JOURNAL OF AVIAN BIOLOGY 34: 237–242, 2003
Structural plumage colour and parasites in satin bowerbirds
Ptilonorhynchus 7iolaceus: implications for sexual selection
Stéphanie M. Doucet and Robert Montgomerie
Doucet, S. M. and Montgomerie, R. 2003. Structural plumage colour and parasites
in satin bowerbirds Ptilonorhynchus 6iolaceus: implications for sexual selection. – J.
Avian Biol. 34: 237– 242.
We investigated whether variation in structural plumage coloration in satin bowerbirds, Ptilonorhynchus 6iolaceus, could reveal the intensity of infection from parasites,
as predicted from models of parasite-mediated sexual selection (PMSS). To do this,
we captured adult male, female, and juvenile male satin bowerbirds in Queensland,
Australia, and objectively measured individual plumage reflectance from four body
regions using a spectrometer. We quantified both ectoparasite load and the intensity
of infection from blood parasites. In iridescent blue adult males, plumage reflectance
is unimodal, with a single peak in the ultraviolet, while in greenish females and
juveniles, plumage reflectance is bimodal, with peaks in both the ultraviolet and green
portions of the spectrum. In adult males, the intensity of infection from blood
parasites was best predicted by plumage brightness (total reflectance), with brighter
males having fewer parasites. Similarly, juvenile males exhibiting greater UV chroma
(proportion of reflectance in the UV) had fewer blood parasites. Our findings support
a key prediction of PMSS models and provide the first evidence that a structural
colour ornament can signal the intensity of infection from blood parasites.
S. M. Doucet (correspondence) and R. Montgomerie, Department of Biology, Queen’s
Uni6ersity, Kingston, ON, K7L 3N6, Canada. Present address of S. M. Doucet:
Department of Biological Sciences, 331 Funchess Hall, Auburn Uni6ersity, Auburn, AL
36849, USA E-mail: [email protected]
Females may prefer elaborate male ornaments, despite
obvious costs to the bearer, if these ornaments reveal
aspects of male quality, such as reduced parasite loads
(Hamilton and Zuk 1982). By mating with parasite-free
males, females may gain direct benefits (e.g., avoidance
of parasite transmission, quality of paternal care;
Clayton 1991) and/or indirect benefits (parasite resistance genes for their offspring; Hamilton and Zuk
1982). Many investigations of parasite-mediated sexual
selection (PMSS) have justifiably focused on the
plumage coloration of birds, as this is a conspicuous
visual signal that can be objectively quantified (see
Endler and Lyles 1989) and is directly involved in mate
choice in several species (reviewed in Andersson 1994).
To date, intraspecific investigations of PMSS have
disproportionately involved carotenoid- and, to some
extent, melanin-based signaling systems (reviewed in
Møller et al. 1999). Recent comparative and intraspecific studies have, however, highlighted the important
©
role that structural plumage coloration may play in
sexual selection (e.g., Andersson and Amundsen 1997,
Hunt et al. 1998, Owens and Hartley 1998, Keyser and
Hill 2000). Indeed, structural colours have the potential
to reveal different kinds of information about the signaler since they are produced by feather microstructure
rather than pigmentation (e.g., McGraw et al. 2002).
In this study we investigated PMSS in satin bowerbirds, Ptilonorhynchus 6iolaceus. We first characterized
variation in the plumage reflectance and parasite infection intensities in adult males, females, and juveniles,
then assessed the potential for their structural plumage
coloration to serve as an indicator of the intensity of
infection from parasites. We thus predicted a negative
relation between parasite loads and the quality of
plumage colour among individuals. Borgia and Collis
(1989) previously reported a negative relationship between ectoparasite load and mating success in satin
bowerbirds. They also attempted to score plumage col-
JOURNAL OF AVIAN BIOLOGY
JOURNAL OF AVIAN BIOLOGY 34:3 (2003)
237
oration visually but they could find no consistent differences between males (Borgia and Collis 1989), which is
not surprising given the limitations of short-wave human vision compared to that of birds (Cuthill et al.
2000). Borgia and Collis (1989, 1990) proposed that
their data favored a parasite avoidance model of PMSS
in satin bowerbirds. Our findings suggest that, at least
for haematozoan parasites, a different sexual selection
mechanism might be operating.
Methods
We studied a population of satin bowerbirds throughout their breeding season, from September to December
2000, in Mount Baldy State Forest, near Atherton
(17°30%S, 145°30%E), Queensland, Australia. The distribution of this subspecies, P. 6. minor, is limited to the
highlands of northeastern Queensland (Pizzey and
Knight 1997). Using mist-nets baited with blue objects,
we captured 11 adult males, 9 juvenile males, and 6
females and fitted each bird with a unique combination
of one stainless steel band and two colour bands. We
initially sexed green birds using plumage characteristics
(Vellenga 1980), and confirmed sex assignments by
collecting a small blood sample from each individual
and performing molecular analysis using sex-specific
DNA primers (Griffiths et al. 1998).
reflectance over this range, and UV chroma (a measure
of spectral saturation) as the proportion of total reflectance occurring in the UV region.
UV/green colour variables
In contrast to adult males, reflectance spectra for the
greenish plumage of females and juvenile males are best
described as bimodal, with peak reflectance values in
the UV/blue (300 – 450 nm) and green/yellow (450 –700
nm) regions (Fig. 1). Thus, we calculated both ‘blue’
and ‘green’ hues for females and juvenile males as the
wavelength of maximum reflectance in each of those
two regions of the spectrum, respectively, and total
brightness as described for adult males. We calculated
UV chroma as described above and green chroma as
the proportion of total reflectance in the green region,
from 512 to 575 nm. We selected wavelength ranges for
these chroma computations by considering the intersec-
Plumage colour
We measured plumage reflectance of four body regions
(wing coverts, rump, mantle, breast) on each individual,
using an S2000 spectrometer and PX-2 pulsed xenon
lamp (Ocean Optics, Dunedin, Florida, USA). Measurements were taken perpendicular to the feather surface. All reflectance readings were expressed as the
proportion of reflectance from a Spectralon® white
standard, an almost perfect reflector. We obtained five
reflectance spectra for each region, moving the probe at
least 5 mm between readings. We used an average
spectrum from the five readings for each body region in
the following analyses.
UV/blue colour variables
To facilitate sex-, age-, and parasite-based comparisons
of plumage reflectance, we calculated colour variables
according to three dimensions of colour vision: hue,
brightness, and chroma. For adult males, reflectance
spectra are unimodal, with a peak in the UV region
(Fig. 1). To estimate hue, we used the wavelength
corresponding to the maximum reflectance in the range
of spectral sensitivity in birds (300 –700 nm; Cuthill et
al. 2000). We calculated total brightness as the mean
238
Fig. 1. Plumage reflectance spectra from four different body
regions in adult and juvenile male satin bowerbirds. These are
mean reflectance curves from individual males chosen because
their brightness, chroma, and hue values were representative of
the means for each of their plumage types. Female spectral
reflectance is similar to that of juvenile males except for the
breast regions (see text). Note the different scales on the
reflectance axes.
JOURNAL OF AVIAN BIOLOGY 34:3 (2003)
tion of curves describing the relative sensitivities of four
cone types in passerines (Cuthill et al. 2000).
across the four body regions we measured, resulting in
a single measure of each colour variable for each bird.
We used log(y +1) transformations to normalize the
distributions of some variables when appropriate.
Assessing parasite infection
To assess ectoparasite load, we counted the number of
Myrsidea ptilonorhynchi lice around the head and eyes
of the birds we captured. These lice presumably aggregate in these body regions because they can’t easily be
preened away from those areas. This species of louse
belongs to a suborder (Amblycera) of blood-feeding lice
known to influence the fitness of their hosts (Clayton
1991) and appears to be the only common ectoparasite
on satin bowerbirds (Borgia and Collis 1989, 1990).
To assess the intensity of infection from blood parasites, we collected a small blood sample from each bird
by piercing the brachial vein, drawing blood into a
capillary tube, and thinly smearing it onto a glass slide.
We fixed and stained the slides using the Hema 3™
staining procedure (Fisher Scientific). All slides were
scored by the same observer, blind to the identity of the
bird being scored. We observed the stained slides under
oil immersion at 1250 × magnification, scanning each
slide for haemosporidian parasites until approximately
10,000 red blood cells had been surveyed. The only
common parasite we encountered was Haemoproteus
sp. in the form of extra –erythrocytic, developing, and
mature gametocytes (Campbell 1988). Thus, we restricted our blood parasite analyses to the total number
of mature Haemoproteus parasites per 10,000 red blood
cells. Neither ectoparasite load (r = 0.26, n = 26, P =
0.19) nor the intensity of infection from blood parasites
(r =0.06, n =26, P =0.76) varied significantly with
date of sampling.
Statistical analyses
In all analyses we used the grand mean of the calculated colour variables (hue, chroma or brightness)
Results
Plumage colour in satin bowerbirds
The UV/blue hue of adult males peaked at significantly
shorter wavelengths than that of both females and
juvenile males (Tukey-Kramer post hoc tests, P B 0.05),
which were not significantly different from each other
(Table 1). There were also significant differences in
both the total brightness and UV chroma of adult
males compared to females and juvenile males (TukeyKramer test, P B 0.05), with females and juveniles exhibiting greater overall reflectance but relatively less
reflectance in the UV region (Table 1). The green
chroma, UV chroma, and green hue of juvenile males
and females were remarkably similar (Table 1).
In satin bowerbirds, juvenile males retain a cryptic
green plumage, similar to that of females, until they
reach six years of age, when they moult into the iridescent blue characteristic of adult males (Vellenga
1980). The breast plumage of juvenile males has been
described as more greenish than that of females (Vellenga 1980) and females indeed had significantly
brighter (i.e., paler green) breasts than juvenile males
(ANOVA, F1,14 = 8.70, P = 0.01).
Parasite infection
The prevalence of ectoparasitic infection was 91%,
100% and 89% among adult males, females, and juvenile males, respectively, with significant differences in
the intensity of ectoparasitic infection among these
age/sex categories (ANOVA, F2,23 = 3.63, P =0.04; Fig.
2). Adult males had significantly larger ectoparasite
Table 1. Hue, chroma, and brightness values for adult male, female, and juvenile male satin bowerbirds; mean 9 SE, CV
(range). Green hue and chroma were not calculated for adult males.
Colour variables
UV/blue hue (nm)
Green hue (nm)
Total brightness (%)
UV chroma
Green chroma
Adult males
Juvenile males
Females
ANOVA
n= 11
n =9
n=6
F2,23
3629 6.8
6.3 (326–391)
–
389 94.9
3.8 (371–408)
566 93.7
2.0 (543–580)
12.7 90.4
9.8 (10.6–14.5)
0.18 9 0.01
11.0 (0.15–0.21)
0.22 90.003
4.7 (0.20–0.23)
396 95.0
3.1 (378–415)
565 99.1
3.9 (538–592)
13.6 9 0.6
10.9 (11.7–17.2)
0.17 90.01
12.0 (0.16–0.20)
0.22 9 0.01
7.8 (0.19–0.23)
4.79 0.2
14.6 (3.6–5.6)
0.329 0.01
6.2 (0.27–0.35)
–
JOURNAL OF AVIAN BIOLOGY 34:3 (2003)
P
8.8
0.001
0.08
0.78
181
B0.0001
146.9
B0.0001
0.001
0.97
239
ward stepwise procedure. For adult males, we constructed two models, one for each type of parasite, with
parasite load as the dependent variable and mean total
brightness, UV chroma, and UV/blue hue as potential
predictor variables. In adult males, total plumage
brightness was the only significant predictor of the
intensity of infection from blood parasites, explaining
over 35% of the variation in infection intensity (standardized beta = − 0.67, P =0.03; P \ 0.1 for all other
variables). Thus, brighter adult males had fewer blood
parasites (Fig. 3). None of the plumage variables were
significant predictors of ectoparasite load (P \ 0.15).
For juvenile males, we included green hue, UV/blue
hue, total brightness, UV chroma, and green chroma as
potential predictor variables. UV chroma was the only
significant predictor of the intensity of infection from
blood parasites, explaining over 50% of the variation in
infection intensity (standardized beta = − 0.71, P =
0.03; P \ 0.15 for all other variables). Thus juvenile
males whose plumage exhibited greater UV saturation
were infected with fewer blood parasites (Fig. 3). None
Fig. 2. Box plots illustrating the variation in ectoparasite and
blood parasite infection intensities in adult male (n =11),
female (n = 6), and juvenile male (n = 9) satin bowerbirds. Box
plots show 10th, 25th, 50th, 75th, and 90th percentiles with
horizontal lines. Significant (P 5 0.05) differences are identified
by different letters above the boxes.
loads than juvenile males (Tukey-Kramer test, P B
0.05), but there were no significant differences between
females and either adult males or juveniles (P \ 0.05).
Based on an assessment of 10,000 red blood cells, the
prevalence of infection from blood parasites was 64%,
100% and 78% among adult males, females, and juvenile males, respectively. However, a more detailed inspection of the blood smears of the six individuals that
scored zero in this analysis revealed that they had all
been exposed to blood parasites; four were infected
with at least one mature Haemoproteus and two carried
gametocytes. There were no significant differences in
the intensities of blood parasite infections among the
age/sex categories (ANOVA, F2,23 =1.10, P =0.35, Fig.
2). There was also no relationship between ectoparasite
load and the intensity of infection from blood parasites
among all birds (r=0.009, N =26, P =0.96) or when
each age/sex category was considered separately (P =
0.60, P =0.15, P =0.73 for adult males, females, and
juveniles, respectively).
Can plumage colour reveal the intensity of
infection from parasites?
To determine which plumage characteristics could best
reveal the intensity of infection from parasites, we
constructed multiple regression models using a back240
Fig. 3. Significant negative relationships between the intensity
of blood parasite infections in adult (r = −0.63, P = 0.04,
n= 11) and juvenile (r = −0.71, P =0.03, n =9) male satin
bowerbirds and structural plumage colour variables. Model II
regression lines are shown in each case.
JOURNAL OF AVIAN BIOLOGY 34:3 (2003)
of the plumage variables were significant predictors of
ectoparasite load in juvenile males (P \0.5). The small
number of females sampled did not allow us to perform
multivariate analyses while including the five plumage
variables measured.
Discussion
As we have shown here, there is considerable UV
reflectance in both the iridescent blue plumage of adult
male satin bowerbirds and the greenish plumage of
females and juvenile males. The coefficients of variation
(CV) for the structural plumage colour variables of
adult males tended to be higher than those of females
(Table 1), as would be predicted for traits under the
influence of sexual selection and the CVs also tended to
be higher for UV/blue than for green colour variables.
The iridescent plumage colour of males is produced
within the barbules by constructive interference from
arrays of melanin granules and/or air vacuoles suspended in barbule keratin (Fox 1976, Prum 1999). In
contrast, the greenish appearance of females and juvenile males likely results from a combination of coherently scattered light waves in the spongy, medullary
layer of feather barbs and carotenoid pigmentation
(Fox 1976, Prum 1999, S. Doucet pers. obs.). The
UV/blue peak in adult male reflectance spectra occurred at shorter wavelengths than that of females and
juveniles and coincides with the peak sensitivity of
UV-sensitive cones in passerine birds (Cuthill et al.
2000). This difference between reflectance spectra in the
UV range may result from the different production
mechanisms for these two types of structural colour.
Overall, the plumage coloration of females and juvenile
males was similar; the only detectable difference was
revealed in the brighter breasts of females.
Ectoparasite prevalence was high among all satin
bowerbirds, but adult males had significantly more
ectoparasites than either females or juveniles. These
results provide an interesting contrast to the findings of
Borgia and Collis (1990) who showed that, in three
consecutive years, adult males had fewer ectoparasites
than females and juveniles in the southern subspecies of
satin bowerbirds, P. 6. 6iolaceus. A number of different
factors may account for this variation in ectoparasite
intensities, including considerable climate and habitat
differences between the two subspecies distributions.
There were no significant differences between sex and
age classes in the intensity of infection from blood
parasites, possibly resulting from the high individual
variation in infection intensities (range 0 –196 Haemoproteus parasites per 10,000 red blood cells). We found
no evidence of a correlation between degrees of infection from ecto- and endoparasites. The lack of a relationship between lice and blood parasites may reflect
JOURNAL OF AVIAN BIOLOGY 34:3 (2003)
their different transmission modes or the different selective pressures influencing susceptibility to these
parasites.
Our findings support an important intraspecific prediction of PMSS models: that within species, there
should be a negative association between parasite load
and male showiness (e.g., Hamilton and Zuk 1982). In
adult males, total plumage brightness was a significant
predictor of the intensity of infection from blood parasites, as was UV chroma in juvenile males. Because
satin bowerbirds are long-lived, and the birds in our
study were all exposed to blood parasites, low parasite
intensities may be related to parasite resistance in this
species.
Female satin bowerbirds make multiple courtship
visits to several males, eventually narrowing down the
pool of copulation partners to one or two males (Uy et
al. 2000). Choosiness is therefore likely to incur some
costs to females in terms of time and energy constraints
(Gibson and Langen 1996) and the increased likelihood
of aggressive encounters and forced copulations with
other males (Uy et al. 2000). Yet, male satin bowerbirds
provide only gametes to females, suggesting that female
choosiness may be driven by the potential to gain
indirect fitness benefits. Our findings suggest that females could choose relatively parasite-free males by
assessing male plumage colour, and as a result acquire
parasite-resistance genes for their offspring, if parasite
resistance is heritable (Hamilton and Zuk 1982).
Obtaining parasite resistance for offspring is of no
small consequence. Although much of the evidence
emphasizing the pathogenicity of Haemoproteus parasites originates from laboratory studies of domesticated
birds, reports of dead or moribund wild birds with
acute haemosporidian infections generally support laboratory findings (Atkinson and van Riper 1991). In
blue tits Parus caeruleus, females medicated with an
antimalarial substance exhibited a reduced level of
Haemoproteus infection, and had fewer nestling deaths
and greater fledging success than unmedicated control
females (Merino et al. 2000). Given that younger birds
are more susceptible to infections (Scott and Edman
1991) and more likely to die from acute haematozoan
infections (Atkinson and van Riper 1991), the consequences of acquired genetic resistance could be even
more pronounced in nestlings. Thus, choosing parasiteresistant copulation partners could thus influence lifetime reproductive success in female satin bowerbirds by
impacting offspring survival.
This is the first study to show an association between
structural plumage colour and blood parasites (see also
Doucet and Montgomerie 2003), although a link between structural plumage colour and feather mites has
been documented (Harper 1999). In our study, neither
adult male plumage nor juvenile plumage was a significant predictor of ectoparasite load. This finding may
not be surprising, given that ectoparasite load likely
241
varies on a shorter time scale than chronic infection
from blood parasites. Thus, ectoparasite load is more
likely to signal current condition. Indeed, we show
elsewhere that males with fewer ectoparasites build
higher quality bowers, another trait likely to be influenced by current condition (Doucet and Montgomerie
2003). As Borgia and Collis (1989, 1990) remarked,
ectoparasites are quite visible in satin bowerbirds, and
the fact that females avoid mating with highly parasitized males suggests that these ectoparasites may fall
under a ‘parasite avoidance’ rather than ‘good genes’
model of parasite-mediated sexual selection.
Acknowledgements – We thank M. Bhardwaj for excellent field
assistance, D. Westcott for indispensable help with establishing this project, D. J. Mennill for assistance in the field, K. J.
McGraw for help with carotenoid extractions, three anonymous reviewers for numerous helpful suggestions, the Australian Bird & Bat Banding Scheme and the Queensland
Department of Environment for permission to work on satin
bowerbirds, the Queensland Department of Natural Resources
for permission to work in State Forests, and CSIRO Australia
and the Tropical Forest Research Center in Atherton for
logistical support. Funding was provided by the Natural Sciences and Engineering Research Council of Canada in the
form of a PGS A scholarship to S.M.D. and both research and
equipment grants to R.M.
References
Andersson, M. B. 1994. Sexual Selection. – Princeton University Press, Princeton.
Andersson, S. and Amundsen, T. 1997. Ultraviolet colour
vision and ornamentation in bluethroats. – Proc. R. Soc.
Lond. B 264: 1587 –1591.
Atkinson, C. T. and van Riper, C. III. 1991. Pathogenicity and
epizootiology of avian haematozoa: Plasmodium, Leucocytozoon, and Haemoproteus. – In: Loye, J. E. and Zuk, M.
(eds). Bird-parasite interactions: ecology, evolution, and
behaviour. Oxford University Press, Oxford, pp. 19 – 48.
Borgia, G. and Collis, K. 1989. Female choice for parasite-free
male satin bowerbirds and the evolution of bright male
plumage. – Behav. Ecol. Sociobiol. 25: 445 –454.
Borgia, G. and Collis, K. 1990. Parasites and bright male
plumage in the satin bowerbird (Ptilonorhynchus 6iolaceus).
– Amer. Zool. 30: 279 –285.
Campbell, T. W. 1988. Avian Hematology and Cytology. –
Iowa State University Press, Ames.
Clayton, D. H. 1991. The influence of parasites on host sexual
selection. – Parasit. Today 7: 329 –334.
Cuthill, I. C., Partridge, J. C., Bennett, A.T. D., Church, S. C.,
Hart, N. S. and Hunt, S. 2000. Ultraviolet vision in birds.
– Adv. Study Behav. 29: 159 –214.
242
Doucet, S. M. and Montgomerie, R. 2003. Multiple sexual
ornaments in satin bowerbirds: UV plumage and bowers
reveal different aspects of male quality. – Behav. Ecol. 14:
503 – 509.
Endler, J. A. and Lyles, A. M. 1989. Bright ideas about
parasites. – Trends Ecol. Evol. 4: 246 – 248.
Fox, D. L. 1976. Animal biochromes and structural colours,
2nd edition. – University of California Press, Berkeley.
Gibson, R. M. and Langen, T. A. 1996. How do animals
choose their mates? – Trends Ecol. Evol. 11: 468 – 470.
Griffiths, R., Double, M. C., Orr, K. and Dawson, R. G.
1998. A DNA test to sex most birds. – Mol. Ecol. 7:
1071 – 1075.
Hamilton, W. D. and Zuk, M. 1982. Heritable true fitness and
bright birds: A role for parasites? – Science 218: 384 – 387.
Harper, D. G. C. 1999. Feather mites, pectoral muscle condition, wing length and plumage coloration in passerines. –
Anim. Behav. 58: 553 – 562.
Hunt, S., Bennett, A. T. D., Cuthill, I. C. and Griffiths, R.
1998. Blue tits are ultraviolet tits. – Proc. R. Soc. Lond. B
265: 451 – 455.
Keyser, A. J. and Hill, G. E. 2000. Structurally based plumage
coloration is an honest signal of quality in male blue
grosbeaks. – Behav. Ecol. 11: 202 – 209.
McGraw, K. J., Vonnegut, E. A., Dale, J. and Hauber, M. E.
2002. Different plumage colors reveal different information: how nutritional stress affects the expression of
melanin- and structurally based ornamental coloration. –
J. Exp. Biol. 205: 3747 – 3755.
Merino, S., Moreno, J., Sanz, J. J. and Arriero, E. 2000. Are
avian blood parasites pathogenic in the wild? A medication
experiment in blue tits (Parus caeruleus). – Proc. R. Soc.
Lond. B 267: 2507 – 2510.
Møller, A. P., Christe, P. and Lux, E. 1999. Parasitism, host
immune function, and sexual selection. – Quart. Rev. Biol.
74: 3 – 20.
Owens, I. P. F. and Hartley, I. R. 1998. Sexual dimorphism in
birds: why are there so many different forms of dimorphism? – Proc. R. Soc. Lond. B 265: 397 – 407.
Pizzey, G. and Knight, F. 1997. The Field Guide to the Birds
of Australia. – HarperCollins Publishers, Sydney.
Prum, R. O. 1999. The anatomy and physics of avian structural colours. – In: Adams, N. and Slotow, R. (eds). Proc.
22nd Int. Ornith. Congr., Durban, Univ. Natal., pp. 1669 –
1686.
Scott, T. W. and Edman, J. D. 1991. Effects of avian host age
and arbovirus infection of mosquito attraction and bloodfeeding success. – In: Loye, J. E. and Zuk, M. (eds).
Bird-parasite interactions: ecology, evolution, and behaviour. Oxford University Press, Oxford, pp. 179 – 204.
Uy, J. A. C., Patricelli, G. L. and Borgia, G. 2000. Dynamic
mate-searching tactic allows female satin bowerbirds
Ptilonorhynchus 6iolaceus to reduce searching. – Proc. R.
Soc. Lond. B 267: 251 – 256.
Vellenga, R.E. 1980. The moults of the satin bowerbird
Ptilonorhynchus 6iolaceus. – Emu 80: 49 – 54.
(Recei6ed 12 August 2002, re6ised 17 December 2002, accepted
1 January 2003.)
JOURNAL OF AVIAN BIOLOGY 34:3 (2003)