FORAGING BEHAVIOR OF GENTOO AND CHINSTRAP AS DETERMINED BY NEW RADIOTELEMETRY TECHNIQUES

WAYNE Z. TRIVELPIECE,JOHN L. BENGTSON,1 SUSAN G. TRIVELPIECE, AND NICHOLAS J. VOLKMAN PointReyes Observatory, 4990 Shoreline Highway, Stinson Beach, California 94970 USA

ASSTRACT.--Analysisof radio signalsfrom transmittersaffixed to 7 Gentoo( papua) and 6 Chinstrap (P. )penguins allowed us to track penguins at sea. Signal charac- teristics allowed us to distinguish among 5 foraging behaviors: porpoising, underwater swimming, horizontal diving, vertical diving, and restingor bathing. GentooPenguins spent a significantly greater portion of their foraging trips engaged in feeding behaviors than Chinstraps,which spent significantly more time traveling. Gentooshad significantly longer feeding dives than Chinstraps(128 s vs. 91 s) and significantlyhigher dive-pauseratios (3.4 vs. 2.6). These differencesin foraging behavior suggestGentoo and Chinstrappenguins may have different diving abilities and may forage at different depths. Received3 June 1985, accepted24 April 1986.

THE trophic relationships among Pygoscelis among behaviors during foraging trips. This penguins,the Ad61ie(P. adeliae),Chinstrap (P. method improved our understanding of pen- antarctica),and Gentoo (P. papua),have been a guin feeding efficiencies,ranges, and traveling major focal point of researchin recent years, speeds,and permitted preliminary compari- particularlywith respectto the ecologyof their sons of Gentoo and Chinstrap forag- major prey , (Euphausiasuperba). To ing behaviors.

date, however, our knowledge of the ' METHODS feeding ecologyis largely derived from stom- ach samples obtained ashore (Emison 1968; This studywas conductedat a breedingrookery at Croxall and Furse 1980; Croxall and Prince Point Thomas, King George Island, South Shetland 1980a;Volkman et al. 1980, 1986;Lishman 1985). Islands, Antarctica (62ø10'S, 58ø30'W), between 26 Diving depth is one aspectof penguin for- January and 12 February 1984. We attached radio aging behavior that has been investigated in transmitters(1.5 cm in diameter, 5 cm long, 25 g) to the backsof 7 Gentoo and 6 Chinstrap penguins rear- somedetail. Multiple depth recorders,logging ing chicks. We securedthem to back feathers with the number of dives within set depth ranges, Devconepoxy and two plastictie-wraps. A 30-cmlong have been deployed on King (Aptenodytespata- antennacurved upward from the penguin'sback to gonica;Kooyman et al. 1982), Chinstrap (Lish- assureit was well above the water when the penguin man and Croxall 1983),and Gentoo (Costapers. surfaced.We triangulated penguins' locations,at 15- comm.) penguins. Maximum diving depths min intervals, with radio receivers (164 MHz) from have been reported for Emperor (A. forsteri; two huts, 200 m above sea level and approximately Kooyman et al. 1971), Black-footed(Spheniscus 3 km apart. We used a null-peak, 4-element double demersus;Wilson and Bain 1984), and Gentoo Yagi antenna receiving system.A third receiver was (Adams and Brown 1983) penguins. Feeding coupled to a strip-chart recorder and continuously monitored a single penguin during its foraging trip. range also has been investigated, but indirect- Signalsfrom foraging penguinswere receivedonly ly, using nest relief intervals (Williams and when the bird was on the surface and the antenna Siegfried 1980,Ainley et al. 1984, Croxall et al. was not submerged.Thus, strip chartsprovided pro- 1984). files of the surface vs. underwater time during pen- We report a new method of tracking pen- guins' foraging trips. Penguins were tracked be- guins at sea that allowed us to differentiate tween 0900 and 2300, and the two species were alternatedover the studyperiod. Krill was apparent- ly plentiful in Admiralty Bay during this time as all * Present address: National Marine Mammal Labo- penguinsforaged exclusivelyin the bay, within 10 ratory, NOAA/NMFS, 7600 San Point Way N.E., Se- km of the rookery. attle, Washington 98115 USA. All data are expressedas means + standard error.

777 The Auk 103: 777-781. October 1986 778 TI•IVEI•PIECEœTAL. [Auk,Vol. 103

Statisticalanalysis was done using the Mann-Whit-

ney U-test or G-test with Yates correction. •rRAC#INGNUTS MINUTE FIXœS •rOIiAG•ld G PATH RESULTS

Foragingbehavior patterns.--The foraging track of 3875 and sectionsfrom its chartrecording are illustratedin Fig. 1. We were ableto distinguishseveral types of diving pat- terns from the strip-chart records. We believe these indicate different behaviors such as rest- ing at the surface,two modesof traveling (por- poising and underwaterswimming), and two feeding behaviors(horizontal and vertical div- ing). Restingat the surfaceincluded time spent bathing. Porpoisingalways was used by penguins leaving from or returning to the immediatevi- cinity of the beach.Outward-bound porpoising often was followed by bathing, and both be- haviors were confirmed visually. Rapid por- poising, then bathing, is a constantfeature of Ad•lie Penguindeparture at Cape Crozier in the RossSea (Ainley 1972). Underwater swim- ming, the primary method of traveling usedby penguins, accountedfor 73% of all traveling time. It consistedof dives,averaging 50 s each, followed by surface pauses,averaging 12 s. Horizontal diving behavior followed under- Fig. 1. The foraging trip of Gentoo male 3875 in water swimming and precededvertical diving AdmiraltyBay. The penguin'spath wasplotted from in 5 of the 6 penguins for which we have com- compassbearings triangulated by radio and taken at plete records.Penguins that exhibit this for- 15-min intervals from the two tracking huts. Forag- aging pattern moved considerabledistances by ing behaviorswere copiedfrom a continuousstrip alternating periods of long dives with one or chart. Whenever the penguin was submerged,the two short dives. The short dives were of ap- signal was lost and the strip-chartpen returned to proximately the same duration as the dives of baseline;upon surfacing,the radiosignal caused the underwater swimming. We hypothesize that pen to deflect upward and remain there until the horizontal diving may be the primary method next dive. The raggedlook of surfaceperiods during horizontaland verticaldiving wascaused by changes of searchingbehavior used by penguins. The in the orientation of the antenna because of bird long dives of this behavior pattern may be to movement and from waves washing over the bob- explore the deeper layers of the water column bing penguin, which interfered with the signal's for krill. When no prey are located, the pen- propagation.The paper speedof the strip chart was guin moves to a new location by short "trav- 16 mm/min. Data on surface times and dive times eling dives" before diving again to searchfor were taken directly from the chartsby converting prey. Vertical diving, like underwater swim- linear distance to time. ming, was characterizedby regular intervalsof dive-to-surface pause times. The duration of dives and pauses were significantly longer, all underwaterswimming dives plus the short however, and the penguins remained in a lo- dives during horizontal diving behavior. Por- calized area during vertical diving (Fig. 1). poisingwas not consideredto be diving. For analysisof the foraging behavior of the Comparisonsof Gentoo and Chinstrap foraging be- penguins, feeding dives were defined as all haviors.--GentooPenguins spent a significantly divesduring vertical diving, plus the long dives greater portion of foraging trips engagedin during horizontal diving; traveling dives were feeding behaviors(horizontal and vertical div- October1986] PenguinForaging Behavior 779

ing) than in travelingbehaviors (porpoising and TABLE1. The percentage of foraging time spent in underwater swimming) than did Chinstrap different behaviorsby Gentoo and Chinstrap pen- Penguins (G = 5.47, P < 0.025; Table 1). Gen- guins.a toos made significantly longer vertical dives Foraging behaviorsb (%) than Chinstraps(128 + 4.5 vs. 91 _+4.9 s; U = Traveling Feeding 42.00,P < 0.005)and had significantlylonger Foraging maximum dive times (189 _+ 8.4 vs. 130 _+ 4.3 trip (h) P U H V R s; U = 12.00, P < 0.01; Table 2). Gentoos also Gentoo Penguinsc had significantlyhigher dive/pauseratios dur- 3894 F 5.7 5 12 53 21 ing vertical diving (3.4 + 0.2 vs. 2.6 + 0.2; U = UBM 7.1 3 24 6 60 6 36.50, P < 0.005; Table 2). Gentoo Penguins 3875M 5.5 6 15 34 44 1 made a mean 193 dives during 6.1 h of forag- Mean 6.1 5 17 31 44 3 ing, 90 of which (47%) were feeding dives; ChinstrapPenguins c Chinstrapsmade 182 divesduring 5.3 h, 74 of 3896M 4.8 9 18 11 57 4 which (41%)were feeding dives(Table 2). 3963 F 4.4 13 17 22 46 2 3893 F 6.8 10 47 7 32 4 We calculatedtraveling speedsfor penguins Mean 5.3 11 27 14 45 3 returning from feeding areas to the breeding rookeries.The mean (+SE) travelingspeed for aFrom analysesof strip-chart recordings of com- 3 Gentooand 3 Chinstrappenguins was 4.5 + plete foraging trips. bp = porpoising,U = underwater swimming, H = 0.4 km/h (Table 2). Penguin return trips varied horizontal diving, V = vertical diving, R = resting from 20 to 64 min. Strip-chartrecordings as- or bathing on surface. suredus penguinswere traveling (i.e. porpois- c M = male, F = female. ing or underwaterswimming) throughout their returns. tions from their histogramsshow that 445 of DISCUSSION 1,110 dives made by ChinstrapPenguins were recordedin the shallowestdepth range(0-7 or Usingradiotelemetry to trackpenguins at sea 0-10 m). Thisrepresents 40% of all dives,many and to distinguishamong their behaviorspro- of which may have been dives associatedwith vided detailed recordsof the activitiesof pen- traveling (i.e. porpoising and underwater guins during foraging trips and allowed two swimming). Dives during underwater swim- important general conclusions.First, a large ming accountedfor 44%of all dives by the pen- percentageof dives recordedusing multiple guins we studied. depth recorders(MDRs) in some other studies Penguins can swim 7.2 km/h (Kooyman et probably are not feeding dives. Therefore, al. 1971);however, the sustainedtraveling we MDRs could overestimate the number of feed- observed consisted of an underwater locomo- ing divesand underestimatethe foragingeffi- tory phasefollowed by a shortpause at the sur- cienciesof penguins.Second, traveling speeds face(see Fig. 1). Analysisof the dive/pause ra- of penguinsat seamay be considerablylower tios showed that penguins were swimming than the 7.2 km/h swimmingspeeds reported during only 76% of the time they were travel- for Ad•lies (Kooyman1975), and penguinfor- ing; the remaining time was spent on the sur- aging rangesbased on theseswimming speeds face.Calculating an overallspeed by taking76% may have been overestimated (Croxall and of 7.2 km/h results in an estimate of 5.5 km/h. Prince 1980b, Croxall et al. 1984). The 4.5 km/h meanreturning speeds observed Multiple depth recorderslog the number of were lower still, perhapsbecause calculations dives within preset depth ranges. Because were basedon straight-line distancesbetween Kooyman et al. (1982) deployed MDRs with feedingand landing areas,whereas penguins minimum thresholds of 5 m, their estimates of swam more erratically. feeding dives probably do not Comparisonsof Gentoo and Chinstrap for- include traveling divesand are likely to reflect aging behaviorsprovided new insights into feeding effort. Lishman and Croxall (1983), their feeding ecology.Gentoo and Chinstrap however, had no lower thresholds set on MDRs penguinsate predominantlykrill during chick placedon ChinstrapPenguins, because Chin- rearing;3-year meanswere 86.5%and 99.4%by strapsforage at shallower depths.Our calcula- wet massof all their food, respectively(Volk- 780 TRIVELPIECEET AL. [Auk, Vol. 103

TABLE2. Comparisonsof Gentoo and Chinstrap penguin foraging behaviors?

Total Total Total No. Mean Mean Maximum Travel foraging no. feeding feeding dive pause Dive/pause dive speed Birdc trip (h) dives time (h) dives time (s) time (s) ratio time (s) (km/h) Gentoo Penguins 3894 5.7 192 4.6 85 131 38 3.5 203 4.2 UB 7.1 231 4.3 92 108 41 2.6 174 5.3 3875 5.5 157 4.3 93 119 34 3.6 190 3.1 2245 .... 135 45 3.0 -- -- 3895 .... 132 30 4.3 -- -- LIB .... 128 31 4.1 -- -- 2219 .... 145 50 2.9 -- -- Mean 6.1 193 4.4 90 128'* 39 3.4** 189' 4.2 Chinstrap Penguins 3896 4.8 139 3.0 86 97 31 3.2 126 4.7 3963 4.4 158 3.3 69 103 47 2.2 139 3.9 3893 6.8 249 3.6 68 99 34 2.9 126 5.7 3885 .... 83 33 2.5 -- -- 2573 .... 71 40 1.8 -- -- 2605 .... 91 33 2.7 -- -- Mean 5.3 182 3.3 74 91 36 2.6 130 4.8 aData for the first 3 Gentoo and Chinstrappenguins are from entire foraging trips and include a strip- chart record.Dive and pausetime data for the remainingpenguins were collectedby timing a minimum of 10 consecutivedive/pause cyclesduring vertical feeding. b * = Differs significantlyfrom ChinstrapPenguin; Mann-Whitney U-test,P < 0.01. ** = P < 0.005. cUB = unbanded penguin.

man et al. 1986). Gentoo Penguins require sig- Penguins exceeded70 m and only 6% of the nificantly more krill to rear their chicks to dives exceeded 45 m (Lishman and Croxall fledging than do Chinstrap chicks (118 kg vs. 1983).A direct relationshipbetween body size 73 kg per breeding pair; Trivelpiece et al. in and diving depth has been postulatedfor pen- press)and have significantlyshorter nest relief guins (Stonehouse1967b). Consistent with this intervals during chick rearing (12.5 h vs. 16.7 idea, Gentoo Penguins are the largest of the h; Volkman et al. 1986). Foraging rangesbased pygoscelids,with mean adult massesduring on nest relief intervals, feeding times, and trav- the chick phase of 5.3 kg, compared with 4.0 eling speedsare estimatedas within 17 km of kg for adult Chinstraps(Volkman et al. 1980). the rookery for Gentoosand within 27 km for In the shallow-water regions of King George Chinstraps (Trivelpiece et al. in press). Thus, Island,krill aggregationsoccur in a broadlayer Gentoos require more krill per day for their from 10 to 80 m deep during the night, and chicks,and they acquire this food from a more descendto a daytime level between an upper restrictedforaging range. We suggestthat Gen- 30-60 m limit and a lower 90-120 m limit (Kal- tooscan do this becauseof their greater diving inowski and Witek 1980). Thus, whereas the abilities. deeper-divingGentoos can exploit any krill ag- Gentoosdove significantly longer and had gregationsthey locate, Chinstrapsmay be un- significantlyhigher dive/pause ratiosthan did able to feed effectively on deep krill swarms Chinstraps(Table 2). Dive/pause ratiosmay in- and therefore must spend more time traveling dicatephysiological diving abilities(Dow 1964), in searchof available prey. which in cormorants (Phalacrocoraxspp.) are correlated with feeding depths (Stonehouse ACKNOWLEDGMENTS 1967a, Ainley et al. 1981). Gentoo Penguins This work was supported by National Science have been caught in trammel nets at 100 m Foundationgrant DPP 81-17205.Radio transmitters, depth (Conroy and Twelves 1972) and are receivers,and other telemetry equipment were gen- known to dive to at least 135 m (Costa pers. erously loaned by the Cedar Creek Bioelectronics comm.); none of 1,110 dives by 4 Chinstrap Laboratoryof the University of Minnesota.We thank October1986] PenguinForaging Behavior 781

Dr. E, Feofiloff for loan of her strip-chart recorder, 191-212 in Antarctic bird studies (O. L. Austin, without which the behavior patternsof the foraging Ed.). Washington, D.C.ß Amer. Geophysical penguins would have gone undetected.We thank Union. personnelof the U.S. AntarcticResearch Program and KALINOWSKI,J.ß & Z. WITEK. 1980. Diurnal vertical Fuerza Aerea de Chile for logistical help. We thank distribution of krill aggregations in the west the PolishAcademy of Sciences,Professor S. Rakusa- Antarctic. Polish Polar Res. 1: 127-146. Suszczewski,and membersof the eighth Polish Ant- KOOYMAN,G. L. 1975. Behavior and physiology of arctic Expedition for their generous hospitality diving. Pp. 115-138 in The biology of penguins throughout our stay at Arctowski Station. D.G. Ain- (B. StonehouseßEd.). BaltimoreßUniversity Park ley, W. and P. Hamnet, E. C. Murphyß and W. J. L. Press. Sladen provided many helpful suggestionsduring --, R. W. DAVISßJ.P. CROXALL,& D. P. COSTA. the manuscript'spreparation. This is contribution 1982. Diving depths and energy requirements #277 of the Point ReyesBird Observatory. of King Penguins.Science 217: 726-727. , C. M. DRABEK, R. ELSNER,& W. ]•. CAMPBELL. LITERATURE CITED 1971. Diving behaviorof the EmperorPenguin. Auk 88: 775-795. ADAMSßN.J., & C. R. BROWN.1983. Diving depths LISHMAN,G.S. 1985. The food and feeding ecology of the GentooPenguin. Condor 85: 503-504. of Ad•lie and Chinstrap penguins at Signy Is- AINLE¾,D.G. 1972. Flockingin Ad•lie Penguins. land,. J. Zool. London205: Ibis 114: 388-390. 245-263. ßD. W. ANDERSONß& P. W. KELLY. 1981. Feed- --, & J.P. CROXALL.1983. Diving depths of the ing ecologyof marine cormorantsin southwest- ChinstrapPenguin. Brit. AntarcticSurv. Bull. 61: ern North America. Condor 83: 120-131. 21-25. ß E. F. O'CONNORß& R. J. ]•OEKELHEIDE.1984. STONEHOUSE,]•. 1967a. Feedingbehavior and diving The marine ecology of birds in the RossSea, rhythms of some New Zealand shagsßPhalacro- Antarctica.Ornithol, Monogr. No. 32. coracidae. Ibis 109: 600-605. CONROY,J. W. H., & E. L. TWELVES.1972. Diving 1967b. The general biology and thermal depthsof Gentoo Penguin and Blue-eyedShag balanceof penguins.Pp. 131-196in Advancesin from the South Orkney Islands. Brit. Antarctic ecologicalresearch (J. B. Cragg, Ed.). Londonß Surv. Bull. 30: 106-108. Academic Press. CROXALL,J. P., & J. R. FURSE. 1980. Food of Chin- TRIVELPIECE,W. Z., S.G. TRIVELPIECE,& N.J. VOLKMAN. strap Penguinsand Macaroni Penguinsat Ele- In press.Ecological segregation of Ad•lie, Gen- phant Island Group,. Ibis too, and Chinstrap penguins at King George Is- 122: 237-245. landßAntarctica. Ecology. ß & P. A. PRINCE, 1980a. The food of Gentoo VOLKMAN,N, J., K. JAZDZEWSKI,W. KITTEL,S.G. TRI- Penguinsand Macaroni Penguinsat South Geor- VELPIECE,& W. Z. TRIVELPIECE. 1986. Adelie, gia. Ibis 122:245-253. Chinstrapand Gentoopenguin diets during chick ß& . 1980b. Foodßfeeding ecologyand rearing at King GeorgeIsland, South Shetland ecologicalsegregation of seabirdsat SouthGeor- Islands. Condor 88: in press. gia. Brit. J. Linnean Soc. 14: 103-131. ß & W. Z. TRIVELPIECE.1980. ßC. RICKETTS,& P. A. PRINCE.1984. Impact of Diets of pygoscelidpenguins at King George Is- seabirdson marine resources,especially krill, of landß Antarctica. Condor 82: 373-378. South Georgiawaters. Pp. 285-317 in Seabirden- WILLIAMS,A. J., & W. R. SIEGFRIED.1980. Foraging ergetics(G. C. Whittow and H. Rahn, Eds.).New ranges of krill-eating penguins. Polar Rec. 20: Yorkß Plenum Press. 159-175. DOT, D.D. 1964. Diving times of wintering water WILSONßR. P., & C. A. R. BAIN. 1984. An inexpen- birds. Auk 81: 556-558. sive depth gaugefor penguins.J. Wildl. Mgmt. EMISON,W. B. 1968. Feeding preferences of the 48: 1077-1084. Ad•lie Penguinat CapeCrozierß Ross Island. Pp.