Content uploaded by Patricia Majluf
Author content
All content in this area was uploaded by Patricia Majluf on Jun 04, 2014
Content may be subject to copyright.
Behavioral Ecology Vol. 10 No. 3: 242–250
Female-solicited extrapair matings in
Humboldt penguins fail to produce extrapair
fertilizations
Michael K. Schwartz,
a,b
Daryl J. Boness,
a
Catherine M. Schaeff,
b
Patricia Majluf,
c
Elizabeth A. Perry,
a
and Robert C. Fleischer
a
a
Department of Zoological Research, National Zoological Park, Smithsonian Institution, Washington,
DC 20008, USA,
b
Department of Biology, American University, 4400 Massachusetts Avenue,
Washington, DC 20016, USA, and
c
Wildlife Conservation Society, Paul de Beaudiez 520, Lima 27,
Peru
The study reported in this paper demonstrated that Humboldt penguins at Punta San Juan, Peru, despite forming pair-bonds,
are not strictly monogamous in their mating behavior: 19.2% of the study males and 30.7% of the study females (21 nests)
engaged in extrapair copulations. The total number of completed matings observed during the course of this study was 106, of
which 17.9% were extrapair copulations. Using DNA fingerprinting we demonstrated that none of these extrapair copulations
resulted in extrapair fertilizations; all 49 offspring were attributed to the putative father. Location of copulations suggested that
females solicited these extrapair copulations because 89.2% of Humboldt penguin within-pair copulations occurred at the home
burrow, yet extrapair copulations took place at a different location based on the sex of the penguin. Extrapair copulations by
males occurred at their nest, whereas females conducted 92% of their extrapair copulations away from the nest. These results
are most consistent with mate-appraisal and epiphenomenal hypotheses.
Key words:
epiphenomenon, extrapair fertilization,
extrapair mating, female choice, Humboldt penguins, mate appraisal,
Spheniscus humboldti. [Behav Ecol 10:242–250 (1999)]
U
ntil recently monogamy has been considered the pre-
dominant mating system among birds (Lack, 1968).
Nevertheless, behavioral studies have shown that many avian
species previously believed to be monogamous actually engage
in extrapair copulations (EPCs; Birkhead et al., 1990; Gowaty
and Bridges, 1991a; Lifjeld et al., 1993; Morton et al., 1990;
Westneat, 1990). Individuals may engage in EPCs to increase
their fitness relative to others of the same sex who only cop-
ulate with their bonded partner. To evaluate the possible
adaptive significance of this reproductive tactic, the frequency
and nature of EPCs need to be known, as does the extent to
which EPCs lead to extrapair fertilizations (EPFs). Develop-
ments in molecular genetics have provided important tools
such as DNA fingerprinting and DNA microsatellite analyses
that allow for accurate assignment of genetic parentage
(Burke and Bruford, 1987; Jeffreys et al., 1985; Primmer et
al., 1995).
Possible explanations for EPC have been postulated in
terms of trade-offs between the costs and benefits in repro-
ductive effort for both sexes. For males it is often believed
that the production of sperm is relatively cheap (Trivers,
1972); hence the costs of extrapair mating are minimal. Un-
less offspring require extensive paternal care, which may com-
pete directly with EPC effort, the male should employ a strat-
egy of seeking EPCs (Westneat et al., 1990, but see Birkhead
and Møller, 1992). A variety of costs to the male for partici-
pating in EPCs have been noted, all of which must be bal-
anced against the potential increased fitness through in-
Address correspondence to M. K. Schwartz, Wildlife Biology Pro-
gram, University of Montana, Missoula MT 59812; USA. E-mail:
mks@selway.umt.edu.
Received 30 January 1998; revised 26 June 1998; accepted 8 August
1998.
q 1999 International Society for Behavioral Ecology
creased lifetime reproductive success (Gowaty, 1985; Trivers,
1972; Westneat, 1990).
For males the magnitude of the costs and benefits are me-
diated through ecological factors and social factors that influ-
ence the value of paternal care (Dunn et al., 1994; Freeman-
Gallant, 1996; Stuchbury and Morton, 1995). For example, a
male that has to make long foraging trips to partition resourc-
es for his offspring will spend less time seeking EPCs, as will
a male whose offspring benefit from nest guarding (Birkhead
and Møller, 1996; Westneat, 1990). It has also been shown that
increases in coloniality and density (Dunn et al., 1994; Gowaty
and Bridges, 1991b; Møller and Birkhead, 1993; Westneat and
Sherman, 1997; Westneat et al., 1990) will increase the male’s
opportunities for EPCs (Westneat et al., 1990).
Although these social and ecological factors often influence
a female as well, the overall costs and benefits are different
than for the male. Female costs include retaliation by the
paired male if he discovers he has been cuckolded, harass-
ment and possible injury by extrapair males, and increased
rates of parasitism or disease transmission (Westneat et al.,
1990). The numerous hypotheses proposed to explain female
participation in EPCs can be divided into two main categories
(reviewed by Birkhead and Møller, 1992; Westneat et al.,
1990): those associated with direct benefits, which immediate-
ly add to a female’s reproductive success, and those associated
with indirect benefits, which aid offspring by conferring ge-
notypic advantages. Specifically, the direct benefits females
might receive are additional courtship feeding and additional
parental care by the extrapair male. The hypotheses based on
indirect benefits for females include obtaining better genes
from higher quality males (Weatherhead, 1994; Weatherhead
and Robertson, 1979), production of a mixed brood with a
maximum variety of genotypes (Westneat et al., 1990), induc-
tion of sperm competition (Lifjeld et al., 1993), the oppor-
tunity for mate appraisal and acquisition in future breeding
seasons (Colwell and Oring, 1989; Wagner, 1991), and ensur-
ance against male infertility (Westneat et al., 1990).
243Schwartz et al. • Extrapair mating in Humboldt penguins
These hypotheses have mostly been tested with respect to
passerine species (e.g., Birkhead et al., 1990; Morton et al.,
1990; Westneat, 1987a,b, 1990) because passerines are often
abundant, relatively easy to capture and observe, and produce
clutches with numerous offspring. The few studies of nonpas-
serine species have shown EPCs to be infrequent and to lead
to few, if any EPFs (e.g., Decker et al., 1993; Warkentin et al.,
1994). Seabirds, in particular, have demonstrated a low rate
of EPF, possibly due to the extensive biparental care needed
to successfully fledge offspring in an environment of unpre-
dictable food resources (Austin and Parkin, 1996; Birkhead
and Møller, 1996; Hunter et al., 1992; Mauck et al., 1995).
Penguins (
Spheniscidae
) provide an extreme example of the
biparental care needed to provision young. For many species
of penguins both adults embark on foraging trips lasting one
or more days, with the foraging trip of one species (the em-
peror penguin,
Aptenodytes forsteri
) extending to as long as 24
days (Williams, 1995). We therefore might expect many pen-
guin species to exhibit strict monogamy or a low level of EPCs
and EPFs.
Despite a considerable number of studies on penguin cop-
ulation behavior, few have focused on extrapair activity (Wil-
liams, 1996). Two recent studies provide evidence for relative-
ly low levels of EPCs. Hunter et al. (1995) found that 9.8% of
female Adelie penguins (
Pygoscelis adeliae
) engage in EPCs,
but genetic analysis was not presented. St. Clair et al. (1995)
demonstrated, by the use of DNA fingerprinting, that extra-
pair males fertilized 7.7% of royal penguins’ (
Eudyptes schle-
geli
) first eggs. None of the penguins from the genus
Sphen-
iscus
have been examined for EPCs.
In this study, we first addressed the basic questions of wheth-
er Humboldt penguins participate in EPCs, and if so, how
frequently. We then used DNA fingerprinting to determine
the extent to which EPC leads to EPF. To identify which fac-
tors might influence the occurrence of extrapair mating ac-
tivity, we examined the nature and the spatial distribution of
this behavior in relation to the sex of each focal animal.
METHODS
Natural history and study site
Based on osteological and molecular data (Olson, 1985; Sibley
et al., 1988; Simpson, 1975), the six modern genera of pen-
guins appear to be related to Procellariiformes (e.g., albatross-
es) and Gaviiformes (e.g., loons). Thus, penguins are often
classified in the superfamily Procellarioidea with Gaviidae,
Procellariidae, and Fregatidae (Fordyce and Jones, 1990). The
Humboldt penguin,
Spheniscus humboldti,
is classified in the
genus
Spheniscus
with the Magellanic, Galapagos, and Cape
penguins. All members of this genus breed along the South
American or African coast line, with the Humboldt penguins
breeding in colonies between the islets off Algarobo, Chile
(338209 S) and Isla Foca, Peru (58129 S; Hays, 1984).
The Humboldt penguin is the least studied of the 17 species
of penguins (Hays, 1986), with little data published on its
breeding cycle in the wild (Williams, 1995). They breed 10
months of the year, commencing in March with pair forma-
tion and nest selection occurring simultaneously (Zavalaga
and Paredes, 1993). After the bond is established and the nest
is nearly finished, mating begins, usually at a nest site, which
is either an excavated burrow, natural crevice, cave, or de-
pression in the surface layer of guano (dried seabird excre-
ment). Many nests are lined with bird feathers and seaweed
brought to the nest by either the male or the female.
Several days after successful breeding, two eggs are laid 3–
4 days apart. Both parents alternately incubate the eggs for
42 days (Zavalaga and Paredes, 1993). Two peaks of egg laying
occur, one in April and another in August (Murphy, 1936).
Some breeding pairs have successive broods, one in each
peak, with a maximum of two offspring per brood (Murphy,
1936; Zavalaga and Paredes, 1993). Because such a high level
of biparental care is needed to successfully rear offspring, it
is generally believed that Humboldt penguins, like most sea-
birds, are monogamous and maintain a pair-bond throughout
the breeding season, similar to other penguin species (Wil-
liams, 1995).
This study was conducted at Punta San Juan, Peru (158229
S, 758129 W), a guano reserve located on a peninsula in south-
ern Peru, containing the world’s largest Humboldt penguin
breeding area. The reserve is protected by a 3.0-m high, 1.2-
km long concrete wall which isolates the peninsula from land-
ward entry, protecting the breeding colonies from disturbanc-
es by humans and other mammalian predators such as foxes
(
Dusicyon griseus
) and domestic dogs (
Canis familiaris
). The
habitat is a dry desert, with strata of sedimentary material cov-
ered by layers of recent guano (Murphy, 1936), a substrate in
which the penguins dig burrows. Offshore, the marine envi-
ronment is best characterized by unpredictable fluctuations,
with high variance in food supply for top-chain marine pred-
ators, especially during El Nin˜o years (Arntz et al., 1991; Idyll,
1973).
Within the reserve there are several distinct Humboldt pen-
guin colonies. The largest one consists of approximately 1100
breeding pairs of penguins (Zavalaga and Paredes, 1995); the
remainder of the colonies mostly comprise 10–50 pairs of pen-
guins. PT colony, the location of our study, is perched on a
20- to 25-m high cliff and consists of approximately 95 pen-
guin burrows, although only 35 breeding pairs used these bur-
rows.
Behavioral observations
During the first breeding peak in 1995 (from the first week
in March until the end of July), behavioral observations of
penguins at PT were conducted from a blind directly opposite
the penguin colony approximately 20 m away. This unique
viewing situation provided an excellent vantage point for
watching all study nests without visual aides. However, we used
a 15–603 spotting scope (Swift) to confirm all penguin iden-
tifications.
Individual penguins were recognizable by either flipper
bands attached during previous studies or by a unique spot-
ting pattern on the penguin’s breast and stomach. The spot-
ting pattern has been shown to remain constant from the first
molt onward (Sholten, 1989; C. B. Zavalaga, personal com-
munication).
Between 7 and 24 March 1995, we successfully identified
more than 125 individual penguins. This included 26 pairs of
birds that were observed throughout the breeding peak to
determine the level of pair-bonding behavior and within-pair
and extrapair copulations. As copulations are relatively rare
events, we observed the penguin colony from dawn to dusk
to maximize the number of copulations observed. Further-
more, one or two observers scanned the colony continuously
throughout the observation period, specifically searching for
copulations or events that were usually precursors to copula-
tion (all-occurrence sampling; Martin and Bateson, 1986).
The fact that most observed copulations exceeded 2 min sug-
gests it is unlikely many copulations would have been missed
using these observation procedures. We were not able to ob-
serve penguins at night. Consequently, our reported rates of
within-pair and extrapair copulations will be underestimated
insofar as copulatory activity occurs at night. However, there
is no reason to expect a bias in estimating either type of cop-
ulation.
244 Behavioral Ecology Vol. 10 No. 3
Pair-bonding in penguins involves several behaviors. We re-
corded the following pair-bonding behaviors, all of which of-
ten preceded mating attempts: mutual vocalizations, mutual
preening, flipper patting, and bill dueling (Boersma, 1977;
see Results). Mating attempts were classified as ‘‘incomplete’’
if the attempted copulation stopped after mounting but be-
fore cloacal contact (cloacal contact was presumed when the
female lifted her tail feathers and the male arched his tail
feathers toward the female’s cloaca, apparently touching cloa-
cas). On the other hand, they were considered ‘‘successful’’
if cloacal contact was achieved during the mating ritual. How-
ever, a significant proportion of the successful copulations
may not have successful sperm transfer (Hunter et al., 1996).
Copulations were also classified as either within-pair or ex-
trapair. Within-pair copulations occurred between pair-bond-
ed individuals, defined as any two birds seen together and
engaging in pair-bonding behavior on more than two occa-
sions at the same location. EPCs were defined as copulations
with an individual other than the pair-bonded mate (Westneat
et al., 1990).
Based on 265 h of behavioral observations during the first
breeding peak in March, no extrapair matings or mating at-
tempts were witnessed after 0900 h nor before 1600 h. There-
fore, between 7 and 26 August 1995, behavioral observations
were only collected between 0530 h and 1000 h and between
1500 h and 1830 h.
Calculation of oviposition
Because Humboldt penguins are burrow nesters and tend to
desert their nests if disturbed during the courtship and mat-
ing period, we did not repeatedly check nests to determine
the exact day of oviposition. Instead, we used weekly culmen-
length data collected on nonstudy birds in the same popula-
tion (R. Paredes and C.B. Zavalaga, unpublished data) to cal-
culate a first-order regression curve for culmen length against
the chick’s age (
y
5 21.36
x
2 40.0;
r
2
5 .95; SE of chick age
determined from culmen length 5 2.29). Using this regres-
sion equation, we calculated the chick’s age, from which we
subtracted 42 days, the mean incubation period at Punta San
Juan (R. Paredes and C.B. Zavalaga, unpublished data). Al-
though we do not have variance estimates on the mean in-
cubation duration, there is evidence that penguins in general
have an incubation period with a low variance (e.g., the gen-
too penguin’s incubation period is reported to be 35.360.1
days,
n
5 48; Williams, 1990). Thus, we believe we provide a
reasonable estimate of hatch date.
Paternity estimations
Blood collection
We collected blood from 21 penguin families (42 adults and
49 chicks). Due to either the extreme depth or narrowness of
the burrows or to the failure of the clutch before hatching,
only 15 pairs of birds included in the behavioral analysis were
accessible for blood collection. Consequently, blood was also
collected from six additional families of penguins for which
no behavioral data were recorded. We captured birds by hand
in their burrows or with the use of a pole with a small noose
on the end (or a sigmoidal hook for chicks) and moved them
out of the colony for sampling.
Birds were bled from the intradigital vein on the top of the
foot, between the second and third digits (Cheney, 1993; G.D.
Miller, personal communication) using a 23- or 25-gauge nee-
dle with a 3-cc syringe. We collected between 1 ml and 2 ml
of blood from adult birds and between 0.1 ml and 0.2 ml from
chicks. Occasionally, when no blood was obtained using the
intradigital veins of the chicks, the metatarsal vein was used.
Blood was mixed with an equal volume of lysis buffer (Fleisch-
er et al., 1994). As all adult birds that were bled were sighted
at the burrow after handling, we presumed there were no
substantial negative effects associated with handling birds.
Laboratory analyses
DNA was extracted from the blood samples, quantified, and
digested using standard techniques detailed in Fleischer et al.
(1994). Four micrograms of digested DNA were electropho-
resed through a 1.0% gel in TBE buffer for 36 h at 47 volts
and transferred to nylon filters (MSI-NT) by vacuum blotting
(Pharmacia, LKB-VacuGene XL). Samples were probed with
radiolabeled Jeffreys 33.15 and 33.6 probes ( Jeffreys et al.,
1985, 1991) following protocols given in Loew and Fleischer
(1996). The membranes were washed up to three times under
stringent conditions (20-min washes at 608C with 250 ml of
wash solution) and then exposed with intensifying screens at
2808C for 1, 4, 33, and 66 h sequentially (see Fleischer et al.,
1994, for details).
Data analyses
We scored all bands between approximately 2 kb and 18 kb
for all individuals. Gels were run with the chick DNA nested
between that of putative parents and with any potential extra-
pair mates, as indicated by the behavioral data, run in adja-
cent lanes to putative family groups. A size marker was run in
the outer lanes, and a line was drawn between fragments of
the same size. No comparisons were made between gels due
to the high level of variation in band mobility curves (Fleisch-
er et al., 1995). We compared band sharing by marking all
bands on a transparency film overlay with four permanent,
colored ink pens (denoting paternal bands, maternal bands,
shared bands, and nonattributable bands).
It is assumed that DNA fingerprinting bands are inherited
according to Mendelian laws (Jeffreys et al., 1985). Therefore,
nestling bands not found in profiles of either putative parent
indicate a mutation in the offspring or a mistake in parent
identification. To establish a criterion for excluding a putative
male as father, we created a frequency plot of the number of
offspring versus the number of nonattributable bands. This
distribution did not differ significantly from a Poisson distri-
bution (
p
. .5). Based on our expected distribution, we would
have expected to see 1.15 offspring with 2 nonattributable
bands and 0.094 offspring with 3 nonattributable bands (using
our calculated mutation rate of 0.0145 mutations per band
per generation). Thus, we established our criterion at a level
similar to Westneat’s (1990) and Lifjeld et al.’s (1993) criteria:
at least one of the presumed parents was excluded as a parent
if an offspring’s DNA fingerprint profile had three or more
nonattributable bands. In addition, we calculated the proba-
bility of assigning an unrelated male as father to be 1.18310
2
4
with Jeffreys’ 33.15 probe (Table 1).
To further confirm that these fingerprinting data could re-
solve paternities, the putative father of the family group was
substituted with a presumably unrelated ‘‘sham’’ male, and
the number of nonattributable bands was recounted (Fleisch-
er et al., 1994). When a sham male is substituted into a mo-
nogamous family group, the number of nonattributable bands
in the offspring’s DNA profile should increase above the
threshold used to resolve paternities (see above). Thus, we
used this analysis to demonstrate our ability to detect extrapair
copulations.
We calculated band-sharing coefficients (S; Lynch, 1988,
1991) between parent–offspring dyads, offspring–offspring dy-
ads, and among presumably unrelated birds. The variance
around the theoretical mean for first-degree relatives was cal-
culated to determine 95% confidence intervals (CI) around
the theoretical mean predicted for first-degree relatives.
245Schwartz et al. • Extrapair mating in Humboldt penguins
Table 1
Results of analysis of penguin DNA fingerprints
Analysis Result
Number of bands (
f
) 17.100
Mean proportion of bands shared by nonrelatives (
x
) 0.340
Mean allele frequency (
q
)
a
0.171
Heterozygosity (
h
)
b
0.991
Sibling band sharing (
s
)
c
0.515
Expected number of maternal bands (
m
)
d
8.718
Expected number of exclusively paternal bands (
p
)
e
8.382
Probability of assigning an unrelated male as father
(
p
u
)
f
1.18 3 10
2
4
Probability of assigning an uncle as father (
p
r
)
g
3.84 3 10
2
3
a
x
5 2
q
2
q
2
(Jeffreys et al., 1985).
b
h
5 2(12
q
)/(22
q
) (Georges et al., 1988).
c
s
5 (415
q
26
q
2
1
q
3
)/4(22
q
) (Georges et al., 1988).
d
m
5
f
(11
q
2
q
2
)/(22
q
) (Georges et al., 1988).
e
p
5
f
2
m
(Burley et al., 1996).
f
p
u
5
x
p
(Burley et al., 1996).
g
p
r
5
s
p
(Burley et al., 1996).
Table 2
Frequency of within-pair copulations and extrapair copulations
observed during the March and August study periods
March August Total
Within-pair copulations 73 (2.81) 14 (1.75) 87
Extrapair copulations 15 (0.58) 4 (0.5) 19
Total 88 18
The numbers in parentheses are the numbers of observed
copulations divided by the total number of nests observed in March
(26) and in August (8).
Figure 1
Temporal distribution of extrapair (light bars) and within-pair (dark
bars) copulations for all nests where the estimated oviposition date
is within the observation period.
When comparing between
S
(i.e., first-degree relatives versus
unrelated individuals),
t
tests were used, and each individual’s
DNA profile was used only once to maintain independence.
RESULTS
Pair-bonding behavior
The majority of pair-bonding behaviors occurred at the nest
between 0 and 10 days before the first within-pair mating
(mean 5 1.7762.9 days, median 5 1.0 day), with the four
most common behaviors being mutual preening, mutual vo-
calization, bill dueling, and flipper patting. Mutual preening
usually occurred after a bout of self-preening and consisted
of one penguin using its bill to preen the other penguin’s
head or neck. Often the preened bird would reciprocate
preening or continue to self-preen. Mutual vocalization and
bill dueling often occurred when one member of the pair-
bond returned to the nest. During mutual vocalization, the
two birds faced each other, threw their heads back simulta-
neously, and commenced braying (similar to the well-docu-
mented ‘‘ecstatic call’’ that occurs in the
Pygoscelis
genus; Wil-
liams, 1995). Similarly, bill dueling displays also occurred
when the pair reunited and the two animals, facing each oth-
er, moved their heads back and forth, banging their bills. Bill
dueling was always followed by flipper patting.
Flipper patting was common and was solely a male behavior
directed toward females. In all cases the male approached the
female from behind and rapidly moved his flippers along her
body. Twenty-one percent (
n
5 33) of the recorded flipper
pattings led to the male rapidly rubbing his bill along the neck
of the female, forcing her horizontal to the ground in the
process, and ultimately mounting her.
Mating behavior
Despite pair-bonding behavior, not all within-pair copulation
attempts succeeded; often the female rejected the male’s ini-
tial attempts at mating or the mating was incomplete. Rejected
mating attempts occurred when the female refused to lie on
her ventrum, preventing the male from mounting her. As the
study was not designed to collect detailed behavioral data, we
were unable to accurately assess the proportion of male mat-
ing attempts that were refused by the female. However, incom-
plete copulations, defined as a mating in which the male
mounted the female but cloacas never touched, accounted for
26.3% of the within-pair copulation attempts which the female
initially accepted.
In addition to copulations between bonded partners, 19
EPCs were witnessed. Therefore, the total number of com-
pleted matings observed during the course of this study was
106, of which 17.9% were EPCs (Table 2). Six of the EPCs
inolved males from one of the study nests; 13 involved focal
females. The relative frequency of EPCs varied slightly be-
tween study nests (from one to three for females, and from
one to two for males). Overall, 42.3% (11/26) of the bonded
pairs observed had at least one mate participating in at least
one EPC, either at the nest or away from it.
Timing of mating behavior relative to oviposition
Copulations that occur during a female’s fertile period but
before eggs are laid are more likely to result in fertilizations
than those occurring either too early or too late. We therefore
examined when copulations occurred relative to oviposition.
Using the regression equation for nestling growth in culmen
length (described in methods) and subtracting 42 days from
each estimated hatch date, we found all observed extrapair
and within-pair copulations occurred before the estimated ovi-
position date calculated for individual nests.
Five study females laid their eggs at the end of our obser-
vation period. For these birds we failed to detect a difference
in the timing of within-pair copulations (11.7 days prior to
egg laying 6 5.51) compared to the timing of EPCs (13.7 days
prior to egg laying 6 4.46; Figure 1), but had low power to
detect a difference (b , 0.5). In all cases except one (nest
246 Behavioral Ecology Vol. 10 No. 3
Table 3
The distribution of within-pair and extrapair copulations relative to
the home burrow
At home
burrow
One burrow
away
More than
one burrow
away
Within-pair 74 6 3
Extrapair
Females 1 0 12
Males 5 1 0
Total 6 1 12
Four within-pair copulations were not used for this analysis because
information on their location was not collected.
104), the pair male was the last one to copulate with his fe-
male. The last observed within-pair copulation occurred on
average 9.8 days (65.17) before the estimated oviposition of
the first egg.
Other patterns of EPCs were examined to determine if any
trends in the timing of within-pair versus extrapair copulation
could be established. Two of the 15 (13.3%) females partici-
pated in EPCs before any recorded pair behavior with their
ultimate mate. In one other case (nest 133, second breeding
peak), after 5 days of pair-bonding behaviors (without ob-
served copulation) with her mate from the previous breeding
peak, the female participated in an EPC. A within-pair mating
did occur subsequently.
Seasonal differences
Because approximately half of the Humboldt penguins at
Punta San Juan breed twice within one year (C. B. Zavalaga,
personal communication), we had the opportunity to exam-
ine differences that may exist between the two breeding
peaks. Of the 26 pairs that were observed in March, 30.7% (
n
5 8) laid a second clutch in August. Seventy-three of 87
(83.9%) within-pair copulations recorded during the two
breeding peaks were recorded in March. However, because
observations in August were made only on those animals that
laid a second clutch and observation periods were shorter, we
standardized these data for the number of nests observed and
by observer effort. Thus, we compared only afternoon and
evening sessions during those days that the nest was active in
March and August. This reduced the proportion of copula-
tions in March (relative to August) to 65.9%.
There was no significant difference in the number of with-
in-pair copulations observed in March (0.2960.41 within-pair
copulations per day per nest) compared to August (0.1560.12
within-pair copulations per day per nest; Wilcoxon,
p
5 .44,
n
5 8). None of the birds engaged in EPCs in both March
and August, nor was the overall frequency of EPCs different
for the two time periods (Wilcoxon,
p
5 .57,
n
5 8). However,
due to the extremely small sample size and hence the ability
to detect a significant difference, the power of the Wilcoxon
test was calculated and determined to be low.
Distribution of matings
All observed copulations except one occurred at a burrow.
The one exception took place on the beach where animals
stage before going to sea to preen and feed. Not all within-
pair copulations occurred at the pair’s burrow, although most
(89.2%) of the 83 within-pair copulations from the 26 study
pairs occurred at the home burrow. Because pairs often en-
gage in multiple copulations, to test whether within-pair cop-
ulations occurred significantly more often at the home bur-
row, a Wilcoxon paired-sample test was performed (
p
, .01).
Interestingly, 66.7% (6/9) of within-pair copulations away
from the home burrow occurred at an adjacent burrow, while
others occurred more than one nest away (Table 3). In con-
trast to within-pair copulations, 68.4% (13/19) of EPCs tend-
ed to occur away from the focal animal’s burrow (Table 3).
Furthermore, only 1 of these 19 matings occurred at an ad-
jacent burrow, the remainder being at least two burrows away.
EPCs might be expected to occur in some parts of the col-
ony more than others (e.g., opportunities for EPCs might be
greater for birds in those areas near the ocean, where all birds
need to travel for food). To further examine the distribution
of EPCs within the colony, we divided the colony into north,
central, and south sections, in accordance with natural breaks
in the topography. Penguins did not preferentially engage in
EPCs in one particular area of the colony versus another (
G
test,
p
5 .25).
Similarly, EPCs did not tend to occur more frequently with
individuals from nests along the path leading from burrows
to the ocean than those from nests off the path (
G
test,
p
.
.5). Female EPCs had the same probability of occurring at
burrows in their path to the ocean (50%,
n
5 12) as in the
other direction, at burrows located away from this path.
To determine if there was a sex difference in location of
EPCs, we used a Fisher’s Exact test to examine the association
between the sex of the animal engaging in an EPC and the
location of the copulation. Most individuals engaged in only
one EPC, which could be easily classified, as either at burrow
or away from burrow. One female had both an at-burrow and
an away-from-burrow EPC. Regardless of whether that fe-
male’s copulations were classified as at burrow or away from
burrow, female EPCs occurred away from their burrows
(92.3% of EPCs), whereas focal male EPCs occurred at their
burrows (83.3% of EPCs;
p
5 .032).
DNA fingerprinting to assess parentage
DNA fingerprinting was completed on a total of 21 family
groups, with a mean of 2.461.3 offspring per year (1.760.47
offspring per clutch), to determine whether putative parents
were the actual parents. Putative parents were identified for
15 of these families using the extensive behavioral data col-
lected at colony PT. Six additional families from a smaller col-
ony for which behavioral data were not available were includ-
ed. For these families, adults were sampled sequentially while
at the nest and presumed to be the parents.
The DNA fingerprints produced from Jeffreys 33.15 probes
had a mean of 17.864.23 bands scored for adults (males,
18.864.1; females, 16.964.1;
t
test,
p
5 .14,
n
5 21) and
16.862.8 bands scored for offspring. For the second probe,
Jeffrey’s 33.6, a mean of 19.661.23 scorable bands were pro-
duced between 2 kb and 18 kb, with no significant differences
in the number scored for adult males and females or adults
and offspring (
t
test,
p
. .5, both cases).
Jeffreys 33.15 produced an
S
of 0.3460.15 among presumed
unrelated individuals (
n
5 126, 95% CI, 0.17–0.52) and was
not significantly different using pairings of presumed unre-
lated individuals from different age–sex classes (Table 4). The
S
between parents and offspring and between full-siblings did
not differ from the theoretical
S
for related animals (Lynch,
1991; Table 5).
In most cases the DNA fingerprints of the putative parents
accounted for all of a nestling’s bands. However, to ensure
that this result only occurred with the parent–offspring com-
binations, the analysis for nonattributable bands was redone
247Schwartz et al. • Extrapair mating in Humboldt penguins
Table 4
Mean band-sharing coefficients between presumed unrelated
individuals calculated from using Jeffreys probe 33.15 with standard
deviations and sample sizes
Unrelated individuals Mean SD
n
Male–male 0.34 0.24 10
Female–female 0.34 0.13 11
Male–female 0.34 0.16 18
Mates 0.37 0.12 21
‘‘Sham’’ male–offspring 0.38 0.08 49
Total unrelated individuals 0.34 0.15
Table 5
Mean band-sharing coefficients between presumed first-degree
relatives calculated from using Jeffreys probe 33.15 with standard
deviations and sample sizes
Related individuals Mean SD
n
Female–offspring 0.68 0.13 21
Male–offspring 0.68 0.11 21
PT colony, parent–offspring 0.69 0.09 30
S3 cave colony, parent–offspring 0.67 0.13 12
Theoretical
a
0.67
b
No significant differences were detected between any of these
pairings.
a
The theoretical band-sharing coefficient is based on Lynch (1991).
b
The 95% confidence interval for this value is 0.51–0.83.
with the putative male being substituted with a presumed un-
related ‘‘sham’’ male. The number of nonattributable bands
produced for the putative parents–offspring analysis was sig-
nificantly lower than the number of nonattributable bands
produced for the sham male (plus parental female)–offspring
analysis (
t
test,
p
, .001, {x-}
real
5 0.2660.49 {x-}
sham
5
4.2362.05; Figure 2A versus 2B). Hence resolution between
putative males and extrapair males was possible.
There were five cases where
S
of chicks and sham males
were outside the upper limit of the 95% CI surrounding the
mean for unrelated individuals (Figure 2b). In four of these
cases, the offspring were from the same nest. Therefore, two
additional sham males were substituted sequentially for the
original sham male, both times significantly reducing
S
to well
within the 95% CI. The
S
between the original sham male
and the putative father was found to be high (0.88), strongly
suggesting a first-order relationship. Parsimony suggests that
the putative father is the biological father. The four offspring–
putative father comparisons produced only one nonattribut-
able band and displayed band-sharing coefficients between
0.82 and 0.83, whereas the apparently related sham male pro-
duced 14 nonattributable bands (in total), despite its high
degree of similarity to the offspring.
When examining putative fathers and offspring, two points
were above the upper limit of the 95% CI for the theoretical
S
of first-degree relatives. In both cases, we considered the
offspring to be products of within-pair fertilizations because
S
with the putative male is higher than with the sham male, and
the number of nonattributable bands with the putative family
is lower. However, we used a second and third sham male to
test these dyads. Again, there were more nonattributable
bands and a lower
S
with the sham males than with the pu-
tative male, failing to provide evidence for exclusion of the
presumed father. There was only one family group that had
two nonattributable bands in one chick (Figure 2A). We used
Jeffreys 33.6 to reexamine family groups in which chicks dis-
played two nonattributable bands with Jeffreys 33.15. The
mean
S
between unrelated individuals was 0.9460.03 (
n
5
25) and between first-degree relatives was 0.9860.02 (
n
5 28).
Although these means are significantly different (
t
test,
p
,
.005), this probe is not useful to exclude presumed parents.
Given the high
S
between unrelated individuals once mater-
nal bands are accounted for in an offspring’s DNA finger-
printing profile, on average less than one band remained for
use in paternal identification (well below the exclusion crite-
ria of three bands). Therefore, Jeffreys 33.6 does not provide
necessary data needed to identify putative males.
Comparison of EPC and EPF rates
The failure of EPCs to yield EPFs might be explained by the
proportional representation of sperm favoring the pair-bond-
ed male, due to multiple copulations by the pair-bonded male.
Therefore, we tested to see if the rate of within-pair copula-
tion is significantly different from the rate of EPC for each
pair of penguins. Within-pair copulations occurred a greater
number of times than EPCs (Wilcoxon,
p
5 .002), yet the
proportion of extrapair to within-pair copulations was differ-
ent from the proportion of EPFs to within-pair fertilizations
(proportions test,
p
, .05).
DISCUSSION
Frequency of extrapair copulations and fertilizations
This study has shown that Humboldt penguins, despite form-
ing pair-bonds, are not strictly monogamous in their mating
behavior, as 19.2% of the study males and 30.7% of the study
females engaged in EPCs (affecting a total of 42.3% of the
nests). None of these EPCs apparently resulted in EPFs. We
were able to assess paternity for 21 families and 49 offspring.
The
S
among unrelated Humboldt penguins was 0.34 and
0.94 for Jeffreys 33.15 and Jeffreys 33.6, respectively. There-
fore, only the Jeffreys 33.15 data were useful for paternity
analysis. Assigning paternity based on both
S
and the presence
of nonattributable bands was straightforward for most of the
offspring.
Since molecular genetic techniques became more accessible
to behavioral ecologists, numerous studies have shown that
behavioral observations alone may not be a good indicator of
an individual’s reproductive success (Yamagishi et al., 1992).
In most cases the observed level of EPCs (Lifjeld et al., 1993;
Møller, 1987; Westneat, 1987a,b) has drastically underestimat-
ed the actual level of EPFs. Our finding of a high rate of EPCs,
but no evidence of these yielding fertilizations is highly un-
usual. Only one other study has reported such results. Hunter
et al. (1992) showed that in another seabird, the Northern
fulmar (
Fulmarus glacialis
), 2.4% of all copulations involved
an extrapair male (16% of the breeding females participated
in EPCs), with EPCs occurring throughout the breeding cycle.
However, none of these copulations led to EPFs.
Prelaying copulation hiatus
The large temporal gap we found between the last breeding
attempt by a pair and the oviposition of the female is consis-
tent with the idea that long-term sperm storage occurs in
some seabirds (Hatch, 1983; Birkhead and Møller, 1992). Al-
though there is a possibility that this long interval is an artifact
of the back-calculation method we used to estimate the date
of egg laying, or an artifact of a small sample size, these results
are consistent with several other seabird studies. Imber (1976)
reported that the grey-faced petrel (
Pterodroma macroptera
)
248 Behavioral Ecology Vol. 10 No. 3
Figure 2
(A) The relationship between band-sharing coefficients from DNA
fingerprinting between 49 chicks and their putative fathers
(determined by behavioral data) and the number of
nonattribuatble bands in chicks. The dotted line is the lower limit
of the 95% confidence interval for unrelated individuals
(determined theoretically from unrelated individuals). (B)
Distribution of the 49 chicks that were DNA fingerprinted versus
‘‘sham’’ males. The dotted line is the upper limit of the 95%
confidence interval for unrelated individuals. The oval encloses a
family of four chicks (see text for details).
laid fertilized eggs 2 months after copulation, a presumed ad-
aptation allowing for extremely long prelaying absences.
Hatch (1987) reported similar results for northern fulmars,
and A. Chiaradia (personal communication) reported similar
results for little blue penguins (
Eudyptes minor
).
If sperm is stored and released sequentially, then the last
male to copulate with the female has an advantage over pre-
vious males. However, if sperm is released simultaneously, mix-
ing can occur, and the likelihood of fertilization will be based
on the relative proportion of each male’s sperm in the fe-
male’s reproductive tract (Hunter et al., 1992). The fact that
we found no EPFs, but saw EPCs, and in all but one case the
pair-bonded male was the last to copulate with a female, sug-
gests a recency effect in sperm storage in Humboldt penguins.
Female control of mating
Westneat et al. (1990) suggested that females act in one of
three ways with respect to EPCs. They either actively solicit,
apparently resist, or passively accept EPC attempts from the
male. Although penguins are capable of severe fights leading
to major injury or occasionally death, in this study females
rejected copulations from both extrapair males and their ma-
tes without any physical punishment, suggesting that female
cooperation is essential for mating to occur.
The majority (89.2%) of Humboldt penguin within-pair
copulations occur at the home burrow, yet this trend does not
hold true for EPCs. When separated by sex, it becomes ap-
parent that EPCs by males tend to occur at their nest. Females,
however, conduct 92% of their EPCs away from the nest. They
show no pattern to where they mate away from the nest. For
example, they do not tend to mate with males along the path
to the ocean, as one might expect if females were simply being
coerced or attracted by males occupying nests along travel
routes.
Female solicitation of EPCs has been reported in only a few
species of birds (Birkhead et al., 1990; Fujioka and Yamagishi,
1981; Hatch, 1987; Møller, 1990; Smith, 1988; Venier et al.,
1993). Venier et al. (1993) suggested that tree swallow females
solicit EPCs because those copulations that occurred when
females visited neighboring nests resulted in more fertiliza-
tions than copulations initiated by males. Hatch (1987) found
female northern fulmars not only soliciting extrapair matings,
but also visiting numerous nests and engaging in pair-bonding
behaviors.
Westneat et al. (1990) argued that soliciting EPCs costs little
for some species, such as female northern fulmars. Northern
fulmars are a synchronous breeding species, producing a sin-
gle egg per year, and require biparental care to successfully
fledge their offspring. If a male discovers that his female has
engaged in an extrapair mating, it is unlikely that he will re-
nest in that year because all females will have already initiated
breeding. His choice then is either to abandon the clutch for
the year, guaranteeing that the offspring at the nest will per-
ish, or to provide parental care to offspring of uncertain pa-
ternity. If he has copulated multiple times with his pair-bond-
ed female, the probability of the offspring being his may be
greater than the probability of it being the extrapair male’s,
and hence he may still provide parental care.
Humboldt penguins differ from northern fulmars in that
the breeding cycle is not as synchronous, and males may be
more likely to abort the clutch and remate if they discover
their pair-bonded female has participated in an EPC. Despite
this, female Humboldt penguins solicited extrapair matings,
suggesting some net benefit to these matings that outweigh
the added risks of being abandoned by her mate.
Hypothesis addressing costs and benefits to female penguins
Several hypotheses about benefits that females might acquire
from extrapair mating have been suggested for other species.
However, the majority of these hypotheses require the exis-
tence of EPF, which we did not find with Humboldt penguins.
Therefore, hypotheses that state that females will seek EPCs
to (1) increase the genetic diversity of the brood or (2) obtain
better genes from an extrapair male (sexy-son hypothesis) or
to induce sperm competition, seem implausible as no EPFs
were found in this study. Two additional hypotheses, which
better explain the results of this study, are presented below in
detail.
Colwell and Oring (1989) suggested that EPCs might be
used for appraisal and acquisition of future mates. They found
that female spotted sandpipers (
Actitis macularia
) engaged in
249Schwartz et al. • Extrapair mating in Humboldt penguins
extrapair mating after their clutch was completed, suggesting
the function of this behavior was not for current fertilization.
Subsequently, in the next breeding season, the spotted sand-
piper females paired with males with whom they had previ-
ously engaged in EPCs. Wagner (1991) found that EPCs in
razorbill (
Alca torda
) females occurred during infertile peri-
ods, again suggesting EPC is used for future mate appraisal
and acquisition. Wagner also argued that the use of EPCs for
mate appraisal is not uncommon and can be applied to other
published studies (e.g., Fujioka and Yamagishi, 1981).
Females using copulation to assess males should participate
in many EPCs with multiple males. Furthermore, Wagner
(1991) predicted that females using this strategy would posi-
tion themselves to receive EPCs, but would successfully resist
insemination. Our data are somewhat consistent with this;
25% of the females copulated with multiple mates, and these
copulations did not lead to fertilizations. However, at least
33% of the female Humboldt penguins that were in a pair-
bond for multiple seasons still participated in EPCs and began
breeding with the same mate the year after this study (C.B.
Zavalaga and R. Paredes, personal communication). If pen-
guins were using EPCs to appraise future mates, we would
have expected to see some mate switching between March and
August. Yet, no females that sought EPCs in the March breed-
ing peak changed mates in the subsequent August breeding
peak.
In both the Colwell and Oring (1989) study and the Wagner
(1991) study, the EPCs were done after the breeding cycle.
This was not the case for Humboldt penguins. While this
alone would not preclude the mate appraisal and mate ac-
quisition hypothesis, the combination of data suggest that
mate appraisal alone does not explain EPCs in Humboldt pen-
guins.
It is possible that EPCs are not the product of selection.
They may occur simply because they are not selected against.
However, if this were true, we would expect that male and
female Humboldt penguins would participate in EPCs in a
random pattern around the colony and their nests. In this
study females appeared to solicit extrapair matings away from
their nest, while males stayed at their nests, suggesting it is
not an epiphenomenal event. It is possible that without selec-
tion, the male’s breeding system drifted in one direction
(waiting at the nest to solicit extrapair mating), while the fe-
male’s breeding system drifted in another (roaming the col-
ony to seeking extrapair mating). Hence EPC as an epiphe-
nomenon is still a likely explanation for these data.
In summary, EPCs occur in Humboldt penguins and appear
to be under female control. Although no single hypothesis
alone explains our results, our data best support the hypoth-
eses that female Humboldt penguin use EPCs either to ap-
praise future mates or that this behavior is epiphenomenal.
We thank Carlos Zavalaga and Rosana Paredes for field help and for
sharing their valuable knowledge of the penguins. We thank L. Paz
Soldan, C. Rivas Medina, M. Boness, E. Winer, ‘‘Tatanka’’ Uerena, and
O. Riofrio for help in the field. Assistance was generously given in the
Molecular Genetics Laboratory at NZP, especially by W. Piper, C. Mc-
Intosh, and S. Goldsworthy, and by many members of the Department
of Zoological Research, NZP. J. Block and J.C. Riveros helped in ac-
quiring the necessary permits, and G. Miller, R. Davis, M. Stetter, and
B. Karesh all provided advice on handling the penguins and collecting
the blood. P. Gowaty provided valuable comments. This project was
possible through support from the Friends of the National Zoo, The
Smithsonian Institution, American University, and the Helmlinger
Fund. It was completed under CITES import permit 800195, CITES
export permit 364 from Peru, INRENA resolucion ministerial 0659-
95-AG, USDA permit 38877, and Ministerio de Agricultura permit
967-95.
REFERENCES
Arntz W, Pearcy WG, Trillmich F, 1991. Biological consequences of
the 1982–83 El Nin˜o in the Eastern Pacific. In: Pinnipeds and El
Nin˜o: responses to environmental stress (Trillmich F, Ono KA, eds).
Berlin: Springer-Verlag; 22–42.
Austin JJ, Parkin DT, 1996. Low frequency of extra-pair paternity in
two colonies of the socially monogamous short-tailed shearwater
Puffinus tenuirostris.
Mol Ecol 5:145–150.
Birkhead TR, Burke T, Zann R, Hunter FM, Krupa AP, 1990. Extra-
pair paternity and intraspecific brood parasitism in wild zebra finch-
es (
Taeniopygia guttata
), as revealed by DNA fingerprinting. Behav
Ecol Sociobiol 27:315–324.
Birkhead TR, Møller AP, 1992. Sperm competition in birds: evolu-
tionary causes and consequences. London: Academic Press.
Birkhead TR, Møller AP, 1996. Monogamy and sperm competition in
birds. In: Partnerships in birds: the study of monogamy (Black JM,
ed). Oxford: Oxford University Press; 323–343.
Boersma PD, 1977. An ecological and behavioral study of the Gala-
pagos penguin. Living Bird 115:43–93.
Burke T, Bruford MW, 1987. DNA fingerprinting in birds. Nature 327:
149–152.
Burley NT, Parker PG, Lundy K, 1996. Sexual selection and extrapair
fertilization in a socially monogamous passerine, the zebra finch
(
Taeniopygia guttata
). Behav Ecol 7:218–226.
Cheney C, 1993. A technique for restraining and drawing blood from
Humboldt penguins. Penguin Conserv 5:1–4.
Colwell MA, Oring LW, 1989. Extra-pair mating in the spotted sand-
piper: a female mate acquisition tactic. Anim Behav 38:675–684.
Decker MD, Parker PG, Minchella DJ, Rabenold KN, 1993. Monoga-
my in black vultures: genetic evidence from DNA fingerprinting.
Behav Ecol 4:29–35.
Dunn PO, Whittingham LA, Lifjeld JT, Robertson RJ, Boag PT, 1994.
Effects of breeding density, synchrony, and experience on extrapair
paternity in tree swallows. Behav Ecol 5:123–129.
Fleischer RC, Fuller G, Ledig DB, 1995. Genetic structure of endan-
gered clapper rail (
Rallus longirostris
) populations in Southern Cal-
ifornia. Conserv Biol 9:1234–1243.
Fleischer RC, Tarr CL, Pratt TK, 1994. Genetic structure and mating
system in the palila, an endangered Hawaiian honeycreeper, as as-
sessed by DNA fingerprinting. Mol Ecol 3:383–392.
Fordyce RE, Jones CM, 1990. Penguin history and new fossil material
from New Zealand. In: Penguin biology (Davis LS, Darby JT, eds)
San Diego, California: Academic Press; 419–446.
Freeman-Gallant CR, 1996. DNA fingerprinting reveals female pref-
erence for male parental care in Savanah sparrows. Proc R Soc
Lond B 263:157–160.
Fujioka M, Yamagishi S, 1981. Extramarital and pair copulations in
the cattle egret. Auk 98:134–144.
Georges M, Lequarre A-S, Castelli M, Hanset R, Vassart G, 1988. DNA
fingerprinting in domestic animals using four different minisatellite
probes. Cytogenet Cell Genet 47:127–131.
Gowaty PA, 1985. Multiple parentage and apparent monogamy in
birds. In: Avian monogamy (Gowaty PA, Mock DW eds). New York:
Chapman and Hall; 11–21.
Gowaty PA, Bridges WC, 1991a. Behavioral, demographic, and envi-
ronmental correlates of extrapair fertilizations in eastern bluebirds,
Sialia sialis.
Behav Ecol 2:339–350.
Gowaty PA, Bridges WC, 1991b. Nest box availability affects extra-pair
fertilization and conspecific nest parasitism in eastern bluebirds,
Sialia sialis.
Anim Behav 41:661–676.
Hatch SA, 1983. Mechanism and ecological significance of sperm stor-
age in the northern fulmar with references to its occurrence in
other birds. Auk 100:593–600.
Hatch SA, 1987. Copulation and mate guarding in the northern ful-
mar. Auk 104:450–461.
Hays C, 1984. The Humboldt penguin in Peru. Oryx 18:92–95.
Hays C, 1986. Effects of the 1982–83 El Nin˜o on Humboldt penguin
colonies in Peru. Biol Conserv 36:169–180.
Hunter FM, Burke T, Watts SE, 1992. Frequent copulation as a meth-
od of paternity assurance in the northern fulmar. Anim Behav 44:
149–156.
Hunter FM, Davis LS, Miller GD, 1996. Sperm transfer in the Adelie
penguin. Condor 98:410–413.
250 Behavioral Ecology Vol. 10 No. 3
Hunter FM, Miller GD, Davis LS, 1995. Mate switching and copulation
behaviour in the Adelie penguin. Behaviour 132:2–16.
Idyll CP, 1973. The anchovy crisis. Sci Am 228:6 22–29.
Imber MJ, 1976. Breeding biology of the grey-faced petrel Pterodroma
Macroptera gouldi.
Ibis 118:51–64.
Jeffreys AJ, MacLeod A, Tamaki K, Neil DL, Monckton DG, 1991.
Minisatellite repeat coding as a digital approach to DNA typing.
Nature 354:204–209.
Jeffreys AJ, Wilson V, Thein SL, 1985. Individual-specific ‘‘finger-
prints’’ of human DNA. Nature 316:76–79.
Lack D, 1968. Ecological adaptations for breeding in birds. London:
Chapman and Hall.
Lifjeld JT, Dunn PO, Robertson RJ, Boag PT, 1993. Extra-pair pater-
nity in monogamous tree swallows. Anim Behav 45:213–229.
Loew S, Fleischer RC, 1996. Applications of molecular methods to the
assessment of genetic mating systems in vertebrates. In: Molecular
zoology, advances, strategies, and protocols (Ferrans JD, Palumbi,
SR, eds). New York: Wiley-Liss; 133–161.
Lynch M, 1988. Estimation of relatedness by DNA fingerprinting. Mol
Biol Evol 7:478–484.
Lynch M, 1991. Analysis of population genetic structure by DNA fin-
gerprinting. In: DNA fingerprinting: approaches and applications
(Burke T, Dolf G, Jeffreys AJ, Wolff R, eds). Switzerland: Birkha¨user
Verlag; 1–19.
Martin P, Bateson P, 1986. Measuring behaviour: an introductory
guide. Cambridge: Cambridge University Press.
Mauck RA, Waite TA, Parker PG, 1995. Monogamy in Leach’s storm-
petrel: DNA fingerprinting evidence. Auk 112:473–482.
Møller AP, 1987. Behavioral aspects of sperm competition in swallows
(
Hirundo rustica
). Behaviour 100:92–104.
Møller AP, 1990. Sexual behaviour is related to badge size in the
house sparrow
Passer domesticus.
Behav Ecol Sociobiol 27:23–29.
Møller AP, Birkhead TR, 1993. Cuckoldry and sociality: a comparative
study of birds. Am Nat 142:118–140.
Morton ES, Forman L, Braun M, 1990. Extrapair fertilizations and the
evolution of colonial breeding in purple martins. Auk 107:275–283.
Murphy RC, 1936. Oceanic birds of South America, vol I. New York:
American Museum of Natural History.
Olson SL, 1985. The fossil record of birds. In: Avian biology (Farner
DS, King JR, Parkes KC, eds). New York: Academic Press; 80–238.
Primmer CR, Møller AP, Ellegren H, 1995. Resolving genetic relation-
ships with microsatellite markers: a parentage testing system for the
swallow
Hirundo rustica.
Mol Ecol 4:493–498.
St. Clair CC, Waas JR, St. Clair R, Boag P, 1995. Unfit mothers? Ma-
ternal infanticide in royal penguins. Anim Behav 50:1177–1185.
Scholten CJ, 1989. Individual recognition of Humboldt penguins.
Spheniscid Penguin Newslett 2:4–8.
Sibley CG, Ahlquist JE, Monroe BL, 1988. A classification of the living
birds of the world based on DNA-DNA hybridization studies. Auk
105:409–423.
Simpson GG, 1975. Fossil penguins. In: The biology of penguins
(Stonehouse B, ed). London: Macmillan; 19–42.
Smith SM, 1988. Extra-pair copulations in black-capped chickadees:
the role of the female. Behaviour 107:15–23.
Stuchbury BJ, Morton ES, 1995. The effect of breeding synchrony on
extra-pair mating systems in songbirds. Behaviour 132:675–690.
Trivers RL, 1972. Parental investment and sexual selection. In: Sexual
selection and the descent of man, 1871–1971 (Campbell B, ed).
Chicago: Aldine; 136–179.
Venier LA, Dunn RO, Lifjeld JT, Robertson RJ, 1993. Behavioural pat-
terns of extra-pair copulation in tree swallows. Anim Behav 45:412–
415.
Wagner RH, 1991. The use of extrapair copulations for mate appraisal
by razorbills,
Alca torda.
Behav Ecol 2:198–203.
Warkentin IG, Curzon AD, Carter RE, Wetton JH, James PC, Oliphant
LW, Parkin DT, 1994. No evidence for extrapair fertilizations in the
merlin revealed by DNA fingerprinting. Mol Ecol 3:229–234.
Weatherhead PJ, 1994. Mixed mating strategies by females may
strengthen the sexy son hypothesis. Anim Behav 47:1210–1211.
Weatherhead PJ, Robertson RJ, 1979. Offspring quality and the po-
lygyny threshold: ‘‘the sexy son hypothesis.’’ Am Nat 113:201–208.
Westneat DF, 1987a. Extra-pair copulations in a predominantly mo-
nogamous bird: observations of behavior. Anim Behav 35:865–876.
Westneat DF, 1987b. Extra-pair fertilizations in a predominantly mo-
nogamous bird: genetic evidence. Anim Behav 35:877–886.
Westneat DF, 1990. Genetic parentage in the indigo bunting: a study
using DNA fingerprinting. Behav Ecol Sociobiol 27:67–76.
Westneat DF, Sherman PW, 1997. Density and extra-pair fertilizations
in birds: a comparative analysis. Behav Ecol Sociobiol 4:205–215.
Westneat DF, Sherman PW, Morton ML, 1990. The ecology and evo-
lution of extra-pair copulations in birds. Curr Ornithol 7:331–369.
Williams TD, 1990. Annual variation in breeding biology of Gentoo
penguins
Pygoscelis papua
at Bird Island, South Georgia. J Zool 222:
247–258.
Williams TD, 1995. The penguins, Spheniscidae. Oxford: Oxford Uni-
versity Press.
Williams TD, 1996. Mate fidelity in penguins. In: Partnerships in
birds: the study of monogamy (Black JM, ed). Oxford: Oxford Uni-
versity Press; 268–285.
Yamagishi S, Nishiumi I, Shimoda C, 1992. Extrapair fertilization in
monogamous bull-headed shrikes revealed by DNA fingerprinting.
Auk 109:711–721.
Zavalaga C, Paredes R, 1993. A study of the Humboldt penguin at
Punta San Juan Peru. New York: Wildlife Conservation Society.
Zavalaga C, Paredes R, 1995. A study of the Humboldt penguin at
Punta San Juan Peru. New York: Wildlife Conservation Society.
