sciences USA 99:13379–13381.

WARHAM, J. 1990. The petrels – their ecology andbreeding systems. Academic Press, London.

WARHAM, J. 1996. The behavior, population biologyand physiology of the petrels. Academic Press,London.

WEIMERSKIRCH, H., O. CHASTEL, Y. CHEREL, J.-A.HENDEN, AND T. TVERAA. 2001. Nest atten-dance and foraging movements of NorthernFulmars rearing chicks at Bjørnøya BarentsSea. Polar Biology 24:83–88.

WELCH, H. E., M. A. BERGMANN, AND T. D.SIFERD. 1992. Energy flow through the marineecosystem of the Lancaster Sound region, arcticCanada. Arctic 45:343–357.

906 MARK L. MALLORY AND MARK R. FORBES