• adaptive divergence;
  • Centrarchid;
  • feeding behaviour;
  • Lepomis gibbosus;
  • morphometrics;
  • natural selection;
  • phenotypic plasticity;
  • resource polymorphism


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Morphological plasticity can influence adaptive divergence when it affects fitness components such as foraging performance. We induced morphological variation in pumpkinseed sunfish (Lepomis gibbosus) ecomorphs and tested for effects on foraging performance. Young-of-year pumpkinseed sunfish from littoral and pelagic lake habitats were reared each on a ‘specialist diet’ representing their native habitat-specific prey, or a ‘generalist diet’ reflecting a combination of native and non-native prey. Specialist and generalist diets, respectively, induced divergent and intermediate body forms. Specialists had the highest capture success on their native prey whereas generalist forms were inferior. Specialists faced trade-offs across prey types. However, pelagic specialists also had the highest intake rate on both prey types suggesting that foraging trade-offs are relaxed when prey are abundant. This increases the likelihood of a resource polymorphism because the specialized pelagic form can be favoured by directional selection when prey are abundant and by diversifying selection when prey resources are restricted.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Many organisms are phenotypically plastic because they can adjust their phenotype over some or all of their lifetime in response to local environmental conditions. Phenotypic responses that confer fitness advantages in alternate environments indicate that plasticity can be adaptive and can evolve when additive genetic variation in reaction norms is present (West-Eberhard, 1989, 2003). Considerable theoretical understanding of the evolution of adaptive phenotypic plasticity has occurred over the last 20 years (Via & Lande, 1985; Lively, 1986; Gavrilets & Scheiner, 1993), but only recently has its influence on adaptive divergence been considered (reviewed in Losos et al., 2000; Schlichting & Smith, 2002; Pigliucci & Murren, 2003; Price et al., 2003). Phenotypic plasticity potentially influences the rate and direction of evolution by affecting the distribution of phenotypes present in a population, which in turn can influence colonization success and population persistence in novel environments, and the subsequent responses to selection (West-Eberhard, 1989, 2003; Schlichting & Pigliucci, 1998; Pigliucci, 2001; Price et al., 2003; Parsons & Robinson, 2006). To better understand how plastic traits influence adaptive divergence requires tests of how plastic responses diverge between populations, and how induced responses affect performance and ultimately fitness.

Divergent plastic trait responses and their fitness consequences have begun to be studied in plants (Donohue et al., 2000, 2001; Schlichting & Smith, 2002) and invertebrates (Morin et al., 1999; Laurila et al., 2002), but rarely in vertebrates. The freshwater fishes of post-glacial lakes are relevant for such studies because they show considerable phenotypic variation in the form of trophic or resource polymorphisms along an ecological gradient often bounded by littoral and pelagic habitats (Robinson & Wilson, 1994; Skúlason & Smith, 1995; Smith & Skúlason, 1996), large amounts of phenotypic plasticity (Robinson & Parsons, 2002) and relatively rapid evolutionary divergence (Bernatchez & Wilson, 1998). Although morphological plasticity has been widely demonstrated in a number of fishes (e.g. Meyer, 1987; Witte et al., 1990; Day et al., 1994; Hjelm et al., 2001; Pakkasmaa & Piironen, 2001; see Robinson & Parsons, 2002 for review), the effects of induced variation on performance are generally unknown with two exceptions. In the Crucian carp (Carassius carrasius), exposure to the large ambush predator Northern pike (Esox lucius) induces a deeper body form (Bronmark & Miner, 1992). Pike swallow their prey whole and so prey girth relative to predator mouth gape limits capture success and therefore likely contributes to prey fitness (Nilsson et al., 1995). Although some details are still unknown, this system remains the best example of predator-induced defences in fishes. Adaptive plasticity of body form to different diets has also been tested in freshwater threespine stickleback (Gasterosteus aculeatus). In five lakes in British Columbia, Canada, coexisting ‘limnetic’ and ‘benthic’ stickleback species pairs have adaptively diverged in less than 10 000 years (McPhail, 1994; Schluter, 2000). Day et al. (1994) demonstrated that the species have divergent plastic morphological responses to shared diet treatments. Day & McPhail (1996) tested if the diet-induced morphological variation was related to prey-specific foraging efficiency. The induced benthic morphology had improved foraging efficiency on benthic prey relative to fish with induced limnetic body forms as expected if plastic responses were adaptive. Although sparse, such studies show that plastic phenotypic responses in fish can affect both survival and foraging performance (Robinson & Parsons, 2002).

More recently, plastic responses have also been shown to diverge between fish populations possibly because they influence performance and allow phenotypes to more closely match local conditions. For example, coexisting ecomorphs of sunfish have divergent plastic responses to prey resources (Parsons & Robinson, 2006) and also predator cues (Januszkiewicz & Robinson, 2006). Here, the influence of environmental cues on morphological development appears to have evolved under selection because in the former example, evolved reaction norms have converged in three separate populations.

The sunfish resource polymorphism

Numerous lake populations of pumpkinseed sunfish (Lepomis gibbosus) in the Adirondack region of New York state, USA, and the Mazinaw and Kawartha regions of south-eastern Ontario, Canada, exhibit a trophic or resource polymorphism, whereby individuals that inhabit the pelagic habitat appear somewhat specialized for foraging on zooplankton, whereas those that live in the shallow littoral habitat feed primarily on macroinvertebrate prey (Robinson et al., 1993, 2000; Gillespie & Fox, 2003; Jastrebski & Robinson, 2004; McCairns & Fox, 2004). Subtle morphological differences distinguish sunfish from these habitats, with pelagic fish generally having shorter heads, longer overlapping gill rakers, a larger caudal peduncle and maximum body depth displaced forward of that in littoral fish (Jastrebski & Robinson, 2004; Weese, 2004). Variation is continuous with intermediate phenotypes common in both habitats, although extreme forms are only found in their ‘native’ habitat. Similar morphological differences occur between pelagic and littoral fish taxa, suggesting that variation in body form is functionally advantageous (Webb, 1984). The external body form of pumpkinseed sunfish exhibits substantial plastic responses to diet (Robinson & Wilson, 1996; Hegrenes, 1999), and coexisting ecomorphs have divergent plastic responses to the same set of diet treatments (Parsons & Robinson, 2006). This constitutes strong evidence of heritable variation in the plastic responses of polymorphic sunfish. However, we do not know if diet-induced trait responses affect foraging performance in these sunfish. Our goal was to test predictions relating to adaptive morphological plasticity in pumpkinseed sunfish ecomorphs.

Adaptive plasticity evolves when (a) the developmental system has the capacity to produce alternate phenotypes, (b) heritable variation in plasticity exists, (c) reliable cues exist to match phenotype appropriately to environmental conditions, and (d) divergent selection favours alternate phenotypes in different environments because of strong trade-offs in performance (DeWitt et al., 1998). Criteria (a) and (b) have been satisfied in prior studies (Robinson & Wilson, 1996; Hegrenes, 1999; Mittelbach et al., 1999; Parsons, 2002). We expect that criterion (c) is met because benthic macroinvertebrates have profoundly different properties compared with the small zooplankton of the pelagic habitat, such as size and hardness that are experienced by fish during the pursuit, capture and processing of prey. Evidence of adaptive plasticity also allows inference of this criterion.

We test five predictions arising from adaptive plastic morphological responses in pumpkinseed sunfish, focusing on criteria (a) and (d). For (a), we predict (1) that laboratory diets that mimic characteristics of littoral or pelagic prey types will induce divergent body forms. We further predict that feeding sunfish a combination of both prey types will induce body forms intermediate between divergent specialist forms. For (d), we predict that induced phenotypes will be functionally advantageous. In particular, (2) body form responses to diet will conform to functional expectations. Based on prior studies of fish functional morphology and phenotypic plasticity, we expect that sunfish reared on smaller and softer pelagic type prey will have larger eyes, more elongate fusiform bodies, larger caudal peduncles, and longer and deeper head regions relative to sunfish reared on larger and harder littoral-like prey (Webb, 1984; Meyer, 1987; Higham et al., 2006). (3) The two ‘specialized’ phenotypes (e.g. pelagic progeny reared on pelagic-type diets and littoral progeny reared on littoral-type diets) will have the highest foraging performance on their respective native prey reflecting the functional consequences of phenotypic specialization. (4) Specialized phenotypes will experience reduced foraging performance on non-native prey, reflecting functional trade-offs in foraging performance on different prey. (5) The foraging performance of intermediate phenotypes (averaged across prey types) will be inferior to that of the two specialized phenotypes suggesting that diversifying selection, imposed by prey use, favours specialization in this system.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Origins of phenotype classes

We created and compared four phenotype classes using juvenile sunfish sampled from littoral and pelagic (‘source’) habitats of their native lake, split and reared in the laboratory on one of three diets described below. We refer to the phenotype classes as: pelagic specialists (pelagic source fish + pelagic diet), pelagic generalists (pelagic source + mix of pelagic and littoral diets), littoral specialists (littoral source + littoral diet) and littoral generalists (littoral source + mixed diets). We did not rear each source fish type solely on non-native diets because our goals were to produce the most divergent littoral and pelagic phenotypes possible along with biologically realistic intermediate forms that likely exist in the field, and then to compare their relative foraging performances.

Two hundred and eighty young-of-year pumpkinseeds were collected in late July 2000 using minnow traps placed in multiple littoral and pelagic habitats of Lake Ashby, ON, Canada (45°02′N, 77°35′W), where pumpkinseed sunfish are trophically polymorphic (Jastrebski & Robinson, 2004). Fish were estimated to be on average no more than 6 weeks of age based on the time since peak nesting, and were of standard length (20–45 mm), corresponding to ≤ 50 days post-hatch. Fish were transported to the Hagen Aqualab facility at the University of Guelph for rearing under the various diet treatments described below.

Juvenile sunfish were randomly allocated to replicated groups and reared for 6 months (21 August 2000 to 20 February 2001) on one of three possible diets: a ‘pelagic’ diet consisting of live first instar brine shrimp nauplii (Artemia salina), commercial fish flakes, and finely chopped chironomid larvae; a ‘littoral’ diet that consisted of raw liver pieces (approx. 0.5–1.5 cm3 in size) and crushed snails (Larina larina); or a ‘mixed’ diet that consisted of switching between the littoral and pelagic diets on a weekly interval. Diet treatments were presented to whole groups (of 12–36 fish, see Table 1). Each of the four phenotype classes were replicated twice. The eight groups of juvenile fish were reared in fenced half-sections of large rectangular tanks (dimensions of sections: length 250 cm, width 60 cm and depth 30 cm). Homogenous water conditions were maintained among tanks because all shared a single water purification and recirculating system that replaced approximately 10 % of water daily with fresh well water. Water temperature was kept at 22 °C (± 1 °C) and photoperiod on a 12-h light–12-h dark cycle. No cross contamination of diet treatments occurred among groups.

Table 1.   Experimental group sizes for each diet treatment, the prey performance test faced, and the results of a multiple regression of capture success (number of prey captured per bite) against multivariate body shape pooled across diet treatments within a single source and prey test.
Sunfish sourceDiet treatmentForaging performance testPooled regressions
SpecialistGeneralistFd.f.P% variance explained
Littoral1215Amphipod0.54930, 7500.9772.1
Littoral1829Daphnia1.30230, 13500.1282.8
Pelagic3019Amphipod1.02230, 14100.4352.1
Pelagic3620Daphnia1.94930, 16200.0023.5

Variation in foraging performance may result from induced morphological changes or from learned forging behaviour. To reduce the effects of behavioural variation on feeding performance, one replicate group of each phenotype class was randomly selected and acclimated to feed on live cladocerans (Daphnia magna) whereas the other replicate was acclimated with live amphipods (Hyallella azteca) beginning 2 weeks prior to the foraging trials described below. These novel prey, respectively, reflect pelagic and littoral food types found in the diets of wild specimens (Robinson et al., 1993; Gillespie & Fox, 2003; Jastrebski & Robinson, 2004). All feeding was then halted 2 days prior to a foraging trial to standardize and enhance the hunger state of all fish.

Foraging performance trials

Foraging trials tested whether body form was related to foraging ability (predictions 2–5) on live amphipods or daphnia. We chose live daphnia and amphipods because they were found in the diets of wild sunfish populations (Robinson et al., 1993), and as novel prey, should further reduce the effects of learning on foraging performance. Individual fish from each replicate of the four phenotype classes (pelagic specialists, pelagic generalists, littoral generalists and littoral specialists) were observed feeding on the prey type designated in the 2-week pretrial prey acclimation period described above.

In an ‘amphipod’ trial, a single fish (from the replicate group acclimated with amphipods) was placed in a 38-L aquarium containing a sandy substrate, two aquatic plants and fifty H. azteca. For a ‘daphnia’ trial, a single fish (from the replicate group acclimated with daphnia) was placed in a 114-L aquarium containing a gently bubbling airstone and 100 D. magna. Trial duration was 7 min for each ‘amphipod’ trial and 10 min for each ‘daphnia’ trial, starting from the time the fish took its first bite at a prey item. In each trial, we recorded the number of ‘bites’ that the sunfish made at prey defined as a rapid opening and closing of its mouth after prey pursuit. Following a trial, fish were immediately killed in clove oil (100 ppm), blotted wet weight was taken (Denver instruments XP-300 digital scale; Göttingen, Germany) and individuals were fixed in 10 % buffered formalin. One month later, specimens were rinsed in water and stained with alizarin red in a 1 % KOH solution to accentuate morphological structures, and preserved in 70 % ethanol. Following morphometric analysis (described below), stomachs were dissected out and the total number of prey items ingested was counted. Only fish with two or more prey items were deemed informative and were used in our analyses (87 % of the total sample of 208 fish).

We calculated two measures of foraging performance. Capture success was the average number of prey consumed per bite (total number of prey items/total number of bites). Capture rate was the average number of prey caught per minute (total number of prey items/total time foraging). Although these two measures are related, capture success incorporates errors made during foraging (when < 1) whereas capture rate reflects the rate at which prey were consumed. Capture success has the added advantage of being less influenced by prey density whereas capture rate likely has a positive functional response to prey density and so reflects performance relative to the specific prey densities used here.

Behavioural data from both sets of trials were generally normal and were first analysed separately. Variation in performance among phenotype classes (predictions 3–5) was analysed using a one-factor ancova with fish size as the covariate (blotted wet weight in grams) and phenotype class as a factor (four phenotypes). Slopes were all homogenous among phenotype classes in all analyses (interaction between size and treatment effects: all P > 0.18), and so this interaction was subsequently removed for comparisons of phenotype classes. Size-adjusted mean foraging performance is reported for each treatment group.

Trade-offs in foraging performance on different prey types raises the possibility that the mean foraging performance over prey types of generalists is so far below that of specialized forms that diversifying selection would strongly favour specialization (i.e. a concave ‘fitness set’sensuLevins, 1962). We tested for a nonlinear (concave)-shaped performance function, corresponding to a one-sided alternate hypothesis that mean generalist performance averaged over prey types was significantly below the straight line connecting the mean performances of the two specialist forms, using a likelihood ratio test applied to independent data from reciprocal transplant experiments that include intermediate forms (O'Hara Hines et al., 2004). The performance data of fish in the two generalist phenotype classes were pooled for this analysis. The null hypothesis of collinearity among the three bivariate mean values is rejected for large values of the test-statistic, L, which reflects the size of the displacement of the generalist mean away from the straight line connecting the mean values of the specialized forms. The test-statistic was evaluated using a χ2 approach (O'Hara Hines et al., 2004).

Quantifying body shape

External body form was assessed using the geometric morphometric method of thin-plate spline (TPS), which can detect subtle variation in morphology (Bookstein, 1991). One advantage of the TPS technique is that it implements D'Arcy Thompson's concept of Cartesian grid deformations (Thompson, 1917), which allows a more holistic visual interpretation of shape differences among different groups (Rohlf & Marcus, 1993; Parsons et al., 2003).

Stained specimens along with a reference scale were photographed using a Nikon 950 digital camera (Nikon Corp., Tokyo, Japan) with a telephoto lens mounted on a standard camera stand. To prevent curvature in body form that can result from preservation specimens were pinned to a wax tray in a straightened position. Seventeen homologous landmarks (Fig. 1) were digitized from each image using tpsDig (Rohlf, 2001a). The Cartesian grid coordinates of all landmarks were then analysed using tpsRelw (Rohlf, 2001b), which performs a generalized procrustes adjustment procedure (GPA) that translates, rotates and scales all specimens to a common orientation and size (Bookstein, 1991). A consensus (or average) configuration was then calculated from the standardized data set and used as a benchmark for comparison against each individual. Partial warps estimated for each individual using tpsRelw (Rohlf, 2001b) reflect the amount and direction of bending energy required to change the consensus form to that of a given individual. Variation in warp scores is continuous and can be analysed using standard statistical procedures (Parsons et al., 2003).


Figure 1.  Locations of the 17 homologous landmarks used in all geometric morphometric analyses recorded from digital images of the left side of each sunfish. Refer to this diagram for clarification of landmark locations in the deformation grid diagrams that demonstrate changes in body form.

Download figure to PowerPoint

One advantage of geometric morphometric techniques is that shape variation can be studied independent of size. However, scaling all individuals to a common size does not eliminate variation caused by allometric effects that may vary between phenotype classes. We statistically removed allometric variation from the data set by linearily regressing partial warp scores against geometric centroid size and estimating residual warp variation. All partial warp scores used in subsequent analyses and statistical results refer to this residual shape variation.

Relating variation in body shape to foraging performance

Variation in body form among the four phenotype classes was assessed by summarizing the multivariate geometric shape information of every individual (partial warp scores) using a four-group discriminant function analysis (prediction 2). This approach estimated correlated shape information with the greatest ability to distinguish the four phenotype classes. We generated deformation grids to visualize the plastic responses of source type of sunfish to diet treatment (prediction 1) by regressing the partial warps (residual shape data) of all fish against a ‘dummy’ diet treatment grouping variable using Tpsregr (Rohlf, 2000).

We also tested if variation in foraging performance within a source group was related to body form for each prey type (prediction 3), by performing separate linear regressions of capture success against multivariate shape for four subsets of the data using Tpsregr (Rohlf, 2000). The four subsets pooled fish from source and foraging prey test (Table 1), and are: both diet treatments of littoral source fish tested with amphipods (n = 27), both diet treatments of littoral source fish tested with daphnia (n = 47), both diet treatments of pelagic source fish tested with amphipods (n = 49), both diet treatments of pelagic source fish tested with daphnia (n = 56).


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References


Average survival within treatment groups over 6 months was 66%, which is relatively high for experiments that require the collection, handling and rearing of young-of-year sunfish under artificial conditions (e.g. Robinson & Wilson, 1996). Neither rearing diet nor source habitat influenced survival, indicating that shape variation among phenotype classes was not a result of differential mortality (two-factor anova of relative survival: rearing diet effect F2,2 = 0.19, P = 0.83; source habitat effect F1,3 = 0.10, P = 0.77; interaction F3,1 = 0.12, P = 0.95).

Variation in body shape among phenotype classes

Diets induced variation in the body form of juvenile sunfish. The discriminant function analysis (DFA) of the partial warp scores indicated that morphological variation could be used to discriminate between the four phenotype classes (Wilk's λ = 0.33, F90,437 = 2.16, P < 0.001) (Fig. 2). Specimens from the pelagic and littoral specialist groups were at the most extreme ends of the morphological spectrum on the first canonical root axis which represented 65% of the total variation in external body form. Sunfish from the pelagic specialist class had a narrower mid-body, a longer snout, larger buccal region, larger eyes and a deeper caudal peduncle compared with the littoral specialist class (Fig. 3b,d).


Figure 2.  Scatterplot depicting variation in external body form between the four phenotype classes on the first two canonical roots of a discriminant function analysis. The four groups are littoral specialists reared on a littoral diet (solid circles), littoral generalists reared on a mixed diet (open circles), pelagic specialists reared on pelagic diets (solid triangles) and pelagic generalists reared on a mixed diet (open triangles). Interpretations of body form are shown in Fig. 3. The first and second canonical roots explained 65 % and 22 % of the variation in body form respectively.

Download figure to PowerPoint


Figure 3.  Diet-induced variation in the body form of littoral (a and b) and pelagic (c and d) pumpkinseed sunfish. Deformation of the gridlines represents the observed difference in body form between each treatment group compared with the mean form. Littoral source fish were grown on a mixed generalist (a) and littoral specialist (b) diet treatment. Pelagic source sunfish were grown on either a mixed generalist (c) or pelagic specialist (d) diet. Deformation grids were estimated by regressing all partial warp scores against a ‘dummy’ treatment variable. For clarity, shape variation has been exagerated 3×.

Download figure to PowerPoint

A post hoc analysis of mean-squared Malhalanobis distances on the first two canonical axes revealed that the greatest pair-wise group differences were between the pelagic and littoral specialist classes (Bonferonni adjusted P = 0.01), followed by the pelagic specialist–littoral generalist, pelagic specialist–pelagic generalist comparisons (both P = 0.01). No other paired contrasts were significant. This indicates that the mixed diet treatment induced a body form closer to the littoral specialist than to the pelagic specialist phenotype classes, and also that the mean body form of the two generalist classes did not significantly differ from each other.

Relating body shape to foraging performance among phenotype classes

If diet-induced variation in body form is adaptive, then form should influence function and variation in body form should be related to prey-specific foraging performance. We found stronger evidence in support of this prediction for capture success than for capture rate (Fig. 4a,b). Pelagic specialists were the most efficient class at capturing pelagic zooplankton prey as assessed by both measures of performance. Pelagic specialists captured 2.7 more daphnia per minute (34% increase) than the average rate of the other three treatment types (ancovaF3,98 = 4.07, P < 0.01). Capture success was also higher (0.2 more daphnia per bite, 19% higher) for the pelagic specialists than the average of the other three treatment types (ancovaF3,98 = 3.23, P < 0.05). In feeding on amphipods, we found weak evidence that the pelagic specialists also ate on average 0.5 more amphipods per minute (67% more) than the average rate of the other three phenotype classes (ancovaF3,71 = 2.64, P = 0.056). Amphipod capture success seemed highest in littoral specialists (0.17 more amphipods per bite, 28% higher) but this was not a statistically significant difference (ancovaF3,71 = 1.76, P = 0.16). This suggests that other factors besides body form may have influenced amphipod capture.


Figure 4.  Scatter plots displaying, (a) mean capture success (prey individuals captured per bite) and (b) mean capture rate (prey individuals ingested per minute) for each phenotype class and prey type. Error bars represent ± 1 standard error. In (a), note that both specialist phenotype classes experience strong trade-off in capture success across prey types, and that generalist phenotypes have generally poor capture success averaged over both prey types compared with specialist forms. In (b), no trade-off occurs in prey capture rate because the pelagic specialist phenotype class consistently has the highest intake rate on both prey types. The four groups are littoral specialists reared on a littoral diet (solid circles), littoral generalists reared on a mixed diet (open circles), pelagic specialists reared on pelagic diets (solid triangles) and pelagic generalists reared on a mixed diet (open triangles).

Download figure to PowerPoint

Specialized forms experienced strong trade-offs in foraging success across prey types, as shown by the negative relationship in Fig. 4a. We also found strong evidence that the mean success of the generalists averaged over prey types was significantly less than that of either specialist class, reflecting that the performance trade-off function was concave in shape. Mean performance of the two generalist types combined was 0.68 prey captured per bite compared with the specialized littoral and pelagic phenotypes of 0.73 and 0.76 respectively (L = 3.54, P < 0.05). The positive relationship in Fig. 4b indicates that the mean capture rate of pelagic specialists of 4.6 prey per minute was higher than in any other phenotype class (capture rates of littoral specialists, and littoral and pelagic generalists averaged over prey types were 2.9, 3.1 and 3.0 respectively). Thus, although specialist forms experienced trade-offs with respect to foraging success, this did not translate into similar trade-offs in foraging rate at the prey densities used here.

Relating body shape to foraging performance within ecomorphs

Diet-induced body form responses appear to have functional consequences particularly in pelagic ecomorphs reared on pelagic prey, but was feeding performance also related to phenotypic variation within ecomorphs? Patterns of variation in body form that were associated with highest feeding performance on one prey type tended to be associated with lowest performance on the other type of prey. Figure 5 depicts variation in the body form of the most and least efficient foragers (measured as capture success) on each prey type separately for littoral and pelagic source sunfish. As above, the strongest statistical evidence that capture success was related to body form occurred in pelagic source fish tested on daphnia prey, all other regressions were nonsignificant (Table 1). The most efficient pelagic consumers of daphnia had a dorsal-ventrally compressed body, longer snout and larger eyes relative to fish that were less efficient foragers (compare h with g in Fig. 5). Although not statistically significant, the visual evidence suggests that this relationship held for littoral source fish feeding on daphnia as well (compare f with e in Fig. 5). A separate cluster analysis did not divide groups on the basis of the upward and downward bending seen in Fig. 5e–h. Although, there was no statistical evidence for an association between amphipod capture success and body form, the visual deformation grid evidence seems to be reversed with respect to feeding on amphipod prey. The least efficient consumers of amphipods tended to be fish with a more compressed body, longer snout and larger eyes (compare a and c with b and d in Fig. 5). High capture success of amphipod prey was associated with a more robust body, a shortened blunt snout and smaller eyes again generally regardless of the origins of the fish (littoral source fish in a and b and pelagic source fish in c and d of Fig. 5). The visual (but not statistical) consistency of the association between foraging performance and body form regardless of the habitat origins of these sunfish suggests that diet-induced variation in body form improved foraging performance on both prey in both source types of sunfish.


Figure 5.  Variation in morphology related to capture success (estimated as the number of prey consumed per number of bites made). Deformation grids in the left and right columns, respectively, represent body form associated with the lowest and highest capture success on a particular prey type by each phenotype class. Deformation grids (a)–(d) represent morphological variation associated with foraging on amphipods, whereas (e)–(h) represent variation associated with foraging on daphnia. Variation in littoral source fish on amphipod prey is represented by (a) and (b), whereas feeding on daphnia prey is represented by (e) and (f). For pelagic source fish (c) and (d) represent variation associated with feeding on amphipod, whereas (g) and (h) represent variation in body form associated with consuming daphnia prey. Note that the foraging-associated variation shown here mirrors the pattern of diet-induced variation shown in Fig. 3. For clarity, shape variation has been exaggerated 10×.

Download figure to PowerPoint


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Phenotypic plasticity can contribute to biodiversity in two important ways: when a single genotype develops different phenotypes in response to heterogenous conditions, and when plastic developmental systems genetically diverge. Both mechanisms appear to have contributed to the post-glacial evolution of trophic polymorphism in pumpkinseed sunfish. External body form responds strongly to diet (Robinson & Wilson, 1996) and predation cues (Januszkiewicz & Robinson, 2006). Additionally, Parsons & Robinson (2006) found evidence of evolutionary divergent plastic body form responses between that littoral and pelagic sunfish ecomorphs that coexist in Ashby Lake. Pelagic ecomorphs have larger responses to diet, a greater covariation of plastic responses among individual traits, and different patterns of covariation compared with littoral ecomorphs. Similar plastic responses to diet in three populations suggests that some morphological responses have evolved under selection in the pelagic habitat. Our current goal was to test if diet-induced variation in body form can influence feeding performance, and if so, whether the phenotypes induced by prey specialization could be favoured by diversifying selection.

We found that pumpkinseed sunfish exhibited predictable changes in external body form in response to diet (see also Robinson & Wilson, 1996; Hegrenes, 1999). In both source ecomorphs, native specialist diets induced more divergent and specialized body forms compared with sunfish fed on a generalist diet of both prey types which induced similar intermediate body forms. Diet-induced body forms generally conformed to our functional expectations both among and within phenotype classes. Prey capture has three general components: prey search, capture and processing, and morphological variation may influence aspects of all of these. Prey search embodies detection and pursuit, which are expected to be influenced by overall body form to optimize mobility and energetic costs of swimming in different habitats. Comparative studies suggest that a slender body form, larger eye and caudal locomotion are optimal traits in a pelagic predator searching for patchily distributed zooplankton prey whereas deeper body forms and paired fin locomotion are optimal for more cryptic and evasive prey in structured habitats (Webb, 1984). Prey capture in sunfish involves suction feeding where prey are drawn into the mouth in a flow field generated external to the head. However, a trade-off exists between maximizing water velocity and volume in the flow field (Higham et al., 2006) that should translate into shape differences between fish specializing on larger evasive prey like amphipods vs. smaller nonevasive daphnia prey. Optimal shape in amphipod specialists should involve larger/deeper buccal regions and a larger mouth, whereas daphnia specialists should have smaller mouths and an ability to accelerate water quickly. As expected, fish reared on the littoral diet developed smaller eyes, deeper bodies, blunter shorter heads and larger mouths relative to the fish reared on the pelagic-like diet (Fig. 3b,d). Thus, sunfish have plastic developmental systems that can respond appropriately to dietary cues.

Foraging performance depended on an interaction between the induced predator phenotype and prey type. ‘Specialized’ sunfish phenotypes (e.g. pelagic and littoral progeny reared on their native diets) had the highest capture success on their native prey reflecting the benefits of phenotypic specialization. Specialized phenotypes also experienced greatly reduced capture success on non-native prey, reflecting that the benefits of morphological specialization on the native prey type came at the expense of reduced foraging performance on the alternate prey type. The evidence for these effects was stronger for zooplankton compared with amphipod prey. The intermediate body forms induced by a generalist diet experienced severe penalties in mean foraging success compared with that of the specialized phenotypes. If we assume that foraging performance is a surrogate of fitness in these sunfish (Robinson et al., 1996), these results suggest that prey resources can generate diversifying selection that favours specialized body forms. These are some of the first evidences in fishes that alternate-induced phenotypes can be favoured by diversifying selection, an important prerequisite for the evolution of adaptive plasticity. However, polymorphism in pumpkinseed sunfish is not simply a result of ‘classical’ adaptive plasticity whereby a genotype expresses adaptive developmental responses to local conditions in a heterogeneous environment, but instead appears to reflect the genetic divergence of plastic developmental systems between sunfish ecomorphs adapting to different lake habitats (Parsons & Robinson, 2006).

Our results expand on a similar test of adaptive plasticity in ‘limnetic’ form of the threespine stickleback by Day & McPhail (1996), where diet-induced body forms also experienced foraging performance trade-offs between prey types. Closely related ‘benthic’ stickleback forms (analogous to littoral sunfish ecomorphs here) tend to have smaller responses to diet treatments and so phenotypic plasticity has also diverged between stickleback ecomorphs (Day et al., 1994). However, the role of plastic developmental responses in the divergence of sticklebacks is unknown because neither the performance of diet-induced morphotypes in the benthic form has been studied, nor has the feeding performance of diet-induced generalist been compared with that of specialist stickleback forms as we have carried out here. By comparing the foraging performance of a variety of diet-induced morphotypes using these less divergent sunfish ecomorphs, we have been able to demonstrate that plastic trait responses can have functional consequences that contribute to trophic polymorphism.

Three of our results may seem inconsistent with this interpretation. First, no trade-offs in capture rate were detected across prey types because pelagic specialists consistently had the highest intake rate on both prey types. Second, we have little evidence that body form affected the capture success of amphipod prey. Third, why is the pelagic ecomorph not more frequent in these systems? One explanation is that our results reflect the artificial conditions of our study. Trade-offs in prey capture rate may be avoided when prey densities are high, as was likely here. Assessing this possibility would require experiments that manipulate prey densities. Why did phenotype not more strongly influence capture success on amphipod prey? Pumpkinseed sunfish are formidable predators on large hard-bodied prey (Lauder, 1983) and favour snails when available (Mittelbach, 1981, 1988). Feeding-related traits in the pumpkinseed head also exhibited strong plastic responses to the dietary presence or absence of snails (Wainwright et al., 1991). This suggests that foraging on hard-bodied prey requires morphological specialization. Amphipods are a common macroinvertebrate prey of sunfish, but their smaller size and weaker exoskeleton may provide little defence against large fish predators, unlike snails. Instead, amphipods rely on behavioural tactics to hide from fish, and when disturbed use rapid and erratic swimming and hiding to avoid capture (K. J. Parsons, personal observation). The structural simplicity of our aquaria probably hindered the relatively high density of amphipod prey from hiding. Consequently, amphipods were probably fairly easy to capture and variation in sunfish body form may have had little affect on capture success. Alternatively, pelagic specialists may face few trade-offs when consuming certain littoral prey (Robinson & Wilson, 1998). Again, foraging trials with a range of amphipod densities and more realistic littoral conditions that include better refuges may better test the effects of body form on foraging success with amphipod prey. Finally, the magnitude of phenotypic effects were also probably influenced by incomplete or even partially inappropriate plastic responses in the absence of other environmental cues experienced by sunfish under natural conditions (Ehlinger, 1989; Horne & Goldman, 1994), and the short duration of our foraging tests. Nevertheless, our study was able to show that induced variation in body form can affect foraging performance in sunfish.

That pelagic pumpkinseed ecomorphs are not more frequent in nature given the foraging advantages shown here suggests that other factors also influence their abundance. For example, the bioenergetic costs of pelagic ecomorphs may be higher than those of littoral ecomorphs because of greater search costs, or because pelagic ecomorphs have poorer conversion efficiencies (e.g. Trudel et al., 2001; Derome et al., 2006). The foraging advantages of pelagic ecomorphs may be offset by increased risk of predation in pelagic compared with littoral habitats (Houston et al., 1993; Januszkiewicz & Robinson, 2006). Alternatively, pelagic zooplankton resources may not be as abundant or as stable as littoral resources. For example, pelagic plankton diversity and abundance in lakes vary strongly throughout the year (e.g. Ashton, 1985; Salmaso, 1996). Competitor abundance also can reduce the availability of pelagic resources (Robinson et al., 2000). All of these factors may limit the abundance of pelagic pumpkinseed ecomorphs.

Development of body form

There was some evidence that the generalist diet treatment on average induced a body form somewhat closer to that of the littoral specialist form rather than one symmetrically intermediate between divergent specialist phenotypes (although considerable variation occurred in each phenotype class). This suggests that the specialized pelagic body form does not develop as a simple graded response to increasing zooplankton in the diet, but instead may develop more like a course-grained step function when some internal threshold is surpassed (see Hazel et al., 1990). A strong cue, such as committing to complete planktivory, seems to be required to develop a specialized planktivore body form. This developmental strategy may be possible because pelagic specialists appear to suffer less trade-offs in capture success between prey (e.g. phenotype did not appear to influence foraging performance strongly on amphipod prey) and no costs with respect to capture rate. Prior laboratory studies have found that pelagic ecomorphs consume snails as well as littoral ecomorphs (DeWitt et al., 2000). An explanation for only weak performance trade-offs across prey types is that some traits related to planktivory (such as gill raker architecture) may be developmentally or functionally independent of traits used for feeding on macroinvertebrates (e.g. pharyngeal jaw structure). Strong site and habitat-specific fidelity by both ecomorphs (McCairns & Fox, 2004) may also guarantee the strong environmental signal necessary to develop the specialized planktivore form. Thus, many prerequisites exist that would favour the evolution of induced responses that fine-tune phenotype, especially under pelagic conditions.

Selection, trophic divergence and the evolution of morphological plasticity

Various studies of sunfish in post-glacial lakes suggest that selection is stronger in pelagic compared with that in littoral habitats, and that selection is diversifying, favouring different phenotypic optima. First, in a comparative field study involving sunfish from Paradox L., Robinson et al. (1996) found that condition factor and growth rate, both important components of fitness in fishes, were related to body form in fish sampled from pelagic but not littoral habitats. Second, Jastrebski (2001) also found that variation in onset of spawning, mean age at maturity and growth rate were more strongly related to phenotype in pelagic compared with littoral samples of sunfish from Ashby lake. Third, DeWitt et al. (2000) found that littoral and pelagic ecomorphs were equally good at consuming snails, a primarily littoral prey. Our results provide a proximal explanation for the persistent association between phenotype and performance under pelagic but not littoral conditions: the consumption of pelagic prey is influenced more by body form than by feeding on littoral prey.

These results consistently suggest that specialized morphs have higher fitness than intermediate forms particularly in the pelagic habitat and perhaps that different adaptive peaks are present in the littoral and pelagic habitats of many post-glacial lakes (e.g. Schluter, 1995; Robinson et al., 1996; Kondrashov, 1998). Such an adaptive landscape is expected to favour the evolution of trophic polymorphism (Skúlason & Smith, 1995; Robinson et al., 1996). If selection is strongly influenced by prey capture success as measured here, then diversifying selection is more likely because of the substantial inferiority of fish with intermediate body forms. If selection is influenced more by capture rate, say when resources are abundant, then selection should also favour pelagic specialists because of their superior prey intake rate. In either case, an adaptive peak is indicated in the pelagic habitat. It seems likely that selection imposed by resource use in the ancestral littoral habitat (Jastrebski & Robinson, 2004) is weakly stabilizing because pumpkinseed sunfish are already close to their phenotypic optimum in that habitat (Charlesworth et al., 1982; Via, 1993).

Parsons & Robinson (2006) proposed that selection in the pelagic habitat should favour genotypes whose developmental responses to local environmental conditions provide fitness benefits. Referring to our previous hypotheses, this is facilitated in sunfish because: (1) considerable phenotypic variation is induced by diet (and probably other factors), an effect likely amplified by strong habitat fidelity in nature (McCairns & Fox, 2004). This seems plausible as (2) induced phenotypic variation has large foraging consequences that likely affect fitness and promote diversifying selection, and (3) heritable variation exists in the form of plastic responses to diet (Parsons & Robinson, 2006). These attributes may have allowed pumpkinseed sunfish to colonize, persist and adapt to the novel conditions of the pelagic habitat in some post-glacial lakes.

Skúlason & Smith (1995) suggested that flexible behaviour allows a population to first exploit alternate available habitats or resources, and that subsequent environmentally induced variation in morphology could enhance responses to selection promoting adaptive divergence and perhaps species formation. An unclear aspect of this model is how selection sorts genetic variation when a large component of phenotypic variation is environmentally induced. Our work on trophically polymorphic pumpkinseed sunfish offers one solution: populations can adaptively diverge with respect to how developmental genotypes respond to environmental cues (Parsons & Robinson, 2006). Selection can only sort environmentally responsive developmental genotypes, however, if diet-induced variation in body form improves fitness through greater prey consumption, as we have shown here. Thus, the divergent evolution seen here appears not to be a choice between ‘nature’ (genetic divergence) vs. ‘nurture’ (phenotypic plasticity) but rather a combination of the two; the nature of nurture has diverged.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

We are grateful for the assistance provided by J. O'Hara-Hines in testing the collinearity of foraging performances among phenotype classes; K. Peiman, S. Csinos and J. Kurtis in rearing of sunfish and C. Jastrebski in the field collection of juvenile sunfish. This work benefited from discussion with D. Bolnick, W. Cresko, A. Edelsparre, M. Ferguson, B. Kristjánsson, R. McLaughlin, D. Noakes, S. Skúlason, P. Phillips, T. Tunney, P. Wimberger and P. Wright. Also, two anonymous reviewers greatly improved this manuscript. Financial support was supplied by the Natural Sciences and Engineering Research Council of Canada Discovery grants programme, and a Premier's Research Excellence Award of Ontario to BWR.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  • Ashton, P.J. 1985. Seasonality in southern hemisphere freshwater phytoplankton assemblages. Hydrobiology 125: 179190.
  • Bernatchez, L. & Wilson, C.C. 1998. Comparative phylogeography of Nearctic and Palearctic fishes. Mol. Ecol. 7: 431452.
  • Bookstein, F.L. 1991. Morphometric Tools for Landmark Data: Geometry and Biology. Cambridge University Press, Cambridge, UK.
  • Bronmark, C. & Miner, J.G. 1992. Predator-induced phenotypic change in body morphology in crucian carp. Science 258: 13481350.
  • Charlesworth, B., Lande, R. & Slatkin, M. 1982. A neo-Darwinian commentary on macroevolution. Evolution 36: 474498.
  • Day, T. & McPhail, J.D. 1996. The effect of behavioral and morphological plasticity on foraging efficiency in the threespine stickleback (Gasterosteus sp.). Oecologia 108: 380388.
  • Day, T., Pritchard, J. & Schluter, D. 1994. A comparison of two sticklebacks. Evolution 48: 17231734.
  • Derome, N., Duchesne, P. & Bernatchez, L. 2006. Parallelism in gene transcription among sympatric lake whitefish (Coregonus clupeaformis Mitchell) ecotypes. Mol. Ecol. 15: 12391249.
  • DeWitt, T.J., Sih, A. & Wilson, D.S. 1998. Costs and limits of phenotypic plasticity. Trends Ecol. Evol. 13: 7781.
  • DeWitt, T.J., Robinson, B.W. & Wilson, D.S. 2000. Functional diversity among predators of a freshwater snail imposes an adaptive trade-off for shell morphology. Evol. Ecol. Res. 2: 129148.
  • Donohue, K., Messiqua, D., Pyle, E.H., Heschel, M.S. & Schmitt, J. 2000. Evidence of adaptive divergence in plasticity: density and site-dependent selection on shade avoidance responses in Impatiens capensis. Evolution 54: 19561968.
  • Donohue, K., Pyle, E.H., Messiqua, D., Heschel, M.S. & Schmitt, J. 2001. Adaptive divergence in plasticity in natural populations of Impatiens capensis and its consequences for performance in novel habitats. Evolution 55: 692702.
  • Ehlinger, T.J. 1989. Learning and individual variation in bluegill foraging: habitat specific techniques. Anim. Behav. 38: 643658.
  • Gavrilets, S. & Scheiner, S.M. 1993. The genetics of phenotypic plasticity. V. Evolution of reaction norm shape. J. Evol. Biol. 6: 3148.
  • Gillespie, G.J. & Fox, M.G. 2003. Morphological and life-history differentiation between littoral and pelagic forms of pumpkinseed. J. Fish Biol. 62: 10991115.
  • Hazel, W.N., Smock, R. & Johnson, M.D. 1990. A polygenic model for the evolution and maintenance of conditional strategies. Proc. R. Soc. Lond. B Biol. Sci. 242: 181187.
  • Hegrenes, S.G. 1999. Diet-induced phenotypic plasticity of feeding morphology in the Genus Lepomis. PhD thesis, Illinois State University, Normal, IL.
  • Higham, T.E., Day, S.W. & Wainwright, P.C. 2006. Multidimensional analysis of suction feeding performance in fishes: fluid speed, acceleration, strike accuracy and the ingested volume of water. J. Exp. Biol. 209: 27132725.
  • Hjelm, J., Svanback, R., Bystrom, P., Persson, L. & Wahlstrom, E. 2001. Diet dependant body morphology and ontogenetic reaction norms in Eurasian perch. Oikos 95: 311323.
  • Horne, A.J. & Goldman, C.R. 1994. Limnology, 2nd edn. McGraw-Hill, Toronto, ON, Canada, 576 pp.
  • Houston, A.I., McNamara, J.M. & Hutchinson, J.M.C. 1993. General results concerning the trade-off between gaining energy and avoiding predation. Philos. Trans. Biol. Sci. 341: 375397.
  • Januszkiewicz, A. & Robinson, B.W. 2006. Divergent walleye (Sander vitreus) mediated inducible defenses in the Centrarchid pumpkinseed sunfish (Lepomis gibbosus). Biol. J. Linn. Soc. In press.
  • Jastrebski, C.J. 2001. Divergence and selection in trophically polymorphic pumpkinseed sunfish (Lepomis gibbosus). MSc Thesis, University of Guelph, Guelph, ON, Canada.
  • Jastrebski, C.J. & Robinson, B.W. 2004. Natural selection and the evolution of replicated trophic polymorphisms in pumpkinseed sunfish (Lepomis gibbosus). Evol. Ecol. Res. 6: 285305.
  • Kondrashov, A.S. 1998. Measuring spontaneous deleterious mutation process. Genetica 102/103: 183197.
  • Lauder, G.V. 1983. Functional and morphological bases of trophic specialization in sunfishes (Teleostei: Centrarchidae). J. Morphol. 178: 121.
  • Laurila, A., Karttunen, S. & Merilä, J. 2002. Adaptive phenotypic plasticity and genetics of larval life histories in two Rana temporia populations. Evolution 56: 617627.
  • Levins, R. 1962. Theory of fitness in a heterogeneous environment. 1. The fitness set and adaptive function. Am. Nat. 96: 361373.
  • Lively, C.M. 1986. Canalization versus developmental conversion in a spatially variable environment. Am. Nat. 128: 561572.
  • Losos, J.B., Creer, D.A., Glossip, D., Goellner, R., Hampton, A., Roberts, G., Haskell, N., Taylor, P. & Ettling, J. 2000. Evolutionary implications of phenotypic plasticity in the hindlimb of the lizard Anolis sagrei. Evolution 54: 301305.
  • McCairns, R.J.S. & Fox, M.G. 2004. Habitat and home range fidelity in a trophically dimorphic pumpkinseed sunfish (Lepomis gibbosus) population. Oecologia 140: 271279.
  • McPhail, J.D. 1994. Speciation and the evolution of reproduction isolation in the sticklebacks (Gasterosteus) of southwestern British Columbia. In: The Evolutionary Biology of the Threespined Stickleback. (M. A.Bell & S. A.Foster, eds), pp. 399426. Oxford Science Publications, Oxford.
  • Meyer, A. 1987. Phenotypic plasticity and heterochrony in Cichlasoma managuense (Pisces, Cichlidae) and their implications for speciation in cichlid fishes. Evolution 41: 13571369.
  • Mittelbach, G.G. 1981. Foraging efficiency and body size: a study of optimal diet and habitat use by bluegills. Ecology 62: 13701386.
  • Mittelbach, G.G. 1988. Competition among refuging sunfishes and effects of fish density on littoral zone invertebrates. Ecology 69: 614623.
  • Mittelbach, G.G., Osenberg, C.W. & Wainwright, P.C. 1999. Variation in feeding morphology between pumpkinseed populations: phenotypic plasticity or evolution? Evol. Ecol. Res. 1: 111128.
  • Morin, J.P., Moreteau, B., Petavy, G. & David, J.R. 1999. Divergence of reaction norms of size characters between tropical and temperate populations of Drosophila melanogaster and D. simulans. J. Evol. Biol. 12: 329339.
  • Nilsson, P.A., Bronmark, C. & Pettersson, L.B. 1995. Benefits of a predator-induced morphology in crucian carp. Oecologia 104: 291296.
  • O'Hara Hines, R.J., Hines, W.G.S. & Robinson, B.W. 2004. A new statistical test of fitness set data from reciprocal transplant experiments involving intermediate phenotypes. Am. Nat. 163: 97104.
  • Pakkasmaa, S. & Piironen, J. 2001. Water velocity shapes juvenile salmonids. Evol. Ecol. 14: 721730.
  • Parsons, K.J. 2002. Morphological plasticity as a factor in the evolution of trophically polymorphic pumpkinseed sunfish (Lepomis gibbosus). MSc thesis, University of Guelph, Guelph, ON, Canada.
  • Parsons, K.J., Robinson, B.W. & Hrbek, T. 2003. Getting into shape: an empirical comparison of traditional truss-based morphometric methods with a newer geometric method applied to New World cichlids. Environ. Biol. Fishes 67: 417431.
  • Parsons, K.J. & Robinson, B.W. 2006. Replicated evolution of integrated plastic responses during early adaptive divergence. Evolution 60: 801813.
  • Pigliucci, M. 2001. Phenotypic Plasticity: Beyond Nature and Nurture. John Hopkins University Press, Baltimore, MD.
  • Pigliucci, M. & Murren, J. 2003. Genetic assimilation and a possible evolutionary paradox: can macroevolution sometimes be so fast as to pass us by? Evolution 57: 14551464.
  • Price, T.D., Qvarnstrom, A. & Irwin, D.E. 2003. The role of phenotypic plasticity in driving genetic evolution. Proc. R. Soc. Lond. B 270: 14331440.
  • Robinson, B.W. & Parsons, K.J. 2002. Changing times, spaces, and faces: tests and implications of adaptive morphological plasticity in the fishes of northern postglacial lakes. Can. J. Fish. Aquat. Sci. 59: 18191833.
  • Robinson, B.W. & Wilson, D.S. 1994. Character release and displacement in fishes: a neglected literature. Am. Nat. 144: 596627.
  • Robinson, B.W. & Wilson, D.S. 1996. Genetic variation and phenotypic plasticity in a trophically polymorphic population of pumpkinseed sunfish (Lepomis gibbosus). Evol. Ecol. 10: 122.
  • Robinson, B.W. & Wilson, D.S. 1998. Optimal foraging, specialization, and a solution to Liem's paradox. Am. Nat. 151: 223235.
  • Robinson, B.W., Wilson, D.S., Margosian, A.S. & Lotito, P. 1993. Ecological and morphological differentiation of pumpkinseed sunfish in lakes without bluegill sunfish. Evol. Ecol. 7: 451464.
  • Robinson, B.W., Wilson, D.S. & Shea, G.O. 1996. Trade-offs of ecological specialization: an intraspecific comparison of pumpkinseed sunfish phenotypes. Ecology 77: 170178.
  • Robinson, B.W., Wilson, D.S. & Margosian, A.S. 2000. A pluralistic analysis of character release in pumpkinseed sunfish (Lepomis gibbosus). Ecology 81: 27992812.
  • Rohlf, F.J. 2000. Tps Regression v.1.2. Free Computer Software for Exploring the Relationship between Shape and One or More Independent Variables. Ecology and Evolution, SUNY at Stony Brook. URL
  • Rohlf, F.J. 2001a. Tps Dig v.1.28. Free Computer Software for Collecting Landmark Data from Images. Ecology and Evolution, SUNY at Stony Brook. URL
  • Rohlf, F.J. 2001b. Tps Relative Warps v.1.22. Free Computer Software for Calculating Partial and Relative Warp Scores from Landmark Data. Ecology and Evolution, SUNY at Stony Brook. URL
  • Rohlf, F.J. & Marcus, L.F. 1993. A revolution in morphometrics. Trends Ecol. Evol. 8: 129132.
  • Salmaso, N. 1996. Seasonal variation in the composition and rate of change of the phytoplankton community in a deep subalpine lake (Lake Garda, Northern Italy). An application of nonmetric multidimensional scaling and cluster analysis. Hydrobiology 337: 4968.
  • Schlichting, C.D. & Pigliucci, M. 1998. Phenotypic Evolution: A Reaction Norm Perspective. Sinauer Associates Inc., Sunderland, MA.
  • Schlichting, C.D. & Smith, H. 2002. Phenotypic plasticity: linking molecular mechanisms with evolutionary outcomes. Evol. Ecol. 16: 189211.
  • Schluter, D. 1995. Adaptive radiation in sticklebacks: trade-offs in feeding performance and growth. Ecology 76: 8290.
  • Schluter, D. 2000. The Ecology of Adaptive Radiation. Oxford University Press, Oxford.
  • Skúlason, S. & Smith, T.B. 1995. Resource polymorphisms in vertebrates. Trend. Ecol. Evol. 10: 366370.
  • Smith, T.B. & Skúlason, S. 1996. Evolutionary significance of resource polymorphisms in fish, amphibians and birds. Annu. Rev. Ecol. Syst. 27: 111133.
  • Thompson, D.W. 1917. On Growth and Form. Cambridge University Press, Cambridge.
  • Trudel, M., Tremblay, A., Schetagne, R. & Rasmussen, J.B. 2001. Why are dwarf fish so small? An energetic analysis of polymorphism in lake whitefish (Coregonus clupeaformis). Can. J. Fish. Aquat. Sci. 58: 394405.
  • Via, S. 1993. Adaptive phenotypic plasticity: target or by-product of selection in a variable environment? Am. Nat. 142: 352365.
  • Via, S. & Lande, R. 1985. Genotype-environment interaction and the evolution of phenotypic plasticity. Evolution 39: 502522.
  • Wainwright, P.C., Osenberg, S.W. & Mittelbach, G.G. 1991. Trophic polymorphism in the pumpkinseed sunfish (Lepomis gibbosus Linnaeus): effects of environment on ontogeny. Funct. Ecol. 5: 4055.
  • Webb, P.W. 1984. Body form, locomotion and foraging in aquatic vertebrates. Am. Zool. 24: 107120.
  • Weese, D. 2004. Tests of replicated phenotypic divergence in polymorphic pumpkinseed sunfish (Lepomis gibbosus), using microsatellite markers. MSc thesis, University of Guelph, Guelph, ON, Canada.
  • West-Eberhard, M.J. 1989. Phenotypic plasticity and the origins of diversity. Annu. Rev. Ecol. Syst. 20: 249278.
  • West-Eberhard, M.J. 2003. Developmental Plasticity and Evolution. Oxford University Press, Oxford. 794. pp.
  • Witte, F., Barel, C.D.N. & Hoogerhoud, R.J.C. 1990. Phenotypic plasticity of anatomical structures and its ecomorphological significance. Neth. J. Zool. 40: 278298.