Current Biology 18, 769–774, May 20, 2008 ª2008 Elsevier Ltd All rights reserved DOI 10.1016/j.cub.2008.04.027
Reverse Evolution of Armor Plates
in the Threespine Stickleback
Jun Kitano,1Daniel I. Bolnick,2David A. Beauchamp,3
Michael M. Mazur,3Seiichi Mori,4Takanori Nakano,5
and Catherine L. Peichel1,*
1Division of Human Biology
Fred Hutchinson Cancer Research Center
Seattle, Washington 98109
2Section of Integrative Biology
University of Texas
Austin, Texas 78712
3U.S. Geological Survey
Washington Cooperative Fish & Wildlife Research Unit
School of Aquatic and Fisheries Sciences
University of Washington
Seattle, Washington 98105
Ogaki, Gifu 503-8550, Japan
Research Institute for Humanity and Nature
Kamigyo-ku, Kyoto 602-0878, Japan
Faced with sudden environmental changes, animals must
of the mechanisms underlying rapid adaptation is crucial
not only for the understanding of natural evolutionary pro-
cesses but also for the understanding of human-induced
evolutionary change, which is an increasingly important
problem [1–8]. In the present study, we demonstrate that
the frequency of completely plated threespine stickleback
fish (Gasterosteus aculeatus) has increased in an urban
freshwater lake (Lake Washington, Seattle, Washington)
within the last 40 years. This is a dramatic example of
‘‘reverse evolution,’’  because the general evolutionary
trajectory is toward armor-plate reduction in freshwater
sticklebacks . On the basis of our genetic studies and
simulations, we propose that the most likely cause of re-
verse evolution is increased selection for the completely
plated morph, which we suggest could result from higher
levels of trout predation after a sudden increase in water
transparency during the early 1970s. Rapid evolution was
facilitated by the existence of standing allelic variation in
Ectodysplasin (Eda), the gene that underlies the major plate-
morph locus . The Lake Washington stickleback thus
provides a novel example of reverse evolution, which is
probably caused by a change in allele frequency at the major
plate locus in response to a changing predation regime.
Results and Discussion
Reverse Evolution of Armor Plates in Lake
The threespine stickleback (Gasterosteus aculeatus) pro-
vides a good model system for elucidation of the ecological
and genetic mechanisms underlying phenotypic evolution
[12, 13]. One dramatic and prevalent phenotypic change in
these fish is the reduction of armor plates, which cover the
lateral body surface, that occurred repeatedly after freshwa-
ter colonization 12,000 years ago . Whereas ancestral
marine sticklebacks typically have a continuous row of lat-
eral plates (completely plated morph), freshwater stickle-
backs usually have a reduction in lateral plates resulting in
a gap in the middle part of the plate row (partially plated
morph) or a loss of both the middle and posterior plates
(low-plated morph). The major gene responsible for reduc-
tion of the stickleback lateral plates across the world is
Ectodysplasin (Eda) . There are two major alleles of Eda
found in stickleback populations, and they are here referred
to as the complete allele and the low allele. Most marine
sticklebacks are homozygous for the complete allele, al-
though marine sticklebacks that are heterozygous carriers
of the low allele are found at a low frequency . It is pro-
posed that when marine sticklebacks colonize freshwater
environments, strong selection results in an increase in the
frequency of the low Eda allele, leading to the prevalence
of low-plated fish in freshwater.
In contrast to the prevalence of the low-plated morph in
many freshwater environments [10, 11], we found a high fre-
quency of completely plated sticklebacks in Lake Washing-
ton, an urban freshwater lake in Seattle [14–16]. In 2005, we
found that all three lateral-plate morphs were present, with
49% completely plated morphs, 35% partially plated morphs,
and 16% low-plated morphs (Figures 1A and 2C). Although
a previous study had also shown that all three morphs were
present in Lake Washington in 1968–1969, only 6% were
classified as completely plated morphs (Figure 1A) .
Instead, the low-plated morph, with a mode of seven plates,
was the most common morph until the late 1960s (Figures
1B and 1C). In 1976, bimodal peaks appeared, one corre-
sponding to fish with seven plates and another correspond-
ing to fish with 32 plates (Figure 1C) . The frequency of
fish with 33 plates was even higher in the 2005 sample
(Figure 1C). The increase in completely plated fish in the
2005 sample did not reflect bias in the sampling methods
(n = 322, c2= 6.6949, d.f. = 4, p = 0.1529) or in the seasonal
(Figure S1, available online) or geographical (Figure 2C) distri-
bution of differently plated sticklebacks. These data demon-
strate that the frequency of plate-morph phenotypes has
changed dramatically in Lake Washington within the past
40 years, which is equivalent to 40 generations in this stickle-
back population .
Genotyping of the 2005 samples at the Eda locus revealed
a strong association between plate phenotype and Eda geno-
(Figure S2, Table S1). By ANOVA, the Eda genotype explains
75.2% of the variance in plate number in the Lake Washington
stickleback. This is close to the percentage of phenotypic var-
iance in plate number explained by the Eda locus in laboratory
crosses (76.9%) . Thus, the increase in the completely
plated phenotype in Lake Washington is probably the result
of an increase in the frequency of the Eda complete allele,
given the previously established link between plate phenotype
and Eda genotype in stickleback populations across the
Gene Flow is Not the Primary Cause of Armor-Plate
Evolution in Lake Washington Sticklebacks
Most marine sticklebacks in Puget Sound are completely
plated (Figure 2C), with high frequencies of the complete Eda
allele (Figure 2D). Because marine sticklebacks can now mi-
grate into the lake through the Lake Washington Ship Canal
(Figure 2B and Figure S3), which was built in 1917 , an in-
crease in migration might have contributed to the increase of
lateral plates in the Lake Washington stickleback. In order to
test this hypothesis, we collected sticklebacks in neighboring
marine environments (Puget Sound), in multiple points in Lake
Washington, and in neighboring streams (Figure 2) and geno-
typed them with 15 microsatellite markers (Table S2). Genetic
STRUCTURE . Within the marine, lake, and stream fish that
were genotyped, the most probable number of genetic clus-
ters (K) was three (Figure S4). Estimation of ancestry for each
Figure 1. Lateral-Plate Evolution in the Lake Washington Stickleback
(A) Temporal change in the frequency of the completely plated (black bar), partially plated (gray bar), and low-plated (white bar) morphs in Lake Washington
sticklebacks. Sample sizes are shown above the graph. Right panels show representative images of the three stickleback morphs. Skeletal structures are
visualized by alizarin red staining. Scale bars represent 10 mm.
(B) Representative images of sticklebacks collected via midwater trawling during March 1957 and March 2006 in the northern pelagic zone of Lake
(C) Histograms of lateral-plate number for sticklebacks collected in 1957, 1968–1969, 1976, and 2005. Sample sizes are the same as those in Figure 1A. The
most-common plate number is also shown in each panel as a mode. Among sticklebacks collected in 1968–1969, 22% had more than 12 plates, but the
individual plate counts for each fish are not available . Plate number was counted from the left side of the fish except in the 1976 sample, for which
only right-side plate-number data were available . For the 1957 data, museum specimens in the University of Washington Fish Collection were analyzed.
The frequency of morph was significantly different between successive sampling time points (p < 0.05) except between 1957 and 1968–1969 samples
(c2= 2.375, p = 0.305).
Current Biology Vol 18 No 10
individual revealed that the sticklebacks in Lake Washington
have two main genetic sources (Figure 2E). Sticklebacks sam-
pled from areas near the ship canal were genetically similar
to marine sticklebacks (indicated with green in Figure 2E),
whereas those sampled from areas close to the streams
were more similar to neighboring stream sticklebacks
(indicated with blue in Figure 2E). However, there was no
significant correlation between probability of marine ancestry
and plate number (Pearson correlation r = 0.005, p = 0.967)
(Figure S5). Multidimensional scaling of the genetic-distance
matrix also confirmed the lack of association between geno-
types at neutral loci and plate number (Figure S6). Thus, the
increase in armor plates in Lake Washington sticklebacks
does not result simply from the presence of marine stickle-
backs in the lake.
It might be still possible that an increase in long-term
migration from Puget Sound has contributed to the overall in-
crease in the completely plated morph in the lake. To test this
possibility, we first estimated migration rates (m; fraction of
migrantsper generation)fromthegenetic dataandthen exam-
ined whether the empirically estimated m can explain the ob-
served plate evolution. We used both Isolation with Migration
(IM) and LAMARC software [21–23] to estimate the m between
a Lake Washington population and a Puget Sound marine
Figure 2. Genetic and Morphological Variation around Lake Washington
(A) Map of Washington State. Blue dots indicate the collection sites of marine stickleback from Puget Sound. Lake Washington is highlighted by a square
and magnified in Figure 2B.
(B) Map of Lake Washington and neighboring streams. Numbers indicate the sampling sites in Lake Washington: Point 1: Union Bay; Point 2: northern
east channel; Point 9: northern pelagic zone (Area 2 in ).
(C) Variation in plate-morph frequencies among populations. Each column indicates the frequency of the completely plated (black bar), partially plated (gray
bar), and low-plated (white bar) morphs for each stickleback population. Numbers in parentheses indicate sample size. The frequency of different morphs
was not significantly different among different points within Lake Washington (n = 322, c2= 17.0, d.f. = 14, p = 0.255).
(D) Variation in the allele frequency of Eda among populations. The black bar indicates the frequency of the complete Eda allele, and the white bar indicates
the frequency of the low Eda allele at Stn382. Numbers in parentheses indicate sample size.
(E) Genetic structure of sticklebacks collected in Lake Washington, Puget Sound and neighboring streams. The three different genetic clusters are shown in
different colors. Each individual is represented by a thin column that is partitioned into colored segments indicating the estimated proportion of ancestry
from each cluster. Numbers in parentheses indicate sample size.
Reverse Evolution of Armor in Stickleback
population (Table S3). The m of Puget Sound sticklebacks into
Lake Washington was estimated as 3.03 3 1024(IM) or 1.77 3
1023(LAMARC), whereas the m of Lake Washington stickle-
backs into Puget Sound was estimated as 6.43 3 1024(IM)
or 1.20 3 1023(LAMARC). Then, we developed deterministic
numerical simulations to calculate the m required for the
observed change of plate phenotype under different selection
regimes (Figure S7). In the absence of selection (s = 0), migra-
tion would need to be 0.148 to explain the observed change
from 1969 to 1976 and 0.035 to explain the observed change
from 1969 to 2005. These values are inconsistent with our
low (m < 1023) migration-rate estimates, suggesting that there
was a period of selection that favored the completely plated
morph in Lake Washington.
Changes in Selection Regime in Lake Washington
By using the empirically estimated values of m, we found that
a selection coefficient s (strength of selection for the com-
lutionary shift from 1969 to 1976 (from 6% completely plated
morphs to 40.2% completely plated morphs) (Figure 1A). This
suggests that the complete morph had 58%–72% greater fit-
ness than that of the low-plated morph during this period. To
explain the transition between 1976 and 2005 (from 40.2%
completely plated morphs to 49% completely plated morphs),
an s of 0.01–0.03 is required (Table 1). We thus conclude that
there was a period of very intense selection for the completely
low-level fitness advantage (1%–3%) of the completely plated
morph over the low-plated morph.
One of the dramatic ecological changes that occurred in
Lake Washington during the early 1970s is increased water
transparency as a result of the mitigation of eutrophication in
the late 1960s. Water transparency in the lake was 1–2 m Sec-
chi depth (the maximum depth at which a white Secchi disk
is visible from the water surface) during 1955–1971, and it
increased to 3.4 m in 1973 and then to 6–7 m from 1976 to
the present [14, 15]. Previous behavioral experiments have
demonstrated that an increase in water transparency signifi-
cantly increases the reaction distance of visual predators to
their prey, thus leading to increased predation pressure on
prey fish . Cutthroat trout (Oncorhynchus clarki) are visual
predators, extremely sensitive to subtle changes in water
transparency , and are the primary predators of threespine
sticklebacks in both the littoral and pelagic zones of Lake
Washington [16, 25, 26]. Therefore, we used a visual-foraging
model, which calculates the search volume by cutthroat
trout as a function of light intensity and turbidity [27, 28], to
investigate a possible change in the stickleback predation
regime. This analysis demonstrated that the increase in lake
transparency created an 8-fold increase in the visual-search
volume of cutthroat trout and also expanded the depth range
over which effective visual foraging could occur (Figure 3).
Most of the expanded search volume was achieved during
1972–1975, when the mean Secchi-disk transparency in-
creased to 3.4 m. Although the cutthroat trout population
in Lake Washington did not increase between 1971 and
2006 (Figure S8), our model suggests that an increase in lake
transparency could have changed the predation regime by
increasing encounter rates between sticklebacks and cut-
Predation by toothed predators, such as cutthroat trout, is
thought to favor completely plated sticklebacks because the
posterior lateral plates can protect the stickleback from being
injured and swallowed [29, 30]. Reimchen predicted that the
completely plated morph would occur in open-water habitats
of high clarity where capture rather than pursuit defenses
predominate . Consistent with this hypothesis, we have
shown that the increase in the frequency of completely plated
morphs occurred during the time when the water clarity in-
creased dramatically in Lake Washington, a relatively deep
and large lake (with a surface area of 8.76 3 107m2and a max-
imum depth of 65.2 m). Further supporting the hypothesis that
an increase in predation by cutthroat trout has contributed to
the rapid evolution of Lake Washington stickleback, recent
stickleback samples are larger than historical stickleback sam-
ples (Figure 1B and Table S4). Larger body size can protect
against predation by gape-limited predators such as cutthroat
trout [31, 32]. Although salinity and water temperature have
lution [33, 34], we can exclude a role for these abiotic factors in
the evolution of Lake Washington sticklebacks (Supplemental
We have reported a dramatic example of ‘‘reverse evolution’’
, in which there has been an increase in completely plated
sticklebacks in a freshwater lake. Our data demonstrate that
selection for the complete morph was particularly strong dur-
ing the early 1970s, suggesting that the main increase in the
frequency of completely plated fish might have occurred dur-
ing a time period of less than a decade. Armor reduction has
also been shown to occur within only a few decades after
the introduction of marine sticklebacks into freshwater [35–
37].Thus, sticklebacks canrespond toenvironmental changes
byeitheranincrease oradecreaseinlateralplateswithin afew
decades. Rapid phenotypic evolution in sticklebacks provides
us with a great opportunity to further investigate the mecha-
nisms by which animals can respond to rapidly changing
The rapid evolution of armor plates in Lake Washington
sticklebacks might have been enabled by the presence of
standing genetic variation at the major plate locus . With-
out standing variation, a sudden increase in predation might
have led to population extinction before a new mutation
appeared [6, 39]. Although an increase in gene flow was
not the primary cause of armor evolution, gene flow from
the marine population might have enabled rapid armor evolu-
tion by contributing to standing genetic variation within the
lake [7, 40]. This work provides an example of a rapid pheno-
typic change that does not result from phenotypic plasticity,
which has been proposed as a major mechanism of contem-
porary evolution [4, 8]. Thus, investigation of the genetic
mechanisms that underlie adaptive phenotypes is essential
Table 1. Estimation of the Strength of Selection (s) for the Completely
Dominance of Plate-Morph Fitness
h = 0h = 0.5h = 1
s = 0.708/0.720
s = 0.013/0.015
s = 0.597/0.606
s = 0.017/0.020
s = 0.582/0.591
s = 0.026/0.031
Values of s during different time periods were calculated for different values
of the dominance of plate-morph fitness (h). Migration rates estimated by
LAMARC (left side in each cell) and IM (right side in each cell) were used.
Frequencies of 6%, 40.2%, and 49.0% completely plated morphs were
used for 1969, 1976, and 2005.
Current Biology Vol 18 No 10
for a better understanding of rapid evolutionary change
[2, 4, 8, 41].
Although we suggest that reverse evolution in the Lake
and additional factors might have also contributed. As in many
other cases of rapid evolution , it is often difficult to tease
lution. We demonstrated, however, that changes in water clar-
ity are able to influence predator-prey interactions; further
attention should be given to the influence of water clarity on
predator-prey interactions and animal evolution. In addition,
this work highlights the importance of investigation of the rela-
tionships between environmental changes, species interac-
tions, and the genetic basis of phenotypic evolution, both to
better understand the mechanisms of animal evolution and to
inform conservation efforts.
Detailed experimental procedures, a supplemental discussion, eight fig-
ures, and five tables are available with this paper online at http://www.
Figure 3. Changes in Cutthroat-Trout Foraging Volume in Lake
(A–C) Depth-specific search volumes of cutthroat trout are shown dur-
ing peak eutrophication, from 1955–1971 (A); during initial recovery,
from 1972–1975 (B); and during the current transparency regime, from
1976–2006 (C). Different diel periods are indicated with white bars for
day, hatched bars for dusk, and black bars for night.
(D) Changes in the search volume by trout are shown as the percentage
ofthecurrentsearchvolume.Blackandwhitecircles indicate thesearch
volumes at the depths of 0–60 m and 10–20 m, respectively.
This work was supported bya Uehara Memorial Fellowship (J.K.), bythe
Packard Foundation (D.I.B.), by Seattle Public Utilities (D.A.B.), by the
Water and People Project (S.M.), and by a Burroughs Wellcome Fund
Career Award in the Biomedical Sciences (C.L.P.). Sampling was con-
ducted under the Washington Department of Fish and Wildlife permits
to C.L.P. (05-049, 06-159, 07-047) and D.A.B. (06-115) and an ESA Sec-
tion 10a permit #1376 from the National Oceanic and Atmospheric Ad-
ministration to D.A.B. All experiments were approved by Institutional
Animal Care and Use Committees (Fred Hutchinson Cancer Research
Center IACUC no. 1575; University of Washington IACUC no. 3286-01).
We thank D. Schluter, D. Kingsley, J. Gow, and B. Stein for comments
on the manuscript; K. Maslenikov and T.W. Pietsch for access to the
University of Washington Fish Collection; the Seabeck Conference Cen-
touck, and N. Hurtado for technical assistance; and T.P. Quinn, P. West-
ley, F. Goetz, L. Gilbertson, W. Aron, T. Flagg, D. Maynard, Y. Kitano, C.
Sergeant, N. Overman, A. Bruner, A. Wark, P. Pagels, members of the
Peichel and Beauchamp labs, and many field assistants for discussions
and sampling assistance.
Received: January 18, 2008
Revised: March 31, 2008
Accepted: April 8, 2008
Published online: May 15, 2008
1. Palumbi, S.R. (2001). The Evolution Explosion: How Humans Cause
Rapid Evolutionary Change (New York: W.W. Norton & Company).
2. Smith, T.B., and Bernatchez, L. (2008). Evolutionary change in human-
altered environments. Mol. Ecol. 17, 1–8.
3. Carroll, S.P., Hendry, A.P., Reznick, D.N., and Fox, C.W. (2007). Evolu-
tion on ecological time-scales. Funct. Ecol. 21, 387–393.
4. Hendry, A.P., Ferrugia, T.J., and Kinnison, M.T. (2008). Human
influences on rates of phenotypic change in wild animal populations.
Mol. Ecol. 17, 20–29.
5. Majerus, M. (1998). Melanism: Evolution in Action (Oxford: Oxford
6. Reznick, D.N., Ghalambor, C.K., and Crooks, K. (2008). Experimental
studies of evolution in guppies: a model for understanding the evolu-
tionary consequences of predator removal in natural communities.
Mol. Ecol. 17, 97–107.
7. Stockwell, C.A., Hendry, A.P., and Kinnison, M.T. (2003). Contemporary
8. Gienapp, P., Teplitsky, C., Alho, J.S., Mills, J.A., and Merila, J. (2008).
Climate change and evolution: disentangling environmental and genetic
responses. Mol. Ecol. 17, 167–178.
9. Teoto ´nio, H., and Rose, M.R. (2001). Perspective: reverse evolution.
Evolution Int. J. Org. Evolution 55, 653–660.
10. Bell, M.A., and Foster, S.A. (1994). The Evolutionary Biology of the
Threespine Stickleback (Oxford: Oxford University Press).
biology. Trends Ecol.Evol.18,
Reverse Evolution of Armor in Stickleback
11. Colosimo, P.F., Hosemann, K.E., Balabhadra, S., Villarreal, G., Dickson,
M., Grimwood, J., Schmutz, J., Myers, R.M., Schluter, D., and Kingsley,
D.M. (2005). Widespread parallel evolution in sticklebacks by repeated
fixation of Ectodysplasin alleles. Science 307, 1928–1933.
12. Kingsley, D.M., and Peichel, C.L. (2007). The molecular genetics of
evolutionary change in sticklebacks. In Biology of the Three-Spined
Stickleback, S. O¨stlund-Nilsson, I. Mayer, and F.A. Huntingford, eds.
(Boca Raton: CRC press), pp. 41–81.
13. Schluter, D. (2000). The Ecology of Adaptive Radiation (New York:
Oxford University Press).
14. Edmondson, W.T. (1991). The Uses of Ecology: Lake Washington and
Beyond (Seattle: University of Washington Press).
15. Edmondson, W.T. (1970). Phosphorus, nitrogen, and algae in Lake
Washington after diversion of sewage. Science 169, 690–691.
16. Beauchamp, D.A., Vecht, S.A., and Thomas, G.L. (1992). Temporal, spa-
Northwest Sci. 66, 149–159.
17. Hagen, D.W., and Gilbertson, L.G. (1972). Geographic variation and en-
vironmental selection in Gasterosteus aculeatus L. in the Pacific North-
west, America. Evolution Int. J. Org. Evolution 26, 32–51.
fish population in southern Lake Washington. University of Washington
19. Colosimo, P.F., Peichel, C.L., Nereng, K., Blackman, B.K., Shapiro,
M.D., Schluter, D., and Kingsley, D.M. (2004). The genetic architecture
of parallel armor plate reduction in threespine sticklebacks. PLoS
Biol. 2, 635–641.
20. Pritchard, J.K., Stephens, M., and Donnelly, P. (2000). Inference of
population structure using multilocus genotype data. Genetics 155,
21. Beerli, P., and Felsenstein, J. (1999). Maximum-likelihood estimation of
migration rates and effective population numbers in two populations
using a coalescent approach. Genetics 152, 763–773.
22. Nielsen, R., and Wakeley, J. (2001). Distinguishing migration from isola-
tion: a Markov chain Monte Carlo approach. Genetics 158, 885–896.
23. Kuhner, M.K. (2006). LAMARC 2.0: maximum likelihood and Bayesian
estimation of population parameters. Bioinformatics 22, 768–770.
24. Mazur, M.M., and Beauchamp, D.A. (2003). A comparison of visual prey
detection among species of piscivorous salmonids: effects of light and
low turbidities. Environ. Biol. Fishes 67, 397–405.
25. Nowak, G.M., Tabor, R.A., Fresh, K.L., and Quinn, T.P. (2004). Ontoge-
netic shifts in habitat and diet of cutthroat trout in Lake Washington,
Washington. N. Am. J. Fish. Manage. 24, 624–635.
26. Beauchamp, D.A. (1994). Spatial and temporal dynamics of piscivory:
implications for food web stability and the transparency of Lake Wash-
ington. Lake Reservior Manage. 9, 151–154.
27. Beauchamp, D.A., Baldwin, C.M., Vogel, J.L., and Gubala, C.P. (1999).
Estimating diel, depth-specific foraging with a visual encounter rate
model for pelagic piscivores. Can. J. Fish. Aquat. Sci. 56 (Suppl. 1),
28. Mazur, M.M., and Beauchamp, D.A. (2006). Linking piscivory to spatial-
temporal distributions of pelagic prey fishes with a visual foraging
model. J. Fish Biol. 69, 151–175.
predator (Oncorhynchus) and implications for the evolution of lateral
plates. Evolution Int. J. Org. Evolution 46, 1224–1230.
30. Reimchen, T.E. (2000). Predator handling failures of lateral plate morphs
in Gasterosteus aculeatus: implications for stasis and distribution of the
ancestral plate condition. Behaviour 137, 1081–1096.
31. Reimchen, T.E. (1991). Trout foraging failures and the evolution of body
size in stickleback. Copeia 1991, 1098–1104.
32. Moodie, G.E.E. (1972). Predation, natural selection and adaptation in an
unusual threespine stickleback. Heredity 28, 155–167.
33. Heuts, M.J. (1947). Experimental studies on adaptive evolution in
Gasterosteus aculeatus L. Evolution Int. J. Org. Evolution 1, 89–102.
34. Hagen, D.W., and Moodie, G.E.E. (1982). Polymorphism for plate
morphs in Gasterosteus aculeatus on the east coast of Canada and
an hypothesis for their global distribution. Can. J. Zool. 60, 1032–1042.
35. Bell, M.A., Aguirre, W.E., and Buck, N.J. (2004). Twelve years of contem-
poraryarmorevolution inathreespinesticklebackpopulation. Evolution
Int. J. Org. Evolution 58, 814–824.
36. Klepaker, T. (1993). Morphological changes in a marine population of
threespine stickleback, Gasterosteus aculeatus, recently isolated in
freshwater. Can. J. Zool. 71, 1251–1258.
37. Kristja ´nsson,B.K.,Sku ´lason,S.,andNoakes,D.L.G.(2002).Rapiddiver-
gence in a recently isolated population of threespine stickleback
(Gasterosteus aculeatus). Evol. Ecol. Res. 4, 659–672.
38. Reimchen, T.E., and Nosil, P. (2002). Temporal variation in divergent se-
lection on spine number in threespine stickleback. Evolution Int. J. Org.
Evolution 56, 2472–2483.
39. Barrett, R.D.H., and Schluter, D. (2008). Adaptation from standing
genetic variation. Trends Ecol. Evol. 23, 38–44.
40. Garant, D., Forde, S.E., and Hendry, A.P. (2007). The multifarious effects
ofdispersal andgeneflowoncontemporary adaptation.Funct. Ecol.21,
41. True, J.R. (2003). Insect melanism: the molecules matter. Trends Ecol.
Evol. 18, 640–647.
Current Biology Vol 18 No 10