Behavioral Ecology
doi:10.1093/beheco/ars030
Advance Access publication 14 March 2012
Original Article
Nice guys finish last: same-sex sexual behavior
and pairing success in male budgerigars
Puya Abbassi and Nancy Tyler Burley
Department of Ecology and Evolutionary Biology, 321 Steinhaus Hall, University of California, Irvine, CA
92697-2525, USA
In budgerigars (Melopsittacus undulatus), a social parrot in which females are socially dominant, males of all ages engage in a set of
behaviors with other males that closely resembles the repertoire used in heterosexual courtship. One adaptive hypothesis for this
tendency, the ‘‘courtship practice hypothesis,’’ posits that males with greater experience in same-sex activities develop superior
skills that increase their courtship success with females. To test this hypothesis, we measured individual variation in tendency of
subadult males to engage in such behaviors and asked whether birds’ relative participation in same-sex activities predicted adult
pairing success in competitive pairing trials conducted after birds reached adulthood. Males’ tendency to participate in same-sex
activities was correlated across samples collected a month apart. Directly contrary to the courtship practice hypothesis, however,
males that participated in same-sex interactions more often were less likely to obtain mates. Pairing success was also predicted by
2 sexually dimorphic facial traits, namely iris prominence and cere color. We hypothesize that the tendency of males to engage in
same-sex interactions may represent a mechanism of mutual assessment of male abilities, especially those involved in group
foraging efforts (the ‘‘leader assessment hypothesis’’) and suggest that increased investigation of allopreening and other seemingly affiliative behaviors that occur outside the heterosexual pair bond may advance understanding of social organization in
flocking birds. Key words: courtship practice hypothesis, iris prominence, leadership assessment hypothesis, parrots, same-sex
sexual behavior, social foraging. [Behav Ecol]
INTRODUCTION
S
ame-sex sexual behavior consists of actions, directed by an
individual toward a conspecific of the same sex, which
would be classified as courtship or pairing behaviors if displayed toward an opposite-sex conspecific (Bailey and Zuk
2009). While such same-sex interactions have been documented in a wide range of species, including over 90 species
of birds (MacFarlane et al. 2010), their causes, functions, and
consequences have been little studied by evolutionary biologists, perhaps as a result of the tendency to view such behaviors
as mistakes (Bagemihl 1999). Nevertheless, several categories of
adaptive hypotheses have been posed that may explain many of
the more striking examples of well-documented same-sex sexual behavior (reviewed in Bailey and Zuk 2009). ‘‘Social glue’’
hypotheses posit that same-sex interactions, promote alliances,
or reduce tensions that develop in group-living species. ‘‘Intrasexual conflict’’ hypotheses predict that same-sex interactions
reduce mating success of competitors or function to establish
dominance hierarchies. ‘‘Courtship practice’’ hypotheses posit
that same-sex activities improve courtship or copulatory skills of
their participants. Even where same-sex activities do not confer
a reproductive benefit to individuals that initiate them, expression of such behaviors may result from constraints emerging
from evolved genetic mechanisms such as overdominance
(Gavrilets and Rice 2006), sexually antagonistic selection
(Camperio-Ciani et al. 2004), or common genetic bases underlying tendencies to court individuals of the same and opposite
sex (Logue et al. 2009).
Address correspondence to N.T. Burley. E-mail: [email protected].
Received 4 July 2011; revised 12 February 2012; accepted 13
February 2012.
The Author
Author 2012.
2012. Published
Published by
by Oxford
Oxford University
UniversityPress
Presson
onbehalf
behalfofof
© The
the
Society for
for Behavioral
Behavioral Ecology.
Ecology. All
All rights
rights reserved.
reserved.
the International
International Society
For
permissions,
please
e-mail:
[email protected]
For permissions, please e-mail: [email protected]
Here, we report results of an inquiry into the possible function of the tendency of subadult male budgerigars to interact
with one another using behaviors that are major components
of the heterosexual courtship repertoire of adult males. An important step toward understanding the likely significance of
such behaviors involves ascertaining whether a trait under
study has variable expression within a population and, if so,
whether trait expression displayed by individuals is consistent
over some period of time (Bell et al. 2009 and references
therein). Accordingly, we ask here whether individuals show
consistent differences in tendency to engage in behaviors with
same-sex individuals. We also investigate whether relative tendency to engage in same-sex sexual behaviors as subadults
predicts, along with several other phenotypic attributes, the
ability of adult males to acquire mates when mating opportunities are constrained by limited availability of females.
Consideration of the social behavior of budgerigars, briefly
summarized below, led us to expect that ‘‘courtship practice’’ is
the most promising hypothesis to explain apparently affiliative
same-sex activities observed (Stamps et al. 1990) in captive
colonies of budgerigars. Previous researchers have focused
on the possibility that repetition of motor functions involved
in courtship and copulation during same-sex practice results
in polished motor skills that are attractive to females (e.g.,
Drosophila—McRobert and Tompkins 1988 and bison—
Vervaecke and Roden 2006). Another possibility is that practice with same-sex conspecifics helps individuals overcome any
tendency to avoid making courtship overtures to potential
mates, perhaps through fear of approaching females. We term
this the ‘‘confidence’’ variant of the courtship hypothesis, in
keeping with literature recognizing fearfulness/confidence
as a common behavioral axis in nonhuman animals (e.g.,
Gosling 2001). Both variants of the courtship practice are
considered in our investigation.
776
2
Behavioral Ecology
Social behavior of budgerigars
Budgerigars (Melopsittacus undulatus) are small Australian parrots that live in flocks of variable size. Although they are not
well studied in nature due to their unpredictable movements
over long distances, the behavior repertoire found in nature
(Wyndham 1980, 1981) approximates that observed in laboratory settings (see below). Like most parrots whose habits are
known, budgerigars typically display social monogamy and biparental care (Forshaw 2002), and, at least in captivity, their
pair bonds often endure across multiple reproductive attempts (Brockway 1964a). Females have primary responsibility
for obtaining and defending nesting cavities, incubation, and
brooding. Males regularly allofeed their nesting mate and
sometimes directly feed their dependent young (Wyndham
1980; Stamps et al. 1985).
In captive social groups, adult females regularly dominate
males (Masure and Allee 1934; Hile et al. 2005) and are much
more aggressive to conspecifics of both sexes than males
(Brockway 1964a; Hile et al. 2005). In contrast to females,
adult males are quite gregarious (Brockway 1964a, 1964b;
Stamps et al. 1990) and frequently engage in behaviors with
one another that mirror those occurring during heterosexual
courtship (Table 1). In nature, males congregate in flocks
while females are nesting (Wyndham 1980). Recently fledged
offspring of both sexes affiliate with their father and/or a
clutch-mate of the same or opposite sex (Stamps et al. 1990).
However, affiliative tendencies of females begin to wane well
before sexual maturity, whereas males quickly extend their interactions to include nonkin males. That young males do not
typically court females suggests that these male–male interactions do not constitute courtship ‘‘mistakes’’ (failure to discriminate the sexes). Also, aggregations of subadults are common in
nature (Wyndham 1981), so the opportunity for such interactions is not an artifact of captive housing conditions.
In sum, female budgerigars are much less sociable than
males and, at least as adults, are socially dominant. Young
males, as well as adult males, have extensive social contact with
same-sex individuals and may experience limited opportunity
to gain courtship experience with the less sociable females.
These observations formed the basis for our decision to test
the courtship practice hypothesis as an explanation for the tendency of young male budgerigars to participate in activities that
closely resemble heterosexual, adult courtship in this species.
Experimental approach
We conducted competitive mating trials in which we released
groups of adult males—whose subadult tendencies to participate in same-sex sexual behaviors tendencies had been
quantified—into an aviary containing a smaller number of
unmated adult females. The test population had a male-biased
sex ratio to ensure that some males would remain unmated.
Competitive mating trials were chosen over more commonly
used dyadic mate choice trials because the latter severely limit
physical contact between chooser and stimuli that might provide choosers with information about important aspects of the
stimuli’s courtship abilities gained through same-sex practice.
One undesirable aspect of competitive mating trials is that,
while they demonstrate relative mate-getting ability, interpretation of results can be complicated by the tendency of individuals of the more numerous sex to compete among themselves,
raising the question of the extent to which pairing patterns result from intrasexual agonistic interactions or intersexual selection. The strong dominance of females in this species, however,
assures that they have ‘‘the final say’’ in choice of social mate.
Another limitation of competitive mating trials is that participants in such trials necessarily vary in numerous traits that may
influence their mate-getting ability. Accordingly, we included
several other phenotypic traits of adult males in analyses to
evaluate the outcome of competitive mating trials.
We created 2 indices to quantify same-sex interactions among
subadults. The ‘‘participation index’’ reflects each male’s relative tendency to engage in same-sex sexual behaviors; the ‘‘initiation index’’ reflects the proportion of all interactions in which
an individual participated that he initiated. We predicted that, if
the motor skills variant of the practice hypothesis is the primary
function of same-sex activities, males with high scores on the participation index would be most successful in obtaining mates
during competitive mating trials. If practice serves to build male
confidence in initiating courtship overtures, we expected that
males successful in obtaining mates would have obtained high
scores on both indices; that is, they would have participated
frequently in same-sex sexual behaviors and tended to be initiators of such interactions.
MATERIALS AND METHODS
Birds
Twenty male and 23 female budgerigars participated in these
experiments. Birds were either bred in our lab or purchased as
juveniles from a wholesale distributor. Only birds that
appeared to be in excellent health were included in the study.
Males were 3–4 months old when introduced to Experiment 1
and 13–19 months old at the beginning of Experiment 2;
females were 12–18 months at the beginning of Experiment 2.
Age was determined from banding records (lab-reared birds)
or by iris color and forehead plumage (for birds purchased
as juveniles—Forshaw 1981). Sex was determined by cere color,
Table 1
Behaviors used in budgerigar courtship and during male–male affiliative interactions
a
b
Behavior
Description
Allopreen
Attempt to allopreen
Head bob
Beak rub
Beak flick
Allofeed
Copulate attempta
Copulateb
One bird preens another
One bird attempts to preen another, but the object rebuffs the actor or moves away
A bird moves its head up and down rapidly in a circular motion (often appears simply vertical in real time)
One bird rubs its beak up and down rapidly against that of another bird
A bird flicks its beak against that of another in a side-to-side motion
A bird appears to regurgitate food to another
One bird attempts to mount a second bird, but the second bird moves away
One bird mounts another for at least 5 s, with apparent cloacal contact
This ethogram was based on papers by Brockway (1964a, 1964b), Hile et al. (2005), and preliminary observations for experiments reported here.
Rarely observed during male–male interactions in current experiments.
Observed only during heterosexual interactions in Experiment 2.
777
3
Abbassi and Burley • Nice guys finish last
which in birds that are 3 months old is purplish-pink in males
and pink or tan in females; also, the nares of females of this age
tend to have white outlines. In adult birds, the cere is blue in
males and reddish brown in breeding females (Brockway
1964a). All males in this study achieved bright blue ceres no
later than 5 months of age. In both experiments, birds wore
unique combinations of plastic color bands to facilitate individual recognition.
We minimized the number of sibling relationships among
birds by selecting birds from different families (lab-reared
birds) whenever possible, and by purchasing young birds from
the wholesale distributor at different times. (Our distributor
turns over approximately 500 budgerigars per week.) Among
male test subjects, there were initially 2 sets of lab-reared sibs,
but one of these birds died during the study; no males were
known to be first cousins. All birds lacked reproductive experience prior to the start of the second experiment, and kin of
the opposite sex were not given the opportunity to pair.
Early in Experiment 1, we realized that 1 bird had been
sexed incorrectly. To maintain the cohort size, we replaced this
bird with the only same-age male available that had been
reared under the same conditions. This male, which had blue
plumage, participated in the remainder of the study. His behavior scores were adjusted to reflect the shortened duration of his
participation. Other birds in the study had green wild-type
plumage.
Housing
Experiments were conducted in an indoor (2.8 m 3 4.4 m 3
2.9 m) flight, which contained numerous perch sites. Most
perches consisted of natural branches that had been stripped
of their leaves; some branches were hung horizontally, whereas
others were potted upright in buckets of cement. In Experiment 1, short dowels were mounted on the walls of the flight
to allow birds to perch alone. In Experiment 2, nest boxes were
provided; short dowels at the entrances to these also permitted
birds to sit alone. The wide variety of perches allowed birds to
interact in the full range of social behaviors described for this
species, but also to avoid social contact whenever they chose to
do so.
A commercial budgerigar seed mix, as well as water, cuttlebone, and oyster shell were supplied ad libitum; additional supplements were provided regularly. Seed was provided in large
trays that allowed 6–8 birds to feed at the same time. Seed was
refreshed and water changed daily. All food resources and water were provided at ground level.
Lighting, maintained on a 14:10 light:dark cycle, was provided by full-spectrum fluorescent bulbs. Temperature in the
flight varied between 20 �C (nighttime) and 27 �C (daytime).
Observations
Observations on birds in the experimental flight were conducted from an adjacent room through a large one-way mirror.
All perching sites, nest boxes (Experiment 2 only) and resources were visible to observers. During each observation session,
a group of 2–4 trained observers applied all-accounts sampling
procedures (Altmann 1974) to score behaviors. Overtly agonistic behaviors were uncommon among males in Experiment
1. In Experiment 2, agonistic interactions were more common
but brief and identity of participants difficult to ascertain.
Accordingly, we limited data analysis to behaviors that are
considered within the heterosexual courtship repertoire listed
in Table 1. Song and call traits, also used in courtship activities, are very fluid in this species (Moravec et al. 2010) and
could not be studied using this methodology.
Experiment 1: same-sex sexual behaviors
To quantify individual variation in tendency of male budgerigars to participate in and initiate same-sex sexual behaviors,
young males were released into the flight and observations
of their behavior were recorded. Males were divided into
2 sets, based on age. Set A (mean age at start = 3.25 months)
consisted of 8 birds and Set B (mean age = 3.33) had 12 birds.
In order to assess consistency of individual behavior, each set
was sampled twice; the second sample was collected a month
after the first. At the beginning of each sampling interval, birds
to be observed were released into the flight simultaneously.
Observations started the next morning. Typically, 2 observers
called out identities of interactants and behaviors as they occurred, and a third person recorded the action in live time.
Although all observations were made during morning hours
when budgerigars are most active, activity levels varied greatly
among days. On sampled days with low activity, observations
were suspended after 1.5 hours; on high-activity days, birds
were observed for 3 hours. (Total observation hours: Set
A—18 hours; Set B—21 hours.)
Behaviors were recorded in bouts, with a bout being defined
as a single behavior from one male directed toward another
male (e.g., a bout of allopreening) at a single location. Bouts
were scored as having ended when there was a lapse of 2 min or
more between interactions; when a third bird interrupted an
interaction; when interactants moved to a new location in the
flight; and/or when interactants began to engage in a different
behavior. The number of behaviors each bird engaged in with
every other bird was totaled for each sampling interval (1 and
2) and constituted the bird’s total number of interactions for
that interval.
In Set A, a few birds tended to form a relationship with a
single individual (‘‘buddy’’), such that most same-sex interactions involved the same partner. To determine whether participation and initiation tendencies would be similar across
sampling intervals if flock composition prevented birds that
had been buddies in the first sample from interacting in the
second, the birds of Set B were split into 2 groups of 6 for the
second observation sample, and birds described as buddies in
the first sample were assigned to different groups in the second sample.
During the interval between samples, test males were kept in
large cages in groups of 2–3 birds. Any birds deemed buddies
were separated at this time.
Participation index and rank
We modified parameters H# and E of the Shannon–Weiner index (Shannon 1948) to quantify 2 components of same-sex
interactions. This approach was adopted to standardize measures of interaction for flocks of varying size. Separate calculations of H# and E were made for each individual male
during each sampling interval. The diversity index (H#) reflects
the relative number of flock members with which a given individual participated in same-sex behaviors during 1 sample.
The index was calculated as
H# ¼ 2
s
X
i¼1
pi ðln pi Þ
where pi equals the number of interactions that a focal budgerigar had with a particular individual bird (ni) divided by
the focal bird’s total number of interactions during a single
sampling interval (N), and S equals the total number of individuals with which the bird could have potentially interacted
(set size minus one). Thus, a bird obtaining a higher value of
H# interacted with a higher proportion of others in its cohort
778
4
and did so more evenly across those with which it interacted,
compared with a bird having a lower H# score.
To standardize this value relative to the number of birds in
a cohort, parameter E (‘‘evenness’’) was calculated as:
E ¼ H # Hmax
where Hmax ¼ lnS. A bird with a higher E value tended to
distribute its interactions more evenly among all potential
partners.
Participation scores for individuals in each sample were calculated as (H# 1 2E) to achieve equal weighting of parameters
because the mean E score was approximately half the value of
the mean H# score. Finally each male’s participation scores
were summed across samples to compute his participation
index. For the purpose of regrouping males for the second
experiment, ranks were assigned to birds on the basis of participation index. The rank of 1 was assigned to the male with
the lowest index.
Initiation index
The participation index does not discriminate between tendencies to initiate interactions versus respond to overtures
of another bird. To obtain a simple comparative measure of
these relative tendencies that is independent of cohort size,
we calculated the percentage of all interactions in which an individual participated that he initiated in each sample, and averaged these 2 values across samples to compute the initiation
index. Individuals with low initiation indices were more often
recipients than originators of interactions.
Experiment 2: competitive pair formation
This experiment was performed to determine whether a male’s
relative tendency to engage in same-sex behaviors as a subadult
predicted his mate-getting ability in adulthood. Prior to the
start of the experiment, males were regrouped into 2 new sets
based on the rank of their participation index. One set consisted of all males with odd ranks, the other included males with
even ranks (Table 2). The 2 sets were tested sequentially.
Before the first set was tested, the observation flight was fitted with 12 nest boxes, and a group of 7 adult females was released into the flight and given time to select nest boxes. After
all females had selected nest boxes (about 2 weeks), the set of
10 males was released into the flight, and the first pair-formation
trial commenced. When this trial was complete, the second trial
was conducted with the even-ranked males. One of the evenranked males died prior to the start of this trial, so the 9 remaining males were tested with a new group of 7 females.
A team of observers conducted all-accounts sampling of courtship behaviors and copulation attempts several times per week,
in 2-h sessions, during each pair formation trial. Male and female visitations to nests were also scored. In the heterosexual
environment, same-sex sexual behaviors (Table 1) continued
to occur among males, but their frequency was not scored due
to sampling limitations.
Each trial continued until all females in each group had
formed a stable relationship with a male and had laid a fertile
clutch, as determined by regular inspection of nest contents.
A ‘‘stable relationship’’ was considered to be present when a female was observed to invite copulations exclusively from a single
male during the egg-laying interval and the week preceding it.
Observations conducted throughout the nesting phase confirmed that males consistently attended their mate and their
mate’s nest throughout development. Specifically, males
joined their mate whenever she emerged from the nest, typically feeding her and engaging in allopreening and (during
the egg-laying period) copulation. Most males inspected their
Behavioral Ecology
Table 2
Rankings of males based on Participation index
(Experiment 1) and mating status at the end of Experiment 2
Rank
Male ID
Set
Participation index
(Experiment 1)
Mating status
(Experiment 2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
SW
OH
WW
GG
SB
BB
WS
OB
PW
SR
OS
WB
BS
HW
OO
SO
KS
BO
PR
BW
A
B
B
B
A
B
A
B
B
A
A
B
A
B
B
A
A
B
B
B
0
0.69
1.01
1.10
1.41
2.73
3.87
3.89
4.09
4.28
4.36
4.36
4.58
4.65
4.92
5.02
5.23
5.31
5.75
6.25
Mated
Mated
Unmated
Mated
Mated
Unmated
Mated
Mated
Mated
Unmated
Unmated
Unmated
Mated
(Died)
Unmated
Mated
Unmated
Mated
Unmated
Unmated
Rank was used to select birds for Experiment 2 sets: birds with odd
ranks comprised one set, those with even ranks the other. SB and SO
were full sibs, as were HW and BW. GG had blue plumage. Bold face
denotes males that obtained 2 mates.
mate’s nest during egg laying, and all fed their incubating/
brooding mate on the nest.
Follow-up on unsuccessful males
To ascertain whether males that failed to obtain mates were
able/willing to form heterosexual bonds, the subset of males
that did not pair with a female (n = 9) in Experiment 2 was
retested with a novel set of females (n = 9). Courtship behaviors and copulation attempts were measured as described
above. Observations ended after all remaining test males
had heterosexually paired.
Additional phenotype measures
Prior to the start of Experiment 2, male cere color and body
mass were scored. These measurements were repeated after
the end of the experiment (approximately 7 months later).
At this time, the size of the yellow cap, wing length, and iris
prominence were also measured.
Cere color was assessed using the Munsell Book of Color.
Measurements were taken of hue, value (relative lightness/
darkness of color), and chroma (brightness or saturation) of
the central area of the cere just below the nares.
Cap size was estimated using a flexible tape measure to determine the distance (millimeter) from the top of the cere,
measured at the midpoint between the eyes, to the end of
the unbarred yellow feathers at the top of the head. Based
on this measure, Wyndham (1981) reported that cap size of
wild budgerigars is mildly sexually dimorphic. We measured
a cohort of same-aged females to determine whether this was
also true of our birds.
Body mass was measured using an electronic balance; condition was calculated as the residual of the linear regression
of body mass (averaged across 2 measurements) on the cube
of wing length (Schulte-Hostedde et al. 2005).
Iris prominence was rated on a 4-point scale, where 1 = least
prominent and 4 = most prominent, irides. Prominence
779
5
Abbassi and Burley • Nice guys finish last
reflects a combination of iris size and coloration: birds assigned
a prominence score of 4 had large irides that were nearly white
and contrasted strongly with their small pupils; less prominent
irides were smaller and/or had slightly grayish or tan colors
(Figure 1). Birds that were judged intermediate between 2
categories were assigned average values (i.e., 2.5 and 3.5).
Measurements were made on birds in a relaxed state under
uniform illumination by 2 observers.
Birds were scored for iris prominence on 3 occasions; there
was a 1-week span between the first 2 measurements and a
6-week span between the second and third measurements. Repeatability of iris prominence was assessed across this span.
Data analyses
Prior to analyses, data distributions for each variable under
study were inspected and, where necessary, subjected to a suitable transformation to improve conformity to the normal
distribution; the most appropriate method of transformation
was identified by executing the STATA ‘‘ladder’’ command.
Where multiple measures of physical traits were made, analyses relied on means across all measurements. Data for the
male that died after the completion of Experiment 1 was
excluded from analyses.
Pearson tests were used to determine whether birds’ tendencies to participate in and/or initiate same-sex activities were
correlated across samples and to assess whether variables entered in the final regression model were correlated.
Repeatability (r) of iris prominence was calculated as:
r ¼ S2 A S2 1 S2 A
where S2 (within-group variance) = the mean square within
groups (MSw), S2A (between-group variance) = (MSA 2
MSW)/n0; MSA is the mean square between groups, and n0
is the number of observations per individual (Sokal and Rohlf
1981).
Sex differences in iris prominence scores and cap sizes were
assessed using unequal variance t-tests (Ruxton 2006).
Prior to pooling behavior data across the samples in Experiment 1 (to calculate participation and initiation indices), we
performed REML linear mixed models to assess whether scores
differed by set (A or B), sample (first or second), or their interaction. Male identity was included as a level 2 variable in
these analyses.
A series of reverse stepwise logistic regression analyses
(P-to-remove = 0.15) was performed to assess whether male
mating status (0 = unmated; 1 = mated) achieved during Experiment 2 was predicted by aspects of the birds’ adult physical phenotype (round 1) or subadult behavior (round 2).
Variables that survived elimination in the first 2 rounds were
entered in a third round, and the stepwise procedure repeated. Multiple rounds were employed because the statistical
application (STATA 9.2) was unable to estimate the full set of
coefficients and their corresponding P values when all variables were entered simultaneously, making it impossible to initiate the stepwise procedure.
The first round focused on identification of physical traits
that influenced mating status. Included variables were: age
at the beginning of participation in Experiment 2; mass, condition; cere color (hue, value, and chroma); cap size; and iris
prominence. At the conclusion of the reverse stepwise procedure, each variable that had been dropped was individually
reentered with the set of ‘‘successful’’ variables to make sure
that its failure to be included in the model was not the result
of the presence of extraneous variables entered in the model at
the time the variable was originally excluded (e.g., Byers 2007).
Interaction effects among variables that survived the round
were included in the stepwise procedure.
A second logistic procedure, identical in outline to that used
above, was performed to evaluate the contribution of behavior
variables. Variables entered into this model included: set (A or B)
in which bird participated in Experiment 1; participation index; initiation index; and—to consider the possibility that intermediate affiliation indices are most favorable for pairing
success—the absolute value of the deviation from the mean affiliation index.
In a third round, variables that survived the 2 previous
rounds were combined, and the stepwise procedure repeated
to produce the final model. For the sake of brevity, only the
final model is reported here. All results reported here include
the blue-plumaged male, as the models generated with and
without his inclusion were very similar.
RESULTS
Experiment 1: same-sex sexual behaviors
The range of same-sex sexual behaviors observed among subadult males held in a single-sex environment was qualitatively
identical to courtship behaviors previously observed in mixedsex flocks, with the notable exception of copulatory behaviors
(Table 1). No heterogeneity between sets (A vs. B), sampling
intervals (1 vs. 2), or their interaction was identified for either
behavior variable (participation or initiation), so data were
pooled across sets and samples. Participation scores were positively correlated between sampling periods (Pearson r = 0.68,
N = 19, P = 0.001; Figure 2); initiation scores for the 2 samples
were not significantly correlated (r = 0.32, N = 19, P = 0.12).
A bird’s age at the beginning of the experiment was not
correlated either with its participation index or initiation
index (P’s . 0.8).
Experiment 2: Competitive pair formation
In total, 10 of the 19 surviving males paired during this experiment. Four of the males mated with 2 females each (Table 3).
All males that were not successful in establishing a heterosexual
pair bond during Experiment 2 obtained a mate and began to
reproduce during the subsequent follow-up period.
Iris prominence repeatability
Iris prominence scores (Figure 1) of males showed high repeatability of the 3 measurements taken over a span of 7 weeks
(measurement repeatability r = 0.72, df = 17, P , 0.0001).
Figure 1
Representative eye prominence scores: 1 (A), 2 (B), 3 (C), and 3.5
(D). All birds illustrated are males.
Traits that predicted pairing success
In the final round of the logistic regression procedure, we considered all variables that had been successful in the previous
2 rounds (for physical traits: iris prominence, cere color,
and cap score; for behavioral traits: participation index and
780
6
Behavioral Ecology
Table 4
Means and standard errors (SE) of traits that entered final logistic
model
Mated males
Figure 2
Correlation between participation scores for males across the
2 sampling intervals of Experiment 1. Pearson’s r = 0.68, N = 19,
P = 0.001.
initiation index). The final best model contained 3 variables:
males with low participation indices, high iris prominence
scores, and darker ceres tended to pair (Tables 3 and 4).
There were no significant correlations among these variables
and no significant interaction effects in the final model.
Sex differences in iris prominence and cap size
Iris prominence scores of females were lower than those of
males (Figure 3). Cap size was smaller in females (females:
N = 21, mean 6 S.E. = 13.21 6 0.25; males: N = 18, mean 6
S.E. = 15.78 6 0.24; unequal variance t-test, t = 7.45, df =36.98
with Satterthwaite’s correction, P , 0.0001).
DISCUSSION
Our experiments generated results contrary to our expectations for both variants of the courtship practice hypothesis,
as males with lower scores on the participation index were
more likely to pair with females, not less: thus, extent of
same-sex ‘‘courtship practice’’ negatively predicted success in
pairing. Males’ participation scores were correlated across samples and were independent of the identity of males included in
samples; these results indicate that there is individual variability in the tendency of subadult males to participate in same-sex
sexual behaviors.
While our sample size is modest, the fact that trends were opposite those predicted by the hypothesis argue against the likelihood that Type II error accounts for our failure to support the
courtship practice hypothesis. Thus, despite the fact that the social organization of this species makes the courtship practice hypothesis quite plausible (see Introduction), we conclude that it
does not explain the tendency of immature male budgerigars to
engage in same-sex sexual behavior. Below, we briefly discuss alternative hypotheses for male–male sexual behavior in this species in light of our findings here.
The follow-up to Experiment 2 established that males that
had been unsuccessful in securing a mate readily paired with
a heterosexual partner when given an opportunity to do so.
Unmated males
Trait
Mean
SE
Mean
SE
Iris prominence
Beak value
Participation index
2.52
3.05
3.00
0.21
0.16
0.62
1.90
3.42
4.32
0.22
0.09
0.53
Also, we failed to observe male–male copulation attempts in
either experiment, and we found no evidence that the
unmated males in Experiment 2 established stable associations
with male partners. Therefore, although these males participated in same-sex interactions at higher levels as subadults,
there was no indication that these males had homosexual mating preferences.
Because we evaluated male mate-getting ability through
competitive pair formation, it was important to include consideration of variation in physical aspects of male phenotype on
mating ability. Our findings, though ancillary to the main purpose of the experiment, are noteworthy if only because mate
choice is a topic that has received relatively little study in this
species. Two of the male physical traits we included, cere and
cap phenotypes, have been previously studied, albeit with
approaches that differ from ours. 1) Many years ago, CinatThompson (1926) suggested that male cere color influences
female choice in budgerigars. Our results are consistent with
her hypothesis, as males that acquired mates in Experiment 2
tended to have darker ceres. 2) Pearn et al. (2001, 2003)
showed that female budgerigars prefer males displaying UVreflectant caps over those whose caps had been made nonUV reflectant by chemical application. To our knowledge, no
experiments have examined effects of naturally varying levels
of UV reflectance of male cap phenotype on female preference, but it is possible that larger caps result in larger UV
displays. (UV reflectance of breast plumage also influences
male mating attractiveness—Griggio, Hoi, et al. 2010.) While
our final model did not include cap size as a predictor of male
mate-getting ability, we cannot discount its possible impact on
attractiveness, especially given our modest sample size, our
inability to measure UV reflectance directly (due to lack of
access to a spectrophotometer), and our finding of sexual
dimorphism in cap size. UV reflectance patterns of plumage
may be a better predictor of condition in this species (Griggio,
Table 3
Behavioral and physical traits predicting male mating status
Trait
Coefficient
Z
P
Participation index
Iris prominence
Beak value
21.15
4.92
24.42
21.87
1.82
21.66
0.061
0.069
0.096
Stepwise Logistic regression (P-to-remove = 0.15), v2 = 13.08, N = 18,
Model P = 0.004.
Figure 3
Distribution of iris prominence scores for 18 male (black fill) and
22 female (white fill) budgerigars. Unequal variance t-test, t =3.11,
df = 31.67 with Satterthwaite’s correction, P = 0.004.
781
7
Abbassi and Burley • Nice guys finish last
Zanollo, et al. 2010) than the residual of mass on body size,
which did not predict mating success of males in this study.
Iris prominence, the remaining trait we found to influence
male mate-getting ability (Table 3), has not (to the best of our
knowledge) been previously investigated even though many
species have brightly colored irides that provide contrast with
pupils. Here, we found this trait to be sexually dimorphic,
with males having higher average prominence scores than
females (Figure 3), and males that obtained mates in Experiment 2 having higher scores than those that did not (Tables 3
and 4). These findings, as well as evidence that humans subconsciously attend to pupil diameter of conspecifics (Tombs
and Silverman 2004), lead us to suggest that the social significance of eye traits deserves increased attention by behavioral
ecologists who study birds and other visually orienting vertebrates.
Alternative hypotheses
Two major alternative hypotheses that have been suggested as
adaptive explanations for the occurrence of same-sex sexual
behaviors are the social glue and intrasexual conflict hypotheses (Table 2 in Bailey and Zuk 2009). Further work would be
needed to test the applicability of these ideas to budgerigars,
of course, but observations here and those made in the course
of previous experiments (especially Hile et al. 2005) suggest
that some possibilities are less likely than others. First, we can
discount one variant of the intrasexual conflict hypothesis:
physical interference of heterosexual copulation through
same-sex mounting (Preston-Mafham 2006) was not observed.
Alliance formation (Mann 2006), a major variant of the ‘‘social
glue’’ hypothesis, also seems unlikely, as there is no evidence
that budgerigars of either sex form same-sex coalitions to secure resources (including mates) or repel invaders either in the
lab or under field conditions. Moreover, since food resources
are not defensible and females are the primary nest defenders,
it is hard to envision a scenario leading to male coalitions. Too
little information is available to currently assess the likelihood
that male–male behavior interactions function in flock cohesion or to reduce stress (Mareike et al. 2007; Aureli and Yates
2010).
As an additional alternative, we suggest that same-sex interactions function as a low-cost mechanism that allow males to
assess their own physical condition and that of others (e.g.,
Arnott and Elwood 2009); for example, variation in intensity
of head bobbing or beak rubs could indicate relative vigor.
While such assessment is typically thought to function to minimize costs of fighting (Parker 1974), we envision a role for it
in foraging decisions, as described below.
Like many species that forage for seed in the open, budgerigars derive protection from foraging in flocks. Although data
were not quantified in the current study, we have observed that
males in heterosexual flocks tend to initiate group foraging
events and that individual males vary in their tendency to lead
to the ground. Moreover, there appear to be ‘‘preferred leaders,’’ individuals who are quickly followed to the ground by
flock mates. These observations lead us to suggest that male
budgerigars assess the relative leadership qualities of others
through the various behaviors in which they engage, adjusting
willingness to follow when they assess changes in condition
(‘‘leadership assessment hypothesis’’). Consistent with both
our observations and this hypothesis, Soma and Hasegawa
(2004) reported that subordinate budgerigars were less likely
to be foraging leaders than were more dominant ones in an
experiment in which agonistic interactions over food were
effected through providing food in small hoppers.
One question arising from the leadership assessment hypothesis is why birds should tend to prefer to follow some for-
aging leaders over others. Individuals preferred as leaders may
be superior food finders (Mateos-Gonzalez et al. 2011) or may
be relatively safe to follow due to superior risk-assessment
abilities or more alert responses when predators arrive at foraging sites. Even though individuals in superior condition
may be at lower risk than others in flock situations, foraging
leadership is likely not cost-free, so the question of why
individuals might choose to be leaders is also significant. Indeed, in some species, individuals with fewer energy reserves
are forced to initiate group foraging bouts despite the enhanced predation risk this role entails (e.g., Fischhoff et al.
2007). In other cases, however, dominant individuals (social
mammals—Peterson et al. 2002; King et al. 2008) or those with
phenotypic markers suggesting superior foraging abilities
(rock sparrows—Tóth and Griggio 2011) tend to lead. Such
leaders may be rewarded by priority of access to resources,
greater control over their feeding schedule, and/or increased
mating access. We expect that male budgerigars that are good
foraging leaders have less need to evaluate the abilities of
others, and may tend to repel overtures from others by delivering hard beak flicks or through aggressive allogrooming
(Harrison 1965). If true, these tendencies could generate the
negative association observed here between participation in
same-sex activities and mating success.
Undoubtedly, there are additional hypotheses to account for
same-sex ‘‘sexual’’ behaviors among group-living birds. Yet,
while careful studies of allogrooming and other seemingly
affiliative contact behaviors have contributed much to our understanding of primate social structure (e.g., Sparks 1967;
Seyfarth and Cheney 1984; Dunbar 1991; Perry 1996; Barrett
et al. 1999; Manson et al. 2004; Silk et al. 2004), little comparable work has been done on group-living birds (but see
Radford and Du Plessis 2006; Mareike et al. 2007; Aureli and
Yates 2010). Future investigations of contact behaviors among
same-sex individuals may reveal the existence of more complex
social relationships among flock members in birds, especially in
groups in which sophisticated cognitive abilities have been
documented. We offer the leadership assessment hypothesis
as one avenue of approach to this understudied topic.
FUNDING
This research was supported by the University of California,
Irvine.
We thank Diane Livio and students in Bio 199, especially John Pulliam,
for assistance with bird maintenance and data collection. Brad Hawkins,
Judy Stamps, Pauline Yahr, and 2 anonymous reviewers made helpful
comments on the manuscript. Pauline Yahr suggested the ‘‘courtship
confidence’’ hypothesis.
REFERENCES
Altmann J. 1974. Observational study of behavior: sampling methods.
Behaviour. 49:227–267.
Arnott G, Elwood RW. 2009. Assessment of fighting ability in animal
contests. Anim Behav. 77:991–1004.
Aureli F, Yates K. 2010. Distressed prevention by grooming others in
crested black macaques. Biol Lett. 6:27–29.
Bagemihl B. 1999. Biological exuberance: animal homosexuality and
natural diversity. London: Profile Books.
Bailey NW, Zuk M. 2009. Same-sex sexual behavior and evolution.
Trends Ecol Evol. 24:439–446.
Barrett L, Henzi SP, Weingrill T, Lycett JE, Hill RA. 1999. Market
forces predict grooming reciprocity in female baboons. Proc R
Soc B Biol Sci. 226:665.
Bell MA, Hankison SJ, Laskowski KL. 2009. The repeatability of behaviour: a meta-analysis. Anim Behav. 77:771–783.
Brockway BF. 1964a. Ethological studies of the budgerigar (Melopsittacus undulatus): non-reproductive behavior. Behaviour. 22:193–222.
782
8
Brockway BF. 1964b. Ethological studies of the budgerigar: reproductive behavior. Behaviour. 23:294–324.
Byers BE. 2007. Extrapair paternity in chestnut-sided warblers is correlated with consistent vocal performance. Behav Ecol. 18:130–136.
Camperio-Ciani A, Corna F, Capiluppi C. 2004. Evidence for maternally inherited factors favouring male homosexuality and promoting female fecundity. Proc R Soc B Biol Sci. 271:2217–2221.
Cinat-Thompson H. 1926. Die geschlechtliche Zuchtwahl beim Wellensittich (Melopsittacus undulatus Shaw). Biol. Zentralbl. 46:543–552.
Dunbar RIM. 1991. Functional significance of social grooming in primates. Folia Primatol. 57:121–131.
Fischhoff IR, Sundaresan SR, Cordingley J, Larkin HM, Sellier M-J,
et al. 2007. Social relationships and reproductive state influence
leadership roles in movements of plains zebra, Equus burchellii.
Anim Behav. 73:825–831.
Forshaw JM. 1981. Parrots of the world. Melbourne (Australia): Landsdowne Editions.
Forshaw JM. 2002. Australian parrots. 3rd revised ed. Queensland
(Australia): Avi-Trader Publishing.
Gavrilets S, Rice WR. 2006. Genetic models of homosexuality: generating testable predictions. Proc R Soc B Biol Sci. 273:3031–3038.
Gosling SD. 2001. From mice to men: what can we learn about personality from animal research? Psychol Bull. 127:45–86.
Griggio M, Hoi H, Pilastro A. 2010. Plumage maintenance affects
ultraviolet colour and female preference in the budgerigar. Behav
Process. 84:739–744.
Griggio M, Zanollo V, Hoi H. 2010. UV plumage color is an honest
signal of quality in male budgerigars. Ecol Res. 25:77–82.
Harrison CJO. 1965. Allopreening as agonistic behavior. Behaviour.
24:161–209.
Hile AG, Burley NT, Coopersmith CB, Foster VS, Striedter GF. 2005.
Effects of male vocal learning on female behavior in the budgerigar,
Melopsittacus undulatus. Ethology. 111:901–923.
King AJ, Douglas CMS, Huchard E, Isaac NJB, Cowlinshaw G. 2008.
Dominance and affiliation mediate despotism in a social primate.
Curr Biol. 18:1833–1838.
Logue DM, Mishra S, McCaffrey D, Ball D, Cade WH. 2009. A behavioral syndrome linking courtship toward males and females predicts
reproductive success from a single mating in the hissing cockroach,
Gromphaldorhina portentosa. Behav Ecol. 20:781–788.
MacFarlane GR, Blomberg SP, Vasey PL. 2010. Homosexual behaviour
in birds: frequency of expression is related to parental care disparity
between the sexes. Anim Behav. 80:375–390.
Mann J. 2006. Establishing trust: socio-sexual behaviour and the development of male-male bonds among Indian Ocean bottlenose dolphins. In: Sommer V, Vasey PL, editors. Homosexual behaviour in
animals. Cambridge (UK): Cambridge University Press. p. 107–130.
Manson JH, Navarrete CD, Silk JB, Perry S. 2004. Time-matched
grooming in female primates? New analyses from two species. Anim
Behav. 67:493–500.
Mareike S, Bugnyar T, Schoegl C, Heinrich B, Kotrschal K, Möstl E.
2007. Corticosterone excretion patterns and affiliative behavior over
development in ravens (Corvus corax). Horm Behav. 53:208–216.
Masure RH, Allee WC. 1934. Flock organization of the shell parakeet
Melopsittacus undulatus Shaw. Ecology. 15:388–398.
Mateos-Gonzalez F, Quesada J, Senar JC. 2011. Sexy birds are superior
at solving a foraging problem. Biol Lett. 7:668–669.
McRobert SP, Tompkins L. 1988. Two consequences of homosexual
courtship performed by Drosophila melanogaster and Drosophila affinis
males. Evolution. 42:1093–1097.
Behavioral Ecology
Moravec ML, Striedter GF, Burley NT. 2010. ‘‘Virtual parrots’’ confirm
mating preferences of female budgerigars. Ethology. 116:961–971.
Parker GA. 1974. Assessment strategy and the evolution of fighting
behavior. J Theor Biol. 47:223–243.
Pearn SM, Bennett ATD, Cuthill IC. 2001. Ultraviolet vision, fluorescence and mate choice in a parrot, the budgerigar Melopsittacus
undulates. Proc R Soc B Biol Sci. 268:2273–2279.
Pearn SM, Bennett ATD, Cuthill IC. 2003. The role of ultraviolet-A
reflectance and ultraviolet-A induced fluorescence in budgerigar
mate choice. Ethology. 109:961–970.
Perry S. 1996. Female-female social relationships in wild white-faced
capuchin monkeys, Cebus capuchinus. Am J Primatol. 40:167–182.
Peterson RO, Jacobs AK, Drummer TD, Mech LD, Smith DW. 2002.
Leadership behavior in relation to dominance and reproductive
status in gray wolves, Canis lupus. Can J Zool. 80:1405–1412.
Preston-Mafham K. 2006. Post-mounting courtship and the neutralizing of male competitors through ‘‘homosexual’’ mountings in the
fly Hydromyza livens F. (Diptera: Scatophagidae). J Nat Hist.
40:101–105.
Radford AN, Du Plessis MA. 2006. Dual function of allopreening in
the cooperatively breeding green woodhoopoe, Phoeniculus purpureus. Behav Ecol Sociobiol. 61:221–230.
Ruxton GD. 2006. The unequal variance t-test is an underused alternative to Student’s t-test and Mann-Whitney U test. Behav Ecol.
17:688–690.
Schulte-Hostedde AI, Zinner B, Millar JS, et al. 2005. Restitution of
mass-size residuals: validating body condition indices. Ecology.
86:155–163.
Seyfarth RM, Cheney DL. 1984. Grooming, alliances and reciprocal
altruism in vervet monkeys. Nature. 308:541–543.
Shannon CE. 1948. A mathematical theory of communication. Bell Sys
Tech J. 27:379–423. ; 623–656.
Silk JB, Alberts SC, Altmann J. 2004. Patterns of coalition formation by
adult female baboons in Amboseli, Kenya. Anim Behav. 67:573–579.
Sokal RR, Rohlf FJ. 1981. Biometry. 2nd ed. San Francisco (CA): W.H.
Freeman.
Soma M, Hasegawa T. 2004. The effect of social facilitation and social
dominance on foraging success of budgerigars in an unfamiliar
environment. Behaviour. 141:1121–1134.
Sparks JH. 1967. Allogrooming in primates: a review. In: Morris D,
editor. Primate ethology. London: Weidenfield and Nicholson. p.
190–225.
Stamps J, Clark A, Arrowood P, Kus B. 1985. Parent-offspring conflict
in budgerigars. Behaviour. 94:1–40.
Stamps J, Kus B, Clark A, Arrowood P. 1990. Social relationships of
fledgling budgerigars, Melopsittacus undulatus. Anim Behav.
40:688–700.
Tombs S, Silverman I. 2004. Pupillometry: a sexual selection approach. Evol Hum Behav. 25:221–228.
Tóth Z, Griggio M. 2011. Leaders are more attractive: birds with bigger
yellow breast patches are followed by more group-mates in foraging
groups. PLoS One. 6:e26605.
Vervaecke H, Roden C. 2006. Going with the herd: same-sex interaction and competition in American bison. In: Sommer V, Vasey PL,
editors. Homosexual behaviour in animals: an evolutionary perspective. Cambridge (UK): Cambridge University Press. p. 131–153.
Wyndham E. 1980. Dirurnal cycle, behaviour and social organization
of the budgerigar Melopsittacus undulatus. Emu. 80:25–33.
Wyndham E. 1981. Breeding and mortality of budgerigars Melopsittacus undulatus. Emu. 81:240–243.