en
t
i
R
eJacanas represent the extreme in avian sex-role reversal. We studied the division of labour between the two
sexes in parental care of the eggs and chicks in wattled jacanas, Jacana jacana, in the Republic of Panama.
Males were the sole incubators of the eggs (515 clutches). Males were also the sole caretakers of chicks for
97% of broods (NZ 252). Caretaking activities included continuous close association with the young,
leading them to foraging locations, brooding them during cold temperatures and rain, and maintaining
nearly constant vigilance against threats of predation. Although females were also observed to perform all
of these direct chick care behaviours, they did so very rarely (3% of broods) and only when the male was
unavailable to provide such care himself. This happened either when females laid new clutches for males
that were still tending dependent chicks, or when predators killed males, leaving broods defenceless and
untended. The only regular contribution to chick care by females was providing defence against predators.
The male was usually the ?rst parent to detect a predator, but he recruited the female to join him in
confronting threats that he could not deter alone. We conclude that, although males perform the
overwhelming majority of parental care in wattled jacanas, females provide critical backup support. By
displaying parental behaviours facultatively, only when the male is unable to provide them adequately
himself, female wattled jacanas are able to enhance the survival of their eggs and young while typically
expending only minutes per day in such efforts.
2004 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd.
Parental investment refers to anything done for an off-
spring that increases its chance of survival while de-
creasing the parents? ability to produce additional
offspring (Trivers 1972). It includes the cost of producing
the gamete plus any additional time or effort expended on
behalf of the offspring until it becomes independent. In
the vast majority of animal species, parental investment
by females exceeds that by males. In birds, for example,
females produce extremely large and costly eggs, and
when there is an asymmetry in egg care behaviour by
the two sexes, it is usually the female that performs the
majority (or all) of the incubation (Lack 1968; Clutton-
Brock 1991).
Differences in the relative parental investment of the
two sexes often cause females to be unavailable for future
reproductive events longer than males. The potential
reproductive rate of males then exceeds that of females
(Clutton-Brock & Vincent 1991). Females and the parental
investment they provide become valuable ?resources? for
males, and males compete among themselves for access to
females (Williams 1966; Trivers 1972; Andersson 1994).
In a small minority of animal species, the asymmetry of
relative parental investment is reversed. Among birds,
females always invest more in gametes than males, but
males of some species expend much more time and effort
in the care of eggs and young than do females. Such sex-Division of labour in par
sex-role-reversed shoreb
STEPHEN T. EMLEN*?
*Department of Neurobiology a
ySmithsonian Tropical Research Ins
(Received 24 February 2003; in
?nal acceptance 27 August 2003; published o
ANIMAL BEHAVIOU
doi:10.1016/j.anbCorrespondence and present address: S. T. Emlen, Department of
Neurobiology and Behavior, Cornell University, Ithaca, NY 14853,
U.S.A. (email: ste1@cornell.edu). P. H. Wrege is at the Department of
Ecology and Evolutionary Biology, Cornell University, Ithaca, NY
14853, U.S.A.
84
0003?3472/03/$30.00/0 2004 The Antal care behaviour of a
ird, the wattled jacana
& PETER H. WREGE*?
d Behavior, Cornell University
itute, Balboa, Republic of Panama
tial acceptance 14 May 2003;
nline 25 August 2004; MS. number: A9563)
, 2004, 68, 847?855
hav.2003.08.034role reversal in the provisioning of parental care behaviour
has been reported in about 35 species of birds in 13
families (Jenni 1974; Ridley 1978; Oring 1982, 1985).
One of these avian families, Jacanidae, comprises eight
species of shorebirds, all of which have adapted to live on
mats of aquatic vegetation in freshwater marshes and
swamps throughout the tropics. Seven of the eight species
7
ssociation for the Study of Animal Behaviour. Published by Elsevier Ltd.
848of jacana show nearly complete sex-role reversal. Females
lay extremely small eggs relative to their body size (Ross
1979), thus reducing the costliness of egg production.
Males appear to be the sole incubators of eggs and to
provide the majority of care for the dependent young
(Jenni 1974, 1996). As would be predicted, the potential
reproductive rate of females greatly exceeds that of
males (Butchart 2000; Emlen & Wrege 2004) and females
compete among themselves for access to males and the
attendant parental care that males provide. Females are
also much larger than males in all sex-role-reversed
jacanas, with female mass being 148?183% that of males
(reviewed in Emlen & Wrege 2004). In six of these species,
females form pair bonds with multiple males, a mating
system known as simultaneous polyandry (Jenni 1996).
Males in such associations share their female with other
males, termed ?comates?.
In this paper, we quantify the roles of male and female
wattled jacanas, Jacana jacana, in providing parental care
to their eggs and young. The research was part of a broad,
7-year study of the behavioural ecology of the species. We
con?rm that males are the sole providers of direct care for
all clutches and for the overwhelming majority of broods.
However, we also document that females are fully capable
of showing the full suite of chick care behaviours under
rare but predictable circumstances. Females also play a
regular role in predator defence of both the eggs and
young. We conclude that females play an important role
in supplementing the parental care activities of the male.
We speculate that the reason why division of labour is not
more complete in jacanas is that females are able to
contribute signi?cantly to nest and chick survival without
signi?cantly compromising their ability to engage in other
activities.
METHODS
Study Area
We studied the parental behaviour of wattled jacanas at
a study site on the Chagres River, 2.5 km northwest of the
town of Gamboa, in the Republic of Panama. The area is
part of the Soberania National Park, which lies within
a zone of semideciduous lowland rain forest. At our study
site, the river was approximately 0.9 km wide; it supported
extensive mats of ?oating aquatic vegetation on which
many hundreds of wattled jacanas resided. The Chagres
River is a permanent body of water and jacanas bred on it
during all months of the year. The river runs into, and
provides the main water source for, the Panama Canal.
Jacanas established territories, constructed their nest
platforms and reared their young on a ?oating vegetative
habitat that consisted primarily of three species: hydrilla,
Hydrilla verticillata, water hyacinth, Eickhornia crassipes,
and water lettuce, Pistia stratiotes. Resident jacanas from
the study site spent virtually all of their time within the
con?nes of their defended territories and, to our knowl-
edge, never ventured ashore. The habitat was essentially
ANIMAL BEHAVIOUR, 68, 4two-dimensional and afforded great visibility, both to the
birds and to human observers. Hyacinth extended lessthan 1 m above the water surface, whereas the dominant
water lettuce grew only 10?15 cm in height.
Study Population
We focused our attention on a 13-ha area (approximate-
ly 325! 400 m) located in the centre of the Chagres River.
The site was part of a larger area of contiguous habitat that
stretched over 0.6 km east?west and 0.3 km north?south.
On a monthly basis, an average of 18 resident females
(range 4?29), 30 resident males (range 9?40), and a vari-
able-sized population of nonresident ??oaters? occupied
the site. The network of territories of our study birds
merged imperceptibly into neighbouring territories of less-
studied individuals. No part of the study area was closer
than 0.2 km to the nearest shore and water depth typically
varied between 4 and 9 m.
Data were obtained during seven ?eld seasons, typically
beginning in January and averaging 7.5 months (1990:
March?June; 1991 and 1993: January?June; 1992: Febru-
ary?July; 1994 and 1995: January?December; 1996: Feb-
ruary?June). Depending on the time of year, the research
team varied from one to six persons, with effort concen-
trated between February and June.
Capture and Marking
We attempted to capture all resident jacanas, including
chicks and juveniles, as well as many ?oaters, on the study
area. We did so using a variety of techniques, including
the use of mist nets, walk-in traps and playbacks of
?comfort? calls to chicks. By the end of each ?eld season,
an average of 78% of all resident females and 90% of all
resident males had been banded. Overall, a total of 150
females, 311 males and 172 juveniles were captured, meas-
ured and had blood samples collected. Each bird was ?tted
with a unique combination of one to three colour leg
bands plus a numbered aluminium band. All bands were
placed above the tarsal?metatarsal joint to increase visi-
bility.
Nest Checks
We monitored 515 nests, 342 (66%) of which were
found before or during egg laying. Nests were checked
every 2 days throughout incubation until the day of hatch.
Broods of chicks were checked every 4 or 5 days until 25
days of age, and less frequently thereafter.
Behavioural Observations of Mating System
and Parental Care
We obtained behavioural data using focal, scan and
ad libitum techniques (Altmann 1974). Each resident
female was observed at least twice weekly, for 2 or more
hours, to con?rm her number of mates and to gauge
reproductive activity. A female was considered to be paired
with a male when she met four criteria: (1) she consis-
tently foraged with the male; (2) she defended most or all
849of the male?s territory against other females; (3) she
responded to the ?rally? calls of the male by ?ying to the
male?s location; and (4) she solicited the male for cop-
ulations. A female was considered to be polyandrous when
she met the four criteria with two or more males
simultaneously.
To quantify the role of the two sexes in incubation, we
conducted 38 scan samples at 17 nests. At 5-min intervals,
for a total of 57 h, we recorded the location of the female
and male relative to their nests. Incubation constancy was
calculated as the percentage of scans during which either
bird was at the nest site. Additionally, we recorded the sex
of the incubating bird present during regular (alternate
day) checks of the contents of all nests (NO 2500 visits).
To quantify the roles of male and female in nest and
chick defence, we conducted focal samples on 15 territo-
ries for a total of 45 h. The identity of each intruder
( purple gallinule, Porphyrula martinica, common gallinule,
Gallinula chloropus, female jacana or male jacana) onto the
territory was recorded, as was the behavioural response
of the resident pair of jacanas. From these data we
calculated the frequency of intrusions by nest and chick
predators, the responses of both male and female jacanas
to such predators, and the success of any antipredator
defences. Differences between the sexes in the probability
of being the ?rst to detect and respond to an intruder were
examined, for each intruder type, by chi-square tests.
Differences in the responses of each sex to different types
of intruder (whether ?rst responders recruited their mates
for joint defence) were tested by logistic regression anal-
ysis (single degree of freedom orthogonal contrasts) incor-
porating band number as a factor to control for unequal
numbers of observations of different individuals. We ?rst
compared responses to male conspeci?c intruders against
those to all types of predators ( gallinules and female
jacanas), combined. We then omitted male conspeci?c
intruders and compared responses to gallinules against
those to intruding female conspeci?cs.
Finally, many ad libitum observations were made during
the course of daily research activities on the study site.
Behavioural observations were concentrated in the morn-
ing hours from ?rst light until 1100 hours and again in
the afternoon from 1500 hours until dark. Observations
were made from canoes, using binoculars and spotting
telescopes. A total of 137 person-months of observations
were compiled during the study.
Statistical Tests
The SAS statistical software system, version 8.1 and
SYSTAT version 5.2 were used for statistical analyses.
Mean values are reportedGSE, unless otherwise indicated.
Experiment-wise type I error rates were held at the 0.05
level.
RESULTS
Males as the Sole Caretakers of EggsParental care in wattled jacanas was typi?ed by marked
division of labour and extreme sex-role reversal. Only themale incubated the clutch, which was almost invariably
four eggs. Of 342 clutches found before laying was
complete, 339 had a ?nal clutch size of four eggs whereas
three had three eggs. The incubation interval (measured
from the laying of the last egg until hatch) was 25 days.
Females were never seen to incubate. In 57 h of scan
sampling of 17 females paired to incubating males, no
female ever ventured closer than 8 m to an active nest.
Furthermore, in none of more than 2500 checks of nest
contents was the female ever observed to incubate or to be
at the nest site. On more than 30 occasions, we continued
incubation observations until dark and, in every instance,
the male was the member of the pair on the nest. Females
roosted solitarily, elsewhere on the territory.
On nine occasions, incubating males disappeared ( pre-
sumably predated), leaving clutches untended. In no
instance was a female observed to incubate, or even to
defend, these unguarded nests, and the clutches disap-
peared by the following day.
Male incubation constancy was extremely low. Males
with active nests spent, on average, 41% of their time at
the nest. This included time spent both incubating and
shading the eggs.
Despite this low incubation constancy, males lost weight
while tending the eggs. We captured and weighed ?ve
individual males both early and late during their incuba-
tion. Each of the ?ve males lost weight as incubation
progressed, with an average weight loss of 0.3 g per day
(Table 1). Assuming a constant rate of loss, this would
represent a loss of 7.5 g over the 25-day incubation period,
or 8% of average male body mass.
Hatching usually occurred synchronously, with all four
eggs typically hatching within a few hours of one another.
As each chick emerged, the male removed the eggshells,
carrying and dropping them several metres from the nest
area.
Males as the Primary Providers of Chick Care
The precocial chicks were mobile within hours of hatch-
ing and were actively led away from the nest by themale. It
was the male that set the direction of travel and de-
termined the location of foraging. He brooded the chicks
under his wings in the cold and rain. The male occasion-
ally led very young chicks back to the nest platform at
night and brooded them there for several nights following
hatching. During their ?rst few weeks of life, the chicks
remained within a few metres of the male. As they became
older, they wandered more widely, although they contin-
ued to forage as a loose unit and to follow and respond to
the male?s vocalizations and other behaviours until they
were about 2 months old.
The male maintained vigilance over the dependent
chicks at all times and communicated to them when
conditions merited alert or alarm. To minor disturbances
and distant predators, he uttered repetitive ?kek-kek? (alert)
calls and led the chicks in a direction away from the
danger. If a potential predator (or human) approached
EMLEN & WREGE: PARENTAL CARE IN WATTLED JACANASmore closely, he gave a staccato rattle (danger) call to
which the chicks responded by running short distances
thus nearly 3 months (4 days of receiving eggs, 25 days of ?danger? and ?comfort? calls. Only after her two young
b
)
o
850incubation and 50?60 days of chick care).
Females as Backup Providers of Chick Care
Female jacanas typically did not associate at all with
dependent chicks. This pattern occurred for 235 of 242
observed broods (97%). On very rare occasions, however,
a female provided full and active care for her young. Such
instances occurred in only seven (3%) of the observed
broods. These seven exceptions occurred under two very
speci?c circumstances: ?rst, when a male was simulta-
neously tending both a brood of young and a clutch of
eggs, and second, when predation upon a male with
chicks left a brood defenceless and untended.
We observed 17 cases where a female laid a clutch of
eggs for a male before the young from his previous brood
had reached independence (Table 2). In none of these 17
cases did the female have another, ?unoccupied?, mate
available to receive the clutch of eggs. Thus, the males that
received these clutches were, in essence, attempting to
reached independence 28 days after the death of her mate
did she accept a new male mate onto the territory.
We examined the in?uence of age and number of chicks
on a female?s decision to provide or not to provide active
chick care. We used the sample of 12 males performing
?double duty? for which we had suf?cient data (O3 ob-
servation periods) plus the sample of three broods for
which the male parent disappeared, leaving chicks un-
tended. When broods were partitioned into two groups
based on chick age ( young broods:!35 days; old broods:
O35 days) at the time the male received the new clutch or
was predated, females were signi?cantly more likely to
provide active care to the young broods ( ?ve of six young
broods received care versus two of nine old broods;
c12Z 5.40, P! 0.025). In a logistic regression analysis
incorporating both age and number of young, age was
nearly signi?cant in predicting active care by the female
(Wald?s c12Z 3.57, PZ 0.06), whereas number of young
was not (Wald?s c12Z 0.39, PZ 0.53). Because the sample
size was small, the corresponding power of the regressionand then either crouching and freezing in the vegetation
or sinking under the water. In the latter situation, the
chicks could remain immobile under the water with only
their bills extending above water for periods as long as
25 min (see also Bosque & Herrera 1999). The male sig-
nalled when the danger had passed by giving a soft,
repetitive, clucking (comfort) call to which the chicks
responded by reappearing and following him once again.
Chicks remained in the care of the male until they were
capable of ?ight. This occurred at 8?10 weeks of age. We
used the attainment of ?ight as our criterion for de?ning
chick independence (the end of parental care). Young
birds generally left their natal territories within 1?2 weeks
of ?ying. Occasionally, juveniles were tolerated on the
natal territory for up to 8 months.
The time required for a male to successfully rear a clutch
from receipt of eggs to independence of the brood was
Table 1. Weight loss of male wattled jacanas during incu
Male ID Date
Day of
incubation Weight (g
496 5 April 2002 6 92
24 April 1992 25 90
630 18 February 1993 2 95
10 March 1993 22 94
669 13 March 1993 2 89
4 April 1993 24 83
670 21 March 1993 5 100
9 April 1993 23 90
704 27 April 1993 4 96
18 May 1993 25 85
491y 13 May 1991 NA 80
10 June 1991 NA 86
NAZ nonapplicable.
*Percentage of weight loss was calculated over the full 25
yBird 491 was a nonbreeding male and served as a contr
ANIMAL BEHAVIOUR, 68, 4perform ?double duty?, alternating their activities between
incubating a clutch of eggs and tending and protectinga mobile brood of chicks. In four of these instances, the
female provided direct care to the chicks and thereby freed
the male to spend more time incubating the clutch. These
females foraged with the chicks, brooded them during the
morning cold and afternoon rain showers, and alerted
them to impending threats (e.g. the approach of purple
gallinules and common gallinules).
We also observed ?ve cases where males were predated
while caring for broods of dependent young. In each of
the three cases where the chicks did not disappear along
with the male, the female mate took over full care of the
chicks following the male?s death and continued to pro-
vide active care until the young either became indepen-
dent or were lost to predation. One female provided such
care for a full month. During this time, she showed the
entire suite of species-speci?c parental care behaviours,
including leading, brooding, defending and giving of the
ation
Interval between
weighings (days)
Actual weight
change ( g) %Weight loss*
19 
2 2.9
20 
1 1.3
22 
6 7.7
18 
10 13.9
21 
11 13.6
28 C6
-day incubation period.
l.analysis was weak, and this result should be considered as
suggestive rather than de?nitive.
female played no role and the male, by himself, de?ected
51from another female (Emlen et al. 1989). Male jacanas
constitute no known threat to eggs and young, but are
a threat to the territory ownership of the resident male.
the threat away from the nest, led the chicks away from
the threat, or distracted the potential predator while the
young ran to safety. The results were similar for intruding
Table 3. Differences between male and female responses to various predation threats. Data are from 45 h of focal
sampling on 15 nesting pairs of jacanas
Type of threat
Heterospecific predators* Female jacana intruders Male jacana intruders
Sex of first responder
Male 44 (71%) 149 (68%) 548 (86%)
Female 18 (29%) 69 (32%) 90 (14%)
Total 62 218 638
Did male give rally call?
Yes 24 (55%) 103 (69%) 31 (6%)
No 20 (45%) 46 (31%) 517 (94%)
Total 44 149 548
Did the female join in predator defence?
Yes 21 (88%) 74 (72%) 16 (52%)
No 3 (12%) 29 (28%) 15 (48%)
Total 24 103 31The Role of Both Sexes in Predator Defence
Predation threats to jacana eggs or chicks were frequent.
During 45 h of focal observations on 15 nesting pairs, we
recorded 280 territorial intrusions by potential predators,
or an average of just more than six intrusions per hour.
Both male and female breeders played major roles in
defending the nest and chicks. We partitioned the focal
data according to whether the threat was from a hetero-
speci?c predator ( purple and common gallinules), a con-
speci?c predator (intruding female) or neither (male
jacana). Female jacanas were considered potential preda-
tors because they routinely destroy both eggs and small
young when they successfully intrude and take over a male
292 602 8/1995 3
179 602 7/1994 1
20 795 8/1996 1*Heterospecific predators included purple gallinules, Po
chloropus.They were included for purposes of comparison. We asked
(1) which parent ?rst responded to the threat, (2) whether
the male called to the female for assistance in the defence,
and (3) whether the female, if called upon, joined the
male in attacking the source of the threat.
Male jacanas were signi?cantly more likely than females
to be the ?rst to respond to an intruder on the territory
(Table 3). This was true regardless of the type of intruder
(heterospeci?c intruder: c12Z 10.9; female jacana in-
truder: c12Z 29.4; male jacana intruder: c12Z 328.8; all
P! 0.0001). More than two-thirds (44/62) of potential
heterospeci?c predators were ?rst detected by the breed-
ing male. In roughly one-half (20/44) of these cases the
X No
X Unknown
X NoTable 2. Behaviour of female wattled jacanas when the male was incubating a clutch of eggs while still caring for
dependent young
Female ID Male ID
Date
(month/year)
Number
of young
Age of young (days) Female
parental
care21?25 26?30 31?35 36?40 41?45 46?50
259 546 4/1994 3 X Yes
136 475 4/1990 1 X Unknown
148 467 4/1990 1 X Unknown
22 814 8/1995 3 X Yes
284 602 9/1996 3 X No
20 795 10/1996 3 X Yes
180 496 5/1992 1 X Unknown
2606 735 10/1996 3 X Unknown
247 665 5/1994 2 X No
393 667 5/1994 4 X No
284 602 9/1996 3 X No
919 879 9/1996 4 X No
138 535 2/1991 3 X Yes
284 841 7/1995 2 X No
EMLEN & WREGE: PARENTAL CARE IN WATTLED JACANAS 8rphyrula martinica, and common gallinules, Gallinula
852female jacanas. Again, two-thirds (149/218) of the ?rst
sightings were by males, and in roughly one-third (46/
149) of these instances, the male quietly led the chicks
away from the potential danger.
The female was the ?rst responder to the remaining
predator intrusions. When females responded to conspe-
ci?c females, they almost always attacked the intruder.
When they responded to gallinules, they more often
blocked the approach to, or de?ected the intruder from,
the location of the nest or young. This behaviour by the
female provided the male with time to lead the chicks to
safety.
Often, when the male was the ?rst responder, he alerted
the female to the presence of the intruder and, together,
they confronted the source of the threat (95 of 193 in-
stances). This occurred primarily when a purple or com-
mon gallinule or a trespassing female jacana came very
close to the nest or the location where the chicks were
hiding. In these cases, the male gave a shrill, high-pitched
rally call to which the female typically responded by ?ying
immediately to the male?s location. Responses to intrud-
ing female jacanas did not differ from those to hetero-
speci?c predators. In both cases, males were equally likely
to call to the female (Wald?s c12Z 1.90, NS) and, if alerted,
the female was equally likely to join the male in defending
against both intruder types (Wald?s c12Z 2.65, NS).
The form of the joint defence depended on the type of
predator. Against gallinules, both male and female actively
attacked the threatening individual, running at it with
wings outstretched and striking it repeatedly from the air
with their feet. In contrast, only the female actively at-
tacked an intruding female jacana. Males maintained a
close presence, but restricted their activities to running at
the female with outstretched wings or facing the intruder
in a crouched position with wings outstretched and head
and neck feathers erected.
Conspeci?c male intruders were treated very differently.
Themale resident was the sole responder in 81% (517/638)
of such instances (Table 3). Separate orthogonal contrasts
(see Methods) indicated, in comparison to other types of
intruders, that (1) resident males were signi?cantly more
likely to be ?rst responders to intrusions by male jacanas
(Wald?s c12Z 196.7, P! 0.0001), and (2) males were
signi?cantly less likely to alert their females to such
intrusions (Wald?s c12Z 31.3, P! 0.0001). In only 2.5%
(16/638) of incursions by trespassing males did the
breeding pair show a joint defence in chasing the male
intruder from the territory. In contrast, joint defence
occurred in 34% of intrusions by gallinules or trespassing
female jacanas (21 of 62, and 74 of 218 cases, respectively).
Most defensive attacks against potential predators were
successful. Of the 280 observed predator detections, in-
cluding the 95 more severe cases where joint defence oc-
curred, only four resulted in actual loss of eggs or chick(s).
DISCUSSION
The Role of the Male Parent
ANIMAL BEHAVIOUR, 68, 4Male wattled jacanas performed the overwhelming
majority of parental care. Incubation lasted 25 days andyoung became independent (i.e. they could ?y) at 50?60
days of age. A successful breeding cycle thus occupied
nearly 3 months of a male?s time.
Males were the sole incubators of the eggs. In the course
of 57 h of scan sampling, and in more than 2500 nest
checks, no female was ever observed to incubate eggs. On
nine occasions when a male disappeared (presumably
predated) while incubating eggs, in no instance did the
remaining female incubate the untended eggs. We suggest
that incubation is no longer present in the females?
physiological or behavioural repertoire. Incubation has
not been con?rmed in females of any of the sexually
dimorphic jacana species (pheasant-tailed jacana, Hydro-
phasianus chirurgus: Hoffmann 1949; northern jacana,
J. spinosa: Jenni & Collier 1972; Jenni & Betts 1978;
African jacana, Actophilornis africanus: Tarboton 1993;
bronze-winged jacana, Metopidius indicus: Butchart 2000;
comb-crested jacana, Irediparra gallinacea: Mace 2000).
The cost of incubation to the male is presumably low.
The high solar radiation that typi?es the nesting habitat
of jacanas relieves the male of much of the incubation
load (Jenni 1974; Tarboton 1995; Butchart 2000; Mace
2000). Scan samples indicated that males tending clutches
spent only 41% of the daylight hours at the nest site,
either incubating eggs or shading them from the sun.
Osborne & Bourne (1977), in a study of J. jacana in
Suriname, reported a value of approximately 60%. Simi-
larly low estimates of incubation constancy have been
found in the other sex-role-reversed jacana species (44%
for northern jacana, Betts & Jenni 1991; 47% for pheas-
ant-tailed jacana, Hoffman 1949; 53% for African jacana,
Tarboton 1993; 56% for bronze-winged jacana, Butchart
2000; 54% for comb-crested jacana, Mace 2000).
Incubation in wattled jacanas was not without some
costs, however. Based on a sample of ?ve males that were
recaptured both early and late in incubation, we estimated
that males lose approximately 8% of their body weight
during their 25 days of incubating. This is consistent
with detailed time budget studies of wattled jacanas by
Osborne & Bourne (1977) who reported that incubating
males spend approximately 50% less time foraging than
do nonbreeding males. Similar reductions in time spent
foraging by incubating males have been reported for
northern jacana (Betts & Jenni 1991) and comb-crested
jacana (Mace 2000).
Males were also the primary caretakers of the chicks. Of
242 broods monitored during this study, females were
only observed to provide direct chick care to seven (3%).
Chicks thus spent the overwhelming majority of their
time under the direct care of the male parent. During the
?rst 3?4 weeks after hatching, the chicks followed the
male closely. It was the male who determined the chicks?
foraging and resting locations. He brooded them under his
wings in the cold of early morning and during Panama?s
frequent afternoon rainstorms. He also provided nearly
constant vigilance, sending the chicks scurrying for cover
at the threat of danger, and calling them back to him
when the threat had passed.
As the chicks became older, the male continued to
perform all of these parental behaviours, and the male and
chicks still moved as a unit, but the distance the chicks
853foraged from the male increased. Similar patterns of male
care have been reported for all of the sexually dimorphic
jacanas (Hoffmann 1949; Jenni & Collier 1972; Tarboton
1992; Butchart 2000; Mace 2000). Male parental care
in wattled jacanas continued until the chicks were able
to ?y, which occurred at approximately 50?60 days of
age. Young were tolerated by the male parent after this
age, but they typically left the territory and joined the
mobile population of ?oaters a few weeks after attaining
?ight.
We assumed that the cost of chick care in wattled
jacanas is low because the young are precocial and able
to forage for themselves within hours of hatching. There is
no provisioning of young jacanas in the sense of adults
bringing food to the young. Rather, the male leads the
chicks to different areas of the territory where they feed
themselves. Time budget studies of northern jacanas
(Betts & Jenni 1991) show that, although males do reduce
their time spent foraging when caring for chicks, the
reduction is minor because they are able to forage for
themselves while simultaneously tending the young.
Chick care in wattled jacanas is nondepreciable
(Wittenberger 1979; Walters 1982) in the sense that the
cost of providing care for multiple chicks is not signi?-
cantly greater than the cost of providing care for a single
chick. This is true whether leading chicks to favoured
foraging sites, brooding them in inclement weather,
communicating danger or comfort calls, or even driving
off a potential predator. Thus, whereas ef?cient chick care
by a male jacana is incompatible with simultaneous
incubation of a second clutch, it is not incompatible with
simultaneous foraging and vigilance for himself.
The Role of the Female Parent
In marked contrast to the incubation behaviour and
nearly continuous chick care behaviour provided by the
male, female wattled jacanas spent minimal time in par-
ental activities. Females never incubated eggs and spent,
on average, only minutes per day in chick-related behav-
iours. The importance of this female care, however, was
much greater than its minimal quantity might suggest.
As such, female parental care represents a previously
underappreciated aspect of jacana breeding biology.
There were two ways in which female wattled jacanas
directly contributed to clutch and chick survival. First, on
very rare instances, females displayed the same suite of
chick care behaviours as those regularly shown by males.
And second, females routinely provided essential predator
protection.
Active female chick care only occurred under two highly
speci?c but predictable conditions: (1) when a male mate
was tending chicks while simultaneously incubating a
clutch of eggs, and (2) when a parental male disappeared
permanently, leaving a brood untended.
Although extremely rare in their occurrence, these ob-
servations con?rm that all of the basic chick care behav-
iours of leading, brooding and providing vigilance are
present in the behavioural repertoire of female wattled
jacanas. However, females only displayed active chick carewhen their mates were temporarily or permanently un-
available, and even then, only under circumstances when
the chicks were likely to die if no care was provided at all.
Thus, females were more likely to care for chicks of a
deceased male than for a male doing ?double duty?, and
were signi?cantly more likely to care for younger (pre-
sumably more vulnerable) chicks than for older ones
(Table 2). We hypothesize that female parental care in
wattled jacanas is conditionally expressed only in circum-
stances when the bene?t of providing such care is
exceptionally high.
Among jacanas, active chick care by females is not
unique to wattled jacanas. Rare observations of brooding
by females have been reported for northern jacanas (Jenni
& Betts 1978; Betts & Jenni 1991) and pheasant-tailed
jacanas (Hoffmann 1949), but not for African jacanas
(Tarboton 1992) or comb-crested jacanas (Mace 2000,
personal communication). In their study of northern
jacanas, Jenni & Betts (1978; Betts & Jenni 1991) noted
that females brooded and accompanied chicks only after
times of unusually heavy rain when, they speculated, food
availability was low. They also reported one instance where
a female tended chicks of an incubating male doing
?double duty?. These observations are consistent with
active female care being both rare and conditional upon
the male?s inability to provide suf?cient parental care
alone.
Mutual Dependency of the Pair in
Predator Defence
Female wattled jacanas played a regular and essential
role in defence against nest and chick predators. This was
only possible, however, because of the territorial and
parental behaviours of the male. By being territorial, the
male was vigilant to intruders, and by tending the nest
and young, he was vigilant to predation threats. Not
surprisingly, males were the ?rst to detect and respond to
greater than two-thirds of predator intrusions (by purple
gallinules, common gallinules and conspeci?c females). In
roughly one half of such cases, the male, unaided by the
female, was able to de?ect the predator from the nest
vicinity or to lead the chicks away to safety.
However, male jacanas depended on their female mates
to back them up in cases where the predation threat
escalated beyond their ability to deal with it alone. Under
such circumstances, the male alerted his female to the
danger by giving a shrill, high-pitched vocalization that
usually brought her rapidly to his side. The pair, together,
then confronted the predator.
Such joint defence was the response to one-third of all
intrusions by potential predators ( gallinules or female
conspeci?cs). Because of her much larger body size (wat-
tled jacana females and males, on average, weigh 129 g
and 88 g, respectively; Emlen & Wrege 2004), the female
was presumably more effective than the male in deterring
predators. Stephens (1984) con?rmed this assumption for
the northern jacana. He reported that purple gallinules
EMLEN & WREGE: PARENTAL CARE IN WATTLED JACANASwere driven furthest when attacked by pairs, less far by
females alone, and least far by attacking lone males.
854Wattled jacanas thus have a two-tiered system of
predator defence. The male typically is the ?rst responder
but, if needed, he can count on the female for backup
support. Similar behaviours, whereby males recruit their
females to jointly attack predators, have been reported for
all of the sex-role-reversed jacanas (Hoffmann 1949; Jenni
& Collier 1972; Jenni & Betts 1978; Stephens 1984;
Tarboton 1992; Butchart et al. 1999; Mace 2000).
Responses to intruding male jacanas were very different.
Females were signi?cantly less likely both to be the ?rst to
respond to, and to join in defending against, a male tres-
passer. This is consistent with the fact that male jacanas do
not constitute a known threat to the female or to the eggs
or chicks, but only to the territorial ownership status of
the resident male.
Several features of the jacanas? habitat facilitated the
ability of females to provide backup support in predator
defence even though they frequently foraged far from the
vicinity of the nest or young. First, the ?oating mats of
vegetation on which jacana territories were established
seldom exceeded 1 m in height, thus allowing the birds
easy visibility over large areas. Second,male territories were
small, averaging only 40 m in diameter (unpublished data).
Third, the expansiveness of the vegetation meant that
male territories often abutted one another in a checker-
board matrix. Both the high density and the two-dimen-
sional aspect of male territories increased the effectiveness
of predator detection and defence. A jacana typically could
see directly across the entirety of its own territory and those
of several immediate neighbours. Thus, when a male
alerted his female to a serious threat, the female often
had line-of-sight access to information that enabled her to
locate, assess and rapidly respond to that threat.
Jacana sex roles in predator defence were thus mutually
interdependent. The female relied on her mate to alert her
to the presence of most intrusion threats, and the male
relied on his female to back him up when predation
threats escalated. This mutual dependency made it possi-
ble for the female to play a critically important role in
increasing the survival probability of her eggs and chicks
without having to spend more than minutes per day
engaging in such behaviours.
This two-tiered system of predator defence was highly
effective. Of the 280 intrusions by potential predators,
only four resulted in the loss of eggs or chicks. Yet, par-
adoxically, despite this high success in predator deterrence,
rates of nest loss were high in our study population. Fully
one-half of all clutches were lost prior to hatching and two-
thirds of hatchlings disappeared prior to reaching inde-
pendence (unpublished data). The most common cause of
such loss was predation.
We suggest that the reason for the high predation losses
was not an inability of parental pairs to successfully defend
against those predators that were detected early enough for
deterrence to be effective. Rather, we suspect that most
losses occurred when lapses in vigilance made it possible
for predators to gain close access to the nest or chicks
without being detected. Stephens (1984) similarly sug-
gested that successful predation by purple gallinules only
ANIMAL BEHAVIOUR, 68, 4occurs on northern jacanas when the eggs or chicks are
temporarily unattended by the male parent.Parental care in wattled jacanas is thus typi?ed by a
marked division of labour, but with two unusual twists.
First, although females no longer incubate eggs, they
retain the ability to perform all aspects of chick care. They
display such chick care, however, only under exception-
ally rare circumstances when their mates are physically
unable to perform these behaviours themselves. Second,
although almost completely emancipated from all direct
care of eggs or young, females nevertheless play a critically
important role in predator defence. By relying on the male
to provide vigilance for the eggs and young, yet respond-
ing when alerted to threats by the male, female wattled
jacanas reap the bene?ts of a system of male uniparental
care that can be facultatively transformed into one of
biparental defence when required.
Acknowledgments
We thank the Instituto Nacional de Recursos Naturales
Renovables of Panama for permission to work in the
Soberania National Park, and the Smithsonian Tropical
Research Institute for facilitating the research. Numerous
persons assisted with the ?eld work, and we thank
Orianna Batista, Sandra Black, Stuart Butchart, Manuel
Demaria, Natalia Demong, Ceinwen Edwards, Richard
Eibach, Suzanne Kelley, Brian Kurzel, Jessica Lewis, Tom
Mazorlig, Ramiel Papish, Iris Saxer, Darius Sarshar, Manuel
Velez and Michael Webster for their efforts. The manu-
script bene?ted from comments by Natalia Demong and
Stuart Butchart. This work was conducted under the
auspices of the Smithsonian Tropical Research Institute
and was funded by the National Science Foundation, the
Smithsonian Tropical Research Institute and a Fulbright
Scholar Fellowship to S.T.E.
References
Altmann, J. 1974. Observational study of behaviour: sampling
methods. Behaviour, 49, 227?267.
Andersson, M. 1994. Sexual Selection. Princeton, New Jersey:
Princeton University Press.
Betts, B. J. & Jenni, D. A. 1991. Time budgets and the adaptiveness
of polyandry in northern jacanas. Wilson Bulletin, 103, 578?597.
Bosque, C. & Herrera, E. A. 1999. ?Snorkling? by the chicks of the
wattled jacana. Wilson Bulletin, 111, 262?265.
Butchart, S. H. M. 2000. Population structure and breeding system
of the sex-role reversed, polyandrous bronze-winged jacana,
Metopidius indicus. Ibis, 142, 93?102.
Butchart, S. H. M., Seddon, N. & Ekstrom, J. M. M. 1999. Yelling
for sex: harem males compete for female access in bronze-winged
jacanas. Animal Behaviour, 57, 637?646.
Clutton-Brock, T. H. 1991. The Evolution of Parental Care. Princeton,
New Jersey: Princeton University Press.
Clutton-Brock, T. H. & Vincent, A. C. J. 1991. Potential re-
productive rates and the operation of sexual selection. Quarterly
Review of Biology, 67, 437?456.
Emlen, S. T. & Wrege, P. H. 2004. Size dimorphism, intrasexual
competition and sexual selection in wattled jacanas (Jacana
jacana), a sex-role reversed shorebird in Panama. Auk, 121, 391?
403.Emlen, S. T., Demong, N. J. & Emlen, D. J. 1989. Experimental
induction of infanticide in female wattled jacanas. Auk, 106, 1?7.
Hoffmann, A. 1949. Uber die Brutpflege des polyandreschen
Wasserfans, Hydrophasianus chirurgus. Zoologische Jahrbucher-
Abteilung fu?r Systematik Okologie, 78, 368?403.
Jenni, D. A. 1974. Evolution of polyandry in birds. American
Zoologist, 14, 129?144.
Jenni, D. A. 1996. Family Jacanidae ( jacanas). In: Handbook of the
Birds of the World. Vol. 3 (Ed. by J. del Hoyo, A. Elliott & J. Saratal),
pp. 276?291. Barcelona: Lynx Edicions.
Jenni, D. A. & Betts, B. J. 1978. Sex differences in nest construction,
incubation, and parental behaviour in the polyandrous American
jacana (Jacana spinosa). Animal Behaviour, 26, 207?218.
Jenni, D. A. & Collier, G. 1972. Polyandry in the American jacana
(Jacana spinosa). Auk, 89, 743?765.
Lack, D. 1968. Ecological Adaptations for Breeding in Birds. London:
Methuen.
Mace, T. R. 2000. Time budget and pair-bond dynamics in the
comb-crested jacana, Irediparra gallinacea: a test of hypotheses.
Emu, 100, 31?41.
Oring, L. W. 1982. Avian mating systems. In: Avian Biology. Vol. 6
(Ed. by D. S. Farner, J. R. King & K. C. Parkes), pp. 1?92. New York:
Academic Press.
Oring, L. W. 1985. Avian polyandry. In: Current Ornithology. Vol. 3
(Ed. by R. F. Johnston), pp. 309?351. New York: Plenum.
Osborne, D. R. & Bourne, G. R. 1977. Breeding behavior and food
habits of the wattled jacana. Condor, 79, 98?105.
Ridley, M. W. 1978. Paternal care. Animal Behaviour, 26, 904?932.
Ross, H. A. 1979. Multiple clutches and shorebird egg and body
weight. American Naturalist, 113, 618?622.
Stephens, M. L. 1984. Interspecific aggressive behavior of
the polyandrous northern jacana (Jacana spinosa). Auk, 101,
508?518.
Tarboton, W. R. 1992. Aspects of the breeding biology of the
African jacana. Ostrich, 63, 141?157.
Tarboton, W. R. 1993. Incubation behaviour of the African jacana.
South African Journal of Zoology, 28, 32?39.
Tarboton, W. R. 1995. Polyandry in the African jacana: the roles
of male dominance and rate of clutch loss. Ostrich, 66,
49?60.
Trivers, R. L. 1972. Parental investment and sexual selection. In:
Sexual Selection and the Descent of Man 1871?1971 (Ed. by
B. Campbell), pp. 136?179. Chicago: Aldine.
Walters, J. R. 1982. Parental behavior in lapwings (Charadriidae)
and its relationship with clutch sizes and mating systems.
Evolution, 36, 1030?1040.
Williams, G. C. 1966. Adaptation and Natural Selection. Princeton,
New Jersey: Princeton University Press.
Wittenberger, J. F. 1979. The evolution of mating systems in birds
and mammals. In: Handbook of Behavioral Neurobiology. Vol. 3
Social Behavior and Communication (Ed by P. Maler & J.
Vendenberg), pp. 271?349. New York: Plenum.
EMLEN & WREGE: PARENTAL CARE IN WATTLED JACANAS 855