Content uploaded by Charlotte Récapet
Author content
All content in this area was uploaded by Charlotte Récapet
Content may be subject to copyright.
MELANIN-BASED COLORATION AS
A SIGNAL OF ALTERNATIVE LIFE-HISTORY
STRATEGIES IN FERAL PIGEON COLUMBA LIVIA
March – July 2010
Feral pigeons in Paris - Steve Greaves 2008
Charlotte Récapet,
École Normale Supérieure, Paris.
Supervisors: Julien Gasparini and Lisa Jacquin,
Evolutionnary Ecology, UMR 7625 CNRS-UPMC, Paris 6 University.
2
METHODS 5
1) EXPERIMENTAL SET-UP 5
2) MEASUREMENT OF EUMELANIC COLORATION 6
3) EVALUATION OF INVESTMENT IN IMMUNITY, MAINTENANCE AND REPRODUCTION 7
4) STATISTICAL ANALYSES 8
RESULTS 10
1) IMMUNE RESPONSE 10
2) MAINTENANCE 11
3) REPRODUCTION 13
4) TRADE-OFFS 14
MAINTENANCE AND IMMUNITY 14
IMMUNITY AND REPRODUCTION 14
MAINTENANCE AND REPRODUCTION 14
DISCUSSION 16
TRADE-OFF BETWEEN MAINTENANCE AND IMMUNE RESPONSE 16
TRADE-OFF BETWEEN IMMUNE RESPONSE AND REPRODUCTION 16
TRADE-OFF BETWEEN MAINTENANCE AND REPRODUCTION 17
CONCLUSION 19
ACKNOWLEDGMENTS 19
REFERENCES 20
3
Introduction
Fluctuations of environmental conditions and selective pressures over time or space
have been hypothesised to drive genetic diversification and genetic diversity maintenance in
natural populations (Gillespie & Turelli 1989, Byers 2005). However, empirical evidences of
this disruptive selection are still rare (Hedrick 2006). Mechanisms explaining the maintenance
of polymorphism in the wild remain therefore to be tested. A species is said to be
polymorphic when individuals from the same age- and sex-class can exhibit variations in a
trait (at least two phenotypes must have a frequency higher than 0.5 %) that are genetically
determined, independently of environment and body condition (Roulin 2004). Variations in
melanin-based coloration in birds often follow Mendelian laws (see Roulin 2004 for a review)
and have been shown to be highly heritable in some species (feral rock pigeon Columba livia,
Johnston & Janiga 1995; barn owl Tyto alba, Roulin et al. 1998; tawny owl Strix aluco,
Gasparini et al. 2009a; guillemot Uria aalge, Jefferies & Parslow 1976). As they can be easily
measured, variations in melanin-based coloration constitute a good trait to study selective
pressures potentially responsible for genetic polymorphism maintenance.
Variations in melanin-based coloration has been shown to be related to variations in
physiological and life-history traits (see Roulin 2004 for a review). For example, dark reddish
females tawny owls Strix aluco maintain a higher level of antibodies for a longer period of
time compared to pale reddish females when injected with a vaccine (Gasparini et al. 2009a).
Moreover, tawny owls chicks from differently coloured genetic mothers have different
abilities to convert food into body mass. Chicks from paler females were less affected by food
restriction combined with an immune challenge (Piault et al. 2009). Differently coloured
individuals may thus differ in the resolution of physiological and life-history trade-offs, which
open the interesting possibility that melanin-based coloration could be a signal of genetic
strategies facing variable environments (Roulin 2004).
Feral rock pigeon Columba livia constitute an excellent model to study the
maintenance of melanin-based coloration polymorphism. In fact, a large variety of melanin-
based coloration patterns, called morphs, has been artificially selected during domestication
for aesthetic purposes. Some of this coloration diversity has been maintained when pigeons
escaped from captivity and colonized the urban environment, but the factors involved in the
maintenance of coloration polymorphism still remain poorly known (see Johnston and Janiga
1995 for a review). A few studies have tried to investigate this phenomenon. For example,
Obukhova (2001) showed that the proportion of strongly melanic pigeons is higher in
urbanized area, suggesting different selective pressures along an urbanization gradient. A
4
previous study in Paris showed differential parasite resistance between differently coloured
free-living pigeons, with darker individuals being more resistant to blood parasites than paler
ones (Jacquin et al. in press). Parasitism seems thus to be an important environmental factor
involved in the maintenance of melanin-based coloration polymorphism in urban areas. As
mounting an immune response is energetically costly (Lochmiller & Deerenberg 2000), birds
showing a higher resistance to parasites are expected to invest less in other traits such as
maintenance and reproduction. Such a trade-off might therefore have different effects
depending on food resources in the environment. As parasite abundance and food availability
probably fluctuate over time and space, the beneficiary morph will change over time and
space, promoting polymorphism maintenance at a global scale. Thus, spatio-temporal
variations in parasite diversity and food abundance within the urban environment could
promote the maintenance of alternative genetic colour morphs as each morph could be better
adapted to different parasite exposure levels and/or to alternative food availability levels
(Roulin 2004).
In this study, we aimed at comparing life-history traits of differently coloured captive
feral pigeons under various parasitic pressures and food availability conditions. We wanted to
determine whether melanin-based coloration was linked to different resolutions of trade-offs
between maintenance, immunity and reproduction. Individuals were or not submitted to a
non-pathogenic immune challenge and/or to food limitation. Investment in maintenance was
evaluated by measures of body mass. Investment in immunity was measured as the intensity
of specific immune response (specific IgG antibodies levels). Allocation to reproduction was
evaluated both as the reproductive activity and the total reproductive output over the whole
experiment. Trade-offs between these life-history traits are expected to vary between
colorations depending on environmental conditions, as differently coloured individuals may
be differently adapted to alternative environments. Previous studies showed that darker
pigeons were more resistant to parasites (Jacquin et al. in press) and that darker barn owls
Tyto alba lost more weight when immune challenged (Gasparini et al. 2009a). We thus
expected darker pigeons to invest more in immunity at the expense of maintenance and/or
reproduction than paler ones when exposed to an immune challenge, and to be more sensitive
to food restriction (Piault et al. 2009).
5
Methods
1) Experimental set-up
The experiment was conducted from March to July 2010. 120 adult rock pigeons
Columba livia were captured in 3 locations of the Parisian suburbs and kept in 10 outdoor
aviaries. Each aviary contained six males and six females. Birds were submitted to two
treatments: a “food treatment” and an “immune treatment”.
The food treatment was initiated two weeks before the immune treatment, on day -14
(see Table 1). Food was manipulated both in quantity and quality. 60 pigeons (five aviaries)
were food limited (“limited food group”). They were fed with 30 grams wheat (low protein
and lipid diet) per day and animal, which correspond to a basal food quantity to maintain non-
breeding pigeons (Hawkins et al. 2001). When chicks hatched, no food was added during
their first week of life, then 15 g per chick and per day during their second week and 30 g per
chick and per day afterward. 60 other pigeons (5 aviaries) were fed ad libitum (“non-limited
food group”) with mixed corn, wheat and peas (high protein and lipid diet) (see Table 2). All
pigeons were provided with mineral grit and water supplemented with vitamins.
The immune treatment started on day 0. On day 0 and 14 (see Table 1), 60 birds (three
pairs chosen randomly in each aviary, called the “immunised group”) were injected
subcutaneously with 100 µL of a solution containing 0.5 mg.mL
-1
KLH (Keyhole Limpet
Hemocyanin). KLH is an artificial protein that birds do not encounter in their natural
environment, therefore used to stimulate immune responses to a novel antigen in birds
(Hasselquist et al. 2001, Duffy et al. 2002). A first injection at day 0 is expected to cause a
primary immune response, whereas a second injection at day 14 should cause a secondary
stronger and longer immune response. 60 other birds (“non-immunised group”) were injected
with a neutral physiological saline (PBS) (see Table 2).
Day
-14
0
14
28
35
42
Start of food
treatment
First injection
Second
injection
Opening of the
nests (start of
reproduction)
Table 1: Experimental schedule.
6
Body mass and coloration score were balanced between food treatments (Body mass:
χ
2
= 0.016, df = 1, P = 0.90; Coloration score: χ
2
= 0.32, df = 1, P = 0.57), immune treatments
(Body mass: χ
2
= 0.13, df = 1, P = 0.72; Coloration score: χ
2
= 0.31, df = 1, P = 0.58), and
their interaction (Body mass: χ
2
= 0.77, df = 1, P = 0.38; Coloration score: χ
2
= 1.33, df = 1, P
= 0.25).
Immune treatment
Immunised
(injection of KLH)
Non-immunised
(injection of PBS)
Limited
(-)
30
30
Food
treatment
Non-limited
(+)
30
30
Table 2: Number of individuals involved in each experimental group.
2) Measurement of eumelanic coloration
Colour variation (black, red or brown) in vertebrates is due to the deposition of two
different types of melanin pigments: yellow to red phaeomelanins and black eumelanins
(Haase et al. 1992). In this study we focused on eumelanic black coloration because it is the
most widespread coloration in feral pigeon populations. Indeed, feral pigeons display a
continuous variation in this eumelanin-based coloration from white to black (Johnston &
Janiga 1995) which can be divided by human eye in 5 main groups (Johnston & Janiga 1995;
Johnson & Johnston 1989): (0) White or almost white individuals; (1) Blue bar (grey mantle
with two dark wing bars); (2) Checker (a checked mantle with moderate dark spots); (3) T-
pattern (a dark mantle with small grey marks); and (4) Spread (a completely melanic
plumage) (Figure 1). These patterns are mainly genetically determined (Jonhston & Janiga
1995) and differ by the surface of dark area on the wings that corresponds to different melanin
deposition in feathers (Haase et al. 1992).
7
Figure 1: Main eumelanin-based coloration morphs in feral pigeons
However, intermediary morphs are frequent and melanin-based coloration may be
considered as continuous. In this study we thus calculated a continuous score as the
percentage of dark surface on the wing compared to grey surface. All birds were
photographed with a digital camera (Cyber-shot DSC-HX1, Sony) under a standardised white
light in a photo light tent cube. The percentage of dark surface was calculated as the number
of black pixels/number of white pixels x 100 after binary transformation of the picture on the
wings of birds using ImageJ software (U.S. National Institutes of Health, Bethesda, USA).
Dark surface measurements were highly repeatable between photographs of the same
individual (n = 30, 4 photographs per individual, F
29,90
= 40.28, P < 0.001, r = 0.91).
3) Evaluation of investment in immunity, maintenance and reproduction
The aim of this study was to compare the relative investment of each morph in
immunity, maintenance and reproduction.
To assess the quality of the immune response, we took blood samples from the wing
vein of each bird every two weeks from day 0 in order to assay specific anti-KLH antibodies
production. Blood samples were left at 5 °C for 24 hours before centrifugation (10000 r.p.m.
for 10 min). Plasma was then extracted and stored at -20 °C. The ELISA technique was used
to quantify anti-KLH antibodies in plasma.
To assess the investment in maintenance, we measured body mass to the nearest 5g
with a spring balance (Medio-Line 40600, Pesola, Switzerland) every 2 weeks from day -14.
To assess the investment in reproduction, we recorded the reproductive state as the
reproductive activity on a particular week (0: no eggs; 1: incubating eggs; 2: feeding chicks)
as well as the total reproductive output, i.e. the total number of eggs laid at the end of the
experiment (day 126).
8
Anti-KLH antibody assay
High-binding plates (96 wells, flat bottom, Microlon® 600 ; cat. 655101, Greiner Bio-
One, Germany) were coated overnight at 5 °C with 100 µL KLH (40 µg.mL
-1
in 50 mM
carbonate/bicarbonate buffer, pH 9.6), then washed five times with phosphate buffered saline
(0.1 M PBS, pH 7.4). Wells were blocked with 200 µL 3% milk powder (Régilait Bio) in PBS
for two hours at room temperature with agitation. After five washings, two 100 µL dilutions
in PBS with 0.5% milk powder were distributed for each sample (1:500 and 1:5000). A
standard pool of pigeon serums with high concentrations of anti-KLH antibodies, diluted
serially from 1:500 to 1:32000, was added to each plate, with a duplicate for each dilution.
Plates were then incubated either one night at 5 °C or two hours at room temperature with
agitation. After five washings, plates were incubated two hours at room temperature with 100
µL rabbit-anti-pigeon IgG conjugated to horseradish peroxidase (10 mg/ml in PBS, pH 7.2;
cat. RAP/IgG(H+L)/PO, Nordic Immunology, Netherlands), diluted 1:5000 in PBS with 0.5%
milk powder. Plates were washed five times, then 100 µL OPD (o-Phenylenediamine
dihydrochloride, 0.4 mg.mL
-1
with 0.4 mg.mL
-1
urea hydrogen peroxide and 0.05 M
phosphate-citrate in distillated water, pH 5.0 ; cat. P9187, Sigma-Aldrich, USA) was added to
the wells. OPD is the substrate of peroxidase and the reaction produces a coloured compound
with a maximal absorption at 450 nm. After 10 minutes in the dark at room temperature, the
reaction was stopped by adding 50 µL HCl (1 M). Plates were then read at 490 nm in a
microplate reader (Model 680, Bio-Rad Laboratories, UK) and data were computed using
Microplate Manager (Bio-Rad Laboratories). A logistic curve was fitted to the standards to
calculate the unknown concentrations of the samples. Only values within the standard
concentration limits were kept. When the 1:5000 dilution gave a value above 40, a higher
dilution was realised (1:50000). Log-transformed anti-KLH titres were highly repeatable
within-plates (n = 59, 2 wells per sample, F
46,46
= 13.94, P < 0.001, r = 0.93) as well as
between-plates (n = 40, 2 wells per sample, F
29,31
= 38.32, P < 0.001, r = 0.81), showing that
this method is reliable to measure anti-KLH antibody levels.
4) Statistical analyses
Immune response
Antibody levels were log-transformed to obtain a normal distribution. We studied the
dynamics of the immune response separately for both immune treatment groups, by including
time as a linear effect in mixed-effects models and an individual random effect (on slope and
intercept) to account for repeated measures on a same individual over time. Generalized linear
models (GLM) were fitted to study both quantitative (coloration score) and qualitative effects
9
(food treatment) and their interactions on the intensity of the primary and secondary immune
responses (day 14 and day 28) in the immunised group.
Maintenance
Body mass dynamics over the experiment were analysed using restricted maximum-
likelihood fit (REML) of mixed-effects models with food, immune treatment and time as
factors, coloration score as a covariate and all interactions. To account for repeated measures
on a same individual over time, we added an individual random effect on the intercept.
Reproduction
The total number of eggs laid was analysed by fitting GLM with both treatments as
factors and coloration score as a covariate.
Trade-offs
First, we examined the trade-offs between maintenance and immunity by modelling
the relationship between antibody levels and body mass loss in a GLM.
Then we investigated the trade-off between immunity and reproduction by testing the
impact of reproduction on antibody production. As reproduction had not started during the
first month (see Table 1), we studied the effect of reproductive state on the decrease in
antibody levels from day 14 (primary response) to day 84 (stop of the decrease in antibody
levels, see Figure 2). The food treatment was also included as a factor.
Finally we tested the trade-off between maintenance and reproduction by modelling
the link between body mass loss and antibody production between day 0 and day 14 or day 28
in a GLM.
All final models were assessed by a step-wise model selection using the AIC criterion
and an analysis of variance was used to estimate the significance of each effect of the selected
model. Significance levels were set to 0.05 and tests were two-tailed. Each model was
checked for homoscedasticity and normality of the residuals. Cook’s distances were inferior
to one. All analysis were performed with R (R Foundation for Statistical Computing, Vienna).
10
Results
1) Immune response
In the immunised group, anti-KLH antibody level increased significantly after the first
injection (Figure 2; paired t-test between day 0 and day 14: t = 17.50, df = 59, P < 0.001).
Antibody level then decreased continuously throughout the experiment (mixed model, linear
effect of time from day 14 to day 98: F
1,353
= 196.99, P < 0.001). There was no increase in
antibody levels between the primary response on day 14, and the secondary response on day
28, when birds were challenged for a second time with the same antigen (paired t-test between
day 14 and day 28: t = 0.21, df = 59, P = 0.84). Antibody level of non-immunised individuals
showed no variation over time (mixed model; linear effect of time from day 0 to day 98: F
1,411
= 1.06, P = 0.30).
Figure 2: Mean antibody level as a function of time, for immunised birds (solid line) and non-
immunised birds (dotted line).
Immunised individuals from the non-limited food group had a higher secondary
response than individuals from the limited food group (Figure 3; GLM for the immunised
group at day 28; effect of food treatment: F
1,56
= 8.95, P = 0.004). There was a quadratic
relationship between coloration score and the strength of the secondary response (F
2,56
= 6.19,
P = 0.004), but no interaction with the food treatment (F
2,54
= 0.29, P = 0.75). Immune
response is lower for intermediate melanic individuals in both food treatments (Figure 3). A
similar relation to coloration score was found for the primary response (F
2,57
= 3.72, P = 0.03)
but with no effect of the food treatment (F
1,58
= 0.96, P = 0.33).
injection
1
injection 2
○ non-immunised
● immunised
Mean antibody level
11
Figure 3: Log-transformed antibodies levels of immunised individuals at day 28 (secondary
response) as a function of the coloration score, for the non-limited food group (white dots,
dotted line) and the limited food group (black dots, solid line).
2) Maintenance
Food treatment had a significant effect on body mass dynamics (Figure 4; food × time:
F10,1132 = 2.79, P = 0.002). There was a significant difference in mass between the two
treatment groups on day 42 (Table 3), one week after the starting of egg-laying and from day
70 (beginning of chick rearing) to the end of the experiment (Table 3). Both treatment groups
lost mass during chicks rearing (paired t-test between day 56 and day 126, in the unlimited
food group : t = -2.95, df = 57, P = 0.005 ; in the limited food group : t = -6.43, df = 57, P <
0.001), but this effect was more pronounced in the food limited group (body mass loss
between day 56 and day 126 in the limited group: -26.8 ± 4.2; in the non-limited food group: -
15.9 ± 4.4 g). There was no effect of the immune treatment or its interaction with the food
treatment on body mass dynamics (KLH × time: F10,1132 = 0.31, P = 0.98; food × KLH ×
time: F10,1132 = 0.54, P = 0.86).
Table 3: Comparison of mean body mass between the two food-treatment groups by Welch
two-sample t-test
Day
-14
0
14
28
42
56
70
84
98
126
t
-0.67
-0.56
-0.90
-1.06
-2.86
-0.87
-5.79
-2.66
-1.81
-2.76
df
114
115
116
117
118
116
117
117
116
114
P
0.51
0.58
0.37
0.29
0.005
0.39
< 0.001
0.009
0.07
0.007
○ non-limited
● limited
Log-transformed antibody level
12
Figure 4: Variation in mean body mass for the two food-treatment groups. Stars indicate
significant differences between the two treatment groups (Table 3).
Analyses show a significant interaction of food treatment and coloration on body mass
dynamics (food × coloration score × time: F
10,1132
= 2.19, P = 0.016), whereas coloration
alone had no effect (coloration score × time: F
10,1132
= 0.93, P = 0.50). Darker individuals lost
more weight than paler ones in the food limited group but lost less weight than paler ones in
the non-limited food treatment (Figure 5).
Figure 5: Mass loss over the whole experiment (from day -14 to day 126) as a function of the
coloration score, for the non-limited food group (white dots, dotted line) and the limited food
group (black dots, solid line).
First
chick
First egg
Injection 2
Injection 1
Start of
food
treatment
*
*
*
*
○ non-limited
● limited
○ non-limited
● limited
13
3) Reproduction
Food limited birds laid significantly less eggs (n = 60, mean number of eggs laid per
individual ± SE: 3.30 ± 0.35) than the non-limited food group (n = 60, 4.23 ± 0.45) (see Table
4). The number of eggs laid was also positively correlated with coloration score when all
treatment groups were considered together (slope ± SE = 0.045 ± 0.017) (see Table 4).
However this relationship differed between food and immune treatment groups (Table 4).
Under food limitation or immune challenge, the number of eggs laid tended to be higher for
darker individuals, whereas under mild environmental conditions (non-immunised or non-
limited food group) differently coloured individuals laid a similar number of eggs (Figure 6).
F
1,112
p
Food
6.79
0.010
*
Immune treatment
1.75
0.19
Coloration score
6.69
0.011
*
Food × Immune treatment
2.44
0.12
Food × Coloration score
6.10
0.015
*
Immune treatment × Coloration score
4.18
0.043
*
Food × Immune treatment × Coloration score
4.14
0.044
*
Table 4: Type III-Ancova on the number of eggs laid over the whole experiment with food and
immune treatment as factors and coloration score as covariate. The final model was assessed
by a step-wise model selection using the AIC criterion. Adding initial body mass as a
covariate and its interactions did not improve the model (AIC = 622 in both cases after model
simplification). Sex and its interactions had no significant effect (P > 0.10).
Immune treatment Food treatment
Figure 6: Number of eggs laid as a function of coloration in immune treatment groups (left;
immunised: black dots, solid line; non immunised: white dots, dotted line) and food treatment
groups (right; limited: black dots, solid line; non limited: white dots, dotted line).
○ non-limited
● limited
○ non-immunised
● immunised
14
4) Trade-offs
Maintenance and immunity
There was no significant link between the increase in antibody levels and body mass
changes for the primary response (F
1,57
= 0.08, P = 0.78) nor the secondary response (F
1,57
=
0.01, P = 0.91).
Immunity and reproduction
The decrease in antibody levels from day 14 to day 84 in the immunised group was
linked to the reproductive state (having eggs or chicks) on day 84 (GLM with food treatment
as factor, reproductive state as covariate and their interactions; reproductive state: F
1,55
= 8.86,
P = 0.004). Birds with no eggs kept circulating antibodies for a longer time in their plasma
than birds incubating eggs or raising chicks (Figure 7). This correlation was independent of
melanin-based coloration.
Figure 7: Mean decrease in log-transformed antibody level between day 84 and day 14
(primary response) as a function of reproductive state on day 84, in the immunised group.
Maintenance and reproduction
Body mass loss over the whole experiment was not related to the number of eggs laid
(GLM with food treatment and sex as factors and number of eggs as a covariate: F
1, 109
= 0.24,
P = 0.63). Nevertheless changes in body mass were negatively influenced by reproductive
state (Table 5; Figure 8). Differently coloured birds were similarly affected by reproduction in
Decrease in antibody level
15
their body mass dynamics (reproductive state × coloration score × time: F
1,1173
= 0.93, P =
0.33). Females were more negatively affected by reproduction than males (Table 5).
Table 5: Mixed-effects final model (AIC criterion selection) explaining body mass changes
over time (individual as a random intercept effect).
There were no differences in body mass dynamics between birds without eggs and
birds incubating eggs (Figure 8; t = 1.20, df = 1179, P = 0.23). But birds that were feeding
chicks exhibited significantly higher body mass losses than non reproductive birds (Figure 8; t
= -2.10, df = 1179, P = 0.036).
Figure 8: Mean mass (± SE) loss between day 126 and day -14 (start of food treatment) as a
function of reproductive state on day 126.
F
df
p
Time
5064.84
1, 1184
< 0.001
***
Food
1.08
1, 115
0.30
Sex
13.72
1, 115
< 0.001
***
Coloration score
0.35
1, 115
0.55
Reproductive state
12.18
1, 1184
< 0.001
***
Time × Food
1.01
1, 1184
0.31
Time × Coloration score
5.03
1, 1184
0.025
*
Time × Reproductive state
14.20
1, 1184
< 0.001
***
Food × Sex
0.37
1, 115
0.54
Food × Reproductive state
22.87
1, 1184
< 0.001
***
Sex × Reproductive state
40.65
1, 1184
< 0.001
***
Time × Food × Reproductive state
3.52
1, 1184
0.061
○ non-limited
● limited
16
Discussion
Trade-off between maintenance and immune response
Birds injected with KLH had a higher immune response in the non-limited food group
than in the limited food group. This could be due to a negative effect of food limitation on
antibody production, as that process is known to be energetically demanding (Lochmiller &
Deerenberg 2000). However it is impossible to draw such a conclusion in our experimental
design since individuals in the non-limited food group had already been injected with KLH
antigen in year 2009, whereas individuals in the limited food treatment were not. Therefore
the higher antibody production in the non-limited food group could be due to a stronger
reactivation of the immune system in 2010 when challenged with the same antigen. Moreover,
body mass changes of the immunised individuals were not related to the intensity of the
immune response. Therefore we provide no evidence of a direct trade-off between
maintenance and immune response.
Moreover, differently coloured individuals exposed or not to an immune challenge lost
similar weight. This is surprising as Gasparini et al. (2009a) found a significant effect of
immune challenge on body mass loss in dark barn owls but not in pale ones. However, our
study takes place in artificial aviaries where predation and climate pressure may be milder
than in the wild, making the expression of a trade-off between immunity and maintenance
difficult to bring to light. It is also possible that the immune challenge was not strong enough
to cause significant differences in weight losses between different treatment groups and
differently coloured individuals.
Trade-off between immune response and reproduction
Birds with no eggs maintained higher antibody levels in their plasma for a longer time
than birds incubating eggs or raising chicks, suggesting a trade-off between reproduction and
immunity. This correlation was independent of coloration. Two mechanisms are possible:
investment in reproduction impairs antibodies production, or the cost of the immune response
decreases investment in reproduction. As there was no clear effect of the immune treatment
alone or in interaction with the food treatment on the number of eggs laid, the first proposal
seems more likely. This is consistent with studies on pied flycatchers Ficedula hypoleuca that
found no effect of maternal immunisation with LPS on the size of replacement clutch after
eggs removal (Grindstaff et al. 2006). However, further studies evaluating hatching and
fledging success could show an impact of immune challenge on reproductive output. In fact,
Ilmonen et al. (2000) found that female pied flycatchers injected with a vaccine had lower
17
feeding rates and fledging success (fledglings / hatched eggs) than females injected with a
saline.
Another interesting result is that intermediate morphs had a lower immune response
than lighter and darker individuals. This could be an advantage for extreme morphs and result
in disruptive selection under high exposure to parasites. However a higher antibody
production would have a higher cost and may result in a lower fitness when exposition to
parasites and/or parasites virulence are low. Moreover, other components of the immune
system might be differently related to melanism. In fact, when feral pigeons were injected
with phytohaemagglutinin PHA (an antigen eliciting a cellular response), lighter birds had a
lower response than darker birds (Jacquin et al. in press). Therefore, different components of
the immune system seem to be linked to melanin-based coloration in a complex way
(Gasparini et al. 2009b). Moreover, our study failed to show any significant link between
antibody production and body mass loss or reproductive output. The impact of these
differences in humoral response on fitness thus remains to be tested, for example by testing
directly parasite resistance.
Trade-off between maintenance and reproduction
After the start of reproduction, individuals from both food treatment groups decreased
in mass, but body mass dynamics were different between groups. Individuals from non-
limited food group exhibited a slow and continuous mass loss, whereas individuals from the
limited food group showed rapid and discontinuous weight losses at the peaks of chicks
feeding. Effect of food limitation was visible only in combination with reproduction. Food
limitation also had a negative impact on investment in reproduction, measured as the number
of eggs laid over the whole experiment. On the other side, analyses suggest a negative impact
of chicks feeding on body mass changes, but no impact of incubating eggs. Our results
therefore suggest a trade-off between reproduction and maintenance.
Furthermore, differently coloured individuals reacted differently to the food treatment.
Darker individuals lost less mass than paler ones when food was unlimited, but lost more
mass than paler ones when food was restrained. Darker individuals seem thus to better convert
food into body mass in good food conditions, whereas paler ones would do better in harsh
food condition. This is consistent with previous studies conducted by Piault et al. (2009) and
Roulin et al. (2008), who demonstrated that chicks from darker barn owls and tawny owls
seemed to grow better in good food conditions than paler ones, but to lose more weight in bad
food conditions. This could be due to a higher metabolic rate of darker eumelanic individuals
caused by pleiotropic effects of melanocortin ligands on energy homeostasis control centres
(Ducrest et al. 2008).
18
However, investment in reproduction showed an opposite pattern: when food was
limited, darker individuals laid more eggs than paler ones. This suggests that pale individuals
kept a relative good body condition in harsh conditions maybe by investing less in
reproduction, whereas darker individuals produced more eggs, but at the expense of a greater
body mass loss. It thus seems that dark and pale individuals adopt different strategies facing
variable environments (Figure 9). One possible scenario is that when energy is limited, darker
individuals adopt a “r” or “quick” strategy, with higher reproductive rate and lower survival
rate, whereas lighter individuals would adopt a “K” or “slow” strategy favouring survival to
reproduction (Wilson & Mac Arthur 1967; Promislow & Harvey 1990). Consistently,
behavioural studies conducted on the same birds showed that darker females invest more time
in reproductive activities but less in grooming necessary to feathers maintenance (Bouche et
al., unpublished data).
Figure 9: Hypothetic pattern of investment in reproduction and maintenance in variable
environments for dark and light coloured individuals as suggested by this study.
It can be argued that the number of eggs laid do not necessarily reflect the
reproductive success as some eggs could be sterile or chicks mortality be high. As eggs were
cross-exchanged between nests for further experiments on parental effects, I did not evaluate
the proportion of eggs hatched and the number of youngs fledged. Murton et al. (1974)
showed that males of different morphs varied in their ability to rear chicks rather than the
number of eggs laid in a year. It would thus be interesting to further investigate the
components of breeding success under various environmental conditions and compare the
number of fledglings between differently coloured individuals.
Evaluating the possible impact of both strategies on fitness would also require to
investigate long term effects of body mass loss on survival probability, especially in the wild.
In fact, birds in captivity do not have to invest energy in moving between foraging sites and
Parasitic pressure
Food limitation
(Parasitic pressure ?)
Food limitation
DARK
DARK
light
light
Maintenance
Reproduction
19
they are not subject to predation that could decrease survival rate in a non-linear way, since
birds in poor condition are more prone to be captured. Ongoing demographic studies in wild
populations of the Parisian area will permit to compare survival rates between differently
coloured individuals and to evaluate the effects of temporal variations in the environment.
Conclusion
We found evidences of trade-offs between reproduction and maintenance, as well as
between reproduction and immunity. These correlations are often found in natural systems
and has been shown to result from structural and energetic limitations (e.g. Norris & Evans
2000). In agreement with this, the relationships found in our study were similar between
experimental conditions and between colorations, suggesting that they are physiologically
constrained.
However, differently coloured individuals seemed to adjust their life-history traits to
experimental conditions in different ways. Darker individuals maintained a higher
reproductive output whatever the food limitation and immune challenge, with a negative
impact on body mass. On the contrary, lighter individuals seemed to adapt their reproductive
output to environmental conditions, and showed lower body mass losses in harsh food
conditions (Figure 9). The link between melanin-based coloration and resistance to parasites
is more complex and seem to depend on the component of the immune system considered.
Because of such adjustment in energy allocation, differently coloured individuals may attain
similar fitness facing the same parasite and food conditions by adopting different strategies.
Further studies on the field are now called for to test whether such strategies could lead to
different population dynamics in variable environments and how this could explain coloration
polymorphism maintenance in urban environment.
Acknowledgments
I would like to thank L. Jacquin and J. Gasparini for associating me in their research
project and helping me to complete the lab work as well as the writing of this report; C.
Haussy for introducing me to ELISA procedures; and S. Perret, L. Blottière, P. Bouche and S.
de la Bardonnie as well as the other students at the Foljuif research station who helped me
during manipulations on the birds.
20
References
BOUCHE P., GASPARINI J., JACQUIN L. (2010). Coloration mélanique et stratégies d’histoire
de vie: une étude comportementale chez le Pigeon biset, Columba livia. Rapport de Master 1.
Université de Bordeaux.
BYERS D.L. (2005) Evolution in heterogenous environments and the potential of maintenance
of genetic variation in traits of adaptive significance. Genetica 123, pp. 107 – 124
DUCREST, A. L., KELLER, L. AND ROULIN, A. (2008). Pleiotropy in the melanocortin system,
coloration and behavioural syndromes. Trends in Ecology and Evolution 23, pp. 502 – 510.
DUFFY D.L., BALL G.F. (2002). Song predicts immuno-competence in male European
starlings (Sturnus vulgaris). Proceedings of the Royal Society London B 269, pp. 847 – 852.
GASPARINI J., BIZE P., PIAULT R., WAKAMATSU K., BLOUNT J.D., DUCREST A.-L.,
ROULIN A. (2009a). Strength and cost an induced immune response are associated with a
heritable melanin-based color trait in female tawny owls. Journal of Animal Ecology 78, pp.
608 – 616.
GASPARINI J., PIAULT R., BIZE P., ROULIN A. (2009b). Synergistic and antagonistic
interaction between different branches of the immune system is related to melanin-based
coloration in nestling tawny owls. Journal of Evolutionary Biology 22, pp. 2348 – 2353.
GRINDSTAFF J.L., HASSELQUIST D., NILSSON J.-Å., SANDELL M., SMITH G.S., STJERNMAN
M. (2006). Transgenerational priming of immunity: maternal exposure to a bacterial antigen
enhances offspring humoral immunity. Proceedings of the Royal Society London B 273, pp.
2551 – 2557.
GILLESPIE J.H., TURELLI M. (1989) Genotype–environment interactions and the
maintenance of polygenic variation. Genetics 114, pp. 333 – 343.
HASSELQUIST D., WASSON M. F., WINKLER D. W. (2001). Humoral immunocompetence
correlates with date of egg-laying and reflects work load in female tree swallows. Behavioral
Ecology 12, pp. 93 – 97.
HAWKINS ET AL. (2001). Laboratory birds, refinements in husbandry and procedures, Fifth
report of the BVAAWF/FRAME/RSPCA/UFAW Joint Working Group on Refinement.
35, pp. 109 – 115.
HEDRICK P.W. (2006). Genetic polymorphism in heterogeneous environments: the age of
genomics. Annual Review of Ecology, Evolution and Systematics 37, pp. 67 – 93.
ILMONEN P., TAARNA T., HASSELQUIST D. (2000). Experimentally activated immune
defence in female pied flycatchers results in reduced breeding success. Proceedings of the
Royal Society London B 267, pp. 665 – 670.
JACQUIN L., LENOUVEL P., HAUSSY C., DUCATEZ S., GASPARINI J. (in press). Melanin-
based coloration is related to parasite intensity and cellular immune response in an urban free
living bird: the feral pigeon.
JEFFERIES D.J., PARSLOW J.L.F. (1976). The genetics of bridling in guillemots from a study
of hand-reared birds. Journal of Zoology London 179, pp. 411 – 420.
JOHNSTON R. F., JANIGA M. (1995). Feral pigeons. Oxford University Press, Oxford.
LOCHMILLER R.L., DEERENBERG C. (2000). Trade-offs in evolutionary immunology: just
what is the cost of immunity? Oikos 88, pp. 87 – 88.
MACARTHUR R.H., WILSON E.O. (1967), The Theory of Island Biogeography, Princeton,
Princeton University Press.
21
MURTON R. K., THEARLE R. J. P., COOMBS C. F. B. (1974). Ecological studies of the feral
pigeon Columba livia var. III. Reproduction and plumage polymorphism. Journal of Applied
Ecology 11, pp. 841 – 854.
NORRIS K., EVANS M.R. (2000). Ecological immunology: life history trade-offs and immune
defense in birds. Behavioral Ecology 11, pp. 19 – 26.
OBUKHOVA N. Y. (2001). Geographic variation of color in the synanthropic Blue Rock
Pigeon. Russian Journal of Genetics 37, pp. 649 – 658.
OHLSSON T., SMITH H.G., RÅBERG L., HASSELQUIST D. (2002). Pheasant sexual ornaments
reflects nutritional conditions during early growth. Proceedings of the Royal Society London
B 269, pp. 21 – 27.
PIAULT R., GASPARINI J., BIZE P., JENNI-EIERMANN S., ROULIN A. (2009). Pheomelanin-
based coloration and the ability to cope with variation in food supply and parasitism. The
American Naturalist 174 (4), pp. 548 – 546.
PROMISLOW D.E.L., HARVEY P.H. (1990). Living fast and dying young: a comparative
analysis of life history variation among mammals. Journal of Zoology 220, pp. 417 – 437.
ROULIN A., GASPARINI J., BIZE P., RITSCHARD M., RICHNER H. (2008). Melanin-based
colorations signal strategies to cope with poor and rich environments. Behav Ecol Sociobiol
62, pp. 507 – 519.
ROULIN A. (2004). The evolution, maintenance and adaptive function of genetic colour
polymorphism in birds. Biology Reviews 79, pp. 1 – 34.
ROULIN A., JUNGI T. W., PFISTER H., DIJKSTRA C. (2000). Female barn owls (Tyto alba)
advertise good genes. Proceedings of the Royal Society London B 267, pp. 937 – 941.
ROULIN A., RICHNER H., DUCREST A.-L. (1998). Genetic, environmental and condition-
dependent effects on female and male plumage ornamentation. Evolution 52, pp. 1451 – 1460.
