Content uploaded by Janos Hennicke
Author content
All content in this area was uploaded by Janos Hennicke on Mar 27, 2014
Content may be subject to copyright.
RESEARCH ARTICLE
C. Herling Æ B. M. Culik Æ J. C. Hennicke
Diet of the Humboldt penguin (
Spheniscus humboldti
)
in northern and southern Chile
Received: 16 March 2004 / Accepted: 10 December 2004 / Published online: 26 January 2005
Springer-Verlag 2005
Abstract The diet of the Humboldt penguin (Spheniscus
humboldti) was examined and compared in two colonies
in Chile. Field work was conducted on Pan de Azu´ car
Island in northern Chile in the breeding season 1998/
1999 and on the Pun
˜
ihuil Islands in southern Chile over
two successive breeding seasons during 1997/1998 and
1998/1999. Penguin diet was studied by stomach-
pumping birds and analysed by species composition, size
and mass of prey. Fish were the dominant prey item at
both sites, the contribution of cephalopod s and crusta-
ceans varying between sites. Th e fish prey consisted
predominantly of school fish, but there were clear lati-
tudinal differences in fish prey taken. Penguins in the
northern colony consumed primarily garfish ( Scombe-
resox saurus), while birds at the southern colony of
Pun
˜
ihuil fed primarily on anchovy (Engraulis ringens),
Araucanian herring (Strangomera bentincki) and silver-
side (Odontesthes regia). The results showed significant
differences in terms of numbers of fish taken betw een the
two breeding seasons at Pun
˜
ihuil. In 1997/1998 penguins
consumed almost exclusively anchovy, while they fed
primarily on silversides in the successive year. Almost all
prey, except stomatopods, were characterised as being
pelagic species that occur in relatively inshore water,
consistent with the foraging behaviour of Humboldt
penguins. The dependence of Humboldt penguins on
commercially exploi ted, schooling prey species makes
the species particularly susceptible to changes in prey
stocks, due to non-sustainable fisheries management.
Introduction
Investigations on seabirds have shown that food selec-
tion has an important influence on the biology of the
birds: it affects bird activity, distribution, energet ics,
competition, breeding success and rate of survival (e.g.
Furness and Monaghan 1987; Montevecchi et al. 1988;
Carey 1996). Knowledge of feeding ecology and diet
composition therefore is an important basis for under-
standing bird biology. Diet co mposition allows estima-
tion of the impacts of changes in food availability, i.e.
due to climatic changes or to anthropogenic influences.
Seabirds breeding along the coast of Chile encounter
different feeding conditions, resulting from climatic dif-
ferences between northern and southern Chile. This is
also the case for the Humboldt penguin (Spheniscus
humboldti Meyen, 1834), with a breeding range along the
Pacific coast of South America from Isla Foca in the
north (5S) to the Pun
˜
ihuil Islands in the south (41S)
(Williams 1995). Along the coastline the Humboldt
penguin faces many different ecological and climatic
conditions from dry coastal desert climate and nutrient-
rich coastal upwelling regions in the north to lush tem-
perate rainforests and poor coastal upwelling in the
south (e.g. Lalli and Parsons 1997 ; Hennicke 2001). The
coastline of Chile is influenced by the cold, nutrient-rich
Humboldt Current, flowing northwards from Antarctica
(Colling 2001), bringing waters with colder temperatures
than typical for the prevailing latitudes (Dietrich et al.
1975). In the north of Chile the cold sea-surface tem-
peratures are being intensified by the coastal upwelling
that usually brings up cold, nutrient-rich waters from
below the thermocline (Colling 2001). This, in turn,
allows high primary productivity (Arntz and Fahrbach
Communicated by O. Kinne, Oldendorf/Luhe
C. Herling Æ B. M. Culik
Institut fu
¨
r Meereskunde,
Du
¨
sternbrooker Weg 20,
24105 Kiel, Germany
J. C. Hennicke (&)
Zoologisches Institut und Museum,
Martin-Luther-King-Platz 3,
20146 Hamburg, Germany
E-mail: janos.hennicke@uni-hamburg.de
Fax: +49-40-428385980
Marine Biology (2005) 147: 13–25
DOI 10.1007/s00227-004-1547-8
1991), resulting in high amounts of fish such as the
plankton-eating anchovy (Engraulis ringens) (Lalli and
Parsons 1997). The upwelling system of the Humboldt
Current is regularly affected by recurrent periods of
anomalous climatic conditions known as El Nin
˜
o
(ENSO: El Nin
˜
o Southern Oscillation), where the
nutrient-rich surface waters are being replaced by warm
nutrient-poor waters (Colling 2001) and fish becomes
unavailable to seabirds (e.g. Hays 1986; Boersma 1998).
Humboldt penguin populations have dropped dra-
matically (Hays 1986; Luna-Jorquera 1996a). The
commercial fishery that occurs especially along coastal
upwelling systems like those of South America and
South Africa (Duffy et al. 1987a; Montevecchi et al.
1988) is potentially capable of harming penguin popu-
lations as in the case of the African and Magellanic
penguins (Crawford and Shelton 1981, cited in Scolaro
et al. 1999.). The commercial fishery is the biggest threat
to Humboldt penguins nowadays (Ellis et al. 1998).
Other factors that severely affect the populations are
entanglement and drowning in nets (Simeone et al. 199 9)
and recurring food-shortage during El Nin
˜
o events (e.g.
Hays 1986). Humboldt penguins have recently been
classified as vulnerable, the population remaining at low
levels since the early 1980s (Luna-Jorquera and Culik
2000).
Although the Humboldt penguin is one of the main
predators of the Humboldt Current, only little is
known about its feeding preferences along the coast off
Chile, with its cli matic differences between the north
and the south. Until know, the prey of Humboldt
penguins has, to our knowledge, only been documented
by Murphy (Lalli and Parsons 1936) and Wilson et al.
(1989, 1995). The substantial population decline to-
gether with expanding human activities, including
commercial fisheries, suggests that investigations on
feeding preferences of the Humboldt penguin have long
been overdue.
In the present study we investigated the diet of
Humboldt penguins in two Chilean colonies and over
two successive breeding seasons in one colony in order
to examine annual and regional differences. Our aims
here were to characterise the penguins diet, compare
site- and year-specific variation and to assess implica-
tions for competition with commercial fisheries. Overall,
the study shall help to better understand the role of the
Humboldt penguin in the marine ecosystem.
Materials and methods
Study areas
Field work was conducted in two breeding colonies of
the Humboldt penguin (Spheniscus humboldti) in Chile,
which are 1,500 km apart: Pan de Azu´ car Island
(2609¢S; 7041¢W) in Pan de Azu´ car National Park in
northern Chile during the breeding season 1998/1999
and Pun
˜
ihuil Islands in the south (4155 ¢S; 7402¢W)
during the breeding seasons 1997/1998 and 1998/1999.
Pan de Azu´ car Island is located about 1.2 km from the
coast, at the edge of the Atacama Desert. The cold
temperate surface water along the coast is maintained by
the Humboldt Current, flowing northwards from Ant-
arctica (Colling 2001). The island is about 2 km long
and 1 km wide. Pan de Azu´ car is one of the most
important breeding colonies of Humboldt penguins in
northern Chile (Luna-Jorquera 199 6a), with 800 birds
counted on the island in 1998/1999 (Hennicke, unpub-
lished data). The Pun
˜
ihuil Islands are located on the
west coast of Chiloe
´
Island in southern Chile. The
islands are exposed to high precipitation and low
temperatures all year round, with temperatures seldom
exceeding 11C during summer. As opposed to Pan de
Azu´ car, the area surrounding the Pun
˜
ihuil Islands is not
influenced by the upwelling ecosystem of the Humboldt
Current (Lalli and Parsons 1997). Isla Chica, the
smallest of the islands, is about 0.4 km offshore. The
largest island, Isla Grande, is located about 0.7 km from
the coast. The Pun
˜
ihuil Islands have the only known
mixed colony of Humboldt and Magellanic penguins
(Spheniscus magellanicus) in the world (Simeone and
Schlatter 1998), with 150 Humboldt and 500 Magellani c
penguins counted in the year of study (Hennicke,
unpublished data).
Collection of food samples
Stomach samples were obtained from birds returning to
the colonies by the water offloading technique of Wilson
(1984). In the breeding season 1997/1998 penguins were
flushed two to five times. In 1998/1999 penguins were
only flushed twice. During the breeding season 1997/
1998 stomach samples on Pun
˜
ihuil Island were collected
as soon as penguins return ed back from the sea. For this
purpose each nest was individually checked every half
hour. In 1998/1999 stomach samples in both colonies
were obtained from penguins returning from a foraging
trip, caught on their way back from the shore to the
nests. Most of the penguins were breeding adults
returning to their nests; some were moulted and no
longer breeding.
In Pan de Azu´ car field work was conducted in two
intervals. The first sampling period was between 28
November and 2 December 1998. The second sampling
period was 1 week later between 9 and 18 December
1998.
Analysis of food samples
The digested part of the stomach content was separated
from the undigested, i.e. still identifiabl e material like
whole fish, fish tails, fish heads, squid, or crustaceans.
The digested chyme was weighed and immediately
placed in 95% isopropanol for storage and transport to
the laboratory for later analysis. The undigested mate-
14
rial was identified directly whenever possible and then
dried and stored for transport. Analyses were conducted
at the Institut fu
¨
r Meereskunde, Kiel, Germany. Each
sample was sorted into fish, cephalopod and crustacean
components. For analysing the digested material, the
chyme was drained through a 90-lm siev e and then
examined for the presence of any diagnostic prey
remains like otoliths, vertebrae, cephalopod beaks, or
crustacean carapaces. Diet composition from each site
was assessed from complete samples in four ways:
1. frequency of occurrence,
2. number of individuals,
3. wet mass of prey, and
4. length of prey.
Fish
Loose otoliths were collected from stomach samples,
and those still enclosed in fish crania were removed.
Otoliths were identified by direct comparison with a
reference collection compiled for this purpose by Dr. C.
Moreno, Universidad Austral de Chile, Valdivia, Chile,
and by using the identification key from Gellona et al.
(1995). The identification was verified at Universidad de
Concepcio
´
n, Chile, by Dr. M. Araya. Because only few
otoliths were found, fish species were also determined by
examining vertebrae. Loose vertebrae were collected
from the samples, and those still enclosed in fish bodies
were removed. The vertebrae were identified by dir ect
comparison with a reference collection. For this purpose
samples of anchovy (Engraulis ringens), pilchard
(Strangomera bentincki) and silverside (Odontesthes
regia) were caught, weighed and measured in Chile and
sent to Germany, where the vertebral columns were
removed from the fish samples.
Other species were identified with the key by Gellona
et al. (1995). For identification purposes, the caudal
vertebrae were used, because they show less variations
than abdominal vertebrae and are therefore easier to
compare. Furthermore caudal vertebrae are very robust
and more resistant to digestion than other vertebrae
(Watt 1997).
Due to the low number of otoliths found in the
samples, the first caudal vertebrae were used to esti-
mate the number of fish consumed. When no first
vertebrae were present, the anterior or, when no
anterior caudal vertebrae were present, the posterior
caudal vertebrae were counted. The total number was
divided by the number of anterior or posterior caudal
vertebrae, which was counted on average in fresh fish
samples to obtain the total number of fish in the
sample.
Loose otoliths were paired by species and size
together with the otoliths of intact heads enclosing
paired otoliths. Otoliths were considered to represent an
additional fish, when repre senting a species that was not
already represented by the vertebrae.
The length of otoliths and vertebrae were determined
using a calibrated reticule in a microscope. Regressions
relating vertebra length or otolith length, respectively, to
fish length were established to allow estimation of the
original fish size. Regressions were obtained from the
biometric data derived from the otolith and vertebra
collection from fresh fish samples of Engraulis ringens,
Strangomera bentincki and Odontesthes regia. Regres-
sions relating fish length and body mass were obtained
from direct measurements of fresh fish. Lengths of fish
specimens other than the three species mentioned above
were calculated using the proportional method (Watt
1997):
l
1
¼ q
1
l
2
/q
2
ð1Þ
where l
1
is the length of the unknown fish from the
sample (mm), q
1
is the length of vertebrae from the
sample (mm), l
2
is the length of the reference fish (mm)
and q
2
is the length of vertebrae of the reference fish
(mm).
This method requires knowledge of the relationship
between fish length and length of the first caudal verte-
brae (when not present, other caudal vertebrae) from at
least one specimen of the identified species. Regressions
relating fish length and body mass were obtained in this
case from the literature or from the internet (http://
www.fishbase.org).
Cephalopods
All loose cephalopod beaks were removed from the
stomach contents, sorted into upper and lower beaks,
counted and stored in alc ohol. Intact buccal masses were
counted separately before extraction of lower beaks.
Intact specimens and beaks were identified from the
literature with the keys by Clarke (1986), Rocha (1997),
Wolff (1982, 1984) and where possible by comparison
with reference material obtained from Dr. U. Piatkow-
ski, Institut fu
¨
r Meereskunde, Kiel, Germany. The total
number of cephalopods was estimated from the com-
bined number of loose lower beaks and intact buccal
masses. Regressions were obtained from the literature
and applied to all intact beaks in order to estimate the
size of the cephalopods. For Loligo gahi:
ln M gðÞ¼2.25 þ 2.39 ln LRL mmðÞ ð2Þ
and
ln ML mmðÞ¼4:23 þ 1:01 ln LRL mmðÞ ð3Þ
both equations according to Piatkowski et al. (2001);
and for Dosidiscus gigas:
ln M gðÞ¼7:4 þ 2:48 ln LRL cmðÞ ð4Þ
and
ML mmðÞ¼44:2 þ 357:9 LRL cmðÞ ð5Þ
15
both equations according to Wolff (1984), where LRL is
the lower rostral length, ML is the dorsal mantle length
and M is the body mass.
Crustaceans
All crustaceans were counted and identified. Only sto-
matopods and isopods were found. The stomatopods
were identified to the lowest taxonomic level possible
from literature with the keys by Manning (1968) and the
Food and Agriculture Organization (FAO 1993 ). The
identification of the isopods to species or genus level was
not possible, due to rat e of digestio n and small size. The
lengths of all intact stomatopod specimens were mea-
sured from the tip of the rostrum to the tip of the uro-
pod. A regression from a related species (Oratosquilla
nepa) of the same family, obtained from the literature,
was used to estimate the body mass of the stomatopods:
log M ¼4.9665 þ 2.9661 log L ð6Þ
according to James and Thirumilu (1993), where L is the
length (mm) and M is the body mass (g).
Statistics
All data were statistically treated using SPSS 10.0 for
Windows. Data were tested for normality by using
Kolmogorov–Smirnov’s adaptation test (Lamprecht
1992). Due to the non-normal distributed data only non-
parametric tests were used. Two independent random
samples were compared by Mann–Whitney U-test.
Frequencies were compared by v
2
-test. When N was too
small or the expectation values were <5.0, relative fre-
quencies were tested by Fisher’s test. All Fisher tests
were two-tailed. In all tests, P<0.05 was accepted as
indicating statistical significance.
Results
During the breeding season 1997/1998, 11 stomach
samples were collected on the Pun
˜
ihuil Islands; in the
following season, 1998/1999, the stomachs of 8 penguins
could be flushed. At Pan de Azu´ car, 22 stomach samples
were collected in 1998/1999. During the first sampling
period the stomachs of 7 penguins were flushed; 15
samples were taken during the second sampling period.
Frequency of occurrence
At both sites, fish was the predominant prey item in
terms of frequency of occurrence. At Pan de Azu´ car and
Pun
˜
ihuil it appeared in every sample (Table 1). Cepha-
lopods were also found at both sites. This prey taxon
occurred in approximately 55% of penguin stomach
samples in Pun
˜
ihuil in the breeding season 1997/1998
and in 63% of the samples in 1998/1999. In Pan de
Azu´ car cephalopods occurred in 41% of the samples.
Penguins at Pun
˜
ihuil rarely consumed crustaceans in
both seaso ns, while this prey taxon occurred in 45% of
penguin stomach samples from Pan de Azu´ car.
Fish
Overall, Humboldt penguins fed predominantly on
pelagic prey species, particularly those occurring in
aggregations. There were well-defined , area-specific
dietary differences.
At Pun
˜
ihuil penguins fed in both years on anchovy
(Engraulis rigens), Araucanian herring (Strangomera
bentincki), silverside (Odontesthes regia), common hake
(Merluccius gayi) and Inca scad (Trachurus murphyi).
The compositions of species were the same in both
breeding seasons, but there were significant differences
in the occurrence of silversides in the samples. In 1997/
1998 penguins rarely consumed silversid es (9%), while
this species occurred in approximately 88% of penguins
stomach samples in the succ essive breeding season
(P=0.002, Fisher’s test) (Table 2).
At Pan de Azu´ car penguins fed on anchovy and Inca
scad and additionally on garfish (Scomberesox saurus)
and South American pilchard (S ardinops sagax). No
Araucanian herring, silverside, or common hake were
found.
Garfish occurred in every penguin stomach sample
from Pan de Azu´ car during both sampling periods.
Anchovy was the second most important fish species and
occurred in >40% of penguin stomach samples during
both sampling periods (Table 3).
Table 1 Frequency of occurrence (%) of fish, cephalopods and crustaceans in the stomach contents of Humboldt penguins (Spheniscus
humboldti) in Chile
Pun
˜
ihuil Pan de Azu´ car
1997/1998 1998/1999 All samples First sampling period Second sampling period
Samples (n) 11 8 22 7 15
Fish 100 100 100 100 100
Cephalopods 55 62.5 41 71 18
Crustaceans 18 25 45 71 23
16
Cephalopods
All intact cephalopod beaks from the stomach samples
of Pun
˜
ihuil were identified as Loligo gahi. At Pan de
Azu´ car two species were identified from all intact
cephalopod beaks in the samples; Dosidiscus gigas
occurred in 32% of penguin stomach samples, while
L. gahi occurred in approximately 5% of the samples.
Marked changes in the cephalopod diet were measured
between the two sampling periods at Pan de Azu´ car.
During the first sampling period cephalopods occurred
in 71% of penguin stomach samples, while this prey
species occurred only in approximately 18% of the
samples 1 week later (Table 1).
Crustaceans
At Pun
˜
ihuil nearly all crustacean carapaces found in the
samples were identified as isopods. The degree of
digestion combi ned with the small size prevented further
identification. At Pun
˜
ihuil isopods only occurred in 2 of
the 11 samples (18%) in the first year 1997/1998 and in 2
of the 8 samples (25%) in the following year (Table 1).
Only one stomatopod was found in one sample during
the season 1998/1999. It was identified to the family level
as a Lysiosquillidae.
At Pan de Azu´ car stomatopods occurred in 3 of the
22 samples, all taken during the first sampling period.
All stomatopods belonged to the family Squillidae.
Some were identified to the genus level as Pterygosquil-
lidae or as the species Pterygosquillidae gracilipes. Iso-
pods occurred in 71% of the samples at Pan de Azu´ car
during the first sampling period, while this prey species
occurred only in approximately 23% of the samples
during the second sampling period.
Composition by numbers
At Pun
˜
ihuil a total of 195 prey items were analysed in
1997/1998, and 145 prey items, in the following season
(1998/1999). A total of 412 prey items were examined at
Pan de Azu´ car. At both sites penguins fed predomi-
nantly on fish in terms of numbers.
At Pun
˜
ihuil, in 1997/1998, fish accounted for 93% of
all items eaten, followed by cephalopods (6%) and
crustaceans (1%), respectively. In 1998/1999 fish
accounted for almost 88% of all items consumed. At
Pan de Azu´ car fish constituted >75% of the items
eaten, followed by crustaceans (17%) and cephalopods
(8%).
Marked changes in the diet were recorded betw een
the two sampling periods at Pan de Azu´ car. The
proportion of fish consumed was lower during the
first sampling period than in the second. From the total
of 164 prey items counted, 79 were identified as fish in
the first season (48%), whereas in the second seaso n,
from a total of 248 prey items , 229 were analysed as fish
(92%) (P<0.001, v
2
-test). Crustaceans accounted for
almost 37% (n=63) of all items consumed during the
first sampling period, while penguins rarely consumed
crustaceans during the second sampling period (3%)
(n=8) (P<0.001, v
2
-test). Cephalopods constituted
the remaining 13% (n=22) of the diet in the first
sampling period and 4% (n=11) during the second
sampling period, respectively. This difference was sig-
nificant (P<0.01, v
2
-test).
Fish
At Pun
˜
ihuil a total of 309 fish were counted (181 in
1997/1998 and 128 in 1998/1999). In Pan de Azu´ car a
total of 308 fish were counted.
There was a significant seasonal difference in the
numbers of anchovies (P<0.001, v
2
-test) and silversides
(P<0.001, v
2
-test) eaten by the penguins at Pun
˜
ihuil
(Fig. 1). In 1997/1998 penguins fed almost exclusively
on anchovy (E. rigens). From the total of 181 fish con-
sumed 111 were identified as anchovy, which constituted
61% of the fish diet in terms of numbers. In 1998/1999,
from the total of 128 fish counted, anchovy made up
only 27% (n=35), while silversides (O. regia) accounted
for almost 37% (n=48) of the fish consumed.
At Pan de Azu´ car the penguins fed almost only
(>70%) on garfish (S. saurus). Anchovy was the second
most abundant fish species, accounting for 22% of all
308 fish consumed. The othe r two fish species contrib-
uted only to a smaller degree to the diet with 5%
Table 3 Frequency of occurrence (%) of the different fish species in
the stomach contents of Humboldt penguins (Spheniscus hum-
boldti) at Pan de Azu´ car over all samples and during both sampling
periods
All
samples
First
sampling
period
Second
sampling
period
Garfish
(Scomberesox saurus
scombroides)
100 100 100
Anchovy
(Engraulis ringens)
45 43 47
Inca scad
(Trachurus murphyi)
14 14 13
South American pilchard
(Sardinops sagax)
9147
Table 2 Frequency of occurrence (%) of the different fish species in
the stomach contents of Humboldt penguins (Spheniscus hum-
boldti) at the Pun
˜
ihuil Islands
1997/1998 1998/1999
Anchovy (Engraulis ringens) 91 62.5
Araucanian herring
(Strangomera bentincki)
91 50
Silverside (Odontesthes regia) 9 87.5
Common hake (Merluccius gayi) 18 12.5
Inca scad (Trachurus murphyi)9 25
17
analysed as Inca scad ( T. murphi) and 2% as pilchard (S.
sagax). There were significant differences between the
two sampling periods in the numbers of garfish
(P<0.001, v
2
-test) and anchovy (P<0.001, v
2
-test) eaten
(Fig. 2). During the first period, from the total of the 79
fish counted approximately half of the fish were ancho-
vies (46%), garfish being the next most abundant group
of fish, with 30 items consumed (38%).
During the second sampling period garfish accounted
for 82% (n=188) of all 229 fish consumed, followed by
anchovy accounting for 14% (n=31). The contribution
of T . murphy varied significantly (P<0.001, v
2
-test) be-
tween the samplings (Fig. 2). From all 248 prey items
eaten, including cephalopods and crustaceans during the
second samp ling period, the proport ion of garfish con-
sumed was more than four times higher (76%, n=188)
than in the first sampling period (18%, n=30), where a
total of 164 prey items were identified (P<0.001, v
2
-
test).
Cephalopods
At Pun
˜
ihuil, numbers of cephalopod beaks, all identified
as L. gahi, were almost equal during both breeding
seasons. From the total of 195 prey items identified in
the first season, 12 were identified as cephalopods. In
Fig. 1 Composition by
numbers (%) of the fish
component of the diet of
Humboldt penguins (Spheniscus
humboldti) at Pun
˜
ihuil in 1997/
1998 and 1998/1999
Fig. 2 Composition by
numbers (%) of the fish
component of the diet of
Humboldt penguins (Spheniscus
humboldti) at Pan de Azu´ car in
the samples during the first
sampling period and during the
second sampling period
18
1998, 13 of the 145 prey items consumed were cephalo-
pods. At Pan de Azu´ car, of the 33 cephalopod counted
in all samples, only 15 beaks were identified as D. gigas
and 4 as L. gahi. Identification of the beaks was difficult
due to advanced digestion or because the beaks were
broken. More cephalopods were consumed during the
first sampling period (P<0.01, v
2
-test).
Crustaceans
The contribution of the two groups of crustaceans, iso-
pods and stomatopods, to stomach contents was differ-
ent between sites, as well as between the two sampling
periods at Pan de Azu´ car. During both seasons at
Pun
˜
ihuil and during the second sampling period at Pan
de Azu´ car, crustaceans constituted <5% of numbers of
all prey items consumed and these were primarily iso-
pods. At Pan de Azu´ car stomatopods occurred in large
numbers during the first sampling period and co n-
stituted 31% of the number of all prey items consumed.
Isopods, which occurred in >70% of the samples during
the first sampling, constituted only 7% of all items eaten
in terms of numbers.
Composition by wet mass and length of prey
Fish predominated in the diet by wet mass at both sites
followed by cephalopods and crustaceans, respectively.
During both seasons at Pun
˜
ihuil the diet was almost
exclusively fish, with cephalopods and crustaceans
combined accounting for <5% of the mass.
At Pan de Azu´ car cephalopods were eaten in larger
quantities during the first sampling period (36%), but
much less during the second sampling period (10%).
Contribution to the diet by stomatopods in terms of wet
mass was greatest during the first sampling period at Pan
de Azu´ car, but during both samplings crustaceans con-
tributed only to a small degree on a mass per stomach
basis.
Fish
At Pun
˜
ihuil most indivi dual fish were relatively small,
mean lengths ranging between 10.84 and 28.88 cm and
mean masses ranging between 8.84 and 156.12 g (Ta-
ble 4). Size of anch ovy showed no significant differences
between seasons (P=0.171, Mann–Whitney U-test). At
Pun
˜
ihuil in 1997/1998 anchovy contributed most of the
fish prey mass (47%), with silversides being more
important (61%) than anchovies during the successive
year (Fig. 3). Araucanian herrings were the second most
important fish prey at Pun
˜
ihuil during both seasons and
showed no significant differences in size between seasons
(P=0.065, Mann–Whitney U-test).
At P an de Azu´ car, in all samples, mean lengths ran-
ged between 5.4 and 15.06 cm and mean mass between
2.24 and 25.04 g (Table 4). The sizes of garfish eaten by
Humboldt penguins during the first and second sam-
pling periods were significantly different (P<0.000,
Table 4 Mean length (cm) and mass (g) of the different fish species of in the stomach content of Humboldt penguins (Spheniscus
humboldti) at Pun
˜
ihuil (1997/1998 and 1998/1999 samplings) and Pan de Azu´ car (all, first and second samplings)
N Length (cm) SD Weight (g) SD
Pun
˜
ihuil 1997/1998
Engraulis ringens 111 11.09 1.93 8.84 4.46
Strangomera bentincki 65 11.04 1.28 12.04 3.30
Odontesthes regia 2 19.73 47.45
Merluccius gayi 2 28.88 156.12
Trachurus murphyi 1 14.7 30.41
Pun
˜
ihuil 1998/1999
Odontesthes regia 48 19.52 3.58 49.73 30.13
Strangomera bentincki 41 10.97 1.72 11.19 6.95
Engraulis ringens 35 10.96 2.59 10.42 4.20
Trachurus murphyi 3 10.84 2.75 13.53 9.45
Merluccius gayi 1 13.51 11.19
Pan de Azu´ car, all samples
Scomberesox saurus 218 15.06 3.99 7.08 7.73
Engraulis ringens 67 8.36 1.39 3.76 1.88
Trachurus murphyi 16 13.23 2.89 25.04 15.89
Sardinops sagax 7 5.40 1.18 2.24 1.1
Pan de Azu´ car, first sampling
Scomberesox saurus 30 19.25 5.60 15.17 12.10
Engraulis ringens 36 8.34 1.38 3.71 1.67
Trachurus murphyi 11 13.91 3.52 26.20 19.29
Sardinops sagax 2 4.10 1.04
Pan de Azu´ car, second sampling
Scomberesox saurus 188 14.40 3.22 5.79 5.88
Engraulis ringens 31 8.37 1.44 3.82 2.13
Trachurus murphyi 5 13.32 0.43 22.51 2.20
Sardinops sagax 5 5. 93 0.49 2.72 0.74
19
Mann–Whitney U-test). Garfish predominated in the
fish diet during both sampling periods in terms of wet
mass, but the contribution of garfish to the fish prey was
lower during the first sampling period (52%) than dur-
ing the second sampling period (82%). Pilchards con-
tributed only to a small degree to mass (on a per
stomach basis) during both sampling periods, with 0.2%
during the first sampling period and 1% during the
second sampling period (Fig. 4).
Cephalopods
Most cephalopods were relatively small, with a mean
length of 5.2±1.4 cm and a mean mass of 5.5±3.0 g at
Pun
˜
ihuil in the first year and a mean length of
6.4±1.4 cm and mean mass of 8.7±4.3 g in the suc-
cessive season. The length was significantly different
between the two seasons (P=0.000, Mann–Whitney
U-test). At Pan de Azu´ car individual cep halopods were
somewhat larger in size, with a mean length of 10 cm.
Discussion
Methodology
Problems associated with differential retention and
digestion of diagnostic prey remains are inherent in any
study of seabird diet. Digestion and passage rates of
hard prey remains like otoliths or cephalopod beaks are
Fig. 4 Composition by wet
mass (%) of the fish component
of the diet of Humboldt
penguins (Spheniscus humboldti)
at Pan de Azu´ car during the
first sampling period and the
second sampling period
Fig. 3 Composition by wet
mass (%) of the fish component
of the diet of Humboldt
penguins (Spheniscus humboldti)
at Pun
˜
ihuil in 1997/1998 and
1998/1999
20
subject to variation. This variation is affected by the
relative size and thickness of the prey remains, their
chemical composition and the acidity of the predator’s
stomach (Jobling and Breiby 1986). In general, the
keratinous cephalopod beaks are more liable to rete n-
tion than the calcareous otoliths. Accumulation of
cephalopod beaks can result in overestimat es of cepha-
lopods (e.g. Van Heezik 1989). In the case of consumed
fish, with small, fragile, hyaline otoliths, it is probable
that some were digested completely and the numbers
and their contribution to the diet were therefore
underestimated. By using vertebrae for fish identifica-
tion, we tried to avoid underestimation of the relative
importance of fish, since the vertebrae showed little or
no sign of digestion. Investigations on seals showed that
the rate of identification is much higher when using ot-
oliths as well as vertebrae (Pierce et al. 1991). Most
cephalopod beaks present in the stomachs of Humboldt
penguins (Spheniscus humboldti) in this study were still
associated with flesh. This suggests that these prey items
were caught recently, reducing the bias from accumu-
lation.
Penguin stomachs in the breeding season 1997/199 8
at Pun
˜
ihuil were flushed up to five times, whereas in the
following breeding season stomachs were flushed only
twice. Since the aim of this study was not the compari-
son of the absolute numbers of prey items ingested, but
the composition of the diet and the contribution of the
different pre y species to it , we only compared percent-
ages. This enabled us to compare the different breeding
seasons at Pun
˜
ihuil despite the different flushing inten-
sities.
Prey composition
In spite of the problems and biases which act in favour
of underestimating the relative importance of fish and
overestimating that of cephalopods, fish were still the
dominant prey in Humboldt penguin stomachs. Hum-
boldt penguins in this study took mainly anchovies
(Engraulis ringens), Araucanian herring (Strangomera
bentincki), silversides (Odontesthes regia) and garfish
(Scomberesox saurus) (Figs. 1, 2). Contribution of the
different species to the diet differed between sites and
seasons. Penguins in the north eat mainly garfish and
anchovies, which is consistent with the study on Hum-
boldt penguins by Wilson et al. (1995), who found at
Chan
˜
aral, in the north of Chile, that garfish was the
dominant prey item and occurred in 94% of penguin
stomach samples, anchovies being the second most
abundant fish species. Of all items eaten, the proportion
of garfish consumed in the present study was more than
four times higher during the second sampling period
than during the first (76% vs. 18%). Even though the
contribution of anchovies to the total numbers of fish
consumed was higher during the first sampling period,
there were no significant differences between the pro-
portion of anchovies of all items eaten during both
sampling periods. Thus, the low amount of garfish
during the first sampling was compensated during the
second by preying on cephalopods and crustaceans,
mainly stomatopods. In northern Chile pilchards (Sar-
dinops sagax) contributed only to a small degree to the
penguin’s diet, in this study as well in the study by
Wilson et al. (1995).
At Pun
˜
ihuil seasonal shifts between prey taxa
occurred. In 1997/1998, penguins consumed almost
exclusively anchovies, while they fed primarily on sil-
versides in the successive year (Fig. 1). The second most
abundant fish species preyed on at Pun
˜
ihuil was in both
years Araucanian herring. Silversides and Ara ucanian
herring were absent in the study of Wilson et al. (1995).
Instead, South American pilchard (S. sagax) was an
important prey species, which, in turn, was absent in our
investigations.
The principal prey species of this study are predom-
inantly found in inshore areas (http://www.fishbase.org),
which is consistent with the study by Culik and Luna-
Jorquera (1997). They reported the foraging range of
satellite-tracked Humboldt penguins to be between 2
and 92 km from Pan de Azu´ car; 90% of all satellite
locations came from a range of 35 km around the island;
50% came from a range of 5 km around the island.
Comparison with diet studies of congeners
The Humboldt penguin is similar to the other penguins
of the genus Spheniscus in specialising on pelagic school
fish. African penguins (S. demersus) have been reported
to feed on anchovies (Engraulis capensis), pilchards
(Sardinops ocellata) and pelagic gobies (Sufflogobius
barbatus) (Crawford and Shelton 1981, cited in Scolaro
et al. 1999; Randall and Randall 1986). Magellanic
penguins (S. magellanicus) off Argentina take anchovies
(Engraulis anchoita), sprats (Sprattus fuegensis), hake
(Merluccius hubbsi), silversides (Odontesthes spp.) and
squid (Frere et al. 1996; Scolaro et al. 1999). Sprats were
the main food item of Magellanic penguins in southern
Chile. Other prey were cephalopods like Loligo spp. or,
rarely, other fish species (Wilson et al. 1995; Radl and
Culik 1999). Galapagos penguins (S. mendiculus) were
reported to feed on pilchard (Sardinops sagax) (Mills
2000). In previous investigations Humboldt penguins
have been reported to feed on garfish (Scomberesox
spp.), anchovy (Engraulis ringens), pilchard (Sardi nops
sagax) and squid (Todarodes fillipovae) (Wilson et al.
1989, 1995).
The studies of Frere et al. (1996) and Scolaro et al.
(1999) on Magellanic penguins off Argentina show
clearly different feeding preferences over their breeding
range. Penguins in the north feed mainly on anchovy
(Engraulis anchoita), whereas penguins in the south of
Argentina consume primarily sprat (Sprattus fuegensis)
(Frere et al. 1996) and squid (Scolaro et al. 1999 ).
African penguins show similar feeding preferences along
their breeding range. African penguins breeding in the
21
north exclusively take pelagic gobies (Sufflogobius
barbatus) (Crawford and Shelton 1981, cited in Scolaro
et al. 1999); farther south the main prey is anchovy
(Engraulis capensis) (Randall an d Randall 1986). Previ-
ous studies on Humboldt penguins in Chile show similar
results. In the north Humboldt penguins ate almost
exclusively garfish; in central Chile they consumed
anchovy, pilchard and squid; and in the south the pen-
guins consumed anchovy and pilchard (Wilson et al.
1989, 1995). Differences in prey species taken by
Spheniscus penguins between sites in the studies men-
tioned above, as well as in our study, seem to be a
reflection of the availabili ty of pelagic school fish in the
foraging region. Already Wilson et al. (1995) and
Scolaro et al. (1999) noted that inter-site variability is
greater than inter-specific differences in prey species
taken by Spheniscus penguins.
Seasonality
Some fish species show seasonal variation in their dis-
tribution (e.g. Parrish et al. 1989; Castillo et al. 1996).
Even in regions with frequent and strong upwelling
events like in northern Chile, fish species like pilchards
(Sardinops spp.) show migratory behaviour between the
winter spawning season and summer and fall (Parrish
et al. 1989). Seasonal differences in the abundance of
prey species lead to variation in their availability
throughout the year. Seasonality in the diet of Sphenis-
cus penguins has already been documented (e.g. Wilson
1985; Scolaro et al. 1999).
The high abundance of cephalopods and crustaceans
during the first sampling period at Pan de Azu´ car may,
therefore, indicate seasonal differences in the diet of
Humboldt penguins and a change of prey species during
the breeding season. Similar results were found by
Scolaro et al. (1999). Magellani c penguins in Argentina,
at Monte Leon, fed almost exclusi vely on cephalopods
in November, whereas in December cephalopods only
accounted for about 30% of all the prey items con-
sumed. This shows that the prey composition of the
penguins can change in short periods of time.
In 1997/1998, penguins at Pun
˜
ihuil consumed almost
exclusively anchovy, while they fed primaril y on silver-
sides (Odontethes regia ) in the successive year (Fig. 1).
Fish of the genus Odontesthes have already been docu-
mented to occur in the diet of Magellanic penguins (e.g.
Scolaro et al. 1999). The seasona l differences in the diet
of the Humboldt penguin may reflect a change in the
availability of fish species during the season, as was also
documented for Magellanic penguins off Argentina
(Scolaro et al. 1999). Magellanic penguins consumed
squid almost exclusively for most of the breeding season,
but silversides were the major prey item in December
(<60%). Because our investigations during the first
season were conducted in February, but in March dur-
ing the successive season, the differences we found may
reflect seasonal as well as annual variability.
Oceanographic conditions
Previous studies have shown that physical oceano-
graphic conditions like sea surface temperature and
salinity have an influence on distribution and availability
of marine resources (e.g. Castillo et al. 1996; Hansen
et al. 2001). Differences in the oceanographic conditions
along the coast off Chile, resulting from climatic con-
ditions like wind, insolation and rainfall, which in
addition vary seasonally, seem to lead to different spatial
fish distributions. The relationship between physical
oceanographic conditions and pelagic fish distributions
off northern Chile has been studied by Castillo et al.
(1996); they showed that the distribution of three pelagic
fish species was associated with the occurrence and
intensity of thermal and haline fronts. Annual and
regional differences in the prey species taken in our study
may therefore also be related to different thermal fronts
during the sampling periods.
In the case of El Nin
˜
o events, when oceanographic
conditions change and sea surface temperatures
increase, fish stocks have been reported to migrate to
cold-water areas (e.g. Parrish et al. 1989; Arntz and
Fahrbach 1991). The impacts of poor fish availability on
penguins during El Nin
˜
o events have been widely stud-
ied (e.g. Hays 1986; Duffy et al. 1987b; Boersma 1998).
The consumption of stomatopods at Pan de Azu´ car can
possibly be related to the El Nin
˜
o phenomenon in 1997/
1998, which extended late into 1998. Stomatopods have
never previously been recorded in the diet of Humboldt
penguins. These benthic crustaceans seem to be atypical
prey for these seabirds and may reflect poor availability
of fish in the beginning of the post–El Nin
˜
o breeding
season. Duffy et al. (1987b) observed a shift from fish to
cephalopod prey in Magellanic penguins as a result of an
El Nin
˜
o event. It might be that the oceanographic con-
ditions at our study location changed between the two
sampling periods, as a La Nin
˜
a event was expected at the
end of 1998 (http://www.wmo.ch/nino/updat.html#
intro). However, since no detailed information on the
oceanographic conditions was collected in the current
study, this explanation has to remain speculative.
Commercial fisheries
The commercial fishery of Chile intensively exploits of
the main prey species of Humboldt penguins, e.g .
anchovy, silversides, garfish and Araucanian herring. In
1997, the first year of our study, the total catch of these
species was 2.2 million tons (Sernap 1998). This repre-
sents 34% of the overall catch of the entire Chilean
industrial fishery in the same year.
Anchovy is a commercially harvested species
(Whitehead et al. 1988), with Chilean landings of
>0.5·10
6
t in 1998 (Sernap 1999). In the region in the
north of Chile, where Pan de Azu´ car is located (re-
gion III), the total catch of anchovy was 54,847 t, wi th
peaks of landings from January to August (Sernap
22
1999). Even when the landings of anchovy decreased
towards the end of the year, anchovies were the most
harvested fish species in this region. Because schools of
garfish are smaller than those of anchovies (http://
www.fishbase.org), garfish are only to a small degree of
interest to the commercia l fisheries (Sernap 1998, 1999).
There has been no report on the commercial exploitation
of garfish in region III in the north of Chile (Sernap
1999). The predominance of garfish in the north might
indicate that the birds could switch from one prey spe-
cies to another if overfishing of anchovies occurs.
At Pun
˜
ihuil the abundance of anchovy in the penguin
stomach contents was high in the first season and lower
in the second. The total catch of anchovy in the years
and months of the study show the opposite results.
Landing of anch ovy in region X of Chile, where our
study area was located, were low in 1998, with no
landings reported from January to April, whereas
landings of anchovy ad ded up to >1,800 tons in the
same months of the successive year (Sernap 1999, 2000).
The lower amount of anchovy in the penguins’ diet in
March 1999 at Pun
˜
ihuil may be related to the high
numbers of anchovy landings, especially in the first
4 months of the year. The total catch of silversides was
different between years, with total landings in region X
of 416 tons in 1998 and 3,271 tons in 1999 (Sernap 1999,
2000). During 1998 the total landings of silversides were
50 tons or less in 9 months and >50 tons in 3 months of
the year. These results suggest that the abundance of
silversides was low during that year, but was higher in
the suc cessive year, which is shown in the greater
amounts of silversides in the penguins’ diet and the nets
of the fisheries in 1999.
Food requirements during foraging
Knowledge of diet composition is an important basis for
understanding key parameters of seabird energetics and
breeding performance (Furn ess and Monaghan 1987).
Because, e.g., cephalopods have a lower energy content
than fish, the greater the proportion in the diet, the
lower the energy per stomach load. Studies have shown
that African penguins digest cephalopods only half as
fast as anchovy (Wilson et al. 1985) and that chicks grow
more slowly when fed with squid compared to anchovy
(Heath and Randall 1985). Duffy et al. (1987b) con-
nected the massive mortality of Magellanic penguins
during El Nin
˜
o events with a change in the prey spec-
trum, which shifted from anchovies to cephalopods.
Knowledge on the composition of the diet is therefore
important to estimate potential difficulties for the pen-
guins in meeting their energetic demands.
Energy requirements of adult Humboldt penguins at
sea and ashore have been investigated earlier (e.g. Luna-
Jorquera 1996b; Culik and Luna-Jorquera 1997; Culik
et al. 1998; Luna-Jorquera and Culik 2000). Luna-
Jorquera (1996b) calculated that the energetic require-
ments of Humboldt penguins for the 70 days of the
chick-rearing period, including time at sea and ashore,
are 2,240 kJ on average per day per individual penguin.
Assuming a diet almost exclusively of fish (on a wet
mass basis), as found at Pun
˜
ihuil, which has a higher
energy density than a diet containing cephalopods or
crustaceans and which resulted in higher reproductive
success (Hennicke, unpublished data), a mean energy
content of 5.93 kJ g
1
wet mass (mean energy content
for anchovy, Fitzpatrick et al. 1988) and an assimilation
efficiency of 77.3% (Guerra 1992), an individual breed-
ing penguin would need to consume 489 g day
1
to
balance its mean energy requirements. For the relatively
small colony at Pun
˜
ihuil of about 150 Humboldt pen-
guins, the fish consumed during the 70 days of breeding
add up to 5.1 tons. If these demands exceed the food
availability in adjacent waters, they may lead to breeding
failure and consequently to population decline.
The total population of Humboldt penguins in
Chile is estimated at 40,000 individuals (Bernal, per-
sonal communication). The energetic demands of the
total population during the breeding season sum up to
1,400 tons of fish. The intensive commercial exploita-
tion of the main prey species of the Humboldt pen-
guin shows that the commercial fishery is potentially
capable of harming the penguin population if overf-
ishing occurs. Because the potential threat to Hum-
boldt penguins is serious, continued monitoring of
feeding preferences and diet composition could do
much to alert authorities to changes in the marine
ecosystem on which these birds depend and to help
improve fisheries management to avoid und ue com-
petition between man and penguins.
Acknowledgements We are grateful to the people who offered their
help, particularly to Dr. C. Moreno, University of Valdivia, Chile,
for compiling the reference collection of otoliths; Dr. M. Araya,
University Concepcio
´
n, Chile, for helping with identification of the
otoliths; Dr. G. Luna-Jorquera and F. Sapulveda, University of
Coquimbo, Chile, for their support in getting fresh fish samples and
in identifying the different vertebrae; Dr. U. Piatkowski, Institut
fu
¨
r Meereskunde, University Kiel, Germany, for helping with
identification of the squid beaks; and Professor Dr. D. Adelung
and Dr. R. Wilson, Institut fu
¨
r Meereskunde, University Kiel,
Germany, for their overall support. The study was conducted with
permission of the Chilean Ministry of Fisheries (SERNAP) and the
Chilean Nature Conservancy (CONAF), and complied with the
laws of Chile.
References
Arntz WE, Fahrbach E (1991) El Nin
˜
o—Klimaexperiment der
Natur. Birkha
¨
user, Basel, Switzerland
Boersma PD (1998) Population trends of the Ga
´
lapagos penguin:
impacts of El Nin
˜
o and La Nin
˜
a. Condor 100:245–253
Carey C (ed) (1996) Avian energetics and nutritional ecology.
Chapman and Hall, New York
Castillo J, Barbieri MA, Gonzalez A (1996) Relationships between
sea surface temperature, salinity and pelagic fish distribution off
northern Chile. ICES J Mar Sci 53:139–146
Clarke M (1986) A handbook for the identification of cephalopod
beaks. Clarendon, Oxford
Colling A (2001) Ocean circulation, 2nd edn. Prepared by an Open
University Course Team, Butterworth-Heinemann, Oxford
23
Crawford RJM, Shelton PA (1981) Population trends for some
southern African seabirds related to fish availability. In: Cooper
J (ed) Proceedings of the symposium on the birds of the sea and
shore, 1979. African Seabird Group, Cape Town, pp 15–41
Culik BM, Luna-Jorquera G (1997) Satellite tracking of Humboldt
penguins (Spheniscus humboldti) in northern Chile. Mar Biol
128:547–556
Culik BM, Luna-Jorquera G, Oyarzo H, Correa H (1998) Hum-
boldt penguins monitored via VHF telemetry. Mar Ecol Prog
Ser 162:279–286
Dietrich G, Kalle K, Krauss W, Siedler G (1975) Allgemeine Me-
ereskunde, 3rd edn. Gebru
¨
der Borntraeger, Berlin
Duffy DC, Wilson RP, Ricklefs RC, Broni SC, Davis H (1987a)
Penguins and purse seiners: competition or co-existence? Nat
Geol Res 3:480–488
Duffy DC, Arntz WE, Tovar HS, Boersma PD, Norton RL (1987b)
A comparison of the effects of El Nin
˜
o and the Southern
Oscillation on birds in Peru and the Atlantic Ocean. In: Proc
1986 Int Ornithol Cong. Ottawa, Canada
Ellis S, Croxalll JP, Cooper J (eds) (1998) Penguin conservation
assessment and management plan. Report. IUCN/SSC Con-
servation Breeding Specialist Group, Appl Valley, Minn., USA
FAO (Food and Agriculture Organization) (1993) Field guide to
the commercial marine and brackish-water resources of the
northern coast of South America. Species Identification sheets
for fishery purposes, FAO, Rome
Fitzpatrick LC, Guerra C, Aguilar R (1988) Energetics of repro-
duction in the desert nesting sea gull Larus modestus. Estud
Oceanol Fac Recur Mar Univ Autofagasta 7:33–39
Frere E, Gandini P, Lichtschein V (1996) Variacon latitudinal en la
dieta del pinguino de Magallanes (Spheniscus magellanicus)en
la costa Patagonia, Argentina. Ornitol Neotrop 7:35–41
Furness RW, Monaghan P (1987) Seabird ecology. Blackie, Glas-
gow
Gellona FF, Corte
´
s RM, Vignoli MLV (1995) Claves osteolo
´
gicas
para peces de Chile central. Un enfoque arqueolo
´
gico. Uni-
versidad de Chile, Santiago de Chile
Guerra C (1992) Efectos de la nidificacio
´
n estival/invernal, sobre
para
´
metros seleccionados de la historia de vida del pingu
¨
ino
de Humboldt Spheniscus humboldti. Informe Proyecto
Fondecyt 90-0599, Universidad de Antofagasta, Antofagasta,
Chile
Hansen JE, Martos P, Madirolas A (2001) Relationship between
spatial distribution of the Patagonian stocks of Argentine an-
chovy, Engraulis anchoita, and sea temperatures during late
spring to early summer. Fish Oceanogr 10:193–206
Hays C (1986) Effects of the El Nin
˜
o on Humboldt penguin colo-
nies in Peru. Biol Conserv 36:169–180
Heath RGM, Randall RM (1985) Growth of jackass penguin
chicks (Spheniscus demersus) hand reared on different diets.
J Zool (Lond) 205:91–105
Hennicke J (2001) Variabilita
¨
t im Jagdverhalten des Humboldt-
pinguins (Speniscus humoldti) unter verschiedenen ozeanogra-
phischen Bedingungen. PhD thesis, Christian-Albrechts-
Universita
¨
t, Kiel, Germany
James DB, Thirumilu P (1993) Population dynamics of Oratosqu-
illa nepa in the trawling grounds off Madras (Tamil Nadu, In-
dia). J Mar Biol Assoc India 35:135–140
Jobling GM, Breiby A (1986) The use and abuse of fish otoliths
in studies of feeding habits of marine piscivores. Sarsia
71:265–274
Lalli CM, Parsons TR (1997) Biological oceanography—an intro-
duction, 2nd edn. Butterworth-Heinemann, Oxford
Lamprecht J (1992) Biologische Forschung: von der Planung bis
zur Publikation. Pareys Studientexte 73, Verlag Paul Parey,
Berlin
Luna-Jorquera G (1996a) Conservation assessment and manage-
ment plan (CAMP): the Humboldt penguin Spheniscus hum-
boldti. In: Croxall J, Cooper J, Ellis S (eds) Penguin
conservation assessment and management plan workshop,
Cape Town, South Africa. IUCN/SSC Conservation Breeding
Specialist Group, Apple Valley, Minn., USA
Luna-Jorquera G (1996b) Balancing the energy budget for a warm-
blooded bird in a hot desert and cold seas: the case of the
Humboldt penguin. PhD thesis, Christian-Albrechts-Universi-
ta
¨
t, Kiel, Germany
Luna-Jorquera G, Culik BM (2000) Metabolic rates of swimming
Humboldt penguins. Mar Ecol Prog Ser 203:301–309
Manning RB (1968) A revision of the family Squillidae (Crustacea,
Stomatopoda), with the description of eight genera. Bull Mar
Sci 18:105–142
Mills KL (2000) Diving behaviour of two Gala
´
pagos penguins
Spheniscus mendiculus. Mar Ornithol 28:75–79
Montevecchi WA, Birt VL, Cairns DK (1988) Dietary changes of
seabirds associated with local fisheries failures. Biol Oceanogr
5:153–161
Murphy RC (1936) Oceanic birds of South America. American
Museum of Natural History, New York
Parrish RH, Serra R, Grant WS (1989) The monotypic sardines,
Sardina and Sardinops: their taxonomy, distribution, stock
structure, and zoogeography. Can J Fish Aquat Sci 46:2019–2036
Piatkowski U, Pu
¨
tz K, Heinemann H (2001) Cephalopod prey of
king penguins (Aptenodytes patagonicus) breeding at Volunteer
Beach, Falkland Islands, during austral winter 1996. Fish Res
(Amst) 52:79–90
Pierce GJ, Boyle PR, Diack JSW (1991) Identification of fish ot-
oliths and bones in faeces and digestive tracts of seals. J Zool
(Lond) 224:320–328
Radl A, Culik BM (1999) Foraging behaviour and reproductive
success in Magellanic penguins (Speniscus magellanicus): a
comparative study of two colonies in southern Chile. Mar Biol
133:381–393
Randall RM, Randall BM (1986) The diet of jackass penguins
Spheniscus demersus in Algoa Bay, South Africa, and its bearing
on population declines elsewhere. Biol Conserv 37:119–134
Rocha R (1997) Cephalopods in Chilean waters—a review. Mala-
col Rev 30:101–113
Scolaro JA, Wilson RP, Laurenti S, Kierspel M, Gallelli H, Upton
JA (1999) Feeding preferences of the Magellanic penguin over
its breeding range in Argentina. Colon Waterbirds 22:104–110
Sernap (1998) Anuario estadı
´
stico de pesca, 1997. Servicio Nac-
ional de Pesca, Ministerio de Economia, Fomento y Recon-
struccion, Valparaiso, Chile
Sernap (1999) Anuario estadı
´
stico de pesca, 1998. Servicio Nac-
ional de Pesca, Ministerio de Economia, Fomento y Recon-
struccion, Valparaiso, Chile
Sernap (2000) Anuario estadı
´
stico de pesca, 1999. Servicio Nac-
ional de Pesca, Ministerio de Economia, Fomento y Recon-
struccion, Valparaiso, Chile
Simeone A, Schlatter RP (1998) Threats to a mixed-species colony
of Spheniscus penguins in southern Chile. Colon Waterbirds
21:418–421
Simeone A, Bernal M, Meza J (1999) Incidental mortality of
Humboldt penguins Spheniscus humboldti in gill nets, central
Chile. Mar Ornithol 27:145–149
Van Heezik YM (1989) Diet of the fjordland crested penguin during
the post-guard phase of chick growth. Notornis 36:151–156
Watt J (1997) Guide to the identification of North Sea fish
using praemaxillae and vertebrae. ICES Coop Res Rep
220:129–190
Whitehead PJP, Nelson GJ, Wongratana T (1988) Clupeoid fishes
of the world (suborder Clupeoidei). An annotated and illus-
trated catalogue of the herrings, sardines, pilchards, sprats,
shads, anchovies and wolf-herrings, part 2: Engraulidae. FAO
Species Catalogue vol 7, FAO, Rome
Williams TD (1995) The penguins. Spheniscidae. Bird families of
the world. Oxford University Press, Oxford
Wilson RP (1984) An improved stomach pump for penguins and
other seabirds. J Field Ornithol 55:109–112
Wilson RP (1985) Seasonality in diet and breeding success of the
jackass penguin Spheniscus demersus. J Ornithol 126:53–62
Wilson RP, La Cock GD, Wilson M-P, Mollagee F (1985) Dif-
ferential digestion of fish and squid in the jackass penguin
Spheniscus demersus. Ornnis Scand 16:77–79
24
Wilson RP, Wilson M-P, Duffy DC, Araya B (1989) Diving
behaviour and prey of the Humboldt penguin Spheniscus
humboldti. J Ornithol 130:75–79
Wilson RP, Duffy DC, Wilson M-P, Araya B (1995) The ecology of
species replacement of Humboldt and Magellanic penguins in
Chile. Gerfaut 85:49–61
Wolff GA (1982) A beak key for eight eastern tropical Pacific
cephalopods species with relationships between their beak
dimension and size. Fish Bull (Wash DC) 80:357–370
Wolff GA (1984) Identification and estimation of size from the
beaks of 18 species of cephalopods from the Pacific Ocean.
NOAA Tech Rep NMFS 17
25
