Behavioral Ecology Vol. 12 No. 6: 716-725
© 2001 International Society for Behavioral Ecology
Nestling aggression in broods of a siblicidal kingfisher, the laughing kookaburra
Botany and Zoology, Australian National University, Canberra, ACT 0200, Australia
Address correspondence to S. Legge. E-mail: sarah.legge{at}anu.edu.au .
Received 3 August 2000; revised 14 February 2001; accepted 19 February 2001.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Third-hatched nestling in broods of the laughing kookaburra (Dacelo novaeguineae) are often killed by aggressive attacks from their older siblings within days of hatching. By installing surveillance cameras inside nest hollows, we examined nestling aggression over the "siblicidal" period, in particular to identify whether parental behavior and competitive disparities between nestlings affected aggression, and hence the likelihood of siblicide. Aggression decreased as nestlings aged and dominance hierarchies became established. The first-hatched nestling was the most aggressive. Fighting between the first-hatched nestling and its closest rival (second-hatched nestling) increased when the hatch interval between them was short, when the size difference between them at hatching was small, and when the second nestling was female. Female nestlings are faster-growing than males, so young sisters may be an incipient threat requiring preemptive action by older siblings. When the second-hatched nestling was female, the first-hatched also attacked the third-hatched nestling more frequently. Thus the third-hatched nestling seems to experience some of the "overflow" of aggression occurring between its two older siblings. Nestlings in siblicidal broods were not fed less compared to nonsiblicidal broods; this is unsurprising because siblicide occurs when feeding rates are comparatively low. However, siblicidal nestlings were brooded less, and in shorter bouts, which allowed them more time to fight.
Key words: brood reduction, cooperative breeding, hatching asynchrony, kingfisher, siblicide.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Competition between siblings for resources is widespread in the broods of altricial birds. Differences in competitive abilities between nestlings at hatching catalyze the development of dominance hierarchies within broods, which concentrate resource deprivation onto the lowest-ranking member. Nestlings with a large competitive advantage at hatching can use this head-start to further consolidate their dominance, makint it hard for low-ranking nestlings to "improve" their position. In a small proportion of species, dominant nestlings use aggression to intimidate and even kill their siblings (siblicide).
The level of aggression required to form and then maintain the hierarchy
may depend on the asymmetries in the competitive abilities of nestlings.
Fighting should escalate when nestlings are closely matched and dominance is
therefore unclear (Enquist and Leimar,
1983
; Maynard-Smith and
Parker, 1976). The extent of the asymmetry will depend largely on
the age and size differences between the nestlings. In other words, hatching
asynchrony (which determines relative age) and egg size (which determines
relative size at hatching) may be important. In addition, in size dimorphic
birds the hatching sequence of sexes may have an effect.
Empirical studies have confirmed that aggression in broods does increase
when nestlings hatch more synchronously
(Cook et al., 2000;
Forbes, 1991;
Fujioka, 1985;
Machmer and Ydenberg, 1998;
Mock and Ploger, 1987;
Osorno and Drummond, 1995).
There are no data regarding the role of egg size in nestling aggression, but
studies of whether egg size affects fledging success generally find that any
advantages of hatchling size are overwhelmed by age differences
(Magrath, 1992). The effect of
sex sequence on nestling aggression has received scant empirical attention
(but see Golla et al., 1999
for a mammalian example). Aggressive dominance between blue-footed booby
nestlings was determined by their hatching order and early social experience
(Drummond and Osorno, 1992).
Nestling growth and survival was not affected by the order of sexes (although
females grow to be 27% heavier than males), suggesting aggression is also
unaffected (Drummond et al.,
1991). However, sex sequence may affect nest success in bald
eagles, suggesting that aggressive interactions between nestlings could be
influenced by their sexes (Bortolotti,
1986; but see Drummond et al.,
1991).
Aggression between broodmates could also vary because of influences
external to the brood. The most obvious example is the amount of food
delivered to the nest. Hunger could act as a proximate cue stimulating
nestlings to increase aggression, in order to monopolize resources and perhaps
even kill a junior sibling. Evidence supporting this hypothesis is mixed. For
example, blue-footed booby and osprey nestlings fought more when hungry
(Drummond and Garcia Chavelas,
1989; Machmer and Ydenberg,
1998), but great egrets, blue herons and swallow-tailed kites did
not (Gerhardt et al., 1997;
Mock et al., 1987). However,
nestlings may only respond facultatively to the current food supply if it is a
reliable predictor of future food availability
(Forbes and Ydenberg, 1992).
Another external influence which could affect aggression is the brooding
behavior of parents. Nestlings that are left uncovered for longer have more
time to fight. Brooding patterns could be an indirect indicator of food
availability if parents that leave the nest unattended do so because they are
having trouble finding food (Newton,
1979).
Patterns of aggression are relatively straightforward in broods of two
nestlings, because they contain only one dyad. For example, the oldest chick
is always the main aggressor (e.g., Braun
and Hunt, 1983; Drummond et
al., 1991; Gargett,
1978) However, in broods of three each nestling may be affected by
interactions between the other two nestlings (e.g.,
Parker et al., 1989). Very few
studies have examined the distribution of nestling aggression within broods of
three or more. In cattle egrets, fighting is concentrated at the bottom end of
the hierarchy. The costs of losing fights are potentially higher for these
nestlings, because losers incur the penalty of becoming the "designated
victim" of siblicide (Fujioka,
1985; Mock and Lamey,
1991; Ploger and Mock,
1986). In contrast, fighting is more frequent between the oldest
two siblings in three-nestling broods of brown pelicans
(Ploger, 1997). Often only one
nestling fledges from broods in this population of brown pelicans, so the
critical fight may be over who is ranked first, rather than who avoids being
ranked last.
In this study, we examine aggression between the three nestlings in broods
of the laughing kookaburra (Dacelo novaeguineae), a large kingfisher
with a complex breeding biology. As adults they live in cooperative family
groups, but intense aggression between the three broodmates causes the death
of the youngest nestling in one-third of nests within a few days of hatching
(Legge, 2000c). Legge
(2000c) showed that the
probability of siblicide may be affected both by parental behavior and
competitive asymmetries between nestlings. Siblicide was more likely when the
breeding pair did not have male helpers—groups without helpers may brood
less or bring less food to the nest, triggering more aggression between
nestlings. In addition, the third hatchling was more likely to die when it was
much smaller than its siblings (perhaps because it was easier to kill), when
the hatch interval between the first and second nestlings was short (possibly
because brood aggression escalates when dominance between these nestlings is
unclear), and when the sexes of the first and second nestlings were male and
female respectively. Younger sisters could pose a threat to older brothers
because they grow more quickly, and are already 8% larger 10 days after
hatching. Kookaburras appear to have strong control over the sex of their
offspring, and they may manipulate the hatching sequences of sexes to dampen
or promote aggression in their broods
(Legge, 2000c;
Legge et al., 2001).
The four variables that were associated with siblicide were also correlated
with each other, making it unclear which were directly responsible for
precipitating siblicide, and which were simply noncausal correlates of another
factor common to siblicidal nests: breeding females in poor condition. Poor
females are less likely to have male helpers, more likely to lay a small third
egg (because they run out of reserves by the end of egg-laying), and also more
likely to delay incubation after clutch initiation because of nutritional
constraints during the laying period (thus causing the first and second
nestlings to hatch more synchronously;
Legge, 2000c).
In the study reported here we installed tiny surveillance cameras inside the nesting hollows of this kookaburra population to examine whether parental behavior and competitive asymmetries affected aggression between the three nestlings in kookaburra broods. In this way we hoped to determine which factors directly led to siblicide in kookaburra nests. We were particularly interested in whether the sex of nestlings affected aggression between them, as these data are scarce from other species.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Study site and species
Detailed observations of nestling aggression were made between October and December 1997. We used a subset of 11 kookaburra groups from a marked population of about 35 groups that had been monitored since 1994 in the Canberra Nature Park, a eucalypt woodland reserve in south-east Australia. We knew the size, composition and recent reproductive history of every group in this 20 km2 area. Kookaburras live in cooperative groups of two to eight birds, comprising a socially dominant pair and their offspring-helpers of both sex. The mating system is monogamous (Legge and Cockburn, 2000
).
Helpers are involved with territory defense, and all aspects of reproduction:
incubation, brooding, feeding, and guarding nestlings and fledglings
(Legge, 2000b;
Parry, 1973).
Kookaburra nestlings hatch blind and naked, but are aggressive to siblings
within a couple of hours of hatching. They have a downward-pointing hook at
the end of their upper beak that grows out and disappears by the time the
nestlings fledge. This is a rare example of a morphological specialization for
sibling rivalry (Legge,
2000c). Nestlings use this beak hook in lunging pecks and bites to
the backs and heads of their siblings that result in scratches, bruises, and
skin lesions. They also attempt to grasp a sibling firmly around the neck, so
that the aggressor's beak hook presses into the base of the victim's skull.
With this hold, the aggressor shakes vigorously. Nestlings did not appear to
use aggressive or threat displays to intimidate siblings. Nor did we observe
submissive displays; instead victims attempted to avoid aggression by moving
away from their attacker. In a broader survey of this population of
kookaburras, the youngest nestling in broods of three was killed in one third
of nests. Most victims (75%) died before they were 4 days old, and all victims
died before they were 8 days old. Victims showed obvious wounding before their
disappearance, and their heavily bruised bodies were sometimes found in nests
(Legge, 2000c).
Kookaburras prey mainly on invertebrates and small lizards. When feeding nestlings, adults delivered single prey items directly to one nestling in the brood, usually the nestling that was closest to the hollow entrance. Food dropped onto the nest floor was sometimes picked up by adults and re-offered to nestlings, but nestlings never retrieved dropped food themselves.
General field methods
We found nests during incubation, and estimated the hatching date by
candling eggs. We visited nests repeatedly over the hatching period to assign
sequential hatch ranks A, B, and C to each hatchling, and to estimate hatch
intervals between hatchlings (denoted ABinterval, BCinterval, ACinterval) to
within 2 h. Hatchlings could be aged reliably from two features. Immediately
after the nestling emerges from the egg the hatching muscle at the back of its
neck is swollen, and a blistery appearance is evident under their skin. The
neck swelling subsides in 2 h, and the blistery effect disappears over the
next 24 h. Hatchlings were made individually identifiable by tying embroidery
thread around their tibias, or daubing nontoxic paint on their beaks, tibias
or wing chords.
In this population, the ACinterval and BCinterval were both symmetrically
distributed around 36 and 24 h respectively. However, the ABinterval was
distributed tri-modally, with peaks at 4, 12, and 24 h, corresponding to three
likely behavior patterns at the onset of incubation. Breeding females lay eggs
in the morning. They may begin incubation as soon as the first egg is laid (A
and B will hatch asynchronously, about 24 h apart), the night after the first
egg is laid (A and B hatch about 12 h apart), or only after the second egg has
been laid (A and B will hatch relatively synchronously). Previous work on this
population has shown that when A and B hatched close together, the C nestling
was almost always killed, but when the hatch interval was 6 h or more,
siblicide was rare (Legge,
2000c). In this study, we therefore categorized the ABinterval as
"short" (less than 6 h) or "long" (6 h or more).
We took a small blood sample (10-20 µl) within 48 h of hatching and used
a molecular method to find the sex of nestlings
(Griffiths et al., 1998). We
used egg volume as a surrogate measure for hatchling size, since the weight
and skeletal size of new hatchlings is tightly correlated with the volume of
the egg from which they emerged (Legge,
2000c). We measured eggs to the nearest 0.1 mm, and calculated
volume in mm3 as
0.51*Length*Width2
(Hoyt, 1979). Egg volume
varies considerably between clutches, so we calculated a relative egg volume
difference between pairs of eggs in the same clutch [e.g., ABvol = (volume of
A egg—volume of B egg)/mean volume of A and B eggs].
We defined broods as "siblicidal" if the youngest nestling died within 8 days of hatching after being wounded by sibling attacks. If the youngest nestling survived this critical post-hatching period the brood was termed "nonsiblicidal." After the 10 day observation period (see below), we visited nests at weekly intervals to census nestlings, and finally to band them when the oldest nestling was 32 days old. Fledging takes place 35-40 days after hatching.
Installing cameras
Kookaburras use nest hollows which are up to 12 m high and 2 m deep, making
it impossible to watch nestling behavior from the ground. We installed
surveillance cameras into hollows (miniature CCD cameras with infra-red LEDs;
2 x model CA-28A [EIA] and 4 x PCB module series; circuit boards
measured 32 x 32 x 27 mm). Hollows are very different in shape and
size, so each camera was attached to a ball-and-socket joint to allow us to
adjust the orientation of cameras inside the hollow. The ball-and-socket joint
was threaded, and could be screwed onto a double-ended screw fixed to the
hollow wall. We installed the cameras into nest hollows before the eggs
hatched. The procedure took about 1 h. During this time we removed the eggs
from the hollow, keeping them warm (36°C) using a hot water bottle and
thermometer to monitor temperature. No egg breakages or hatching failure
resulted from this procedure. We used 15-17 m of coaxial cable to connect the
surveillance cameras to recorders at the base of the tree (one Sony video
camera recorder CCD-V700E and two Sony recorder/monitors GV-S50E). The video
recorders were powered by 12V rechargeable sealed lead-acid batteries (Yuasa
NP7-12). Three-hour recordings were made on 90-min standard quality video-8
tape (Sony P5-90MP2) by using the long-play setting. At each nest, the cameras
remained in place inside the hollow (thus minimizing disturbance), but the
recorders could be simply unplugged and moved to another site as required.
Sampling
We began video recording at focal nests on the day the last egg hatched. At
each nest, we recorded once per day (for 3 h) for 10 consecutive days, since
in a larger sample siblicide always occurred within this period
(Legge, 2000c). Recordings
were made alternately in the mornings (between 0700-1130) and the afternoons
(between 1330-1830). This protocol allowed us to record at six nests per
day—three in the morning and three in the afternoon. We chose focal
nests carefully to obtain a balanced sample of nests from unassisted pairs
versus groups with at least one male helper, since siblicide is known to occur
more frequently in groups lacking male helpers. We avoided sampling clutches
of two eggs, focusing instead on the modal clutch size of three. Overall, we
recorded the behavior of 33 nestlings from 11 broods
(Table 1). We would have
preferred to balance the sample with regard to the sex of nestlings, but could
not do so because nestlings were sexed from DNA samples retrospectively. Thus
we could not avoid correlations between some key variables. For example, all
unassisted pairs hatched female nestlings second
(Table 1).
|
Videotape scoring
The 110 videotapes were played back on a Sony video cassette recorder
(EV-S1000E) and Sony video monitor with a timer (PVM-2130QM). Most ran for 3 h
and 5 min; the first 5 min were ignored to avoid scoring behavior related to
disturbance caused by visiting the nest tree to switch on the recorder. Four
tapes were between one and a half and 3 h long because of premature battery
failure. Measurements from these tapes were scaled up proportionately to be
comparable with the measures from full-length videos. Two videos of very poor
quality were discarded.
Aggression between nestlings
Any aggressive act (peck, bite, shake) directed from one nestling to
another was called an attack. Nestlings could be identified on screen from the
embroidery thread and paint marks applied to them soon after hatching. From
each 3 h recording, we counted the number of attacks that each nestling
directed towards either of its siblings.
A nestling's position in the hatching sequence may influence the context in which it displays aggression. We categorized attacks as "food-fights" (FF) if they were associated with the arrival of a feeding adult. Otherwise we categorized attacks as "gratuitous violence" (GV). From each day of recording we calculated the proportion of attacks initiated by each nestling that was associated with feeding visits [FF/(FF + GV)]. We then averaged the proportions for each nestling over the entire 10-day observation period, and analyzed these nestling means with respect to their hatch rank.
We examined overall aggression (i.e., FF + GV) in each of the six possible
directional dyads of nestlings separately (A
B, AC, BA,
BC, CA, CB). We analyzed the effect of nestling age (days
after C nestling hatched), and the impact of variables that relate to
competitive asymmetries between nestlings: the hatch intervals (ABinterval,
BCinterval, ACinterval) the sex of chicks (Asex, Bsex, Csex), sex sequences
(ABsex, ACsex, BCsex), and hatchling size differences (ABvol, ACvol, BCvol).
We also tested whether aggression was affected by the presence of male helpers
in the attending group (unassisted pair or group). We noted the time of day
the recording was made (morning/afternoon) and the ambient temperature, but
these variables were never important in analyses and will not be discussed
further.
Differences between siblicidal and nonsiblicidal broods
We compared the ontogeny of aggression in siblicidal with nonsiblicidal
broods over the 10 day recording period. On each day of surveillance, we
summed all attacks made by all nestlings in each brood, and divided this by
the number of directional dyads in the brood (i.e., six dyads in a brood of
three, two dyads in a brood of two). We call this measure "brood
aggression" to distinguish it from dyadic aggression.
We also looked for differences in parental behavior between siblicidal and nonsiblicidal broods. For each 3 h recording session, we calculated the proportion of time that the nestlings were brooded by an attending adult (brooding time in seconds/recording time in seconds). We also counted the number of brooding changeovers (i.e., adult leaves or arrives at the nest). Finally, we counted the number of feeding visits by adults, and divided this by the brood size on that day to get an estimate of the feeding rate per nestling during the recording session. In exploratory analysis, there was no difference in the size of food items brought to the siblicidal versus nonsiblicidal broods. We attempted to record which nestling received the food item brought during an adult's feeding visit, but visibility was often obscured by the incoming adult; these data were of inconsistent quality and were discarded.
Preliminary analyses showed that the proportion of time nestlings were brooded, the number of brooding changeovers, and the nestling feeding rate were all correlated with the age of nestlings. Consequently it was difficult to use brooding patterns and feeding rate as fixed effects in models of aggression because all three measures of parental behavior were collinear with a variable already in the model (nestling age). Instead, we considered whether parents of siblicidal and nonsiblicidal broods had different brooding and feeding patterns, after taking into account nestling age and other potentially important variables (i.e., ambient temperature, time of day, group size).
Analysis
We used a statistical modeling approach. Data from repeated sampling of
nestlings and broods may not be independent. We used Genstat 5 (release 4.1;
Genstat, 1993) to fit mixed
models, incorporating random effects ("brood" or
"nestling" as required) as well as the fixed effects of interest.
The REML procedure was used to estimate variance components. The significance
of terms was assessed by comparing the deviance of a full model with that of a
sub-model excluding the fixed effect of interest, but with the same random
effects. The change in deviance approximates to a Chi-square distribution.
Variables were transformed with natural logarithms or square roots to
normalize residuals. Residual and normal probability plots were used to check
model assumptions. We also checked for autocorrelation effects in the data
after removing the time trend, but found no evidence for persistent serial
dependence.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Aggression
The maximum number of attacks observed in a 3 h observation period was 458 (median 38, 75% quartile 91 attacks). The highest number of attacks made by an individual nestling in 3 h was 408. First-hatched nestlings made more attacks on their siblings than B and C nestlings (Figure 1). Hatch rank also affected the context in which nestlings initiated attacks; B and especially C nestlings initiated a higher proportion of their attacks over food deliveries than A nestlings (ANOVA F2,29 = 3.2; p =.05; mean proportion and SE, A = 0.33 ± 0.08; B = 0.51 ± 0.08; C = 0.60 ± 0.08).
|
Dyadic aggression
Of the six possible directional dyads, we found that aggression varied with
the tested variables in only three dyads (AB, AC, BA). In
these three dyads, aggression always declined as nestlings aged, although the
rate of this decline sometimes interacted with one other variable pertaining
to competitive disparities between nestlings. The full results of the analyses
are presented in Table 2, but
below we summarize the influence of each of the three factors which affect
competitive asymmetries between nestlings (hatch intervals, hatchling size,
and sex of nestlings).
|
The only hatch interval that affected aggression in directional dyads was
the ABinterval. When the ABinterval was short, aggression from BA was
higher in the first few days after hatching but declined more steeply with
nestling age compared to nests where the ABinterval was long
(Table 2,
Figure 2a). Attacks from
AB followed a similar pattern, but the interaction between nestling age
and the ABinterval was not quite significant
(Table 2,
Figure 2b).
|
Initial size differences at hatching had a significant effect on levels of aggression in just one dyad: B attacked A more often when the B hatchling was similar in size or larger than the A hatchling (estimated from their egg volumes; Table 2, Figure 3).
|
The sex of the B nestling affected aggression from A to both its siblings.
The A nestling attacked B more when B was female
(Table 2,
Figure 4a). Since all
siblicidal broods hatched a female second (see
Table 1), the result may have
been confounded by some other attribute common to siblicidal broods. Thus we
restricted the analysis to nonsiblicidal broods, but Bsex was still
significant (Bsex, 
12 = 5.92, p =.015).
Interestingly, the sex of B also affected aggression from AC. When B was
female, A attacked C more in the first few days after hatching, but aggression
declined with nestling age more quickly compared with nests where B was male
(Table 2,
Figure 4b).
|
We found that sex sequences also affected how aggressive A was to its two
siblings. Attacks from AC were much lower when both A and C nestlings
were male (i.e., ACsex = MM), and highest when the ACsex was MF
(Table 2,
Figure 5a). Similarly, the
number of attacks from AB was lower when A and B were both male compared
with other sequences, although this effect was not quite significant
(Table 2,
Figure 5b).
|
Differences between siblicidal and nonsiblicidal broods
Brood aggression (i.e., number of attacks per dyad) generally declined with
nestling age. However, aggression in siblicidal broods was higher initially,
and declined more markedly with nestling age than aggression in nonsiblicidal
broods (age.siblicide, 12 = 10.9, p
<.001; Figure 6). Thus by
day 10, aggression was actually higher in nonsiblicidal broods.
|
The proportion of time that nestlings were brooded decreased as nestlings
aged. In addition, the brooding time in siblicidal broods was lower at first
but declined more slowly over the 10 days compared to nonsiblicidal broods
(age.siblicide, 12 = 7.8, p =.005;
Figure 7a). The proportion of
time spent brooding did not differ between unassisted pairs and groups
(12 = 0.44, p =.51).
|
The change in the number of brooding changeovers over the 10 days was best
described by a quadratic equation. The number of changeovers increased until
the C nestling was 3 days old, then decreased over the remaining 7 days
(day2:21 = 10.4, p =.001;
Figure 7b). For any day, there
were more brooding changeovers in siblicidal broods than nonsiblicidal broods
(21 = 4.2, p =.04;
Figure 7b). There were no
differences in the number of brooding changeovers between pairs and groups
(21 = 2.1, p =.15).
The feeding rate per nestling increased with nestling age
(21 = 43.3, p <.001). There was no
significant difference in the nestling feeding rate between siblicidal and
nonsiblicidal broods (21 = 1.3, p =.25),
nor did the relationship between feeding rate and nestling age differ between
siblicidal and nonsiblicidal broods (age.siblicide,
21 = 0.06, p =.81). If anything, nestlings
in siblicidal broods tended to be fed more (feeds/nestling/recording session,
averaged over all days: siblicidal broods = 4.66 ± 0.50, N =
3; nonsiblicidal broods = 3.94 ± 0.30, N = 8). However, since
nestlings in siblicidal broods were fighting more, they may still have been
hungrier than nestlings in nonsiblicidal broods. In addition, it is possible
that the A and B nestlings in nonsiblicidal broods were getting fed more than
their siblicidal counterparts if they were taking the lion's share of all the
food delivered to their nest.
Adults were never seen to intervene in the fights of nestlings, either by separating nestlings or even by simply brooding them. Yet fights (especially food fights) often took place in full view of the incoming adult. In one nest, the youngest nestling died during a recording session. Shortly afterwards, an adult came in to brood. It discovered the body of the nestling, took it into its beak and swallowed it whole.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Aggression
Overall, the first hatched nestling was the main aggressor, initiating more attacks than either of its siblings. In addition, fighting between B and C (both B
C and CB) was comparatively rare
(Figure 1). The youngest
nestling is assaulted from the moment of hatching, mainly by A, and hence gets
little chance to feed, grow, and develop its motor and fighting skills. Rather
than retaliate against elder siblings, C nestlings apparently tried to avoid
aggression by moving away from attacks. This differs from patterns of
aggression observed in cattle egrets, where fighting is concentrated at the
bottom end of the hierarchy (Fujioka,
1985; Mock and Lamey,
1991; Ploger and Mock,
1986). In cattle egrets the benefits of winning fights may be
greatest for the youngest nestlings since they may avoid being relegated to
the "risky" end of the hierarchy.
In contrast to egrets, aggression in brown pelican broods was concentrated
between the A and B nestlings, possibly because fledging success in this
population was low, and nestlings were fighting for the top rank
(Ploger, 1997). The oldest two
kookaburra nestlings may fight more than other dyads for a similar reason.
Long-term nest monitoring in this population has shown that even in broods
where the C nestling was killed, the B nestling still starved to death in one
third of nests (5/15) later in the nestling period, when growth rates and food
requirements were highest (Legge,
2000c, unpublished data). In addition, fledging weight, which
increases with nestling rank, significantly predicts the probability of
surviving to nutritional independence 2 months after fledging
(Legge, 2000a). So vying for
the top rank may be important for nestling survival, access to food, and
consequently longer-term survival prospects (see also
Vinuela, 1999).
A higher proportion of attacks by the C nestling was associated with feeding visits by adults, whereas a higher proportion of A's attacks involved "gratuitous" violence. Thus, although C fights relatively rarely, these fights are more likely to occur in the context of an immediate reward. When adult kookaburras return with food, they usually perch near the nest for up to several minutes, vocalizing softly with a distinctive call, before entering the hollow (personal observation). This would alert nestlings to the impending arrival of food, and give them time to fight and gain control of the prime feeding position facing the tunnel entrance. When visiting nests to check nestlings, we also noticed that our movements at the hollow entrance (scraping against the wood and casting shadows over the brood) seemed to instigate bouts of fighting.
Dyadic aggression
The analyses of dyadic aggression revealed three broad patterns. First,
aggression decreased as nestlings aged. As dominance hierarchies within nests
are established, overt fighting may become unnecessary
(Enquist and Leimar, 1983).
Second, in dyads where fighting was relatively infrequent (i.e., BC,
CB, CA), aggression did not vary with nestling age nor any other
variable. It appears that the C nestling in particular did not respond to
variation in its competitive environment, and its fate depended mainly on
interactions between its two elder siblings. Third, in dyads where fighting
was frequent (i.e., AB, BA, AC), asymmetries between
nestlings caused by hatching asynchrony, hatchling size differences, and sex
differences were all linked to variation in aggression.
When A and B hatched relatively synchronously, aggression in both
directions (i.e., both AB and BA) was higher initially, but
declined more steeply with nestling age compared to nests where A and B
hatched asynchronously. The fact that A and B fight more when they are closely
matched in age is consistent with the theory of asymmetric contests, which
predicts that fighting should escalate when competitive asymmetries between
the contestants diminish (Enquist and
Leimar, 1983; Maynard-Smith
and Parker, 1976). For example, the B nestling also attacked A
more frequently if the physical size difference between them at hatching was
small or reversed. In our analyses, age and size asymmetries between other
nestlings (i.e., the ACinterval, BCinterval, ACvol, and BCvol) did not affect
aggression in those dyads. Perhaps interactions between the older two
nestlings overwhelm other effects; a larger sample may reveal that asymmetries
between other dyads do have subtle influences on aggression.
The A nestling also attacked B more frequently when B was female. Females
begin outstripping males in mass and size days after hatching. Ten days after
hatching females are 8% heavier than males of similar age (unpublished data).
Therefore to a first-hatched male, a younger sister may constitute a threat to
dominance, requiring preemptive action immediately after hatching. However,
aggression was also very high when the AB sex sequence was FF, and this is not
explained by differences in growth between these two nestlings. Alternatively,
females may simply be more aggressive than males and any dyad involving at
least one female experiences more fighting. In support of this, aggression
from AB and from AC was lowest when both nestlings in the dyad
were male. However, this result must be interpreted cautiously, as only one
nest had an AB sex sequence that was MM, and this nest was also one of the two
nests with an AC sex sequence that was MM. The second nest with an AC sex
sequence of MM experienced siblicide, reducing its contribution to the
analysis.
When B was female, A also attacked C more frequently, especially in the first few days after hatching. Therefore it appears that C may experience some of the "overflow" of aggression between A and B. In other words, it may get "caught in the crossfire." This would explain why a short ABinterval, which inflates aggression in that dyad, results in the death of C.
The small number of studies that have examined aggression between dyads in
three nestling broods have usually ignored the interaction of nestling age
with other key variables (such as hatch asynchrony). The approach we use here
shows that the rate of decline in aggression with nestling age may vary
depending on attributes of the dyad under consideration. In addition, previous
studies have ignored the direction of attacks with a dyad, yet clearly
nestlings within a dyad may respond differently to the same variable. For
example, in our data A attacks B more when the latter is female. but B does
not retaliate. Aggression from AC depended on several variables, but C
did not vary its aggression towards A.
Differences between siblicidal and nonsiblicidal broods
In our sample of broods, all those that experienced siblicide had a female
B nestling and a short ABinterval (Table
1), both of which increased aggression between the oldest two
nestlings. Since A and B fought each other more than other dyads, their
rivalry underlaid the different ontogenies of brood aggression in siblicidal
and nonsiblicidal broods. The frequency of attacks per dyad was higher in
siblicidal broods for the first 5 days after hatching, but declined more
steeply over the entire 10 day period compared with nonsiblicidal broods. By
day 10, the frequency of attacks was actually higher in nonsiblicidal broods
(Figure 6).
The different patterns of aggression in siblicidal and nonsiblicidal broods could occur for several reasons. The competitive asymmetries between nestlings (especially A and B) are different in the two types of nest, and conflict may therefore be resolved in different ways. For example, since the A and B nestlings fight more frequently in siblicidal broods in the few days after hatching, perhaps the question of dominance between them comes to a swifter resolution compared to their nonsiblicidal counterparts.
The different patterns of aggression in siblicidal and nonsiblicidal broods
may also be explained by differences in parental behavior. Like most reports
from other siblicidal species, adult kookaburras were never observed to
intervene in fights between nestlings (but see
Lougheed and Anderson, 1999;
Vinuela, 1999). However, they
could still affect aggression indirectly by modifying feeding or brooding
behavior. Unfortunately we could not explore the direct effect of feeding
rates and brooding patterns on aggression because they were correlated with
nestling age. Nevertheless, we checked whether these parental behaviors
differed between siblicidal and nonsiblicidal broods in ways that would
account for the different patterns of aggression.
If fighting is moderated by hunger, in order to explain the observed patterns of aggression, nestlings in siblicidal broods should be fed less per capita immediately after hatching, but progressively more as they age. However, we found no evidence to support this: nestlings from siblicidal nests were fed similarly or even more than counterparts from nests without siblicide. A hunger cue would be unreliable if siblicide occurs before food is potentially in short supply, as with kookaburras.
To some extent, brooding restricts the ability of nestlings to fight. The brooding schedule of siblicidal broods differed from that of nonsiblicidal broods in ways that would "encourage" more aggression. Siblicidal broods had a higher number of brooding changeovers; in other words they were left unattended more frequently. When a brooding adult leaves the nest, the temperature in the hollow drops. Once nestlings get cold, they became torpid and seemed unable to fight (personal observation). Thus nestlings were able to fight more if left alone for two 5 min blocks than if left alone for a single 10 min block. In addition, nestlings in siblicidal broods were brooded less during the first 5 days after hatching, when aggression is highest. Curiously, equivalence in the brooding time of siblicidal and nonsiblicidal nests was achieved at the same age as equivalence in the levels of brood aggression, implying the two measures are somehow related (5 days, compare Figures 6 and 7a).
These differences in brooding behavior could arise if adults of siblicidal
broods have trouble finding food, and need to leave the nest unattended for
longer periods, or more frequently
(Newton, 1979). However, three
observations make us doubt this explanation for the brooding behavior of
kookaburras. First, nestlings in siblicidal broods were not fed less than
nestlings in nonsiblicidal broods. Second, food is unlikely to be limiting in
the first week after hatching, when the requirements of nestlings are still
low (Legge, 2000b). Moreover,
if food was already a problem at this early stage, we would expect to
see a difference in brooding behavior between unassisted pairs and groups,
because we know that pairs have trouble finding enough food for their young
much later in the nestling period, when growth rates and feeding rates are
highest (Legge, 2000c).
However, pairs did not brood less than groups, suggesting that early brooding
patterns are not affected by foraging constraints. Third, the decrease in
brooding time over the observation period was shallower in siblicidal broods,
yet the opposite pattern would be predicted because foraging should take up
progressively more time as nestlings grow.
Instead, we suggest two alternatives. In siblicidal broods, where
aggression is high because the A and B nestlings are closely matched, the
energetic expenditure of nestlings increases. This may force parents to leave
the nest more frequently to hunt (Mock and
Ploger, 1987; Osorno and
Drummond, 1995). In other words, heightened aggression between
nestlings may explain the brooding patterns of siblicidal broods, rather than
the brooding patterns affecting the aggression. Alternatively, the adults
attending siblicidal broods may deliberately modify their brooding
behavior to promote brood aggression. As already mentioned, adults returning
with food seemed to encourage fighting by calling softly outside the nest for
a period before entering. Unfortunately we are unable to examine whether the
frequency or length of this "fight-solicitation" behavior by
adults differed between siblicidal and nonsiblicidal broods.
Summary
Siblicide in kookaburras has previously been shown to be associated with
four correlated variables: the number of male helpers attending the nest, the
relative size of the third nestling at hatching, the hatch interval between
the first and second nestlings, and the sex sequence of the first and second
nestlings. The study presented here shows that the number of male helpers and
the size of the C hatchling do not actually affect aggression between
nestlings, suggesting that these variables simply reflect the quality or
condition of the breeding female. In contrast, the hatch interval and sex
sequence of the oldest two nestlings may be proximately responsible for
siblicide of the third nestling because the latter gets "caught in the
crossfire" of escalated aggression between A and B. Parents may
"engineer" an aggressive environment in the brood by adjusting
hatch intervals and the sexes of offspring in order to decrease the
competitive asymmetry between the oldest two nestlings. In addition, parents
of siblicidal broods do not interfere with aggression, and may even modify
their brooding patterns to encourage fighting between nestlings.
|
ACKNOWLEDGEMENTS |
|---|
This manuscript was prepared by S.L. and A.C. from the Honors thesis of the late Anjeli Nathan. Chris Boland, Michelle Hall, and Rob Heinsohn made helpful comments on the manuscript. Bruce Barrie, Neil Kaines, Elsie Krebs and Alan Muir provided invaluable technical support. We know that Anjeli would have thanked Oggy, Vis, and Liz Nathan, and especially Daniel Ebert, for their support and encouragement. This research was funded by a grant from the Country Women's Association to A.N., an Australian Research Council grant to A.C., and ANU and Overseas Postgraduate Scholarships to S.L. Permits were acquired from the ANU Animal Experimentation Ethics Committee, Canberra Nature Park, and the Australian Bird and Bat Banding Scheme, all of whom were extremely helpful.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Bortolotti GR, 1986. Influence of sibling competition on nestling sex ratios of sexually dimorphic birds. Am Nat 127: 495-507.
Braun BM, Hunt GLJ, 1983. Brood reduction in black-legged kittiwakes. Auk 100: 469-476.
Cook MI, Monaghan P, Burns MD, 2000. Effects of
short-term hunger and competition asymmetry on facultative aggression in
nestling black guillemots Cepphus grylle. Behav Ecol
11: 282-287.
Drummond H, Garcia Chavelas C, 1989. Food shortage influences sibling aggression in the blue-footed booby. Anim Behav 37: 806-819.
Drummond H, Osorno JL, 1992. Training siblings to be submissive losers: dominance between booby nestlings. Anim Behav 44: 881-893.
Drummond H, Osorno JL, Torres R, Garcia Chavelas C, Larios, HM, 1991. Sexual size dimorphism and sibling competition: implications for avian sex ratios. Am Nat 138: 623-641.
Enquist M, Leimar O, 1983. Evolution of fighting behaviour: decision rules and assessment of relative strength. J Theor Biol 102: 387-410.
Forbes LS, 1991. Hunger and food allocation among nestlings of facultatively siblicidal ospreys. Behav Ecol Sociobiol 29: 189-195.
Forbes LS, Ydenberg RC, 1992. Sibling rivalry in a variable environment. Theor Pop Biol 41: 135-160.
Fujioka M, 1985. Sibling competition and siblicide in asynchronously-hatching broods of the cattle egret Bubulcus ibis. Anim Behav 33: 1228-1242.
Gargett V, 1978. Sibling aggression in the black eagle in the Matapos, Rhodesia. Ostrich 49: 57-63.
Genstat Committee, 1993. Genstat 5 Release 3 Reference Manual. Oxford: Clarendon Press.
Gerhardt RP, Gerhardt DM, Vasquez MA, 1997. Siblicide in swallow-tailed kites. Wilson Bull 109: 112-120.
Golla W, Hofer H, East ML, 1999. Within-litter sibling aggression in spotted hyaenas: effect of maternal nursing, sex and age. Anim Behav 58: 715-726.[ISI][Medline]
Griffiths R, Double MC, Orr K, Dawson RJG, 1998. A simple DNA test to sex most birds. Mol Ecol 7: 1071-1075.[Medline]
Hoyt DF, 1979. Practical methods for estimating the volume and fresh weight of bird eggs. Auk 96: 73-77.[ISI]
Legge S, 2000a. The effect of helpers on reproductive success in the laughing kookaburra. J Anim Ecol 69: 714-724.
Legge S, 2000b. Helper contributions in the cooperatively-breeding laughing kookaburra: feeding young is no laughing matter. Anim Behav 59: 1009-1018.[ISI][Medline]
Legge S, 2000c. Siblicide in the cooperatively breeding laughing kookaburra. Behav Ecol Sociobiol 48: 293-302.
Legge S, Cockburn A, 2000. Social and mating system of cooperatively breeding laughing kookaburras (Dacelo novaeguineae). Behav Ecol Sociobiol 47: 220-229.
Legge S, Heinsohn R, Double M, Griffiths R, Cockburn A,
2001. Complex sex allocation in the laughing kookaburra.
Behav Ecol 12:
524-533.
Lougheed LW, Anderson DJ, 1999. Parent blue-footed boobies suppress siblicidal behavior of offspring. Behav Ecol Sociobiol 45: 11-18.
Machmer MM, Ydenberg RC, 1998. The relative roles of hunger and size asymmetry in sibling aggression between nestling ospreys, Pandion haliaetus. Can J Zool 76: 181-186.
Magrath RD, 1992. The effect of egg size on growth and survival of blackbirds: a field experiment. J Zool Lond 227: 639-653.
Maynard-Smith J, Parker GA, 1976. The logic of asymmetric contests. Anim Behav 24: 159-175.
Mock DW, Lamey TC, 1991. The role of brood size in regulating egret sibling aggression. Am Nat 138: 1015-1026.
Mock DW, Lamey TC, Ploger BJ, 1987. Proximate and ultimate roles of food amount in regulating egret sibling aggression. Ecology 68: 1760-1772.
Mock DW, Ploger BJ, 1987. Parental manipulation of hatch asynchrony in cattle egrets: an experimental study. Anim Behav 35: 150-160.
Newton I, 1979. Population ecology of raptors. Berhampsted: Poyser.
Osorno JL, Drummond H, 1995. The function of hatching asynchrony in the blue footed booby. Behav Ecol Sociobiol 37: 265-273.
Parker GA, Mock DW, Lamey TC, 1989. How selfish should stronger sibs be? Am Nat 133: 846-868.[ISI]
Parry V, 1973. The auxiliary social system and its effect on territory and breeding in kookaburras. Emu 73: 81-100.
Ploger BJ, 1997. Does brood reduction provide nestling survivors with a food bonus. Anim Behav 54: 1063-1076.[Medline]
Ploger BJ, Mock DW, 1986. Role of nestling aggression in food distribution to nestling cattle egrets (Bulbulcus ibis). Auk 103: 768-776.
Vinuela J, 1999. Sibling aggression, hatching asynchrony, and nestling mortality in the black kite (Milvus migrans). Behav Ecol Sociobiol 45: 33-45.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||