THE AUK

A QUARTERLY JOURNAL OF

ORNITHOLOGY

VoL. 79 APRIL, 1962 No. 2

THE BIOLOGY OF THE GIANT PETREL MACRONECTES GIGANTEUS

JOHN WARHAM

THis accountof the Giant Petrel (Macronectesgiganteus) is based on investigationson MacquarieIsland from 29 December1959 to 12 March 1961 while the author was biologistto the 1960-1961 AustralianNational Antarctic ResearchExpedition (ANARE). MacquarieIsland (latitude 54030' S; longitude159 ø E) is about 1,300 km (800 miles) south of Tasmania and just north of the Antarctic Con- vergence. It lies approximately north and south and is about 34 km (21 miles) long by 4.8 km (3 miles) wide. A narrowbeach terrace borders much of the coastline,but inland the groundrises steeply in screesto a plateau averagingabout 260 meters (800 feet) abovesea level. There are no trees,and the vegetationis composedmostly of mosses,Maori Cabbage (Stilbocarpa), and tussockgrasses (Poa spp.); there is very little growth on the exposedparts of the plateau. The climate is wet with much mist on the high ground,and, as the island lies in the west wind belt, galesare frequent at all seasons. The mean monthly temperatures range from 3 ø C (37 ø F) in June to 7ø C (44 ø F) in January, and even in winter snowseldom remains for long on the groundat sealevel. A generalaccount of the island and of ANARE activities there has been given by Law and Burstall (1956). Macronectes giganteusis a large petrel of the sub-Antarctic zone. It is a surfacenester; too large for natural enemiesto attack it when ashore, it is active on land by day. Most recent studiesof the speciesare con- cernedwith migrationof juvenilesas revealedby the recoveriesof birds bandedas nestlings.On Macquarie Island 7,366 chicksand 1,157 adults have been banded up to March 1961, including 2,946 chicks during the courseof the presentwork. Recoverieshave been discussedby Howard (1956) and Ingham (1959). Sladenand Tickell (1958) and Tickell and Scotland(1961) have dealt with recapturesof birds marked by the Falk- land IslandsDependencies Surveys. Murphy (1936: 584-596) summarized

139 The Auk, 79-' 139-160. April, 1962

140 W^R•r^•x, Biology of Giant Petrel [ Auk [ Vol. 79 what was known of the speciesat that time. Various voyagers,e.g., Biermannand Voous( 1950: 35-39), recordedthe distributionof the Giant Petrel at sea. Falla (1937: 137-145) describedobservations made at Macquarie Island, Kerguelen, at sea, and on the coast of Antarctica. Downeset al. (1959: 69-78) have publishedobservations made at Heard Island. Severalmembers of the ANARE parties contributedto the present study. Messrs.K. Watsonand A. Evans helpedwith the bandingof the chicksin 1961,a projectthat wouldhave been impracticable single handed, and Messrs.J. McNally and A. Thomasalso assisted. Dr. R. Carrick, Dr. D. L. Serventy,Mr. S. Davies, and Miss S. A. Ingham kindly read the roughdraft, and the latter's summariesof the biologicallogs of previous partieshave been of help for comparativepurposes. Mr. W. B. Hitchcock providedthe bandingdata. The paper has alsobeen improvedas a result of a helpful discussionwith Dr. R. A. Falla.

THE ANNUAL CYCLE

Most Giant Petrels at Macquarie Island laid their eggs from about 5 to 11 October. A minority laid earlier--mostly from 21 August to 10 September.These did not nest coloniallylike the later breeders,but rather as solitary pairs or in small, very scatteredgroups. Such early nesters hatchedtheir eggsfrom about 19 Octoberto 6 November,and their chicks flew from about 1 February to 5 March. Most of the colonialpairs hatched their eggsbetween 5 to 20 December,and their chicksfirst flew between26 March to 30 April. Long beforethe last chicksleft there was an outburstof nest building, either at the edgesof the coloniesor in entirely new places. This was seenduring the first week of February and was attributed to submature birds. Thus on 6 February 1961 a small party of Giant Petrels was dis- turbed from a group of four freshly made nests. The birds were in dark gray-brownplumage, had dark heads, and were away from any estab- lished rookery. During the next three weeksa revival of nest building was noted at many places, but especially around the edges of existing colonies. This activity may have begun as early as 16 December. On 28 February six new nestswere found whoseoccupants were mostly rather dark-plumagedbirds. They were in pairs, braying and displaying mutually. Dark or freckle-headedpetrels predominatedamong the other autumnalnesting parties, and the absenceof light-headedbirds strongly suggestedthat such autumn nest makers belongedto a subadult class of the Giant Petrel community.They couldhardly have beenbreeding birds becausesuccessful pairs were still busy feeding chicks. The only other Vol.^uk 79 ] W^R•.•M, Biology of Giant Petrel 141 bird that frequentlymakes nests in the autumn at MacquarieIsland is the GentooPenguin (Pygoscelispapua), which is presentthroughout the year. Somebreeding stations of the Giant Petrel, suchas thoseon islandsoff Antarctica, are desertedduring the southernwinter, but at Macquarie Island the rookerieswere occupiedat all seasons.Successful breeders may have goneto sea after the departureof the chicks,but the colonieswere soonreoccupied by adults and continuedto be used during the winter. At that seasonmany Giant Petrels sat in pairs in the coloniesby day and engagedin energeticmutual display. Such activitieswere seeneven when the groundwas under snow,and pale-headedbirds predominated.This indicatesthat they were all mature individuals,a hypothesissupported by the captureon 2 June of a bird that had been bandedon its nest at the samecolony in November 1954 and by three similar recoverieson 25 July 1959 by S. Csordas.

Four Giant Petrels banded as chicks have now been recovered on the island, three during the course of the present study:

130-03555, a white-phase bird banded 13 March 1955 and shot on 5 August 1958, three years and five months after banding.

130-o0160,banded 24 February 1957 north of Aurora Point and shot when a dark- eyed, dark-brcwn bird at Bauer Bay on 30 August 196o, three years and six months later.

13o-o8175, banded 22 February 1956 at Major Lake, was caught and released at Hasselboro Bay on 4 October 196o, four years and seven months after banding. It was describedby its captors as a "uniformly grey bird."

130-o2806,banded 24 February 1957 north of Aurora Point, was caught and released on lO March 1961, four years later on a rookery near Douglas Point. It was then a dark-eyed, dark gray-brown bird with a dark head and may have been one of several autumn-nesting birds on this rookery.

The last of these recaptures,and the absenceof other birds banded as chickson the colonies,suggests that the Giant Petrel doesnot breed until at least five years old; this is in agreementwith Downes et al. (1959: 75), who came to the same conclusion from the lack of recoveries of chicks at Heard Island. Tickell and Scotland (1961: 261) report a banded bird nestingat eight yearsold but give no detailsof its age at first breedingor of its sex or plumage.

POPULATION AND NEST SITES

A censusof breederswas attempted during January and February 1961 when all the rookerieswere inspectedand the chicksbanded. A total of 2,846 weremarked with bandsissued under the AustralianBanding Scheme and inscribed "Write Wildlife CSIRO Canberra Australia." Twenty-one 142 W^Ra^•r,Biology of GiantPetrel [ Vol.^uk 79 other chicks too small to band were also counted. There were more than these2,867 chickson the island, as a few isolatednests on both west and east coastswere known to have been overlooked; a yard-by-yard search of the ruggedtussock-clad rocks along the indented coastlinewas im- practicable. It is estimatedthat not more than 150 such nestlingswere missed. Per contra, someof the countedchicks must have died through various natural causesafter banding. In all, the chick production for the 1960-1961 seasonis believedto have been about 3,000 birds. Assumingthat the breedingsuccess of Giant Petrels was 57 per cent that year, then about 5,300 eggswere laid during the season,the product of 10,600 breedingbirds. However, from the data available, the breeding successcould have been nearer 70 per cent, which givesonly 8,600 breed- ing birds. The total population,allowing for all classesof nonbreeders, was certainly much greater than this. The petrels bred in 70 colonies. All but one of these colonieswere on the west side of the island where they were exposedto the strongprevail- ing winds. Most were quite small, producingan average of 41 chicks. The three largestheld 167, 160, and 126 young birds, the smallestonly four to six at the time of banding. The greatest concentrationsof rookeries in both the 1959-1960 and 1960-1961 seasonswere in the area of Caroline Valley in the southwest corner,where 606 chickswere counted, and alongthe coastalrocks between Mawsonand Aurorapoints, where there were 940 chicksduring the census. Previousrecords suggest that these have always been the best breeding areas. Complete details of the 1960-1961 colonies are deposited at ANARE headquartersin Mdbourne. From previousplots showingthe sitesof someof the breedingcolonies it is clear that the locationsof many of the larger and more accessible rookeriesare relatively permanent. Severalwere evidently on the same sitesin 1952 and 1954 as in 1958-1961. A morecomplete mapping by K. Keith in 1956 reveals that 12 of the larger colonieswere in the same placesthen as they were during my stay on the island. Someshifting occurs;large coloniesmay split into two, and small onesmay coalesce. In the region of Major Lake the sites of the rookerieshave changed severaltimes between1952 and 1960. One colony near Mawson Point in which about 20 chicks were reared in the 1959-1960 season was not re- occupiedthe next year; insteada new site was usedabout 400 metersto the south,apparently by the samebirds. The faithfulnessof Giant Petrels to a particular rookery or breeding area was shownby six recapturesof breedingbirds during 1959-1961 at the samecolony or in the samegeneral area wherethey were bandedas Vol.Auk 79 ] W^R•^•, Biology of Giant Petrel 143 adultsseveral years before. Most adultsrecaptured before 1960 werealso foundwhere banded, but a few recordssuggest shifts of one to three km. Postbreedingoccupation of the rookeries,apparently by the mature birds, must tend to stabilizethe sitesfrom year to year. It may also tend towardthe continueduse of the samenest by the samepair duringsucces- sive seasons.Many of the early solitary breedersdid not choosethe same placesin successiveseasons, but suchbirds were not seennear their nests during the winter months. Repeateduse of the same nestsby the same pairs is usual in burrowingpetrels and occursin at least one surfacenester --Fulmarus glacialis (Carrick and Dunnet, 1954).

DISPLAY

Giant Petrels are easily frightened by man. For closeobservation con- cealmentor long-rangeviewing was necessary,and much of the following information on behaviorwas obtainedby observationfrom blinds. Written recordswere supplementedby a frame-by-framestudy of motion pictures.

Threat

Aggressiveposturing was the mostcommon type of display. Threatening occurredamong birds in rookeriesand amongmembers of partiesgathered around food either on land or sea. It was not always clear just where threat endedand courtshipor greetingceremony began, and many threats beganmildly and ended in combat. A low-intensityattitude used by birds on their neststo discouragein- trudersinvolved braying criesand head-swayingmovements in which the neck was upstretchedand the nape feathersruffled. If this was ineffective, sitting birds lungedforward and struckwith their bills. Well-grownchicks also threatened in this manner. Much more impressivewas the forward-threat posture, a grotesque attitude often seenamong birds attempting to feed on a dead seal (Figure 1). The threateningbird squatted,stood, or ran with its tail fully ex- panded and tilted so that the tip was directed forward. The neck was arched, the nape feathers much ruffled, and the bill lowered. The out- stretchedwings were alsoarched, and their tips might trail on the ground. The posturingbird either facedits rival or had its back to it with the tail uplifted. This forward threat seemed to reflect the bird's intention to attack. Suchdisplay was seldomignored by the rival (usually a dominantbird on a seal carcass). The responseof such a "sealmaster"was the upright- threat attitude describedbelow followed perhaps by attack. Particularly noteworthy was the effectivenessof forward-threat display in inducing 144 W^R•ta, Biology oj Giant Petrel [ Vol.auk 79 •.uk ] W^•-a•, Biology of Giant Petrel 145 Vol. 79 J ' attack at a distance;dominant birds were seento stop feedingand to drive off potentialrivals displaying 20 metersaway. Furthermore,the factthat the threateningpetrel presented only the undersideof the fanned tail did not reducethe vigorwith whichit wasattacked. It seemedrather thatthe uplifted tail releasedaggression in the "sealmaster," but no experi- ments were made to test this belief. Petrels that fed with their heads and necksimmersed within a sealkept their tailscocked high. Suchbirds seemedvery consciousof the weaknessof their position;they werenervous and were quick to switchto attack when threatened. In disputesover food it wasnot alwaysthe feedingbird that wasthe moreaggressive; often the forward-threateningpetrel dashed toward the carcassto fight the dominantbird. The latter eitherretained its position andthe disputantretired or the newbird wonand reigned in the other's place.Sometimes threats alone sufficed to evicta "sealmaster,"and there was no fighting. The changeoverof feedingbirds appearedto be con- tinuous.As the dominantones became satiated, their belligerencedwindled, and they werereplaced by morevigorous or hungrierbirds. None was everseen to displaceanother without first adoptingthe forward-threat posture. The upright-threatattitude is illustratedby Downeset al. (1959: 76). Asin forwardthreat, the wings were spread and bowed, the tail fannedand uptilted,and the napeand back feathers extensively ruffled. Now, how- ever,the neck was vertical, the head held high, and the bill at about45 ø . The headwas waved vigorously from sideto side,rapidly or slowlyac- cordingto theindividual, so that the bill almosttouched the bend of each wing. At the sametime an expiratoryneiõhing cry wasmade, wavering in time with the head oscillations. Thisdisplay occurred when a birdseemed to be tryingto repelseveral others,whereas forward threat seemed to bedirected at a singlerival. Up- rightthreats were often used by GiantPetrels immediately after alighting in a rookery,perhaps because they had difficulty in landingon "neutral" ground.The displaywas also used almost invariably by the victorof a fightover food or by a victorious"sealmaster" that returnedto a carcass to findseveral bystanders feeding in its absence.Adoption of the upright- threatposture appeared to reestablishthe dominantbird's claim to the feedingplace. Whenfighting, the birds'beaks often be.came interlocked. The ad- versariesthen grappledtogether, breast to breast,straining upwards. Sometimesone grasped the other'snape and hung on tenaciously.A bird that obtainedsuch a grip usuallybecame the victor. Many fightswere seenwithout any wounds being apparent, but the severity of somestruggles suggestedthat seriousinjuries must sometimes result. 146 WARa•r, Biology of Giant Petrel I_[ Vol.t•uk 79

Both forwardand uprightthreatening with uptiltedtails occurredat sea when mobs of birds were attracted to a carcass. In forward threat the neck lay on the surface, and attack followed a violent dash across the water. In upright threat the birds kept their headshigh and waved them from side to side just as on land. During banding operationsmany adults were found with their chicks, and someremained with them. They had to be pushedoff the nestsbefore the chicks could be banded. Most parents merely gave low-intensity threats and returned when we moved off, but one male that had shown little alarm when its own chick was handled,suddenly attacked the bander as he was engagedwith the next chick. Such aggressiontoward man was quite exceptional.

Mutual Display This sexualdisplay was seenthroughout the year. It occurredduring the winter reoccupationof the rookeries,before egg laying, in the breeding season,and during the autumn nest-buildingphase. The actionsused were very similar to those describedfor Fulmarus glacialis and F. glacialoides and involved billing, mutual nibbling of heads and necks, and bowing. Suchdisplays often involvedonly two birds of oppositesex, but three or more membersof small groupswould display among themselvesin this way. Much mutual displayduring and immediatelyafter nestingseemed to be between adult nonbreeders. The birds of a pair generallysat closetogether and either sex initiated the display. Typically, one lifted its head and pointed the partly opened bill skywardsbefore turning it down toward its companion,leaning for- ward, and, with a swayingaction of the head, reached toward the other's head and neck. Both then wobbled and stretched out their heads, and a soft braying call was heard. Their tails were often spread and waggled from side to sidebut were not uptilted. Their bills were slightly ajar, but their napeswere not ruffled. Head wobblingoften led to mutual preening of heads,throats, and flanks when featherswere sometimesdislodged. The birds' beaks often touched, and each bout of activity lasted about 10 seconds.One pair on 26 July 1960 performed22 timesbetween 09:21 and 09:58 hours. At intervalsone bird, often the female, tucked its beak into the scapularsas if it were sleeping. Beforeegg laying, pairedpetrels generally kept togetherin one spot,but somewandered and displayedmutually in variousplaces. Suchwanderers often returned to one particular site; presumablythis was where they would nest. Head wavingwas frequentlyseen when a Giant Petrel was approached by another, not necessarilyby its mate, and such actions appearedto Vol.Auk 79 ] WARItAlv[,Biology ' of GiantPetrel 147 form a greetingceremony since mutual preeningand billing often fol- lowed. Mutual displaywas also seenamong birds restingon the beaches while waiting to feed at a dead seal.

Coltion

Coition was first seenon 24 Augustand last seenon 1 October. On both these occasionsthe femaleswere already coveringeggs. On 1 October,when her eggwas not morethan one day old, a sittingfemale was the object of preliminary circlingby a male thought to be her mate. The femalesat low in her nest,bobbing her headslightly and makinga short, mewingcry. At intervalsthe male dippedhis beak toward her; she swivelledaround to face him, and occasionallyshe tried to peck him. Eventuallyhe climbedonto her back and stayedthere for five minutes, trying to strop his bill acrosshers in the manner described below;but shewas unresponsiveand he had to stop. More usually,once the male had mounted,he held his bill vertically, pointing downward,and, by a series of swingingmovements to either side, wiped his bill acrossthat of his mate. Their beaks touchedthree or four times per second,and the resultingclicks were quite audible from 100 metersaway. The male'stail was swungfrom side to side, and at intervals the female gently thrust her openedbill upwardsinto the feathersof his throat, breast,or flanks. Most malesonly openedtheir wings slightly during these preliminaries,but toward the climax ex- tended them to either side. The bill movementsthen ceased,and the male reached forward and remained quite still while he swung his tail sidewaysto bringthe cloacaetogether. Some attempted copulations lasted for 5-10 minutes,but it was not possibleto distinguishbetween effective and ineffective matings. The seldomdescribed copulation of the Atlantic Fulmar (Fisher, 1952: 333) is apparentlyrather similar exceptthat beak actionsof the males are not recorded; in Fulmarus glacialiodes,however, they do occur (Pr•vost, 1953b).

Aerial Display This was the mostpuzzling of the Giant Petrel'sdisplays. It was seen in every month and did not seemto be more frequent at any seasonor at any time of day (vide Murphy, 1936: 593). The performanceresembled the upright-threatdisplay and was usedboth by singlebirds far from other Giant Petrels and when two were flying closetogether. Most such displaysbegan when a flying bird rose a few meters on stiff, somewhatdrooped wings, and then descendedin a short curve. The actionwas rather like that of a displayingturtle dove. The petrel'sneck 148 W^R•^•r,Biology of GiantPetrel [ Vol. 79 and head were stretchedand raised,the nape feathersruffled, and the tail fanned. Often the legswere loweredand the websexpanded. As the bird glided along,its head was waved from side to side and the usual neighing cry was heard. In a few instancesaerial display appearedto be the culminationof a short chase. If so, it was usuallythe hindermostbird that displayed,al- thoughboth might do so. Singlebirds were seento display, and after a wide circuit over the coastalterrace, repeated the performancewhen they reachedthe sameplace as before. Suchbirds might be over rookeries,but often no possiblepoint of interestcould be seen; e.g., one petrel displayed vigorouslywhen about 100 metersup and directly above the writer. A small group of Giant Petrels was feeding on a dead rabbit on 29 March 1960. Severalothers circled above,and as they flew past drooped their wings, lowered their feet, swayed their heads, and brayect. Here aerial displayseemed to be associatedwith threat. Quite distinctwas the mild chasingthat occasionallyinvolved two birds when one followedanother for a few hundredmeters. Close pursuitsor synchronousflights suchas occur amongpetrels of the genusPterodroma or amongPhoebetria albatrosses were never seenin Macronectes.

THE NEST AND INCUBATION

The kind of nest at Macquarie Island varied with the situation. Soli- tary nests, and those in coloniesamong coastal rocks or tussock,were built of grassand were mostly quite large. Those on the open mosscon- sistedof depressionslined with plant debris and grasses,while those on the plateau,which were very exposed,were saucerlikehollows containing little or no nestmaterial. Usually at suchsites there were no plants nearby exceptmosses or Azorella, and any vegetationthat was broughtin would soondisappear in the strongwinds. Such situationsappear to resemble the nest siteson Kerguelenreported by Paulian (1953: 163). No solitary nestsof the early breederswere found on the plateau; all these nested near the sea,mostly in relativelysheltered places among rocks and tussock grass. How the nestinghollows were made was not discovered,but the birds beforeegg laying collectedgrasses that they droppedto one side and later added to their nests. Somebirds tore blades from the Poa stools; similar behaviorwas noted amongthe autumn-nestingbirds. During incubation the petrelsadded to their nestsany fragmentsof vegetationwithin reach. Becauseno birds were color marked before egg laying, it was impossible to discoverwhether there was a prelaying exodussuch as occurswith some burrowing petrels and with the fulmars Fulmarus glacialis and Daption capensis. Vol.Auk79 ] WARItAM,Biology of GiantPetrel 149

Many incubating Giant Petrels deserted their eggs when closely ap- proached. Such eggswere quickly taken by the skuas. For information on incubation,the nestsoccupied by the bolder birds that stayed on their eggswere studied. Two different groups were used: one, consistingof early breeders,gave information about the r61esof the sexesand, later, of the length of the fledging period. Eleven such pairs included birds sufficiently tame to be marked on the forehead with red enamel. This, when well presseddown into the feathers,remained visible throughoutthe breedingseason. Once one memberof eachpair had been marked, it was easy to seewhich was on duty. Even if not color marked individual Giant Petrels could also be recognizedby the patterns of their irides and by the colors of their beaks. No two birds were alike in all these features. The secondgroup gave data on the length of the incubationperiod and was usedlater to obtain growth curvesfor the chicks. Once the datesof egg laying had been found and the nest sites pegged,these birds were not disturbeduntil hatchingwas imminent. Regular inspectionswere then begun. The first eggsseen on 21 Augustwere in solitary nests. Laying of lone pairs continueduntil about 10 September.Very few eggswere then laid until those of the main body, whose laying extended from about 1-15 Octoberwith a peak from 5-11 October. Theseearly breeders laid beforethe bulk of the skuashad returnedfrom their winter migration;indeed these Giant Petrelswere the first of the island'sbirds to lay. But soonany eggsleft unguardedwere immediately lost to patrollingskuas and to Wekas (Gallirallusaustralis). However,if I promptlyhid sucheggs under the nestingmaterial most birds resumed incubation when they returned, and lossesthrough disturbancewere therebyreduced. The bolder Giant Petrelstended to shuffleback off their nestswhen I camenear, or they jerked upwardsas if to fly but did not do so. When lifted with a stick so that the egg could be checked, many brayedquietly. Somewent throughthe motionsof vomitingbut disgorgednothing; others discharged the contentsof their stomachsand then either fled or remained. Many apparentlytame birds revealedtheir nervousnessby shiveringas penguins,albatrosses, and otherpetrels do in similar circumstances. Incubationwas undertakenby both sexes. Each sat continuouslyfor severaldays until relievedby the other,a routinethat obtainswith other petrels;e.g., Puffinus tenuirostris (Marshall and Serventy,195 6: 492). A few female Giant Petrels sat for two days after laying, but the males usuallyundertook the first lengthyshift. Daily checksat the 12 solitary nestsstudied were not possible;only approximatefigures were obtained for the lengthsof the incubationshifts. Thesevaried from two to 12 days' 150 WARttAXI,Biology o! Giant Petrel [ Auk I_ Vol. 79 duration, and usually the male had five shifts on the egg and the female four shifts betweenlaying and hatching. Despite the lack of daily inspec- tions it is not thought that any changeoverswere missed,and checksat night revealedthe samebirds sitting as at dusk. A particularly long spell was undertakenby one female, which sat for 25-28 days. On the day before her abandonmentI noted that this hitherto tame bird was becoming timid and that her desertion seemed imminent. During the many visits made to these early nests while incubation was proceeding,only one bird was found at each nest, and no nest reliefs were seen. Nest reliefs were, however, witnessedlater at the rookeries. The changeoverusually took place without ceremony,the sitting bird raised its wingsand jumped off to reveal the egg for a moment. Then its mate walked on and sat down, and the relievedbird flew off. One male sat by its mate for over an hour before taking over and at about 10-minute intervalsreached forward toward the female to indulgein mutual display. Eleven determinationsof the incubationperiod were made. They were 58.5; 61.5; 58.5; 58.5; 60.5; 58.5; 60.5 (each -+ 1.5 days); and 61; 58; 59; and 59 days (each -+ 1 day). The shortestpossible period was 57 days; the longestpossible period, 62 days. These are the true incubation periods, for this surface-nestingpetrel cannot leave its eggs uncovered as do many burrowing species;this thereby increasesthe time between laying and hatching. The figuresare substantiallyless than that given by Downeset al. (1959: 73) of approximately70 days at Heard Island but agreewith Pr•vost's (1953a) one recordof 60 days for a bird at Terre Ad•lie. Data for other fulmars are 40-57 days for the Atlantic Fulmar (Fisher, 1952), 43-44 days for the Antarctic Fulmar (Pr•vost, 1958), and about 45 days for the Cape Petrel (Downes et al., 1959).

THE NESTLIN½

Birds of either sex were sitting when hatchingbegan, and from one to six dayselapsed between the starringof the eggand the birth of the chick. Seveneggs shortly before hatchingweighed 210, 220, 230, 240, 250, and 260 g and two just-dry chickseach weighed 220 g. Lone pairs hatched their eggsbetween 19 October and 5 November. For the colonialbirds this event fell between3 and 20 December,but a few chickswere not hatcheduntil 1 January. The youngwere covered by either parent in turn for about 15-24 days (8 records),after which they were not broodedeven if a parent was present. During this guard stage the on-duty bird often preenedits chick. From time to time adults rose from their neststo beat their wings and to shake their feathers. The voiceof the newly hatchedchick was a quiet peep; this changed Vol.Auk79 ] WARItAl'•f,Biology of GiantPetrel 15 1 to a deep,guttural snortingduring the nonguardstage, and later a bray- ing note was added. Parent-chickfeeding behavior was similar to that of other Procellariidae, e.g., to Puffinus carneipes(Warham, 1958). The chick used a deep cry as it reachedforward to peck at the parent'shead, bill, and flanks. The latter swallowedseveral times and then openedits beak. The chick's cries now got louder, its actions more persistent,and the parent then flatteneditself somewhatto the ground,opened its bill widely to allow the chick to insert its own beak acrossthat of the adult bird. Calling stopped as the food was transferred with scissoringmovements of the mandibles. The chick then withdrew, both birds swallowed, and the processwas then repeated. A seriesof meals was usually given, the nest- ling's beak being insidethe parent'sfor about four secondsat eachmeal. Individual variations were noted; some chicks begged for a long time beforetheir parentsresponded, yet other adults did so readily and some even initiated feeding by opening their bills before their chicks had started to call. When the nestlingswere old enoughto be left unattended,they slept much of the time. On warm days they gapedand panted heavily. In wet or windy weather they hid their heads in the down of their backs and assumeda sphericalshape. There were sometimesdroplets on the tips of their bills, perhapsthe resultof salt excretionby the nasalglands. Many chicksadded to their nestsby reachingout to collect dead grassesand other fragmentsfrom the hollowsbetween the nests. During this stage of the breedingcycle the parentswere very seldomseen at their nests; this suggeststhat the adult's feeding visits are brief. The chicksgrew rapidly and at 50 daysattained adult weightof about 4.5 kg. Weight variations resemblethose of other petrels and of alba- trosses,showing a rapid increaseto a peak greaterthan adult weight, followed by a steady decline to fledging. The study chicks flew on attaining weightsof 3.6-5.6 kg. There was no evidencethat feedingcon- tinued until they flew. Rather, there appeared to be a short desertion period. The first featherswere hidden by down, but at about 40 days of age the wing and tail featherswere both about 2.5 cm in length. At 60 days the primarieswere about nine cm and the rectricessix cm in length. By 80 days thesewere 15 and 11 cm, respectively,and at 100 days they were 25 and 15 cm. Somechicks departed at this stage. White-phasechicks started life with white down, and thosethat were to have black spotsin their plumageshowed these with the unfoldingof the first feathers. Some white-phasechicks had legs of a pale grayish- flesh color. But most had pale-blue legs with flesh-coloredwebs, and 152 W^Ra^M,Biology of GiantPetrel [ Vol.Auk 79 many had gray legsjust like thoseof normal chicks. Pale-footednestlings were thought to result from matings between two white-phaseadults, and because such pairs were scarce, pale-footecl chicks were rare. A few normal dark-phaseyoung had patticoloredwebs shadedwith dark brown on their outer edges. Many chicksstayed in their nestseven when they seemedable to walk, but othersstarted to wanderwhen about 85 days old; for sometime before their departureyoung birds in the shiny-blackjuvenal plumagewere found severalmeters from their nests. One of the study chickswould bray when approachedand after being weighedwould waddle 10 meters back to its nest. This tendency to move became noticeable on calm days in early February when down-free birds ready to leave dotted the coastal moss terraces. Thesewere the young from the solitary nestsawaiting the winds they needed for takeoff. Their dependenceon strong westerliesmeant that spellsof calm weather at fledgingtime strandedthe young birds. There was no evidencethat any walked down to the shoreand swamout to sea as shearwaters often do. What was believed to be a Giant Petrel's first flight was seen on 11 April. The bird stood on a ridge with its wings out so that a westerly wind with gusts to about 20 knots tended to lift it from its feet. The fledglingbeat its wingsbetween gusts but seemedunable to get airborne. Eventually, after severalpoorly timed attempts, it took off with a short run, risingon flappingwings and gainingheight as it movedforward over falling ground. The bird made no attempt to sail on stiflenedwings in the mannerusual with this species,and it lookedrather unstable. Beating its wingssteadily it drifted southat about 45ø to the breezeuntil it was over the sea and its migrationhad probablybegun. Othersevidently did not get away at their first attempts. Thus the chick from nest 51 was found on the mossabout one-half km from its natal rookery. It was free of down and had doubtlessreached the mosson its first flight. Seven determinationsof the fledgingperiod were made. They take no accountof any strandingdue to calms,and, as both hatchingand flying dateswere not known accurately,the figuresare only approximate. They are (in days): 102 ñ 4; 103 ñ 3; 106 ñ 2; 108 ñ 4; 110 ñ 3; 113 ñ 4; and 117.5 ñ 1.5 days. By 15 April only occasionalstragglers were seen near the rookeries;the rest of the young had gone.

BREEDING SUCCESS

Three colonieswere observed in an attempt to determine the ratio betweenthe number of eggslaid and the number of chicksreared to the bandingstage. A countof incubatingbirds was madeat eachcolony on 30 October1960, when egglaying would have beencomplete, and another ^uk ] W^R•r^•, Biology of Giant Petrel 153 Vol. 79 ] count of the chickswas made on 12 January 1961. As the rookerieswere not visited between these checks, any lossesmust have been due to natural causes.

At rookery N.15 50 eggsproduced 38 chicks: 76% success. At rookery N.69 104 eggs produced 40 chicks: 39% success. At rookery N.60 114 eggsproduced 75 chicks: 66% success.

The wide variationsin thesefigures, for whichI can offer no explanation, suggestthat they are not reliableas indicatorsof the over-all efficiencyof breedingon the island during that season. Most nestingsapparently failed duringincubation; if the egg hatched the chance of fledging was good. Although skuas were often seen near the rookeries,neither they nor the numerousferal cats are known to have taken properlyguarded eggs or chicks. The latter, at the start of the nonguardstage, were quite smallbut seemedcapable of deterringenemies by spitting oil. During banding,very light chicks,which appearedto be starvingto death, were sometimeshandled. Others,though of normalweight, had oily, matted downthat was of little protectionagainst the weather. These chicks too seemedlikely to die.

Two-EGG CLUTCHES

In the 1960-1961 season,four and possiblyfive nests (0.14-0.18 per cent) were found with two eggsor two chicks. That thesewere not the result of two femaleslaying in one nest was suggestedby the improb- ability of a bird leaving its egg without losing it to the skuas,by the obviousterritorality of the breeders(they threaten other Giant Petrels that comenear), and by the discoveryof a white-phaseadult brooding two similarwhite-phase chicks (Figure 2). The two otherchicks believed to belongto one pair were about one-half meter apart. No nest could be seenfor the extra bird, and as both were of the samesize and wore a similar,rather distinctive,brown-gray down, they werejudged to be the offspringof the same female. Of the nestswith two eggsone was not reexamined.The secondpair hatchedone egg, but the chickwas found dead in the neston 23 December; the unhatchedegg was holed and containeda well-developedbut dead embryo.At the thirdnest (Frontispiece) neither of the eggs(66.1 X 99.3 and 65.1 x 103.0mm) hatched;one was slightly developed and the other disappearedbetween 16 and 22 December.Both theselatter pairscon- tinued incubationfor over a week after their neighborshad hatchedtheir eggs. 154 WAR•IA1V•,Biology o! Giant Petrel Auk Vol. 79

Figure 2. White-phase adult Giant Petrel brooding two white-phase chicks.

Neither of the pairs of birds with two chickssucceeded in rearing both. A secondinspection on 6 February 1961 revealed that one chick had dis- appearedwithout trace from each nest.

Marshall and Serventy (1956) have described the rapid collapse of the arian testes after ovulation in PuIfinus tenuirostris and state that this would make it impossible for these petrels to lay a secondegg if the first were lost. They add: "In fact, the great size of the one egg, occupying a large part of the abdominal space, and the probable long period between insemination and egg-laying, throws serious doubt on the validity of observations that two-egg clutchessometimes occur in Fulmarus glacialis seealso Fisher (1952)." Serventy (pers. comm.) suggeststhat in thesetwo egg layings of Macronectes it is more likely that two o6cytes were released from the ovary and both fertilized at one insemination or at inseminationsvery close together, rather than that twinning occurred in a single o6cyte after its release. It may be noted that Vol.^uk 79 J] W^RH^M,Biology o] Giant Petrel 155 accordingto Marshall and Serventy (1956), the singleegg of P. tenulrostrisrepresents about 16 per cent of the female'sbody weight whereasin the Giant Petrel the egg at about 250 g representsonly about 6 per cent of her body weight of 4.3 kg. This makes more understandablethe holding of two eggsin the body cavity just before laying. Two-egg clutcheshave been reported also in the Atlantic Fulmar. Fisher (1952: 94) recordsthat on the Westmann Islands up to 10 per cent of the nests may con- tain two eggsand again (p. 461) reports a site where one long and one normal egg were found in 1947 and again in 1948. The latter evidently represented layings of two eggs by the same female, and some of the layings on the Westmanns must surely have been genuine two-egg clutches. However, there seem to be no records of any Atlantic Fulmars that have hatchedtwo eggs,and Gudmundsson,quoted by Fisher, noted that the Westmann birds were incubating only one of their eggs, the other lying cold beside them. This was not the situation with Macronectes; the sitting birds coveredboth eggs,although whether their incubation patcheswere always large enoughproperly to accommodateboth is unknown. With the two-egg clutchesthat failed, lack of covering power, particularly by the smaller females, might have been decisive,but the incubation of two eggsis evidently effective sometimesor the chicks in Figure 2 could not have been hatched.

PLUMAGE AND MOLT

Apart from white-phasebirds that are white from hatching,three main plumagetypes were distinguished: i. Juvenalplumage in which the featherswere a glossyblackish-brown; eye dark brown; bill horn-colored;nails pale green. 2. Body feathersdark brown, lighter aroundthe baseof the beak; eye brown or gray; beak and nails pale yellow-greenwith the nostril horn- colored.No breedingbirds were seenin this plumage,which is evidently that of iramatures.It is typical of the Giant Petrelsthat occuroff the southern coasts of Australia and New Zealand. 3. Body feathers gray-brown,lighter around the base of the beak and on the cheeksand throat; breast and flanks more or lessmottled with pale gray and brown or wholly pale gray; the head either rather dark, freckled gray and brown,or whollypale gray; eye varying from brownto pale gray; bill greenishto yellow-brown.This wasthe plumageof the breedingbirds, and it was very variable. That the lone pairs nested consistentlysix to seven weeks earlier than the majority suggeststhat the lone pairs belongedto a different age group than the rest. Suchearly nesterswere light, dark or freckle headed, but the colorsof the soft parts were more constant. Nearly all were gray eyed or had brown eyes fleckedwith gray, and their bills were generally pale yellow-brown,reddish towards the tips and with black smudgeson the insidesof both nails. Colonial breederswere different, for most had pale or freckled headsand very many were dark eyed, althoughsome had gray irides. The greatestdifference was in bill color. Most of the colonial 156 W.aRHA•, Biology of Giant Petrel [ Vol. 79 birds had pale, yellow-greenbills without black marks on the nails and with horn-colored nasal tubes. Becauseof bill and eye color the lone nestingpairs might well be judgedto be very old birds. The youngMacronectes has brown irides,and among breedersevery variety between that and the light-gray iris was seen. Many had brown eyesirregularly fleckedwith gray so that a bird's eyeswere seldomidentical. Evidently the brownpigment is graduallylost with age,and the high proportionof light-eyedbirds amongthe lone breert- ers suggeststhat they were old birds. Again, the greenishbills of the colonial birds were closer to those of the juveniles than to those of the solitary pairs. On the other hand, the plumage changesfrom blackish throughbrown to grayishsurely representstages to maturity so that we should expect the later-breeding,light-headed birds to be the older. In- cidentally, the suggestionof Falla (1937: 138) that these petrels become light headedthrough the bleachingaction of seal bloodwas not confirmed. Giant Petrel feathersplaced in fresh seal blood were not visibly changed after prolongedimmersion. That the solitarypairs retain the tendencyto nest aloneand in advance of the othersis indicatedby male 00208, which bred in 1960-1961 in the same area as that in which it was banded as a solitary breeder in 1957. This recoverymay be significantin view of the tendencyamong sea birds for establishedpairs to nest earlier than inexperiencedones.

Some counts were made of the white-phase adults in the Giant Petrel population. Of 557 birds counted at rookeries,44 or 8 per cent were white. In November 1952 F. Soucek and M. Taylor noted that 32 (10 per cent) of 305 adults banded were white, and in December 1959 A.M. Gwynn recorded 10 white birds (9 per cent) among 110. Previously Falla (1937: 145) noted 20 white birds (10 per cent) among 200. Downes et al. (1959: 77) state that at Heard Island such birds compriseabout one-half of 1 per cent of the population but do not indicate how many they counted. Previous estimates of white-phase birds are, for West Antarctica, 5-12.5 per cent, and for South Geor•:ia and the South Orkney Islands, 2 per cent (Murphy, 1936: 588). Care of the body surface involved preening and bathing, although preeningwas seldomseen. The uropygialgland was exposed,the bill oiled, and then workedamong the feathers. Similarly, the side of the head was rubbedacross the nipple of the gland and then rubbed on the feathersof the back and body. After feedingon a carcass,Giant Petrels removedseal fat and blood by vigorousand protracted bathing in the sea. The wings were opened and the head ducked under so that water sluiced over the back and tail. The coastal fresh-water lagoons were also popular bathing places, as were the endsof Prion Lake, the southernpart of Waterfall Lake, and the easternbeach of Major Lake on the plateau. Parties of petrelswere very VoL^uk 79 ]/ W^RH^•, Biology of Giant Petrel 157 common too in early spring on the coastal lakes between Mawson and Aurora points and, at all seasons,on the water of Caroline Cove. The nestsof many birds in early October containedcontour feathers, and it seemedthat a gradualreplacement of the body plumagewas taking place. From early October to the end of February birds on the wing showedgaps in the inner flight quills. In October and November this conditionseemed to be restrictedto nonbreeders,but five breedingbirds examinedon 22 December1960 had the three innermostprimaries freshly replacedbut not full grown. They werenew, shiny, and gray in color,and the adjacent secondarieswere also new. Biermann and Voous (1950: 36) record birds at sea showingthe loss of the innermostprimaries as early as 28 December,and two collectedon 20 Januaryand 11 February showed a heavy molt of primaries,secondaries, and tail feathersand a heavy body molt. "Consequently,the birds have a generally mixed body plumage consistingof old dull brownfeathers and newlyacquired fresh dark ones." These conditionswould also apply to many Macquarie Island birds on those dates. A nonbreedingadult caughton 28 February 1961 was in almost com- pletely newplumage---a silvery gray bird with only a few scatteredbrown featherson the body. The edgesof the new featherswere lighter and gave the bird a scaly appearance. The innermostprimaries were short and fresh, and those secondariesadjacent to the axillaries had also been re- newed. The tail was molting, and the central feathers were short and growing. At this date many of the petrels flying over the island had gaps toward the tips of their wings. Replacementof the flight feathersappears to proceedfrom the center of the wing outwards.

FOOD AND FEEDING

The food disgorgedby the Giant Petrel chickscontained bird intestines, penguin feet, tongues,and feathers, blubber and seal meat, cephalopod beaksand tentacles,and, morerarely, Euphausiaand amphipods.The legs of SootyShearwaters (Pu//inus griseus)and prions (Pachyptila sp.) were also found in the nests. Downes likewise noted this high proportion of sea birds in food remainsfrom Heard Island. The majority of these victims are certainly not caught on land, and it is strangethat large and active specieslike shearwatersand penguins can be captured, the more so as many of the remainsare of adult birds. During the parturition of the Elephant Sealsin SeptemberGiant Petrels fed on the placentaeand on the many dead pups. At all seasonsthey fed on the carcassesof seals that died on the beaches,a source of food that was probably of great value during the winter months. At this time too rabbits were eaten. Although Giant Petrels occasionallymade shallow 158 W^R•^•,Biology o] GiantPetrel [ Vol.Auk 79 swoopstoward rabbits, none was known to have killed a healthy animah those eaten were believed to have been killed by cats. Whenever these petrels fed on a carcassa form of "peck order" was quickly establishedwith one or more birds dominatingthe rest until, with partial satiation, others displacedthem. The number of "sealmasters" able to find places to feed at one time dependedon the size of the body. Many Giant Petrels appeared to fly regular beats, and when one de- scendedto a source of food others seemedto be attracted by its descent and flew down to investigate for themselves. A freshly killed seal was found within an hour, and within three hours 30 birds were present. Another seal from which no blood was flowing was not attacked for 24 hours, which suggeststhat the sight of blood is important for prompt recognitionof a dead seal. Seals that died at the water's edge posed a problem, for the petrels found difficulty in balancingon the carcassas it rolled in the swell. Feeding birds disturbed by man generally regurgitatedas they fled. Dominican Gulls (Larus dominicanus)promptly flocked down to gather up the discardedfood. The petrels destroyedmany chicks of the Royal Penguin (Eudyptes chrysolophusschlegeli) in early February during descentof the latter to the beachesbefore going to sea. At Lusitania Bay they also killed the young King Penguins (Aptenodytespatagonica) in their winter craches. Awkward thoughthey are on land, the petrels were agile enoughto enter a crachewhose birds openedto admit them.

OIL SPITTING

All fulmars eject proventricular oil, which is used to repel aggressors. At Macquarie Island most young and some adult Giant Petrels could throw oil up to two meters--more if the wind was behind them--but many adults could not do so or disgorgedonly small quantities of oil. Usually, when approached,a breedingbird merely made retching sounds and jerked its head without disgorging;if caught it would then throw up the stomachcontents. Some adults disgorgednothing when handled at their nests. Ejection of oil by flying birds as reportedfor the Atlantic Fulmar was never seen. Fisher(1952: 393) quotesan instancein whichan AtlanticFulmar chick squirtedoil througha holein its shell,but noneof the many newly hatched Macronecteshandled at Macquarie Island did this. By the end of the guard stage, however, the chicks could spit effectively, and on entering a colonyone wasgreeted by a barrageof oil from the nearbybirds, which revolved on their nests to face the intruder. There was, nevertheless,con- siderablevariation in the chick's ability to spit oil. Thus nestling 91, a Vol.^uk 79 ] W^R}r^M,Biology of Giant Petrel 159 male weighedregularly, never did so, nor did at least one of its parents. But female 92 nearby was a persistentspitter. Oil ejection was most marked early in the nonguardstage and gradually declinedwith age; after 25 February even 92 stoppedspitting when inspected. This "drying-up" processoccurs in other oil spitters such as the young Diomedea exulans, and it is not simply that regularly weighedchicks becomeconditioned to handling,for young Giant Petrels caught when ready to fly seldomeject oil even if they have never been handled before. Here again there were exceptions,and a few chicksdid remaincapable of spittinguntil they flew. Chick 10, a female,squirted oil on 11 February when about 112 days old, althoughit had not been fed sinceabout 28 January. Such behaviorwas quite unusual. No direct information was gathered on the effectivenessof this device as a defenseagainst the only natural predatorsnow active on the island, the skua and other Giant Petrels. A skua walking among chicks was threatenedby severalof them, which it ignored, but it was impossibleto see whether they actually spat at the bird. No Giant Petrel was seen to attack a chick of its own species,but Falla (pers. comm.) has seen a solitary guarded chick on the Auckland Islands ringed by adult Giant Petrels that were intent on its destruction.

SUMMARY

1. In the 1960-1961 seasonabout 3,000 chicksof Macronectesgiganteus were rearedat Macquarie Island (54 ø S; 159ø E), and the populationis estimatedat lessthan 9,000 to over 10,000 breedingbirds. 2. No birds bandedas chickshave yet been found breeding; maturity is probablynot reachedfor at least five years. 3. There are two classesof breedingbirds. A small proportionof pairs nestssolitarily about six weeksbefore the main body. The latter breed in 70 colonies. 4. Somebirds, believed to be immature, build nestsin the autumn but do not lay eggsat that time. 5. The adult breedersapparently do not migrate, since the coloniesare occupiedalmost daily during the winter. 6. The sites of the larger rookeriesseem fairly permanent from year to year, and there is evidencethat breeding birds remain faithful to their own colony. 7. Display includes several dramatic threat postures,more restrained courtshipbehavior, bill stroppingduring coition, and an aerial display the significance of which is not clear. 8. Both sexes incubate alternately, the males taking the first shift. Total period is 58-61 days. 160 WARI•A•, Biology of Giant Petrel [ Vol.Auk 79

9. Both sexesfeed the chick, which is left unguardedat 15-24 days, reachesadult weightat 50 days,and flies at 102-117 days. 10. About 0.1-0.2 per cent of nestsheld two eggsor two chicks,prob- ably the productof one female. No pair fledgedtwo chicks. 11. Variations in the plumageand in the colorsof the soft parts are described.Evidence is presentedthat the solitarybreeders are very old birds. 12. Chicksare fed largely on seabirds. A form of socialhierarchy exists amongmembers of groupsat a food source. 13. Oil spitting is well developedin the chick after it is no longer guarded. Somechicks and many adultsare unableto spit, and there is a declinein the chick'sability as it feathers. LITERATURE CITED BIERMA•WV,W. H., and K. H. Voous. 1950. Birds observed and collected by the whaling expedition of the "Willem Barendsz." Ardea, Suppl. 123 pp. CARRICX,R., and G. M. Do•wVEw. 1954. Breedingof the Fulmar Fulmarus glacialis. Ibis, 96: 356-370. DOW•ES,M. C., E. H. M. EAnEY,A.M. GWY•Wv,and P.S. You•o. 1959. The birds of Heard Island. ANARE Rep. Ser. B. 1. 135 pp. FAnnA,R.A. 1937. Birds. BANZAR Exp. 1929-1931. Rep. Ser. B. 2. 288 pp. FISHER, J. 1952. The Fulmar. Collins, London. 496 pp. HOWARD,P.F. 1956. Banding of Giant Petrels at Heard and Macquarie Islands-- II. Emu, 56: 401-404. I•OHA•, S.E. 1959. Banding of Giant Petrels by the Australian National Antarctic Research Expeditions, 1955-1958. Emu, 59: 189-200. LAW, P. G., and T. BURSWAnk.1956. Macquarie Island. ANARE Interim Rep. No. 14. 48 pp. MURPHY, R.C. 1936. Oceanic birds of South America. New York. 1245 pp. MARSHAnk,A. J., and D. L. SERVE•WY.1956. The breedingcycle of the Short-tailed Shearwater Puffinus tenuirostris (Temminck), in relation to trans-equatorial migration and its environment. Proc. Zool. Soc. Lond., 127: 489-510. PAUnIA•, P. 1953. Pinnep&des,Cdtac•s, Oiseaux des Iles Kerguelen et Amsterdam. M•m. de L'Inst. Sci. de Madagascar, 8: 111-234. PR•VOST,J. 1953a. Notes sur l'•cologiedes p•trels de Terre Ad•lie. Alauda, 21-' 205 -222. PR•VOST,J. 1953b. Notes sur la reproducticn du fulmar antarctique Fulmarus glacialoidesA. Smith. A!auda, 21: 157-164. PR•vosw,J. 1958. Note compl•mentaire sur l'•cologie de p•trels de Terre Ad•lie. Alauda, 26: 125-130. SnADE•,W. J. L., and W. L. N. T•cxEnn. 1958. Antarctic bird banding by the Falkland Islands DependenciesSurvey, 1945-1957. Bird Banding, 29: 1-26. TicxEnn, W. L. N., and C. D. SCOTn.a•D. 1961. Recoveries of ringed Giant Petrels Macronectesgiganteus. Ibis, 10•a: 260-266. •VARH^*•,J. 1958. The nesting of the shearwaterPuffinus carneipes.Auk, ?$: 1-14. Antarctic Division, Department of External Affairs, 187 Collins Street, Melbourne, Australia.