How did some ancestral sea stickleback populations come to live exclusively in freshwater

Biol J Linn Soc Lond. Author manuscript; available in PMC 2013 Mar 1.

Published in final edited form as:

PMCID: PMC3351840

NIHMSID: NIHMS330997

Abstract

For over a century, evolutionary biologists have debated whether and how phenotypic plasticity impacts the processes of adaptation and diversification. The empirical tests required to resolve these issues have proven elusive, mainly because it requires documentation of ancestral reaction norms, a difficult prospect where many ancestors are either extinct or have evolved. The threespine stickleback radiation is not limited in this regard, making it an ideal system in which to address general questions regarding the role of plasticity in adaptive evolution. As retreating ice sheets have exposed new habitats, oceanic stickleback founded innumerable freshwater populations, many of which have evolved parallel adaptations to their new environments. Because the founding oceanic population is extant, we can directly evaluate whether specific patterns of ancestral phenotypic expression in the context of novel environments (plasticity), or over ontogeny, predisposed the repeated evolution of “benthic” and “limnetic” ecotypes in shallow and deep lakes, respectively. Consistent with this hypothesis, we found that oceanic stickleback raised in a complex habitat and fed a macroinvertebrate diet expressed traits resembling derived, benthic fish. Alternatively, when reared in a simple environment on a diet of zooplankton, oceanic stickleback developed phenotypes resembling derived, limnetic fish. As fish in both treatments grew, their body depths increased allometrically, as did the size of their mouths, while their eyes became relatively smaller. Allometric trajectories were subtly but significantly impacted by rearing environment. Thus, both environmental and allometric influences on development, along with their interactive effects, produced variation in phenotypes consistent with derived benthic and limnetic fish, which may have predisposed the repeated genetic accommodation of this specific suite of traits. We also found significant shape differences between marine and anadromous stickleback, which has implications for evaluating the ancestral state of stickleback traits.

Introduction

Although considered intermittently for over a century (Baldwin, 1902; Morgan, 1896; Osborn, 1896; Simpson, 1953; Waddington, 1953), the recent emergence of evolutionary developmental biology has stimulated a renewed interest in how environmental influences on phenotypic development (phenotypic plasticity) might impact evolutionary processes (Ghalambor et al., 2007; Pigliucci, 2007; Price et al., 2003; Robinson & Dukas, 1999; Schlichting, 2004; West-Eberhard, 2003; Wimberger, 1994). Because the environment mediates how genetic variation translates into heritable, phenotypic variation, the environment directly affects the distribution of phenotypes upon which natural and sexual selection act. As such, phenotypic plasticity should substantially influence the processes of adaptation, speciation and diversification, a theory that is gaining empirical support (reviewed in Pfennig et al., 2010).

The goal of this study is to understand whether phenotypic plasticity can in part cause an adaptive radiation to exhibit parallel ecotypic variation: the repeated evolution of specific suites of traits in similar environments (reviewed in Schluter, 2000). One explanation for such “replicate radiations” is that similar selective pressures lead to similar adaptive outcomes in newly isolated lineages (Schluter & Nagel, 1995), which seems particularly likely if a source population provides the same standing genetic variation to multiple colonizing populations (Barrett & Schluter, 2008; Colosimo et al., 2005). However, a new environment not only imposes new selection pressures, it also determines how standing genetic variation translates into the phenotypic variation upon which those selection pressures act (Gibson & Dworkin, 2004; Le Rouzic & Carlborg, 2008; McGuigan et al., 2010a; Schlichting, 2008). When an ancestral stem group repeatedly encounters similar environments, the same phenotypes are expressed via plasticity and then exposed to similar selection pressures, leading to the evolution of parallel ecotypic variation (the “flexible stem hypothesis,” West-Eberhard, 2003). A key prediction of this hypothesis is that the phenotypes associated with derived ecotypes will resemble those that initially arose via plasticity in the ancestral colonists.

Testing the flexible stem hypothesis requires observing ancestral development in the context of derived environments—a challenge since ancestors are typically extinct or now adapted to derived conditions. Several recent studies have circumvented this problem either by inferring ancestral patterns of plasticity from the traits of derived taxa (e.g., Aubret et al., 2007; Gomez-Mestre & Buchholz, 2006; Ledon-Rettig et al., 2008; Losos et al., 2000), or by identifying systems in which both ancestral and derived populations are extant, permitting direct comparisons between ancestral plasticity and derived, adaptive variation (e.g., Scoville & Pfrender, 2010; Wund et al., 2008).

The threespine stickleback fish (Gasterosteus aculeatus) offers an outstanding opportunity to compare ancestral reaction norms to derived ecotypes. This small, holarctic fish has experienced an expansive freshwater radiation, much of which has taken place since the last glacial recession, which began 15 – 20,000 years ago. The stickleback radiation is characterized by parallel ecotypic variation along a number of phenotypic axes, including reproductive behavior (Foster et al., 2008; Shaw et al., 2007), defensive armor (Bell & Ortí, 1994; Moodie & Reimchen, 1976; Reimchen, 1994); foraging behavior (Hart & Gill, 1994), and morphology (Hart & Gill, 1994; Walker, 1997; Walker & Bell, 2000). Moreover, the oceanic population that produced the freshwater radiation is still widespread, allowing us to determine how ancestral, oceanic phenotypes developed in the context of novel, freshwater environments. Morphological and genetic evidence indicate that modern oceanic stickleback have changed little since giving rise to the freshwater radiation (Cresko, 2000; Deagle et al., 1996; Haglund et al., 1992; Mäkinen et al., 2006; Orti et al., 1994; Taylor & McPhail, 1999; Walker & Bell, 2000; Withler & McPhail, 1985); thus, modern oceanic stickleback can be used to infer ancestral reaction norms.

We are specifically interested in whether the repeated evolution of “benthic” and “limnetic” foraging ecotypes resulted in part from recurring patterns of developmental plasticity in their ancestors. In numerous lakes throughout the freshwater radiation, stickleback morphology has diverged in relation to the amount of shallow, littoral habitat present in a lake (Aguirre, 2009; Hart & Gill, 1994; Spoljaric & Reimchen, 2007; Walker, 1997; Willacker et al., 2010). Benthic stickleback inhabit shallow, relatively eutrophic lakes with much littoral habitat. They forage on zooplankton until they reach about 1 cm in length and then switch to feeding on benthic macroinvertebrates (Hart & Gill, 1994). Their adult morphology reflects this lifestyle, with deep, maneuverable bodies, deep heads with short snouts (useful for sucking invertebrates from the substrate, Carroll et al., 2004; Webb, 1984), large mouths and short, widely spaced gill rakers (Hart & Gill, 1994; Walker, 1997; Willacker et al., 2010). Limnetic stickleback live in deep, relatively oligotrophic lakes and eat primarily zooplankton for their entire lives (Hart & Gill, 1994). They have a number of morphological adaptations associated with cruising through the water column, searching for and handling tiny prey. These include a long, streamlined head and body, relatively large eyes (although not exclusively, see Willacker et al., 2010), and long, closely-spaced gill rakers that help them retain small prey in their buccal cavities as water exits their gills (Hart & Gill, 1994; Walker, 1997). This ecotypic variation is likely adaptive, given that it has been linked to foraging performance in shallow and deep water environments (Robinson, 2000); moreover stickleback of intermediate phenotype are outperformed by the more extreme forms in their respective environments (Hatfield & Schluter, 1999; Schluter, 1993).

If the flexible stem hypothesis is correct, oceanic stickleback should develop phenotypes resembling those of derived, benthic and limnetic ecotypes when reared in environments that mimic shallow and deep lakes, respectively. Previously, we performed a partial test of this prediction by rearing marine stickleback on a diet of either chironomid larvae, reflecting a benthic diet, or zooplankton, reflecting a limnetic diet (Wund et al., 2008). While plasticity of some traits did support the flexible stem hypothesis (e.g., head shape), other traits were either not plastic (e.g., body shape), or exhibited plastic variation contrary to predictions (e.g., eye size, gill raker length). Our interpretations of these outcomes might have been limited by the relatively simplistic nature of our experiment: we only considered a single environmental variable (diet) and we examined plasticity in only small fish of roughly uniform size. In order to come to a complete and accurate understanding of phenotypic evolution, it is important to more thoroughly consider the mechanisms of phenotypic development, which include the interacting inputs of environmental variation (plasticity), ontogeny and allometric relationships among traits (the “developmental norm of reaction,” Schlichting & Pigliucci, 1998).

In the present study, our aim is to more thoroughly characterize morphological plasticity in ancestral threespine stickleback in order to test the flexible stem hypothesis. This experiment diverges from our previous work in three important ways. First, to more completely represent natural environmental variation, alternative treatments differ not only in diet, but also in habitat structure. Second, by sampling individuals at regular intervals during the study we evaluate both the independent and interactive effects of the environment and ontogenetic allometry on ancestral trait development. We predict that the benthic treatment, with a diet of macroinvertebrates and a complex habitat structure, would lead to deep heads and bodies, relatively small eyes, and short, widely-spaced gill rakers. Conversely, when reared on a diet of zooplankton in a structurally simple environment, oceanic stickleback would develop narrow heads and bodies, relatively large eyes and long, closely-spaced gill rakers. Body shape in wild-caught, marine stickleback varies with size (Spoljaric & Reimchen, 2011), and thus we characterized allometric changes in shape and asked whether the relationship between size and shape differed across the two rearing environments. Heterochronic shifts in body shape allometry may also have played a role in the evolution of benthic and limnetic stickleback (West-Eberhard, 2003) so it is important to establish ancestral allometric relationships to determine whether benthic- and limnetic-like phenotypes emerge over the course of individual ontogeny. Finally, oceanic stickleback exhibit two distinct life-history strategies, with some individuals being fully marine, and others living in the ocean but breeding in freshwater (“anadromy;” Baker, 1994). The marine form breeds in shallow (~ 0.25 – 0.5m depth) tide pools, whereas anadromous stickleback move into freshwater streams to breed. While habitat differences exist early in life (e.g., salinity, flow, prey), it is unknown as to whether the ecology of the two forms differ once they move out to the ocean after 1–2 months of age. Relative to the freshwater radiation, there is little morphological variation among oceanic stickleback (Walker & Bell, 2000); however, given that our work depends so strongly on our choice of a representative ancestor, one goal of this study was to characterize any differences in morphology (or morphological plasticity) between these two forms that could influence the choice of representative ancestors in future studies.

Methods

Field Collection, Generation of Laboratory-reared Families (Crosses) and Fish Maintenance

Adult stickleback were collected in June of 2008 using unbaited minnow traps. Two representatives of the ancestral stickleback were sampled: a fully marine population (Mud Bay), which breeds in tide pools near Homer, AK (59°38’16.5”N, 151°29’56.3”W) and an anadromous population that breeds in a freshwater creek, Rabbit Slough, near Wasilla, AK (61°32’9.87”N, 149°15’13.78”W). Five full-sib families were generated from each population. Crosses were made by extruding eggs from gravid females, who were then euthanized with an overdose of MS-222. After males were euthanized in MS-222, their testes were removed, macerated and sperm was spread over the eggs to induce fertilization. Fertilized embryos were washed each day with clean embryo medium (distilled water, 0.5 ppt Instant Ocean® sea salt). Within 2–3 days post-fertilization, clutches were disinfected in a 1% iodine solution and before day 5, were shipped overnight to Clark University in Worcester, MA. Fry were kept in Petri dishes until they absorbed their yolk sacs, after which they were fed “Platinum” ® brine shrimp nauplii (Artemia spp., Argent Chemical Co., Redmond, WA). As they grew, fry were moved to 0.45 L and then 0.9 L jars. After several weeks, they were transferred to 37 L aquaria in a recirculating aquarium system, and maintained on an 8 hour: 16 hour light:dark cycle at a temperature of 17°C. The short day length ensured that the fish would not become reproductive during the course of the study (nor any of the other fish in our facility), while the water temperature was maintained in a range to promote growth. Fish were fed brine shrimp nauplii until they became large enough to eat blended bloodworms (chironomid larvae). At this point, fish were placed in their respective experimental aquaria and the experiment began (fish ~1 cm standard length, and 2.5 months old).

Experimental Setup

The experiment was conducted in 37 L aquaria in the recirculating systems with two treatments: benthic and limnetic. Twenty-four individuals from each of the ten families were divided into two adjacent tanks, each representing one of the two treatments. Each tank contained 12 full siblings, and all families were represented in both treatments (Fig. 1.). To simulate the structural complexity of a benthic environment, benthic tanks contained gravel bottoms, two inverted, plastic “strawberry” baskets and nine small, plastic plants, 8 of which were anchored in the tops of the plastic baskets with the 9th being rooted in the gravel substrate. The arrangement of these structures was identical for all benthic tanks. Fish in this treatment were fed bloodworms twice a day, with the food being distributed within the gravel, plants and baskets, thus requiring fish to maneuver through a complex habitat and search for food. Bloodworms were briefly blended for the first 50 days of the experiment so they would be small enough for the fish to ingest. Subsequently, the fish were large enough to eat full sized, unblended worms. Three times a week, benthic fish were fed nutrient rich “Cyclop-eze” ® flakes (Argent Chemical Co., Redmond, WA).

Experimental design. Two fish from each aquarium were preserved every 6 weeks, for a total of 24 weeks (8 fish per aquarium).

For the limnetic treatment, aquaria were kept completely bare to simulate an open, freshwater environment. Fish in this treatment were also fed twice daily, and continued their diet of “Platinum” brine shrimp nauplii for the duration of the experiment. This treatment was also supplemented three times a week with Cyclop-eeze, but in this case, as whole freeze-dried crustaceans, a neutral-bouyancy, planktonic food. Thus, both treatments received the same type of dietary supplement, but which was size-appropriate to each treatment. This strategy was designed to minimize morphological differences due to diet quality (Wimberger, 1993). Every 6 weeks for 24 weeks, two fish were haphazardly chosen from each tank, euthanized with MS-222 and preserved in 10% formaldehyde. Fish were then stained with Alizarin red S and stored in 60% isopropanol in individual vials. Three Rabbit Slough families suffered some mortality due to a microsporidial infection in some individuals, so out of the 160 possible individuals included in the experiment, 151 survived (Mud Bay, Benthic: N = 40; Mud Bay, Limnetic: N = 40; Rabbit Slough, Benthic: N = 37; Rabbit Slough, Limnetic: N = 34).

Morphometric analyses

The 151 specimens were photographed with a 10 Megapixel Canon Digital Rebel XTi under identical lighting and exposure conditions. Slightly bent specimens were pinned in place to ensure that all photos were taken from an identical lateral view. Geometric morphometrics (reviewed in Zelditch et al., 2004) was used to characterize shape variation among subjects. We used the same set of 16 landmarks as in our previous work (Wund et al., 2008; Fig. 2), which were selected to capture variation in both cranial and postcranial external morphology. Landmarks were identified on each specimen using TPSdig v. 2.10 (Rohlf, 2004). A single observer (SV) placed all landmarks, and was blind to treatment and population of origin. To determine the location of landmark 3, a straight line was created between landmarks 1 (the tip of the snout) and 2 (the center of the eye), and landmark 3 was placed at the intersection between this line and the edge of the eye.

The 16 digital landmarks used to describe fish shape in left lateral view: (1) anterior tip of upper lip; (2) center of eye; (3) eye radius as defined by line created by points 1 and 2; (4) supraoccipital notch lateral to dorsal midline; (5) farthest posterior point in skull; (6) junction of head to body on ventral midline; (7) posterior edge of angular; (8) posterior end of mouth; (9) anterior junction of first dorsal spine; (10) base of dorsal fin ray; (11, 12, 13) caudal peduncle; (14) base of first anal fin ray; (15) junction of pelvic spine on pelvic girdle; (16) upper edge of junction of pectoral fin to body.

Generalized Procrustes Analysis superimposition was used to obtain shape variables that are independent of size, rotation and translation (Rohlf, 1990; Walker, 1997; Zelditch et al., 2004). A consensus configuration of all specimens, specimen centroid sizes (a measure of size independent of shape), as well as principal and partial warps from a thin plate spline bending energy matrix were generated using IMP geometric morphometric software (Sheets, 2003a; Sheets, 2003b). The partial warps (“shape variables”) of the bending energy matrix can be analyzed with conventional multivariate statistical analysis (Zelditch et al., 2004). Multivariate analysis of covariance (MANCOVA, Sokal & Rohlf, 1995) was performed on partial warp scores and uniform components to determine if shape varied between populations, treatments and among families (nested within population), with centroid size as a covariate. Centroid size and preservation stage were highly correlated (r = 0.934; P < 0.001), so stage was not included in the MANCOVA models to avoid multicollinearity. It is also worth noting that we chose to preserve fish at six week intervals to achieve a final range of sizes to use as a continuous covariate, not because they represented some inherent ontogenetic stages. We also included population x treatment, population x size, treatment x size and population x treatment x size interactions in our initial models to determine whether populations and/or treatments exhibited different allometric changes in shape, or if populations exhibited different patterns of plasticity across treatments. Backwards elimination was used to remove interaction terms that did not explain a significant amount of variation in shape. A one-way ANCOVA was performed on standard lengths and centroid sizes to determine whether fish from different populations and/or treatments differed in size.

As a complement to our MANCOVA approach, we also employed discriminant function analysis (Kleinbaum et al., 1988) to identify axes of shape variation that best distinguished populations and treatments. Because MANCOVA indicated a significant treatment x size interaction (see below), we performed a DFA to look for treatment differences using size-adjusted residual Procrustes coordinates. A regression of Procrustes coordinates on size was performed separately for each treatment, and then DFA was performed on the combined residuals from these two analyses. Thin plate spline was used to generate deformation grids that permit a visual interpretation of shape variation along the discriminant functions. DFA and thin plate spline analyses were performed using MorphoJ v. 1.02h (Klingenberg, 2011).

Gill raker measurement

In addition to examining external morphology, we also examined variation in the gill rakers, an important internal trophic character that helps retain food in the buccal cavity as water exits through the gills. We dissected out the first gill arch on the left side of all fish and placed them in glycerin on a microscope slide. Once mounted and viewed under a dissecting microscope, gill arches were photographed via MIAS2006 and measured in ImageJ version 1.43r (http://rsbweb.nih.gov/ij/). ImageJ’s sequential line tool was used to measure the lengths of the 2nd, 3rd, and 4th rakers on the descending (long) branch of the first arch, which are typically the longest gill rakers. These three lengths were then averaged for each fish. We also measured the linear distance along the gill arch spanned by these three gill rakers. All fish were scored blind by a single observer (SW) with respect to treatment and population of origin. We were able to dissect out and measure gill rakers from 147 of the 151 specimens used in the morphometric analyses.

Nested ANCOVAs were used to determine if population, family within population, treatment or head centroid size (covariate) could explain variation in average gill raker length or spacing. Both of these characters were more highly correlated with head centroid size (r = 0.97 and 0.96, respectively) than with standard length (r = 0.91 and 0.89, respectively). Size-adjusted gill raker lengths and spacing were generally normally distributed within population x treatment groups (Shapiro-Wilks test for normality: all P > 0.1, except benthic treatment/Mud Bay population, in which P = 0.046). Population and treatment were treated as fixed effects, and family nested within population as a random effect. As with the MANCOVAs described above, we also included population x size, population x treatment, treatment x size and population x treatment x size interactions in initial MANCOVA models, and used backwards elimination to remove non-significant interaction terms from subsequent analyses.

Results

The 151 fish included in the study ranged in size from a standard length of 17.0 mm to 46.3 mm (mean ± sd = 31.65 ± 8.5 mm), spanning all four preservation stages. The two populations did not differ in size, whether measured by standard length (F1, 149 = 1.096; P = 0.297), or morphometric centroid sizes (F1, 147 = 1.111; P = 0.293). Fish from the two treatments also did not differ in size (F1, 147 = 2.285; P = 0.126).

Shape

Shape differed between the two populations, among families within populations and between the two treatments, and these differences were associated with moderate to large effect sizes (Table 1). Fish also changed shape as they grew, and this allometric affect differed between the two environmental treatments (as indicated by a significant treatment x centroid size interaction; Table 1). No other interactions were significant. With respect to population differences, Mud Bay fish exhibited a more streamlined shape than Rabbit Slough fish (T2 = 159.0; P < 0.001; Mahalanobis D = 2.06; Fig. 3). With respect to plasticity, fish in the limnetic treatment had shallower bodies and more elongate heads than fish in the benthic treatment (T2 = 122.5; P < 0.001; Mahalanobis D = 1.802; Fig. 4). Despite a significant treatment x size interaction in the MANCOVA (Table 1), fish in both treatments exhibited similar changes in shape as they grew, with larger fish exhibiting relatively deeper bodies, smaller eyes, larger mouths, and a more dorsal positioning of the more posterior portion of the skull (Figs. 5a and 5b).

Wireframe graphs depicting the average body shapes of Rabbit Slough fish (black) and Mud Bay fish (grey). Deformations are exaggerated 4× to illustrate differences.

Wireframe graphs depicting the average body shapes of limnetic treatment fish (black) and benthic treatment fish (grey). Deformations are exaggerated 4× to illustrate differences.

Wireframe graphs depicting change in shape as fish grew (small fish = grey, large fish = black). Given the significant treatment x size interaction, allometric shape changes are depicted separately for the (A) limnetic and (B) benthic treatments. Deformations exaggerated 4× to illustrate differences.

Table 1

MANCOVA results for shape

Gill rakers

When adjusted for head centroid size, gill raker length and spacing differed between treatments, but not between populations, although the effect size of treatment on gill raker spacing was quite low (Tables 2 and ​3). For a given body size, fish in the limnetic treatment had longer, more closely-spaced gill rakers than those in the benthic treatment (Figure 6). Gill raker length also differed significantly among families (Table 2). Interactions among treatment, population and head size did not affect either gill raker length or spacing (all P > 0.05). Considering the partial effect sizes of these traits, the environment had a much larger effect on gill raker length than spacing.

2-D reaction norm graphs depicting differences in (A) the length of the three longest gill rakers on the 1st gill arch and (B) the mean distance spanned by these gill rakers. Dashed lines = Mud Bay, solid lines = Rabbit Slough. Values are LS means adjusted for head centroid size, ± 95% CI.

Table 2

ANCOVA results for gill raker length

Table 3

ANCOVA results for gill raker spacing along the gill arch

Discussion

The environment directly impacts how genetic variation translates into the expressed phenotypic variation upon which selection pressures act, so patterns of plasticity that emerge following a change in selection pressures should influence the nature of adaptive evolution. The results of this study support a key prediction of this hypothesis, that the phenotypes revealed by plasticity in a colonizing population will resemble the type of adaptive phenotypes that ultimately evolve. When reared in aquaria with no structure and only planktonic prey, oceanic stickleback exhibited features similar to what is observed in the derived, limnetic ecotype: long heads, narrow bodies, and long, narrowly spaced gill rakers. When reared in an environment simulating a shallow lake, oceanic stickleback exhibited benthic-like features: a deep head and body, and short, widely-spaced gill rakers. Thus, given that plasticity consistently revealed these patterns of variation, it is not surprising that in lake after lake, selection would produce the same general outcomes. Both body depth and head depth also increased disproportionately with overall size, so selection on allometric trajectories should further reinforce evolution along these phenotypic trajectories. Indeed, it has been proposed that the evolution of the limnetic ecotype represents a heterochronic shift in allometric growth, with adult limnetics retaining a juvenile-like, narrow body shape (West-Eberhard, 2003).

Considering the effect sizes of in our MANCOVAs and ANCOVAs, it is clear that allometric effects on shape were quite substantial. However, available evidence suggests that Rabbit Slough stickleback harbor little to no additive genetic variance in the slope of static allometry (the relationship between size and shape among individuals of the same age; McGuigan et al., 2010b), so additive variance in the slopes of ontogenetic allometry (allometries of individuals of different ages/sizes) is likely not present, since the latter leads to the former (Shingleton et al., 2007). Allometry, and the expression of genetic variance thereof, might be environmentally contingent. A key strategy in our experimental design was to consider the interactive effects between the environment and allometry, espousing Schlichting and Pigliucci’s “developmental reaction norm perspective” (“DRN;” 1998). Doing so demonstrated that allometric effects on body shape are subtly, but significantly contingent upon the developmental environment. Experiments currently underway, examining developmental plasticity within freshwater benthic and limnetic stickleback, are exploring whether the ancestral DRN established in the present study was altered via genetic accommodation in the course of adaptation to freshwater environments.

The observed plasticity in head shape was generally consistent with our previous work (Wund et al., 2008), however, there are a number of notable differences between the outcomes of the two experiments, which we attribute to our expanded experimental approach in the present study. In the present study, plasticity of body shape and gill raker morphology were consistent with the flexible stem model, and eyes no longer appeared larger in the benthic treatment as in our previous experiment. These results are also broadly consistent with those of Day and colleagues (Day & McPhail, 1996; Day et al., 1994), who examined diet-induced plasticity of head morphology in sympatric, freshwater benthic and limnetic stickleback populations in British Columbia. However, our results contrast with those of another study that examined plasticity in response to habitat structure alone, which found that wild-caught, riverine stickleback held in spatially complex environments had shallower bodies than those held in spatially simple environments (Garduño-paz et al., 2010). The discrepancy between these two sets of results possibly reflects differences in experimental design, the use of laboratory-bred versus wild-caught fish, and/or the use of riverine versus marine populations. These discrepant outcomes highlight the care that must be applied to designing and interpreting developmental plasticity experiments. By design, the outcomes are contingent upon the environments chosen, a decision that must be made based upon the biology of the organisms and the objectives of the study. Our first experiment laid the groundwork for testing the flexible stem hypothesis, but led to perplexing outcomes in some instances. Moving forward, the developmental reaction norm approach detailed by Schlichting and Pigliucci (Schlichting & Pigliucci, 1998) has led to a more thorough characterization of ancestral patterns of development in relevant ecological settings.

The two populations we examined, one marine (Mud Bay) and one anadromous (Rabbit Slough), differed in body shape, with Rabbit Slough fish having generally deeper bodies than Mud Bay fish. The population effect size was nearly twice as large as that due to plasticity (although the effects of plasticity were manifest both directly, and via an interaction with allometry). Previous work on wild-caught fish has documented relatively little morphological variation among oceanic sampling sites, even between stickleback from the Atlantic and Pacific basins (Walker & Bell, 2000). The fish in this study were reared in freshwater conditions, which could have revealed intrinsic differences between marine and anadromous fish that remain cryptic in the wild. In a recent experiment using Rabbit Slough stickleback, the same anadromous population used in the present study, additive genetic variance was significantly higher when were reared in freshwater versus saltwater (McGuigan et al., 2010a). Perhaps marine stickleback respond differently than anadromous stickleback when exposed to freshwater rearing conditions. It is not currently known why some oceanic stickleback breed in freshwater while others breed in saline tide pools and little is known about their habits in the open ocean or how they choose breeding sites as adults. Our results indicate that marine and anadromous stickleback cannot be considered equivalent with respect to ancestral character state determination; given the importance of stickleback as a model system for studying evolutionary processes, it is important to understand why any such differences exist, and they affected patterns of freshwater diversification.

Although mainly considered on theoretical grounds for the better part of a century, recent empirical evidence from a variety of systems indicates that phenotypic plasticity can indeed impact patterns of diversification (reviewed in Pfennig et al., 2010). Plasticity can facilitate a population’s persistence under novel circumstances, and can allow individuals to explore new ecological niches, bringing newly-expressed phenotypes under novel selective regimes (e.g., Aubret et al., 2007; Badyaev, 2009; Tebbich et al., 2010; Yeh & Price, 2004). The adaptations that result from this process will necessarily reflect the expressed phenotypic variation in the ancestral populations—variation that was induced by a change in the developmental environment (West-Eberhard, 2003). A number of recent studies have found that patterns of plasticity within taxa parallel evolved differences between taxa (e.g., Gomez-Mestre & Buchholz, 2006; Kolbe & Losos, 2005; Losos et al., 2000; Magalhaes et al., 2009; Stauffer & van Snick Gray, 2004), suggesting a role for plasticity in diversification. Other studies have inferred ancestral reaction norms from derived taxa in order to evaluate plasticity’s role in adaptation (e.g., Ledon-Rettig et al., 2008). The strength of the stickleback system, and others like it (e.g., Badyaev, 2009; Parsons & Robinson, 2006; Scoville & Pfrender, 2010), is that ancestral reaction norms can be directly observed in extant taxa, and then linked to patterns of divergence in their descendants (McCairns & Bernatchez, 2010; Shaw et al., 2007; Wund et al., 2008; the present study).

Here, we have strengthened support for the hypothesis that ancestral plasticity predisposed the evolution of benthic and limnetic stickleback ecotypes. Inter-population and interspecific divergence in body shape and trophic morphology characterize many lacustrine fishes, and plasticity has been implicated, either directly or indirectly, in many of these instances (e.g., Adams & Huntingford, 2004; Collyer et al., 2007; Magalhaes et al., 2009; Parsons & Robinson, 2006; Stauffer & van Snick Gray, 2004; Wimberger, 1994, for review). These plastic responses can manifest in complex ways, impacting the functional integration among traits and their development over ontogeny, so understanding phenotypic evolution requires consideration of the complex interactions among traits, across environments and through time (Schlichting & Pigliucci, 1998). When comparing ancestral and derived traits to further elucidate mechanisms of adaptation, it is important to consider how the ancestral traits were expressed not in the ancestral environment, but rather in the context of the derived selective environment. When a population’s adaptive landscape changes, plasticity could lead to an immediate, within generation shift in the distribution of heritable phenotypic variation—it is this altered distribution of traits upon which novel selection pressures will act, thus influencing the subsequent adaptive process. The stickleback radiation is an outstanding model for addressing how ancestral development in novel environments impacts subsequent adaptive radiation (Foster & Wund, 2011; McGuigan et al., 2010a; Shaw et al., 2007; Wund et al., 2008).

Acknowledgments

Frank von Hippel and his laboratory, as well as the staff at the University of Alaska, Anchorage, provided generous material support during our fieldwork in Alaska. MW and SF are part of a NESCent working group exploring the role of polyphenism in diversification and innovation, and all members provided helpful insights into the analyses and their interpretations, including: Ehab Abouheif, Tami Cruickshank, Ian Dworkin, Cris Ledon-Rettig, Armin Moczek, David Pfennig, Carl Schlichting, Sonia Sultan, and Mary Jane West-Eberhard. Stickleback were collected under permit from the Alaska Department of Fish and Game. During the course of this experiment, Wund was supported by a National Institutes of Health Ruth R. Kirchstein NRSA postdoctoral fellowship.

Contributor Information

Matthew A. Wund, Department of Biology, The College of New Jersey, 2000 Pennington Road, Ewing, NJ 08628 USA.

Sophie Valena, Department of Biology, Indiana University, Bloomington, 101 E. 3rd Street, Bloomington, Indiana, 47405, USA.

Susan Wood, Department of Biology, The College of New Jersey, 2000 Pennington Road, Ewing, NJ 08628 USA.

John A. Baker, Department of Biology, Clark University, 950 Main Street, Worcester, MA 01610, USA.

References

• Adams CE, Huntingford FA. Incipient speciation driven by phenotypic plasticity? Evidence from sympatric populations of Arctic charr. Biological Journal of the Linnean Society. 2004;81:611–618. [Google Scholar]
• Aguirre WE. Microgeographic diversification of threespine stickleback: body shape-habitat correlations in a small, ecologically diverse Alaskan drainage. Biological Journal of the Linnean Society. 2009;98:139–151. [Google Scholar]
• Aubret F, Bonnet X, Shine R. The role of adaptive plasticity in a major evolutionary transition: early aquatic experience affects locomotor performance of terrestrial snakes. Functional Ecology. 2007;21:1154–1161. [Google Scholar]
• Badyaev A. Evolutionary significance of phenotypic accommodation in novel environments: an empirical test of the Baldwin effect. Philosophical Transactions of the Royal Society of London Series B-Biological Sciences. 2009;364:1125–1141. [PMC free article] [PubMed] [Google Scholar]
• Baker JA. Life history variation in female threespine stickleback. In: Bell MA, Foster SA, editors. The Evolutionary Biology of the Threespine Stickleback. Oxford: Oxford University Press; 1994. pp. 144–187. [Google Scholar]
• Baldwin JM. Development and Evolution. Macmillan; New York: 1902. [Google Scholar]
• Barrett RDH, Schluter D. Adaptation from standing genetic variation. Trends in Ecology & Evolution. 2008;23:38–44. [PubMed] [Google Scholar]
• Bell M, Orti G. Pelvic reduction in threespine stickleback from Cook Inlet lakes: geographical distribution and intrapopulation variation. Copeia. 1994;1994:314–325. [Google Scholar]
• Carroll AM, Wainwright PC, Huskey SH, Collar DC, Turingan RG. Morphology predicts suction feeding performance in centrarchid fishes. Journal of Experimental Biology. 2004;207:3873–3881. [PubMed] [Google Scholar]
• Collyer ML, Stockwell CA, Adams DC, Reiser MH. Phenotypic plasticity and contemporary evolution in introduced populations: evidence from translocated populations of white sands pupfish (Cyrpinodon tularosa) Ecological Research. 2007;22:902–910. [Google Scholar]
• Colosimo PF, Hosemann KE, Balabhadra S, Villarreal G, Jr, Dickson M, Grimwood J, Schmutz J, Myers RM, Schluter D, Kingsley DM. Widespread parallel evolution in sticklebacks by repeated fixation of ectodysplasin alleles. Science. 2005;307:1928–1933. [PubMed] [Google Scholar]
• Cresko WA. The ecology and geography of speciation: a case study using an adaptive radiation of threespine stickleback in Alaska. Clark University; 2000. [Google Scholar]
• Day T, McPhail JD. The effect of behavioural and morphological plasticity on foraging efficiency in the threespine stickleback (Gasterosteus sp.) Oecologia. 1996;108:380–388. [PubMed] [Google Scholar]
• Day T, Pritchard J, Schluter D. A comparison of 2 sticklebacks. Evolution. 1994;48:1723–1734. [PubMed] [Google Scholar]
• Deagle BE, Reimchen TE, Levin DB. Origins of endemic stickleback from the Queen Charlotte Islands: Mitochondrial and morphological evidence. Canadian Journal of Zoology. 1996;74:1045–1056. [Google Scholar]
• Foster SA, Shaw KA, Robert KL, Baker JA. Benthic, limnetic and oceanic threespine stickleback: profiles of reproductive behavior. Behaviour. 2008:485–508. [Google Scholar]
• Foster SA, Wund MA. Ancestral plasticity and adaptive evolution in the threespine stickleback radiation. In: Hallgrímsson B, Hall BK, editors. Epigenetics: Linking Genotype and Phenotype in Development and Evolution. Los Angeles: University of California Press; 2011. [Google Scholar]
• Garduño-paz MV, Couderc S, Adams CE. Habitat complexity modulates phenotype expression through developmental plasticity in the threespine stickleback. Biological Journal of the Linnean Society. 2010;100:407–413. [Google Scholar]
• Ghalambor CK, McKay JK, Carrol SP, Reznick DN. Adaptive versus non-adaptive phenotypic plasticity and the potential for contemporary adaptation in new environments. Functional Ecology. 2007;21:394–407. [Google Scholar]
• Gibson G, Dworkin I. Uncovering cryptic genetic variation. Nature Reviews Genetics. 2004;5:681–690. [PubMed] [Google Scholar]
• Gomez-Mestre I, Buchholz DR. Developmental plasticity mirrors differences among taxa in spadefoot toads linking plasticity and diversity. Proceedings of the National Academy of Sciences. 2006;103:19021–19026. [PMC free article] [PubMed] [Google Scholar]
• Haglund TR, Buth DG, Lawson R. Allozyme variation and phylogenetic relationships of Asian, North American, and European populations of the threespine stickleback. Copeia. 1992;1992:432–443. [Google Scholar]
• Hart PJB, Gill AB. Evolution of foraging behaviour in the threespine stickleback. In: Bell MA, Foster SA, editors. The Evolutionary Biology of the Threespine Stickleback. Oxford: Oxford University Press; 1994. pp. 207–239. [Google Scholar]
• Hatfield T, Schluter D. Ecological speciation in sticklebacks: environment-dependent hybrid fitness. Evolution. 1999;53:866–873. [PubMed] [Google Scholar]
• Kleinbaum DG, Kupper LL, Muller KE. Applied Regression Analysis and Other Multivariable Methods. PWS-Kent; Boston: 1988. [Google Scholar]
• Klingenberg CP. MorphoJ: an integrated software package for geometric morphometrics. Molecular Ecology Resources. 2011;11:353–357. [PubMed] [Google Scholar]
• Kolbe JJ, Losos JB. Hind-limb length plasticity in Anolis carolinensis. Journal of Herpetology. 2005;39:674–678. [Google Scholar]
• Le Rouzic A, Carlborg O. Evolutionary potential of hidden genetic variation. Trends in Ecology & Evolution. 2008;23:33–37. [PubMed] [Google Scholar]
• Ledon-Rettig CC, Pfennig DW, Nascone-Yoder N. Ancestral variation and the potential for genetic accommodation in larval amphibians: implications for the evolution of novel feeding strategies. Evolution & Development. 2008;10:316–325. [PubMed] [Google Scholar]
• Losos JB, Creer DA, Glossip D, Goellner R, Hampton A, Roberts G, Haskell N, Taylor P, Ettling J. Evolutionary implications of phenotypic plasticity in the hindlimb of the lizard Anolis sagrei. Evolution. 2000;54:301–305. [PubMed] [Google Scholar]
• Magalhaes IS, Mwaiko S, Schneider MV, Seehausen O. Divergent selection and phenotypic plasticity during incipient speciation in Lake Victoria cichlid fish. Journal of Evolutionary Biology. 2009;22:260–274. [PubMed] [Google Scholar]
• Mäkinen HS, Cano JM, Merilä J. Genetic relationships among marine and freshwater populations of the European three-spined stickleback (Gasterosteus aculeatus) revealed by microsatellites. Molecular Ecology. 2006;15:1519–1534. [PubMed] [Google Scholar]
• McCairns RJS, Bernatchez L. Adaptive divergence between freshwater and marine sticklebacks: insights into the role of phenotypic plasticity from an integrated analysis of candidate gene expression. Evolution. 2010;64:1029–1047. [PubMed] [Google Scholar]
• McGuigan K, Nishimura N, Currey M, Hurwit D, Cresko WA. Cryptic genetic variation and body size evolution in threespine stickleback. Evolution. 2010a;65:1203–1211. [PubMed] [Google Scholar]
• McGuigan K, Nishimura N, Currey M, Hurwit D, Cresko WA. Quantitative genetic variation in static allometry in the threespine stickleback. Integrative and Comparative Biology. 2010b;50:1067–1080. [PubMed] [Google Scholar]
• Moodie G, Reimchen TE. Phenetic variation and habitat differences in Gasterosteus populations of the Queen Charlotte Islands. Systematic Zoology. 1976;25:49–61. [Google Scholar]
• Morgan CL. On modification and variation. Science. 1896;4:733–740. [PubMed] [Google Scholar]
• Orti G, Bell MA, Reimchen TE, Meyer A. Global survey of mitochondrial DNA sequences in the threespine stickleback: evidence for recent migrations. Evolution. 1994;48:608–622. [PubMed] [Google Scholar]
• Osborn HF. A mode of evolution requiring neither natural selection nor the inheritance of acquired characters. Transactions of the New York Academy of Science. 1896;15:141–142. [Google Scholar]
• Parsons KJ, Robinson BW. Replicated evolution of integrated plastic responses during early adaptive divergence. Evolution. 2006;60:801–813. [PubMed] [Google Scholar]
• Pfennig DW, Wund MA, Snell-Rood EC, Cruickshank T, Schlichting CD, Moczek AP. Phenotypic plasticity’s impacts on diversification and speciation. Trends in Ecology & Evolution. 2010;25:459–467. [PubMed] [Google Scholar]
• Pigliucci M. Do we need an extended evolutionary synthesis? Evolution. 2007;61:2743–2749. [PubMed] [Google Scholar]
• Price TD, Qvarnström A, Irwin DE. The role of phenotypic plasticity in driving genetic evolution. Proceedings of the Royal Society of London, B. 2003;270:1433–1440. [PMC free article] [PubMed] [Google Scholar]
• Reimchen TE. Predators and morphological evolution in threespine stickleback. In: Bell MA, Foster SA, editors. The Evolutionary Biology of the Threespine Stickleback. Oxford: Oxford University Press; 1994. pp. 240–276. [Google Scholar]
• Robinson BW. Trade offs in habitat-specific foraging efficiency and the nascent adaptive divergence of sticklebacks in lakes. Behaviour. 2000;137:865–888. [Google Scholar]
• Robinson BW, Dukas R. The influence of phenotypic modifications on evolution: the Baldwin effect and modern perspectives. Oikos. 1999;85:582–589. [Google Scholar]
• Rohlf FJ. Rotational fit (Procrustes) methods. In: Rohlf FJ, Bookstein FL, editors. Proceedings of the Michigan Morphometrics Workshop. Ann Arbor: University of Michigan Museum of Zoology; 1990. pp. 227–236. Special Publication No. 2. [Google Scholar]
• Rohlf FJ. TPSdig. 2.10 2004. [Google Scholar]
• Schlichting C. Hidden reaction norms, cryptic genetic variation, and evolvability. Annals of the New York Academy of Sciences. 2008;1133:187–203. [PubMed] [Google Scholar]
• Schlichting CD. The role of phenotypic plasticity in diversification. In: DeWitt TJ, Scheiner SM, editors. Phenotypic Plasticity: Functional and Conceptual Approaches. New York: Oxford Univerisity Press; 2004. pp. 191–200. [Google Scholar]
• Schlichting CD, Pigliucci M. Phenotypic Evolution: a Reaction Norm Perspective. Sinauer; Sunderland, MA: 1998. [Google Scholar]
• Schluter D. Adaptive radiation in sticklebacks: size, shape and habitat use efficiency. Ecology. 1993:74. [Google Scholar]
• Schluter D. The Ecology of Adaptive Radiation. Oxford University Press; Oxford: 2000. [Google Scholar]
• Schluter D, Nagel LM. Parallel speciation by natural selection. The American Naturalist. 1995;146:292. [Google Scholar]
• Scoville AG, Pfrender ME. Phenotypic plasticity facilitates recurrent rapid adaptation to introduced predators. Proceedings of the National Academy of Sciences of the U S A. 2010;107:4260–4263. [PMC free article] [PubMed] [Google Scholar]
• Shaw KA, Scotti ML, Foster SA. Ancestral plasticity and the evolutionary diversification of courtship behaviour in threespine sticklebacks. Animal Behaviour. 2007;73:415–472. [Google Scholar]
• Sheets HD. IMP CoordGen6. 2003a. f ed. [Google Scholar]
• Sheets HD. IMP Regress6. 2003b. n ed. [Google Scholar]
• Shingleton AW, Frankino WA, Flatt T, Nijhout HF, Emlen D. Size and shape: the developmental regulation of static allometry in insects. Bioessays. 2007;29:536–548. [PubMed] [Google Scholar]
• Simpson GG. The Baldwin effect. Evolution. 1953;7:110–117. [Google Scholar]
• Sokal RR, Rohlf FJ. Biometry. W. H. Freeman and Company; New York: 1995. [Google Scholar]
• Spoljaric MA, Reimchen TE. 10 000 years later: evolution of body shape in Haida Gwaii three-spined stickleback. Journal of Fish Biology. 2007:70. [Google Scholar]
• Spoljaric MA, Reimchen TE. Habitat-specific trends in ontogeny of body shape in stickleback from coastal archipelago: potential for rapid shifts in colonizing populations. Journal of Morphology. 2011 n/a-n/a. [PubMed] [Google Scholar]
• Stauffer JR, van Snick Gray E. Phenotypic plasticity: its role in trophic radiation and explosive speciation in cichlids (Teleostei: Cichlidae) Animal Biology. 2004;54:137–158. [Google Scholar]
• Taylor EB, McPhail JD. Evolutionary history of an adaptive radiation in species pairs of threespine sticklebacks (Gasterosteus): insights from mitochondrial DNA. Biological Journal of the Linnean Society. 1999;66:271–291. [Google Scholar]
• Tebbich S, Sterelny K, Teschke I. The tale of the finch: adaptive radiation and behavioural flexibility. Philosophical Transactions of the Royal Society of London Series B-Biological Sciences. 2010;365:1099–1109. [PMC free article] [PubMed] [Google Scholar]
• Waddington CH. Genetic assimilation of an acquired character. Evolution. 1953;7:118–126. [Google Scholar]
• Walker JA. Ecological morphology of lacustrine threespine stickleback Gasterosteus aculeatus L. (Gasterosteidae) body shape. Biological Journal of the Linnean Society. 1997;61:3–50. [Google Scholar]
• Walker JA, Bell MA. Net evolutionary trajectories of body shape evolution within a microgeographic radiation of threespine stickleback (Gasterosteus aculeatus) Journal of the Zoological Society of London. 2000;252:293–302. [Google Scholar]
• Webb PW. Body form, locomotion and foraging in aquatic vertebrates. American Zoologist. 1984;24:107. [Google Scholar]
• West-Eberhard MJ. Developmental Plasticity and Evolution. Oxford University Press; New York: 2003. [Google Scholar]
• Willacker J, von Hippel F, Wilton P, Walton K. Classification of threespine stickleback along the benthic-limnetic axis. Biological Journal of the Linnean Society. 2010;101:595–608. [PMC free article] [PubMed] [Google Scholar]
• Wimberger PH. Effects of vitamin C deficiency on body shape and skull osteology in Geophagus brasiliensis: implications for interpretation of morphological plasticity. Copeia. 1993;1993:343–351. [Google Scholar]
• Wimberger PH. Trophic polymorphisms, plasticity, and speciation in vertebrates. In: Strouder DJ, Fresh KL, Feller RJ, editors. Theory and Adaptation of Fish Feeding Ecology. Columbia, S.C: Uinversity of South Carolina Press; 1994. [Google Scholar]
• Withler RE, McPhail JD. Genetic variability in freshwater and anadromous sticklebacks (Gasterosteus aculeatus) of southern British Columbia. Canadian Journal of Zoology. 1985;63:528–533. [Google Scholar]
• Wund MA, Baker JA, Clancy B, Golub J, Foster SA. A test of the ‘flexible stem’ model of evolution: ancestral plasticity, genetic accommodation, and morphological divergence in the threespine stickleback radiation. The American Naturalist. 2008;172:449–462. [PubMed] [Google Scholar]
• Yeh PJ, Price TD. Adaptive phenotypic plasticity and the successful colonization of a novel environment. American Naturalist. 2004;164:531–542. [PubMed] [Google Scholar]
• Zelditch ML, Swiderski DL, Sheets HD, Fink WL. Geometric Morphometrics for Biologists. Elsevier; San Diego: 2004. [Google Scholar]