Behavioral Ecology Vol. 11 No. 1: 93-101
© 2000 International Society for Behavioral Ecology
Asymmetries in male aggression across an avian hybrid zone
a College of Forest Resources and Burke Museum, PO Box 353010, University of Washington, Seattle, WA 98195, USA b Burke Museum and Department of Zoology, PO Box 353010, University of Washington, Seattle, WA 98195, USA
Address correspondence to S. F. Pearson, who is currently at the Department of Zoology, PO Box 118525, University of Florida, Gainesville, FL 32611, USA. E-mail: spearson{at}zoo.ufl.edu .
Received 9 December 1998; revised 9 June 1999; accepted 7 August 1999.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Recent studies suggest that competitive asymmetries are causing the hybrid zones between hermit and Townsend's warblers to move, such that Townsend's warblers are replacing hermit warblers. Here we examine the contribution of male aggression to this competitive asymmetry by measuring aggressive responses to mounts. We presented male mounts of the two parental species to Townsend's and hermit warblers outside the zone and to hybrids within the zone. Outside the zone, Townsend's males are more aggressive to both conspecific and heterospecific mounts than are hermit males. This asymmetry should move the zone in the direction inferred from previous studies. Hybrids fall between parentals in their aggressiveness, which should accelerate the movement of the zone. Remarkably, we found no relationship between phenotype and aggression in individual males at a locality within the hybrid zone. The forces of selection and dispersal that maintain narrow hybrid zones should generate such a correlation if aggressive differences between the parental species are genetically controlled. We resolve this conflict by proposing a behavioral model of competitive sorting within localities. If birds are sorted across the hybrid zone according to competitive abilities, and competitive interactions within neighborhoods are more or less complete, then the correlation between phenotype and aggression within any given neighborhood will be eliminated. We tested this model by examining the relationship between phenotype and aggression across the zone. Warblers in hybrid neighborhoods on the Townsend's side of the zone are more aggressive than warblers in hybrid neighborhoods on the hermit side, indicating that competitive sorting is occurring.
Key words: competitive sorting, Dendroica occidentalis, Dendroica townsendi, hybrid zones, interspecific competition.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Hermit (Dendroica occidentalis) and Townsend's (D. townsendi) warblers hybridize in three geographically separate areas in the Pacific Northwest of the United States: the Olympic Mountains of Washington, the southern Cascades of Washington, and the Cascades of Oregon south of Mt. Hood (Rohwer and Wood, 1998
). The two Washington hybrid zones are only three to four
times wider than estimates of root-mean-square dispersal, implying that
selection is preventing the zones from widening (Barton and Hewitt,
1985,
1989;
Rohwer and Wood, 1998).
Asymmetries in the character transition curves describing these zones suggest
that Townsend's warblers have a selective advantage over hybrids and hermits
(Rohwer and Wood, 1998). These
results also suggest that parental fitness asymmetries keep these zones narrow
(Hewitt, 1988) and that
Townsend's warblers are replacing hermit warblers, moving the zones
southward.
These warblers and their hybrids overlap extensively in habitat
requirements and are interspecifically territorial
(Pearson and Manuwal, 2000).
Thus, an aggressively superior parental species should have a selective
advantage. At a locality near the center of the Washington Cascades hybrid
zone, a higher proportion of Townsend's males maintained territories
throughout the breeding season than did hermit and hybrid males
(Pearson, 2000), suggesting
that Townsend's are aggressively superior. To yield a fitness advantage,
superiority in holding territories must also be accompanied by success in
attracting mates. At this same field locality, Townsend's males were also more
successful at attracting mates (including hybrid and hermit females) than were
hybrid and hermit males (Pearson,
2000).
We assess the relative fighting abilities (hereafter
"aggressiveness") of parentals and hybrids across localities
separated by 435 km (Figure 1).
Aggressiveness was measured by the responses of territorial males toward two
kinds of mounts, pure Townsend's males and pure hermit males. Mounts were
presented with song playback to simulate territorial intrusions. Experimental
mount presentations appear to provide reliable comparisons of fighting
abilities among territorial birds and are frequently used to quantify
aggressive differences among males when contests cannot be staged or observed
(Rohwer, 1978;
Rohwer and Røskaft,
1989; Røskaft and
Rohwer, 1987). In studies of blackbirds, differences in aggressive
responses to mounts accurately predicted the winner of staged contests between
these same males (Rohwer,
1982).
|
Differences in response to conspecific mounts should measure maximal aggressive differences between parentals. Differences in response to heterospecific mounts should measure the outcome of interspecific fights during initial contacts between parentals. If differential male dispersal is contributing to the greater introgression of Townsend's warbler characters into hermit warbler populations than vice versa, then Townsend's males should be more aggressive to heterospecific mounts than are hermit males. We also measured the response of hybrids to mounts of both parentals. The aggressiveness of hybrids relative to parentals will affect the rate of zone movement. If hybrids are less aggressive to mounts than Townsend's but more aggressive to mounts than hermits, then the movement of the zone should be accelerated, but if hybrids are less aggressive than both parentals, the movement of the zone should be slowed.
We also examined the relationship between phenotype and aggression within
and between localities in the hybrid zone. Morphological traits are strongly
associated within individuals in these warbler hybrid zones; thus hybrids from
midpoint localities intermediate in one character tend to be similarly
intermediate in other characters (Rohwer
and Wood, 1998). This association between presumably nonepistatic
characters is attributed to the dispersal of parentals and parental-like
backcrosses into the zone, thus providing a constant source of linked sets of
genes. It is for this reason that character correlations at any locality
within the zone remain high, instead of falling to zero as they should if
mating were random (Barton and Hewitt,
1985). We expect an association between aggression and phenotype
if the aggressive differences between male hermit and Townsend's warblers have
some genetic basis.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Natural history and field localities
See Pearson (2000
) for a
description of natural history. Our field presentations of mounts were
conducted from 1994 through 1998. We presented mounts to pure hermit warblers
near Cathlamet, in the southwest corner of Washington
(Figure 1). At this locality,
Hermit warblers were found on steep south-facing slopes and ridge tops
dominated by Douglas fir, 300-600 m in elevation. We presented mounts to pure
Townsend's warblers at two localities, one slightly northeast of the
Washington Cascades hybrid zone, near Tieton, and the other far to the east of
the zone in the Selkirk Mountains (Figure
1). At Tieton, on the dry side of the Cascade Mountains just
northeast of the hybrid zone, Townsend's warblers were found in relatively wet
stands, 1000-1500 m in elevation, dominated by grand fir (Abies
grandis) and Douglas fir. In the relatively dry environment of the
Selkirk Mountains, Townsend's warblers inhabited wet stream corridors,
1000-1600 m in elevation, dominated by Douglas fir, grand fir, and Engelmann
spruce (Picea engelmannii). We presented mounts to all phenotypes at
a locality near the phenotypic center of the Washington Cascades hybrid zone
near Randle. At this locality warblers were found on steep south-facing slopes
and ridge tops, 300-1250 m in elevation, dominated by Douglas fir (see
Pearson and Manuwal,
2000).
To examine the relationship between phenotype and aggression across the
zone, we conducted additional presentations during 1997 and 1998 at five
localities along the Washington Cascades transect of Rohwer and Wood
(1998:
Figure 5). To measure mean
phenotypes (see below) for these localities, we computed average hybrid
indices for the responding males.
|
Mount presentations
At each locality, territorial males were presented with a hermit warbler
mount and a Townsend's warbler mount. From 1995 to 1997 every territorial male
was also presented with a chestnut-backed chickadee (Poecile
rufescens) mount to control for song (see below). All warbler mounts were
males at least 2 years old (aged after
Jackson et al., 1992), in full
breeding plumage. Mounts were presented on green poles, approximately 2 m
above the ground, and within 5 cm of vegetation on which the responding male
could alight without being forced to contact the mount. Mounts were posed in
an aggressive posture with wings drooped, tail fanned and beak slightly open,
as though they were singing.
We presented one mount per territory per day and systematically altered the sequence of presentations from one territory to the next. Several different mounts of each species were used, so differences in the strength of attacks cannot be attributed to particular mounts. Presentations were conducted between 0515 and 1300 h, and all presentations on a particular territory were conducted at the same location and at the same time of day. To attract males to our mounts, we played a locally recorded song throughout the 10 min of each presentation from a small, black speaker 1 m below the mount. We stood approximately 12 m from the mount during presentations and started the playback after a 3-min quiet period. During each presentation a running account of the male's response was dictated into a tape recorder. We worked in a different drainage each year to avoid presenting to the same birds more than once. No presentations were conducted during windy or rainy conditions, and only males present during all presentations on their territories were included in our analysis. All presentations were conducted by S.F.P.
Quantifying aggression
Seven behavioral response variables
(Table 1) were chosen following
a set of trial mount presentations. Time out of view measured the response of
males that counter-sang from the canopy but never approached the mount
(usually remaining more than 15 m away). Flight distances were counted in two
categories (>5 m and <5 m) because most birds either made a limited
number of longer flights or many short flights around their territory. Number
of wing flicks measured the agitation of males when they were perched. Number
of flights over the mount (with occasional strikes) measured agitation in
males that flew a lot. We summed the number of hits and pecks delivered to the
mount by males while flying or while perched on the mount. Time spent within 2
m of the mount measured willingness of males to remain close to the mount;
some alighted next to the mount and sang quietly, while others were more
active, making frequent 1- to 2-m flights over the mount.
|
We excluded two response variables that were measured in the field. Number of chips was excluded because we failed to distinguish between male and female chips during our first field season. Number of songs was excluded because males that closely approached the mount usually sang quietly; unfortunately, we failed to distinguish these quiet songs of highly aggressive males from the loud songs of males that did not approach the mount.
Controlling for song
Using song in conjunction with mounts confounds interpretation because some
of the aggressive response could be attributed to the song rather than to the
mount. Thus we included a third, neutral mount in each set of presentations
conducted from 1995 to 1997. Chestnut-backed chickadees occupy the same canopy
and subcanopy habitat as these warblers, and neither warblers nor chickadees
show aggression toward each other (Pearson and Rohwer, personal observations).
We played warbler song with chickadee mount presentations, so any differences
between responses to chickadee and warbler mounts can be attributed to the
effect of the warbler mounts and not to the playback of warbler song.
Song varies geographically, so locally recorded songs were essential to ensure that we attracted males to our mounts. Local songs were recorded the day before starting mount presentations using a Sony Professional Walkman recorder with a Sennhieser directional microphone. Neither the male we recorded nor any of his immediate neighbors received mount presentations. New recordings were made for each locality and for each field season. Thus, differences in aggressive responses among phenotypes cannot be attributed to particular recordings.
Song variation within the hybrid zone could further complicate our analysis. However, preliminary analyses of song recordings made within the hybrid zone indicate that all birds regardless of phenotype sing the same song (Bard S, personal communication); the phenotypes of the males used for this analysis ranged from pure hermits (score = 0.03) to pure Townsend's (0.92). Because song does not vary with phenotype within the zone, we used the same song for all of our presentations including our examination of the relationship between phenotype and aggression across the hybrid zone.
Controlling for breeding phenology
Mount presentation experiments were conducted between 19 May and 6 June
during the 5 years of presentations. Each year we conducted our mount
presentations during a different breeding stage and used song category as an
indicator of breeding stage. Hermit and Townsend's warblers change song types
when they pair (Pearson and Rohwer,
1998); thus, song type is a good indicator of a male's pairing
status. Because male aggressiveness can vary with breeding stage
(Wingfield et al., 1987), we
conducted our presentations during one of the following four stages each year:
(1) before female arrival, before any male had switched song category; (2)
during female arrival; males had not switched song category but females were
present on approximately one-third of the territories; (3) just before and
during clutch initiation; approximately half the males had switched song
category and females were present on most territories; and (4) during
incubation; approximately 90% of the males had switched song category. Our
goal was to conduct this work during the first three breeding stages, but in
the first year of our field work, most presentations fell into the fourth
category. All mount presentations on the transect across the hybrid zone were
conducted during a 5-day interval before female arrival.
Measuring phenotype within the hybrid zone
To quantify the phenotype of responding males, we used photographs of the
voucher specimens that define the hybrid index of Rohwer and Wood
(1998). Six of their eight
characters could be scored with binoculars in the field: extent of yellow on
the crown, intensity of yellow on the breast, back color, bib corner,
mid-flank streaking, and lower flank streaking. Variation in the six
characters was scaled from 0 to 1; the sum of these standardized scores was
also scaled to vary from 0 and 1. By definition, males with scores of 0.0-0.25
are hermit warblers, 0.25-0.75 hybrids, and 0.75-1.0 Townsend's. Thus,
"hybrid" refers to both F1 and backcrossed
individuals.
Statistics
In responding to mounts, some males flew over the mount, some spent most of
their time on the mount pecking the back of the head, while others failed to
approach the mount and only counter sang. The seven behavioral responses to
mounts (Table 1) were combined
into a single score of aggression using unrotated principle components
extracted from a Spearman correlation matrix
(Langston et al., 1990;
Rohwer, 1978).
The principal components analysis used to quantify aggression was computed from a balanced sample of pure Townsend's warblers (the Tieton sample), pure Hermit warblers (the Cathlamet sample), and hybrids (the midpoint sample from Randle). The character loadings from this analysis were used to compute the scores of aggression (principal component 1 scores) for these birds and for the birds on the transect across the hybrid zone; we excluded the presentations to pure Townsend's in the Selkirk Mountains and the many presentations to hybrids in the transect across the zone. We used a separate principal components analysis to compare the two pure Townsend's warbler localities (Selkirk and Tieton) because we did not want the responses of hermits and hybrids to affect these comparisons.
The following tests use principal component 1 scores (hereafter "scores of aggression" or "aggressiveness") from the principal components analysis described above. To examine the effect of breeding stage on male aggression, we used a two-way ANOVA. For comparisons of different mounts within localities, we used a randomized block ANOVA (because two or three mounts were presented to each male) and Scheffé's multiple comparisons for contrasts between pairs of mounts when three mounts were used. To evaluate the response to song (chickadee mount) among localities, we used an analysis of variance with Scheffé multiple comparisons. A two-sample t test was used for comparisons of aggressiveness between the parental localities. We used a two-way ANOVA to compare the response of Townsend's males at Tieton and the Selkirks to the two warbler mounts. We used linear regressions to examine the relationship between aggression and phenotype within the hybrid zone.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Validating the principal components analysis
Principal component 1 clearly measures aggressive differences among males (Table 1). As would be expected, pecks and hits and time spent within 2 m of the mount receive strong positive loadings, whereas time out of view receives a strong negative loading.
If the three classes of phenotypes (hybrids, hermits, and Townsend's) had different ways of responding to mounts (e.g., used different variables), then combining them into a single measure of aggression might be invalid. We examined this possibility by performing separate principal components analyses for each of the three localities (Townsend's from Tieton, hermits from Cathlamet, and hybrids from the midpoint locality at Randle). The component loadings for the seven behavioral variables were remarkably similar (Table 1), indicating that the same set of response variables measures aggression for each phenotypic group equally well. Thus, we combined these groups into a single principal components analysis to obtain the character loadings to compute our scores of aggression for most of the analyses that follow.
Breeding stage and aggression
We found no effect of breeding stage/year on aggressiveness (F =
1.6, df = 3, 69, p =.21) and no interaction between breeding
stage/year and locality (F = 0.75, df = 3, 69, p =.53). For
these comparisons we used pure hermits (Cathlamet) responding to hermit mounts
and pure Townsend's (Tieton) responding to Townsend's mounts. Because breeding
stage/year did not affect aggressiveness and because most of our presentations
took place in early spring, we ignore breeding stage/year in subsequent
analyses.
Comparisons between mounts within localities
Male hermit warblers outside the hybrid zone were 57% more aggressive
(lowest mean set to 0) to hermit mounts than to Townsend's mounts
(Figure 2; p =.015).
Their response to mounts of chickadees and Townsend's warblers was similar
(Figure 2; p =.681),
suggesting that they failed to see the Townsend's mounts as threatening. The
overall ANOVA comparing responses of the hermit males at this locality to all
three mounts was highly significant (F = 6.3, df = 2,38, p
=.003).
|
Male Townsend's warblers at Tieton, just northeast of the hybrid zone, were
strongly aggressive to both of the warbler mounts, showing no less aggression
to hermit mounts than to Townsend's mounts
(Figure 2; p =.71).
They were more aggressive to both warbler mounts than they were to chickadee
mounts (both p 
.001). Again, the overall ANOVA comparing the
responses of Townsend's males at this locality to the three mounts was highly
significant because the chickadee mount elicited little aggression (F
= 11.98, df = 2, 38, p <.0001).
Hybrid males from Randle, near the center of the zone, were more aggressive
to both of the warbler mounts than they were to the chickadee mount
(Figure 2; p
.018). As was true for pure Townsend's, these hybrid males responded with
equal aggression to the two warbler mounts
(Figure 2; p =.74).
Again, the overall ANOVA comparing the responses of warblers at this locality
to the three mounts was significant (F = 6.4, df = 2, 37, p
=.003).
Comparisons between the pure Townsend's localities
The pure Townsend's males at Tieton, just northeast of the hybrid zone,
responded as aggressively to hermit warbler mounts as they did to conspecific
mounts. Although the warblers at this locality are phenotypically pure
(Rohwer and Wood, 1998), some
individuals contain hermit warbler haplotypes (Rohwer S, Bermingham E,
personal communication). Thus, past movement of the Washington Cascades hybrid
zone or gene flow out of it could account for Townsend's males at Tieton
recognizing hermit males as threats. To evaluate this possibility, we
presented the same set of mounts to Townsend's males in the Selkirk Mountains,
more than 300 km east of the Cascade hybrid zones
(Figure 1). In the Selkirks,
Townsend's warblers are phenotypically pure and carry no hermit haplotypes
(Rohwer S, Bermingham E, unpublished data).
The aggressive response of Selkirk Townsend's warblers to all three mounts was almost identical to the response of Tieton Townsend's warblers (Figure 3; F = 0.135, df = 1,173, p =.71). Furthermore, there was no interaction between locality and the type of mount presented (Figure 3; F = 0.162, df = 2,173, p =.85).
|
Maximal aggressive differences between localities
We sought to compare maximal aggressive differences between localities;
thus, we used the response of Townsend's and hermits to conspecific mounts and
the average response of hybrids to both parental mounts. For this comparison
to be valid, the responses to chickadee mounts had to be the same for all
three groups (hybrids, hermits, and Townsend's). Unfortunately, this was not
the case (Figure 2; F
= 3.2, df = 2, 48, p =.051): hybrids responded more strongly to the
chickadee mount than did hermits (p =.051). Most of the stronger
response of hybrids was due to two hybrid males striking the chickadee mount
on their first flights over it, something no parental did. We could compare
the aggressive responses of males at the parental localities because they did
not differ in their response to chickadee mounts (p =.536).
Tieton Townsend's were 35% more aggressive toward Townsend's mounts than were Cathlamet hermits toward hermit mounts (Figure 4a; t = -2.35, df = 75, p =.021). Thus, Townsend's males seem to be more aggressive than hermit males. These differences were also apparent in all behavioral variables (Table 2). Pure Townsend's warblers at Tieton were more active (more flights, pecks, and wing flicks) and spent little time out of view than were pure hermit warblers at Cathlamet (Table 2). For both localities and all behaviors, the response to chickadee mounts was less aggressive than the response to conspecific mounts (Table 2).
|
|
For both the principal components analysis (Figure 2) and the individual variables (Table 2), the aggressive ranking of the three phenotypic classes was: Townsend's > hybrids > hermits. To include hybrids in this ranking, we averaged the response of hybrids to the two warbler mounts (Figure 2). Inclusion of hybrids in this rank ordering is valid only if the two males at the hybrid locality that struck the chickadee mount were anomalies. This rank ordering is supported below, where the chickadee control was not needed.
Heterospecific comparisons between localities
Differences in aggression toward mounts of the other species should inform
us of the relative ability of the parental species to disperse and establish
territories in regions dominated by the other species. Tieton Townsend's were
61% more aggressive toward hermit mounts than were Cathlamet hermits toward
Townsend's mounts (Figure 4b;
t = -4.39, df = 76, p <.0001). This should give
Townsend's a substantial advantage in initial contacts.
Relationship between phenotype and aggressiveness within the hybrid
zone
Our results from mount presentations away from the hybrid zone suggest that
Townsend's warblers are more aggressive than hermit warblers. We further
explore this relationship at a single locality near the phenotypic center of
the hybrid zone by examining the correlation between aggression and phenotype.
For this comparison we used the average response of males from Randle, near
the center of the zone, to mounts of the two parentals. There was no
relationship between phenotype and aggressiveness
(Figure 5; F = 0.42,
r =.13, df = 1,23, p =.52). This result was surprising
because we expected the substantial difference in aggression between parentals
to have a genetic component. If there is a genetic component, then the forces
of selection that are keeping these zones narrow should produce linkage
disequilibrium between aggression and phenotype.
Competitive sorting across the zone offers a possible explanation for this
surprising result. These warblers are patchily distributed across the
landscape in the breeding season and hold interspecific territories
(Pearson and Manuwal, 2000).
If competitiveness is important in the acquisition of territories within
habitat neighborhoods, then males within a neighborhood may be similar in
aggressiveness. Males can hold their territories within a neighborhood only if
they are aggressively superior to males that challenge them, and males seeking
to establish territories can displace owners in these neighborhoods only if
they are aggressively superior. These competitive interactions should result
in the resident males of any given neighborhood being similar in
aggressiveness, regardless of their phenotype. However, neighborhoods should
be rank ordered in aggression across the hybrid zone because Townsend's males
are more aggressive than hermit males. All birds will tend to move away from
Townsend's side of the zone where most males are highly aggressive. However,
reasonably aggressive birds should not have to move far away from the
Townsend's side of the zone to find territories that they can hold. But less
aggressive birds should have to disperse farther toward the hermit side of the
zone before they find territories they can hold.
We tested this model of competitive sorting in 1997 and 1998 by conducting mount presentations at five localities across approximately three-quarters of the zone. For both parental mounts, mean aggressive response and mean phenotypic score were significantly correlated, even though the use of locality means gives us only five points for these analyses (Figure 6; hermit mount: F = 14.1, df = 1,3, r =.91, p =.03; Townsend's mount: F = 9.6, df = 1,3, r =.87, p =.054). These results are consistent with the idea that competitive sorting organizes males into neighborhoods that are characterized by similarities in competitive ability. Thus the association between aggression and phenotype is lost within neighborhoods (Figure 5) but evident between neighborhoods (Figure 6).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Competitive asymmetries and movement of the zone
Character transition curves suggest that the hybrid zones between hermit and Townsend's warblers are moving such that Townsend's warblers are replacing hermit warblers (Rohwer and Wood, 1998
), and the experiments described here implicate differences in
aggression between the parental species in the movement of these zones. At
localities outside the Washington Cascades hybrid zone, Townsend's warblers
are more aggressive toward conspecific mounts than are hermit warblers
(Figure 4a). Townsend's males
outside the zone are far more aggressive toward hermit mounts than hermit
males are toward Townsend's mounts (Figure
4b). Thus, Townsend's males should have a large competitive
advantage over hermit males in acquiring territories. These results, along
with differences in clutch size (Pearson
and Rohwer, 1998) indicate that Townsend's warblers are replacing
hermit warblers.
If aggressive differences among males are contributing to the apparent
movement of these hybrid zones, then Townsend's males must be better able to
disperse into and across the zone and to establish territories than hermit
warblers. In addition, Townsend's males that do successfully disperse must
breed. Rohwer and Wood (1998)
provide evidence of substantial dispersal, and Pearson
(2000) shows that Townsend's
and Townsend's-like males near the center of the zone are more likely to breed
than hermit males.
The competitive ranking of hybrids relative to parentals will affect the width and movement of the zone. By all measures of aggression comparing parental and hybrid localities, hybrids were intermediate between parentals in aggressiveness (Figure 2, Table 2). Additionally, we found that hermit-like males were least aggressive and Townsend's-like males were most aggressive when comparing phenotype and aggression across five localities spanning the hybrid zone (Figure 6). The intermediate ranking of hybrids should accelerate the movement of the zone.
Why are Townsend's more aggressive than hermits?
Systematic changes in densities of territorial males could explain
differences in aggression across this zone. At high densities territorial
males may have high testosterone levels, especially during territory
establishment (Beletsky et al.,
1990). Because testosterone is intimately linked to territorial
behavior (Wingfield et al.,
1987), density differences between localities could entrain
differences in aggression. If this is true, Townsend's males should occur at
the highest densities, hybrids at intermediate densities, and hermit males at
the lowest densities. During 5 years of field work within the zone, we noticed
no systematic change in densities across the hybrid zone that would be
consistent with this explanation.
If population densities decrease from the Townsend's side of the zone to
the hermit side of the zone, then Townsend's males within the zone should be
less aggressive than Townsend's males outside the zone, and hermit males
should show the opposite trend. In a post hoc test we compared the
aggressiveness of the parental species within the hybrid zone to the parental
species outside the zone and found no difference for either comparison
(p 
.44). Contrary to the expected relationship, pure Townsend's
within the zone where slightly more aggressive than pure Townsend's outside
the zone, and pure hermits within the zone where slightly less aggressive than
pure hermits outside the zone.
Aggressive differences between parental males cannot be explained by
differences in arrival dates or by differences in habitat. Males of all
phenotypes return to the breeding grounds at the same time in spring
(Pearson and Rohwer, 1998) and
occupy similar habitats (Pearson and
Manuwal, 2000). Thus, preemption of the best habitats through
early arrival cannot explain the aggressive superiority of Townsend's males.
We suspect that their higher aggressiveness is an evolved consequence of
differences in their evolutionary history.
Why are Townsend's aggressive to hermits and not vice versa?
The high levels of aggression shown by Townsend's males outside the hybrid
zone toward hermit warbler mounts is remarkable
(Figure 2a). Most species that
are interspecifically territorial recognize heterospecifics only in regions of
sympatry (Catchpole, 1978;
Järvi et
al., 1978; Prescott,
1987; Reed, 1982);
the same is true for species that hybridize
(Baker, 1991;
Emlen et al., 1975;
Rohwer, 1972).
There a several possible explanations for the strong aggressive response by
Townsend's males toward heterospecific mounts. Because the Tieton locality
lies just northeast of the Washington Cascades zone, Townsend's warblers at
this locality could have assimilated genes coding for aggressiveness toward
hermit warblers either as a result of past movement of the zone through this
region or as a result of gene flow. To evaluate this possibility, we measured
the aggressiveness of Townsend's males at a locality more than 300 km east of
the current hybrid zones in the Selkirk Mountains. It is unlikely that gene
flow from the hybrid zone could have affected this population because the
males do not contain hermit haplotypes. Furthermore, it is unlikely that a
hybrid zone between these warblers ever moved through the Serlkirks because
these warblers do not make contact in this region, and the Cascade and Olympic
Mountain zones appear to be moving from north to south
(Rohwer and Wood, 1998;
Rohwer, unpublished data). Interestingly, Townsend's males from the Selkirk
Mountains were no less aggressive toward hermit mounts than Townsend's males
from Tieton (Figure 3). Thus
the high levels of aggression shown by Townsend's males toward hermit warblers
probably evolved before these warblers made contact and is not a consequence
of recent gene exchange.
Competitive interactions between these species during the winter could also
explain this aggressive asymmetry. During winter, the ranges of these warblers
overlap in the mountains of Mexico and Guatemala
(Howell and Webb, 1995), where
they are found together in mixed-species flocks
(Hutto, 1987), and they may
compete for food resources (Greenberg et
al., 1993). Mistaken identity is a third possible explanation for
this aggressive asymmetry. The subdued plumage of yearling Townsend's males
bears some resemblance to the plumage of hermit warblers. Thus, Townsend's
males may treat hermit males as yearling Townsend's males.
Phenotype and aggression within the zone
If aggressive differences between these species are based in part on
genetic differences, then aggression and phenotype should be correlated among
the individuals at a hybrid locality. We failed to find this relationship
(Figure 5), and we attribute
this failure to competitive sorting. Linkage disequilibrium theory predicts an
association between aggression and phenotype both within neighborhoods and
across landscapes (Barton and Gale,
1993). However, competitive sorting can eliminate the association
between aggression and phenotype within neighborhoods. Mean aggressive
response of warblers at localities spanning three-quarters of the hybrid zone
support competitive sorting. Males at more Townsend's-like localities were
much more aggressive than males at more hermitlike localities
(Figure 6). The strength of
this relationship was striking given the small sample size (see Results). This
organization of aggressiveness across the zone may be achieved by competition
within neighborhoods, causing males to move greater and lesser distances
across the zone, according to their fighting abilities. The behavioral
mechanisms that underlie competitive sorting are explained in the Results.
The clear pattern we obtained from mount presentations across the 125 km width of this avian hybrid zone suggests that mount presentations offer a useful approach for comparing aggressive differences within and between species over large geographic areas. We know of no other study of a hybrid zone where behavioral differences between parentals and hybrids have been systematically addressed across the zone. Our measures of aggression within and between localities spanning this warbler hybrid zone suggest that competitive sorting organizes dispersal at the landscape level.
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
We thank the University of Washington Burke Museum staff and especially Chris Wood for logistical support. Keen insight from Matthias Leu and Dave Manuwal led to many productive discussions of this work. The manuscript benefited from comments by Sharon Birks, Luke Butler, Sergei Drovetski, Chris Filardi, Jon Herron, Matthias Leu, Dave Manuwal, Catherine Smith, Gary Voelker, Ron Ydenberg, and three anonymous reviewers. This research was supported by a Burke Museum Eddy Fellowship in Ornithology and by direct contributions from Garrett Eddy to the University of Washington Burke Museum.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Baker MC, 1991. Response of male indigo and lazuli buntings and their hybrids to song playback in allopatric and sympatric populations. Behaviour 119:225-242.
Barton NH, Gale KS, 1993. Genetic analysis of hybrid zones. In: Hybrid zones and the evolutionary process (Harrison RG, ed.. Oxford: Oxford University Press;13 -45.
Barton NH, Hewitt GM, 1985. Analysis of hybrid zones.Annu Rev Ecol Syst 16:113-148.[Web of Science]
Barton NH, Hewitt GM, 1989. Adaptation, speciation and hybrid zones. Nature 341:497-503.[Medline]
Beletsky LD, Orians GH, Wingfield JC, 1990. Steroid hormones in relation to territoriality, breeding density, and parental behavior in male yellow-headed blackbirds. Auk 107:60-68.
Catchpole CK, 1978. Interspecific territorialism and competition in Acrocephalus warblers as revealed by playback experiments in areas of sympatry and allopatry. Anim Behav 26:1072-1080.
Emlen ST, Rising JD, Thompson WL, 1975. A behavioral and morphological study of sympatry in the indigo and lazuli buntings of the Great Plains. Wilson Bull 87:145-179.
Greenberg R., Caballero CM, Bichier P, 1993. Defense of homopteran honeydew by birds in the Mexican highlands and other warm temperate forests. Oikos 68:519-524.
Hewitt GM, 1988. Hybrid zones—natural laboratories for evolutionary studies. Trends Ecol Evol 3:158-167.
Howell SNG, Webb S, 1995. A guide to the birds of Mexico and northern Central America. New York: Oxford University Press.
Hutto RL, 1987. A description of mixed-species insectivorous bird flocks in western Mexico. Condor 89:282-292.
Jackson WM, Wood CS, Rohwer S, 1992. Age-specific plumage characters and annual molt schedules of hermit warblers and Townsend's warblers. Condor 94:490-501.
Järvi T, Radesäter T, Jakobsson S, 1978. Aggressive responses of two hole-nesting passerines, Parus major and Ficedula hypoleuca, to the playback of sympatric species song.Ornis Fenn 55:154-157.
Langston NE, Freeman S, Rohwer S, Gori D, 1990. The evolution of female body size in red-winged blackbirds: the effects of timing of breeding, social competition, and reproductive energetics.Evolution 44:1764-1779.
Pearson SF, 2000. Behavioral asymmetries in a moving
hybrid zone. Behav Ecol
11:84-92.
Pearson SF, Manuwal DA, 2000. Influence of niche overlap and territoriality on hybridization between hermit and Townsend's warblers. Auk 117:175-183.
Pearson SF, Rohwer S, 1998. Influence of breeding phenology and clutch size on hybridization between hermit and Townsend's warblers. Auk 115:739-745.
Prescott DRC, 1987. Territorial responses to song playback in allopatric and sympatric populations of alder (Empidonax alnorum) and willow (E. traillii) flycatchers. Wilson Bull 99:611-619.
Reed TM, 1982. Interspecific territoriality in the chaffinch and great tit on islands and the mainland of Scotland: playback and removal experiments. Anim Behav 30:171-181.
Rohwer S, 1972. Significance of sympatry to behavior and evolution of great plains meadowlarks. Evolution 27:44-57.
Rohwer S, 1978. Passerine subadult plumages and the deceptive acquisition of resources: test of a critical assumption.Condor 80:173-179.
Rohwer S, 1982. The evolution of reliable and unreliable badges of fighting ability. Am Zool 22:531-546.
Rohwer S, Røskaft E, 1989. Results of dyeing male yellow-headed black birds solid black: implications for the arbitrary identity badge hypothesis. Behav Ecol Sociobiol 25:39-48.
Rohwer S, Wood C, 1998. Three hybrid zones between hermit and Townsend's warblers in Washington and Oregon. Auk 115:284-310.[Web of Science]
Røskaft E, Rohwer S, 1987. An experimental study of the red epaulettes and the black body colour of male red-winged blackbirds. Anim Behav 35:1070-1077.
Wingfield JC, Ball GF, Dufty AM Jr, Hegner RE, Ramenofsky M,1987 . Testosterone and aggression in birds. Am Sci 75:602-608.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  M. Krosby and S. Rohwer A 2000 km genetic wake yields evidence for northern glacial refugia and hybrid zone movement in a pair of songbirds Proc R Soc B, February 22, 2009; 276(1657): 615 - 621. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. A. Duckworth and A. V. Badyaev From the Cover: Coupling of dispersal and aggression facilitates the rapid range expansion of a passerine bird PNAS, September 18, 2007; 104(38): 15017 - 15022. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. M. den Hartog, S. R. de Kort, and C. ten Cate Hybrid vocalizations are effective within, but not outside, an avian hybrid zone Behav. Ecol., May 1, 2007; 18(3): 608 - 614. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. F. Pearson Behavioral asymmetries in a moving hybrid zone Behav. Ecol., January 1, 2000; 11(1): 84 - 92. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||