THE

FORAGING

ECOLOGY
DAVID

OF PERUVIAN
DUFFY

SEABIRDS

CAMERON

PercyFitzPatrick Instituteof AfricanOrnithology, University CapeTown, of Rondebosch South 7700, Africa

ABSTRACT.--Most feeding by seabirdsin the Peruvian CoastalCurrent, an upwelling of high productivityoff the west coast SouthAmerica,takesplacein groups.The majorprey of is an anchovy(Engraulis ringens), which occursin large shoalsand is exploitedmainly by three species: the Peruvian Booby (Sulavariegata), Peruvian Brown Pelican (Pelecanus occidentalis thagus), GuanayCormorant(Phalacrocorax and bougainvillii). Other foragingsituations have different species' compositions, and these appear to be related to the size, depth, and duration of availability of prey. Dominanceinteractionsbetween species may be important in structuringflocksthat are scavenging feedingon planktonswarms. or Interspecific piracy seemsunimportant in flocks foraging on fish shoals.Certain speciesusually arrive first at new feeding situations.Thesespecies may be usedas guidesby other species merely may or be faster and thus reach food sourcesfirst. Studies of foraging of seabirdsshould be a valuable addition to the study of the distribution of birds at sea. Received April 1982, 2 resubmitted March 1983,accepted July 1983. 18 8

SEABIRDS have some of the largest foraging ranges of any vertebrates:birds with eggs or young may travel 1,000 km from the nest to feed (Fisher and Lockley 1954, Harris 1977,
Nelson 1979, Dunnet and Ollason 1982). De-

tation of patchesof food (e.g. Hoffman et al.

1981)?
METHODS

spite the fact that they searchover such large areas,they forage over very small areason local concentrations prey (see Brown 1980 for a of review). The study of thesepatchesand their use by birds should improve our understanding of the ecologyand distribution of seabirds at sea.For example,Erwin (1977) linked differencesin use of patchesby three larids to differencesin their nesting ecology. Hoffman et al. (1981) showed that the presenceof some species may affectthe foragingof others;flocks are more than aggregationsof noninteracting
birds attracted The structure to a common food source. of seabird flocks has never been

The studyarea.--I gathered most of my observations on Isla Mazorca (1123'S, 7745'W), Departamento de Lima, Peru. The island lies approximately

15 km offshore an areaof frequentupwelling(Zuta in and Urquizo 1972)and is the siteof a largeand longestablishedcolony of guano birds, especiallyPeruvian Booby(Sulavariegata), Peruvian Brown Pelican (Pelecanus occidentalis thagus), and GuanayCormorant (Phalacrocorax bougainvillii) (Murphy 1925;Duffy 1981, 1983a). Other observations were madeon 12 boattrips
between Isla Mazorca and the port of Huacho (1107'S,7744'W) and from the shore at La Puntilia,
Callao Harbor (1205'S, 7744'W).

The Peruvian Current is one of the most productive of marine ecosystems (Cushing 1971).It extends
from northern Chile to northern Peru and out to the

investigated in low-altitude upwelling areas. The presentpaper reportson the numbersand behaviorsof seabirdspeciespresent in different types of feeding flocksin the Humboldt or
Peruvian Coastal Current off the west coast of

South America. Three general questionswere investigated: what percentage birds for(1) of age in flocksover patchesof prey, (2) doesthe species'compositionof flocks differ in accordance with the type of feeding situation, and (3) are flocksmerely aggregations indepenof dently attractedbirds or do the species interact to enhanceor inhibit the discoveryor exploi800

Galapagos Islands, Ecuador(Murphy 1936). The upwelling is confinedto a relativelynarrowbandalong the coast(Cushing 1971). The dominant fish species of the upwelling is the Peruvian anchovyor anchoveta (Engraulis ringens), which makesup about90%of pelagicfish stocksin the current, basedon relative abtmdances eggsand larvae (Santander1981).The of anchoveta is eaten by a wide variety of seabirds, predatory fish, and marine mammals (Coker 1920, Murphy 1936,Jordan and Fuentes1966,Paulik 1971). A commercial fishmealindustry,which beganin 1955, took catchesof up to 12 million metric tons of anchoveta, becoming the world's largest single-species fishery before the fish stockand industry collapsed

The Auk 100: 800-810. October 1983

October1983]

Foraging Peruvian of Seabirds

801

TABLE Percentage of different 1. use foraging situations Peruvian by seabirds = <0.1). (+

Percentage foraging situation in
Zoo-

Number

ScavI1 48 94
.... ....

Sea- plank"Other"

Species
HumboldtPenguin(Spheniscus humboldti)
Waved Albatross(Diomedea irrorata) Wilson's Storm-Petrel(Oceanites oceanicus)

observedShoal enging lions ton

18
52 54
6 -

88 52 100 100

Elliot'sStorm-Petrel (Oceanites gracilis)

Markham's Storm-Petrel (Oceanodroma markhami)

1
1

Unidentifiedstorm-petrels CapePetrel(Daption capense) Sooty Shearwater (Puffinus griseus) PeruvianDiving-Petrel(Pelecanoides garnoti) Magnificent Frigatebird (Fregata magnificens) NeotropicalCormorant(Phalacrocorax olivaceus) Guanay Cormorant bougainvillii) (Ph. Red-legged Cormorant gaimardi) (Ph. Blue-footed Booby (Sula nebouxii) Peruvian Booby variegata) (S.
Peruvian Pelican (Pelecanus occidentalis thagus) Red-neckedPhalarope(Phalaropus 1obatus) PomarineJaeger(Stercorarius pomarinus)

21 2 61,234 61 1 11 115,542 114 2 402,802

3,341 1 11

5
....

+ -

+ -

95
I00

98 7 ....
.... 99 51

2 93 100
100

50
95 69 .... 27

+ +

1 49 50 5 30
100

72

Jaeger (Stercorarius sp. sp.) GreyGull (Larus modestus)

Band-tailedGull (L. belcheri)

7 1,429
554

71
41

43

28
12

4
3 36 I 60 .... .... 26 94.4 65 20

49
71 11 -

3
12 I + -

16
14 21 44 -

28
51 79 55 40 100 IO0 34 4.7 137 27

Kelp Gull (L. dominicanus) Franklin'sGull (L. pipixcan)

Sabine'sGull (L. sabini) "Comic"Tern (Sterna sp.)

94 743
86 1,328

Elegant Tern (Sterna elegans)

Sandwich Tern (S. sandvicensis) Peruvian Tern (S. lorata) Inca Tern (Larosterna inca)

5
2 2 7,195

I 0.1 67 6

8 0.2 49 I0

30 0.6 37 7

Percentage total numberof birds(n = 561,605) of

Number of observations each foraging type for Number of species each foraging type in

I "terrestrial-seizing" birds for in 1972-1973 (Idyll 1973;Duffy 1983b). pilchard surface-diving.added A on zone.Fordipping (Sardinops wasthought havereplaced sagax) to the feeding landorin theintertidal I duration contact of with water, anchoveta duringthe 1970's (Walshet al. ! 980),but species, measured usinga spring-operated stopwatch and subtracting the data are equivocal. time.Fordivingand plunging Observations.--I madeasmanyobservations ma- 0.2 s for my reaction of I measured duration submergence the of from rine birds as possibleduring the courseof other species,
fieldworkbetweenSeptember 1977and March 1978. I recorded foraging all birdsencountered. Species are
listed in Table I.

first contact until the bird's head emerged. I mea-

suredratesof dipping or surface-seizing followby

Most seabirds in the Peruvian Coastal Current are

ing individualbirdsuntil they completed atI0 tempts ceased or feeding. recorded I interspecific
interactions that included association with marine

relativelyeasyto identify. The exceptions the are

white "comic" terns (Sternaspp.) and the jaegers

(Stercorarius Unidentified spp.). birdsaretreated sep-

mammals, aggressive displacement from potential sources food,kleptoparasitism, orderof arrivof and
al at food sources. Not all of these data could be col-

arately make onlya minuscule but up proportion of

birdsobserved. determinednumbersin smallgroups I

(less than20-50)by directcounts individuals. of In larger groups,I estimated subgroup a and then
countedthe number of such subgroups. defined I

lected simultaneously, in the results, sample so, the sizes maydifferbetween differentmeasurements.
I assumed that the foraging behaviorsand situations I encounteredwere proportional to their true

I to as feeding techniques following simplified by a version abundances.attempted observe wide a variety situations possible watching as by from of Ashmole's(1971) terminology: piracy, dipping, of foraging observations at surface-seizing, plunging, pursuit-plunging, and both shipsand land;by attempting

802

DAVID CAMERON DUFFY

[Auk, Vol. 100

TABLE Species'occurrences 2. over dense shoalsof fish (+ = < 0.1).

All birds
Number Number/ at shoals

Frequency
at shoals

Species Peruvian Booby Guanay Cormorant SootyShearwater

Peruvian Pelican Inca Tern

present 381,957 114,380 60,036

2,320 1,888

occurrence 6,160 2,288 7,504

122 118

(%) 68 20 !1
0.4 0.3

(%) 95 77 12
29 25

Grey Gull
Franklin's Gull

584
270

117
39

0.!
+

8
!1

Red-legged Cormorant
Wilson's Storm-Petrel Waved Albatross Band-tailed Gull "Comic" tern

58
51 25 23 8

12
25 4 7 3

+
+ + + +

8
3 9 5 5

Jaegersp. Peruvian Diving-Petrel Pomarine Jaeger Elegant Tern Kelp Gull Humboldt Penguin Blue-footedBooby Storm-petrel sp.

5 4 3 3 3 2 i !

1.3 0.6 1.0 3 3 1 1 1

+ + + + + + + +

5 5 5 1.5 1.5 3 1.5 1.5

night, using a full moon or bioluminescence; by and combining observations from before (September-October) and during (November-March) the breeding season.The 7-month period should also have been long enough to sample any seasonalvariations in foraging.
RESULTS

seen were feeding alone. After fieldwork, I

found that the records could be divided into

five arbitrary but recognizableforaging situations:shoal-foraging, scavenging, foragingover sealions,feeding on zooplankton swarms,and
miscellaneous other situations. These are re-

ferred to as shoal, scavenging, sealion, zooplankton, and "other" foraging groups throughout this paper. This study was conductedafter the collapse Foraging overshoals fish.--Shoalsof anchoof of the anchoveta stock (Idyll 1973). This col- veta and other fish were frequently fed upon lapse has affected the populations of guano by seabirds the watersaround Mazorca.Bird in birds (Duffy 1983b),but food did not appear to flockswere characteristically very dense. I asbe in short supply around Isla Mazorcaduring sumed other such dense bird flocks were also my study. Foraging trips by Guanay Cormo- feeding on fish shoals,even when I could not
rants and Peruvian Boobies took less than 3 h

(Duffy 1983a).Vogt (1942) considered trips over 6 h to indicate food shortages. frequently obI served shoalsof fish, and birds feeding upon
them, around the islands. After October, an-

seethe fish.Of all individuals,94%foragedover shoalsof fish. Twenty species participated(Table 1), but 98% of the individuals over shoals

chovetas were frequent in regurgitations or stomach casts the cormorant,booby,and pelof ican. I assumethat the foraging behaviors I observed were similar to those used before the

were of three species: Peruvian Booby,Guanay Cormorant, and SootyShearwater (Table2). Inca
Terns and Brown Pelicans were less numerous

than shearwatersbut more consistentlypresent. Other species comprised0.1%or lessof the

flocks. The mean flock size over shoals was

collapseof the anchovetastock.

FORAGING SITUATIONS

8,363.5 birds (SD = 12,629.3; n = 65).

Many large groupsforaging over shoalspersisted for 2-3 h or more. In other cases, instead of at least 28

A total of 561,605 individuals

of large groups, the birds formed a mosaicof smaller aggregations, which persistedfor only
about 15 min. Peruvian Boobies initiated 9 of

speciesoccurred in 355 observationsof birds foraging (Table 1). Only 48 of all individuals

10 such ephemeral groups.

October1983]

Foraging Peruvian of Seabirds

803

TABLE Species' 3. occurrences over sealions.

All Num-Numwith Frewith

TABLE Species' 4. occurrence zooplanktonswarms. at

All
at

Freat

birds quency
ber bet/ seaseaNumNum-

birds quency
zoozoo-

pres- occur- lions lions

bet
Species
Inca Tern

bet/ plank- plankoccurfence

66.4

Species
Inca Tern

ent
567

fence
13.8

(%)
80.6

(%)
84

preso ent
2,191

ton (%)
62

ton (%)
89

Grey Gull

43

4.3

6.1

20

SootyShearwater
Band-tailed Gull

42
16

6.0
1.4

6.0
2.2

17
22

Grey Gull
"Comic" tern

621
580

28.2
30.5

17.5
16.4

59
51 22

Kelp Gull Peruvian Booby

Franklin's Gull "Comic" tern Wilson's Storm-Petrel Peruvian Pelican

11 11
5 4 3 I

1.1 2.2
1.0 2.0 3.0 1.0

1.6 1.6
0.7 0.6 0.4 0.1

20 10
10 4 2 2

Band-tailed
Sabine's

Gull

87

10.9

2.4

SootyShearwater
Gull

27
18

5.4
9.0

0.8
0.5

13.5
5

Kelp Gull

13

2.6

0.4

13.5

gull species on the water while feeding. The sat

speciesoccupied characteristicpositionsaround

Foraging oversealions.--Thesouthern sealion

lOtariabyronia (fiavescens)] providedfood for the

birds, either by chasing fish to the surface or by thrashing large fish and cephalopods to pieces.Birds usually fed only in a very small area immediately surrounding the sealion. Ten species,but only 0.1% of all individuals, foraged over sealions (Table 1). The major bene-

the cone of hovering birds. Grey Gulls (with Kelp and Band-tailed gulls if present) sat on
the surface below the cone, Inca Terns were at

ficiary of sealion activity was the Inca Tern:

80% of the birds over sealions belonged to this species (Table 3). The other nine species oc-

curred irregularly and in small numbers. Average flock size over sealions was 14.3 birds
(SD = 45.7; n = 49).

Foragingon zooplankton swarms.--Dipping or surface-seizingof zooplankton involved seven species and 0.6% of all individuals (Table 1). Only three specieswere common: Inca Terns, "comic" (mainly Arctic) terns, and Grey Gulls made up 96% of all neuston-feeders (Table 4). Sooty Shearwatersand Kelp, Band-tailed, and Sabine'sgulls were rare and irregular. The prey I observedwere 1-cm, reddish crustacea swarming in densities of up to 1,000/m 2 at or just below the surface.Some patchespersisted for hours, but most were short-lived,

lasting no more than 5 min. Most of the neuston-feeding took place in the austral spring,
before anchoveta became terns common hovered in stomach in "cones"

the center of the hovering birds, and "comic" terns were at the periphery. Grey Gulls had the highest feeding rates,followed by Inca and then "comic" terns (Table 5). The higher rate of the Grey Gull was not due entirely to the differencein foragingmethods (surface-seizingvs. hovering + dipping): Band-tailedand Sabine'sgulls had lower rates than the terns, even though Band-tailedGulls foraged in the samemanner as Grey Gulls and Sabine'sGulls used a foraging method intermediate between dipping and surface-seizing. During one observationperiod when several species were foragingover short-livedpatches of zooplankton, Inca Terns arrived before Grey Gulls at surfacepatches7 of 7 times (binomial, P = 0.008). The mean group size during all zooplankton feeding was 95.9 (SD = 83.2; n = 37). Scavenging.--At sea,birds scavengedfish offal, sealioncorpses, and nestlingsthat had fallen from the guanoislands.On land, they scavenged bird corpses, pellets regurgitated by cormorants,and regurgitations from nestlings. Seven species,but only 0.1% of all individuals in this study, scavenged(Table 1). Band-tailed
Gulls were the most numerous, both in num-

pellets of cormorants.
"Comic" and Inca

of birds over zooplankton patches:a few birds

would be close to the surface, and a much

greater number would be above and outside. Hovering was punctuated by irregular bouts of dipping. Sabine's Gulls landed briefly on the water to peck at prey, whereas the three other

bers and in frequency of occurrence(Table 6). Kelp Gulls were frequent, but rarely more than two birds were present. Inca Terns occurred infrequently but in large numbers. Kelp Gulls specialized on corpses(43 of 67; 67%), whereas only 57 of 272 (21%) Band-tailed Gulls were necrophagous. The difference is significant (2X2 contingency table, correctedfor

804

DAVID CAMERON DUFFY

[Auk,Vol. 100

TABLE Feedingrates(dips or pecksper second). 5.

Number of

TABLE Species' 6, occurrences scavenging in situations.

All

Fre-

Species

0.355 0.166 0.153 0.120

SD
0.181 0.062 0.146 --

birds

Num- Num- birds quency

ber ber/ scavscav-

Grey Gull
Inca Tern Arctic Tern Band-tailed Gull Sabine's Gull

0.584 0.466

477
452 147 71 7

50
53 18 1!

Species
Band-tailed Inca Tern Gull

pres- occur- enging enging ent rence (%) (%)

272 97 5.7 13.9 17.2 51 18 70 10

Franklin's

Gull

86

16

7.5

Kelp Gull Grey Gull

Peruvian Pelican

67 7
3

2.4 1.4
1.5

12.6 1.3
0.6

42 7.4
3.0

continuity; X2 = 19.1; P < 0.001). The most important sourcesof food for Band-tailed Gulls while scavengingwere regurgitated pellets of
cormorants (101 of 272; 37%). Inca Terns and

Franklin's and Grey gulls fed on small particulate offal in the water.

plunge, 1; pursuit-plunge, 1; surface-dive,3 (or 4 if a single piracy record for Guanay Cormorant is excluded)]. Two speciesused 2 techniques, 4 speciesused 3, and 4 speciesused 4.

This gradient in food size paralleled the apparent dominance hierarchy. Kelp Gulls invariably displacedBand-tailedand Grey gulls. Band-tailedGulls displacedFranklin's Gulls and Inca Terns, and Franklin's Gulls displacedInca
Terns. Reversals were not observed.

When foragingmethods rankedaccording are to their mostcommonusage(Table 8), surfaceseizing proves to be the dominant foraging methodof 6 species, surface-diving 4 species, of
dipping of 3 species,piracy of 2 species,pursuit-plungingof 1 species, and plunging of 1
species. Species that primarily used surface-

Mean group size was 8.1 birds (SD = 19.8;

n = 67).

"Other" foraging.--This group includes those recordsthat did not fall into the precedingcategories. Twenty-seven speciesand 4.7% of all individuals were involved in foraging situations lumped into this category (Table 1). Among the different situationswere the few inshore recordsfrom sandy bottoms,very dispersed feeding on what were probably small anchoveta shoals(most of the Peruvian Booby, Sooty Shearwater, and Guanay Cormorant records),solitary foraging, and looseaggregations of apparently solitary species(Humboldt Penguin and Red-legged Cormorant). The most common species were the Peruvian Booby
(73.6%) and Inca Tern (Table 7). The 25 other speciescomprisedlessthan 18%of the individuals in this category.No species occurredin all

diving had the narrowest range of foraging methods, only 1.25 per species.Species that primarily used terrestrial seizing (four gulls) had the widest range, averaging four techniques each. Surface-seizingspeciesaveraged 3.6 methods; dipping species, 3.5; pirating species, 2.8; plunging species, 2.4; and pursuitplunging species,2.0. The speciesvaried greatly in the amount of time spent on individual feeding attempts, based on duration of contact with water (Table 5, Fig. 1). Dipping by Inca Terns took only 0.25
s (SD = 0.325; n = 68) while Humboldt Pen-

"Other" foraging situations. Even the most common by number, the Peruvian Booby, occurred only 36% of the time. Fourteen of the speciesoccurred in less than 5% of the observations. Group size was not calculated because of the artificial nature of this category.
]ORAGING BEHAVIOR

guins were submergeda mean time of 75.0 s (SD = 44.9; n = 14). Only three species had submergence durations of over 20 s (Fig. 1): solitary Guanay Cormorants,Red-legged Cormorants, and Humboldt Penguins. Dives by GuanayCormorantsforaging at anchoveta shoals were much shorter (9.55 s; SD =

4.45; n = 144)than dives by Guanayselsewhere

(32.85 s; SD = 22.15; n = 16; P < 0.001, t = 11.04;

F = 24.77; correction unequal for variances). On

the other hand, Peruvian Boobies plunging at
shoalsshowed no significant differencesin durations of submergence(shoal= 3.12 s; SD =
1.29; n = 234; all other situations = 2.99 s; SD = 1.17; n = 254; P > 0.05).

Seven foraging techniqueswere used by the species observedin this study (Table 8). Seven species usedonly a single technique[piracy, 2;

The final component of foraging behavior measured was intraspecific group size in dif-

October 1983]

Foraging PeruvianSeabirds of

805

T^BLE Species' 7. occurrences "Other" foraging situations(+ = < 0.1). at

All birds
Number Number/ at "Other"

Frequency
at "Other"

Species Peruvian Booby

Inca Tern

present 20,834
2,452

occurrence 425
84

(%) 73.6
8.7

(%) 36
21

Guanay Cormorant Sooty Shearwater

Peruvian Pelican
"Comic" tern

1,162 1,129
1,017
736

193 188
39
52

4.1 4.0
3.6
2.6

4 4
19
10

Franklin's

Gull

382

32

1.3

Grey Gull
Band-tailed Gull Sabine's Gull

173
156 68

16
10 11

0.6
0.5 0.2

8
12 4

Peruvian Diving-Petrel Humboldt Penguin

Waved Albatross

57 56
27

3.8 2.5
2.7

0.2 0.2
0.1

11 16
7

Storm-petrel sp. Red-legged Cormorant NeotropicalCormorant Pomarine Jaeger Jaegersp. Cape Petrel
Peruvian Sandwich Tern

20 16 11 8 2 2
2 2

6.7 2.0 3.7 2.0 1.0 1.0

2.0 2.0

0.1 + + + + +
+

2 6 2 3 1 1
0.5 0.5

Tern

Elegant Tern Blue-footed Booby Northern Phalarope

Elliot's Storm-Petrel

2 1 1
1

1.0 1.0 1.0

1.0

+ + +
+

1 0.5 0.5
0.5

ferent foraging situations.Varianceswere very large, group sizesranging over four ordersof magnitude. Of 18 specieswith sufficientrecords (Fig. 2), 10 specieshad mean group sizes
of less than 10/occurrence, and 5 had means of

gest that, even in the highly productive Peruvian Coastal Current, most food for seabirds

10-100/occurrence. No specieshad mean group

sizes of 100-1,000/occurrence, but the three

most common species(Guanay Cormorant, Peruvian Booby,and SootyShearwater)had group

sizes of 1,000-10,000/occurrence.

occursin patches.Similar patterns of aggregation have been observedin other oceanographic zones listed by Ashmole (1971): the tropical Pacific(Gould 1974, King 1974), boreal Pacific (Porter and Sealey 1981, Schneider 1982), and BenguelaCurrent (Duffy pers. obs.).
Because anchoveta were the main food taken

by birds during the study period and because Group size varied between foraging situa- anchovetatypically occurin shoals(Vogt 1942), tions (Tables2-4, 6-7), the largestintraspecific most shoalsexploited by the birds were probgroupstending to occurin shoal-foraging, fol- ably comprisedof anchoveta.Anchoveta have lowed by "Other," zooplankton, scavenging, very patchy distributions. In one survey, 55%
and sealions.

of the total biomass was confined to 13% of the

DISCUSSION

survey area, while, in another, 36%of the biomasswas concentrated only 3.4%of the area in
(Johannessonand Vilchez 1980). Similar con-

FORAGING IN GROUPS

centrationshave been reported for the northern anchovy (Engraulis mordax) California off
(Mais 1974).Densitieswithin patchesmay range up to 2120 metric tons/km2 in Peru (Johannesson and Vilchez 1980). Anchoveta are therefore both patchy enough and sufficientlyabundant

Over 99%of all birds foragedin groups.Solitary foraging was rare: only Humboldt Penguins (13.5% of 39 observations),Red-legged Cormorant (12% of 73), and Peruvian Diving

Petrels(100%of 3; seealsoMurphy 1936)foragedalone with any frequency.Thesedatasug-

to explainthe largeaggregations typicalof birds in the Peruvian coastalupwelling area.

806

DAVID CAMERON DUFFY

[Auk, 100 Vol.

TABLE Percentage of foraging 8. use techniques Peruvian by seabirds = dominant (* method eachspecies). for
Put-

Terres-

Sur-

suit-

Sur-

Sam-

Species
Humboldt Penguin
Waved Albatross

trialface- Plung- plungPiracy Dipping seizing seizing ing ing

.......
100' .....

face dive
100'

pie size
37
17

SootyShearwater Diving-Petrel
Peruvian Pelican

4 .....
18

---

---

59*
63*

-19

36 tOO*
--

----

22 3
498

Peruvian Booby Guanay Cormorant Neotropical Cormorant Red-leggedCormorant

Pomarine and unidentified

2 1 ...... ...... 100' t

-.....

--

91*

--

-99* 100' 100'

---

786 152 36 73 15 508

349

Jaeger Grey Gull

Band-tailed Gull

...... 3
1

3
39

94*
49*

---

---

11

Kelp Gull
Franklin's Gull Sabine's Gull Arctic Tern Inca Tern

7
15 --1

1
77* -85* 93*

16
t ----

75*
4 tOO* ---

---15 5

------

------

73
26 24 426 483

Number of species

for which each technique

is the dominant 2 3 0 6 1 1 4

PREY CHARACTERISTICS AND

FORAGING SITUATIONS There were considerable differences in the

would also facilitate finding shoals (Olson 1964).Two other deeply foraging species, the Humboldt Penguin and Red-legged Cormorant, shouldalsohavebeen able to reachdepths
where anchoveta are abundant, but neither

species' compositions among foraging situations (Tables 2-4, 6-7) and in the use of different types of patchesby each species(Table 1). Are these reflections of differences in prey composition and behavior or of limitations in the foraging behaviors of the bird species?
First, I consider the anchoveta. Most shoals

specieswas common at shoals. The penguin formerly fed at shoals(Paessler1922, Murphy 1936).Its presentrarity may reflecta recentserious population decline becauseof incidental

mortality in fishing nets and from poaching

and becauseof the collapse of the anchoveta

occurwithin 40 m of the surfaceand many are found at the surfaceduring the day, but most occurbetween 10 and 20 m depth (Jordan1976,
Johannesson and Vilchez 1980). Surface-feed-

stocks(Duffy et al. in press).Red-leggedCormorants appeared to specialize in foraging in inshore waters with rocky substrates(Coker

1920, Murphy 1936, Brown et al. 1975, pers.

obs.).

ing species,such as gulls and terns, occurred

at fish shoals, so some anchoveta must occur at

The zooplanktonthat I observed being taken

as prey and in the water at zooplankton foraging situations were small (approximately 1 cm), locally very abundant, and present near the surfacefor only short periods of at most 5 min. The most common birds at zooplankton situations were small gulls and terns, which foragedby dipping and surface-seizing. Larger speciesmay have found it energetically inefficient to eat such small prey and to make frequent moves from one short-lived patch to another. Species that forage in groups would

the surface. Other shallowly foraging birds, such as Waved Albatross,jaegers,and the pelican, stole fish from seabird species that foraged at depths where anchoveta were most abundant (Duffy 1980).
The most abundant birds at shoals were three

species that can plunge, pursuit-plunge, or surface-dive after deeper prey if necessary: Pethe ruvian Booby, Guanay Cormorant, and Sooty Shearwater (Table 8). The tendency for these species to forage and feed in large numbers

October 1983]

ForagingPeruvian of Seabirds

807

36t
20-

69

40-

80

lOO -

S.D.

Fig. 1. Durationsof submergence somemore commonPeruvianseabirds of (meansand standarddeviations;numbers= samplesizes).

also have found the patchestoo small: individuals would have done better to forage alone or in smaller groups. Scavengingand feeding opportunities provided by sealions were also short-lived and
small, which would favor small numbers of

surface. Deeper foraging methods might also have made birds more vulnerable to predation by sealions.
SPECIES' INTERACTIONS AND FORAGING GROUPS

Are the speciesin a foraging group present small birds. The exceptionwould be scavenging at corpses that were rich and relatively per- becauseeach has respondedindependently to sistent food sources. In this case, dominance in a common stimulus or doesthe presenceof one interspecific interactions rather than agility speciesaffect either the discoveryor exploitawould be the most useful attribute. The largest tion of prey by other species? Peruvian Boobies tended to be the first to of the gulls, the Kelp Gull, was the corpse-speforage at anchoveta shoals,and Inca Terns were cialist, displacing the other species. The commonspecies both scavenging in and first at zooplankton swarms. They may have sealionsituationsfed by surface-seizing dip- found the patchesof prey and then attracted or ping. Deeper foraging would have been of lit- other species, the boobyand tern may mereor tle use in obtaining fish offal, corpses, bits ly have respondedto new patchesalready apor of fish thrashedto piecesby a sealionon the parent to other species. Cormorantsand other

808
Jaeger spp. 5

DAVID CAMERON DUFFY

[Auk, 100 Vol.

Pomarine Jaeger 7

Humboldt Penguin 14
Kelp Gull 39
16

Waved Albatross

Diving Petrel
-

19
3

Neotrop. Cormorant

Red-legged Cormorant 31
Sabine's Gull 5

Band-tailed

Gull

67

Grey Gull
Franklin's Comic Tern

49
Gull 22 40 Inca Tern
Brown

122
56

Pelican

Guanay Cormorant 57

Society Shearwater 26

' ' ' ' ' ' ' lO ' ' ' ' ' ' lOO i , i , i , lOOO

Per uviar] Booby 118

' ' ' ' ' ' oooo

Fig. 2. Intraspecific group sizes combined from all feeding situations (mean and standard deviation; numbers = sample sizes;x-axis = log-scale).

speciesfrequently ignored plunging by a single booby.Like species the boreal Pacificusin ing Black-legged Kittiwakes(Rissa tridactyIa) to locate patches(Hoffman et al. 1981), however, Peruvian seabirdsmay respondonly to repeated plunging or dipping by "nuclear" (Sealy 1973)or "catalyst" species (Hoffman et al. 1981). Feeding at a patch may be enhanced in a number of ways by the presenceof other species (e.g.Hoffmanet al. 1981).If successful foraging requiresdisruptionof a shoaland its antipredator defenses,massplunging or surface-diving would probablysend fish fleeing from one bird to another, destroy the fish-to-fishorientation necessary coordinatedevasiveaction(Shaw for 1978), or perhaps cause overuse of anaerobic muscletissueso that fish could no longer use bursts of speed to escape(Blaxter 1969). Such anaerobicstress could conceivablyeven lead to
death (Parker and Black 1959). All of these would appear to be more effective with increases in the size of the bird flock relative to

kleptoparasitismis the most visible. The average number of pirates (pelicans, jaegers, and
albatrosses)was so low at shoal and "Other"

foraging situations(Tables2 and 7), relative to the numbersof potential hosts,that piracy does not appear to be a decisive factor in flock foraging. Displacement from prey, particularly carrion, may be much more important in determining which species are present at scavenging situations.Big species, suchas the Kelp Gull, monopolize large carrion, while small scraps and offal go to smaller,more agile birds suchas the Inca Tern. Size may alsoplay a part in groupsfeeding on zooplankton. The larger, surface-seizing gulls settleover the densest part of zooplankton swarms and have the highest foraging rates, while the smaller speciesdip their prey around the periphery, where prey are presumably less abundant. The smaller, fasterspecies have an advantage new patches, at which they can exploit before the larger birds
arrive.

the sizeof the shoalbeing attacked. There must Although one or two speciesare characterbe, however, some upper limit at which the isticof certainforagingsituations, muchlarga food per bird decreases with further increases er number of species minor, occasional are parin flock size. ticipants. It is difficult to imagine how these Among the potential negative effectsof the speciescould coevolvewith others to take adpresence of other speciesin foraging groups, vantageof the foraging situationsstudiedhere.

October 1983]

ForagingPeruvian of Seabirds

809

These peripheral speciesmay themselvesbe major participants other foraging situations in

than thosestudied here. For example, Neotropical Cormorants, Peruvian Brown Pelicans, Peruvian Boobies, Franklin's Gulls, and Black

Summer seabird distribution in Drake Passage, the Chilean Fjordsand off southernSouthAmerica. Ibis 117: 339-356.

COKER,R. C.
56:449-511.

1920. Habits and economic relations

of the guanobirdsof Peru. Proc.U.S. Natl. Mus.

CUSHING, D.H. 1971. Upwelling and productionof

Skimmers(Rynchops niger)seemto be the main

participants foraginggroups estuaries in in and

bays (pers. obs.). Brown (1981) identified a groupof surface-feeding storm-petrels (Hydrobatidae) as characteristic of warmer, offshore

fish. Adv. Mar. Biol. 9: 255-334.

DUFFY, D.C. 1980. Patterns of piracy among Peruvian seabirds: depth hypothesis. a Ibis 122:521535.

waters.Foragingover dolphins (Delphinidae) seemsto be more important in northern waters

(Brown 1981, pers. obs.) but was never seen farther south during this study. Feeding seabirdsare much rarer than birds flying or sitting on the water, so distributional studieshave concentratedon nonfeeding birds. Closer study of feeding groups may serve as a valuable adjunct to distributional studies, because is while feeding that seabirds it have their
closest links to the marine environment.

1981. Seasonal changes the seabirdfauna in

of Peru. Ardea 69: 109-113.

1983a. The ecology of tick parasitism on densely nesting Peruvian seabirds.Ecology 64:
110-119.

1983b. Environmental uncertainty and

commercial fishing: effects on Peruvian guano
birds. Biol. Conserv. 26: 227-238.

-, C. HAYS, M. A. PLENGE. press. The con& In

servation status of Peruvian seabirds. Seabird

ACKNOWLEDGMENTS

I thank

El Ministerio

de Pesca and PESCA

PERU

for permissionto work on the islands and for providing transportation. Cabrera,M. Plenge,and P. D. Yengle providedessential adviceand supportin Peru.
R. G. B. Brown, J. Croxall, H. Horn, P. Prince, D. Schneider,and J. Wiens all improved the manuscript

Workshop.I.C.B.P. Technical Publ. DUNNET, M., & J. C. OLLASON. G. 1982. The feeding dispersal of Fulmars Fulmarusglacialisin the breedingseason. Ibis 124:359-361. ERWlN, M. 1977. Foraging and breeding adaptaR. tions to different food regimesin three seabirds: the Common Tern Sterna hirundo, Royal Tern Sternamaxima,and Black Skimmer Rynchops niger.Ecology58: 389-397.
FISHER, & R. M. LOCKLEY.1954. Seabirds.London, J.,
Collins.

with their comments wife prepared the illustraMy

tions. Field work was funded by The National Science Foundation (DEB77-16077), The International Council for Bird Preservation (Pan American), and

GOULD, J. 1974. Sooty Tern (Sternafuscata). P. Pp. 6-33 in Pelagicstudiesof seabirds the central in
and western Pacific Ocean (W. B. King, Ed.).
Smithsonian Contrib. Zool. 158.

PrincetonUniversity Data analysis and writing were supportedby the BenguelaEcologyProgrammeof

the South African National Committee for Oceano-

HARRIS, M.P. 1977. Comparative ecology of seabirdsin the Galapagos Archipelago.Pp. 65-76 in

Evolutionary ecology Stonehouse C. Per(B. and

rins, Eds.). London, MacMillan.
HOFFMAN, W., D. HEINEMANN, & J. A. WIENS. 1981.

graphicResearch
LITERATURE CITED

The ecology seabird of feedingflocksin Alaska.

Auk 98: 437-456.

ASHMOLE, P. 1971. Seabirdecologyand the maN. rine environmentPp. 223-287 in Avian biology, vol. 1 (D. S. Farner and J. R. King, Eds.). New
York, Academic Press

IDYLL,C. P. 1973. The anchovy crisis. Sci. Amer.

228: 22-29.

JOHANNESSON, & R. VILCHEZ. 1980. Note on hyK.,

droacoustic observations changes distribuof in

tion and abundanceof somecommonpelagic fish

BLAXTER, H. S. 1969. Swimming speeds of fish. J. Proceedings the F.A.O. conferenceon fish beof haviour in relation to fishing techniques and tactics. FAO Fish. Repts. 62: 69-100.
BROWN,R. G. B. 1980. Seabirds as marine animals.

species the coastal in watersof Peru,with special emphasis anchoveta.I.O.C. WorkshopRept. on

28: 287-323.

JORDAN, 1976. Biologiade la anchoveta.parte 1. R.

Resumn de conocimiento actual. Reuni6n de

Pp. 1-39 in Behavior of marine animals, vol. 4 (J.

Burger,B. L. Olla, and H. E. Winn, Eds.).New

York, Plenum Press 1981. Seabirds in northern Peruvian waters,
November-December
Callao. volumn

trabajo sobre el fenomeno conocidocomo "El Nifio." Guayaquil,Ecuador.

--, & H. FUENTES. 1966. Laspoblacionesde aves

1977. Bol. Instit. Mar. Perif34-42

guanerasy su situaci6nactual. Inf. Instit. Mar

Per. 10: 1-31.

extraordinario:

, F. COOKE, K. KINNEAR,& E. L. MILLS 1975. P.

KING, W.B. 1974. Wedge-tailed Shearwater (Puffinus

810

DAVID CAMERON DUFFY pacificus). 59-95 in Pelagic Pp. studies seabirds of

in the central and eastern Pacific Ocean (W. B.

[Auk,Vol. I00

of flocking in Barkley Sound, British Columbia,

in 1979. Colonial Waterbirds 4: I04-113.

King, Ed.). Smithsonian Contrib. Zool. 158. MAIS,K. G. 1974. Pelagicfish surveys the Caliin
fornia Current.
Putnams.

SANTANDER, H. 1981. Patrones de distribucion y fluctuacionesde desovesde anchoveta y sardina.

Bol. Instit. nario: Mar Peril-Callao. volum6n extraordi180-192.

California

Fish. Game Bull. 162.

MURPHY,R. C. 1925. Bird islands of Peru. New York,

1936. Oceanic birds of South America. New

York, Amer. Mus. Nat. Hist.

NELSON, 1979. Seabirds:their biology and ecolB. ogy. New York, A & W Publ., Inc.
OLSON, F. C. W. 1964. The survival value of fish

SCHNEIDER, D.C. 1982. Fronts and seabird aggregations in the southeastern Bering Sea. Marine Biol. Prog. Ser. I0: 101-103. SEALY, G. 1973. Interspecificfeeding assemblages S.
of marine
762-802.

birds

off British

Columbia.

Auk

90:

schooling.J. du Cons. 29: 115-116. PAESSLER, 1922. In der umgebung Coronel's R. (Chile) beobachtetev6gel. J. Ornithol. 70: 430482.

SHAW, 1978. Schooling fishes.Amer. Sci. 66: 166E.

175.

VOGT,W. 1942. Aves guaneras.Bol. Comp. Admora.

Guano 18: 1-132.

PARKER, R., & E. C. BLACK.1959. Muscular fatigue R. and mortality in troll-caught chinook salmon

WALSH,J. J., T. E. WHITLEDGE, E. ESAIAS, L. SMITH, W. R.

S. g. HUNTSMAN, H. SANTANDER, & B. R. DE

(Oncorhynchus tshawytscha). Fish. Res. Board J.

Canada 16: 95-106.

MENDIOLA.1980. The spawning habitat of the Peruviananchovy.Engraulis ringens. Deep-Sea Res.

27A: 1-27.

PAULIK,G.J. 1971. Anchovies, birds, and fishermen in the Peru Current. Pp. 156-185 in Environ-

ment, resources, pollution, and society(W. W.

Murdoch, Ed.). Stamford, Connecticut, Sinauer.

ZUTA, S., & W. URQUIZO. 1972. Temperatura promedio de la superficiedel mar frente a la costa peruana,periodo1928-1969.Bol.Instit. Mar Perfl.
Callao 2(6): 459-520.

PORTER, M., & S. G. SEALY. 1981. Dynamics of J. seabird multispeciesfeeding flocks:chronology