? 2005 The Society for the Study of Evolution. All rights reserved.
Evolution, 59(2), 2005, pp. 361–373
PARALLEL EVOLUTION OF SEXUAL ISOLATION IN STICKLEBACKS
JANETTE WENRICK BOUGHMAN,1,2HOWARD D. RUNDLE,3,4AND DOLPH SCHLUTER5,6
1Department of Zoology, University of Wisconsin, Madison, Wisconsin 53706
3Department of Zoology, University of Queensland, Brisbane, Queensland 4072, Australia
5Department of Zoology, University of British Columbia, Vancouver, British Columbia V6T1Z4, Canada
how natural and sexual selection cause sexual isolation. Here, we investigate the roles of divergent natural and sexual
selection in the evolution of sexual isolation between sympatric species of threespine sticklebacks. We test the
importance of morphological and behavioral traits in conferring sexual isolation and examine to what extent these
traits have diverged in parallel between multiple, independently evolved species pairs. We use the patterns of evolution
in ecological and mating traits to infer the likely nature of selection on sexual isolation. Strong parallel evolution
implicates ecologically based divergent natural and/or sexual selection, whereas arbitrary directionality implicates
nonecological sexual selection or drift. In multiple pairs we find that sexual isolation arises in the same way: assortative
mating on body size and asymmetric isolation due to male nuptial color. Body size and color have diverged in a
strongly parallel manner, similar to ecological traits. The data implicate ecologically based divergent natural and
sexual selection as engines of speciation in this group.
Mechanisms of speciation are not well understood, despite decades of study. Recent work has focused on
sexual isolation, speciation.
Body size, courtship behavior, divergent natural selection, divergent sexual selection, nuptial color,
Received March 10, 2004. Accepted November 22, 2004.
How new species arise is one of the most significant un-
answered questions in evolutionary biology. Current research
has shifted focus from the geographic context of speciation
(i.e., allopatry vs. sympatry) to the mechanisms that drive
the evolution of reproductive isolation between populations
(Schluter 2001). Evidence is rapidly accumulating from di-
verse taxa that divergent natural and sexual selection are
important causes of speciation in nature (e.g., Funk 1998;
Schemske and Bradshaw 1999; Rundle et al. 2000; Schluter
2000, 2001; Via et al. 2000; Boughman 2001; Podos 2001;
Nosil et al. 2002). Less is known, however, about the details
by which this occurs. For example, what kinds of traits cause
reproductive isolation and what forms of selection act on
them? Is speciation primarily the by-product of adaptive di-
vergence, or does selection also commonly reinforce pre-
mating isolation in sympatry? When sexual selection is in-
volved, is the divergence of male mating traits and female
preferences arbitrary with respect to environment or ulti-
mately the product of ecologically based divergent selection
(e.g., sensory drive; Endler 1992, 1993; Boughman 2002)?
Answers to such questions will require an understanding of
the speciation process that goes beyond simply inferring a
role for natural or sexual selection.
Here we focus on the evolution of sexual isolation. We
address two alternative hypotheses by which natural and sex-
ual selection may be involved in the evolution of sexual
isolation that differ in the role of ecologically based selection.
In the first hypothesis, termed ecological speciation, adap-
tation to the environment drives the evolution of sexual iso-
lation, including the divergence among populations in both
mating signals and preferences. Both natural selection and
sexual selection can be involved, as long as differences in
mating traits and preferences (hereafter ‘‘mating traits’’) that
confer reproductive isolation arise ultimately due to ecolog-
ically based divergent selection (Schluter 2000, 2001). Eco-
logical speciation is a general hypothesis that includes a va-
riety of mechanisms involving contrasting selection acting
on populations inhabiting separate environments or niches.
For example, under this hypothesis mating traits may diverge
between populations as a by-product of natural selection
adapting the populations to their different environments
(Mayr 1942; Dobzhansky 1951). They may also diverge due
to sexual selection caused by different signaling environ-
ments (Endler 1992; Boughman 2002), or due to spatially
varying selection on secondary sexual traits (Lande 1982;
Day 2000) or on communication systems (Ryan and Rand
1993). Evolution of mating traits in correlation with envi-
ronment is predicted not only by natural selection but by
several forms of sexual selection, including sensory drive,
context-dependent good genes, and spatial variation in the
optimal male trait (Schluter 2000).
In the second hypothesis, termed nonecological speciation,
differences in mating traits do not depend on differences in
environment. Several mechanisms may again be involved.
For example, mating traits may diverge between populations
due to genetic drift (Gavrilets and Boake 1998), differences
among populations in the amount and patterns of genetic
variation (Schluter et al. 2004), or nonecological sexual se-
lection driving traits in arbitrary directions (e.g., Parker and
Partridge 1998). However, attention has focused on sexual
selection because it has long been thought to be a powerful
mechanism fostering speciation, and recent comparativestud-
ies implicate it (reviewed in Panhuis et al. 2001). Noneco-
logical sexual selection causes the mating traits that confer
reproductive isolation to evolve haphazardly with respect to
environment. This occurs because selection on mating traits
arises due to interactions between the sexes (West-Eberhard
1983) and thus there is no reason to expect parallel change
JANETTE W. BOUGHMAN ET AL.
in separate populations, even when they inhabit similar en-
vironments. Evolution of mating traits in arbitrary directions
is predicted by various models of sexual selection such as
Fisher’s runaway (Lande 1981; Turner and Burrows 1995;
Pomiankowski and Iwasa 1998; Higashi et al. 1999) and sex-
ual conflict (Arnqvist et al. 2000; Gavrilets 2000; Arnqvist
and Rowe 2002; Rice and Chippindale 2002). Because the
directions of evolutionary change can be diverse, allopatric
populations are very likely to differ in mating traits and con-
sequently lead to sexual isolation (Turelli et al. 2001). This
is the more traditional view of how sexual selection causes
reproductive isolation. We refer to this form of speciation as
sexual selection in arbitrary directions.
We distinguish between these hypotheses using the concept
of parallel speciation. Parallel speciation is the parallel evo-
lution of reproductive isolation in populations descended
from a common ancestor that have colonized similar novel
environments (Schluter and Nagel 1995). Parallel speciation
occurs when reproductive isolation evolves as the by-product
of ecologically based divergent selection. It predicts that the
traits determining reproductive isolation should evolve in
correlation with the environment. We can thus discriminate
between the two hypotheses above by testing a central con-
trasting prediction: whether sexual isolation evolves in par-
allel. Finding extensive parallel evolution of sexual isolation
would support the hypothesis that ecologically based diver-
gent selection is the ultimate cause of mating trait divergence
(ecological speciation). In contrast, finding nonparallel evo-
lution of sexual isolation would support the hypothesis that
nonecological processes cause mating trait divergence (e.g.,
sexual selection in arbitrary directions). Determining the ex-
tent of parallel evolution in sexual isolation is thus an im-
portant step toward understanding how various forms of se-
lection contribute to speciation in nature.
Parallel speciation is not a perfect test for distinguishing
ecological and nonecological speciation for two reasons.
First, ecological selection could cause nonparallel divergence
of mating traits that confer reproductive isolation if replicate
environments vary somewhat in ecology. Thus, although find-
ing parallel evolution of sexual isolation strongly implicates
ecologically based divergent selection in the speciation pro-
cess, a failure to do so does not rule it out. Doing so requires
that the alternative hypothesis of adaptation to locally unique
environments also be rejected. One way to distinguish be-
tween these possibilities is to compare the divergence be-
tween replicate populations in both mating and ecological
traits (traits involved in resource acquisition and predator
defense). Sexual selection in arbitrary directions should cause
nonparallel evolution of mating traits only, whereas nonpar-
allel adaptation to local environments will be reflected in
ecological traits as well.
Here, we study parallel evolution of sexual isolation be-
tween sympatric limnetic and benthic species of threespine
sticklebacks. These species pairs inhabit six lakes in south-
western British Columbia, Canada, and evolved indepen-
dently following the separate colonization of each lake by
the marine threespine stickleback (Gasterosteus aculeatus)
following the retreat of the glaciers at the end of the Pleis-
tocene 10,000–12,000 years ago (McPhail 1993, 1994). In
each lake the two ecotypes are ecologically and morpholog-
ically highly differentiated and exploit alternate foraging
niches: the limnetic forages primarily on zooplankton in the
open water and the benthic forages on invertebrates in the
littoral zone (Bentzen and McPhail 1984; McPhail 1984,
1992, 1994; Schluter and McPhail 1992).
Sexual isolation of limnetic and benthic sticklebacks has
evolved, to a large extent, in parallel. Previous work has
shown that, despite separate origins, limnetics from these
lakes are all reproductively compatible, as are benthics,
whereas limnetics and benthics are reproductively isolated
no matter what combination of lakes they are from (Rundle
et al. 2000). Such parallel evolution of reproductive isolation
strongly implicates divergent natural selection in speciation
because all limnetics have adapted to a similar environment
that is distinct from that of benthics, and sexual isolation has
evolved in correlation with environment. Additional evidence
implicates sexual selection in the evolution of sexual isola-
tion but shows that it also arises from differences in envi-
ronment (Boughman 2001). These results support the first
hypothesis above, arguing that sexual isolation has evolved
ultimately due to ecologically based divergent selection.
Even so, questions remain about the contribution of various
forms of natural and sexual selection to stickleback specia-
tion. In particular, we found substantial variation among pairs
of populations in their reproductive compatibility that is not
obviously correlated with environment (e.g., see fig. 2, Run-
dle et al. 2000; fig. 4, Boughman 2001). This suggests that
nonecological processes, such as sexual selection in arbitrary
directions, may also contribute to sexual isolation.
Here, we explore this possibility by testing the specific
prediction that the mating traits determining sexual isolation
have evolved in parallel. We do this to gain insight into the
relative roles of ecological and nonecological selection in the
origins of these species. We focus on the sympatric species
pairs because it is these combinations of populations for
which reproductive isolation matters in nature; gene flow is
only possible in sympatry and selection in sympatry has likely
been important in the evolution of reproductive isolation
(Rundle and Schluter 1998, 2004; Albert and Schluter 2004).
While several traits important to sexual isolation have been
identified (Nagel and Schluter 1998; Boughman 2001), the
extent of their parallel evolution has not been directly tested,
nor have the specific traits conferring sexual isolation in each
pair been identified. We consider two morphological (body
size and male nuptial color) and several behavioral traits
(male and female courtship behavior), chosen because they
have been implicated in previous studies (size: Nagel and
Schluter 1998; color: Boughman 2001; behavior: Ridgway
and McPhail 1984; McLennan and McPhail 1990). We ask
about the relative contributions of these mating traits to sex-
ual isolation and explore the form of selection causing their
We examine the extent of parallel speciation by addressing
the following questions. First, are the same traits responsible
for sexual isolation in all pairs? This should be true if mating
traits evolve due to ecologically based divergent selection
(ecological speciation), but not if sexual selection in arbitrary
directions drives speciation. Second, has divergence in mat-
ing traits between sympatric limnetics and benthics occurred
in parallel? If sexual isolation is evolving in parallel due to
PARALLEL SEXUAL ISOLATION
divergent selection, not only should the same traits be re-
sponsible for sexual isolation in separate lakes, but the di-
rection of evolutionary change (i.e., increase or decreasefrom
ancestor to descendent) and the extent of divergence between
sympatric species should also be similar. In contrast, sexual
selection in arbitrary directions is not expected to produce
similar outcomes in replicate populations, especially not in
correlation with environment.
MATERIALS AND METHODS
Populations used in this study originate in coastal British
Columbia, Canada, in the Georgia Strait region (for locations
of study populations see map in Schluter and McPhail 1992).
We studied limnetics and benthics from three lakes: Enos,
Paxton, and Priest Lakes. Despite the phenotypic similarity
among lakes of all limnetics and of all benthics, two separate
lines of genetic evidence indicate that the species pairs from
these three lakes are the result of separate colonization events
of the marine stickleback into freshwater (Rundle and Schlu-
ter 2004), as originally proposed by McPhail (1993). The
first comes from mitochondrial DNA (mtDNA; Taylor and
McPhail 1999). The limnetic-benthic pairs within each lake
are characterized by unique assemblages of mtDNA haplo-
types. The majority of these differ from common marine
haplotypes by a single restriction site. In contrast, mtDNA
haplotypes from different lakes always differ by more than
one site. These patterns suggest that the separate assemblages
in each lake trace their origin independently to the marine
environment and not to each other. The second line of evi-
dence comes from an analysis of allelic variation at six mi-
crosatellite loci (Taylor and McPhail 2000). If limnetics and
benthics each arose only once and then colonized these lakes,
genetic variation should be structured into limnetic and ben-
thic classes. However, almost none of the variation (2.2–
4.4%, not significantly different from zero) can be partitioned
between these classes. In addition, although poorly resolved,
the maximum-likelihood phylogeny suggests independent or-
igins for limnetics and benthics from these lakes and is a
significantly better fit to the data than one enforcing mono-
phyly of either. For these reasons, we study the parallel evo-
lution of reproductive isolation by focusing on limnetics and
benthics from these three lakes, treating each pair as an in-
dependent evolutionary replicate. Throughout the paper,
when referring to all limnetics or all benthics we use the term
‘‘ecotype,’’ and when referring to sympatric limnetics and
benthics we use the terms ‘‘species’’ or ‘‘species pair.’’
All fish used for morphological and behavioral measure-
ments were caught in the wild using minnow traps or seine
nets. We collected courtship data during 1467 mating trials
conducted in April to July of eight years: 1992, 1993, 1996,
1997, 1999–2002. Not all data were collected for all popu-
lations. Males and females were collected from the wild with
minnow traps and transported to Vancouver, British Colum-
bia, where they were held in 102-L aquaria and maintained
on a 16:8 L:D cycle at 18?C. All fish were fed once per day
with frozen Artemia and chironomid larvae.
Mating traits and reproductive isolation were assessed in
no-choice mating trials (described in detail in Nagel and
Schluter 1998; Boughman 2001). Single males were placed
in 102-L aquaria and allowed to build a nest. Once males
had completed a nest and were actively courting, a single
gravid female was introduced into the aquarium and inter-
acted directly with the male until spawning occurred or for
20 min. We did not allow females to deposit eggs in a male’s
nest, so as soon as a female entered a nest we gently squeezed
her tail with long forceps to induce her to exit. We used
females only once. We conducted multiple trials with most
males, in most cases pairing each male with a single female
from his own population and with a single female from one
or more other populations. Males were used only once in
each analysis because we did not repeat combinations. On
completion of each trial, we verified that females were indeed
ready to spawn by gently squeezing the abdomen and looking
for ripe eggs in the oviduct. Trials in which females were
deemed nonreceptive by this method were excluded. We es-
timated population means for behavioral traits using only
trials within population. In contrast, we estimated the extent
of sexual isolation using only trials between species within
We recorded behavior with an event recorder (Observer,
Noldus Technologies, Wageningen, The Netherlands), and
recorded the following male courtship behavior: zig-zag,bite,
and lead (described in Rowland 1989). We calculated an
index of courtship aggressiveness as: N(zig-zag)/N(zig-zag)
? N(bite). Female response behavior included follow and
examine (Rowland 1989). To adjust the number of recorded
behaviors for varying trial duration, we calculated the rate
of each behavior per minute. We also estimated female re-
sponsiveness to males as the proportion of follows that led
to nest examination: N(examine)/N(follow). We calculated
female preference for male color by regressing the rate of
nest examination on male color score. The area of red color
for live males was scored prior to behavioral trials by eye
on a scale of 0 (no red coloration) to 5 (large area of intense
red coloration). Our measure of body size was standard
length, which we measured with vernier calipers accurate to
0.02 mm. We collected color, courtship, and body size data
on Paxton benthics and limnetics (N ? 63 and 78, respec-
tively), Enos benthics and limnetics (N ? 34 and 64), Emily
limnetics (N ? 44), and color and size data on Priest benthics
and limnetics (N ? 20 each).
To measure ecological traits (gill raker number and plate
number) we anesthetized fish in MS-222, placed them in 10%
formalin for at least a week, stained them with alizarin red,
and then stored them in 40% isopropyl alcohol. We counted
the number of gill rakers on the first gill arch. Previous studies
have shown that the number of gill rakers predicts the effi-
ciency of foraging on plankton or benthic invertebrate prey
(Schluter 1993). We counted the number of armor plates on
the left side of the body. Armor plates are a key adaptation
against vertebrate predators (Bell et al. 1993; Reimchen
JANETTE W. BOUGHMAN ET AL.
1994) and differ between ecotypes (Schluter and McPhail
1992; Vamosi 2002). We collected data on ecological traits
for benthics and limnetics from Paxton (N ? 24 and 25,
respectively), Enos (N ? 30 and 40), Priest (N ? 39 and 23),
Emily (N ? 39 and 39), and Hadley Lakes (N ? 43 and 10).
We analyzed trait means for each population by one-way
ANOVA, and we compared means with the Tukey-Kramer
method. For mating trials with the male and female of op-
posite ecotypes, we also calculated the difference between
partners in mating traits that could underlie sexual isolation.
These include size difference, color difference, and courtship
difference. We calculated size difference between male and
female partners in standard length (male length minus female
length). Thus positive values indicate that males are larger,
whereas negative values indicate that females are larger. We
plot these means in Figure 2, but use the absolute value of
means in all analyses. We use absolute values because we
are interested in how the magnitude of difference between
ecotypes affects sexual isolation and have no a priori reason
to consider one ecotype as the standard. For traits expressed
only in males (e.g., nuptial color and courtship score), we
assigned to every female the average value of the male trait
in her own population. We then subtracted the value of her
partner and took the absolute value of this difference. The
values we calculated are thus the difference between her male
partner’s value and her expectation for males in her own
population (best estimated as her own population mean). A
significant size difference term would indicate that females
reject heterospecific males who differ substantially in size,
for instance, that large benthics reject small limnetics. A
significant color or courtship difference term would indicate
that females reject heterospecific males when they differ from
the conspecific mean, for instance, that limnetic females re-
ject dull benthic males or those that court aggressively. We
measured reproductive compatibility as the proportion of tri-
als in which spawning occurred during trials between lim-
netics and benthics (male and female of opposite ecotypes).
To evaluate the prediction that the same traits are respon-
sible for sexual isolation in all pairs, we tested whether the
traits that confer reproductive isolation varied among lakes.
These analyses used data from trials with the male and female
of opposite ecotypes. To do this required a two-step process
in which we first identified mating traits that confer repro-
ductive isolation using stepwise regression and then tested
if their effects varied among lakes by testing interactionterms
in a logistic regression. Testing interaction terms is a direct
test of the first prediction, that sexual isolation is based on
the same traits in all pairs. First, we conducted a stepwise
regression to find the model that best explained variation in
spawning between ecotypes, using size and male mating
traits, female ecotype, and lake in the full model. We included
all measured mating traits to identify the traits that actually
contribute to sexual isolation, as previous work had suggested
several traits might be important but had not considered them
jointly (Rundle and Schluter 1998; Nagel and Schluter 1998;
Boughman 2001). The reduced model included differences
in size, color, and courtship (as calculated above) as contin-
uous variables, with female ecotype and lake as categorical
Next, because the response variable for sexual isolation is
binomial (spawn or no-spawn) we used logistic regression
with the reduced model to test the significance of terms. The
direct test of the prediction for parallel sexual isolation is to
test the significance of the interaction between lake and the
phenotypic variables in the model. Significant interaction
terms indicate that mating traits have variable effects on sex-
ual isolation among pairs, while nonsignificant interaction
terms suggest that all pairs use the same traits to discriminate
against heterospecifics. We also asked if limnetics and benth-
ics differ in the basis of sexual isolation by including inter-
actions between female ecotype and size, color, and courtship
differences. A significant interaction here indicates that the
ecotypes differ in how they use a trait for reproductive iso-
lation. The effects could be in different directions, or one
ecotype may use the trait while the other does not.
We also delved into how body size differences influence
mating probability. Because our analyses showed that the
magnitude of difference in body size affected the probability
of spawning between ecotypes, we used multiple regression
to investigate whether this size effect was due to male size,
female size, or their difference. We found no effect of male
size or female size, so used only the difference variable in
all analyses. We looked for size assortative mating between
ecotypes by calculating the mean size difference in trials with
and without spawning. Then, we compared this to the within-
ecotype pattern. We analyzed both between- and within-eco-
type size differences using t-tests.
We transformed data for all analyses to better meet as-
sumptions of ANOVA using square-root transformations for
behavioral and count data and log transformations for size
data. However, we present back-transformed values in tables
and figures for ease of interpretation.
We then turned to testing the second prediction, that di-
vergence in mating traits has occurred in parallel for sym-
patric pairs. For all traits, we quantified the degree of parallel
evolution between pairs using the quantity ?, which compares
the magnitude of the difference in a trait between ecotypes
(limnetic or benthic) averaged across the three pairs, with the
magnitude of the interaction between ecotype and pair
(Schluter et al. 2004). Parallel evolution is indicated by a
significant main effect of ecotype and weak interaction be-
tween ecotype and pair. The quantities are provided by fitting
the two-factor fixed-effects ANOVA model to the data,
? ? ? ? ? ? ? ?? ? ? ,
j ijkij ijk
where Yijkis the trait value of individual k, modeled as the
sum of a constant (?), the effects of its ecotype i (?i) and
pair j (?j), the deviation (??ij) resulting from an interaction
between its ecotype and pair, and a random error term (?ijk).
The measure of parallel evolution is quantified as the dif-
ference between the variance components for ecotype and the
interaction relative to their sum,
A ? AB
A ? AB
PARALLEL SEXUAL ISOLATION
benthic pairs. Data are for trials between ecotypes (benthics and
limnetics). Interaction terms test for parallel slopes of absolute dif-
ferences in mate recognition traits on ecotype or lake. Populations
include Paxton, Enos, and Priest benthics and limnetics. Data were
analyzed with logistic regression on transformed variables.
Predictors of spawning between ecotypes for limnetic-
Source of variation df
Absolute length difference
Absolute male color difference
Absolute courtship difference
Length difference ? lake
Male color difference ? lake
Courtship difference ? lake
Length difference ? female ecotype
Male color difference ? female ecotype
Courtship difference ? female ecotype
* P ? 0.05; ** P ? 0.01.
on spawning between benthic and limnetic sticklebacks. (A) Dif-
ference in size for all populations, and (B) for both ecotypes. (C)
Difference in nuptial color for all populations, and (D) for both
ecotypes. The absolute value of mean difference between male and
female ? standard error is shown for trials where spawning did
(closed circles) and did not (open circles) occur. Populations include
Enos, Paxton, and Priest benthics (EnosB, PaxB, and PriestB, re-
spectively) and limnetics (EnosL, PaxL, and PriestL, respectively).
Values shown in (B) and (D) are means for each ecotype.
The effect of differences in body size and nuptial color
A ?? , (3)
a ? 1
represents the main effect of ecotype, and
(a ? 1)(b ? 1)
represents the interaction between ecotype and pair. The con-
stants a and b are the numbers of ecotypes and pairs, re-
The quantity ? ranges between ?1 and ?1. A value of ?1
represents pure parallel evolution, occurring when traitmeans
of all pairs shift between ecotypes by precisely the same
amounts. A value of 0 represents no parallel evolution and
occurs when shifts in trait means between ecotypes are un-
correlated among pairs. A value of ?1 represents pure an-
tiparallel evolution, occurring when shifts of trait mean be-
tween ecotypes in one pair are exactly opposed by shifts in
trait means of a second pair.
The quantity ? is estimated by substituting the fitted es-
timates of ecotype main effects,
(equal sample size for all ecotype and pair combinations),
the fitted estimates may be extracted using the mean squares
from the fitted ANOVA model and the formulas for the ex-
pected mean squares (e.g., Sokal and Rohlf 1995, p. 333).
In the more typical case of unequal sample sizes the fitted
of the two-factor linear ANOVA model. We obtained these
in S? using the dummy.coef command.
Confidence limits for were calculated using the bootstrap
(Efron 1982). On each iteration a new sample of the data was
generated by resampling with replacement from the original
data. This was repeated 1000 times, leading to 1000 resam-
pled values for . The standard deviation of the 1000 values
is the standard error of . The fraction of the 1000 values of
falling at or below zero provides an approximate P-value
for a one-tailed test of the null hypothesis that ? ? 0.
We used S? (Mathsoft, Seattle, WA) for all statistical
ij, and interactiondeviations
ijinto the above equations. In the case of a balanced design
ijmust be extracted from the coefficients
Parallel Evolution of Sexual Isolation
Limnetics and benthics are reproductively isolated from
one another (assortative mating by ecotype:
0.0001), as found in previous studies. Our analyses indicate
that body size and color play a role, but we found no evidence
that differences in courtship behavior contribute to sexual
isolation (Table 1). There are no main effects for phenotypic
traits, nor are there interactions between the traits and lake.
That we did not find such interactions suggests that all pairs
use traits in the same way to avoid heterospecific matings,
or that any differences among pairs are small relative to other
effects. This confirms the first prediction of the ecological
hypothesis, that the same traits are responsible for sexual
isolation in all pairs.
We did find significant interactions between ecotype and
both body size and color differences, suggesting that lim-
netics and benthics differ in how these traits cause sexual
isolation (Table 1, Fig. 1). Females of both ecotypes reject
heterospecific males who differ substantially in size, although
the effect is stronger for benthic females, and Enos limnetic
females do not show this pattern. Thus, size contributes to
sexual isolation in both ecotypes and leads to assortative
? 26.6, P ?
JANETTE W. BOUGHMAN ET AL.
Courtship, color, and ecological variables are square root transformations, whereas length variables are log10transformed. Sample size
is given first for benthics and then limnetics.
Mean values (on transformed scale) for limnetics and benthics, P-values for differences between ecotypes, and sample sizes.
No. gill rakers
No. armor plates
1.13 ? 0.063
1.43 ? 0.195
1.39 ? 0.119
1.86 ? 0.082
0.35 ? 0.058
1.65 ? 0.050
1.96 ? 0.097
2.21 ? 0.087
1.14 ? 0.059
0.66 ? 0.042
0.71 ? 0.066
0.22 ? 0.056
0.22 ? 0.121
?0.06 ? 0.057
0.83 ? 0.048
0.50 ? 0.040
0.60 ? 0.084
0.09 ? 0.050
1.77 ? 0.004
1.75 ? 0.003
0.51 ? 0.059
1.65 ? 0.004
1.69 ? 0.003
0.64 ? 0.031
4.31 ? 0.013
1.69 ? 0.049
4.86 ? 0.013
2.43 ? 0.050
1N is the number of populations rather than the number of individuals for this variable.
netic-benthic pairs. Lakes include Paxton and Enos for all traits,
and Priest for size and morphological traits as well. ? ranges from
?1 (anti-parallel evolution) to ?1 (pure parallel evolution). See
text for details.
Test of parallel divergence in phenotypic traits for lim-
No. gill rakers
No. armor plates
mating. Oddly, size differences contribute to sexual isolation
between ecotypes (t10? ?2.6, P ? 0.026) even though there
is no evidence for size assortative mating within ecotype (t12
? ?0.5, P ? 0.65).
In contrast, color has asymmetric effects on sexual isola-
tion (Fig. 1). Benthic males have less color on average than
limnetic males. Limnetic females reject the dullest benthic
males, who differ most from conspecific males. Therefore,
large color differences increase sexual isolation. The com-
bination of reduced nuptial color in benthic males and strong
color preference in limnetic females contributes to sexual
isolation between limnetic females and benthic males. Color
differences have the opposite effect on sexual isolation for
benthics. Benthic females reject the dullest limnetic males,
who are most similar to conspecific males. Therefore, it is
small rather than large color differences that increase sexual
isolation between benthic females and limnetic males. In both
ecotypes, females are more likely to mate with bright het-
erospecific males than with dull ones. The similar tendency
to reject dull males contributes to sexual isolation in one
direction but not in the other.
Parallel Divergence of Phenotypic Traits
Mating, behavioral, and ecological traits all show sub-
stantial parallel evolutionary divergence between ecotypes.
This is evidenced both by the number of traits that have
diverged and the extensive parallel nature of this divergence.
Limnetics and benthics differ for 12 of 14 traits, including
all male mating, body size and ecological traits, and two
female mating traits (Table 2).
The extensive divergence between sympatric limnetics and
benthics is fully in parallel for size, color, and ecological
traits (Table 3; Figs. 2, 3) and partly in parallel for behavioral
traits (Figs. 4, 5). Although divergence in behavioral traits
is more variable among lakes, our results provide no evidence
that variation in these traits significantly influence sexual
isolation between ecotypes. Thus, the available evidence sug-
gests that divergence in behavioral traits does not contribute
to sexual isolation. The two traits that do contribute to sexual
isolation between ecotypes—size and color—show strong
parallel divergence in sympatry.
Our primary aim was to understand the nature of selection
acting on the traits that confer sexual isolation in threespine
stickleback species pairs. In particular, we were interested in
PARALLEL SEXUAL ISOLATION
netic sticklebacks from Paxton, Priest, Emily, and Enos Lakes. (A)
male length, (B) female length, and (C) length difference between
male and female. Positive differences indicate larger males and
negative differences indicate larger females. Lines connect lim-
netics and benthics from the same lake.
Mean body size (? standard errors) for benthic and lim-
limnetic sticklebacks from Paxton, Priest, Emily, Enos, and Hadley
Lakes: (A) number of gill rakers, and (B) number of armor plates.
Lines connect limnetics and benthics from the same lake.
Ecological trait means (? standard errors) for benthic and
the relative roles of ecological and nonecological selection
in the evolution of sexual isolation. Two lines of evidence
support our first hypothesis: that the ultimate cause of sexual
isolation is adaptation to environment. First, all three pairs
have independently evolved to recognize mates based on dif-
ferences in body size, and all limnetics use nuptial color.
Second, both of these mating traits have diverged between
benthics and limnetics in parallel, as have all size and eco-
logical traits. The eight behavioral traits that we examined
show more haphazard evolutionary divergence and do not
appear to contribute to sexual isolation in sympatry. We fur-
ther explore these lines of evidence and their implications
Parallel Sexual Isolation
Despite measuring a number of behavioral traits that form
key components of stickleback courtship, sexual isolation in
all three species pairs depended on the same two morpho-
logical traits: assortative mating based on differences in body
size and asymmetric isolation based on male nuptial color.
This confirms that sexual isolation has evolved in parallel.
All benthics use the same traits as do all limnetics; however,
limnetics and benthics base sexual isolation to a different
extent on these two traits. Benthics base sexual isolation
primarily on differences in body size, whereas limnetics base
sexual isolation on a combination of differences in body size
and nuptial color.
Females of both ecotypes spawn most readily with bright
heterospecific males, leading to asymmetric sexual isolation
based on color. Previous work has shown that there is evo-
lutionary change in the strength but not direction of female
color preferences: female preference for colorful males is
strong in limnetics but weak in benthics (Boughman 2001).
The opposite effect of color differences on sexual isolation
for limnetics and benthics is consistent with expectations
from open-ended preference functions, where sexualisolation
is predicted to be asymmetric (Lande 1981; Turelli et al.
2001). The ecotype with strong female preferences and ex-
aggerated male traits (limnetics) is expected to have stronger
reproductive isolation because females will reject hetero-
specific males with low trait values. However, females of the
JANETTE W. BOUGHMAN ET AL.
and limnetic sticklebacks from Paxton, Priest, Emily, and Enos
Lakes. Traits include: (A) male nuptial color score, (B) number of
bites per minute, (C) number of zig-zags per minute, (D) courtship
score, and (E) number of times male led female toward the nest
per minute. Lines connect limnetics and benthics from the same
Male mating trait means (? standard errors) for benthic
and limnetic sticklebacks from Paxton, Priest, Emily, and Enos
Lakes. Traits include: (A) number of times female followed male
toward the nest per min, (B) number of times female examined the
nest per min, (C) female responsiveness, and (D) preference for red
nuptial color. Lines connect limnetics and benthics from the same
Female mating trait means (? standard errors) for benthic
ecotype whose males have low trait values (benthics) are
expected to accept heterospecific mates with high trait values,
allowing gene flow. This asymmetry means that color dif-
ferences alone could not restrict gene flow. Sexual isolation
requires a combination of differences in body size and color.
Asymmetric sexual isolation based on differences in male
nuptial color is likely the extension of within-population mat-
ing patterns, and ecologically based sexual selection is almost
certainly involved. In contrast, at present it is unclear if size
based sexual isolation is an extension of within-population
mating behavior, because we found no evidence for size pref-
erences or size assortative mating within ecotype. Thus, sex-
ual selection on body size within species does not appear to
give rise to sexual isolation between species. Additional ex-
periments need to be done to test directly for such prefer-
ences, especially given that the range of sizes within a species
is not as great as that between species so it might be more
difficult to detect. However, if no such within population size
preferences or assortative mating are found, this raises the
possibility that size assortative mating between ecotypes
evolved as a consequence of secondary contact. This could
occur through a number of mechanisms, including reinforce-
ment or direct selection on mate preferences (Servedio 2000,
2001, 2004). Evidence for both processes has been found for
sticklebacks (Rundle and Schluter 1998; Albert and Schluter
Several studies have found differences in courtship be-
havior between stickleback populations, as we did here
(Ridgway and McPhail 1984; Foster and Baker 1995; Hay
and McPhail 2000; Ishikawa and Mori 2000). For example,
Ridgway and McPhail (1984) found strong reproductive iso-
lation between Enos Lake limnetics and benthics and de-
scribed courtship differences, but they did not test that it was
courtship that caused the reproductive isolation. In their study
males courted conspecific and heterospecific females differ-
ently. Others have shown that males adjust courtship in part
based on female body size, especially when females are much
larger (Nagel and Schluter 1998; Rundle and Schluter 1998).
Ridgway and McPhail (1984) did not consider size or color
in their trials. If courtship behavior depends on size traits,
then including both variables in a statistical model estimates
their independent effect on spawning. Our analyses use this
approach and find that differences in courtship behavior do
not have an effect on spawning that is independent of the
size effect. We did not manipulate male traits, but rather
PARALLEL SEXUAL ISOLATION
selected males haphazardly with respect to size and color,
endeavoring to use males from all parts of the size and color
distributions for each population. Thus, we controlled statis-
tically but not experimentally for correlated effects. Future
work could use experimental manipulations of color, size,
and courtship behavior to investigate their independent ef-
fects on sexual isolation. Future work could also explore the
possibility that some traits we did not measure, such as body
shape or chemical cues, may contribute to sexual isolation.
Traits Conferring Sexual Isolation
Variation in the two morphological traits we measured
function in sexual isolation, whereas no behavioral traits do
so. Is this likely to be a common pattern? Are mating traits
that confer reproductive isolation more likely to be morpho-
logical than behavioral? Certainly differences in body size
and color have been found to confer reproductive isolation
in several taxa. For example, differences in color pattern
isolate mimetic Heliconius butterflies (Jiggins et al. 2001;
Naisbit et al. 2001), Colias butterflies (Ellers and Boggs
2003), Ficedula flycatchers (Saetre et al. 1997; Dale et al.
1999), and many haplochromine cichlid species (e.g., See-
hausen et al. 1997; Couldridge and Alexander 2002; Allender
et al. 2003). In contrast to our findings, in most of these
systems color contributes to sexual isolation in a symmetric
way. Differences in body size also isolate many taxa, in-
cluding sympatric cichlids (Schliewen et al. 2001), skinks
(Richmond and Reeder 2002), Darwin’s finches (Ratcliffe
and Grant 1983), lake whitefish (Lu and Bernatchez 1999),
and lake and anadromous forms of sockeye salmon (Foote
and Larkin 1988). Body size contributes asymmetrically in
sockeye because males of the smaller lake form steal spawn-
ings of the larger male anadromous sockeye (Wood and Foote
What about behavioral traits? Much of the evidence that
behavioral traits confer sexual isolation comes from studies
of acoustic signals, including anurans (Ryan and Wilczynski
1991; Pfennig 2000; Hobel and Gerhardt 2003), crickets(Otte
1992; Gray and Cade 2000; Shaw and Lugo 2001), Dro-
sophila (Gleason and Ritchie 1998), and Crysoperla lace-
wings (Wells and Henry 1998). However, stickleback court-
ship is movement based. Despite commonly found differ-
ences among populations in courtship behavior in many taxa,
we have little data to suggest that movement behavior un-
derlies sexual isolation. Spiders do provide some examples
(Stratton and Uetz 1986; Miller et al. 1998; Hebets and Uetz
2000). The paltry number of examples for movement could
reflect a bias in what is studied. Perhaps movement traits are
more difficult to quantify and study than acoustic or mor-
phological traits, so fewer studies do so. But if this pattern
is real, why should some kinds of traits more commonly
confer sexual isolation? Part of the answer might lie in the
nature of selection acting on those traits.
The Nature of Selection on Mating Traits
Traits under ecological selection may be especially likely
to contribute to sexual isolation for two reasons. First, spe-
ciation should be facilitated when reproductive isolation is
based on ecological traits that differentiate species (e.g.,
Dieckmann and Doebeli 1999; Servedio 2000, 2004) such as
host choice in phytophagous insects that also mate on their
host (Funk 1998; Dres and Mallet 2002; Nosil et al. 2002,
2003). Second, speciation should be facilitated when eco-
logical traits and mating traits diverge in concert (Schluter
2000; Turelli et al. 2001). This is especially likely when
mating traits are under ecologically dependent selection and
habitats differ. Body size fits the first scenario and nuptial
color the second. Courtship behavior fits neither.
Body size affects exploitation of the different niches that
limnetics and benthics inhabit (Bentzen and McPhail 1984;
Schluter 1993). Therefore, natural selection arising from re-
source use and competition for those resources contributes
to divergence in body size (Schluter and McPhail 1992;
Schluter 1994). We find here that this adaptive divergence
in size is an essential component of sexual isolation. Assor-
tative mating on body size has also been found in anadro-
mous-stream pairs of sticklebacks (McKinnon et al. 2004),
who also show strong parallel evolution of body size. Thus,
body size appears to be a trait under strong ecological se-
lection that causes sexual isolation between species in both
limnetic-benthic pairs and anadromous-stream pairs.
Differences in nuptial color and color preference are due,
in part, to ecologically mediated sexual selection arising from
differences in light environment (Boughman 2001), because
the environment determines effective signaling colors, color
perception, and preference (Endler 1992; Fleishman et al.
1997; Fuller 2002; Leal and Fleishman 2002). Given that
limnetics and benthics mate in different light environments,
both color and color preference have diverged between them.
Thus, sexual isolation based on differences in color appears
to be a case of ecological and mating traits evolving in cor-
relation with the same aspect of environment.
Ecologically Based Selection and Genetic Constraints
An alternative to the hypotheses we test here is that parallel
evolution of sexual isolation results because closely related
populations share biases in both their standing genetic var-
iation and their production of new, heritable variation (Hal-
dane 1932; reviewed in Schluter et al. 2004). Such genetic
biases may influence evolutionary change as populations col-
onize new habitats. For example, some traits may have more
genetic variance than others, and evolution is expected to
proceed most rapidly for traits with the greatest genetic var-
iance. However, these biases should result in parallel or di-
vergent evolution among replicate populations only to the
extent that selection acts in a parallel or divergent manner
on this variance. Although shared genetic biases may cause
nonecological processes to produce similar trait changes in
independent populations, these changes will not be correlated
with environment. Parallel evolution in correlation with en-
vironment is only expected when selection is ecologically
Substantial data demonstrate that competition between
ecotypes and divergent natural selection causes divergence
in ecological traits and reproductive isolation (Schluter and
McPhail 1992, 1993; Schluter 1993, 1994, 1995, 2003; Hat-
field and Schluter 1999; Rundle et al. 2000; Pritchard and
Schluter 2001). We find here that limnetics and benthics have
JANETTE W. BOUGHMAN ET AL.
undergone parallel divergence in ecological traits and body
size, which suggests not only that each ecotype adapts to a
particular niche, but that the magnitude of difference between
them is important. This would occur if competition drives
the ecotypes apart, and a constant magnitude of difference
decreases the intensity of competition to the point where
evolutionary change halts (Pacala and Roughgarden 1985;
Pritchard and Schluter 2001; Gray and Robinson 2002; Pfen-
nig and Murphy 2003). Our data bolster the already strong
case for ecological selection.
Sexual Selection and Speciation: Arbitrary or Ecological?
We found differences among lakes in absolute trait values
for size and color, which could be due to differences in local
ecology or to sexual selection acting independently of en-
vironment. We can rule out sexual selection in arbitrary di-
rections by comparing ecological and mating traits. Sexual
selection is expected to affect only mating traits, but instead
we find that ecological traits and mating traits show similar
patterns of divergence. Thus, adaptation to locally unique
environments is more likely.
However, our results suggest that sexual selection acts on
behavioral traits at least partly independently of environment.
Thus, sexual selection in arbitrary directions may be oper-
ating in our populations. Yet, despite extensive evolutionary
change in courtship behavior, behavioral traits do not appear
to confer sexual isolation. Therefore, this arbitrary diver-
gence arising from interactions between the sexes does not
appear to play a major role in speciation for stickleback pairs.
This is a surprise. Several models of sexual selection and
speciation suggest that it is just this haphazard evolutionary
change that enables reproductive isolation to evolve. In fact,
this is the traditional view of how sexual selection is likely
to lead to sexual isolation. Further work will be required to
test this definitively, but the present results provide no evi-
dence that this mode of speciation by sexual selection op-
erates for stickleback pairs.
At present, it is unclear if this is an unusual or common
outcome. Data from some systems supports the arbitrary
mode of sexual selection and speciation. Examples include:
egg-sperm recognition in broadcast spawners (e.g., Palumbi
1999; Swanson and Vacquier 2002), male-female reproduc-
tive tract coupling (e.g., Eberhard 1992; Presgraves et al.
1999), and conspecific sperm precedence (e.g., Price 1997;
Brown and Eady 2001; Dixon et al. 2003). However, these
examples are primarily for fertilization traits. There appear
to be relatively few examples of arbitrary divergence in pre-
mating, preinsemination traits with concomitant evidence
that such divergence plays a critical role in reproductive iso-
lation. This is, however, an area of very active research and
more data become available all the time.
Evidence is accumulating that environmentally dependent
sexual selection contributes to divergence in mating traits
and sexual isolation. A number of examples invoke sensory
drive, where mating signals adapt to local environments to
enhance their transmission (Ryan and Wilczynski 1991; Mar-
chetti 1993; Endler and Thery 1996; Uy and Borgia 2000;
Boughman 2002; Leal and Fleishman 2002; McNaught and
Owens 2002), or perception adapts to local environments and
results in altered preferences (Boughman 2001; Smith et al.
2004). Changes in light environment have even led to the
collapse of premating isolation in some Lake Victoria cich-
lids (Seehausen et al. 1997). Premating isolation is based
heavily on conspicuous color differences between species in
both Lake Victoria (Seehausen et al. 1998; Couldridge and
Alexander 2002) and Malawi cichlids (Allender et al. 2003;
Knight and Turner 2004). It is increasingly recognized that
many forms of sexual selection may depend, at least in part,
on the environment. This recognition lies at the heart of the
view we test here that ecologically dependent selectioncauses
divergence in mating traits. Ecological differences among
populations are likely to be ubiquitous, so investigating this
mode of speciation is likely to prove fruitful.
Many other mating traits that confer sexual isolation appear
to be by-products of adaptation, as is the case for body size
in sticklebacks. These include Darwin’s finches, where beak
shape is an adaptation for exploiting seeds and also affects
vocal tract resonance properties, causing differences among
species in song that contribute to sexual isolation (Podos
2001). There are likely to be many examples for host races
of phytophagous insects (Funk et al. 2002). Even cuticular
hydrocarbon differences between populations of Drosophila
have adaptive function, as shown elegantly by Greenberg et
al. (2003) and others (Markow and Toolson 1990; Blows
2002). Such differences cause sexual isolation between sev-
eral Drosophila species (e.g., Coyne 1996; Higgie et al. 2000;
Etges and Ahrens 2001). In all these cases, sexual selection
does not act on sexual isolation alone, although it could con-
tribute to trait divergence in concert with natural selection.
In conclusion, we found evidence that ecologically based
sexual and natural selection have caused sexual isolation in
species pairs of threespine sticklebacks. The basis for sexual
isolation is similar across multiple species pairs, and the traits
that confer reproductive isolation have evolved in parallel.
Both findings invoke a strong role for ecologically based
divergent selection and provide additional data in support of
the hypothesis that the ultimate cause of speciation in this
system is adaptation to environment. Whether this holds true
for other systems remains to be determined.
Thanks to S. Morgan, B. Harvey, K. Rosalska, A. Agrawal,
and M. McDonald for help with data collection. Comments
by U. Candolin, J. Mallet, K. Shaw, and three anonymous
reviewers improved the paper. JWB was supported by the
Natural Science Foundation and the National Science and
Engineering Research Council of Canada; HDR by Killam
(University of British Columbia), Eastburn (Hamilton Com-
munity Foundation), and Australian Research Council Fel-
lowships; and DS by the Natural Science and Engineering
Research Council of Canada.
Albert, A. Y. K., and D. Schluter. 2004. Reproductive character
displacement of male stickleback mate preference: reinforce-
ment or direct selection? Evolution 58:1099–1107.
Allender, C. J., O. Seehausen, M. E. Knight, G. F. Turner, and N.
Maclean. 2003. Divergent selection during speciation of Lake
PARALLEL SEXUAL ISOLATION
Malawi cichlid fishes inferred from parallel radiations in nuptial
coloration. Proc. Natl. Acad. Sci. USA 100:14074–14079.
Arnqvist, G., and L. Rowe. 2002. Antagonistic coevolution between
the sexes in a group of insects. Nature 415:787–789.
Arnqvist, G., M. Edvardsson, U. Friberg, and T. Nilsson. 2000.
Sexual conflict promotes speciation in insects. Proc. Natl. Acad.
Sci. USA 97:10460–10464.
Bell, M. A., G. Orti, J. A. Walker, and J. P. Koenings. 1993. Evo-
lution of pelvic reduction in threespine stickleback fish: a test
of competing hypotheses. Evolution 47:906–914.
Bentzen, P., and J. D. McPhail. 1984. Ecology and evolution of
sympatric sticklebacks (Gasterosteus): specialization for alter-
native trophic niches in the Enos Lake species pair. Can. J. Zool.
Blows, M. W. 2002. Interaction between natural and sexual selec-
tion during the evolution of mate recognition. Proc. R. Soc.
Lond. B 269:1113–1118.
Boughman, J. W. 2001. Divergent sexual selection enhances re-
productive isolation in sticklebacks. Nature 411:944–948.
———. 2002. How sensory drive can promote speciation. Trends
Ecol. Evol. 17:571–577.
Brown, D. V., and P. E. Eady. 2001. Functional incompatibility
between the fertilization systems of two allopatric populations
of Callosobruchus maculatus (Coleoptera: Bruchidae). Evolution
Couldridge, V. C. K., and G. J. Alexander. 2002. Color patterns
and species recognition in four closely related species of Lake
Malawi cichlid. Behav. Ecol. 13:59–64.
Coyne, J. A. 1996. Genetics of sexual isolation in male hybrids of
Drosophila simulans and D. mauritiana. Genet. Res. 68:211–220.
Dale, S., T. Slagsvold, H. M. Lampe, and G. P. Saetre. 1999. Pop-
ulation divergence in sexual ornaments: the white forehead patch
of Norwegian pied flycatchers is small and unsexy. Evolution
Day, T. 2000. Sexual selection and the evolution of costly female
preferences: spatial effects. Evolution 54:715–730.
Dieckmann, U., and M. Doebeli. 1999. On the origin of species by
sympatric speciation. Nature 400:354–357.
Dixon, S. M., J. A. Coyne, and M. A. F. Noor. 2003. The evolution
of conspecific sperm precedence in Drosophila. Mol. Ecol. 12:
Dobzhansky, T. 1951. Genetics and the origin of species. Columbia
Univ. Press, New York.
Dres, M., and J. Mallet. 2002. Host races in plant-feeding insects
and their importance in sympatric speciation. Philos. Trans. R.
Soc. B 357:471–492.
Eberhard, W. G. 1992. Species isolation, genital mechanics, and
the evolution of species-specific genitalia in three species of
Macrodactylus beetles (Coleoptera, Scarabeidae, Melolonthi-
nae). Evolution 46:1774–1783.
Efron, B. 1982. The jackknife, the bootstrap, and other resampling
plans. Society for Industrial and Applied Mathematics, Phila-
Ellers, J., and C. L. Boggs. 2003. The evolution of wing color: male
mate choice opposes adaptive wing color divergence in Colias
butterflies. Evolution 57:1100–1106.
Endler, J. A. 1992. Signals, signal conditions, and the direction of
evolution. Am. Nat. 139:125–153.
———. 1993. Some general comments on the evolution and design
of animal communication systems. Philos. Trans. R. Soc. B 340:
Endler, J. A., and M. Thery. 1996. Interacting effects of lek place-
ment, display behavior, ambient light, and color patterns in three
Neotropical forest-dwelling birds. Am. Nat. 148:421–452.
Etges, W. J., and M. A. Ahrens. 2001. Premating isolation is de-
termined by larval rearing substrates in cactophilic Drosophila
mojavensis. V. Deep geographic variation in epicuticular hydro-
carbons among isolated populations. Am. Nat. 158:585–598.
Fleishman, L. J., M. Bowman, D. Saunders, W. E. Miller, M. J.
Rury, and E. R. Loew. 1997. The visual ecology of Puerto Rican
anoline lizards: habitat light and spectral sensitivity. J. Comp.
Physiol. A 181:446–460.
Foster, S. A., and J. A. Baker. 1995. Evolutionary interplay between
ecology, morphology and reproductive-behavior in threespine
stickleback, Gasterosteus aculeatus. Environ. Biol. Fish. 44:
Foote, C. J., and P. A. Larkin. 1988. The role of male choice in
the assortative mating of anadromous and non-anadromous sock-
eye salmon (Oncorhynchus nerka). Behaviour 106:43–62.
Fuller, R. C. 2002. Lighting environment predicts the relative abun-
dance of male colour morphs in bluefin killifish (Lucania goodei)
populations. Proc. R. Soc. Lond. B 269:1457–1465.
Funk, D. J. 1998. Isolating a role for natural selection in speciation:
host adaptation and sexual isolation in Neochlamisus bebbianae
leaf beetles. Evolution 52:1744–1759.
Funk, D. J., K. E. Filchak, and J. L. Feder. 2002. Herbivorous
insects: model systems for the comparative study of speciation
ecology. Genetica 116:251–267.
Gavrilets, S. 2000. Rapid evolution of reproductive barriers driven
by sexual conflict. Nature 403:886–889.
Gavrilets, S., and C. R. B. Boake. 1998. On the evolution of pre-
mating isolation after a founder event. Am. Nat. 152:706–716.
Gleason, J. M., and M. G. Ritchie. 1998. Evolution of courtship
song and reproductive isolation in the Drosophila willistoni spe-
cies complex: Do sexual signals diverge the most quickly? Evo-
Gray, D. A., and W. H. Cade. 2000. Sexual selection and speciation
in field crickets. Proc. Natl. Acad. Sci. USA 97:14449–14454.
Gray, S. M., and B. W. Robinson. 2002. Experimental evidence
that competition between stickleback species favours adaptive
character divergence. Ecol. Lett. 5:264–272.
Greenberg, A. J., J. R. Moran, J. A. Coyne, and C. I. Wu. 2003.
Ecological adaptation during incipient speciation revealed by
precise gene replacement. Science 302:1754–1757.
Haldane, J. B. S. 1932. The causes of evolution. Longman, London.
Hatfield, T., and D. Schluter. 1999. Ecological speciation in stick-
lebacks: environment-dependent hybrid fitness. Evolution 53:
Hay, D. E., and J. D. McPhail. 2000. Courtship behaviour of male
threespine sticklebacks (Gasterosteus aculeatus) from old and
new hybrid zones. Behaviour 137:1047–1063.
Hebets, E. A., and G. W. Uetz. 2000. Leg ornamentation and the
efficacy of courtship display in four species of wolf spider (Ar-
aneae: Lycosidae). Behav. Ecol. Sociobiol. 47:280–286.
Higashi, M., G. Takimoto, and N. Yamamura. 1999. Sympatric
speciation by sexual selection. Nature 402:523–526.
Higgie, M., S. F. Chenoweth, and M. W. Blows. 2000. Natural
selection and the reinforcement of mate recognition. Science 20:
Hobel, G., and H. C. Gerhardt. 2003. Reproductive character dis-
placement in the acoustic communication system of green tree
frogs (Hyla cinerea). Evolution 57:894–904.
Ishikawa, M., and S. Mori. 2000. Mating success and male courtship
behaviors in three populations of the threespine stickleback. Be-
Jiggins, C. D., R. E. Naisbit, R. L. Coe, and J. Mallet. 2001. Re-
productive isolation caused by colour pattern mimicry. Nature
Knight, M. E., and G. F. Turner. 2004. Laboratory mating trials
indicate incipient speciation by sexual selection among popu-
lations of the cichlid fish Pseudotropheus zebra from Lake Ma-
lawi. Proc. R. Soc. Lond. B 271:675–680.
Lande, R. 1981. Models of speciation by sexual selection on poly-
genic traits. Proc. Natl. Acad. Sci. USA 78:3721–3725.
———. 1982. Rapid origin of sexual isolation and character di-
vergence in a cline. Evolution 36:213–223.
Leal, M., and L. J. Fleishman. 2002. Evidence for habitat partioning
based on adaptation to environmental light in a pair of sympatric
lizard species. Proc. R. Soc. Lond. B 269:351–359.
Lu, G. Q., and L. Bernatchez. 1999. Correlated trophic speciali-
zation and genetic divergence in sympatric lake whitefish eco-
types (Coregonus clupeaformis): support for the ecological spe-
ciation hypothesis. Evolution 53:1491–1505.
Marchetti, K. 1993. Dark habitats and bright birds illustrate the role
of the environment in species divergence. Nature 362:149–152.
Markow, T. A., and E. C. Toolson. 1990. Temperature effects on
JANETTE W. BOUGHMAN ET AL.
epicuticular hydrocarbons and sexual isolation in Drosophila
mojavensis. Pp. 315–331 in J. S. F. Barker, W. T. Starmer, and
R. J. MacIntyre, eds. Ecological and evolutionary genetics of
Drosophila. Plenum, New York.
Mayr, E. 1942. Systematics and the origin of species. Columbia
Univ. Press, New York.
McKinnon, J., S. Mori, B. K. Blackman, L. David, D. M. Kingsley,
L. Jamieson, J. Chou, and D. Schluter. 2004. Evidence for ecol-
ogy’s role in speciation. Nature 429:294–298.
McLennan, D. A., and J. D. McPhail. 1990. Experimental inves-
tigations of the evolutionary significance of sexually dimorphic
nuptial coloration in Gasterosteus aculeatus (L.): The relation-
ship between male color and female behavior. Can. J. Zool. 68:
McNaught, M. K., and I. P. F. Owens. 2002. Interspecific variation
in plumage colour among birds: species recognition or light
environment? J. Evol. Biol. 15:505–514.
McPhail, J. D. 1984. Ecology and evolution of sympatric stickle-
backs (Gasterosteus): morphological and genetic evidence for a
species pair in Enos Lake, British Columbia. Can. J. Zool. 62:
———. 1992. Ecology and evolution of sympatric sticklebacks
(Gasterosteus): evidence for a species-pair in Paxton Lake, Tex-
ada Island, British Columbia. Can. J. Zool. 70:361–369.
———. 1993. Ecology and evolution of sympatric sticklebacks
(Gasterosteus): origin of the species pairs. Can. J. Zool. 71:
———. 1994. Speciation and the evolution of reproductive isolation
in the sticklebacks (Gasterosteus) of south-western British Co-
lumbia. Pp. 399–437 in M. A. Bell and S. A. Foster, eds. The
evolutionary biology of the threespine stickleback. Oxford Univ.
Press, Oxford, U.K.
Miller, G. L., G. E. Stratton, P. R. Miller, and E. Hebets. 1998.
Geographical variation in male courtship behaviour and sexual
isolation in wolf spiders of the genus Schizocosa. Anim. Behav.
Nagel, L. M., and D. Schluter. 1998. Body size, natural selection,
and speciation in sticklebacks. Evolution 52:209–218.
Naisbit, R. E., C. D. Jiggins, and J. Mallet. 2001. Disruptive sexual
selection against hybrids contributes to speciation between Hel-
iconius cydno and Heliconius melpomene. Proc. R. Soc. Lond.
Nosil, P., B. J. Crespi, and C. P. Sandoval. 2002. Host-plant ad-
aptation drives the parallel evolution of reproductive isolation.
———. 2003. Reproductive isolation driven by the combined ef-
fects of ecological adaptation and reinforcement. Proc. R. Soc.
Lond. B 270:1911–1918.
Otte, D. 1992. Evolution of cricket songs. J Orthoptera Res 1:25–49.
Pacala, S., and J. Roughgarden. 1985. Population experiments with
the Anolis lizards of St. Maarten and St. Eustatius. Ecology 66:
Palumbi, S. R. 1999. All males are not created equal: fertility dif-
ferences depend on gamete recognition polymorphisms in sea
urchins. Proc. Natl. Acad. Sci. USA 96:12632–12637.
Panhuis, T. M., R. K. Butlin, M. Zuk, and T. Tregenza. 2001. Sexual
selection and speciation. Trends Ecol. Evol. 16:364–371.
Parker, G. A., and L. Partridge. 1998. Sexual conflict and speciation.
Philos. Trans. R. Soc. Lond. B 353:261–274.
Pfennig, D. W., and P. J. Murphy. 2003. A test of alternative hy-
potheses for character divergence between coexisting species.
Pfennig, K. S. 2000. Female spadefoot toads compromise on mate
quality to ensure conspecific matings. Behav. Ecol. 11:220–227.
Podos, J. 2001. Correlated evolution of morphology and vocal sig-
nal structure in Darwin’s finches. Nature 409:185–188.
Pomiankowski, A., and Y. Iwasa. 1998. Runaway ornament diver-
sity caused by Fisherian sexual selection. Proc. Natl. Acad. Sci.
Presgraves, D. C., R. H. Baker, and G. S. Wilkinson. 1999. Co-
evolution of sperm and female reproductive tract morphology
in stalk-eyed flies. Proc. R. Soc. Lond. B 266:1041–1047.
Price, C. S. C. 1997. Conspecific sperm precedence in Drosophila.
Pritchard, J. R., and D. Schluter. 2001. Declining interspecific com-
petition during character displacement: summoning the ghost of
competition past. Evol. Ecol. Res. 3:209–220.
Ratcliffe, L. M., and P. R. Grant. 1983. Species recognition in
Darwins finches (Geospiza, Gould). 1. Discrimination by mor-
phological cues. Anim. Behav. 31:1139–1153.
Reimchen, T. E. 1994. Predators and morphological evolution in
threespine stickleback. Pp. 240–276 in M. A. Bell and S. A.
Foster, eds. The evolutionary biology of the threespine stick-
leback. Oxford Univ. Press, Oxford, U.K.
Rice, W. R., and A. K. Chippindale. 2002. The evolution of hybrid
infertility: perpetual coevolution between gender-specific and
sexually antagonistic genes. Genetica 116:179–188.
Richmond, J. Q., and T. W. Reeder. 2002. Evidence for parallel
ecological speciation in scincid lizards of the Eumeces skilton-
ianus species group (Squamata: Scincidae). Evolution 56:
Ridgway, M. S., and J. D. McPhail. 1984. Ecology and evolution
of sympatric sticklebacks (Gasterosteus): mate choice and re-
productive isolation in the Enos Lake species pair. Can. J. Zool.
Rowland, W. J. 1989. The ethological basis of mate choice in male
threespine sticklebacks, Gasterosteus aculeatus. Anim. Behav.
Rundle, H. D., and D. Schluter. 1998. Reinforcement of stickleback
mate preferences: sympatry breeds contempt. Evolution 52:
———. 2004. Natural selection and speciation in sticklebacks. Pp.
1–21 in U. Dieckmann, H. Metz, M. Doebeli, and D. Tautz, eds.
Adaptive speciation. Cambridge Univ. Press, Cambridge, U.K.
Rundle, H. D., L. M. Nagel, J. W. Boughman, and D. Schluter.
2000. Natural selection and parallel speciation in sympatric
sticklebacks. Science 287:306–308.
Ryan, M. J., and A. S. Rand. 1993. Species recognition and sexual
selection as a unitary problem in animal communication. Evo-
Ryan, M. J., and W. Wilczynski. 1991. Evolution of intraspecific
variation in the advertisement call of a cricket frog (Acris cre-
pitans, Hylidae). Biol. J. Linn. Soc. 44:249–271.
Saetre, G.-P., S. G. Weller, W. L. Wagern, P. S. Soltis, and D. E.
Soltis. 1997. A sexually selected character displacement in fly-
catchers reinforces premating isolation. Nature 387:589–592.
Schemske, D. W., and H. D. Bradshaw. 1999. Pollinator preference
and the evolution of floral traits in monkeyflowers (Mimulus).
Proc. Natl. Acad. Sci. USA 96:11910–11915.
Schliewen, U., K. Rassmann, M. Markmann, J. A. Markert, T. D.
Kocher, and D. Tautz. 2001. Genetic and ecological divergence
of a monophyletic cichlid species pair under fully sympatric
conditions in Lake Ejagham, Cameroon. Mol. Ecol. 10:
Schluter, D. 1993. Adaptive radiation in sticklebacks: size, shape,
and habitat use efficiency. Ecology 74:699–709.
———. 1994. Experimental evidence that competition promotes
divergence in adaptive radiation. Science 266:798–801.
———. 1995. Adaptive radiation in sticklebacks: trade-offs in feed-
ing performance and growth. Ecology 76:82–90.
———. 1996. Adaptive radiation along genetic lines of least re-
sistance. Evolution 50:1766–1774.
———. 2000. The ecology of adaptive radiation. Oxford Univ.
Press, Oxford, U.K.
———. 2001. Ecology and the origin of species. Trends Ecol. Evol.
———. 2003. Frequency dependent natural selection during char-
acter displacement in sticklebacks. Evolution 57:1142–1150.
Schluter, D., and J. D. McPhail. 1992. Ecological character dis-
placement and speciation in sticklebacks. Am. Nat. 140:85–108.
———. 1993. Character displacement and replicate adaptive ra-
diation. Trends Ecol. Evol. 8:197–200.
Schluter, D., and L. M. Nagel. 1995. Natural selection and parallel
speciation. Am. Nat. 146:292–301.
Schluter, D., E. A. Clifford, M. Nemethy, and J. S. McKinnon.
PARALLEL SEXUAL ISOLATION
2004. Parallel evolution and inheritance of quantitative traits.
Am. Nat. 163:809–822.
Seehausen, O., J. J. M. van Alphen, and F. Witte. 1997. Cichlid
fish diversity threatened by eutrophication that curbs sexual se-
lection. Science 277:1808–1811.
Seehausen, O., F. Witte, J. J. M. van Alphen, and N. Bouton. 1998.
Direct mate choice maintains diversity among sympatric cichlids
in Lake Victoria. J. Fish Biol. 53:37–55.
Servedio, M. R. 2000. Reinforcement and the genetics of nonran-
dom mating. Evolution 54:21–29.
———. 2001. Beyond reinforcement: the evolution of premating
isolation by direct selection on preferences and postmating, pre-
zygotic incompatibilities. Evolution 55:1909–1920.
———. 2004. The evolution of premating isolation: local adapta-
tion and natural and sexual selection against hybrids. Evolution
Shaw, K. L., and E. Lugo. 2001. Mating asymmetry and the direc-
tion of evolution in the Hawaiian cricket genus Laupala. Mol.
Smith, C., I. Barber, R. J. Wootton, and L. Chittka. 2004. A receiver
bias in the origin of three-spined stickleback mate choice. Proc.
R. Soc. Lond. B 271:949–955.
Sokal, R. R., and F. J. Rohlf. 1995. Biometry: the principles and
practice of statistics in biological research. Freeman, New York.
Stratton, G. E., and G. W. Uetz. 1986. The inheritance of courtship
behavior and its role as a reproductive isolating mechanism in
two species of Schizocosa wolf spiders (Araneae: Lycosidae).
Swanson, W. J., and V. D. Vacquier. 2002. Reproductive protein
evolution. Annu. Rev. Ecol. Syst. 33:161–179.
Taylor, E. B., and J. D. McPhail. 1999. Evolutionary history of an
adaptive radiation in species pairs of threespine sticklebacks
(Gasterosteus): insights from mitochondrial DNA. Biol. J. Linn.
Taylor, E. B., and J. D. McPhail. 2000. Historical contingency and
ecological determinism interact to prime speciation in stickle-
backs, Gasterosteus. Proc. R. Soc. Lond. B 267:2375–2384.
Turelli, M., N. H. Barton, and J. A. Coyne. 2001. Theory and spe-
ciation. Trends Ecol. Evol. 16:330–343.
Turner, G. F., and M. T. Burrows. 1995. A model of sympatric
speciation by sexual selection. Proc. R. Soc. Lond. B 260:
Uy, J. A. C., and G. Borgia. 2000. Sexual selection drives rapid
divergence in bowerbird display traits. Evolution 54:273–278.
Vamosi, S. M. 2002. Predation sharpens the adaptive peaks: survival
trade-offs in sympatric sticklebacks. Ann. Zool. Fenn. 39:
Via, S., A. C. Bouck, and S. Skillman. 2000. Reproductive isolation
between divergent races of pea aphids on two hosts. II. Selection
against migrants and hybrids in the parental environments. Evo-
Wells, M. L. M., and C. S. Henry. 1998. Songs, reproductive iso-
lation, and speciation in cryptic species of insects: a case study
using green lacewings. Pp. 217–233 in D. J. Howard and S. H.
Berlocher, eds. Endless forms: species and speciation. Oxford
Univ. Press, Oxford, U.K.
West-Eberhard, M. J. 1983. Sexual selection, social competition,
and speciation. Q. Rev. Biol. 58:155–183.
Wood, C. C., and C. J. Foote. 1996. Evidence for sympatric genetic
divergence of anadromous and nonanadromous morphs of sock-
eye salmon (Oncorhynchus nerka). Evolution 50:1265–1279.
Corresponding Editor: K. Shaw