Oecologia (2002) 130:485Ð495 DOI 10.1007/s00442-001-0836-x
REVIEW
Gerald L. Kooyman Evolutionary and ecological aspects of some Antarctic and sub-Antarctic penguin distributions
Received: 24 May 2001 / Accepted: 27 September 2001 / Published online: 13 November 2001 © Springer-Verlag 2001
Abstract Penguins probably originated in the core of edge. Less is known about penguins during the pelagic Gondwanaland when South America, Africa, and Ant- phase between breeding cycles. What we do know is sur- arctica were just beginning to separate. As the continents prising in regard to their dispersal, which ranges from drifted apart, the division filled with what became the hundreds to thousands of kilometers from the breeding southern ocean. One of the remaining land masses colonies. moved south and was caught at the pole by the Earth’s rotation. It became incrusted with ice and is now known Keywords Aptenodytes á Eudyptes á Platform transmitter as East Antarctica. Linking it to South America was a terminal á Pygoscelis á Time-depth recorder series of submerged mountain ranges that formed a neck- lace of islands. The northern portion of the necklace, called the Scotia Arc, is now the “fertile crescent” of the Introduction Southern Ocean. The greatest numbers and biomass of penguins are found here as well as that of krill, the pri- Penguins are one of the oldest, most aquatic, and argu- mary prey species of most penguins, and many other ably most mystifying groups of birds. Their adaptation to marine predators. Today penguins are found throughout the sea was most likely a conversion from a procellariid the sub-Antarctic islands and around the entire Antarctic type of flying bird, such as a diving petrel (Simpson continent. Using satellite transmitters and time-depth re- 1976), to flying underwater, as modern penguins do. corders, while taking advantage of the parental dedica- There is no cladogram for penguins at the present time tion of breeding birds, numerous investigators have de- (Fordyce and Jones 1990). Most classification schemes scribed foraging habits of several species of penguins. are based on morphological characters. The exception is The information obtained is labor intensive and costly so the DNA-hybridization study used for the biochemical that studies are restricted to certain species, areas and classification of birds by Sibley and Ahlquist (1990). seasons. Here I review the patterns evident among six of Biochemical evidence indicates a much more detailed the most abundant and completely studied of the pen- classification with the addition of infraclasses, parvclass- guins. The variation in behavior is considerable from es and so on. The major difference from the classifica- those species that seldom dive deeper than 20 m in tion of other authors (Mayr and Amadon 1951; Cracraft search of prey to those that will dive to depths >500 m to 1981) is the placement of penguins in the order Ciconii- catch mesopelagic fish and squid. Foraging trips from formes rather than Sphenisciformes (Sibley and Ahlquist breeding colonies vary among species and with the sea- 1990). In the Sibley/Ahlquist classification, the order son. Often the birds travel no more than 30 km and at Ciconiiformes includes 1,027 species of birds. other times the trips may exceed 600 km. Sub-Antarctic Fossil evidence places the origin of penguins at least species often reach more productive waters near or with- in the Eocene (Fordyce and Jones 1990), and Simpson in the Antarctic Polar Front zone, where the mixing of suggests the origin was likely as early as the Cretaceous Antarctic and sub-Antarctic waters provide rich resourc- (Simpson 1976). The hybridization evidence agrees with es for their prey. Antarctic species usually remain close a divergence from procellarids about 47 million years to shore, along the continental slope, or near the sea ice ago (mya) (Sibley and Ahlquist 1990). All fossil locali- ties are in the present range of penguins, but conditions G.L. Kooyman (✉) have changed considerably. Over the past 50 million Scholander Hall, 0204, Scripps Institution of Oceanography, University of California, San Diego, La Jolla, CA 92093, USA years the continental land masses have shifted and frag- e-mail: [email protected] mented (Fig. 1). Presumably the birds evolved on the Fax: +1-858-5641305 edge of the continental land mass of Gondwanaland, but 486 tinent and adjacent waters became colder. The ultimate conclusion of this drift occurred when Antarctica reached its present position, South America and Austra- lia assumed their separation from Antarctica and the cir- cumpolar Southern Ocean formed and cooled (Eastman 1993; Knox 1994). Penguin diving ability enables some of them to forage at depths unavailable to other sea birds, but it is with some sacrifice. Their flightlessness has placed important restrictions on their distribution. Awkward on land, and relatively defenseless against even moderately sized terrestrial predators, their breeding habitat is restricted, with few exceptions, to cold temperate islands or isolated shorelines of the island continents of Australia and Ant- arctica. Only the Emperor Penguin, by breeding on sea ice, has separated itself completely from the terrestrial habitat. As the southern seas cooled, only two species evolved in ways to cope with high latitude conditions: the Emperor Penguin and Adélie Penguin. In large part, the other 15 species of penguins are associated with New Zealand, South Africa, Australia, southern South America, and islands of the sub-Antarctic region. This review will address penguin foraging behavior during the critical time when constrained by the need to provide food to rapidly growing chicks. During incuba- tion of the egg and nurturing of the chick, penguins share duties, which change from long periods at sea during in- cubation to shorter periods as the chick matures and its Fig. 1 Distribution of the continents about 100 million years ago food requirements increase. If the foraging trip of the (mya) when Gondwanaland was separating (top), and at about 60 mya (bottom), shortly before the earliest known fossil penguin partner is too long, the mate will abandon the egg or of 50 mya. The hatched area represents the central region where brooded chick. Even when the chick no longer needs its penguins probably evolved parent’s physical protection, an exceptionally long fast because of a delayed return by the adults will result in the weakening of the chick. If it does not starve, it may there is no evidence of the location of their origin be killed by nearby colony predators such as skuas (Fordyce and Jones 1990). Most likely penguins fre- (Catharacta spp.) or giant petrels (Macronectes spp.). To quented near-shore islands which gave them protection maintain the chick’s condition it is vital that the parents’ from any continental predators that may have existed. foraging trips to sea become increasingly successful as With few exceptions they continue to enjoy (and require) the chick matures. High food intake is necessary to this isolation. Over the course of tens of millions of maintain body condition adequate to discourage preda- years penguins differentiated into many more species tors, to enhance overall growth and feather development, than exist today. There are 17 extant species, and 32 and therefore encourage the molt of the down and ulti- known extinct species (Simpson 1976). Incidentally, this mately the chick’s independent departure to sea. ratio is similar to that of the great avian diversity loss apparent in the Pleistocene. Over the course of 2 million years of great climatic changes, the number of species Tracking penguin foraging patterns declined by half to the present-day number of more than 9,600 species (Brodkorb 1971, as cited in Proctor and Perhaps the most significant technical advance in the Lynch 1993). Some of the earlier penguin species study of the behavior of marine birds and mammals at were probably substantially larger than the present-day sea has been the ability to track their movements with giant, the Emperor Penguin (Aptenodytes forsteri), which platform transmitting terminals (PTT). These transmit- is twice the mass of its congener the King Penguin ters are tuned to a nominal frequency of 401.650 MHz. (A. patagonicus) and 30 times the mass of the smallest They range in mass from about 150 to 400 g. Much of present-day penguin, the Little Penguin (Eudyptula this mass results from the thick casting necessary to minor). withstand pressures that occur when aquatic animals Penguins spread to many islands of the southern seas, dive to depth. The linear dimensions also vary with the as well as the southern continental coastlines of South size and number of batteries needed. A typical size is America, Africa and Australia. During this time the con- about 5 cm wide by 4 cm thick and 15 cm long. These 487 Table 1 Summary of the annual cycle and foraging range during conservative. Phase durations are the underlined values that are each phase of the cycle. Body mass is the average for the male and first in columns 3Ð7. Distances in kilometers are the bold last val- female parent during the crèche stage if reported. Unknown dis- ues of columns 3Ð8. References are cited by number in column 9, tances are indicated with a question mark. Dispersal patterns are and defined in the note below the table. The time and duration of just being determined by various means and the data presented are the cycles are from Williams (1995) Species; Breeding Incubation Guard Creche Fledging Molt Winter Referencesa mass (kg) dispersal
Macaroni, Oct/Nov 35/NovÐ 25/DecÐ 35/Jan/ 60/Feb/? 25/Mar/ AprÐSep/? (1) Eudyptes Dec/400 Jan/30 50 Colony chrysolophus; 4.6, 3.9 Royal, Sep/Oct 35/OctÐ 21/Dec/ 43/JanÐ 65/Feb/? 26/FebÐMar/ AprÐSep/? (2) E. schlegeli; Nov/650 100 Feb/300 Colony- 4.5, 4.0 Antarctica King, Nov/Dec 54/Dec/ 31/Jan/ 280/FebÐ 313/DecÐ 31/JanÐFeb/ Mar-Oct/ (3) Aptenodytes 650 689 Nov/650 Jan/? colony 2,300 patagonicus; unsexed, 13 brood Chinstrap, Oct/Nov 36/Dec/? 29/Jan/? 25/Feb/? 54/FebÐ 13/FebÐMar/ Mar-Oct/ (4) Pygoscelis Mar/? colony 2,100 antarctica; 4.4, 3.9 Adélie, Oct 33/NovÐ 22/Jan/ 29/JanÐ 51/Feb/? 20/FebÐMar/ Mar-Oct/ (5) P. adeliae; Dec/340 135 Feb/30 1200 1,500 unsexed, 4.2 Emperor, Apr/May 65/MayÐ 45/JulyÐ 105/SepÐ 150/Dec/ 35/JanÐMar/ MarÐApr/ (6) A. forsteri; July/100 Aug/100 Dec/200 3,000 1,200 1,500 unsexed, 25 a (1) Barlow and Croxall, unpublished; (2) Hull et al. 1997; (3) Moore et al. 1998; (4) Wilson et al.1998; W.Trivelpiece, personal com- munication) (5) Kerry et al. 1995; Davis et al. 1996; (6) Kirkwood and Robertson 1997; Kooyman et al. 1996; Kooyman et al. 2000 dimensions limit the species of animals to which the TDR, which must be recovered when the penguin returns units can be applied. For all species, but especially the to the colony. A drawback of the STDR is that they are small, fast swimming penguins, results must be interpret- much larger, i.e. about 250 g compared to 20Ð60 g for ed with caution because of the increased drag and extra the TDR. swimming effort induced by the transmitter (Culik et al. 1994; Bannasch et al. 1994; Hull 1997). The PTT’s transmission power is usually 0.25Ð1 W. Results of various studies This signal is received by one of the U.S. National Oce- anic and Atmospheric Administration (NOAA) satellites. Following is a review of data from six species of pen- Fortunately for high-latitude studies several locations are guins (Tables 1, 2). It is clear from the original publica- possible per day since these polar orbiting satellites tions that there is considerable variability in foraging be- make 5Ð10 passes daily within reception range for PTT havior. Rather than attempt to give details of specific transmission. The number of signals received per pass years, I have in most cases given dates the data were col- influences the accuracy of the location calculations. The lected, and the range of the results. most accurate locations are within 1 km radius of the PTT. Supplementing the geographic determination of bird Sub-Antarctic species movements are measurements from time-depth recorders (TDR), often called an archival recorder because the de- One of the most intensively studied species has been the vices now measure several other variables besides pres- Macaroni Penguin (Eudyptes chrysolophus) breeding on sure and time. These recorders give a pressure profile of Bird Island of the South Georgia Island complex. Breed- the dive and hence a statistical estimate of the foraging ing commences in October/November with males return- depths, foraging time within a 24-h day, and the duration ing to colonies a week or so before females. Incubation of the dive. Other transducers incorporated into the TDR is composed of three long shifts of 10Ð20 days of which can record light intensity, water temperature and swim the male assumes the first and last shifts. In the advanced speed. Some versions (STDR) can transmit dive data to a stages of chick development the foraging trips usually satellite. The STDR overcomes the disadvantage of the range from 10 h to 48 h (Williams 1995). 488 Table 2 Variability of penguin diet, as measured in most cases during the chick phase of the breeding cycle. F Fish, K krill, S squid
Species Diet (% mass) Year Place name Location Reference
Macaroni 96Ð98 K 1977Ð1996 Bird Island, South Georgia 54.0¡S, 38.0¡W Croxall et al. 1997 Macaroni 13 K, 1994 Bird Island, South Georgia 54.0¡S, 38.0¡W Croxall et al. 1999 67 amphipod, 15 F Macaroni 87 K, 5 F 1983 Marion Is. 46.9¡S, 37.9¡E Brown and Klages 1987 Macaroni 62 K, 25 F, 13 S 1984 Marion Is. 46.9¡S, 37.9¡E Brown and Klages 1987 Royal 51 K, 45 F 1984Ð1986 Macquarie Is. 54.5¡S, 158.9¡E Hindell 1988a King >90 F 1991Ð1994 South Georgia 54.2¡S, 36.8 W Olsson and North 1997 King 87 F, 12 S, <1 K 1984, 1985 Marion Is. 46.8¡S, 37.9¡E Adams and Klages 1987 King 57 F, 32 S winter Possession Is. 46.4¡S, 51.8¡E Cherel et al. 1996 1990Ð1992 King 72 F; 28 S overall 1992 Heard Is. 53.1¡S, 73.5¡E Moore et al. 1998 41Ð100 F; 0Ð59 S seasonally King 98 F, 2 S 1984, 1985 Macquarie Is. 54.5¡S, 158.9¡E Hindell, 1988b Chinstrap 99.6 K 1977 King George Is. 62.1¡S, 58.5¡W Volkman et al. 1980 Chinstrap 68 K 1980 King George Is 62.1¡S, 58.5¡W Jablonski 1985 Chinstrap 17 K, 65 F 1981 King George Is 62.1¡S, 58.5¡W Jablonski 1985 Chinstrap 97 K, 2.9 F 1981 Signy Is. 60.7¡S, 45.6¡W Lishman 1985a Chinstrap 99.8 K 1982 Signy Is. 60.7¡S, 45.6¡W Lishman 1985a Chinstrap 99 K day; 1994 Seal Is., South Shetland Is 60.9¡S, 55.4¡W Jansen et al. 1998 74 K overnight Adélie 93 K, 1.4 F 1979 King George Is 62.1¡S, 58.5¡W Jablonski 1985 Adélie 92 K, 3.3 F 1980 King George Is 62.1¡S, 58.5¡W Jablonski 1985 Adélie 56 K, 4.4 F 1981 King George Is 62.1¡S, 58.5¡W Jablonski 1985 Adélie 60 K, 39 F 1965, 1966 Cape Crozier 77.5¡S, 169.6¡E Emison 1968 Adélie 99.6 K 1978 King George Is 62.2¡S, 58.5¡W Volkman et al. 1980 Adélie 98.3 K 1981 Signy Is. 60.7¡S, 45.6¡W Lishman 1985a Adélie 99 K 1982 Signy Is. 60.7¡S, 45.6¡W Lishman 1985a Adélie 66 K, 32 F 1983 Magnetic Is. 68.5¡S, 77.7¡E Puddicombe and Johnstone 1988 Adélie 85 K, 15 F (Jan) 1989 Hukuro Cove 69.0¡S, 39.6¡E Watanuki et al. 1993 49 K, 51 F (Feb) Adélie 86 K, 15 F (Jan) 1990 Hukuro Cove 69¡0 S, 39.6¡E Watanuki et al. 1993 2 K, 98 F (Feb) Adélie 76 K, 24 F (Jan) 1991 Hukuro Cove 69.0¡S, 39.6¡E> Watanuki et al. 1993 75 K, 25 F (Feb) Emperor 52 K, 38 F, 10 S 1986 Drescher Inlet 72.9¡S, 19.4¡W Klages 1989 Emperor 55 F, 45 S 1988 Auster 67.4¡S, 64.1¡E Robertson et al. 1994 31 F, 69 S Taylor 67.5¡S, 60.9¡E Emperor 68 K, 26 F, 5 S 1993 Auster 67.4¡S, 64.1¡E Kirkwood and Robertson (August) 10 K, 1997 20 F, 65 S (Nov) Emperor 8 K, 92 F (Nov) 1992 Cape Washington 74.7¡S, 165.4¡E Cherel and Kooyman 1998 Emperor 0.5 K, 96 F, 1993 Coulman Is. 73.3¡S, 169.6¡E Cherel and Kooyman 3 S (Nov) 1998
Preliminary results from satellite-tracking obtained in Croxall et al. 1993). The summer diet was almost exclu- 1999 indicate that during incubation both sexes traveled sively Antarctic krill (Euphausia superba) at Bird Island, 400 km or more from the colony to the north and north- south Atlantic (Table 2). Diet was much more variable at west; some tracks crossed the Antarctic Polar Front Marion Island, which is north of the APF in the Indian (APF) to reach sub-Antarctic waters (K. Barlow, unpub- Ocean. There they consumed cephalopods, lanternfish, lished data) (Fig. 2). These results show a much greater and euphausiids (Table 2). foraging range than previously assessed by shipboard in- In the closely related Royal Penguin (Eudyptes vestigators (Croxall et al. 1993; Trathan et al. 1998). Fe- schlegeli) which breeds only on Macquarie Island (north males foraging during the guard stage (chicks attended of the APF), all trips from the Sandy Bay colony were to by at least one parent at all times) rarely traveled >30 km the southeast, mostly to the north of the APF (Fig. 2). from the colony. During the crèche stage (chicks left un- Here the surface temperature ranges from 2¡C to 6¡C guarded by parents), foraging area and range were more (Hull et al. 1997). During incubation the first foraging variable and birds frequently traveled up to 50 km from trips lasted about three weeks, and the birds ranged from the colony. 250 km to 650 km from the colony (Table 1). During the Daytime diving depths were 20Ð35 m (mean 29 m) guard stage the shortest foraging trips lasted 3 days, and and nighttime dives averaged <6 m (maximum 11 m; reached about 100 km from the colony. Later, during the 489 crèche stage, trips lengthened again to >300 km and a vation) (Table 2). While diving almost continuously week in duration (Fig. 2) (Hull et al. 1997). while at sea, the depth range of the birds tracks the daily Royal Penguins dive predominantly during daylight, light cycle closely. At night the diving depths are <50 m spending most of the time hunting at depths <60 m in (Kooyman et al. 1992; Putz and Bost 1994). Based on waters 4,000Ð5,000 m deep (Hull et al. 1997). The diet stomach temperature changes the main feeding time is consists primarily of myctophid fish, dominated by not during the shallow night dives, but during the day Krefftichthys anderssoni, and euphausiids, particularly when the fish are thought to be the deepest and most in- Euphausia vallentini (Table 2). These species commonly active (Putz and Bost 1994; Bost et al. 1997; Handrich occur at the APFZ. et al. 1997). Diet analyses completed at South Georgia, The breeding cycle of the King Penguin (A. patagon- Marion, Crozet, and Macquarie islands all showed the icus) was first described by Stonehouse (1960) at South main species consumed to be lanternfish (K. anderssoni) Georgia. The species breeds on most sub-Antarctic is- (Table 2). It is noteworthy that Royal Penguins, as stated lands; the cycle begins in early spring with the moult, above, also feed in the APFZ during the day, but dive no which takes about a month to complete. Courtship and deeper than 60 m, and frequently consume the same fish egg-laying take place normally during November and species, in addition to euphausiids. December. The single chick hatches in January/February and is fed until the beginning of the winter, by which time it can be as heavy as the parent. The chick remains Antarctic species in the colony throughout the winter. During winter peri- ods of food shortage chicks mostly fast with sporadic Chinstrap Penguins (Pygoscelis antarctica) are found feedings. Regular parental feedings are resumed in the mainly in the Scotia Arc with the highest population spring and continue until fledging in December and Jan- on the South Sandwich Islands. With the exception uary. Thereafter, the parents molt and soon after may be- of the scattered and small colonies of Gentoo Penguins gin a new breeding cycle, although this will occur con- (P. papua), the Chinstrap Penguin is the most northerly siderably later than in the previous year (Stonehouse occurring of the Antarctic penguins, and some of these 1960; Williams 1995). The chick of this cycle usually colonies are very large. After arrival and laying, the in- dies and its early loss enables the adults to breed at the cubation tasks are shared by the adults in shifts more propitious time of the following November. 4Ð15 days long (Table 1). Foraging trips in 1993 and In 1985, when a small or no recording device was 1994, while raising chicks ranging from 4 to 22 h from attached, foraging trips of study birds were 6.1Ð7.1 days colonies in the South Shetland Islands (Jansen et al. at Saint Andrews Bay, South Georgia (54.4¡S, 36.2¡W). 1998). During foraging trips the diving depths averaged Similarly, in a 1989 study at Possession Island (46.4¡S, 31 m and the deepest in 1988 was 121 m (Bengtson et al. 51.8¡E), Crozet Archipelago, the cycles were 6 days 1993). during the brooding stage (Kooyman et al. 1992). From Generally the diet during chick provisioning is January to mid-April in 1992 and 1993, also during 90Ð99% krill by weight, mainly E. superba (Table 2). the brooding stage, King Penguins with PTT traveled However, in some instances fish may contribute as much from Possession Island in a southerly direction for 144Ð as 39% (Table 2). At Seal Island there are two types 689 km (Table 1., Fig. 2). The trips lasted 7.9Ð32.4 days, of foragers, those that are out only during the day and averaging 16.8 days (Jouventin et al. 1994; Guinet et al. those that stay overnight. The overnight birds take a 1997). Most of these journeys were to the APFZ between substantial amount of fish that may represent about 25% 45 and 50¡S., where the water temperatures were from by mass of the stomach contents (Jansen et al. 1998) 4¡C to 5¡C (Guinet et al. 1997). Possibly the APFZ was (Table 2). closer in 1989 when the trip durations were shorter. Adélie Penguins (Pygocelis adeliae) are distributed However, as a caution about the importance of the APF, widely around the whole of the Antarctic continent, in- Ainley and Boekelheide (1983) based on ship surveys cluding the Antarctic Peninsula and its associated islands were unimpressed by the number of birds at the APF. north to the South Sandwich Islands (Woehler 1993) Still, birds recorded in 1994 from South Georgia swam (Fig. 2). During the November/December incubation north to 50Ð53¡S, in the region of the APFZ (Fig. 2). both adults tend the two eggs during alternating shifts The cycles lasted from 11 to 30 days. The South Georgia lasting 12 days (Table 1). Throughout the guard stage (54.15¡S, 36.8¡W) birds were experiencing a seriously foraging trips last <1 to 5 days. low food year in which there was nearly 100% loss of In the Ross Sea the birds tracked with PTT in 1988 chicks. The foraging trips were also long compared to ranged up to 272 km from the colony during the incuba- previous years when some were only 2 days (Olsson, un- tion period (Davis and Miller 1992). Birds from the published observation). Apparently, device encumbrance Béchervaise Island colony (67.6°S, 62.8°E), on the and prey abundance are important factors in trip dura- Mawson Coast, tracked with PTT in 1991Ð1993 ranged tion. from 161 to 341 km (Kerry et al. 1995). During the chick The foraging dives of King Penguins range in depth brooding period the distances ranged from 2 to 135 km from 50 m to 300 m (Kooyman et al. 1992; Putz and as the birds traveled to the shelf break (Fig. 2). It is note- Bost 1994; Bost et al. 1997; Olsson unpublished obser- worthy that the foraging cycles for these birds were as 490 491 much as double the duration for birds traveling without mostly krill in the early stages of chick development, PTT. In another study in 1989 and 1991 where VHF and shifting to fish toward the end of chick nurturing. transmitters were used for tracking birds in the guard After egg laying, the Emperor Penguin male assumes stage from a colony at Hukuro Cove (69¡S, 39.6¡E) birds all incubation duties until shortly after hatching only departed for 5Ð20 h. They fed within 1Ð2 km from (Table 1). Over the next 5 months the female and male the colony in a wide variety of conditions from tidal come and go to the colony about 10 times to feed the cracks and ice holes to the ice edge (Watanuki et al. chick. Each trip away tends to get shorter, declining from 1993) (Fig. 2). a length of about 3 weeks in August to as short as 3 days In 1990 and 1991 the foraging depths of Adelie Pen- in December (Kirkwood and Robertson 1997). Around guins ranged from about 3 to 98 m and averaged 26 m summer solstice the chicks fledge by leaving the colony, near Anvers Island, Antarctic Peninsula (Chappell et al. and the parents depart about the same time to molt 1993). Similarly shallow dives averaging from 7 to 12 m (Kooyman et al. 1990). occurred over three seasons of study at Hukuro Cove, During the early chick nurturing period of August to Enderby Land (Watanuki et al. 1993). mid-September 1993, at the Auster colony, trip durations Various studies around the continent show that the averaged 11Ð18 days. The time spent swimming in- Adélie Penguin diet usually consists of 56–93% by creased with day length from 8 h dayÐ1 in August to 12 h weight of krill, (E. crystallorophias and E. superba) dayÐ1 in September. In synchrony with the increase in (Table 2). However, fish are commonly taken, mainly time spent in the water, the dive rates increased from 92 the Antarctic silverfish (Pleuragramma antarcticum). In dives dayÐ1 in August to 149 dives dayÐ1 in September. some circumstances, fish may be 98% of the diet The most common maximum depth ranges for hunting in (Table 2). The differences in diet are seasonal with August and September were 20Ð200 m. The birds trav- elled up to about 100 km from the Auster colony (67.4¡S; 64¡E) (Kirkwood and Robertson 1997) (Fig. 2)