Behavioral Ecology Vol. 12 No. 5: 524-533
© 2001 International Society for Behavioral Ecology
Complex sex allocation in the laughing kookaburra
a Botany and Zoology, Australian National University, Canberra, ACT 0200, Australia b DEEB, Graham Kerr Building, Glasgow University, Glasgow G12 8QQ, UK
Address correspondence to S.L. Legge. E-mail: sarah.legge{at}anu.edu.au .
Received 14 April 2000; revised 1 August 2000; accepted 24 August 2000.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
In groups of the cooperatively breeding laughing kookaburra (Dacelo novaeguineae), offspring sex varied with the type of social group and with hatch rank. Groups with female helpers, especially if all helpers were female, had male-biased clutch and fledging sex ratios. Groups without female helpers (unassisted pairs or male-only helpers) had female-biased clutch and fledging sex ratios. Breeding females responded facultatively to increases in the number of female helpers in their group by producing more male eggs. These biases may occur if breeding females try to limit the number of daughters recruited into their group because unlike male helpers, female helpers depress the breeding success of their parents. Across all nests, two-thirds of first-hatched young were male, two-thirds of second-hatched young were female, and the sex ratio of third-hatched young was even. Hatch rank sex ratios also varied dramatically between different types of social groups, from 16.7% for second-hatched nestlings of unassisted pairs to 100% for first-hatched nestlings of groups with only female helpers. A corollary of the relationship between hatch rank and sex was that hatching sex sequences were distributed nonrandomly: all groups avoided hatching a daughter first followed by a son (FM). Sibling competition is aggressive and sometimes fatal. Since females grow to be 15% larger than males the hatching sequence of sexes could affect nestling growth and mortality. However, an exhaustive analysis found little evidence that growth or survival of males was compromised if hatched after a sister. The small number of FM sequences may only have occurred in nests that were able to ameliorate any negative consequences. Alternatively, when clutch size is small and fledging success unpredictable because of brood reduction, the preferred brood sex ratio may be contingent on the number of fledged young, making it advantageous to order the sexes in the brood.
Key words: cooperative breeding, kingfisher, reverse dimorphism, sex allocation, sex ratio, siblicide.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Fisher (1930
) showed that
when males and females have equivalent reproductive value, parents should
invest equally in sons and daughters. However, situations exist where one sex
may be more profitable than the other, and parents could maximize their
fitness by adjusting their sex allocation
(Frank, 1990). Examples of
nonrandom sex ratios in birds have accumulated only recently, but the types of
biases and their proposed adaptive values have been very varied (reviews by
Hardy, 1997;
Sheldon, 1998). In species
where individuals interact with kin, opportunities arise for competitive
and/or cooperative effects to alter the relative values of each sex. Sex
differences in dispersal may cause one sex to associate more intensively with
siblings or parents, potentially bringing them into competition over resources
and effectively making them more "costly." The "local
resource competition" hypothesis and its variants therefore predict
overproduction of the dispersing sex
(Clark, 1978). Alternatively,
if cooperative interactions between kin increase the fitness of breeders,
parents may preferentially produce the philopatric sex. A version of
"local resource enhancement," the repayment model, has been
applied to cooperatively-breeding birds, many of which live in family groups
with related helpers of one sex (Emlen et
al., 1986; Lessells and Avery, 1997).
Species that breed cooperatively should be excellent models for evaluating
the importance of enhancement and competitive effects, but there are still few
examples. A population-wide bias towards males in red-cockaded woodpeckers
(Gowaty and Lennartz, 1985)
was interpreted as support for enhancement effects because helpers are usually
male, but this result was not repeated in a later study reporting on a larger
sample from the same species (Walters,
1990). Enhancement and resource competition models may be more
usefully applied at the level of individual families, rather than across
populations of cooperative-breeders
(Koenig and Walters, 1999),
since enhancement and competitive effects probably operate to varying extents
among families within a population. Sex ratio variation between families has
been reported for red-cockaded woodpeckers
(Gowaty and Lennartz, 1985),
green woodhoopoes (Ligon and Ligon,
1990), wild dogs (Malcolm and
Marten, 1982), eclectus parrots
(Heinsohn et al., 1997), but
most convincingly for Seychelles warblers
(Komdeur et al., 1997).
Daughters in this species are most likely to help and can have a positive
effect on the reproductive success of their parents. However, too many helpers
can reduce nest success, particularly when food resources are poor. Breeding
females produce daughters in their single-egg clutch when the territory is of
high enough quality to support more birds, but produce the dispersing sex
(sons) when the natal territory is "saturated."
One of the more common patterns of sex ratio bias reported in the
literature is a sequence effect within the clutch, where the sex of eggs
relates to their position in the laying sequence. Such sequence effects could
cause considerable skews in the sex ratio across clutches, if the
"switch" between sexes was moved up or down the sequence, or if
tempered with order-related mortality or clutch size variation
(Krackow, 1999). For example,
if a cooperative breeder hatches the helping sex (say male) at the start of
the clutch sequence, and brood reduction is common or the clutch size is
curtailed in groups lacking male helpers, the brood sex ratio in these groups
becomes relatively male-biased. In contrast, if groups with helpers have
larger clutches or less brood reduction, they will produce clutches with a
balanced or female-biased sex ratio.
Sequence effects may also have important consequences for nestling
interactions in dimorphic birds, particularly if there is sibling aggression.
The sequence of sexes could affect nestling condition, nest productivity, and
fledging sex ratios (Bortolotti,
1986; Dzus et al.,
1996; Edwards and Collopy,
1983; Edwards et al.,
1988; but see Drummond, 1991). In mixed sex clutches, hatching the
smaller sex second may give it the double disadvantage of hatch rank and sex,
leading to reduced growth and/or reduced productivity from those nests,
especially if siblicide is possible. Alternatively, hatching the smaller sex
first could increase sibling competition in the nest if shifting size
asymmetries destabilize dominance hierarchies. In siblicidal species, an older
nestling destined to be smaller than its nest mate may kill its younger
sibling preemptively, to eliminate anticipated conflict.
Here we describe offspring sex ratio patterns in the laughing kookaburra
(Dacelo novaeguineae) in relation to two major features of its
complex social and breeding biology. First, kookaburras breed in nuclear
family groups where a monogamous pair are assisted by up to six
offspring-helpers (Legge and Cockburn,
2000; Parry,
1973). Dispersal, helping behavior and the effect of that help
depends on the sex of the helper. Daughters disperse at a younger age than
sons, sometimes leaving their natal group before they have helped at all
(Legge and Cockburn, 2000).
The contribution of helpers during nesting allows breeders to reduce their
workload, with potential long-term effects on survivorship and reproductive
success. However, females are "poor" helpers, feeding young at a
much lower rate than their brothers. Thus, whereas males have a neutral effect
on the success of each nesting attempt, female helpers actually reduce nest
success (Legge,
2000a,b).
Breeders may be sensitive to the number of males and females already in the
group, for example producing sons if male helpers are lacking, and avoiding
too many daughter-helpers at once. Consequently, we examine the clutch and
fledging sex ratios from groups of different size and composition.
Second, kookaburra broods hatch asynchronously, producing a size hierarchy
that facilitates siblicide. The youngest nestling dies in nearly half of all
nests because of severe aggression from older siblings
(Legge, 2000c). Since
kookaburras are also reverse size dimorphic (adult females 15% larger,
fledgling females 7% larger), the order of sexes in the hatching sequence may
influence the outcome of sibling aggression. Consequently we examine the
relationship between hatch rank and sex, the distribution of sex sequences,
and the effect of sex sequence on fledging success and nestling growth.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The data presented here were collected between December 1994 and February 1998 from a closely monitored population of laughing kookaburras in Canberra Nature Park, a eucalypt woodland reserve in southeast Australia. About 35 groups in territories of 16-224 ha occupied the 20 km2 study area each year. Kookaburras breed in the austral spring and summer, between September and January. Females lay a single clutch per season, usually of three eggs, in a naturally occurring tree hollow. The clutch hatches asynchronously, with intervals from a few hours up to 3 days between successive eggs. Hatch order reflects lay order. Nestlings emerge blind and naked, but use a specially modified "beak hook" to attack their siblings. The youngest nestling was killed in nearly half of nests where three young hatched (46%; Legge, 2000c
). Groups fledged an average of 1.4 young (SD = 1.1,
n = 131 group-years) 32-40 days after hatching. The young remain
dependent on adult group members for food for another 8 weeks
(Legge and Cockburn, 2000;
Parry, 1973).
Group size was defined as the total number of birds attending a nest,
including the breeding pair. Groups were also classified into four
"helper types": groups could have no helpers (the breeding pair is
unassisted), all-female helpers, all-male helpers, or helpers of both sexes.
For details of censusing techniques and general field methods see
(Legge, 2000a;
Legge and Cockburn, 2000).
Nestling sex
Nestling sex was determined from DNA. A small blood sample (10-20 µl)
was taken within 48 h of hatching by puncturing the brachial vein. DNA was
extracted from the blood, and a molecular method was used to identify sex (see
Griffiths et al., 1998 for
details). The sex of 21 dissected birds was correctly predicted using this
method. Unhatched eggs were checked for embryonic material, which was
collected for DNA extraction and molecular sexing (successful for 10/24 eggs
in 10 clutches).
Sex sequences
Near the hatch date (estimated previously by candling eggs) nests were
visited daily to match hatchlings to eggs and assign them sequential hatch
ranks A, B, and C. Nestlings were marked temporarily by tying soft, colored
embroidery thread around the tibia. They were given permanent metal colored
bands on the last nest-visit when the A nestling was 32 days old. Nestling sex
sequences describe the sex of nestlings in order of their hatch rank. For
example, the sex sequence `MF' for A and B nestlings means that a male hatched
first followed by a female second. In some clutches one or two eggs failed to
hatch (24/257 eggs in 22/101 clutches). In these clutches the observed sex
ratio and sex sequences of hatchlings may differ from the sex ratio and sex
sequence of eggs. Within clutches, eggs could be ranked in order of egg volume
[cm3, calculated as 0.00051 x Length x
(Width)2; Hoyt,
1979]. Hatch rank nearly always matched egg volume rank: only 6%
of 180 eggs hatched earlier than their rank assigned from egg volume
(Legge, 2000c). The hatch rank
of unhatched eggs was deduced from their volume, and the ranks of successfully
hatched young were adjusted accordingly. Adjusted hatch ranks and sex
sequences are italicized (A, B, C, AB, etc.) to distinguish them from
the hatch rank and sex sequences assigned to nestlings that had actually
hatched.
Nestling growth
After all eggs had hatched, nests were visited at weekly intervals (usually
days 11, 18, 25, and 32; day 0 is when nestling A hatches) to measure nestling
mass, tibia, and flattened wing chord (nearest g, 0.1 mm, and mm
respectively). Mass gives a general indication of size and condition, tibia
reflects skeletal size, and wing-length indicates feather growth. Fledging
success was defined as the number of young present in the nest on the final
visit, when the A nestling was 32 days old.
Analysis
In descriptive statistics, heterogeneity, goodness-of-fit tests
(
2) and the binomial test were used to check for departures
from equal sex ratios. Other analyses were carried out with a statistical
modeling approach using Genstat 5 release 4.1
(Genstat Committee, 1993).
Some data were potentially nonindependent because groups were represented
between one and four times, and broods contained up to three nestlings. To
account for repeated sampling with an unbalanced design, mixed models were
fitted incorporating random factors ("group" and
"brood," as required) as well as the fixed effects of
interest.
Factors affecting clutch and fledging sex ratios
Variation in clutch and fledging sex ratios was analyzed by defining brood
sex ratio as the binomial response variable (number of males over brood size)
in generalized linear mixed models with a binomial error distribution
(Genstat Committee, 1993).
Estimates of the variance components and fixed effects were initially obtained
using the restricted maximum likelihood procedure (REML). As the standard
errors of the estimates for the random term proved to be large compared to the
estimates, indicating negligible dependency associated with
"group," the models were simplified by omitting the random term
and using generalized linear models. The significance of terms was assessed
using the change in deviance (which approximates to a 2
distribution) associated with dropping that term from a fuller model.
Variables of interest were year, hatch date, group size, and helper type.
Group size and helper type are correlated (e.g., groups without helpers must
have group size two), and were therefore added to models separately. Broods
were only included in the clutch sex ratio analysis if the sex of the entire
clutch, including unhatched eggs, was known (n = 66 clutches; 38
groups). Similarly, broods were only used in the analysis of fledging sex
ratio if the sex of all fledglings was known (n = 82 broods; 46
groups).
Effect of nestling sex on fledging success and nestling growth
When analyzing the effect of brood sex ratios, the sex of siblings and sex
sequences on fledging success and nestling growth, we used the ranks and sexes
of young that had actually hatched (i.e., not adjusted for unhatched eggs),
since the growth and survival of nestlings can only be affected by extant
siblings.
Fledging success was specified as the response variable in a linear mixed model with "group" as the random term. Variables of interest were year, hatch date, group size, helper type, brood sex ratio, sex of hatched nestlings A, B, and C, and sex sequences of hatched nestlings AB, BC, AC, and ABC. Again, group size and helper type were assessed in separate models. The AB sequence was also correlated with group size and helper type (see Results), and was therefore assessed separately. Information on each variable was not available for every nest, so sample sizes in this analysis vary depending on the variables being considered.
To summarize nestling growth, logistic curves of the form [a/(1 +
e-K(age - i)] were fitted to the mass and tibia measurements of
nestlings that survived to fledging age
(Ricklefs, 1971). The three
parameters of the equation describe the asymptote (a), the growth constant
(K), and the inflection point of the curve (i). Using parallel curve analysis,
curves were fitted in steps of increasing complexity to assess first whether
asymptotes differed between nestlings, then whether the nonlinear parameters
(growth constant and inflection point) differed
(Genstat Committee, 1993).
Fitting separate asymptotes for each nestling significantly improved the
model, but adding separate nonlinear parameters for each nestling did not.
Thus, nestlings that achieve a higher asymptote do so by growing more quickly,
reaching their "target" in the same time as a smaller nestling
(Ricklefs, 1971). Logistic
curves were in-appropriate for wing-length because wings were still growing on
the final nest-visit. Instead, the mean rate of length increase (mm/day) was
calculated from the linear phase of growth, between day 10 and the final nest
visit (day 30 to 32, depending on hatch rank of nestling).
The three estimates of nestling growth (mass asymptote, tibia asymptote,
wing growth rate) were used as response variables in linear mixed models. The
REML procedure in Genstat was used to estimate fixed effects and variance
components for the random terms "group" and "brood."
The deviance explained by a full model was contrasted with that of a sub-model
excluding the fixed effect of interest, and the change in deviance was used to
assess the significance of terms (Genstat
Committee, 1993). Data were available for 125 nestlings (61
broods, 39 groups). Variables of interest were nestling sex and hatch rank,
and the number of male and female siblings. Other variables likely to affect
nestling growth were also tested: year, hatch date, group size, and brood
size. This analysis suggested the growth of focal nestlings was not affected
by the sex of their siblings (see Results). A second analysis was performed to
confirm this surprising result. Using only broods of three young (n =
22), the growth estimates of A, B, and C nestlings were modeled separately to
see whether nestling growth of either sex at each hatch rank was affected by
the sex or sex sequences of their siblings.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Overall sex ratios
Across all nests, including those where one or two eggs did not hatch, the sex ratio of 242 eggs from 101 clutches (51 groups) was 47.1% male. In 66 clutches (38 groups) where the sex of every egg (n = 189) was known, the sex ratio was also 47.1%. Ten eggs which failed to hatch had embryonic material suitable for sexing: six were males, four females. At fledging, 48.6% of 175 young were male (88 nests, 49 groups). Using the binomial test, none of these sex ratios differed from parity.
Clutch and fledging sex ratios
Group size had no effect on clutch or fledging sex ratios
(Figure 1a; Eggs,
32 = 6.2, p =.1; Fledglings,
32 = 3.1, p =.38).
Figure 1a suggested that
unassisted pairs laid more female eggs and fledged more daughters than groups
of three or more. However, the differences between unassisted pairs and all
larger groups were also not significant (Eggs, 12 =
3.0, p =.08; Fledglings, 12 = 1.5,
p =.22). In contrast, when helper type was examined in the model
instead of group size, groups lacking female helpers (i.e., no helpers or
all-male helpers) had female-biased clutches, whereas groups with female
helpers (i.e., all-female helpers or helpers of both sex) had male-biased
clutches (Figure 1b;
32 = 13.0, p =.005). By fledging, groups
with all-female helpers had very male-biased broods, but groups of other
helper types had even or slightly female-biased broods
(Figure 1b;
32 = 14.3, p =.003). Clutch and fledging
sex ratios were not affected by year or hatch date.
|
Since the presence of female helpers appeared to have a strong effect on the clutch and fledging sex ratios, we compared the number of males per clutch produced by the same breeding pair when the number of female helpers in their group had increased. These comparisons go forwards and backwards in time, removing the potentially confounding effect of breeder age. When female helpers were added to the group in 13 paired comparisons, the number of male eggs in the clutch increased in seven pairs, was unchanged in six, and never decreased (Wilcoxon paired-sample by rank test, T = 14.0, p =.02; seven paired comparisons go forward in time, six go backward). In 14 comparisons, the number of males in the brood at fledging increased in six pairs, was unchanged in seven, and decreased in only one pair (T = 11.0, p =.09; seven paired comparisons go forward in time, seven go back-ward). Although the sample is small, this is a powerful test of the facultative response of females to changes in the composition of their group. Individual breeding females respond to the addition (or removal) of female helpers in their group by altering the sex ratio of their clutch, and this probably affects the sex ratio of their brood at fledging.
Hatch rank and nestling sex
Across all nests, A nestlings were predominantly male (58/92
males; 63.0%; binomial test p =.02). B nestlings were
predominantly female (26/82 males; 31.7%; binomial test p =.002). In
contrast, the sex of C nestlings did not deviate from parity (27/63
males; 42.9%; binomial test p =.5). These numbers incorporate the sex
and deduced hatch ranks of un-hatched eggs. Sample sizes dwindle for
C nestlings because some were killed by their siblings before a blood
sample could be taken for sexing, and some clutches had only two eggs.
Since clutch sex ratios varied between groups with different helper types,
we examined whether the relationship between hatch rank and sex was consistent
within groups of different helper type. In a generalized model with a binomial
error distribution, nestling sex was affected by both hatch rank
(22 = 10.6, p =.005) and helper type
(32 = 11.6, p =.009), but these variables
did not interact. Thus, groups of different helper types all maintain the
pattern that A nestlings are more likely to be male than B
nestlings (Figure 2), even
though the clutch sex ratio and the mean sex ratio of A nestlings
varies between helper types. Consequently, some helper type and hatch rank
combinations displayed extravagant sex ratio biases. For example, the
B nestling in unassisted pairs was male in only 16.7% of nests
(n = 24), whereas the A nestling in all-female groups was
male in 100% of nests, although the sample for this last category was small
(n = 9).
|
Sex sequences
The sex ratio biases in hatch ranks A and B could arise
if certain AB sex sequences are avoided or overproduced. The observed
distribution of AB sex sequences was significantly nonrandom. In
particular, the FM sequence was very rare (32 =
15.6, p =.001; see totals in Table
1). This distribution of AB sex sequences could result
either because FM was avoided, or because breeders preferred particular sexes
at each hatch rank. In this data set, it is impossible to distinguish between
these two alternatives. For example, if breeding females avoid producing the
FM sequence for the A and B nestlings, the expected overall
sex ratios in hatch ranks A and B would be 67% and 33%,
which is close to that observed (60% and 32% respectively).
|
The distribution of BC sex sequences appeared to be non-random,
although the effect was nonsignificant (32 = 6.9,
p =.08; see totals in Table
1). BC sequences leading with a female were more common
(FF and FM). However, this is expected since B nestlings were
predominantly female.
ABC sequences (Table
2) were not distributed nonrandomly, although the sample is
reduced because we needed to know the sexes of three consecutive eggs rather
than just two (27 = 10.0, p =.19).
However, the patterns revealed in AB and BC sequences were
still reflected in the ABC sequences. The FMM and FMF sequences were
rarest (i.e., those where AB is FM), and the MFF and FFF sequences
were most common (i.e., those where BC is FF). An important point
pertaining to the mechanism of sex allocation to eggs emerges from
Table 2. The sex of sequential
eggs in the clutch can switch from one sex to the other more than once because
MFM and FMF are not absent sequences.
|
The distribution of dyadic sex sequences also differed between groups of
different helper type (AB sex sequences, heterogeneity test
29 = 19.1, p =.02; BC sex
sequences 29 = 21.3, p =.01,
Table 1). Once again, this
complements the observation that the mean sex ratio at each hatch rank
differed between helper types (compare hatch rank sex ratios in
Table 1 to
Figure 2). Notably, in addition
to the general paucity of FM for the AB sequence, the MM sequence was
rare for pairs without helpers and the FF sequence was rare in groups with
all-female helpers. Interpretation of the heterogeneous distribution of
BC sequences is limited by the short-fall in some AB
sequences (e.g., BC sequences starting with a male must be rarer).
However, it appears that the FF sequence was overproduced by groups with
all-male helpers (13/17).
Sex sequence and fledging success
The nonrandom distribution of sex sequences suggested that some sequences
might influence competitive nestling interactions. Therefore we examined the
effect of sex sequences on nestling growth and nest success. Fledging success
was lower in groups with the AB sex sequence MF, and in groups where B was
female, but these differences were not significant (AB sequence,
23 = 6.45, p =.09; B sex,
21 = 3.10, p =.09;
Figure 3). However, note that
the FM sex sequence for the AB dyad, which is the rare or
"avoided" sequence, did not result in lower fledging success,
although the sample for this sequence is small (n = 5). Brood sex
ratio, the sex of A, C, and the sex sequences BC, AC, and ABC did not affect
fledging success. Year, hatch date, and group size also failed to improve the
model. However, fledging success differed between groups with different helper
types, being lowest in groups with only female helpers
(
23 = 9.3, p =.03;
Figure 4).
|
|
Sex sequence and nestling growth
Although nest success was not compromised in nests with the taboo FM
sequence for the AB dyad, sex sequence might still influence nestling growth.
This could be important because fledging weight is a significant positive
predictor of juvenile survival (Legge,
2000a). As expected, female nestlings attained greater mass and
tibia length than males, but there was no difference in the rate of wing
growth between the sexes (Figure
5 and Table 3).
Nestling mass decreased with hatch rank, as did tibia length and rate of wing
growth (Figure 5 and
Table 3). Groups with at least
one helper produced heavier and slightly larger nestlings than unassisted
pairs (Tables 3 and
4), but again, wing growth was
unaffected.
|
|
|
In an exhaustive analysis, there was no evidence that the growth of either sex was adversely affected by the sex of its siblings, or that either sex was a "poorer" competitor (Table 3). Figure 5a,b,c show the mass asymptote, tibia asymptote, and wing growth rates respectively for nestlings in each hatch rank, separated by sex. Although the data suggested that males performed poorly compared to females in C position for all estimates, the sex and hatch rank interaction was not significant in any model. Besides hatch rank, nestling sex did not interact with year, hatch date, group size, brood size, or the number of brothers or sisters in the brood. When the analysis was restricted to broods of three young, growth estimates for nestlings in each hatch rank were not affected by the sex of siblings in other hatch ranks, the sex sequence of other siblings, nor any interactions between these variables and the focal nestling's own sex. Again, males in C position did not fare worse relative to females, regardless of the sex of their elders.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Across the kookaburra population, the sex ratio of eggs and fledglings did not differ from parity. However, there is strong evidence that the sex of nestling kookaburras is nonrandom, and that offspring sex may be facultatively manipulated in response to two major factors: the helper type of the group and the hatch rank of the nestling. Sex ratios for combinations of helper type and hatch rank varied widely, ranging from 16.7% for the second-hatched nestlings in pairs without helpers to 100% male for the first-hatched nestlings of groups with only female helpers. Comparably large biases have only rarely been reported for other birds, most notably peregrines (Olsen P, personal communication), eclectus parrots (Heinsohn et al., 1997
), and Seychelles warblers
(Komdeur et al., 1997).
Group composition and the sex ratio
Groups with female helpers, especially if all helpers were female, had
male-biased clutch and fledging sex ratios. In contrast, groups lacking female
helpers had female-biased clutch and fledging sex ratios
(Figure 1b). Paired comparisons
showed that individual females responded facultatively to increases (or
decreases) in the number of female helpers in their group by producing more
(or fewer) sons in their next clutch. Empirical examples of facultative biases
in sex allocation in response to group composition in cooperative breeders are
still rare. Some studies have demonstrated an enhancement effect, producing
the "helpful" sex when group size is small
(Komdeur et al., 1997;
Ligon and Ligon, 1990), but
this was not evident in kookaburras: although males are the more
"helpful" sex, unassisted pairs produced female-biased
clutches.
Female helpers, but not males, have a negative effect on the fledging
success of their group (Legge,
2000a). Groups with all female helpers fledge fewer young than any
other type of group, including groups with no helpers at all
(Figure 4). This is partly
because female helpers are poor contributors during breeding, but more complex
interactions between group members could also be involved
(Legge, 2000b). Breeding
females may attempt to limit the number of female helpers in their group to
avoid cumulative negative effects on their reproductive success, therefore
producing sons if they already have daughter-helpers. During the study, groups
seldom had more than one female helper (10/132 group-years), and helpers were
rarely all female (12/132 group-years).
If groups are constrained from producing daughters in some years, they may
produce them whenever they have the opportunity, in other words when female
helpers are not already present in the group. This could explain the
female-biased clutches of groups that lack female helpers. Curiously,
unassisted pairs who produce daughters risk becoming all-female groups the
following year, which would lead to reduced breeding success. However,
unassisted pairs live on the smallest territories of 16-50 ha
(Legge and Cockburn, 2000),
which may be unable to support additional birds. If resource competition is
important, it may pay unassisted pairs to produce the sex that disperses at an
earlier age. Although the data are limited, a comparison of dispersal behavior
of sons and daughters produced by pairs versus groups is suggestive.
Considering only those fledglings that were known to have survived to
independence at two months, females hatched to unassisted pairs were more
likely to leave their natal territory by nine months (60%, n = 10)
than males hatched into pairs (17%, n = 12;
21 = 4.6, p =.03). In contrast, the
difference in sex-specific dispersal rates from groups of three or more birds
was much less strong (31% of females leave, n = 26; 23% of males,
n = 40; 21 = 0.6, p =.44). Note
that increasing group size through the production of philopatric sons would
probably not lead to a corresponding acquisition of territory. Throughout the
study, a group's territory changed little in size regardless of variation in
group size (Legge S, unpublished data), probably because territory expansion
requires a simultaneous retraction in territory area of neighboring
groups.
Sex sequences and hatch rank
Nestling sex was strongly dependent on hatch rank. Overall, first-hatched
nestlings were predominantly male, and second-hatched nestlings were
predominantly female. The sex of third-hatched nestlings was unbiased, but
this was unsurprising. Nest productivity is largely determined by competitive
interactions between the oldest two nestlings and the third nestling is the
least likely to fledge. In many groups the third egg is only laid as insurance
against hatch failure of other eggs
(Legge, 2000c). A corollary of
the relationship between sex and hatch rank was that the distribution of sex
sequences was nonrandom. In particular, the FM sequence was rare for the first
and second nestlings, only occurring in five nests out of 73. This leads to
the question of whether sex sequence affects nestling growth and survival.
FM—the double disadvantage?
It is intuitively appealing that sex sequence in kookaburras—a
dimorphic, siblicidal bird—would affect sibling interactions and nest
productivity. Could the avoidance of risky sequences underlie the observed
relationship between hatch rank and sex? In reverse size dimorphic, siblicidal
birds, the male in the FM sequence may suffer a double disadvantage of sex and
hatch rank, resulting in elevated levels of siblicide and lower nest success
(Edwards and Collopy, 1983).
However, in kookaburras there was no evidence that the FM sequence was
detrimental. First, nests with the FM sequence (in AB, BC, or AC position) did
not fledge fewer young. Only one nest out of five with the taboo FM sequence
for A and B nestlings suffered brood reduction, yet the background rate of
siblicide is over 40% (Legge,
2000c). The remaining four nests fledged all their young, and
fledglings in each brood were above average weight for their hatch rank (Legge
S, unpublished data). Second, an exhaustive analysis revealed no convincing
evidence that the growth of either sex was affected by the number or sex of
its siblings, or that either sex was a relatively poor competitor. In
particular, B and C males did not suffer any additional disadvantage beyond
their hatch rank when hatched after a sister.
However, differences in growth and mortality with sex sequence may be
extremely small if breeding females only produce risky sequences when they can
"afford" to. For example, all five of the FM broods occurred in
groups with male helpers, and four occurred in groups of all-male helpers,
which are the most able to raise young (see
Figure 4). Also, although sex
sequence apparently had no impact on nestling growth, the data presented in
Figure 3 do hint that males may
be poorer competitors than females, since they appear to perform slightly
worse when hatched third. Finally, siblicide is particularly common in nests
where the AB sex sequence is MF (Legge,
2000c). Hatching a fast-growing female after a male may
destabilize the age-based dominance hierarchy, elevating levels of aggression
in the nest, and leading to the death of the relatively vulnerable C nestling
in the cross-fire. The preponderance of MF dyads in siblicidal nests implies
that the sex sequence of nestlings does affect competitive interactions.
Evidence from other species
Evidence for sex sequence effects on growth or mortality from other
dimorphic, facultatively siblicidal species is limited, and mixed. To our
knowledge, the only other detailed study (besides the one reported here)
assessing the effect of sex sequence on nestling growth concerns blue-footed
boobies (Drummond, 1991). Like kookaburras, they found that the growth of male
blue-footed boobies was not compromised by the presence of an older, and
eventually larger sister. Indeed, female booby nestlings were more likely to
die in times of food shortage (Torres and
Drummond, 1997), as predicted by a model suggesting that mortality
will be higher in the sex that devotes most energy to growth
(Clutton-Brock et al., 1985).
Unsurprisingly, no biases in sex sequence were found. A study of the growth of
golden eagles suggested male growth might be adversely affected if hatched
after a sister, but the sample was very small (3/5 males had an older sister;
Collopy, 1986). Golden eagle
fledgling sex ratios became relatively more female-biased in poor food years,
but there was no evidence males died from being outcompeted by sisters, or
that sex sequence was manipulated (Edwards
et al., 1988); see also
(Bortolotti, 1989). The best
support for the importance of sex sequence for nestling mortality comes from
bald eagles. Breeding females avoided the risky sequence (MF in this case)
when food was plentiful, but when food availability was low the risky sequence
was more prevalent, possibly to encourage efficient siblicide
(Bortolotti, 1986;
Dzus et al., 1996).
Similarly, evidence for the effects of sex sequence on sibling rivalry in
dimorphic but nonsiblicidal species is mixed. Several raptors have been
observed to produce a preponderance of females early in the laying sequence,
with males later (Dijkstra et al.,
1990; Leroux and Bretagnolle,
1996; Olsen and Cockburn,
1991; but see Bednarz and
Hayden, 1991). Although these sequence effects have been
interpreted in the light of other hypotheses, the fact remains that these
birds are producing sex sequences that should reduce nest productivity if the
double disadvantage was a significant problem. In studies of nestling growth
and/or mortality, the larger sex usually comes off worst when food is limited
(e.g., Cooch et al., 1997;
Griffiths, 1992;
Howe, 1976;
Roskaft and Slagsvold, 1985).
In contrast, female American kestrels were able to out-compete their smaller
brothers, but only under particular circumstances, when food was limited and
monopolizable (Anderson et al.,
1993). The ability of older nestlings to monopolize food may
explain variation in the consequences of sex sequence between species.
Kookaburra nestlings are brought highly monopolizable food—items are 2-4
cm long, and swallowed immediately by a single nestling. Nestlings that use
their age and/or sex advantage to control access to the premium feeding site
at the hollow entrance should be at a distinct advantage.
Why order the sexes of offspring?
Besides the effect of sex sequence on nestling interactions, we can suggest
one alternative explanation for ordering the sexes of offspring in a brood.
Although kookaburras usually lay three eggs, brood reduction claims the C and
sometimes the B nestling in nearly half of kookaburra nests between hatching
and fledging (Legge, 2000c),
and fledging success is generally low (1.4 fledglings per nest). Assuming all
three eggs hatch, breeding females can be fairly confident of fleding one
young, fledging two is possible but uncertain, and fledging three less likely.
In small clutches with variable brood reduction, females may match sex to
hatch ranks with a conditional strategy. That is, if she should only manage to
fledge the A nestling, what sex should it be? If she succeeds in fledging B as
well as A, what sex should B be, given the sex of A, and so on. The strength
of sex biases in each hatch rank should be aligned with how badly the breeder
"wants" a particular sex, and the probability she has of fledging
successively-hatched nestlings. For example, kookaburra groups of all-female
helpers have extremely male-biased clutches, indicating a general preference
for sons. They are also likely to produce just one fledgling. Unsurprisingly,
every first-hatched nestling from broods of all-female groups was a male. In
addition, groups with all-male helpers were the most likely to fledge the C
nestling, and these were the only groups that appeared to control the sex of
the C nestling.
Heterogeneity between groups with different helper types
Because clutch sex ratios varied among groups of different helper type,
there was heterogeneity in the hatch rank sex ratios (and therefore the
distribution of sex sequences) between helper types. For the AB
sequence, one strong pattern was superimposed on the general avoidance of FM
sequence. In contrast to groups with helpers, unassisted pairs also avoided
the MM sequence, nearly always hatching a female second
(Table 1a). Siblicide is more
common in broods of unassisted pairs, and the MF sequence is part of a suite
of variables that characterize siblicidal nests
(Legge, 2000c). Thus, it is
unsurprising that MF dyads are common in broods of unassisted pairs. But this
does not explain why FF is common, and MM rare. It may simply be an artifact
of the general female-biased clutches of unassisted pairs, aiming to produce
dispersive daughters rather than philopatric sons. The MM sequence may be rare
because it would result in a clutch sex ratio of at least 67%. Although the
sample for all-female groups is too low to draw firm conclusions, they may
avoid FF for similar reasons: a clutch with a FF sequence is automatically
female-biased, but all-female groups prefer to produce male-biased
clutches.
Another strong pattern emerges from the heterogeneous distribution of BC sequences. The FF sequence is overrepresented in all-male groups, meaning C eggs are more likely to be female. The C egg has the highest chance of fledging in all-male groups, perhaps explaining why control of offspring sex extends further down the sequence than for other groups. Similarly, groups with helpers of both sex, which generally produce male-biased clutches, may avoid the FF sequence, thus producing more males in third position. After all-male groups, groups with both sexes helping are the next most likely to fledge the C egg. However, the sample for this helper type is rather small.
Conclusions
The sex of kookaburra nestlings may be influenced facultatively by mothers
in response to several selective pressures, demonstrating the potential
complexity of sex ratio variation (see
Cockburn, 1994 for a similarly
complex example in mammals). Enhancement or repayment models at the level of
individual families (Komdeur et al.,
1997) do not explain the observed patterns. Local resource
competition may be important in some contexts (i.e., in unassisted pairs).
However, the make-up of the social group had a strong effect—breeding
females avoid overburdening themselves with daughter-helpers by producing sons
almost exclusively if daughter-helpers are already present in their group.
Finally, there were very strong hatch rank effects, suggesting that the
sequence of sexes in the brood affects the quality or quantity of young.
Although there was little supporting evidence for this, experimental
manipulations are required to create the risky sequence in nests that would
otherwise avoid it.
|
ACKNOWLEDGEMENTS |
|---|
Penny Olsen made helpful comments on the manuscript. This research was funded by Australian National University (A.N.U.) and Overseas Postgraduate Research Scholarships to S.L. and a large Australian Research Council grant to A.C., with permits from the A.N.U. Animal Experimentation Ethics Committee (permit no: F.BTZ.46.94), Canberra Nature Park, the Australian Bird and Bat Banding Scheme and the Australian National Botanic Gardens, all of whom were extremely helpful.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Anderson DJ, Budde C, Apanius V, Martinez Gomez JE, Bird DM, Weathers WW, 1993. Prey size influences female competitive dominance in nestling American kestrels (Falco sparverius). Ecology 74: 367-376.
Bednarz JC, Hayden TJ, 1991. Skewed brood sex ratio and sex-biased hatching sequence in Harris's hawks. Am Nat 137: 116-132.
Bortolotti GR, 1986. Influence of sibling competition on nestling sex ratios of sexually dimorphic birds. Am Nat 127: 495-507.[ISI]
Bortolotti GR, 1989. Sex ratios of fledgling golden eagles. Auk 106: 520-521.
Clark AB, 1978. Sex ratio and the local resource
competition in a prosimian primate. Science
201: 163-165.
Clutton-Brock TH, Albon SD, Guiness F, 1985. Parental investment and sex differences in juvenile mortality in birds and mammals. Nature 313: 131-133.
Cockburn A, 1994. Adaptive sex allocation by brood reduction in antechinuses. Behav Ecol Sociobiol 35: 53-62.
Collopy MW, 1986. Food consumption and growth energetics of nestling golden eagles. Wilson Bull 98: 445-458.
Cooch E, Lank D, Robertson R, Cooke F, 1997. Effects of parental age and environmental change on offspring sex ratio in a precocial bird. J Anim Ecol 66: 189-202.
Dijkstra C, Daan S, Buker JB, 1990. Adaptive seasonal variation in the sex ratio of kestrel broods. Funct Ecol 4: 143-47.
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.
Dzus EH, Bortolotti GR, Gerrard JM, 1996. Does sex-biased hatching order in bald eagles vary with food resources? Ecoscience 3: 252-258.
Edwards TC, Collopy MW, 1983. Obligate and facultative brood reduction in eagles: an examination of factors that influence fratricide. Auk 100: 630-635.
Edwards TC, Collopy MW, Steenhof K, Kochert MN, 1988. Sex ratios of fledgling golden eagles. Auk 105: 793-796.
Emlen ST, Emlen JM, Levin SA, 1986. Sex-ratio selection in species with helpers-at-the-nest. Am Nat 127: 1-8.
Fisher RA, 1930. The genetical theory of natural selection. Oxford: Oxford University Press.
Frank SA, 1990. Sex allocation theory for birds and mammals. Ann Rev Ecol Syst 21: 13-55.[ISI]
Genstat Committee, 1993. Genstat 5 Release 3 reference manual. Oxford: Clarendon Press.
Gowaty PA, Lennartz MR, 1985. Sex ratios of nestling and fledgling red-cockaded woodpeckers (Picoides borealis) favour males. Am Nat 126: 347-353.
Griffiths R, 1992. Sex-biased mortality in the lesser black-backed gull Larus fuscus during the nestling stage. Ibis 134: 237-244.
Griffiths R, Double MC, Orr K, Dawson RJG, 1998. A simple DNA test to sex most birds. Mol Ecol 7: 1071-1075.[Medline]
Hardy ICW, 1997. Possible factors influencing vertebrate sex ratios: an introductory overview. Appl Anim Behav Sci 51: 217-241.
Heinsohn R, Legge S, Barry S, 1997. Extreme bias in sex allocation in eclectus parrots. Proc R Soc Lond B: Biol Sci 264: 1325-1329.
Howe HF, 1976. Egg size, hatching asynchrony, sex and brood reduction in the common grackle. Ecology 57: 1195-1207.[ISI]
Hoyt DF, 1979. Practical methods for estimating the volume and fresh weight of bird eggs. Auk 96: 73-77.[ISI]
Koenig WD, Walters JR, 1999. Sex-ratio selection in species with helpers at the nest: the repayment model revisited. Am Nat 153: 124-130.
Komdeur J, Daan S, Tinbergen J, Mateman C, 1997. Extreme adaptive modification in sex ratio of the Seychelles warbler's eggs. Nature 385: 522-525.
Krackow S, 1999. Avian sex ratio distortions: the myth of maternal control. In: Proceedings 22nd international ornithological congress, Durban, 1999 (Adams NJ, Slotow RH, eds). Johannesburg: Birdlife South Africa.
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.
Leroux A, Bretagnolle V, 1996. Sex ratio variations in broods of Montagu's harriers Circus pygargus. J Avian Biol 27: 63-69.
Lessells CM, Avery MI, 1987. Sex-ratio selection in species with helpers at the nest: some extensions of the repayment model. Am Nat 129: 610-20.
Ligon JD, Ligon SH, 1990. Female-biased sex ratio at hatching in the green woodhoopoe. Auk 107: 765-771.
Malcolm JR, Marten K, 1982. Natural selection and the communal rearing of pups in African wild dogs (Lycaon pictus). Behav Ecol Sociobiol 10: 1-13.
Olsen PD, Cockburn A, 1991. Female-biased sex allocation in peregrine falcons and other raptors. Behav Ecol Sociobiol 28: 417-423.
Parry V, 1973. The auxiliary social system and its effect on territory and breeding in kookaburras. Emu 73: 81-100.
Ricklefs RE, 1971. Avian post-natal growth. In: Avian biology (Farner DS, King JR, eds). New York: Academic Press; 1-83.
Roskaft E, Slagsvold T, 1985. Differential mortality of male and female offspring in experimentally manipulated broods of the rook. J Anim Ecol 54: 261-266.
Sheldon BC, 1998. Recent studies of avian sex ratios. Heredity 80: 397-402.
Torres R, Drummond H, 1997. Female-biased mortality in nestlings of a bird with dimorphism. J Anim Ecol 66: 859-865.
Walters JR, 1990. Red-cockaded woodpeckers: a `primitive' cooperative breeder. In: Cooperative breeding in birds: long-term studies of ecology and behaviour (Stacey PB, Koenig WD, eds). Cambridge: Cambridge University Press; 67-102.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. Abe, Y. Kamimura, and M. Shimada Sex ratio schedules in a dynamic game: the effect of competitive asymmetry by male emergence order Behav. Ecol., November 1, 2007; 18(6): 1106 - 1115. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Korsten, C. M. Lessells, A. C. Mateman, M. van der Velde, and J. Komdeur Primary sex ratio adjustment to experimentally reduced male UV attractiveness in blue tits Behav. Ecol., July 1, 2006; 17(4): 539 - 546. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Allaine Sex ratio variation in the cooperatively breeding alpine marmot Marmota marmota Behav. Ecol., November 1, 2004; 15(6): 997 - 1002. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. L. F. Martins Sex-specific growth rates in zebra finch nestlings: a possible mechanism for sex ratio adjustment Behav. Ecol., January 1, 2004; 15(1): 174 - 180. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. G. Ewen, R. H. Crozier, P. Cassey, T. Ward-Smith, J. N. Painter, R. J. Robertson, D. A. Jones, and M. F. Clarke Facultative control of offspring sex in the cooperatively breeding bell miner, Manorina melanophrys Behav. Ecol., March 1, 2003; 14(2): 157 - 164. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Nathan, S. Legge, and A. Cockburn Nestling aggression in broods of a siblicidal kingfisher, the laughing kookaburra Behav. Ecol., November 1, 2001; 12(6): 716 - 725. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||