Behavioral Ecology Vol. 11 No. 2: 228-238
© 2000 International Society for Behavioral Ecology
The effect of tail streamer length on aerodynamic performance in the barn swallow
University of Stirling, Stirling FK9 4LA, UK
Address correspondence to K. L. Buchanan, Department of Biological and Molecular Sciences, University of Stirling, Stirling FK9 4LA, UK. E-mail: klbl{at}stir.ac.uk .
Received 21 July 1998; revised 6 August 1999; accepted 24 August 1999.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The elongated tail of the male barn swallow (Hirundo rustica) is regarded as one of the classic examples of a male trait exaggerated by female choice. However, recently a hypothesis has been proposed suggesting that the streamers, or elongated outer tail feathers, may aid aerodynamic performance through the Norberg mechanism, providing lift at slow speeds and high angles of attack when the tail is fully spread. The possibility exists that the tail streamer has evolved under natural selection, sexual selection, or a combination of both selection pressures. We tested these three hypotheses by reducing the streamer length of free-flying swallows and measuring their aerodynamic performance, using stereo-video. Measurements of flight performance were made from the digitized three-dimensional flight paths. Five flight variables best described the individual variation in flight performance. Four of these five parameters—mean velocity, mean curvature, maximum agility, and mean rate change of curvature in the XY plane—had significant second-order polynomial relationships with tail streamer manipulation. The first and second principal components (from principal components analysis of the flight variables) also showed similar relationships with streamer manipulation. The combination of a curvilinear relationship between flight performance and streamer length and an aerodynamic optimum between 0 and 20 mm reduction is only predicted if both natural and sexual selection have been acting on streamer morphology. Our data therefore suggest that sexual selection has extended streamer length by around 10 mm beyond its aerodynamic optimum. We suggest that both natural and sexual selection have been important in shaping tail morphology in the barn swallow, and the relative importance of both selection pressures is discussed.
Key words: barn swallows, flight performance, Hirundo rustica, maneuverability, natural selection, Norberg mechanism, sexual selection.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The elaborate male coloration and ornaments found in many sexually dimorphic species were first explained by Darwin (1871
). He suggested that such
traits could evolve despite being costly because they determine access to
mating opportunities, either through male competition or female choice. The
last 20 years has been enormous growth in the number of studies examining the
effects of sexual selection pressures on male behavior or morphology (reviewed
by Andersson, 1994). Of the
many sexually selected male traits now recognized, some of the classic
examples most often cited are the elongated tail feathers of a number of bird
species (Andersson, 1982;
Møller, 1988). In
particular, an established body of literature now exists supporting the
hypothesis that the long tail streamers of the male barn swallow (Hirundo
rustica) have been exaggerated through female choice (Møller,
1988,
1989,
1994;
Saino et al., 1997;
Smith and Montgomerie,
1991).
The possession of long streamers is thought to advertise male quality
because long-tailed males not only have higher annual reproductive success and
survival rates, but also decreased susceptibility to parasitism and more
efficient immune responses (de Lope and
Møller, 1993; Møller,
1988,
1990;
Møller and de Lope,
1994; Saino et al.,
1995). Furthermore, there is evidence that both phenotype
(Møller, 1990,
1991;
Møller et al., 1995)
and genotype (Møller,
1994) influence male tail length.
Theoretical models investigating the limitations to the development of
sexually selected ornaments predict that ornament size should reflect male
quality and should therefore be costly in order to maintain the honesty of the
signal (Fisher, 1930;
Grafen, 1990;
Pomiankowski et al., 1991).
The cost of having elongated tail streamers has been assumed to be impaired
aerodynamic performance through the increased drag from the tail.
Møller and de Lope
(1994) found that males with
relatively long tail streamers caught smaller prey items than males with
shorter streamers, which they interpreted as a decrease in foraging efficiency
for such males. Møller and his co-workers have therefore concluded that
streamer elongation is costly for male swallows in terms of flight efficiency,
which subsequently influences both foraging efficiency
(Møller and de Lope,
1994; Møller et al.,
1998) and predator avoidance
(Møller and Neilsen,
1997).
Elongated tails exist in a number of forms, including streamer tails,
graduated tails where the outer tail feathers are the shortest, with gradual
increases in length toward the central tail feathers, and pin tails, where
only the central two tail feathers are elongated. There is considerable
evidence that both graduated and pintails are costly in terms of aerodynamic
efficiency (Balmford et al.,
1993; Evans and Hatchwell,
1992; Evans and Thomas,
1992). However, the evidence for streamer tails is not as clear
cut. The amount of lift generated by the tail (and therefore the degree of
maneuverability) is proportional to the maximum continuous span of the tail
(Thomas, 1993). As streamer
elongation develops, the maximum continuous span also increases for a forked
tail, increasing the lift:drag ratio. However, further extension of the tail
streamers beyond the maximum continuous span contributes to drag but does not
enhance lift. It has therefore been assumed that elongated streamer tails can
also act as a handicap and that the degree of elongation is in some way
representative of the degree to which aerodynamic performance is reduced.
However, by examining the flexible nature of the streamer, it has become
apparent that the streamer can twist when the tail is spread, causing a
nose-down rotation of the leading edge of the streamer, which creates a vortex
flap at the front edge of the tail
(Norberg, 1994). This
mechanism prevents stalling at high angles of attack, generating a greater
degree of aerodynamic lift and aiding maneuverability in slow, turning flight
(Norberg, 1994). This raises
the possibility that the streamer may have, at least in part, an aerodynamic
function and suggests that the streamer of the barn swallow may have evolved
through natural selection pressures for enhanced maneuverability.
The findings of Møller, his co-workers, and other independent
researchers (Smith and Montgomerie,
1991) have suggested that sexual selection has been the major
force causing streamer elongation in this species, while the role of natural
selection has been primarily to prevent elongation of the tail beyond a
manageable handicap. They do not discount the possible of role of natural
selection in affecting streamer morphology, but maintain that its importance
is secondary to the effects of sexual selection
(Barbosa and Møller,
1999; Møller et al.,
1998). However, the possibility remains that both natural
selection and sexual selection pressures have been responsible for tail
streamer elongation. If this were the case, part of the streamer should have
an aerodynamic function, although sexual selection may have caused the
streamer to exceed its aerodynamic optimum.
Evans and Thomas (1997)
suggested that it would be possible to differentiate between whether the tail
streamers of the barn swallow had evolved purely through sexual selection,
natural selection, or a combination of both selection pressures. Experimental
elongations of streamer length are unlikely to separate the two hypotheses
because both natural and sexual selection predict an increase in cost with
increasing streamer length. However, experimental reductions of streamer
length generate different predictions. If streamer length is purely naturally
selected, reducing its length should increase flight costs, as the observed
length for any individual must approximate to its aerodynamic optimum.
However, if the streamer is purely sexually selected, shortening the length
would produce a decrease in cost, as the handicap is removed. Alternatively,
if the streamer has a naturally selected and a sexually selected component,
shortening the length progressively would be expected to produce a decrease in
cost, followed by an increase in cost as the aerodynamic benefit is removed
(Evans and Thomas, 1997;
Table 1).
|
Recent work examining the flight performance of male swallows with
streamers extended and reduced by 20 mm has shown that flight performance is
affected in the same direction for both experimental groups relative to
controls (Evans, 1998). This
result supports the hypothesis that unmanipulated streamers are within 20 mm
of their naturally selected aerodynamic optimum and has sparked much debate as
to the relative importance of natural and sexual selection for the tail length
of male and female barn swallows (Evans,
1999;
Hedenström
and Møller, 1999). However, these manipulations were not
subtle enough to determine if sexual selection has extended streamer length by
<20 mm. To differentiate between the three hypotheses and to determine the
position of the aerodynamic optimum, we carried out a series of manipulations
on the tail streamers of free flying barn swallows. The flight performances of
experimental and control birds were then assessed using stereo-video filming
to assess the cost of the streamer. The synchronized stereo images were used
to plot flight paths for each bird in three dimensions, and measurements of
individual flight performance were then made from the reconstructed flight
paths.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Individual swallows were caught at 15 different farms in the area around Stirling, UK, between May and July 1997. Individuals were caught with mist-nets at the entrance to barns where pairs were nesting. We took biometric measurements from all individuals, and where individuals were caught more than once (n = 30), we used these data to calculate repeatability of the morphological measures (Lessells and Boag, 1987
). Right (73.6%) and left (71.3%) wing length, right (99.6%)
and left (96.0%) streamer length, length of the midtail (31.8%), and head and
bill length (86.6%) showed highly significant repeatabilities, but mass
(19.4%) and wing area (6.3%) were not significantly repeatable. Birds were
sexed using a combination of presence or absence of a brood patch and the
length of the tail streamers (Svensson,
1992) and marked with a unique combination of dyes on the breast
and belly feathers.
We randomly assigned male and female swallows to one of six experimental
groups or to a control group. Previous manipulations of streamer length in
barn swallows have shown that elongations and reductions of tail length by 20
mm both cause an increase in flight costs
(Evans, 1998); therefore, to
differentiate between the natural selection hypothesis alone and the combined
effects of natural and sexual selection, a range of small streamer reductions
were carried out on individuals, as this would allow the shape of the
relationship between streamer reduction and flight performance to be plotted
more accurately. The experimental birds therefore had their tail streamers
reduced by 2, 4, 8, 10, 15, or 20 mm. Because we were specifically interested
in testing streamer function, we reduced streamer length from the end of the
streamer and then trimmed the streamer to resemble a natural, rounded streamer
end. Previous experiments (Evans,
1998) have shown that basal manipulations, extensively used in
mate choice studies, have quite different effects on flight performance
compared with streamer manipulations. Control birds were handled and marked as
for experimental birds but were not manipulated.
The birds were filmed using two video cameras (Sony Digital Handycam DCR-VX1000E). The cameras were set 1 m apart and held exactly parallel by a supporting bar and fixed mountings, fixed to a tripod. All birds were filmed during brood provisioning of first and second broods at least 24 h and usually <7 days after manipulation. Birds were filmed entering and leaving the building housing the nest site in an attempt to standardize the types of flights analyzed and to focus on tight turn sequences, where the tail was unfurled and the Norberg mechanism could therefore function. A number of individuals were filmed on multiple occasions, in which case average values of flight performance of each individual were calculated.
The stereo-video footage was digitized to obtain the position of the bird
in each frame of flight, at a speed of 25 frames/s using the miroMOTION DC20
digitizer on an Apple Macintosh 9500. We then edited the digitized video to
discard any long video sequences without flights using Adobe Premiere 4.0
(Adobe Systems Inc.) and obtained the two-dimensional stereo-pairs of
coordinates for the position of the center of the head of the bird for each
frame using NIH Image (developed by the U.S. National Institutes of Health and
available on the internet at http://rsb.info.nih.gov/nih-image/). The position
of the bird in synchronous frames, the focal length of the lenses, and the
camera separation were then entered into a computer program to obtain the
three-dimensional coordinates for the flight path. The three-dimensional
coordinates were smoothed using a fourth-difference algorithm
(Rayner and Aldridge, 1985).
We estimated the accuracy of these measures by plotting coordinates for a
stationary object 1 m in size in three dimensions, placed approximately the
same distance from the cameras, as the swallows. We used these data to
calculate the overall error in the positional data, which was estimated to be
6.1% (S.E. ± 2.0%, n = 92). More specifically, mean error in
the positional data both with respect to distance from the cameras and
position in the field of view were <10% and therefore similar to values
calculated from other studies using this technique (Evans et al.,
submitted).
We calculated velocity from the first-order differential of plotted positional data of five consecutive points in three-dimensional space with time. Acceleration was calculated in a similar manner to velocity by estimating the change in velocity over five consecutive positions in three-dimensional space, assuming the rate of change of acceleration is locally constant. Measurement of curvature involves calculating the vector quantities for changes in each plane in three dimensions between each plotted position. The scalar components (dependent on velocity and acceleration) for each of the three planes xy, xz, and yz are then combined to give a measure in three dimensions. The rate of change of curvature is therefore calculated as the first-order differential of the curvature series with time. Details of the exact methods of calculation of the flight parameters are provided elsewhere (Evans et al., submitted).
We estimated the error in the velocity data by measuring an object moving
at a known speed (a tennis ball dropped vertically from a known height
traveling at 6.15 m/s). The error in the calculation was estimated to be 10.2%
(S.E. ± 3.2%, n = 6; Evans et al., submitted). The
three-dimensional flight coordinates were used to calculate the flight
performance parameters for individual flight paths; mean, maximum, and minimum
velocities; mean, maximum, and minimum acceleration; mean and maximum energy
(potential and kinetic); mean and maximum curvature; mean and maximum rate of
change of curvature; mean and maximum curvature in the xy plane; mean and
maximum rate of change of curvature in the xy plane; mean, minimum, and
maximum turn radius; maximum and mean agility (Evans et al., submitted;
Rayner and Aldridge, 1985).
Where multiple flights were obtained for an individual, we used the mean of
each flight variable in the analysis. All filming and digitizing were carried
out blind to the manipulation group of the bird being filmed.
Data were analyzed using MINITAB 10 Xtra. Obviously, the 21 flight variables do not vary independently (Table 2), so we reduced them to the smallest set that explained significant covariation in the others. Initially, we correlated all the variables against each other to identify the flight variables that described significant variation in most other flight variables. We used multiple regression to sequentially combine the flight variables to find the smallest set of variables that explained significant variation in all the others. Each flight variable was used as the dependent variable, in a regression with all the other flight variables, sequentially used as independent variables. The flight variable that explained variation (significant R2) in the highest number of other variables was used as the first factor in the multiple regression model. The variable that explained significant variation in most other variables was in this case mean velocity. We identified the second flight variable by repeating the process outlined above, using each flight variable as the dependent variable in a multiple regression with mean velocity plus all the other flight variables sequentially. The second flight variable was the one which in combination with mean velocity explained variation in the highest number of other variables. This process was continued until all the flight variables were explained by the predictors in this reduced set.
|
Five measures of flight performance best explained most individual variation: mean velocity, maximum curvature, mean acceleration, maximum agility, and mean rate of change of curvature in the xy plane. Another way to deal with a large number of intercorrelated variables is to use principal components analysis. PCA of all the flight variables generated two components, PC1 and PC2, which were the only ones that explained >10% of individual variation in flight performance (PC1 33.4%, PC2 21.9%; Table 3). There was a sharp drop in the amount of variance explained by further PCs, so they were excluded from any further analysis (S-Plus 4 Guide to Statistics, Data Analysis Products Division, Math Soft, Seattle, Washington). The loadings indicate that PC1 was associated with measures of velocity and acceleration, whereas PC2 was associated with measures of flight curvature. The subsequent analyses were based on investigating the nature of the relationships between each of the five flight measures, the two principal components, and tail reduction manipulation and morphology.
|
A general linear model was constructed for each flight variable, with the flight variable as the dependent variable and tail manipulation group as a continuous independent variable. We added date, farm, sex, and individual morphological measures into the models as covariates. We also used interactions between tail manipulation and either sex or morphological measures in the starting models. Variables that failed to explain a significant amount of the variation in the flight variables were eliminated and the model was rerun. The residuals were checked at each stage for a normal distribution and homoscedasticity.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Flight sequences were obtained for 68 individual swallows. The mean velocity of the flight sequences analyzed was 8.59 m/s (S.E. ± 0.388, n = 68), which is well within the range of flight speeds that would allow the Norberg mechanism to operate (Norberg, 1994
). Repeatability estimates were calculated from 2 different flights for each of 16 individuals (mean velocity = 73.6%; mean curvature = 98.1%; mean acceleration = 58.0%; maximum agility = 53.2%; mean rate of change of curvature in the xy plane = 14.2%). The high repeatability of the curvature measures is likely a consequence of filming the birds repeatedly going to and from the nest site. These repeatabilities were significant, except mean rate of change of curvature in the xy plane.
Relationships between flight performance and streamer reductions
Of the five flight variables, four (mean velocity, maximum curvature,
maximum agility, and mean rate change of curvature in the xy plane) had
significant second-degree polynomial relationships with tail streamer
reduction, either alone (mean velocity, mean curvature, and maximum agility)
or in interaction with sex (maximum agility and mean rate change of curvature
in the xy plane) or tail streamer length (mean velocity, maximum curvature,
and maximum agility; Table 4).
In contrast, the fifth variable, mean acceleration, did not have a
second-degree polynomial relationship with manipulation group, although there
was a significant linear relationship. Both principal components also had
significant second-degree polynomial relationships with tail streamer
manipulation (Table 4),
although PC2 also had a significant second-degree polynomial relationship with
manipulation in interaction with both sex and tail streamer length. Therefore,
four of the five flight variables investigated, and both principal components,
showed significant curvilinear relationships with tail streamer manipulation.
It is apparent that the degree to which the manipulation affected flight
performance varied for different amounts of streamer reduction
(Figure 1). The relationships
in Figure 1 demonstrate that
flight performance changed with progressive streamer reductions until a
minimum/maximum beyond which further reductions in the streamer caused the
opposite effect on flight performance. Curvilinear relationships between
streamer reduction and flight performance are predicted by either a natural
selected tail or a tail which is the product of the both natural and sexual
selection. The position of the minimum/maximum of the curve relating tail
reduction and flight performance is therefore crucial for differentiating
between these two hypotheses. It can be seen from
Figure 1 that the
minimum/maximum for each of these variables occurred between 0 and 20 mm
reduction. A sensitivity analysis (Table
5) estimating the 95% confidence intervals around the position of
the minimum/maximum confirmed the position occurred at some reduction of the
streamer length and in almost all cases between 0 and 20 mm reduction.
|
|
|
Sexually selected and naturally selected components of the
streamer
Mean streamer length interacted significantly with manipulation group in
the models for mean velocity, maximum agility, maximum curvature, mean rate
change xy, PC1, and PC2 (Table
4). This means that for all these variables, the relationship
between the degree of streamer reduction and the position of the
minimum/maximum of the curve, and hence the point at which flight performance
was optimized, varied with streamer length. The relationship between flight
performance and streamer length and manipulation is shown in
Figure 2. The minimum/maximum
can be seen clearly in all the figures, and although the effects of streamer
manipulation vary for different streamer lengths, a curvilinear relationship
is found at all streamer lengths.
|
Using the table of coefficients from the final ANOVAs, the equations relating each flight variable with the variables in the model could be produced. Differentiation of this equation with respect to streamer manipulation and setting the first-order differential to zero allowed the evaluation of the minimum/maximum of the curve (Table 5). The position of the minimum/maximum is remarkably similar at different streamer lengths (Figure 2). It is important to recognize that, although the sexually selected component of the tail feather does vary with streamer length, the models consistently suggest that a reduction of between 7 and 15 mm would optimize flight performance.
As predicted from the natural selection and sexual selection combined hypothesis, the minimum for each curve occurred at between 0 and 20 mm reduction for all the flight variables, with the exception of mean acceleration (Table 6).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
There has recently been much discussion concerning the relative importance of both natural and sexual selection in promoting tail streamer elongation in the swallow (Balmford et al., 1993
; Evans, 1999;
Hedenström,
1995;
Hedenström
and Møller, 1999;
Møller et al., 1998;
Thomas and Rowe, 1997). This
experiment was specifically designed to determine whether streamer lengths are
currently at their aerodynamic optimum in our population and, if not, to
quantify the extent they have been exaggerated. The results clearly support
the hypothesis that the tail streamer of the swallow has evolved under the
influence of both natural and sexual selection pressures. Six of the seven
measures of flight performance had significant second-degree polynomial
relationships with streamer shortening, with maxima/minima between 0 and 20 mm
reduction. This type of relationship is only predicted by the process of
sexual selection exaggerating a trait which also has a naturally selected
function (Evans and Thomas,
1997). Therefore, as the tail streamer is reduced, flight costs
are reduced, until the point is reached at which the tail is an optimal length
for the individual; further reductions beyond this point cause an increase in
flight costs. These results are consistent with the recent finding that the
tail streamer reduction can increase flight costs
(Evans, 1998). However, the
incremental reductions used in this experiment have allowed the relationship
between streamer manipulation and flight performance to be quantified more
accurately than was possible in previous experiments
(Evans, 1998). A parallel study has recently been carried out examining the relationship between streamer length and flight maneuverability of individual swallows, using a standardized task through a flight maze (Rowe et al., in preparation). This task allows the maximum maneuverability of the swallows to be determined by assessing their ability to maneuver around obstacles as quickly as possible. This study also attempted to examine maximum maneuverability by assessing flight performance using stereo-video filming of mobbing flight during presentation of a mock predator (Rowe et al., unpublished). Consistent with the results presented here, both tasks demonstrated curvilinear relationships between flight performance and progressive streamer reductions. In addition, they have also both found that a reduction of approximately 12 mm approximated to the aerodynamic optimum tail length for an individual bird.
The flight performance analysis has allowed us to examine a range of
possible flight variables, which show highly consistent results between the
different variables. The only flight variable that did not show this type of
relationship was mean acceleration, which held a linear relationship with
streamer reductions. Because the Norberg mechanism predicts enhanced
maneuverability at slow speeds and high angles of attack, we chose to film
individual birds in turning flight during provisioning visits. Evans and
Thomas (1997) pointed out that
the role of the tail streamer is likely to change as the tail is unfurled
during turns, compared to the furled tail used for straight flight. However,
such flights inevitably involve a reduction of speed as individuals enter the
nest site. For this reason, the sample of flights used for the analysis may
not accurately reflect the acceleration potential of individuals.
Alternatively, this result may be robust, suggesting acceleration costs are
minimized either at no streamer reduction or complete removal; however, this
contrasts with the results from the other flight variables. Further work
examining the effects of such manipulations on a wide range of naturally
occurring acceleration values would be necessary before drawing any firm
conclusions.
We found that the degree of streamer shortening that produced the maximum
benefit in terms of improved flight performance for male barn swallows was
around 10 mm, although the exact length varied slightly with original streamer
length. Our results suggest that this represents the length of the streamer
that has been sexually selected beyond its aerodynamic optimum. This is of
particular interest because many studies have used longer (20 mm)
manipulations (e.g., Møller,
1988,
1989), which would interfere
with the aerodynamic properties of the streamer. Studies of mate choice that
have assumed males to be less attractive because of their experimentally
shortened streamers may in fact have found that females are less attracted to
males with impaired flight performance. Such studies have also used
manipulations of the basal part of the tail streamer, which alters the length
of base of the feather, without altering the length of the streamer itself.
Compared to altering streamer length, this may have quite different
implications for aerodynamic performance
(Evans, 1998).
The results of this study could help to explain why male swallows with
elongated tails have been found to decrease their foraging efficiency, whereas
males with tails shortened by 20 mm showed no difference from control males
(Møller, 1989)
[although this is not supported by a later study at a different field site
(Møller and de Lope,
1994)]. Our results suggest that such a large reduction in
streamer length would impair flight performance and therefore foraging
efficiency. It would be quite reasonable to suppose, therefore, that males
with streamer reductions of 20 mm could be as handicapped as control males.
However, our study examined the effects of tail manipulations only on flight
to and from the nest site. Although we would predict that these effects are
likely to be broadly applicable to other types of flight, this technique has
yet to be used to directly determine the consequences of changes in flight
performance for foraging efficiency.
We found that females also exhibited the same type of relationship between
streamer shortening and aerodynamic performance. This is interesting because
it implies that female tail length is also exaggerated beyond its aerodynamic
optimum. Møller (1993)
suggests that there may be a strong genetic association between the sexes for
tail length, which may be why female swallows have longer tails than juvenile
swallows (Møller et al.,
1998). If such an association exists, it could explain why females
appear to possess tails that are longer than would be expected as a result of
natural selection alone. Alternatively, it is possible that sexual selection
pressures also exist for females, exaggerating their tails beyond the
aerodynamic optimum.
In all cases the direction of the relationship between streamer
manipulation and flight performance was different for individuals with
different initial streamer lengths. Figure
2 demonstrates that the relationships between the flight variables
and streamer manipulation are not consistently the same shape—flipping
between a U- and an n-shaped curve as streamer length changes. Evans and
Thomas (1997) predicted that
changes in flight performance should be either U or n-shaped depending on
whether increases in the flight variable under consideration increased costs
(a U shape) or decreased costs (an n shape). Our data suggest that this view
is naive and that the position of the optimum and whether the relationship is
U shaped or n shaped changes with streamer length. For example, in
Figure 2a, a bird with long
streamers is subject to an n-shaped relationship between tail manipulation and
mean velocity, whereas a bird with short streamers is subject to a U-shaped
relationship. These data are regraphed in
Figure 3 in two dimensions to
show clearly how manipulations affect individuals with different initial tail
lengths. Figure 3 also
demonstrates that, although the shape of the relationship between manipulation
and flight performance changes with initial streamer length, the position of
the turning point is remarkably consistent for different streamer lengths. As
the minimum/maximum of the curves represent the optimum flight speed, this
suggests that the optimum flight speed for birds with short streamers is lower
than that for birds with longer streamers. For birds with short streamers,
increasing flight velocity represents an increase in costs, whereas the
converse is true for birds with longer streamers. This is explicable because
streamer length covaries with other morphological variables. Therefore, birds
with shorter streamers also have shorter wing lengths than birds with longer
streamers. This suggests, not surprisingly, that there may be subtle but
fundamental differences in the flight strategies of birds with different
morphologies. As it seems likely that the different flight variables must
trade off against each other, individual morphologies must represent a range
of optima, with some individuals better adapted, for example, for speed or
agility. As a result there will be no one optimum value for any of the flight
variables, and for some individuals, increases in velocity, for example, will
represent an increase in flight costs, whereas for other individuals they will
represent a decrease in costs.
|
Interestingly, this pattern is also reflected in the relationships between streamer reduction and flight performance for males and females, for the same reason as above. Males and femals have different initial streamer lengths. The effects of manipulating streamer length on flight performance will therefore vary as males and females will generally be at different starting points on the z axis (streamer length) of Figure 2. The shapes of the relationships for males and females are thus different for almost all variables. Where sex does not interact significantly with manipulation (mean velocity, maximum curvature, PC1), there is a significant interaction between original streamer length and manipulation, which we suggest also reflects a difference between the sexes (Table 4). The relative contribution of sex and original streamer length in interaction with streamer manipulation is likely to be affected by the degree of overlap in morphology between the sexes. Furthermore, the relevance of sexual dimorphism may vary between the flight parameters. However, it would appear that our results demonstrate some intrinsic differences between males and females in terms of how streamer reduction affects flight performance.
Sexual selection versus natural selection
The possible mechanisms by which female choice can select and exaggerate
traits that honestly reflect male quality have received much attention
(Fisher, 1930;
Grafen, 1990;
Pomiankowski et al., 1991).
This study provides little evidence for support of either sexual selection or
natural selection as the starting point for streamer elongation. It is
possible, as Møller and his co-workers suggest, that the Norberg effect
has developed as a cost-reducing trait to minimize the aerodynamic
disadvantages of an exaggerated tail streamer
(Møller et al., 1998).
However, it is also possible that the Norberg mechanism can function in
shorter streamers predating the association of female choice and male streamer
length. We believe that further research is needed to evaluate the
contribution of the Norberg mechanism with varying streamer lengths before its
potential contribution to initial streamer elongation can be decisively
dismissed. It would also be of interest to discover whether the Norberg
mechanism operates in other unrelated taxa with similar tail shapes [e.g.,
frigate birds (Frigata minor), arctic terns (Sterna
paradisaea), fork-tailed flycatchers (Muscivora tyrannus)].
Møller et al. (1998)
recently summarized the enormous amount of evidence in support of sexual
selection pressures acting on the tail streamer of the male barn swallow. Our
results in no way contradict this body of work, but they do drawn attention to
the subtlety of the selection forces at work. The importance of the Norberg
mechanism is yet to be fully quantified; a number of questions concerning the
relationship between streamer morphology, feather rotation, and the deflection
of the leading edge of the streamer remain undetermined. However, despite the
fact that the relevance and magnitude of the mechanism are yet to be
quantified, the claim that the importance of the mechanism is secondary to the
effects of sexual selection in shaping the development of the streamer
(Møller et al., 1998)
is premature. The results from our study clearly indicate that individual
flight performance can be increased and then decreased by progressive small
streamer reductions. Such relationships between flight performance and tail
streamer reduction convincingly demonstrate the role that both natural and
sexual selection have played in shaping the tail streamer of the barn swallow.
Such relationships cannot be reconciled to the hypothesis that the streamer is
primarily the product of sexual selection alone. Further work is now needed to
establish the initial mechanism that established streamer elongation in this
species.
|
ACKNOWLEDGEMENTS |
|---|
K.L.B. was supported by a grant from Natural Environment Research Council (GR3/10600). Many thanks to Louise Rowe and Alasdair Sherman for their help with field work and to Adrian Thomas for setting up the flight analysis program. We also thank three anonymous referees for their comments, which greatly helped to improve an earlier draft of this paper.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Andersson M, 1982. Female choice selects for extreme tail length in a widowbird. Nature 299: 818-820.
Andersson M, 1994. Sexual selection. Princeton, New Jersey: Princeton University Press.
Balmford A, Thomas ALR, Jones IL, 1993. Aerodynamics and the evolution of long tails in birds. Nature 361: 628-630.
Barbosa A, Møller AP, 1999. Aerodynamic costs
of long tails in male barn swallows Hirundo rustica and the evolution
of sexual size dimorphism. Behav Ecol
10: 128-135.
Darwin C, 1871. The descent of man and selection in relation to sex. London: Murray.
de Lope F, Møller AP, 1993. Female reproductive effort depends on the degree of ornamentation of their mates. Evolution 47: 1152-1160.
Evans MR, 1998. Selection on swallow tail streamers. Nature 394: 233-234.
Evans MR, 1999. Length of tail streamers in barn swallows—reply. Nature 397: 115-116.
Evans MR, Hatchwell BJ, 1992. An experimental study of male adornment in the scarlet-tufted malachite sunbird. II. The role of the elongated tail in mate choice and experimental evidence for a handicap. Behav Ecol Sociobiol 29: 421-427.
Evans MR, Thomas ALR, 1992. The aerodynamic and mechanical effects of elongated tails in the scarlet-tufted malachite sunbird-measuring the cost of a handicap. Anim Behav 43: 337-347.
Evans MR, Thomas ALR, 1997. Testing the functional significance of tail streamers. Proc R Soc Lond B 264: 211-217.
Fisher RA, 1930. The genetical theory of natural selection. Oxford: Clarendon Press.
Grafen A, 1990. Sexual selection unhandicapped by the Fisher process. J Theor Biol 144: 473-516.[ISI][Medline]
Hedenström A, 1995. Swallows unhandicapped by long tails? Trends Ecol Evol 10: 140-141.
Hedenström A, Møller AP, 1999. Length of tail streamers in barn swallows. Nature 397: 115.
Lessells CM, Boag PT, 1987. Unrepeatable repeatabilities: a common mistake. Auk 104: 116-121.[ISI]
Møller AP, 1988. Female choice selects for male sexual tail ornaments in the monogamous swallow. Nature 332: 640-642.
Møller AP, 1989. Viability costs of male tail ornaments in a swallow. Nature 339: 132-135.
Møller AP, 1990. Effects of a haematophagous mite on the barn swallow (Hirundo rustica): a test of the Hamilton and Zuk hypothesis. Evolution 44: 771-784.
Møller AP, 1991. Sexual selection in the monogamous barn swallow (Hirundo rustica). I. Determinants of tail ornament size. Evolution 45: 1823-36.
Møller AP, 1993. Sexual selection in the barn swallow Hirundo rustica III. Female tail ornaments. Evolution 47: 417-431.
Møller AP, 1994. Sexual selection and the barn swallow. Oxford: Oxford University Press.
Møller AP, Barbosa A, Cuervo JJ, de Lope F, Merino S, Saino N, 1998. Sexual selection and tail streamers in the barn swallow. Proc R Soc Lond B 265: 409-414.
Møller AP, de Lope F, 1994. Differential costs of a secondary sexual character: an experimental test of the handicap principle. Evolution 48: 1676-1683.
Møller AP, Magnhagen C, Ulfstrand A, Ulfstrand S,
1995. Phenotypic quality and molt in the barn swallow,
Hirundo rustica. Behav Ecol 6:
242-249.
Møller AP, Neilsen JT, 1997. Differential predation costs of a secondary sexual character: sparrowhawk predation on barn swallows. Anim Behav 54: 1545-1551.[ISI][Medline]
Norberg RA, 1994. Swallow tail streamer is a mechanical device for self-deflection of tail leading edge, enhancing aerodynamic efficiency and flight manoeuvrability. Proc R Soc Lond B 257: 227-233.
Pomiankowski A, Iwasa Y, Nee S, 1991. The evolution of costly mate preferences. 1. Fisher and biased mutation. Evolution 45: 1422-1430.
Rayner JMV, Aldridge HDJN, 1985. Three-dimensional
reconstruction of animal flight paths and the turning flight of
microchiropteran bats. J Exp Biol 118:
247-265.
Saino N, Møller AP, Bolzern AM, 1995.
Testosterone effects on the immune system and parasite infestation in the barn
swallow (Hirundo rustica): an experimental test of the
immunocompetence hypothesis. Behav Ecol
6: 397-404.
Saino N, Primmer CR, Ellegren H, Møller AP, 1997. An experimental study of paternity and tail ornamentation in the barn swallow (Hirundo rustica). Evolution 51: 562-570.
Smith HG, Montgomerie R, 1991. Sexual selection and the tail ornaments of North American barn swallows. Behav Ecol Sociobiol 28: 195-201.
Svensson L, 1992. Identification guide to European passerines, 4th ed. Stockholm: Svensson.
Thomas ALR, 1993. On the aerodynamics of birds' tails. Phil Trans R Soc Lond 340: 361-380.
Thomas ALR, Rowe L, 1997. Experimental tests on tail
elongation and sexual selection in swallows (Hirundo rustica) do not
affect the tail streamer and cannot test its function. Behav
Ecol 8:
580-1.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  T. L. Hedrick and A. A. Biewener Low speed maneuvering flight of the rose-breasted cockatoo (Eolophus roseicapillus). I. Kinematic and neuromuscular control of turning J. Exp. Biol., June 1, 2007; 210(11): 1897 - 1911. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. J. McGraw, R. J. Safran, M. R. Evans, and K. Wakamatsu European barn swallows use melanin pigments to color their feathers brown Behav. Ecol., September 1, 2004; 15(5): 889 - 891. [Full Text] [PDF]
|
|
|
|
|
|
R. J. Safran and K. J. McGraw Plumage coloration, not length or symmetry of tail-streamers, is a sexually selected trait in North American barn swallows Behav. Ecol., May 1, 2004; 15(3): 455 - 461. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. J. Cuervo, A. P. Moller, and F. de Lope Experimental manipulation of tail length in female barn swallows (Hirundo rustica) affects their future reproductive success Behav. Ecol., July 1, 2003; 14(4): 451 - 456. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Dudley Mechanisms and Implications of Animal Flight Maneuverability Integr. Comp. Biol., February 1, 2002; 42(1): 135 - 140. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. P. Moller and A. Barbosa Flight, fitness, and sexual selection Behav. Ecol., July 1, 2001; 12(4): 511 - 512. [Full Text] [PDF]
|
|
|
|
|
|
K. L. Buchanan and M. R. Evans Flight, fitness, and sexual selection: a response Behav. Ecol., July 1, 2001; 12(4): 513 - 515. [Full Text] [PDF]
|
|
|
|
|
|
L. V. Rowe, M. R. Evans, and K. L. Buchanan The function and evolution of the tail streamer in hirundines Behav. Ecol., March 1, 2001; 12(2): 157 - 163. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. J. Park, M. Rosen, and A. Hedenstrom Flight kinematics of the barn swallow (Hirundo rustica) over a wide range of speeds in a wind tunnel J. Exp. Biol., January 8, 2001; 204(15): 2741 - 2750. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||