Daily Activity Patterns of South Polar and Brown Skuas Near Palmer Station, Antarctica

DAILY ACTIVITY PATTERNS OF SOUTH POLAR AND BROWN SKUAS NEAR PALMER STATION, ANTARCTICA

PAMELA J. PIETZ Departmentof Ecologyand BehavioralBiology, University of Minnesota, Minneapolis,Minnesota 55455 USA

ABSTRACT.--Iconducted a behavioral study of sympatricallynesting South Polar (Cathar- acta maccormicki)and Brown skuas(C. lonnbergi)near Palmer Station, Antarctica. A total of 4,058 bird-hours of observationswas made on eight South Polar, three Brown, and one mixed-speciespair during the 1979-1980 and 1980-1981 austral summers.I used subsetsof these data to analyze various aspectsof skua activity patterns. SouthPolar Skuas exhibited maximum resting and minimumforaging activity during the twilight period around 2400. Brown Skuasappeared to rest and forage more randomly. Membersof the mixed pair exhibited patternssimilar to those of their respectivespecies. Severalhypotheses are suggestedto accountfor speciesdifferences in activitypatterns; most of theserelate dietary differencesand differential foraging abilitiesat low light levels. In both species,agonistic and preening activitiesoccurred at all hours. South Polar Skuas, however, were most often observedbathing in early afternoonand preening in afternoon and twilight. Pairing activity showed no discerniblerelationship with time. Foragingbouts averaged 13 min for Brown Skuaswith feeding territoriesand about 1 h for thosewithout. SouthPolar Skuasaveraged feeding trips of 2-3 h when the seawas open and over 7 h during heavy ice cover. The timing of an individual's activity was correlatednegatively with that of its mate during incubation, brooding, and postbrooding.A nonbreeding pair and breedersthat had sufferednest failure showed more positive correlations.Thus, individual activity patterns were shapedby reproductivepriorities as well as by feeding ecologyand light levels. Re- ceived13 March 1985, accepted18 March 1986.

AN extensive literature exists on the relation- Brown skuas (C. lonnbergi)• where they nest ship between daily activity patterns of birds sympatrically and occasionally form mixed and various light regimes (see reviews by As- pairs. Investigations of allopatric populations choff 1967, Daan and Aschoff 1975, Gwinner suggestthat the activitypatterns of thesespecies 1975, Rusak 1981). Nevertheless, little research are very different. Spellerberg(1969) noted that hasbeen done on activity patternsof birds un- South Polar Skuas on Ross Island (77ø33'S, der natural conditions of continuous light, as 166ø09'E),the southern extreme of their breed- occursin polar regions in summer. Three de- ing range, showeda 24-h cycle of activity even cades ago, Armstrong (1954) noted that such during continuous daylight: "Between 01:00 information from arctic regions was meagre, and 05:00 ... all breeding pairs are roosting, while that from the Antarctic was too scarce to incubating, or brooding and little or no forag- warrant discussion. Since then, antarctic re- ing takes place." On the other hand, Young searchershave examined activity patterns of (1978) documented Brown Skuas feeding reg- Ad•lie Penguins (Pygoscelisadeliae; Muller- ularly during hours of darkness on the Chat- Schwarze1968), Blue-eyed Shags (Phalacrocorax ham Islands, New Zealand (44ø22'S,176ø11'W). atriceps; Bernstein and Maxson 1984), and STUDY AREA AND METHODS Southern Black-backed Gulls (Larus dominican- us; Maxson and Bernstein 1984, Fraser pets. This study was conductedduring the 1979-1980 comm.). Literature on skuas in Antarctica con- and 1980-1981 austral summers on several small is- tains only an incidental comment on timing of activity (Spellerberg 1969). •Pending resolution of their taxonomic status, I At Palmer Station (64ø46'S,64ø03'W) on An- refer to these as separatespecies. For discussionsof vers Island, Antarctica,I had an opportunity to skua systematicsand nomenclature,see Pietz (1984, observeSouth Polar (Catharactamaccormicki) and 1985).

726 The Auk 103: 726-736. October 1986 October1986] SkuaDaily Activity Patterns 727 lands and peninsulaswithin 5 km of Palmer Station, that had been pooledover the 24 h. Finally, weighted Antarctica.The ice-freeareas provided nesting sites hourly meanswere used to calculateangular-linear for nearlya dozenbird species(Parmelee et al. 1977), correlations(Johnson and Wehrly 1977), in a test for includingSouth Polar Skuas(approx. 250 pairs)and 24-h patternsrelated to either sine or cosinefunc- Brown Skuas(up to 12 pairs). From late November tions. to late January there is enough light at Palmer to Agonistic,preening, bathing, and pairing behav- conductfieldwork 24 hours a day, despitethe fact iors occurredtoo rarely to examine variation among that the sun dips below the horizon for a short pe- individuals or pairs. For each of the first three of riod around 2400. thesebehavioral groups,data on all individuals from Skuaswere color-bandedto allow recognitionof all observationdays were combinedby hour within individuals at a distance. Most of these birds had species.For pairing behaviors,only datafor the male been banded in previousyears and thus had docu- of each pair were combined. mented breeding histories. Sexesof birds were de- Dawn-to-dusk observations from 1980 (1,406 bird- terminedby copulatoryposition and courtshipfeed- hours)and the shorterobservation periods of 1981 ing. (805 bird-hours) were included only in calculating Skua feeding habitsalso had been documentedat lengthsof foragingtrips (or time away from the nest Palmer in previousyears (Parmeleeet al. 1978, Neil- territory) and in evaluating correlations of activity son 1983). Brown Skuashere, as in many other areas between pair members. (e.g. Stonehouse1956, Moors 1980, Sinclair 1980, Tri- Activity correlations between pair members were velpieceet al. 1980),concentrated on penguin eggs assessedfor 3,860 bird-hours of observations on 8 and chicksand sometimesdefended penguin colo- South Polar, 3 Brown, and I mixed skua pairs over niesnear their nests as feeding territories. Some South two seasons.Behavioral data were sampledfrom the PolarSkuas feed on penguinsin areaswhere Brown time budgetsonce every 30 min, at the end of each Skuasare few (Parmeleepets. comm.)or absent(e.g. 30-min interval. For each pair at each stage of the Eklund 1961, Le Motvan et al. 1967,Trillmich 1978). breeding cycle, I constructeda 2 x 2 contingency At Palmer, however, most South Polars fed at sea on tableand tabulatedpresence or absenceof a specified fishand krill, a commonhabit in muchof their range behavior by the male and the female. Product mo- (Eklund 1961,Young 1963,Trivelpiece and Volkman ment correlation coefficients and their standard errors 1982). were calculatedfor each table (Bishopet al. 1975); Between8 December1979 and 22 January1980, we the weighted means of these coefficients,with their documentedthe behavior of 4 SouthPolar Skua pairs, associatederror terms, were calculatedacross pairs 2 BrownSkua pairs, and I mixed-speciespair. A total for each species. of 50 bird-daysof observationswas obtainedfor South For the purposesof these analyses,the breeding Polar Skuas, 14 for Brown Skuas, and 8 for the mixed seasonwas divided into (1) pre-egg-laying,(2) incu- pair. All were observedfor 24-h periods. bation (from the time the first egg was laid), (3) Between27 Januaryand 9 March 1980, when in- brooding (from the hatching of the first chick until creasingdarkness prevented 24-h observations,we it wasno longer regularlybrooded), (4) postbrooding watchedeach of the 7 pairs for 6 dawn-to-duskpe- (until the time the first chick flew well enough to riods. From 7 Januaryto 6 March 1981, I made ad- avoid capture),and (5) fledging. A sixth category ditional observationsusing sampleperiods of 3-14 (failed-nest)contained observations on pairs follow- h. EightSouth Polar Skua pairs, 2 BrownSkua pairs, ing lossof eggs or chicks. and I mixed pair constitutedthe samplein the sec- Of the 4 South Polar Skua pairs observed in ond season.The mixed pair, 1 of the Brown, and 4 1979-1980, 1 courted and defended a territory but of the SouthPolar pairs were observedin both years. failed to produceeggs. For purposesof comparison During all observationperiods, the behaviorof each with the three breedingpairs, this pair'sseason was bird wasnoted every 30 s usinga metronome(Wiens dividedinto sectionsreflecting the averagetiming of et al. 1970).From these observations hourly and daily incubationand brooding observedin the breeders. time budgetswere constructedfor each individual. In several cases, related behaviors were combined Only data collectedduring 24-h observationpe- beforeanalysis. Grouped behaviors included resting riodsof the 1979-1980season (1,691 bird-hours) were (i.e. sitting, incubating,brooding), foraging (i.e. pro- usedfor assessingrelationships between activities and curing, transportingand eating prey, activity out of time of day. For restingand foraging behaviors,the view off the nestterritory), pairing (i.e. courtshipand meanand standarderror were calculatedby hour for courtship feeding, copulation, attempted copula- eachindividual and for eachpair. Weightedhourly tion), agonisficdisplays (i.e. long call, alarm call, bent meanswere determinedfor both speciesusing pair neck, wing raising), and agonistic encounters(i.e. data,and for eachsex of both speciesusing individ- running or flying displacement,swooping, aerial ual data. Mean hourly standarderror estimateswere chasing,fighting). weighted acrossindividual or pair standarderrors Assessmentof foragingbehavior presented special 728 P^MELAJ. PIETZ [Auk, Vol. 103

SOUTH POLAR SKUAS RESTING

m FORAGING •] Brown Skuas [] South Polar Skuas I-

MIXED-SPECIES PAIR

0 4 8 12 16 20 0 4 8 12 16 20 24 TIME OF DAY (HOURS) Fig. 2. Daily patternsof restingand foraging.Band width indicates _+i SE of hourly means weighted acrosspairs. The 24-h pattern is repeated to facilitate Fig. i. Daily foraging activity of skuapairs. Each imaging. row representsone individual during one 24-h ob- servationperiod. Horizontal bars denote time spent away from the nestarea; solid barsindicate evidence of foraging (seeMethods) and open bars indicate un- There was no hour in which all birds were al- known location off territory. ways foraging or always at rest. Despite this variation, the activity of South Polar $kuas exhibited a significantcorrelation problems. Direct observationscould be made in the vicinity of each pair's territory, but foraging at sea with time of day (Table 1). On average,both or at penguin colonieson other islandscould only breedingand nonbreedingpairs showeda short be inferred from flight direction or regurgitation of peak in resting behaviors around 2400 (Fig. 2); food when the bird returned. Periods when the birds foraging activity moved from lowest to highest were off their territories and out of view were divid- levels between 0000 and 0500, then gradually ed into those for which there was evidence of for- declined toward midnight. aging (from flight direction or food regurgitation) Restingand foraging of Brown Skuaswas not and those for which there was not. While both types correlatedwith time of day (Table 1). An abrupt of absenceswere includedin total foragingtime, they peak in resting levels at 2300-2400 (Fig. 2) ap- were distinguishedin the representationof daily for- peared as an outlier in relation to the other 23 aging activity (Fig. 1). In calculating lengths of for- h. Although activity seemedto fluctuate more aging bouts,absences of the latter type were exclud- ed. Courtship feedings were included in Fig. 1 sharply than for South Polar Skuas,this prob- because,for much of the early season,this constitut- ably was due to the smaller Brown Skua sample ed the females' primary means of food intake. These size. feedingswere not included in measuresof foraging Male and female South Polar Skuas both bouts. showedthe tendenciesdescribed for the species as a whole, but the pattern was much stronger RESULTS in males (Fig. 3A, Table 1). The lack of a significant correlation between resting and time Restingand foragingbehaviors comprised the of day in femalesreflected their propensityto two largest componentsof skua time budgets. stay on the territory during pre-egg-laying and The timing of these behaviors varied greatly early incubation, while males procured food at from day to day for each individual (Fig. 1). sea for both of them. Because most of the 24-h October1986] SkuaDaily Activity Patterns 729

TABLE1. Angular-linearcorrelations (Johnson and Wehrly 1977) between activity and time of dayfor South Polar and Brown skuas near Palmer Station, Antarctica. a

Resting Foraging Preening Bathing Agonistic Pairing SouthPolar Skua pairs 0.67** 0.83** 0.57* 0.62* 0.34 0.19 Males 0.71 * * 0.77* * _b _ _ _ Females 0.46 0.67* * .... BrownSkua pairs 0.35 0.39 0.42 0.13 0.46 -- Males 0.36 0.43 .... Females 0.07 0.12 .... Mixed-speciespair South Polar Skua male 0.83'* 0.95'* .... Brown Skua female 0.41 0.30 .... ß = p < 0.05 for F with 2 and 21 df; ** = P < 0.005, for F with 2 and 21 df. Not calculated.

observationsof South Polars occurredearly in feeding territoriesadjacent to their nestsfor- the breeding season,patterns of this period agedfor lessthan 13 min per bout, while birds dominated the data set. Nevertheless, females without adjacentfeeding territories averaged showed an abrupt resting peak at 2300-2400, over 60 min (t = -4.58, df = 2, one-tailed P < like that described for Brown Skuas. 0.025).In Fig. 1 thesetwo conditionsare illus- In Brown Skuas,activity of both sexesshowed tratedby BrownSkua pairs 1 and 2, respective- the main features noted for the species(Fig. ly. Similarly, Nellson (1983) documentedav- 3B), but a sex-related difference in pattern eragebout lengthsof 8-13 min for a pair with strength was not apparent. Although some an adjacentfeeding territory and 16-21 min for Brown Skua females may fit the South Polar a pair feeding about 1 km away. trend, the evidence was inconclusive because After all the penguin fledglingshad depart- relatively few observationswere madeearly in ed (late February),differences in the length of the Brown Skua breeding cycle. In an early- foragingbouts between Brown and SouthPo- seasonstudy on nearby King George Island lar skuasdecreased substantially. Brown Skuas (62ø10'S, 58ø30'W), however, Brown Skua fe- that had maintained feeding territories still males spent more time at the nest site and less made short scavengingtrips to the deserted time foraging than males (Trivelpiece et al. penguincolonies (weighted œ = 16.2min, SE = 1980). 2.05, n = 3 pairs;21 bouts).In addition, like the Members of the mixed pair showed tenden- other skua groups,they now made trips away cies of their respectivespecies and sexes(Fig. from their territories that averaged over 3 h 3C, Table 1). The South Polar Skua male exhib- (weighted œ = 180.3 min, SE = 28.42, n = 3 ited strongtime-related patterns, with foraging pairs; 13 bouts). Some observationssuggested highs and resting lows in midmorning. Activ- that thesebirds searchedfor penguins on more ity of the BrownSkua female was not correlat- distant islands. In addition, seaward flight di- ed with time, and her levels of foraging were rectionand subsequentregurgitation of fish to generallylower than thoseof her mate. chicksindicated some foraging at sea.The aver- Foragingbouts, or trips away from the nest ageduration of theselonger trips wasan under- territory, were typically shorterfor Brown than estimatebecause of the inclusionof severaltrips for SouthPolar skuas (Table 2, Fig. 1) (t = -7.83, that started before or ended after the observa- df = 6, one-tailed P < 0.00025). This difference tion period. Nevertheless,the long trips sig- also was observed in earlier seasons. In 1977- nificantlyincreased average bout length in this 1978 Nellson (1983) recorded Brown Skua for- period (on log-transformeddata, paired t = agingbouts that averagedless than 20 min, and -8.89, df = 3, one-tailed P < 0.0025). The Brown in 1976-1977 South Polar bouts that averaged Skua pair without a feeding territory also 1-2 h. showedan increasein bout length (t = -2.64, Further distinctionsoccurred within species. approx.df = 6, one-tailedP < 0.025). During most of the season,Brown Skuaswith For South Polar Skuas,foraging-trip length 730 PAMELAJ.PIETZ [Auk,Vol. 103

RESTING FORAGING A. SOUTH POLAR SKUAS 100.

B. BROWN SKUAS 10050 C. MIXED-SPECIES PAIR

lOO

o 8 16 o 8 16 24 o 8 16 o 8 16 24 TIME OF DAY (Hours)

!•[] Males [] Females

Fig.3. Daily patternsof restingand foragingfor malesand females.Band width indicates+ ! SEof hourly meansweighted acrossindividuals. The 24-h pattern is repeatedto facilitateimaging.

reflectedthe variable availability of their food a period of calm weather and 112 min in a pe- resources.Early in the 1979-1980 season,for- riod of late-season storms (Neilson 1983). The aging at sea was extremely difficult due to ex- impact of extensive ice cover was most dra- tensive pack-ice cover (Pietz 1984), and trips matic in 1977-1978, when birds were some- away from the nest area averaged 7.6 h. Later, times absent from their territories for days, re- when local waters were relatively clear of ice, suiting in a total reproductivefailure (Parmelee foraging trips averaged2.4 h (paired t = 3.57, et al. 1978). df = 3, one-tailed P < 0.025). Members of the mixed-species pair showed Early in the 1980-1981season, when nest ini- bout lengths and diets characteristicof their tiation dates, nest densities, and clutch sizes respective species. The Brown Skua female, indicated that food was more plentiful (Pietz foraging primarily on penguinsat neighboring 1984), average trip lengths were shorter than islands, averaged trips of less than 1 h from in the previous year. Becauseof short obser- mid-Decemberto late February,and over 3 h vation periods in the second year, accuratees- afterwards(t = -3.57, approx. df = 4, one-tailed timates of mean trip length could not be cal- P < 0.025). The South Polar Skua male, forag- culated;however, 25% of the trips were under ing at sea, averaged trips of about 4 h during 1 h during 1980-1981, compared with 12% in the period of extensive ice cover and 2.8 h dur- 1979-1980. ing periods of open water (t = 2.4, df = 23, one- Previous investigators noted the impact of tailed P < 0.025). ice and storms on the foraging efficiency of Finally, female Brown Skuasaveraged long- South Polar Skuas.In 1976-1977 foraging bouts er foraging trips than males (on log-trans- of three South Polar pairs averaged 58 min in formed data, paired t = 4.33, df = 3, two-tailed October1986] SkuaDaily Activity Patterns 731

TABLE2. Average duration of foraging bouts for A. AGONISTIC DISPLAYS (•) South Polar and Brown skuas near Palmer Station, AND ENCOUNTERS (,1•) Antarctica.

No. No. œa SE of of (min) (min) pairs boutsø Brown $kuas Overall 32.3 11.52 4 c 355 Pairswith feeding ,7," territories Penguinsnesting 12.7 0.74 3a 277 • 2 Penguins gone 76.7 17.71 3a 34 Pair without feeding B. PREENING (1•) territory AND BATHING (F'J) Penguinsnesting 61.0 10.52 1 38 Penguinsgone 239.1 66.73 1 6 South Polar Skuas

Overall 198.0 17.43 4 123 4 Heavy ice cover 453.2 77.04 4 24 Ice out 145.1 12.72 4 99 Mixed-speciespair Brown Skua female •- 0 Penguinsnesting 52.8 7.48 1 21 Penguinsgone 216.8 45.28 1 4 South Polar Skua male Overall 185.0 15.11 1 25 Heavy ice cover 244.1 36.12 1 6 .• 4 Ice out 166.4 14.33 1 19 aWeighted means given where sample includes more than one pair. C. COURTSHIP (El) bSummed for all individuals in sample. AND COPULATION ([]) cTwo pairs for 1 yr and one pair for 2 yr. -I 1.0- a One pair for 2 yr and one pair for 1 yr. o :Z: 0.5,

P < 0.05), but male and female South Polar UJ 0 Skuas showed no consistent differences in length of foraging trips (on log-transformed data, paired t = -1.32, df = 3, two-tailed P > 0.5 0.20). This difference may hinge on the sexual size dimorphism that was clearly visible in Brown Skuas (females are larger) but difficult to detect in South Polar Skuas. 0 4 8 12 18 20 24 Agonistic displaysand encountersoccurred TIME OF DAY (Hours) during nearly all hours (Figs. 4A and 5A). No Fig. 4. Timing of agonistic (A), preening and significant correlations were found between bathing (B), and pairing (C) activitiesof 8 South Po- levels of agonisticactivity and time of day (Ta- lar Skuas.In A, bar height representshourly counts ble 1). There was some indication of a lull in of agonisticbehaviors, averaged across50 bird-days agonistic activities during twilight, probably of observations.In B and C, bar height representsthe becausethe fewest potential intruders were ac- percentage of each hour allocated to an activity, tive then (Fig. 2). averagedacross 50 (B) or 25 (C) bird-days. Episodesof preening activity occurredat all hoursin both species,with no decline near 2400 constrainedthe timing of preening as well as (Figs. 4B and 5B). Somelulls in preening activ- resting. ity coincided with peaks in foraging, suggest- Among South Polar Skuas,bouts of bathing ing that time spent off territory occasionally activity showedhighest levels in midafternoon 732 PAMELAJ. PIETZ [Auk, Vol. 103

A. AGONISTIC DISPLAYS (El) In skuas,as in many seabirds,pair members AND ENCOUNTERS ([]) share the tasks of incubation, brooding, and guarding of young. When one bird leaves,the other must remain until its mate returns or risk losing the eggsor chicks.The activity of male and female skuas appeared to be coordinated in this way. It was moststriking in Brown Skuas (Fig. 3B), where a peak in the activity of one sexfrequently concurredwith lowered activity in the other, but was only weakly visible in South Polar Skuas(Fig. 3A). Presumably,with a larger sample of observationdays and observed pairs, these compensatory patterns

I would average out and disappear. Then only the time-related patterns of each species and sex would be apparent. Direct examination of activity within pairs õ showedthat the nature and strength of within- pair correlationsdepended primarily on the stageof breeding (Fig. 6). Most sample correlations were weakly positive during pre-egg- laying and negativeduring incubation,brooding, and postbrooding.Pairs that lost eggs or chicks again showed more positive correlations. For both species,the strongestnegative correlations between activity and breeding stage occurredin resting behavior during incubation and brooding. This reflectedthe dominance of I I I I I I I I I I I I I 0 4 8 12 16 20 24 incubating and brooding in the time budgets of the skuas at these stages(recall that these TIME OF DAY (Hours) two behaviorswere part of the resting category Fig. 5. Timing of agonistic(A) and preening and in the analyses).Members of a nonbreeding bathing (B) activitiesof 4 Brown Skuas.In A, bar South Polar pair generally showedweak posi- height representshourly countsof agonisticbehav- tive correlationsall season,further indicating iors,averaged across 14 bird-daysof observations.In the importance of reproduction to the timing B, bar height representsthe percentageof eachhour of activity between mates. allocatedto that activity,averaged across 14 bird-days. The importanceof nest attendancefor breeding skuas was demonstratedby how seldom both pair members were off the nest territory. For South Polar pairs whose eggsor chickssur- and a sustainedlow in the early morning (Fig. vived, simultaneous absences occurred less than 4B). Brown Skua bathing patterns could not be 0.02% of the observed time during incubation assessedbecause observed bathing was very rare (284 h, 6 pairs), 0.07% during the brooding pe- as well as episodic. The six hours in which riod (112 h, 5 pairs), and about 4% during post- bathingactivity appeared (Fig. 5B) represented brooding (182 h, 4 pairs). By contrast,simulta- only six episodesof bathing. neous absencesrose to 13% (92 h, 6 pairs) for Pairing behaviorsof South Polar Skuas,al- pairs that sufferedtotal nest failure, and aver- though among the rarest activitiesobserved, aged 14%(169 h) in the nonbreedingpair. occurredrandomly (Fig. 4C, Table 1). Brown Skua pairing could not be evaluated because DISCUSSION the pre-egg-laying period, when most courtship and mating occurred,was poorly repre- During this study near Palmer Station, sented in the sample. breeding South Polar Skuasdid not show an October1986] SkuaDaily Activity Patterns 733

A. RESTING B. FORAGING

Z *0.4

*0.2

O o

-0.6

o -o.8 -0.8

Layingtion roodngO ro odln•l . _Ne._st_ 3

;[] BreedingSouth Polar Skua Pairs [] Mixed-SpeciesPair [] NonbreedingSouth Polar Skua Pair [] BrownSkua Pairs

Fig. 6. Correlationsbetween pair membersin resting (A) and foraging (B) skuas.Band width indicates + 1 SE of the meansweighted acrosspairs within eachgroup. The "failed nest" category,shown discontin- uously from the rest of the breeding chronology,represents 6 pairs of South Polar Skuas.

activity lull from 0100 to 0500 as reported for (2) lower air temperaturesfreeze water on birds at Cape Royds, RossIsland (Spellerberg feathersafter birds splash-divefor fish;(3) low- 1969). Instead, averaged data for resting and er anglesof incidenceof light rays,which cause foraging indicated a period of relative quies- more reflection and less penetration of the cence around 2400. In the absence of darkness, water surface, make it more difficult for skuas one might expectskuas to be equally active at to seeprey; (4) prey show behavioralchanges all hours. Instead, South Polar Skuas at Palmer related to light (e.g. vertical movements)that showeda strongrelationship between foraging make them less available to skuas at certain activity and time of day that Brown Skuasdid times of day; or (5) light cuestrigger an intrin- not. sic activity "clock" that may have evolved in The differencesin activity patterns between responseto some other, more strongly rhyth- speciesmost likely stem from differences in mic environment. their feeding ecology.Searching for fish at sea The firsthypothesis was suggested by the ob- presentsproblems for South Polar Skuasthat servation(Spellerberg 1969) that the lowestat- Brown Skuas,feeding on land, do not face. mospherictemperatures at Cape Royds coin- These problems may relate, directly or indi- cided with a 0100-0500 period of skua rectly, to diel changesin light intensity. Even inactivity. Atmospherictemperatures from No- during continuousdaylight, light intensities vember to March ranged between -15øC and may shift markedly; at RossIsland, as much as 5øC. At Palmer, however, summer tempera- 8-fold changesin irradiancewere measuredbe- tures rarely fell below -2øC, and at warmer tween 1200 and 2400 under various degreesof times incubating skuas sometimespanted to cloud cover (Steven Kottmeier pers. comm.). At dissipateexcess body heat.The high frequency Palmer, where some twilight occurs even in of early morning activity in individualsof both midsummer, the range of light intensities was species(Fig. 1) makesit unlikely that heat con- greater. Gallardo and Piezzi (1973) recorded servation was a primary concern for Palmer January light levels near Palmer (64ø53'S, skuas. 62ø53'W) ranging from 30,000-50,000 lux at The secondhypothesis also involves temper- noon to less than 1,000 lux at 2200 when twi- ature as a deterrent to fishing. Young (1963) light began. noted that At the lowest light levels there are several conceivabledeterrents to fishing at sea. It is The plunge into the water immersingmuch of the possiblethat (1) lower temperaturesassociated head and lower part of the body accountsfor the with lower light require birds to conserveheat; ice massesthat encrustedthe upper mandible and 734 PAMELAJ. PIETZ [Auk, Vol. 103

forehead and the jugulum and breast feathers of perba;Permitin 1970,Karl-Hermann Kock pers. many birds. This icing was most commonly ob- comm.).Krill, a secondaryfood sourceof the served during early and late summer when air skuas(Pietz 1984) and a major food of Pleura- temperatureswere lower. Somebirds had so much gramma(Dewitt and Hopkins 1977), migrate ice about the beak and forehead when they returned to the territory that vision was seriously vertically (e.g. Witek et al. 1981). In most areas, impaired. krill aggregatein upper water layers at night and disperseto deeper water in the day. This suggestsa reasonfor Cape Roydsskuas to Such movements would indicate that skuas avoid fishing during the coldest hours of the at Palmer feed when krill are least available. day. At Palmer,where the lowesttemperatures The situation is probably more complicated, were often above freezing, I never saw birds however. First, diel krill migrations may be with ice on their feathers. Nevertheless, the fact much less important when there is continuous that recordsof bathing were rare between 2400 light. Witek et al. (1981) observed no distinct and 0400, and peaked from 1400 to 1600, sug- dependenceof the vertical krill distributionon geststhat skuasmay have avoided getting wet the intensity of daylight. In addition, Pavlov during the early hours of the day. (1974) observeda secondsurface migration at The third hypothesis relates light intensity noon, and Mauchline (1981) noted several and sun position to the visibility of prey below studiesthat suggestkrill move irregularly, per- the water surface.This may be more important hapsreacting to varying concentrationsof phy- at Palmer than RossIsland, becausethe height toplankton. of the midsummer sun ranges much more The fifth hypothesisrelates light intensity to widely at 65øSthan at 78øS.As sun position skua activity and implies that proximal causes moves toward the horizon, the proportion of now may be missingbut that an internal clock light reflected from the water's surface in- still respondsto changing light cues. Experi- creasesand the proportion penetrating the mental evidence of circadianrhythms in skuas water decreases. If the water surface is dis- is lacking; however, 24-h rhythms in their turbed, reflection at low sun angles is even physiologyand behaviormay reflectthe influ- greater (Hutchinson 1957). Coupled with low- encesof an internal clock.At Cape Royds,Spel- er light intensities at this time of day, these lerberg (1969) found that the core body tem- effectsmay severelyreduce skuas'ability to see peratureof an adult male skuapeaked at 42.4øC underwater prey. The high proportion of at midday and fell to 41.0øCat 0200-0400, im- cloudydays at Palmer[mean of about22 cloudy plying a daily cycle.On the other hand, Eklund and 8 parfly cloudy days/month in December- (1942) measured body temperatures of two January of 1979-1980 and 1980-1981 (Anon. South Polar Skuas on the Antarctic Peninsula 1980, 1981)] may reduce the importance of re- (68ø11'S,67ø12'W) and recorded lows of 39.9øC flection off the water, but increasethe impor- in the hoursjust before midnight. Eklund men- tance of light intensity for seeing prey. tioned that the South Polar Skuas observed The feeding habits of South Polar and Brown during continual daylight were "more or less skuasoffer circumstantialsupport for this hy- active for 24-hour periods, and ... could al- pothesis.Penguin colonies provide a closerand ways be observedeating and fighting among more predictablefood sourcethan marine prey, themselvesover seal-meatscraps near the husky allowing Brown Skuas to find their food with sledge-dogkennels." lesssearching and, probably, with less light. These findings, together with individual The fourth hypothesisimplicates diel changes variability noted at Palmer, suggestthat the ac- in marine prey availability; prey are available tivity patterns of South Polar Skuas are not to skuasonly when they occurwithin ! m of controlled by one 24-h pattern. A light-syn- the water surface.The primary food sourcefor chronized internal clock may be involved in South Polar Skuas at Palmer is the nototheniid the timing of activity, but use of this clock fish Pleuragrammaantarcticum (Nellson 1983, may depend on local foraging conditionsand Pietz 1984).Its movementsin responseto light, opportunities.Brown Skuasdisplay this oppor- food, and other environmental factors are not tunistic flexibility to extremes.Brown Skuason well known, but other notothenoid fishes mi- the Chatham Islands,New Zealand, forage pri- grate vertically to feed on krill (Euphausiasu- marily during hours of darkness(Young 1978). October1986] SkuaDaily Activity Patterns 735

This allows them to prey on locally abundant, DEWITT,H. H., & T. L. HOPKINS.1977. Aspectsof night-activepetrels, which they catchon the the diet of the antarcticsilverfish, Pleuragramrna antarcticurn.Pp. 557-567in Adaptationswithin groundnear the petrels'burrows. Similarly, antarcticecosystems: proceedings of the third BrownSkuas on GoughIsland in the SouthAt- SCARsymposium on antarcticbiology, August lantic (40•S, 10øW)feed primarily at night on 1974 (G. A. Llano, Ed.). Houston,Gulf Publ. Co. Procellariiformes(Furness pers. comm.). EICLUND,C. R. 1942. Bodytemperatures of antarctic I believe that light levels ultimately influ- birds. Auk 59: 544-548. ence the activity of South Polar Skuas, al- --. 1961. Distribution and life history studies though the proximatemechanism may vary. of the South-Polar Skua. Bird-Banding 32: 187- The effectof light on air temperaturemay cause 223. the activity lull at 0100-0500on RossIsland GALLARDO,M. G., & R. S. P1EZZI. 1973. Serotonin (hypotheses1 and 2), while the effectof light contentin the pineal gland of the antarcticpen- on prey visibility may producethe midnight guin (Pygoscelispapua). Gert. Comp. Endocrinol. 21: 468-471. low at Palmer (hypothesis3). Alternatively, GWINNER,E. 1975. Circadianand circannualrhythms light may trigger an internal clock that pro- in birds.Pp. 221-285 in Avian biology,vol. V (D. ducesboth patterns.Diet is a proximatefactor S. Farrier,J. R. King, and K. C. Parkes,Eds.). New influencingSouth Polar and Brownskua activ- York, Academic Press. ity patternsat a local level. For both species, HUTCHINSON,G.E. 1957. A treatise on limnology. the largevariance in the timingof individuals' Vol. I, Geography,physics, and chemistry.New activitiessuggests a behavioralflexibility suit- York, John Wiley & Sons. ed to an opportunisticlife style. JOHNSONßR. A., &:T. WEHRLY.1977. Measuresand modelsfor angular-linearcorrelation. J. Royal ACKNOWLEDGMENTS Star. Soc., Ser. B 39: 222-229. LE MORVAN, P., J. L. MOUGIN, & J. PREVOST.1967. I am grateful to George-AnnMaxson for invalu- Ecologiedu Skua antarctique(Stercorarius skua able assistance in data collection, and to the numer- maccormicki)dans l'archipel de Pointe Geologie ous scientistsand station personnelat Palmer who (Terre Adelie). Oiseau 37: 193-220. providedfield support.Special thanks are due to D. MAUCHLINE,J. 1981. Studieson patchesof krill, Eu- F. Parmeleefor financial support,field helpßand ed- phausiasuperba Dana. BiomassHandbook No. 6. itorial advice.I gratefully acknowledgeK. Larntz for Cambridge,England, Sci. Comm. on Antarctic writing computerprograms and for adviceon statis- Res. tical analyses,and D. G. Ainley, E. C. Young, R. W. MAXSON,S. J., & N. P. BERNSTEIN.1984. Breeding Furhess,J. R. Tester, D. F. McKinney, and F. H. Barn- seasontime budgetsof the SouthernBlack-backed well for editorial comments.Funding for this project Gull in Antarctica. Condor 86: 401-409. was provided by National ScienceFoundation grant MooRs, P.J. 1980. Southern Great Skuas on Antipo- DPP77-22096 to D. F. Parmelee. des Island, New Zealand: observations on foods, breeding,and growth of chicks.Notornis 27: 133- LITERATURE CITED 146. ANONYMOUS.1980. Monthly climatesummary. Ant- MULLER-SCHWARZE,D. 1968. Circadian rhythms of arctic J. U.S. 15(1): 11. activityin the Ad•lie Penguin(Pygoscelis adelie) 1981. Monthly climate summary.Antarctic during the australsummer. Antarctic Res. Ser. J. U.S. 16(1): 15. 12: 133-149. ARMSTRONG,E.A. 1954. The behaviour of birds in NEILSON,D. R. 1983. Ecologicaland behavioral as- continuousdaylight. Ibis 96: 1-30. pectsof the sympatricbreeding of the SouthPo- ASCHOFF,J. 1967. Circadian rhythms in birds. Proc. lar Skua (Catharactamaccormicki) and the Brown 14th Intern. Ornithol. Congr.: 81-105. Skua (Catharactalonnbergi) near the Antarctic BERNSTEIN,N. P., & S. J. MAXSON. 1984. Sexually Peninsula.Unpublished M.S. thesis,Minneapo- distinctdaily activitypatterns of Blue-eyedShags lis, Univ. Minnesota. in Antarctica. Condor 86: 151-156. PARMELEE,D. F., N. BERNSTEIN,& D. R. NElLSON. 1978. BISHOPßY. M., S. E. FIENBERG,&: P. W. HOLLAND. 1975. Impact of unfavorableice conditionson bird Discrete multivariate analysis:theory and prac- productivityat PalmerStation during the 1977- tice. Cambridge,Massachusetts, MIT Press. 78 field season.Antarctic J. U.S. 13(4): 146-147. DAAN,S., & J. ASCHOFF.1975. Circadian rhythms of ßW. R. FRASER,& D. R. NEILSON. 1977. Birds 1ocomotoractivity in captive birds and mam- of the Palmer Station area. Antarctic J. U.S. 12(1- mals: their variations with season and latitude. 2): 14-21. Oecologia 18: 269-316. PAVLOV,V. YA. 1974. On the nature of relationships 736 PAMELAJ. PIETZ [Auk, Vol. 103

betweenthe feeding habitsand certainpeculiar- STONEHOUSE,B. 1956. The Brown Skua Catharacta ities of the behaviour in Euphausiasuperba Dana skualonnbergi (Mathews) of SouthGeorgia. Falk- (in Russian).Trudy vses. nauchno-issled.Inst. land Is. Dependencies Surv. Sci. Rept. 14: 1-25. morsk.ryb. Khoz. Okeanogr.99: 104-116. TRILLMICH,F. 1978. Feeding territories and breed- PERMITIN,YU. E. 1970. The consumptionof krill by ing successof SouthPolar Skuas. Auk 95: 23-33. antarcticfishes. Pp. 177-182 in Antarcticecology, TRIVELPIECE,W., R. G. BUTLER,& N.J. VOLKMAN. 1980. vol. 1 (M. W. Holdgate,Ed.). New York, Academ- Feeding territories of Brown Skuas (Catharacta ic Press. lonnbergi).Auk 97: 669-676. PIETZ,P.J. 1984. Aspectsof the behavioralecology ß & J. VOLKMAN. 1982. Feeding strategiesof of sympatricSouth Polar and Brownskuas near sympatricSouth Polar Catharacta maccormicki and Palmer Stationß Antarctic. Unpublished Ph.D. Brown skuasC. lonnbergi.Ibis 124: 50-54. dissertation,Minneapolis, Univ. Minnesota. WIENS, J. A., S. G. MARTINß W. R. HOLTHAUS, & F. A. ß 1985. Long call displaysof sympatricSouth IWEN. 1970. Metronome timing in behavioral Polar and Brown skuas. Condor 87: 316-326. ecologystudies. Ecology 51: 350-352. RUSAK,B. 1981. Vertebratebehavioral rhythms. Pp. WITEK, Z., J. KALINOWSKI, A. GRELOWSKI, & N. 183-213 in Handbook of behavioral neurobiolo- WOLNOMIEJSKI.1981. Studiesof aggregationsof gy. Vol. 4, Biologicalrhythms (J. Aschoff,Ed.). krill (Euphausiasuperba). Meeresforschung 28(4): New York, Plenum Press. 228 -243. SINCLAIRßJ. C. 1980. Subantarctic Skua Catharacta YOUNGßE.C. 1963. Feedinghabits of the SouthPo- antarcticapredation techniques on land and at lar Skua Catharacta maccormicki. Ibis 105: 301-318. sea. Cormorant 8: 3-6. 1978. Behaviouralecology of lonnbergiskuas SPELLERBERG,I.F. 1969ß Incubationtemperatures and in relation to environment on the Chatham Is- thermoregulationin the McCormick Skua. Con- landsßNew Zealand. New Zealand J. Zool. 5: 401- dor 71: 59-67. 416.

A symposiumon Birds of EvergreenForest, organized by the SouthernAfrican OrnithologicalSociety, will be held 8-10 September 1987 at The Wilderness, Cape Province, South Africa. Papersand posterswill be presentedon the following topics:forest bird communities,biogeography of forest birds, population biology of forest birds, and conservationof forest avifaunas.Prospective participants should contactthe Symposium Organising Committee, E.C.W.B.S., P.O. Box 1305, Port Elizabeth 6000, South Africa.

The Third New England Regional Hawk Conferencewill be held 4 April 1987 at the Holiday Inn, HolyokeßMassachusetts. Registration forms are availablefrom HAWKS, P.O. Box212, Portland, Connecticut 06480.There are specialrates for lodging at the Conferencecenter. Registration will be limited.

The Third World Conference on Birds of Prey will be held 22-27 March 1987 at Eilat, Israel. It will be organizedby the World Working Groupon Birdsof Prey in conjunctionwith the IsraelRaptor Information Center and the U.S. Hawk Mountain SanctuaryAssociation. The conferencewill consistof 7 paper sessions on conservation,migration, population biology, educationßand legislation.For further informationwrite to the Hon. Secretaryof the World Working Group: Mr. R. D. Chancellor, 15 Bolton Gardens,London SW5 0AL, U.K.