ASPECTS OF THE ECOLOGY OF SPECIES REPLACEMENT IN HUMBOLDT AND MAGELLANIC PENGUINS IN CHI1.E

Rory P. WILSON, David C. DUFFY, Marie-Pierre WILSON and Braulio ARAYA

INTRODUCTION

Interactions between allopatric species at the edgw of their ranges have been of major interest in studies of biogmgraphy and evolution (Mayr 1963, Endler 1977) and of competition (Macanhut 1972). However. few such studies have been done on seabirds (e.g. Stresemann and Timofeeff-Ressovsky 1947, Udvardy 1963, Smith 1966, Devillers 1976, Devijlers and Tenchuren 1978). Seabirds are panicularly appropriate for such studies as they often nest in discme, conspicuous coloniu; their distributions can be relatively easily mapped at sea, and their marine environments are often well-studied by marine biologists, fisheries scientists and oceanographers, pmviding objective data on environmental limits ta mnges.

Unfortunately, because of the great distances involved and the resulting logistics, it is often difficult to conduct comparative research, so interferences are drawn from several studies. perhaps with differing methods, or ftom studies ar different localities by the ~ a m e author, made at different times. Differences between species may then be the result of the methods used or of different ecological conditions during the studies.

During October and November 1985, we had the opportunity to study a species-replacement, that of the Humboldt Penguin Spheniscus humboldri by the Magellanic Penguin Spheniscus mugellanicus along a latitudinal gradient extending from 29"s to 53% We used a 'snap shot' approach of quick visits to several sites over a shon period of time (cf. Ricklefs rf a!. 1984). Although the time we could consacrate to the study was limited, we hoped to elucidate 'seasonal' differences along the gradient of inneasing latitude that might offer insight as to w h y t h e H u m b o l d t P e n g u i n i s a subaopicaVtemperaE species and the Magellanic Penguin temperatelsub-Antarnic. Such shon term

studies are likely to be inadequate where interspecific differences are minimal. but can serve to indicate species-spaific environmental adaptation and competition avoidence mechanisms where interspecific differences prove to be substantial. Some of the results presented here have been presented in previous publications (Wilson CI ol.

1989, Wilson and Wilson 1990). in the context of other research questions.

The Humboldt Penguin breeds on the Pacific coast of South America from Isla Foca OSO1?'S, Peru (Hays 1984a) south to Isla Pupuya 3PWS Chile (Araya 1983). The Magellanic Penguin breeds from 32'35's (Bike 1977) and exceptionally as far nonh 29'013 (Araya 1983) on the Pacific coask south through Tierra &I Fuego and the Falkland Islands, and north on the Atlantic coast to approximately 4 2 5 (Murphy 1936, lohnsan 1965. Blake 1977, Bwnma 1990). The two species overlap unly in the Pacific, as breeders in central Chile from 32'35's to 24'00'S and as non-breeders from approximately 3 2 5 to 40°S (Blake 1977).

Murphy (1936) suggested that h e two penguins differ in breeding season and tolerance of water temperatures: the Magellanic Penguin breeding in s u e r , the Humboldt Penguin breeding throughout the year. Average ocean temperatures i n the breeding range of the former appear to be below 12.8OC and those in the range of the latter. well above. Johnson (1965) stared that the two spcies differ in nest-site preference; in areas of overlap, one species nests on one side of an island and the second species on the other side.

SWDY AREA AND hlETHODS

We observed Humboldt and hlagellanic penguins at four sites on the Chilean coast; (I) isla Chairaral (29OOI'S. 7l050'W) (Araya and Duffy 1987). a colony of spproxirr~arely 3.D00 pain

.w WU.SON. R.P., DUFFY, D.C., WILSON.M..P.ard ARAYA. B. G m A m 85

of Humboldt Penguins (study dates: 30 October - 3 Sovember 1985jJ2) lsla de Pinguinos at Alganobo (?3"39'S. 7 I 0 4 5 ' W ) , where the re were approximately 120 pairs of Humboldt Penguins and seven pairs of Magellanic Penguins (dates: 27 October and 5-8 November 1985). (3) Isloles Piflhuil (41'433, 74'02'W). two islands off the west coast of Isla Child where one island had approximately 10 pairs of hlagellanic Penguins and the other had 50 pairs of Magellanic Penguins and 20 pairs of Humboldt Penguins (dates: 13 and 16 November 1985). (4) Orway Sound (53°03's, 71°14'w-j, nonh of Punta Arenas. a colony of approximately 200 pairs of Magellanic Penguins (dates: 18-24 November 1985).

At each breeding site, we noted nesting stages of both species, when present. We also examined the liierature to determine monthly nesting activity by the two species throughout their ranges.

Aduh penguins bmoding small chick were removed from their nests and fitted with autoradiographic devices that measured swimming speed and time spent swimming at each speed underwater (and therefore distance swum) (details of design and functioning in Wilson and Bain 1984, Wilson 1985% Adams 1981. A h n s and Wilson 1987. Brown 1987). Attached to feathers along the midline of the back with hoseclamps (Lishman and Croxall 1983), the devices were accurare to within 10% in determination of foraging trip duration and measured swimming speed to within 0.1 kmlh (Wilson and Bain 1984).

We also measured the maximum depth of dive dwing foraging, thmugh use of capillary tubes (internal diameler 1.2 mm), coatcd internally with icing sugar (e.g., Adams and Brown 1983, Burger and Wiison 1988) and taped to the outside of the speed/distance meters. Errors in ihe pmcision of the depth measurements are considered less than 10% (Burger and Wilson 1988). These devices together have a cross-sectional area approximately 2% that of Humboldt and Magellanic penguins, so that the effect of the devices was similar for both species and considered to induce a reduction o f approxima~ely 0.lmls in mean swimming speed (Wilson er ul. 1986. cf. Wilson and Culik 1992). Owall. we fitted a total of 62 speed/distance/depth

devices on Magellanic and Humboldt Penguins; 22 at Chafiaral (Humboldt), 22 at Algarrobo (Humbaldt) and 18 at Otway Sound (Magellanic).

Where possible, we sampled diets of adults of each species at each site, using a wet offloading stomach pump (Wilson 1984). We selected buds returning from the sea with protruding stomachs and clean beaks and plumage, conditions indicative of hi i s recently rehuned from foraging. Random sampling of a d u b at colonies is disruptive and usually pFaluces stomach samples of only bile and otoliths insufficient for analysis. We sampled a total of 36 Humboldr Penguins; 17 individuals at Chafiaral, 15 at Algarrobo and 4 at Child, and 35 MageUanic Penguins; 16 at Chiid and 19 at Otway Sound.

Stomach contents werc suained and placed in plastic bags, then preserved, within fwr h, in alcohol. The samples were sorted within 24 h and p ~ y length (fork length for fish, mantle length for cephalopods, total length for crustaceans) was measured d i i c r l y when possible. All specimens were identified to the lowest possible oxonomic division possible, and voucher specimens retained for confmtion. In highly digested specimens, we removed and swred the otoliths for subsequent identification.

We determined pauerns of arrival and deparIure of foraging birds by conducting 24-h watches of landing sites at ChaAaral, Algarmbo and Otway Sound, recording all birds enlering or leaving the water, as well as numbers of birds in 'beach parties' (cf. Wilson era!. 1989% Wilson and Wilson 1990) and the number arriving or departing to inland nest sites. We could not monitor arrivals and deparmres at Chilo6 because of limited access to the islands during bad weather. At Otway Sound, many Magellaaic Penguins returned during 0500 - 0700 hr. These birds did not have promding stomachs. suggesting that they had not fed recently and had been bathing. Similar bathing was observed at Algmobo and C h ! in he early morning. When birds could bc identified as bathing, they were excluded fmm counts of arriving and depaning bids.

During the 24-h watches and other observation pricds, we recorded interspecific and inmpecific

IW5 GERVALK . SPECIES REPLACEMENT IN HUMBOLDT A N D MAOEL1,ANtC PENGUINS 5 I

interactions of penguins in 'beach parties'. We used contingency tests (Harper 1977) lo examine the relative number of aggressive interactions initiated by Hunlboldt and Magellanic pengins. Once daily, we also scored the number of each species occupying central or pripheral positions and the species' identities of their nearest neighborn. Contingency tests were used to examine the relative proponions of the two s p i e s at the periphery and centre of groups. At C h i l d , we noted group s u e and species composition of penguin flocks returning to the island alter foraging.

RESULTS

T h e l i t e r a t u r e had s u g g e s t e d t h a t t h e northernmost site, Chanaral, would be likely exclusively occupied by Humboldt Penguins, with Alganobo having both species, and the two most southerly sites, C h i l d and Otway Sound, being exclusively occupied by Magellanic Penguins. In fact. Magcllanic Penguins were seen at ChaRaral and Humbfdt Penguins at Chi ld . The C h i l d site represented a 550-km sourhem extension of rhe known breeding range of the Humboldt Penguin (Araya and Millie 1986).

PATTERNS OF ARRIVAL AND DEPARTURE

Differences in foraging periods appeared to be site-specific, rather than characteristic of each species. At the three colonies where we conducted 24.h watches (Chaflaral, Algarrobo and Otway Sound). moa penguins left to forage just after dawn and returned during late afternoon and evening (Fig, I).

Of the 62 penguins fined with speed/depth devices. 47 went to sea during the study periods. Thiny-eight of the devices were recovered, two birds desert4 their nests, and we could not catch seven birds.

Twenty-seven of the depth gauges could be analyzed for maximum deprh; four of h e &vices were broken by the penguins; one was lost; and six

were discarded because of condensation problems within the lumen o l the capiliary (cf. Burger and Wilson 1988). Twenty-three of the speedidismce meters functioned well; one device was lost; two were broken, and twelve units were destroyed by seawater leaks (cf. A d a m 1987).

Mean speed of Humboldt Penguins at ail sites combined (Z = 6.8 kmh. SD - I.?. t i = 1 I) was not significantly different to that of hlagellanic Penguins (f = 7.6 kwh, SD = 1.3, N = 10) (1 - 1.467. P r 0.05). Humboldt Penguins at Isla Chabral did not differ (1 = 0.934, P > 0.05) in their maximum depth of dive (n = 62.: rn. SD = 13.7. N - 9) from Mngelianic Penguins at Punta Arenas (f = 57.8 m, SD = 7.8. N - 12). but both differed significantly (t = 7.382 and 5.34?, respectively, P < 0.05) from those of Humboldt Penguins at A l g m b o (K = 27.3 rn. SD - 9.2. h' - 6).

Mean and median distances traveied also appeared to be site rather than species-specific. Humboldt Penguins at Isla C h a a l (i - 32.2 km. SD = 25.9, median = 22.5, N - 81 and Magellanic Penguins at Otway Sound (x - 23.8 km SD - 28.9, median = 17.1, N = 10) had similar foraging ranges (t = 0.565). Neither was significantly shoner (t - 1.294 and 1,727, respectively, P > 0.05) than those a1 Algarrob (r = 50.6 km. SD = 23.2, median = 44.1, N = 5).

DIETS

Diu varied more between sites than between species. based on both relative aggregate numencal abundance and percentage frequency of occurrence (Table I ) . For example, at Chaaaral. Penguins took mostly g d i s h (Scombere.wr spp.). while their main prey at Algarrobo and Chi106 was anchovy (Enxraulis r ingens) . hlagellanic Penguins took mostly anchovy at C h i l e and herring (Sprarrus fuegensis) at Otway Sound. The main prey of both species at Chilot was anchovy, with sardines (Sordimppr sogar) accounting for the rrst o f the Humholdt diei and sardines and squid (Tudurodes fillipprtvue) that o f the Slagellanic diet.

Prey ranged from 35.6 mm (anchovy) to 270 mnt (garfish) for Humboldi Penguins and 25 mrn (squid) to Id? mm (anchovy) for hlagellanic

1995 GIERVALK SPECIES =LACOMEKT W HVMBOLDT AND MAGELLANK PENGUINS 53

Tabk i. Relative abuodanct (Col. 1) and pmrotage trequency of abundance (Cul. 2 ) of prey of Eiumbddl and Mageihnic Penguins in central and soulhero Chile, Hovernhr 1985.

Humholdr Penguin

Locality Scornbereso Engraulis Sardiaop Sprarms Squid Other

Chaiiaral 94 94 1 I 2 1 6 0 0 0 0 2 6 Algarrobo 0 0 72 100 4 7 0 0 144 1 0 6 Child 0 0 BO 1W 20 100 0 0 0 0 0 0

Magellanic Penguin

Chile 0 0 61 71 6 25 0 0 19 56 14 2S W a y Sound 0 0 0 0 0 0 94 100 6 1 6 0 0

Pcnguins. The two species look anchovy of the same size (Humboldt: x - 109 mm, SD = 15.0, range - 83-143, N = 19; Magellanic: n - 114 mm, SD = 13.4, range 94-142, N - 40, Z - 1.237, P 0.05) and pilcbard (Humboldt: X - I02 mm, SD - 30.4, range 55-126, N - 7; Magellanic: a - 97 mq SD - 20.9, range - 59-141 mm. N-7; Z - 0.358, P z 0.05) a t Chi ld .

BREEDING SEASONS

Our field data (Table 2) showed little difference in nesting stages of Humboldt Penguins between ChaAaral and Algarmh; both had even distributions of nesting stages from eggs to large downy young. At Otway Sound, most of the Magellanic Penguin nests had very small ywng or hatching eggs. At Child, nesting of Humbddt Penguins appeared more advanced than that of Magellanic Penguins. Half of the Humboldt nests had large downy young;

none had eggs. In contrust. small downy young were the commonest contents of Magellanic Penguin nests; eggs were present and there were no large downy young.

Literature repons of nesiing (Appendices 1 & 2) show that Humboldt Penguins nest thmughout the year, with b& winter and spring peaks, while MageUanic Penguins show a strong summer peak. Unfortunately, most o f the Humboldt dam come from Peru and most of the Magellanic &la from Patagunia and the Falklands, with little informarion from the area of overlap which we srudied.

INTER AND INTRA-SPECLFIC lNTERACTIONS

Aggression rates differed significantly beween Humboldt and Magellanic penguins at both A l g m b o and Chilo4 fx2 - 16.9 and x1 = 33.3. respectively, df = I , P < 0.01). At Algmobo. in a

Table 2. N~sting stages or Eumholdl and hIagellonic Penguins breeding a1 diNewnl localities on the Chilean coast during November 1985.

Nest contents (5)

Locality Species Eggs Smdl Medium Large Pre-fledge Tolal fu l l chicks chicks chicks chicks nests

Chanaral hum bold^ 37 26 19 18 0 27 Algmobo Hurnholdt 31 23 n 19 0 26 Chiloe Hurnbldr 0 10 40 SO 0 10

Magellanic 15 44 41 0 0 34 Ofway Sovnd Magellanic XI 59 21 0 0 38

52 WILSON. R.P.. D W P Y , D.C.. WLIJON. M-P. end ARAYA. R. GERFALT 85

beach group of 96.6% Humboldt and 3.4% Magellanic penguins ( N - 178). Hun~boldt Penguins initiated aggression against Magellanic Penguins five times more often than the reverse (total encounters = 14, Humboldt > Humboldt. 0; Huntboldr > Magellanic. 10: Magellanic > I-lurnboldt, 2; Magellanic > Magellanic, 2). Magellanic Penguins were equally aggressive ro conspecific and to Humboldt Penguins, but liumboldt Penguins were not aggressive to conspecifics. At Child, in a beach group of 96.4% Magellanic and 3.6% Humboldt penguins (n = 140). hfageilanic Penguins were almost as aggressive to Humboldi Penguins as Hurnboldt Penguins were to Magellanics (ratio 0.8 to 1.0; total encounters = 38; Humboldt > Humboldt 2; Humboldt . Magellanic, 2 0 Magellanic > Humboldr, 16; Magellanic > Magellanic. 0).

At ChaAaral and Algarrobo. Magellanic Penguins occurred significantly more frequently at ~e peripheries of beach parties than did Humboldt Penguins (Chaharal: X' = 12.1, df- 1; P < 0.01; Humboldt edge. 177; middle, 272; Magellanics edge. 8; middle, 0; Algmbo: X' - 5.8, df - 1; P c 0.05; Humbotdts edge, 99; middle 66; Magellanics edge, 9; middle. 0) but at Child, the reverse occurred: Humboldr Penguins occurred significantly more Frequently at the periphery than did Magellanics (x' = 3.9, df = I; P c 0.05; Humboldtr edge, 8; middle, 5; Magellmics edge, 103; middle, 195). The two species tended to form monospecific groups within beach pmies. Bolh bad conspecific nearest neighburs more o k n [han expected by chance (P r 0.01; df = 1; Alganob, X' 92.6; Humbldl - Humboldt, 236; Humboldt - htagellanic, 4; Magetlanic - Humboldt, 5; Magellanic - Magellanic, 9; Child: x2 = 64.4, Humboldt - Humboidt, 7; Humboldt - Magellanic. 2; Magellanic - Humboldt, 2; Magellanic - Magellanic. 97).

Of 82 groups of penguins returning to Child after foraging, 55 consisted of one bird, 16 of two, five of bee, three of four, and three of five. Gmup size was not significantly different between Humboldt and Magellanic penguins (x' - 2.6, df = I , P > 0.05). Although five mixed-species groups occurred, there was a strong tendency for k k s to be monospecific (e.g., of 16 floclvr of two birds. 12 flocks were exclusively Magellanic and four were exclusively Humboldt. The probability of such monspccific groupings occurring by chance is

tess than 0.001 (based on Humboldts consti~ting 25% and hiagellanics 75% of total individuals observed).

DISCUSSION

Ow study encountered a much greater area of overlap in the ranges of Humboldt and Magellanic penguins than had been dewribed in the literature. This overlap suggesrs that lhe two species do not replace each other i n response to any sharp environmental gradienh such as water lernpererure. Relauve abundance of penguins changed fmm C h a m and Alganobo, with a preponderance of Humboldc Penguins. to Child with a preponderance of Magellanic Penguins. The change-over area provided a valuable opportunity to examine differences berween the two species that might account for their gradual replacement of each other.

Our dam showed no obvious differences in foraging rhythms or foraging behavior between the two species. AU Spheniscus penguins sNdied to date seem to have a nrain peak i n departure from the colonies aruund dawn with birds returning around dusk (Ross 1971, Boersma 1977, Wilson and Wilson 1990). This feature seems invariant of locality (cf Hays 1980b. Wilson and Wilson 1990. Sfolam and Subaro 1991) and is probably due to Ihe fact drat most penguin species stem to need light in &r to forage (Wilson er al. 1993). Foraging performance was similar OI that of the African Penguin, Spheniscus demersus. a close but geographically-isolated congener (mass - 3.0 kg, Williams 1995; speed, R - 6.7 km/h. uncorrected for measurement effects. Wilson er al. 1986a: distance uavelled, x = 40.2 km mean maximum depth, approximately 35 m, excepionally to 130 m, Wilson and Bain 1984. Wilson 1985a, Wilson and Wilson 1990, Wilson and Wilson 1995) which also feeds on anchovy (Wilson 1985b, Duffy and Siegfried 1986). Scolm er al. (in press) recorded that the mean swimming sped of 30 Magellattic Penguins returning to their colony at Punta Lobcria was 6.8 kmm and in a review of penguin swimming speeds in general, Wilson (1995) noled that aU species travelled ac very similar speeds, although there was slight rend for larger s p i e s to swim faster. Given the similarity in body masses of Magellanic and Humboldt Penguins, it is thus unlikely thac species-specific differences in swim speed exist.

1995 GIERVALK SPECLES RePLACBMEtfT W HUMKILT AND MAGELLANIC PENGUINS 55

The general similarity of the three species suggests that Sphenisctu penguins am either unable or do not need to diverge horn one another in foraging techniques, either in allopatry or along the coast of Chile, in the single area of sympatry between two members of the genus. Examination of foraging performance of the geographically-isolated Galapagos Penguin (Spheniscus mcndiculus) would be useful to test {his assumption.

Nesting habitats also appear generally similar throughout the genus. Magellanic, Humbofdt and African penguins nest in a wide variety of habitats: under bushes. in burrows, and under mks. Unlike the other two, Humboldt and Galapagos pengains do not usually nest in the open, perhaps as a result of intense insolation at lower latitudes. La Cock (1988) has shown that surface-nesting in A6ican Penguins is more frequent during the cooler and more overcast austral winter a 33%. suggesting that insolation is likely to contrain surface nesdng of Humbddt and Magellanic penguins at Chanaral (29"s) and Algarrobo (33' 305). our two noRbem shrdy sites.

Nesting seasons diverged sharply between Magellanic and Humboldt penguins, suggesting that this factor is partially responsible for their differences in range. The year-round breeding and winter breeding penk of the Hwnbldt Penguin are similar to those of the African Penguin (La Cocker ul. 1987). raiher than to the suictly summer-nesting regime of the Magellanic Penguin. The Galapago$ Penguin, under the influence of the southern hemisphere Hwnboldt Current, also nesb ptimarily during the ausval winter, although nesting may occur at any time of year (Boenma 1978).

The summer breeding of the Magelianic Penguin appears an adaptation to nesting in the far south. During the austral winler at Punta Arenas, only six hours are available for diurnally foraging penguins to obtain sufficient food for themselves and their Offspring. African Penguins require c a 10 h per day to forage for large chicks (Wilson and Wilson 1990), suggesting that winter breeding would not be possible, if Humboldt and Magellanic penguins have similar foraging time demands. On the other hand. at the nonhem edge of the Magellanic's range, ambient air and ground rernperanves may simply be too hot u, allow summer nesting by this species.

While this may explain the present situation, it d u s not explain why Humboldt Penguins have not evolved to be summer nesters in the southern portions of their range and why Magellanic Penguins do nrr breed in winter fanher north We suggest four explanations. First, winter-breeding Magellanics may in fan be common, but our held work and most visits to the area have occurred in summer, so we have underestimated the extent of Magellanic nesting activity a1 Chanaral and Algmbo. Similarty, winter breeding by Humboldt Penguins may occur a1 Chitoe, despite the heavy winter rains (Darwin 1933. Pickard 1971) which would be expected to inhibit burrow-nesting, because of flooding (cf. Stokes and Boasma 1991).

Second, the time since the two s ~ c i e s came inm contact with one another and with new marine environments along the west coast may have been too shan to allow a shift in the nesting season of the Magellanic Penguin. If we assume, based on their present ranges, that the 'iumboldt Penguin originated on the Pacific coast of Latin America and the Magellanic Penguin on the Atlantic coast. after being geopraphically isolated during one of the Pleistocene glaciations of southern Chile [cf. Devillers 1978, Devillers and Terschuren 1978, Livezey 1986, for other laxa in the area). then the two species of penguin would have come iota contact only once the ice layer remated from the coast nnd the Magellanic Penguin was able to expand south and west.

The last extensive glacial blockage of the Chilean coast occurred as recently as 13,000 B.P. (Mercer 1976) and conwt between the two species may have occurred even later. leaving the distribution and ecologies of the two species still i n flux today. Interestingly, Boersma ef 01. (1990) note that Atfantic populations of Magellanic Penguins are st31 expanding nonhwards, although the antiquity of this trend is unknown. The Magelianic Penguin may also be expanding its range northward on the Pafific coast.

Third, where one species is rare, it may be forced to forage under less favorable conditions than can be expfoired by the more common species. because of inlerspecific interactions. Although Magellanic and Humboldt penguins occur together, our data suggest rhat interspecific aggression results in a tendency to form monospecific aggregations.

% WILSON, R.P., D W . D.C.. WILSON. and ARAYA, 0. CfRFAVT 85

both on land and a! sea. Spheniscus penguins are often gregarious at sea (Siegfried er a / . 1975. Hays 1984b. Wilson er al. 19864). Seleaive aggression by adult African Penguins at sea reduces the numbers of juveniles in adult foraging groups, forcing the young bids to forage less efficiently than do adults (Ryan el al. 1987). In areas where either the Humboldt or Magellanic penguin is much more abundant than the other. similar aggressive behaviour may exclude the rarer species f m m foraging groups.

Finally, competition for nesting sites may pment shifts in breeding season. If each species is dominant where more abundant, this would provide a continuing mechanism to prevent invasion by the less-numerous species. However. while burmw use at Chi106 was heavy, so that interspecific competition could have prevented use by Hwnbaldt Penguins. empty nest sites were common a Algambo, Otway Sound and Chamal.

We cannot determine the relative imponance of these four explanations k a u s e we lack data on historical consuainu. the importance of competition. and on the abundance of the two species throughout the year in their area of sympatry. Further study is necessary at colonies where the relative pmpnions of the two species are intermediate, such as Isla Sanla Maria 365 and lsla Mocha 38%.

Indirect evidence may come from comparisons with detailed ecological studies of other species pairs in southern South America, such as diving-petrels (Pelecanuides spp.) (Johnson 1965). skuas (Carharacra spp.) (Devillers 1978), cormorants (Phalacrocorax spp.) (Devillen and Terschnren 1978, Rasmussen 1991) and steamer-ducks (Tachyeres spp.) (Livezey 1986, Livezey and Humphrey 1986). Even in the absence of such studies, our work suggesrs that the causes of species-replacements may be conhplex and multiple, and that species relationships may be unsurble over time, as suggested by Wiens (1977) for competitive interactions between species.

ACKNOWLEDGEMENTS

Our field work in Chile was assisted by M. B d ,

I. Jordan. R. Navarro, C. Velasque, and C. Venegas, the fishermen of Caleur C h M d , CONAF officials on Chilo&, and the management and memberr of the Confradia Nautica del Pacifico. W. R. Siegfried pmvided necessary sanity during our stay in Punta Arena. N. Klages, B. C. L.ive~ey, W. R. Siegfried and F. Todd provided valuable information or comments. This paper is based in pan on the FiaPatrick Institute's 25th Anniversary Expedition to Chile.

SUMMARY

Humboldt Sphenisc~u humbold~i and Magellanic S Nigellanicus penguins exhibit a wne of syrnparry along the west coast O F South America. We examined aspects of their foraging ecologies at four sios (three in the zone of sympauy) along the Chilean w t between 29O and 5 3 3 and found rhat inter-site differences were more pronounced than inter-specific differences. In contrast, the two species differed in their breeding seasons.

LlTERATURE CITED

Abbot, C. C. 1860. The penguins of the Falklands. Ibis 2: 336.338.

Adams, N. 1. 1987. Foraging range of King Penguins Apmtodytes poragunicus during summer at Marion Island. J. 2001. London 212: 475482.

Adams. N. I. and C. R. Brown. 1983. Diving depths of the Gentoo Penguin Pygoscelis papua. Condor 8 5 503-504.

Adams, N. I, and M.-P. Wilson. 1987. Foraging parameters of Gentoo Penguins Pygoscelis papa at Marion Island Pol. Aicd. 7: 51-56.

Araya M. B. 1983. A preliminary r e p n on the statw and distribution of the Humboldt Penguin in ChilaPp. 125-135 in I. Delacour. Proc. I.F.C.B. Symposium on breeding birds in captivity, Los Angeles.

Araya, M. B. and C. Millie, 1986. Guia de c a m p de 10s aves de Chile. Santiago. Editorial Univetrimria

Amya. M. B. and D. C. Duffy. 1987. Animal introductions to isla Chafzaral, Chile their history and effect on seabirds. Cormorant 15: 3-6.

1995 GlERVALK SPECIES E3'LACEMENT IN HUMEOLDT A N D MAGWLANLC PENGUINS 57

Biraben, M. and M. I. H. Scon 1939. Observaciones s o h el 'pinguino' Sphcniscur magclianicus Fonter. Physis 16: 245-251.

Bhke, E. R. 1977. Manual of Nwmpical b i d . Vol. 1. Spheniscidae (Penguins) to Laridae (Gulls and Allies). Chicago and London. The University of Chicago Press.

Boersma, D. 1977. An ecological and behavioral study of the Galapagos penguin. Living Bird 15: 43-93.

Boenma. D. 1978. Breeding patterns of Galapagos Penguins as an indicator of cceanographic conditions. Science 200: 1481-1483.

Bcenma. D., D. L. Stokes and P. M. Yorio. 1990. Reproductive variability and historical change of Magellanic Penguins Spheniscus magcllonicrcr at Punfa Tombo, Argentina. Pp. 15-43 in L. S. Davis and J. T. Darby. Penguin biology. San Diego, Academic Press.

Boswell, I. and D. Maclvor. 1975. The Magellanic Penguin, Spheniscus mngellanicus. Pp. 271-305 in B. Stonehouse. The biology of penguins. London, Macmillan Press.

Boswell. J. and R. 1. Prytherch. 1972. Some notes on the birds of Punta Tombo, Argentina Bull. Brit. Ornith. Club 92: 118-129. Brown, C. R. 1987. Traveling speed and foraging range of Macamni and Roc!&opper penguins st Mruion Island. J. Field Om. 58: 118-125.

Burger, A. E. and R. P. Wilson. 1988. Capillary deprh gauges an assessment of their accuracy and applicability. I. Field Om. 59: 345-354.

Carrarra, I. S. 1952. L o b s marinos, pinguinos y guaneros del litoral maritimo y islas adyacentes de la Pepublica Argentina. Ministetio de Educacion, Facultad de Ciencias, Vercerinarias, Universidad Nacional de la PI&. La Piata, Argentina

Cobb, A. F. 1933. Birds of the Falkland Islands. London, Gowan's Name Books.

Coker, R. E. 1919. Habits and economic relations of the guano buds of Peru. F'roc. US. National Museum 56: 449-5 1 1.

Darwin, C. 1933. Charles Damin's diary of the voyage of the Beagle. N. Barlow, Cambridge. Cambridge University F'ress.

Devillers, P. 1978. Dismbution and relationships of South American skuas. Le Gerfaut 68: 374-417.

Devillers, P. and I . A. Terschuren. 1978. Relationships between the blue-eyed shags of South America. Le Gerfaut 68: 53-86.

Duffy, D. C. and W. R. Siegfried. 1987. Temporal varialions in food consumption by seabirds of two upwellings. Pp. 327-346 in J. P. Croxall. Seabird feeding ecology and mle in marine ecosystems. Cambridge, Cambridge University Press.

Endler, I. A. 1977. Geographic variation, speciation. and clines. Princeton, Princeton University Press.

Harper, W. M. 1977. Statistics. Plymoulh, M a c d o ~ l d and Evans.

Hays, C. 1984a. The Humboldt Penguin in Peru. Oryx 18: 92-95. Hays, C. 1984b. ?he Humboldt Penguin Spheniscus humboldii in Peru and the effects of the 1982-1983 El Nino. M.S. thesis, University of Florida, Gainesville.

Johnson, R. W. 1x5. The bids of Chile and a d j a n t regions of A r g e n i ~ , Bolivia and Peru. Buenos Aires, Pian Esfablicimientos Graficos, S.A.

La Cock, G. D. 1988. Effect of substrate and ambient temperature on burrowing African Penguins. Wilson Bull. 100: 13 1-13".

La Cock, G. D., D. C. Duffy and I. Cooper. 1987. Population dynanucs of the African Penguin Spheniscus demcrsw a1 Marcus Island in h e Benguela upwelling ecosystem 1979-85. Biol. Conserv. 40: 117-126.

Lishman. G. S. and J . Croxall. 1983. Diving depths of rhe Chinstrap Penguin Pyguscelis unrarcrico. Brit. Anwct. Sum. Bull. 61: 21.25.

Livezey. B. C. 1986. Phylogeny and historical biogeography of steamer.ducks Anatidae Tachyeres. Systenlatic Zool. 35: 458-369.

Livezey, B. C. and P. S . Humphrey. 1986. Flightiessness in steamer-ducks Ana~idae Tachyeres: its morphological bases and probable evolution. Evolution 40: 540-558.

Macanhur, R. H. 1372. Geographic ecology palterns in the disuibution of species. Few York. Harper and Row.

May, E. 1%3. Animal species and eualurion. Cambridge, Harvard Liniversiry Press.

Mercer, J. H. 1976. Glacial history of sourhermost South America. Quarernaty Res. 6: 175- 166.

5 8 WILSON. R.P. D m . D.C.. WaTON. M.4' and ARAYA. 0. GERFAUT 85

. .

~ u r p h y . R. C. 1936. Oceanic birds of South America. New York. Macmillan.

Pettingill, 0. S. 1964 Penguins ashore at ihe Falkland Islands Living Bird 3: 45-64.

Pickard, C. I.. 1971. Some physical oceanographic features of inlets of Chile. I. Fish. Res. Board Canada 28: 1077-1 106.

Rasmussen, P. C. 1991. Relationship between coastal South American King and Blue-eyed shags. Condor 93: 825-839.

Renard, A. 1931. Algunas observaciones sobre aves durante un viaje A la Ria & Deseado. tiornero 4: 412.414.

Ricklefs, R. E., D. C. Duffy and M. Coulter. 1984. Weight gain of blue-footed booby chicks an indication of d n e resources. Ornis Scand. 15: 163166.

Ross. C. J . 1971. The Jackass Penguin on St. Croix a census. Easbem Cape Naturalist 44: 22.24.

Ryan, P. C., R. P. Wilson and I. Cooper. 1987. lntraspecific mimicry and status signals in juvenile African Penguins. Behav. Ecol. Sociobiol. 20: 69-76.

Scolam, I. A. 1978. El Pingiiino de Magelianes Spheniscus mngellunicus IV. Nolas biol6gicas y de componamienm. Pub. Ocas. Scrie Cient. InsL Biol. Anim. 10: 1-6.

Scdar0. J . A. 1984a. Biulogla y selection del habitat do reproduccion de Spheniscus mugellanicus en Patagonia, Argentins. Rcsumen de Tesis Doctoral, Univenidad de Cordoba, Spain. Pp. 1-29.

Scolaro. I. A. 1984b. Revision sobre biologia de la reproduccion del pinguino de Magallanes Spheniscus mgeilanicus. El ciclo biologico anual. Cenuo Nacional Patagonico Convibucion 91: 1-26.

Scolato, J. A, and A. M. Subaro. 1991. Maximum diving depths ofthe MageUanic Penguin. J. Field Om. 62: 204-210.

Siegfried, W. R.. P. C. H. Frost, I. B. Kinahan and J. Cooper. 1975. Social behaviour of lackass Penguins a! sea. Zoologica Africana 10: 87-100.

Simpson. G. G. 1976. Penguins past and prcsen(, here and there. New Haven. Yale University Press.

Slnilll, N. C. 1966. Evolution of some Arctic gulls Larus: an experimental study of isolating mechanisms. Omiihol. Monogr. 4: 1-99.

Stokes, D. L. and P. D. Boersma. 1991. Effects of subsnate on the distribution of Magellanic Penguin Weniscus mageliunicus burrows. AUL 108: 923.933.

Swsemann, E, and N. W . Timofeeff-Ressovsky. 1947. Artentslehung in Geographischen Formenkreisen. I. Der Forrnenkreis Larus argentarus-cachinnam.fuscw.. Biol. Zenv. 66: 57-76.

Udvardy, M. D. 1963. Zoogeographical study of h e Pacific Alcidae. Pp. 85-1 1 1 in I. L. Gressit. Pacific Basin Biogeography A Symposium. Honolulu, Bishop Museum Press.

Vallentin, R. 1924. Zoology. In V. F. Boysen. The Falldand Islands. Oxford, Clarendon Press.

Wiens, J. A. 1977. On competition and variable environments. Am. Na. 65: 590-597.

Williams. T. D. 1995. The ptnguins. Oxford, Oxford University Press.

Wilson, R. P. 1984. An improved stamach pump for penguins and other seabirds. 1. Field Om. 55: 109-112.

Wilson, R. P. 1985a. The Jackass Penguin Sphcniscus demersus as a pelagic predator. Mar. Ecol. PIOF. Ser. 25: 219-227.

Wilson, R. P. 1965b. Seasonality in die1 and breeding success of the Iackass Penguin Sphwiscus demersus. 1. f. Om. 126: 53-62.

Wilson, R. P. and C . A. R. Bain 1984. An inexpensive speed meter for penguins a1 sea. i. Wildl. Manage. 46: 1360-1364.

Wilsan, R. P. and M.-P. Wilson 1990. Foraging ecology of breeding Spheniscus penguins. Pp. 181-206 in L. S. Davis and J . T. Darby. Penguin biology. San Dirgo, Academic Press.

Wilson. R. P., W. S. Grant and D. C . Duffy. 1986a. Recording devices on free-ranging marine animals does measurement affect foraging performance? Ecology 67: 1091-1093.

Wilson, R. P., M.-P. Wilson and L. McQuaid. 1986b. Croup s i u in foraging African Penguins Spheniscw dcmersus. Eihology 72: 338-341.

Wilson, R. P., P. G . Ryan, A. James and M:P. Wilson. 1987. Conspicuous colaration may enhance prey capture in some pirivores. Anim. Behav. 35: 1558-1559.

Wilson, R. P.. M:P. Wilson, D. C . Dufi and B. Araya 1989a. Diving behaviow and prey of the Humboldt Penguin Spheniscus humboldti. I. f. Om. 130: 75-79.

Wilson, R. P., B. Culik N. R. Coria, D. Adelung and H. Spairani. 1989b. Foraging rhythms in Ad6lie Penguins at Hope Bay, Antarctica; determination and conlrol. Polar Biol. 10: 161- 165.

Wilson, R. P. and 8. M. Culik. 1992. Packages on penguins and device-induced data, Pp. 573-580

in I . C. Priede and S. M. Swift. Wildlife Telemetry. Chichester, Ellis Honvood.

Wilsok R. P., K. Puetz, C. A. Bost, B. M. Culik. R. Bannasch. T. Reins and D. Adelung. 1993. Diel dive depth in penguins in relation to die1 venical migraion o f prey whose dinner by candlelight? Mar. Ecol, progr. Ser. 94: 101-104.

Wilson, R. P. and M:P. Wilson. 1995. The foraging behaviow of the African Penguin. Pp. 244-265 in P. Dam, I. Norman and P. Reilly. The Penguins. Surrey, Aushalia. Beany and Sons.

Zapata, A. 1967. Observaciones sobre aves de Puerto Deseado, Provincia de S a m Cmz. Hornero 10: 351-378.

SAMENVATTING

De Humboldt-,Spkenlrciu hunhld t i , en de Magellaan pinpins. Sphenisciu magellaniciu, komen iangr dc wesutust van Zuid-Amerika in sen grwt gebied

sympanisch vow Onderzoek werd verrichr belreffende hun voedscl ecologie op vier verschillende bmedplaalseh waarvan dlie pelepen zijn in de sympalrischt zone sang$ de Chilecnse kusr russen 29O en 53*S. Ei werd vastgcrteld dal vcrxhiilen Nrsen de broedplmren meer uitgesprokcn waren dan de inter-specifieke. De broedreizoenen der wee soonen z ip verrchillead

RESUME

Le Manchor de Humboldt Sphrniscus humhoidn el le Manchor de Magellan Sphcnircvr mngdidnuur vivenr en rympame dam une large zone de la c8re ouesl de I'AmCrique du Sud. De la rrcherche fa1 faire sur leu< kologie de nourrissage dans quatre endroirs de reproducdon difftrents, dont lrois son1 sirues dam la zone 00 ils vivenl en sympaaic enrre 29" er 53's. Les chercheurs on1 mnslate que les differences enne ies lieux & rcproduclion furent pius prononcees que les inler- rp6cifigues. Les periodes de reprodurlion eruenl difftrcnrcs entrc les dcux cspkes.

Available information on the b d i n g Eearon of the Humboldt Penwin

MONTH EVENT LOCALIlY SOCRCE

January Large young lsla Cachaguq Chile F. Todd unpublished

February Lame vouna Isla Santa Rosa. Peru Mumhy 1936

/I ~ni i rc "ear I Brccdinn I € n t h ranee I Murnh, 1936 11 /I March . December Breeding Peru Ha)$ 1984b II /[ March Small chicks lsla Guanape. Pen! Coker 1919

I[ May & October Peak breeding Entire range Johnson 1965

I\ May & October Peak brceding Isla Cachapua. Chile 8. Araya unpublished I Eggs. small chicks lslas Ballestas, Peru Coker 1919

Eggs, small chicks lrla Crande. Chile I 8. Araya unpublished I November EBBS, large chicks Isla Santa Rosa. Pen! Mvrpllj 1936 - December Small-Me chicks isia Cachaguq Chile F. Todd unpublirhed

December EKES. chicks Pan de Azucar. Chile 8. Culik unpubl~shcd -

60 WUSON. R.P.. D W Y , D.C.. WILSON. M.-P.md ARAYA. 8. G W A W 85

Available informadon cn ~e breeding searon of b e Magellanic Penguin

MONTH EVENT

ianuar)

LOCALITY SOURCE 1 Bosweil & Maclver 1975 Young dcpan Punm Tombo

1995 GIERVALK SPECLES RPPLAC@MENT IN HUMBOWT AND MAGELLANLC PENGUINS 61

APPENDIX 2 CONTINUED

WILSON, Rory P.. Jnrfimlflr Mcererkundc. Dutlernbrooker Weg 20, 0-24105 Kiel. C o w .

MONTH

Ocrohr

October - April

Octohr - Mach

Novtmber

DUFFY, David C., Per? FiePatrick Inrlilulc of African Omirhology. Univrmily of Cape Town. Rondebosch 7700. Sowh Africa. Current addras: A h k n N m l a d Heriroge Program, E n v v o m n r and Narural Resources Insriav, Uniuerrip of A h k a Ancharogc, 707 A SIreeI, AIVhwage, AK 99501. Alarka,

W[LSON, Marie.Piem, Percy FitrPmrick IwriNle of Africon Ornirhobgy. Zlnhrrsiry of C q c Town, Rondcbosch 7700. Sourk Africa

EVENT

E g s

Nesdng

Nesdng

Peal; hatching

ARAYA, Braulio A,. ImriNo dc Owanologia, Universidad dc Vdprnaisa, Cmi lh 13-0 Vitia d d Mar. Chilr

Acccpttd 15 October 1995

U3CALITY

Falklands

Titna del Fuego

Falklands

Punta Tombo

SOURCE

Cobb 1933

Murphy 1936

Peninpill 1%4

Bosweil & Vthcrch 1972