Behavioral Ecology Vol. 11 No. 3: 246-259
© 2000 International Society for Behavioral Ecology
Sex-limited expression of ornamental feathers in birds
a Estación Biológica de Doñana, Consejo Superior de Investigaciones Científicas, Pabellón del Perú, Avenida María Luisa s/n, E-41013 Sevilla, Spain b Laboratoire d'Ecologie, CNRS UMR 7652, Université Pierre et Marie Curie, Bât. A, 7ème étage, 7 quai St. Bernard, F-75252 Paris Cedex 05, France
Address correspondence to J. J. Cuervo. E-mail: jcuervo{at}eeza.csic.es .
Received 4 August 1998; revised 26 June 1999; accepted 26 June 1999.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXREFERENCES |
|---|
Extravagant secondary sexual characters show sexual size dimorphism in some species but are completely sex limited in others. Sexual ornamentation has been hypothesized to benefit mainly males through sexual selection, but the costs of secondary sexual characters initially would be experienced by both sexes. The evolution of sexual size dimorphism of ornaments and, eventually, the complete sex-limited expression of these characters, will depend on the effects of sexual and natural selection on the two sexes. A phylogenetic analysis controlling for similarities due to common ancestry of 60 independent evolutionary origins of feather ornamentation in birds was used to investigate ecological factors correlated with sexual size dimorphism and sex-limited expression of secondary sexual characters. When the size of an ornament is large relative to body size, the trait will be particularly costly for females, resulting in selection for increased sexual size dimorphism of the ornament. Indeed, sexual size dimorphism of ornaments was positively related to the relative size of male ornaments but was unrelated to relative size of female ornaments. Species with polygynous and lekking mating systems with little or no male parental care (in particular nest building and incubation) demonstrated sex-limited expression of ornaments as compared to monogamous species. Species with no food provisioning of offspring by the male showed a trend for increased sexual size dimorphism of ornaments. Therefore, large natural selection costs during reproduction imposed by the expression of secondary sexual characters are related to the evolution of sexual size dimorphism of ornaments and eventually their complete loss from females.
Key words: sex limitation, sexual selection, sexual size dimorphism.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXREFERENCES |
|---|
Sexual selection arises as a consequence of variation in mating success, being nonrandomly related to phenotypic characters that are advantageous during competition for mates (Darwin, 1871
). Such characters are termed secondary sexual characters, and
two different processes can account for their evolution: intrasexual
competition (usually male-male competition) and mate choice (usually female
choice of mates) (Andersson,
1994; Darwin,
1871). Observational and experimental evidence suggests that
feather ornaments of birds play an important role in female choice, but they
are of no or little significance in male-male competition (reviews in
Andersson, 1994;
Møller, 1994). Female
choice is therefore presumed to account for the maintenance of extravagant
plumage ornaments in birds.
Feather ornaments of birds are usually large and conspicuous morphological
characters, and they are therefore presumably costly to produce and maintain
(Andersson, 1994;
Møller, 1996). Natural
selection costs of ornamentation have been hypothesized to include the costs
of production of the ornament, but also may include costs of predation due to
increased attraction of predators (e.g.,
Endler, 1980;
Götmark,
1993; Magnhagen,
1991; Møller and
Nielsen, 1997), costs of a suppressed immune system (e.g.,
Folstad and Karter, 1992;
Saino and Møller,
1996), and the physiological costs of carrying an extravagant
exaggerated character (e.g., Saino et al.,
1997; but see Cuervo et al.,
1996b). Although these costs have only been studied in males,
partial expression of male traits by females is also likely to be costly.
The occurrence of ornamental feathers is often limited to the male sex, but
a number of species show partial expression of secondary sexual characters in
females. The degree of sex limitation of ornamental feathers can be classified
as females not expressing ornaments (total sex limitation) or females with
shorter extravagant ornamental feathers than males (partial sex limitation).
Among species with partial sex limitation, we can find different degrees of
sexual size dimorphism of ornaments. Whether ornaments are expressed in
females, and the degree of sexual size dimorphism of these characters, may
depend on the costs to females of developing and carrying an exaggerated
trait. The evolution of sexual size dimorphism is presumably a process
governed by the differential effects of selection on individuals of the two
sexes. Phenotypic characters often have strongly positive genetic correlations
between the sexes (Falconer,
1989), and this is also the case for secondary sexual characters
expressed in both sexes (Møller,
1993; Wilkinson,
1993). Any selection for increased size of a character among
individuals of one sex will therefore result in a correlated response to
selection among individuals of the other sex. However, changes of genetic
correlations are presumably caused by oppositely directed selection pressures
in the two sexes (Lande and Arnold,
1985). The average phenotype of the two sexes together would
evolve on a fast time scale, while sexual dimorphism would evolve on a slow
time scale. For a single character in each sex, the ratio of the slow time
scale to the fast time scale would be (1 + 
)/(1 - ), where
is the additive genetic correlation between the sexes
(Lande, 1980). The rate of
evolution of sexual dimorphism when is high (
0.9) may be one or
more orders of magnitude slower than that for the average phenotype in a
population. The evolution of sexual size dimorphism initially proceeds slowly,
but would eventually increase in speed as the genetic correlation is reduced
due to genetic modifiers that change the expression of the character in
females (Lande and Arnold,
1985). As the female trait eventually is reduced in size, the
natural selection costs of its expression would also be reduced. Vestigial
forms of a male trait may therefore be expressed in females for an extended
period of time (Lande and Arnold,
1985).
Although the theory of evolution of sexual size dimorphism hypothesized
above is relatively well understood (e.g.,
Lande, 1980;
Lande and Arnold, 1985), there
are few empirical tests available. A comparative study of sex differences in
mortality rates of birds revealed a positive relationship between sexual
dichromatism and mortality (Promislow et
al., 1992). However, a subsequent study based on a larger data set
provided a less clear-cut result, with parental care rather than sexual
dichromatism accounting for sex differences in mortality among adult birds
(Owens and Bennett, 1994). A
study of sex differences in mortality in waterfowl with relatively similar
ecology, and a complete absence of male parental care, demonstrated a positive
relationship between male biased mortality and sexual dichromatism, thus
ruling out any confounding influence of male parental care
(Promislow et al., 1994). Such
sex-biased mortality is assumed to be due to sex differences in exposure to
predators during reproduction. A comparative analysis of passerines and sex
roles in parental care showed that dull female plumage was associated with
nest sites with a high risk of predation on females during incubation and
brooding, whereas this was not the case for the brightness of male plumage
(Martin and Badyaev, 1996).
Furthermore, Badyaev (1997) has
recently shown that both male and female coloration correlates with clutch
size in the cardueline finches, but the signs of the correlation differ for
the sexes: males were brighter in species with large clutches, but females of
these species were less bright. Irwin
(1994) found that plumage
dichromatism in Icterinae was greater in polygynous species than in monogamous
ones, in particular due to changes in female plumage brightness.
While all the studies mentioned above have focused on sexual dichromatism in a single avian group, we have studied sexual size dimorphism of ornaments across all avian families. We tested for the ecological correlates of sexual size dimorphism and complete sex-limited expression of secondary sexual characters using birds with extravagant feather ornaments as a model system. Our predictions are based on the assumption that females do not benefit from secondary sexual characters, or, at least, they benefit from these characters through sexual selection to a much smaller degree than males. However, both sexes are impaired by ornaments through natural selection. In males the costs of ornaments through natural selection can be balanced by their sexual selection benefits. In females, however, costs of ornaments have little or no compensation, and ornaments will tend to diminish. Consequently, the larger the natural selection costs imposed, the larger the degree of sex limitation of ornaments.
We tested four predictions concerning sexual size dimorphism and sex-limited expression of secondary sexual characters. First, the expression of ornaments in females will depend on the size of the secondary sexual character in males because only relatively large characters will be sufficiently costly to select for sex-limited expression. In this prediction we are assuming that, everything else being equal, species with the largest ornaments also incur the largest costs due to ornamentation. These natural selection costs due to ornamentation might be balanced by sexual selection benefits in males, but not in females. Obviously, species differing in ornament length may also differ in ecology or life history that affect the cost of ornaments.
Second, the evolution of sex-limited expression of ornaments will depend on
the mating system because a more extreme skew in male mating success from
monogamy over polygyny to lekking is presumably associated with more intense
sexual selection for ornament expression in males
(Darwin, 1871);
Møller and Pomiankowski,
1993). As a consequence, there would be more intense natural
selection against expression of the male trait in females due to the increased
role of the female in reproduction. If females provide most or all parental
care, as in polygynous and lekking species
(Darwin, 1871;
Orians, 1969), there is
particularly strong natural selection against expression of secondary sexual
characters in females, for example, due to predation during reproduction
(Martin and Badyaev, 1996;
Promislow et al., 1992,
1994). We tested for the
importance of different parental duties on the evolution of sex limitation of
extravagant secondary sexual characters.
Third, migration is a widespread but energetically costly activity in
birds. The costs of expression of secondary sexual characters are likely to be
elevated in migratory as compared to resident species, simply due to the costs
of flight with extravagant, aerodynamically non-functional feathers, and this
should affect the sex limited expression of secondary sexual characters
(Balmford et al., 1993).
According to our predictions large natural selection costs due to
ornamentation would be related to a large degree of sex limitation of
ornaments. Therefore, migrants are expected to show greater sexual size
dimorphism of ornaments than nonmigrants. Fourth, the predictability of food
and the foraging mode may potentially also play important roles in the
evolution of sex limitation of ornaments. The relative amount of the energy
budget spent on locomotion may be temporally highly variable if the food
resource is unpredictable in time or space. A large degree of variation in the
costs of locomotion will put upper limits to the degree of ornamentation in
both sexes, but mainly in females. Unpredictable food such as animal food and
expensive foraging modes such as aerial insectivory should be related to a
higher degree of sex limitation than more predictable food and less costly
foraging modes.
We emphasize that each prediction (i.e., the relationship between the degree of sex-limited expression of ornaments and the ecological variables) is based on an assumption of everything else being equal. For example, aerial feeding would be related to large degrees of sex limitation of ornaments as compared to ground foraging if the two groups of species do not differ in other variables (relative ornament size, male provisioning, migratory habits). Obviously, all the variables we are studying might be related to one another. The four predictions were investigated for a number of evolutionary events of extravagant feather ornamentation in birds using a phylogenetic approach.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXREFERENCES |
|---|
Definition of feather ornaments
We have identified independent evolutionary events of feather ornamentation in extant birds. We excluded cases of extravagant feather characters in both sexes when there was no sexual size dimorphism, although mutual sexual selection may account for such exaggerated monomorphism (Jones and Hunter, 1993
). We
admit that more studies have to be performed before we can dismiss mutual
sexual selection as an important factor in the evolution of exaggerated sexual
size monomorphism. For the time being we assume that extravagant sexually size
dimorphic traits are associated with sexual selection, as demonstrated by
numerous observational and experimental studies (see
Andersson, 1994). Sexual size
monomorphism was not the subject of the present study. Furthermore, we have
not considered either feather colors or naked skin patches in the present
study.
Sexually size-dimorphic traits were considered to qualify as secondary
sexual characters if there was a sex difference in their absolute size of at
least 5% because previous studies have considered this cut-off point for
dimorphism
(Höglund,
1989;
Höglund and
Sillén-Tullberg, 1994;
Oakes, 1992). Species
investigated were recorded from extensive searches of the literature (see
Appendix) and major European museum collections (see Acknowledgment).
Representative species of all independent evolutionary events of extravagant
feather ornamentation (see "Phylogenetic information" below) that
were suggested to be sexually size dimorphic in a feather character in the
consulted literature were investigated by measuring 10 males and 10 females,
although a few species could not be measured due to their rarity and hence a
shortage of specimens in the museum collections visited. A total of 82 species
investigated resulted in 12 being considered to be sexually size monomorphic,
according to the criterion stated above, and 70 sexually dimorphic in ornament
size. Because the phylogeny was not known for all species, we could only
include 60 of these species in our analyses; 49 out of these 60 species were
classified as showing partial sex limitation of ornaments. We could not
measure female ornaments for Pteridophora alberti (female specimens
were not available in the museum collections visited), and Rollulus
rouloul (female specimens had ornamental feathers in a poor state). These
two species were classified as showing partial sex limitation of ornaments
because the literature (see references in the Appendix) clearly showed that
females were ornamented, but with much smaller ornaments than males. The
remaining 11 species with a feather character only being expressed in males
were all considered to be sexually size dimorphic with complete sex limitation
of the character. We could not measure female specimens for Pipra
cornuta (they were unavailable in the museum collections visited), but
this species was classified as showing total sex limitation of ornaments
because the literature (see references in Appendix) clearly showed that males
were ornamented but females were not. Female feathers were considered
ornaments when they were larger than expected for a particular feather
tract—that is, longer than ordinary feathers, as compared with other
feathers in the same species or equivalent feathers in closely related
species.
The degree of dimorphism of ornaments {[(male size—female size)/(female size)] x 100} among species with partial sex limitation of ornament expression ranged from 5.0% to 450.9%, with a mean value of 67.9% (SE = 14.0, n = 47 species). Moreover, ornament dimorphism was not due to a general difference in body size between sexes because in the 43 species with males showing longer wings than females, ornament dimorphism (as defined above) was on average 12.2 times larger than wing dimorphism (SE = 2.8), and always more than 1.5 times larger.
Phylogenetic information
In this study we used the phylogeny of Sibley and Ahlquist
(1990) to identify
evolutionarily independent events of extravagant feather ornamentation in
birds. Despite much criticism, the phylogeny of Sibley and Ahlqist
(1990) has been verified in a
large number of cases by independent phylogenetic studies
(Bleiweiss et al., 1995;
Harshman, 1994;
Mooers and Cotgreave, 1994;
Sibley, 1994). For the family
Hirundinidae, information on intrafamilial phylogenetic relationships is
available that allowed discrimination of the number of intrafamilial
independent evolutionary events of extravagant feather ornaments
(Sheldon and Winkler, 1993).
We have only used phylogenetic information based on DNA-DNA hybridization
(Sheldon and Winkler, 1993;
Sibley and Ahlquist,
1990).
Feather ornamentation has evolved independently a large number of times. If no other phylogenetic information was available, we assumed that there was only a single evolutionary event in each family. If ornaments appeared in subfamilies or tribes that were phylogenetically separated, these were counted as evolutionarily independent events. However, if, for example, an extravagant tail had evolved in one species and an extravagant head plume had evolved in another species of the same family, we assumed that they represented two independent evolutionary events, since these traits were obviously developmentally and morphologically independent. If more than a single ornamented species was available within a taxon, we exclusively used abundance as the criterion for choice of species to be used in our analysis, due to more ecological information being available for abundant species.
We have not found resolved phylogenies for all 70 species classified as
ornamented. Because some methods of comparative analysis (Pagel,
1994,
1997) cannot deal with
polytomies (node with more than two descendant nodes), we have excluded from
our analyses 10 of the 70 species in order to achieve a perfectly bifurcated
phylogeny. We have in these cases maximized the number of contrasts. Every
branch in the phylogeny was considered to have the same length. All 60 species
included in the study and their phylogenetic relationships are shown in
Figure 1.
|
Ecological variables
For all bird species considered, we made an extensive search in the
literature for information concerning mating system, parental care, diet,
migration, and additional natural history variables potentially influencing
the evolution of sexual size dimorphism and sex limitation of feather
ornaments. These references are listed in the Appendix. Four different mating
systems have been considered: (1) social monogamy (n = 38 species) if
single males and single females associated for reproduction, (2) polygyny
(n = 6) if at least 5% of the males in one population were associated
with more than a single female for reproduction, (3) polyandry (n =
1) if at least 5% of the females were associated with more than a single male
for reproduction, and (4) lekking (n = 15) if males aggregated at
communal display grounds where females arrived to make their mate choice.
Category 1 was considered monogamy and categories 2-4 were considered polygamy
throughout the analyses. Male parental care was divided into three categories:
nest building, incubation, and feeding of offspring. Species were classified
as having no or some male contribution for each of the three categories (21
species with male nest building, 31 without; 16 species with male incubation,
37 without; 26 species with male young provisioning, 26 without).
Our data show that social mating system is significantly related to male
nest building [omnibus test, likelihood ratio (LR) = 15.38, p
(simulation) <.01], male incubation (LR = 18.86, p <.01), and
provisioning of young by the male [LR = 34.69, p <.01; general
method of comparative analysis for discrete variables (Pagel,
1994,
1997); see "Statistical
procedures" below]. Species with a high skew in male mating success
showed less male contribution to parental care.
Bird species were classified according to their migratory regime as migrants (n = 4 species), partial migrants (n = 18), or residents (n = 37) depending on whether there was no overlap, some overlap, or complete overlap between breeding and nonbreeding ranges due to seasonal movements. Diet was classified in three categories: mainly animal food (n = 23 species), mainly vegetable food (excluding fruit) (n = 23), and mainly fruit (n = 13). Omnivorous species were classified as relying on animal or vegetable food depending on the most important contribution to the diet. For our analyses we pooled vegetable and fruit eaters. The common mode of locomotion while foraging was classified as aerial (n = 13 species), diving (n = 2), swimming (n = 2), perching (n = 22), and ground foragers (n = 20). Aerial foragers obtain all their food from pursuing food (usually invertebrates) in flight. Divers pursue food while diving. Because our emphasis was on the consequences of costly behavior on the evolution of sex-limited expression of extravagant ornaments, for the analyses we separated species with costly foraging modes (aerial, diving) from the others (swimming, perching, ground foragers). Ornament categories and ecological variables of all species are listed in the Appendix.
Measurement of specimens
For most species we measured 10 males and 10 females, although in some
cases (see Appendix) it was impossible to obtain this number of adult
specimens in breeding plumage and good feather condition. Individuals with
broken or worn feathers were excluded. The mean number of specimens per
species and sex was 9.9 ± 0.5 SD, with a minimum value of 7. Specimens
were chosen in the order they appeared in the collections, which prevents any
involuntary bias in sampling. We were especially careful in excluding
specimens in molt by checking all specimens for the presence of feather
quills. If ornaments only appear during part of the year, only specimens from
that period were considered. For each species we measured the length of right
and left flattened wing and the maximum length of right and left sides of
ornaments to the nearest millimeter using a ruler. Measurements were made
according to Svensson (1984).
The size of phenotypic characters of specimens was simply the mean value of
the right and the left character. Summary statistics for all measurements are
given in the Appendix.
All specimens of each species measured belonged to the same subspecies and, when possible, to the same population. In Hydrophasianus chirurgus, females were more ornamented than males due to the polyandrous mating system, and female measurements were therefore included in the analyses as "male measurements" and male measurements as "female measurements."
We assessed the repeatabilities of our measurements in four species
(Anas platyrhynchos, Hirundo rustica, Sturnus unicolor, and
Vanellus vanellus) with different kinds of ornaments and different
body sizes by measuring the same individuals (right and left sides of wings
and ornamental feathers) on 2 different days without knowledge of the results
obtained on the first day. Repeatabilities
(Becker, 1984) ranged from
0.989 to 0.999. In all 16 cases F 188.8 and p
<.0001. For Hirundo rustica and Sturnus unicolor df =
29,30; for Anas platyrhynchos df = 27,28; for Vanellus
vanellus df = 30,31 (wing feathers) or df = 27,28 (crest feathers).
Repeatabilities were large, suggesting that our measurements were sufficiently
precise to allow quantitative analyses.
Statistical procedures
We made two separate types of analyses to investigate the possible
relationship between sex-limited expression of ornaments and different
ecological variables. First, among species with partial sex limitation of
ornaments we investigated the relationship between the degree of sexual size
dimorphism of ornaments and the ecological variables. Second, we compared the
group of species with total sex limitation with the group of species with
partial sex limitation with respect to the ecological variables. These two
types of analyses allowed assessment of ecological variables being associated
with differences in sexual size dimorphism of ornaments as well as differences
in ecological variables being related to total sex limitation of ornaments.
Obviously, if particular ecological conditions affect the evolution of sexual
size dimorphism of ornaments, we should expect the same conditions eventually
to give rise to complete sex limitation.
Relative size of ornaments was calculated for each sex using the
statistical software CAIC to control for similarity due to common descent
(Purvis and Rambaut, 1995).
First, we analyzed log10-transformed ornament length and
log10-transformed wing length together, using the Crunch procedure,
and regressed the contrasts (independent standardized linear contrasts) of the
dependent variable (ornament length) on the contrasts of the independent
variable (wing length) through the origin
(Purvis and Rambaut, 1995).
The expected value of the slope equals the true relation between the two
variables in the absence of phylogenetic effects
(Pagel, 1993). Next, we fitted
the slope of this regression to the original log10-transformed wing
data and calculated the expected values of log10-transformed
ornaments. Original log10-transformed ornament data minus the
expected values will give us the residuals of the regression. These residuals
represent relative ornament length independent of body size for each sex. We
used wing length instead of body mass as a measure of body size because
insufficient body mass data were available in the literature.
To calculate sexual size dimorphism of ornaments, we analyzed relative size
of male ornament and relative size of female ornament again using the Crunch
procedure and regressed the contrasts of male ornament on the contrasts of
female ornament through the origin. As above, we fitted the slope of this
regression to the original relative size of ornaments and calculated residuals
from this line. We have used these residuals as a measure of sexual size
dimorphism of ornaments. Sexual size dimorphism of ornaments was only
calculated for species where both males and females had ornaments. Our method
of calculating sexual size dimorphism of ornaments not only controls for
similarities due to common descent, but also for possible allometric
relationships between the size of the character in the two sexes
(Ranta et al., 1994; Rensch,
1950,
1959). Previous empirical
studies have neglected this allometry effect when investigating the
relationship between the ratio of male size to female size and the size of
females, for example in reptiles (e.g.,
Shine, 1991), birds (e.g.,
Höglund,
1989; Møller,
1986; Payne,
1984), or mammals (e.g.,
Clutton-Brock et al., 1977;
Kappeler, 1991).
To investigate the relationship between the degree of sexual size
dimorphism of ornaments and the ecological variables, we used the Brunch
procedure of the program CAIC (Purvis and
Rambaut, 1995). This procedure allows tests of whether the
evolution of one continuous variable (sexual size dimorphism of ornaments) is
related to the evolution of one categorical variable (all our ecological
variables). We have reduced all the ecological variables to have only two
states. A positive contrast for sexual size dimorphism of ornaments at a node
means that this variable is varying in the same direction as the categorical
variable. Under the null hypothesis that evolution in the continuous variable
has not been linked to the evolution of the categorical variable, we should
expect half the contrasts in the dependent variable to be positive and half
negative and the mean value of the contrasts to be zero. We have tested this
null hypothesis using one sample t tests on the mean contrasts for
each analysis. For example, if polygynous species are coded "1"
and monogamous species "0," a positive mean value would imply that
sexual size dimorphism of ornaments tends to be larger in polygynous species
and a negative value that sexual size dimorphism of ornaments tends to be
larger in monogamous species.
To investigate the total or partial sex limitation of ornaments (a
categorical variable) with respect to the ecological variables (also
categorical variables), we used the general method of comparative analysis for
discrete variables proposed by Pagel
(1994,
1997). Pagel's method develops
maximum likelihood estimates of the rates of change in the discrete characters
and tests the hypothesis of their correlated evolution without relying on
reconstructions of the ancestral character state
(Pagel, 1994). A likelihood
ratio test statistic (omnibus test) is used to discriminate between two models
that are fitted to the data: one allowing only for independent evolution of
the two characters (four parameter model), the other involving correlated
evolution (eight parameter model). The significance of this likelihood ratio
test is assessed using Monte Carlo simulations. Tests of specific directional
hypotheses can also be made. These can include tests of whether changes in one
variable are more or less likely given the state of the other (contingent
changes test), and tests of the temporal ordering and direction of changes
(temporal order test). These hypotheses are tested by forcing certain
parameters (qij) in the matrix of transition
probabilities to take the same value and fitting that model to the data by
maximum likelihood. This model (seven-parameter model) is then compared to the
model of correlated evolution (eight-parameter model) by means of a likelihood
ratio test. Likelihood ratios will be asymptotically distributed as
2 with 1 df (see Pagel,
1994,
1997). It is also possible to
force each parameter of the model to zero and compare the models obtained in
each case to the full model of dependent evolution (i.e., the eight parameter
model). This allows one to construct a flow diagram of evolutionary changes.
Again, every variable has only two states, and we have assumed a model of
punctuated evolution; that is, every branch in the phylogeny is the same
length.
Statistical tests were performed according to Sokal and Rohlf
(1995) and Zar
(1984). All tests are
two-tailed and the level of significance is 5%.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXREFERENCES |
|---|
Ecology and sexual size dimorphism of ornaments
The first series of analyses determined the relationship between sexual size dimorphism of ornaments (only in species with partial sex-limited expression of ornaments) and the ecological factors. Sexual size dimorphism of ornaments was positively related to the relative size of male ornaments [Figure 2; F = 8.99, df = 1,45, r2 =.17, p =.0044, slope (SE) = 0.167 (0.056)]. However, sexual size dimorphism of ornaments was far from significantly related to the relative size of female ornaments [F = 0.00, df = 1,45, r2 =.00, slope = -0.0002 (0.0610)]. Moreover, the difference between the slopes of the two regressions was marginally significant (t = 2.022, df = 44, p =.0496).
|
Our analyses based on standardized linear contrasts did not show any significant difference in sexual size dimorphism of ornaments between socially monogamous and polygynous/lekking species [mean contrast (SE) = 0.032 (0.025), t = 1.29, df = 7]. No significant differences in sexual size dimorphism were found for incubation [mean contrast = -0.014 (0.019), t = -.75, df = 11] or nest building by males [mean contrast = -0.020 (0.023), t = -0.88, df = 10]. However, sexual size dimorphism of ornaments was slightly less in species in which males fed the young as compared to those in which they did not feed [mean contrast = -0.037 (0.016), t = -2.26, df = 8, p =.054].
We found no significant relationship between sexual size dimorphism of ornaments and migration. We used two approaches: first, we combined migratory and partially migratory species [mean contrast (SE) = 0.012 (0.024), t =.49, df = 11], and second, we combined resident and partially migratory species (no test possible because there were only two independent contrasts). Neither diet [mean contrast = 0.027 (0.029), t = 0.92, df = 11] nor foraging mode [mean contrast = 0.018 (0.028), t = 0.64, df = 6] were significantly related to sexual size dimorphism of ornaments.
Ecology and relative size of ornaments
Given that sexual size dimorphism of ornaments depends on the relative size
of ornaments in the two sexes, it is important to test whether the ecological
variables are correlated with relative size of ornaments separately for each
sex. For example, lack of correlation between an ecological factor and sexual
size dimorphism of ornaments could be due to absence of an effect in either
sex, but also by an equal effect of the ecological factor on both sexes.
Moreover, although it has been traditionally assumed that sexual dimorphism
arises because of changes in male traits, some studies have shown that changes
in female traits could be the origin of sexual dichromatism in birds
(Björklund,
1991; Burns, 1998;
Irwin, 1994;
Martin and Badyaev, 1996).
Our analyses based on standardized linear contrasts and including all the species did not show any significant relationship between relative size of male ornaments and mating system [mean contrast (SE) = 0.033 (0.043), t = 0.78, df = 11], male incubation [mean contrast = -0.021 (0.062), t = -0.35, df = 11], nest building by males [mean contrast = 0.028 (0.058), t = 0.48, df = 12], provisioning of young by males [mean contrast = -0.007 (0.061), t = -0.12, df = 10], migration [combining migratory and partially migratory species; mean contrast = -0.053 (0.074), t = -0.71, df = 11], diet [mean contrast = -0.093 (0.059), t = -1.59, df = 13], or foraging mode [mean contrast = -0.028 (0.093), t = -0.30, df = 6].
Similarly, relative size of female ornaments, including only species with ornamented females, was not significantly related to mating system [mean contrast (SE) = 0.019 (0.081), t = 0.23, df = 7], male incubation [mean contrast = -0.001 (0.071), t = -0.01, df = 11], nest building by males [mean contrast = 0.017 (0.064), t = 0.27, df = 10], young provisioning by males [mean contrast = -0.0005 (0.082), t = -0.01, df = 8], migration [combining migratory and partially migratory species; mean contrast = -0.030 (0.079), t = -0.38, df = 11], diet [mean contrast = -0.117 (0.059), t = -1.99, df = 11, p =.072), or foraging mode (mean contrast = -0.017 (0.098), t = -0.17, df = 6].
Ecology and sex-limited expression of ornaments
We investigated the relationship between presence of ornamental feathers in
females and relative size of male ornaments, mating system, male nest
building, male incubation, male provisioning of chicks, migration, diet, and
foraging mode.
The relative size of male ornaments was not significantly related to the
presence of ornaments in females [mean contrast (SE) = -0.112 (0.080),
t = -1.40, df = 9]. Regarding mating system, females of socially
monogamous species showed partial sex limitation of ornaments, and females of
polygynous/lekking species showed total sex limitation of ornaments
significantly more often than expected by chance [omnibus test, LR = 8.01,
p (simulation) =.020). None of the two contingent change tests or the
four temporal order tests of the relationship between mating system and degree
of sex limitation of ornaments (partial or total) was significant (LR 
0.08). The only significant transitions were (polygyny, partial limitation)
(polygyny, total limitation) (LR 3.92, p <.05, in the
two cases; Figure 3a). The
complete elimination (and the acquisition) of ornamentation in females
occurred significantly more often than expected in polygynous/lekking species
but not in socially monogamous species.
|
Sex-limited expression of ornaments was significantly related to nest
building and incubation by the male in a similar way: when males participated
in parental care, females tended to show partially expressed ornaments, but in
the absence of male participation females were significantly more likely to
show total limitation of ornaments (nest building: omnibus test, LR = 13.54,
p <.01; incubation: omnibus test, LR = 9.37, p =.015).
None of the contingent change or temporal order tests was significant (LR
1.74). In both comparisons, the only significant transitions were (no male
care, partial limitation) (no male care, total limitation) (LR
4.54, p <.05, in the four cases;
Figure 3b, c). The complete
elimination (and the acquisition) of ornamentation in females occurred
significantly more often than expected only when males did not invest in
parental care. In contrast, sex limitation of ornaments was not significantly
related to provisioning of young by the male (omnibus test, LR = 4.78). The
nonsignificant p value from the simulation (omnibus test) implies
that the eight-parameter model does not improve the four-parameter model (see
"Statistical procedures"). Because the eight-parameter model will
always fit the data better than any seven-parameter model, this means that it
is impossible to improve the simple four-parameter model of independent
evolution. Therefore, in this case it makes no sense to perform any of the
contingent change or temporal order tests.
Sex limitation of ornamentation was unrelated to migration regime (pooling
partially migratory species either with migratory or resident species), diet
or foraging mode (omnibus test, LR 2.38, ns, in all four cases).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXREFERENCES |
|---|
Sexual selection may account for the evolution of sexual size dimorphism of ornaments in at least two different ways. First, sexual size dimorphism of ornaments could be the result of female ornaments evolving as a correlated response to selection on males because of a positive genetic correlation between the sexes, but females expressing the male trait to a smaller degree than males because of the large natural selection costs of the trait for females. Sexual size dimorphism of ornaments would in this case result from selection for genetic modifiers that control the expression of the male trait in females, with simultaneous selection against such modifiers in males (see Introduction). Responses to selection will in this case ultimately depend on the genetic architecture of the species; if there are few genetic modifiers available, positive genetic correlations will prevent the evolution of sexual size dimorphism (e.g., Meagher, 1992
).
The second explanation suggests that the female trait is an ornament
currently under sexual selection, and different intensities of natural and
sexual selection on males and females will give rise to differential
expression of the trait in males and females
(Cuervo et al., 1996a;
Darwin, 1871;
Hill, 1993;
Jones and Hunter, 1993;
Møller, 1993;
Muma and Weatherhead, 1989;
Trivers, 1972). Empirical
evidence suggests that females with the largest secondary sexual characters in
species with sexually size-monomorphic characters sometimes experience a
mating advantage (Jones and Hunter,
1993), although this is not the case in three sexually dimorphic
species studied so far (Cuervo et al.,
1996a; Hill, 1993;
Muma and Weatherhead, 1989).
Available information thus suggests that sexual dimorphism of ornaments might
be a consequence of different selection pressures combined with genetic
correlations between the sexes rather than sexual selection increasing the
expression of the secondary sexual character in both males and females.
Whatever the mechanism may be in different species, the relative strength of
natural and sexual selection in the two sexes will give rise to
sex-differential expression of ornaments.
A major result of our study relates to the prediction concerning the
relationship between the relative size of male secondary sexual characters and
the degree of sexual size dimorphism of ornaments. We have studied whether the
relative size of male ornaments, after controlling for the effects of
allometry and phylogeny, was a reliable predictor of sex limitation of
ornaments, based on the assumption that relatively large ornaments are more
costly to produce and maintain than small ones. The costs of a relatively
large ornament should disproportionately affect females if females have little
(or no) advantage of large ornaments through sexual selection that could
balance the costs of natural selection. The prediction was tested in two
different ways. First, the prediction that the relative size of an ornament
should affect the evolution of its sexual size dimorphism was investigated by
considering only species with partial expression of the male trait in females.
We found a significant positive correlation between sexual size dimorphism of
ornaments and relative size of male ornaments
(Figure 2). However, the
correlation between sexual size dimorphism of ornaments and relative size of
female ornaments was far from statistically significant. This correlation was
significantly weaker than the same correlation for males, but this result was
not a consequence of different amounts of variation in the size of ornaments
in the two sexes. First, analyses of the coefficient of variation of ornaments
in males and females did not reveal statistically different mean values for
males and females (Cuervo and
Møller, 2000). Second, the degree of divergence in
secondary sexual characters across species was not significantly different in
males and females (Cuervo and
Møller, 2000). These findings indicate that male ornament
size is a much better predictor of sexual size dimorphism of ornaments (and
hence sexual limitation of ornaments) than female ornament size. One
interpretation of this result is that it is the exaggeration of the male
secondary sexual character relative to body size that is causing natural
selection costs of the male trait in females. Hence, selection for genetic
modifiers that control the expression of the male trait in females should
proceed particularly fast in species with extreme exaggeration of male traits.
The positive correlation between sexual size dimorphism and relative size of
the male trait, but not relative size of the female trait, is the first
empirical demonstration that sexual size dimorphism (after controlling for
similarities due to common descent and for the allometric relationship between
male and female character) increases with increasing size of a phenotypic
character in one sex, but not the other (review in
Andersson, 1994).
Second, we studied the difference in relative size of secondary sexual characters in species with complete and partial sex-limited expression of ornaments. This difference was far from statistically significant. The second analysis suggests that complete sex-limited expression of male traits is unaffected by the relative size of the secondary sexual character in males and that natural selection costs of the male character for females may differ depending on the ecological context.
An ecological factor generally accepted to be related to the strength of
sexual selection in males is the social mating system, with sexual selection
supposedly being more intense in lekking than in polygynous species, and more
intense in polygynous than in socially monogamous species
(Andersson, 1994;
Darwin, 1871;
Payne, 1984). Of course,
sexual selection may also operate relatively intensely under monogamy
(Andersson, 1986;
Grafen, 1990;
Kirkpatrick et al., 1990;
Møller and Birkhead,
1994). The partial or total limitation of the expression of female
ornament was found to depend on the social mating system, but the degree of
sexual size dimorphism of ornaments among species was not significantly
related to mating system. The lack of significance of the latter analysis
should be considered with caution because it is based on a small number of
contrasts [power analysis (Cohen,
1988), power = 0.22, d =.065, a = 0.05,
n = 8). Ornamented females were mainly found in socially monogamous
species, whereas an absence of ornaments mainly occurred in polygynous and
lekking species. A previous comparative study of sexual size dimorphism in
exaggerated avian tails in relation to sexual selection also found a positive
relationship between degree of mating skew as determined by the social mating
system and sexual size dimorphism
(Winquist and Lemon,
1994).
Parental care is a time- and energy-consuming activity compared to other
activities during the annual cycle
(Clutton-Brock, 1990). Partial
or total limitation of ornaments in females should be affected by the role of
males in parental care because a larger male contribution would limit the
expression of the secondary sexual character in males and hence reduce the
natural selection costs of the male trait in females, and these costs should
not be different for males and females with similar sex roles during
reproduction (Clutton-Brock and Godfray,
1991; Trivers,
1972; Winkler and Wilkinson,
1988). Therefore, we predicted that sexual size dimorphism of
ornaments should be inversely related to male parental care. When males
contributed to parenting behavior (building the nest or incubating), species
more often only showed partial limitation of ornaments, whereas a complete
lack of male parental care was associated with an absence of exaggerated
traits in females. The relationship did not reach significance for offspring
provisioning by the male. This result is surprising because different kinds of
male parental care tend to be positively correlated across species
(Lack, 1968;
Silver et al., 1985).
Moreover, provisioning of young is generally considered to be the most
energy-demanding activity of parental care
(Clutton-Brock, 1990;
Winkler and Wilkinson, 1988),
and it should therefore be the most important determinant of sex-limited
ornamentation. Among species with ornamented females, sexual size dimorphism
of ornaments tended to be greater when males did not feed the young, but we
did not find a significant relationship for the other two parenting
activities. This finding, however, should be considered with caution because
it is based on a small number of contrasts (power analysis; nest building:
power = 0.13, d = 0.38, a = 0.05, n = 11;
incubation: power = 0.11, d = 0.30, a = 0.05, n =
12; provisioning of young: power = 0.56, d = 1.07, a = 0.05,
n = 9). Winquist and Lemon
(1994) found that all three
male parental care activities were related to sexual size dimorphism of
exaggerated tail feathers. Because monogamy is related to a higher proportion
of male contribution to parental care (see "Ecological
variables"), it is not surprising that the two variables (i.e., social
mating system and male parental care) are related to the degree of sex
limitation of ornamental feathers in a similar way.
Migration, diet, and foraging mode were not significantly associated with
sexual size dimorphism or limitation of ornaments. Although mode of locomotion
and predictability of food were hypothesized to restrict the expression of
secondary sexual characters in females more than in males, this was clearly
not the case. It is therefore unlikely that any of these variables influenced
the associations between sex limitation of ornaments and either mating system
or male parental care. Sexual size dimorphism and sex-limited expression of
secondary sexual characters were significantly associated with the social
mating system and male parental behavior, but not with migration, diet and
foraging mode. This observation suggests that selection pressures during
reproduction rather than during the nonreproductive season affect the
evolution of sexual size dimorphism of ornaments. This suggestion is also
consistent with previous studies of sexual size dimorphism and dichromatism in
birds (Martin and Badyaev,
1996; Promislow et al.,
1992,
1994).
We have found no significant relationship between sexual size dimorphism of
ornaments and the ecological variables, and only slightly less sexual size
dimorphism of ornaments in species in which males fed the young as compared to
those in which they did not feed. This lack of relationship between dimorphism
and the ecological factors could be due to absence of effect on either sex,
but also due to an equal effect of the ecological factors on both sexes. Our
analyses suggest that the former is the most plausible explanation because we
have found no significant relationship between the ecological variables and
relative size of ornaments for each sex separately. However, these results
should be also considered with caution because they are based on a small
number of standarized linear contrasts, and the power of the tests is hence
moderate to low (power analyses; in all 18 tests: power 0.47, d
0.82, a = 0.05, 5 n 14).
In conclusion, relatively large male (but not female) secondary sexual characters, high degrees of polygyny, and absence of male parental care are significantly associated with sex limitation of ornament expression in bird species with extravagant feather ornaments.
|
APPENDIX |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXREFERENCES |
|---|
?
|
|
ACKNOWLEDGEMENTS |
|---|
We are grateful to the curators of the bird collections in Alexander Koenig Museum, Bonn, Germany, British Museum (Natural History), Tring, UK, Doñana Biological Station, Seville, Spain, Natural History Museum, Stockholm, Sweden, and Zoological Museum, Copenhagen, Denmark, for access to specimens. J. Shykoff and an anonymous reviewer kindly provided valuable criticism. We thank M. Pagel for providing the programme DISCRETE and for help with the analyses. Our research was supported by a postdoctoral grant from the European Union (Human Capital and Mobility Program) and by Spanish DGICYT (PB95-0110) to J.J.C. and by Swedish and Danish Natural Science Research Council grants to A.P.M.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXREFERENCES |
|---|
Andersson, M, 1986. Evolution of condition-dependent sex ornaments and mating preferences: sexual selection based on viability differences. Evolution 40: 804-816.[Web of Science]
Andersson M, 1994. Sexual selection. Princeton, New Jersey: Princeton University Press.
Badyaev AV, 1997. Covariation between life history and sexually selected traits: an example with cardueline finches. Oikos 80: 128-138.[Web of Science]
Balmford A, Thomas ALR, Jones IL, 1993. Aerodynamics and the evolution of long tails in birds. Nature 361: 628-631.
Becker WA, 1984. Manual of quantitative genetics, 4th ed. Pullman: Academic Enterprises.
Björklund M, 1991. Coming of age in fringillid birds: heterochrony in the ontogeny of secondary sexual characters. J Evol Biol 4: 83-92.
Bleiweiss R, Kirsch JAW, Shafi N, 1995. Confirmation of a portion of the Sibley-Ahlquist "tapestry." Auk 112: 87-97.[Web of Science]
Burns KJ, 1998. A phylogenetic perspective on the evolution of sexual dichromatism in tanagers (Thraupidae): the role of female versus male plumage. Evolution 52: 1219-1224.[Web of Science]
Clutton-Brock TH, 1990. The evolution of parental care. Princeton, New Jersey: Princeton University Press.
Clutton-Brock TH, Godfray C, 1991. Parental investment. In: Behavioural ecology: an evolutionary approach, 3rd ed. (Krebs JR, Davies NB, eds). Oxford: Blackwell; 234-262.
Clutton-Brock TH, Harvey PH, Rudder B, 1977. Sexual dimorphism, socionomic sex ratio and body weight in primates. Nature 269: 797-800.[Medline]
Cohen J, 1988. Statistical power analysis for the behavioral sciences, 2nd ed. Hillsdale, New Jersey: Lawrence Erlbaum.
Cuervo JJ, de Lope F, Møller AP, 1996a. The
function of long tails in female barn swallows (Hirundo rustica): an
experimental study. Behav Ecol 7:
132-136.
Cuervo JJ, de Lope F, Møller AP, Moreno J, 1996b. Energetic cost of tail streamers in the barn swallow (Hirundo rustica). Oecologia 108: 252-258.[Web of Science]
Cuervo JJ, Møller AP, 1999. Phenotypic variation and fluctuating asymmetry in sexually dimorphic feather ornaments in relation to sex and mating system. Biol J Linn Soc 68: 505-529.
Cuervo JJ, Møller AP, 2000. Evolutionary rates of secondary sexual and non-sexual characters among birds. Evol Ecol (in press).
Darwin C, 1871. The descent of man, and selection in relation to sex. London: John Murray.
Endler JA, 1980. Natural selection on colour patterns in Poecilia reticulata. Evolution 34: 76-91.[Web of Science]
Falconer DS, 1989. Introduction to quantitative genetics, 3rd ed. New York: Longman.
Folstad I, Karter AJ, 1992. Parasites, bright males, and the immunocompetence handicap. Am Nat 139: 603-622.[Web of Science]
Götmark F, 1993.
Conspicuous coloration in male birds: favoured by predation in some species,
disfavoured in others. Proc R Soc Lond B
253: 143-146.
Grafen A, 1990. Sexual selection unhandicapped by the Fisher process. J Theor Biol 144: 473-516.[Web of Science][Medline]
Harshman J, 1994. Reweaving the tapestry: what can we learn from Sibley and Ahlquist (1990). Auk 111: 377-388.[Web of Science]
Hill GE, 1993. Male mate choice and the evolution of female plumage coloration in the house finch. Evolution 47: 1515-1525.[Web of Science]
Höglund J, 1989. Size and plumage dimorphism in lek-breeding birds: a comparative analysis. Am Nat 134: 72-87.[Web of Science]
Höglund J, Sillén-Tullberg B, 1994. Does lekking promote the evolution of male-biased size dimorphism in birds? On the use of comparative approaches. Am Nat 144: 881-889.[Web of Science]
Irwin RE, 1994. The evolution of plumage dichromatism in the new world blackbirds: social selection on female brightness? Am Nat 144: 890-907.
Jones IL, Hunter FM, 1993. Mutual sexual selection in a monogamous seabird. Nature 362: 238-239.
Kappeler PM, 1991. Patterns of sexual size dimorphism in body weight among prosimian primates. Folia Primatol 57: 132-146.
Kirkpatrick M, Price T, Arnold SJ, 1990. The Darwin-Fisher theory of sexual selection in monogamous birds. Evolution 44: 180-193.[Web of Science]
Lack D, 1968. Ecological adaptations for breeding in birds. London: Methuen.
Lande R, 1980. Sexual dimorphism, sexual selection and adaptation in polygenic characters. Evolution 34: 292-305.[Web of Science]
Lande R, Arnold SJ, 1985. Evolution of mating preference and sexual dimorphism. J Theor Biol 117: 651-664.[Web of Science][Medline]
Magnhagen C, 1991. Predation risk as a cost of reproduction. Trends Ecol Evol 6: 183-186.
Martin TE, Badyaev AV, 1996. Sexual dichromatism relative to nest height and nest predation: contributions of females versus males. Evolution 50: 2454-2460.[Web of Science]
Meagher TR, 1992. The quantitative genetics of sexual dimorphism in Silene latifolia (Caryophyllaceae). I. Genetic variation. Evolution 46: 445-457.[Web of Science]
Møller AP, 1986. Mating systems among European passerines: a review. Ibis 128: 234-250.[Web of Science]
Møller AP, 1993. Sexual selection in the barn swallow Hirundo rustica. III. Female tail ornaments. Evolution 47: 417-431.[Web of Science]
Møller AP, 1994. Sexual selection and the barn swallow. Oxford: Oxford University Press.
Møller AP, 1996. The cost of secondary sexual characters and the evolution of cost-reducting traits. Ibis 138: 112-119.[Web of Science]
Møller AP, Birkhead TR, 1994. The evolution of plumage brightness in birds is related to extra-pair paternity. Evolution 48: 1089-1100.[Web of Science]
Møller AP, Nielsen JT, 1997. Differential predation cost of a secondary sexual character: sparrowhawk predation on barn swallows. Anim Behav 54: 1545-1551.[Web of Science][Medline]
Møller AP, Pomiankowski A, 1993. Why have birds got multiple sexual ornaments? Behav Ecol Sociobiol 32: 167-176.[Web of Science]
Mooers AØ, Cotgreave P, 1994. Sibley and Ahlquist's tapestry dusted off. Trends Ecol Evol 9: 458-459.
Muma KE, Weatherhead PJ, 1989. Male traits expressed in females: direct or indirect sexual selection? Behav Ecol Sociobiol 25: 25-31.
Oakes EJ, 1992. Lekking and the evolution of sexual dimorphism in birds: comparative approaches. Am Nat 140: 665-684.[Web of Science]
Orians GH, 1969. On the evolution of mating systems in birds and mammals. Am Nat 103: 589-603.[Web of Science]
Owens IPF, Bennett PM, 1994. Mortality costs of
parental care and sexual dimorphism in birds. Proc R Soc Lond B
257: 1-8.
Pagel M, 1993. Seeking the evolutionary regression coefficient: an analysis of what comparative methods measure. J Theor Biol 164: 191-205.[Web of Science][Medline]
Pagel M, 1994. Detecting correlated evolution on
phylogenies: a general method for the comparative analysis of discrete
characters. Proc R Soc Lond B 255:
37-45.
Pagel M, 1997. Inferring evolutionary processes from phylogenies. Zool Scripta 26: 331-348.[Web of Science]
Payne RB, 1984. Sexual selection, lek and arena behavior, and sexual size dimorphism in birds. Ornithol Monogr 3: 1-53.
Promislow DEL, Montgomerie RD, Martin TE, 1992.
Mortality costs of sexual dimorphism in birds. Proc R Soc Lond
B 250:
143-150.
Promislow DEL, Montgomerie RD, Martin TE, 1994. Sexual selection and survival in North American waterfowl. Evolution 48: 2045-2050.[Web of Science]
Purvis A, Rambaut A, 1995. Comparative analysis by
independent contrasts (CAIC): an Apple Macintosh application for analysing
comparative data. Comput Appl Biosci 11:
247-251.
Ranta E, Laurila A, Elmberg J, 1994. Reinventing the wheel: analysis of sexual dimorphism in body size. Oikos 70: 313-321.[Web of Science]
Rensch B, 1950. Die Abhängigkeit der relativen Sexualdifferenz von der Körpergrösse. Bonner Zool Beiträge 1: 58-69.
Rensch B, 1959. Evolution above the species level. New York: Columbia University Press.
Saino N, Cuervo JJ, Krivacek M, de Lope F, Møller AP, 1997. Experimental manipulation of tail ornament size affects the hematocrit of male barn swallows (Hirundo rustica). Oecologia 110: 186-190.[Web of Science]
Saino N, Møller AP, 1996. Sexual ornamentation
and immunocompetence in the barn swallow. Behav Ecol
7: 227-232.
Sheldon FH, Winkler DW, 1993. Intergeneric phylogenetic relationships of swallows estimated by DNA-DNA hybridization. Auk 110: 798-824.[Web of Science]
Shine R, 1991. Intersexual dietary divergence and the evolution of sexual dimorphism in snakes. Am Nat 138: 103-122.[Web of Science]
Sibley CG, 1994. On the phylogeny and classification of living birds. J Avian Biol 25: 87-92.
Sibley CG, Ahlquist JE, 1990. Phylogeny and classification of birds. A study in molecular evolution. New Haven, Connecticut: Yale University Press.
Silver R, Andrews H, Ball GF, 1985. Parental care in an ecological perspective: a quantitative analysis of avian subfamilies. Am Zool 25: 823-840.
Sokal RR, Rohlf FJ, 1995. Biometry: the principles and practice of statistics in biological research, 3rd ed. New York: W.H. Freeman.
Svensson L, 1984. Identification guide to European passerines, 2nd ed. Stockholm: L. Svensson.
Trivers RL, 1972. Parental investment and sexual selection. In: Sexual selection and the descent of man, 1871-1971 (Campbell B, ed). Chicago: Aldine.
Wilkinson GS, 1993. Artificial sexual selection alters allometry in the stalk-eyed fly Cyrtodiopsis dalmanni (Diptera: Diopsidae). Genet Res 62: 213-222.[Web of Science]
Winkler DW, Wilkinson GS, 1988. Parental effort in birds and mammals: theory and measurement. Oxf Surv Evol Biol 5: 185-214.
Winquist T, Lemon RE, 1994. Sexual selection and exaggerated male tail length in birds. Am Nat 143: 95-116.[Web of Science]
Zar JH, 1984. Biostatistical analysis, 2nd ed. Englewood Cliffs, New Jersey: Prentice-Hall.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||