Research article | Open | Published:
The temporal window of ecological adaptation in postglacial lakes: a comparison of head morphology, trophic position and habitat use in Norwegian threespine stickleback populations
BMC Evolutionary Biologyvolume 16, Article number: 102 (2016)
Studying how trophic traits and niche use are related in natural populations is important in order to understand adaptation and specialization. Here, we describe trophic trait diversity in twenty-five Norwegian freshwater threespine stickleback populations and their putative marine ancestor, and relate trait differences to postglacial lake age. By studying lakes of different ages, depths and distance to the sea we examine key environmental variables that may predict adaptation in trophic position and habitat use. We measured trophic traits including geometric landmarks that integrated variation in head shape as well as gillraker length and number. Trophic position (Tpos) and niche use (α) were estimated from stable isotopes (δ13C, δ15N). A comparison of head shape was also made with two North American benthic-limnetic species pairs.
We found that head shape differed between marine and freshwater sticklebacks, with marine sticklebacks having more upturned mouths, smaller eyes, larger opercula and deeper heads. Size-adjusted gillraker lengths were larger in marine than in freshwater stickleback. Norwegian sticklebacks were compared on the same head shape axis as the one differentiating the benthic-limnetic North American threespine stickleback species pairs. Here, Norwegian freshwater sticklebacks with a more “limnetic head shape” had more and longer gillrakers than sticklebacks with “benthic head shape”. The “limnetic morph” was positively associated with deeper lakes. Populations differed in α (mean ± sd: 0.76 ± 0.29) and Tpos (3.47 ± 0.27), where α increased with gillraker length. Larger fish had a higher Tpos than smaller fish. Compared to the ecologically divergent stickleback species pairs and solitary lake populations in North America, Norwegian freshwater sticklebacks had similar range in Tpos and α values, but much less trait divergences.
Our results showed trait divergences between threespine stickleback in marine and freshwater environments. Freshwater populations diverged in trophic ecology and trophic traits, but trophic ecology was not related to the elapsed time in freshwater. Norwegian sticklebacks used the same niches as the ecologically divergent North American stickleback species pairs. However, as trophic trait divergences were smaller, and not strongly associated with the ecological niche, ecological adaptations along the benthic-limnetic axis were less developed in Norwegian sticklebacks.
Adaptive radiations in which there is a close association between niche use and trophic traits have intrigued scientists for centuries [1, 2]. Here, subtle to strong trophic specialization and habitat use associations are indicators of early to late adaptive steps in taxa such as European whitefish (Coregonus lavaretus) [3, 4] and threespine stickleback (Gasterosteus aculeatus) . However, the time scale for adaptation in wild populations is usually not known. Further, not all taxa show adaptive radiations. Thus, it is important to study mechanisms leading to adaptive radiations, ideally where the time scale for natural selection is known or can be inferred.
In the northern hemisphere post-glacial isostatic rebound has created freshwater lakes of varying ages providing an upper bound on timing of colonization and how long immigrants could be adapting to the lake environment . The ancestral marine threespine stickleback was one such taxon invading freshwater systems as they became available after the retreat of the Pleistocene ice-sheet, representing a remarkable example of parallel evolution and adaptive radiation into these new freshwater environments [7–9]. Many attempts have been made to understand the selective landscape leading to the diversification of freshwater sticklebacks [8–11]. Current adaptations in freshwater sticklebacks may stem from standing genetic variation in the marine population, from selective colonization of individuals preadapted to freshwater, or from new adaptive mutations and selective processes that have arisen post-glacially (0–20 000 years before present) [12, 13]. Few studies have specifically addressed the origin and rate of novel adaptive phenotypic diversity during this postglacial process. These studies show that phenotypic change may occur rapidly even on a contemporary time scale [14–16]. Strong selection on threespine stickleback phenotypes in new environments has been documented [17, 18]. Furthermore, evolutionary constraints due to allometric scaling relationships have been suggested to partly shape phenotypic diversification in postglacial Norwegian freshwater threespine sticklebacks , setting a limit to the extent and rate of phenotypic change.
Even though estimates on trait diversification and selection in threespine sticklebacks exists we still need information on trophic ecology associated with trait changes to understand if and how ecological conditions may affect the rate of adaptive diversification in freshwater lakes. Few comparative studies of trophic transitions between marine and freshwater environments are available for the threespine stickleback. However, the ancestral marine sticklebacks are assumed to be pelagic foragers in a homogeneous environment [8, 20], while anadromous sticklebacks could also spend a large proportion of the year in the ocean feeding on pelagic zooplankton . Conversely, freshwater sticklebacks are assumed to live in a much more heterogeneous environment than populations in marine habitats and to feed on various prey sources in the littoral and pelagic area, e.g. . Thus, divergent niche use between marine and freshwater stickleback, and among freshwater stickleback populations, may lead to selection for different adaptive trophic morphologies in these contrasting environments [1, 10, 22].
Most Holarctic stickleback populations usually contain trophic generalist individuals. In such solitary generalist populations (i.e. one trophic form of stickleback present), phenotypic differentiation along the benthic-limnetic axis (i.e. putative morphological adaptation from littoral-benthic to pelagic areas) can be observed both within and among populations [23, 24]. In a small subset of lakes, sticklebacks have diverged into benthic and limnetic morphs (referred to as species pairs) [1, 25]. Here, contrasting adaptations have accompanied occupation of benthic and limnetic niches with ecological selection against intermediate phenotypes [26, 27]. The limnetic morph forages mainly in the pelagic habitat, having a long fusiform body, large eyes, a long slender head, a terminal mouth and many long gillrakers [7, 21]. In contrast, the benthic morph uses a more spatially complex environment, feeds on larger invertebrates, and has a deeper body and head, a shorter and wider mouth, and fewer and shorter gillrakers [7, 27]. Interestingly, the phenotypic diversification in the solitary lake populations of the threespine stickleback seem to resemble trophic adaptations in the benthic-limnetic species pair [7, 27], with variable position along this axis in different lakes suggesting ongoing trophic adaptation. Matthews et al.  used stable isotopes to compare generalist and benthic-limnetic stickleback morphs/species pairs, and found that divergence in habitat use coincided with shifts in trophic position. The limnetic morph had a higher trophic position than the benthic morph. Sticklebacks with longer gillrakers obtained less carbon from benthic sources. Matthews et al.  implied that gillraker length was a key trait permitting adaptive niche shifts among lake habitats. Gillraker length is associated with food preference, prey items and stable isotopes in sticklebacks , as may be the case for gillraker numbers but to a lesser extent . Thus, freshwater stickleback populations display a wide range of phenotypes ranging from generalists to specialized forms that exploit the benthic-limnetic niche axis, providing an excellent model for studies on rapid post-glacial trophic adaptations.
Individual freshwater systems differ considerably in terms of their morphology and abiotic and biotic conditions, and thus the shape and intensity of natural selection on trophic traits and foraging behavior in the threespine stickleback differ among systems. Individuals in generalist threespine stickleback populations with phenotypes more adapted to the benthic or limnetic niche seem to have higher fitness than those with intermediate morphology [30, 31]. In such, “optimal phenotypes” in a freshwater system may depend on lake morphometry and niche-specific availability of benthic and limnetic prey . Here, intraspecific resource competition can be an important driver for foraging-related phenotypic divergence and niche variation . This adaptive process will likely be affected by piscivorous predators - setting the frame for ecological opportunity [25, 30–32].
In this study, we used a set of twentyfive freshwater and one marine threespine stickleback populations to compare trophic phenotypes between environments in Norway. The main goal was to test specifically if, and to what degree, the freshwater sticklebacks showed trophic trait adaptations to lake niches in the timeframe following the Pleistocene deglaciation. This task was accomplished by first comparing adaptive trophic trait variation (head shape, gill raker number and gill raker length) between the marine and freshwater stickleback populations. Secondly, we tested if and how this trophic trait variation was associated with trophic ecology in freshwater lakes, particularly looking for association with the age of the postglacial lakes. Further, we tested if trophic ecology was associated with the age of postglacial lakes. More specifically, geometric morphometrics were used to characterise head shape, and we counted the number of gillrakers, and measured gillraker length. We used nitrogen stable isotope values (δ15N) of littoral and pelagic invertebrates to estimate individual and population trophic position (Tpos), and carbon stable isotope values (δ13C) to estimate the proportion of dietary carbon from littoral sources (α) as an estimate of long-term habitat use. Furthermore, lake parameters (estimated mean depth and distance to the sea), and estimates of freshwater lake age (when lakes were formed after deglaciation – due to specific isostatic uplift patterns) were used for testing for associations between morphology, habitat use (α) and trophic position (Tpos). We further compared the Norwegian threespine stickleback populations to two sets of the highly specialized benthic and limnetic species pair from North American lakes. Finally, we tested if the stickleback populations in these Norwegian lakes showed similar trait diversity as in the benthic-limnetic species pairs and illuminated how niche use was associated with morphology in Norwegian and a set of North American stickleback populations. Thus, our study approach was to test trophic trait divergence between marine and freshwater lakes, with an emphasis on how trophic ecology and trophic trait associations were potentially modulated by lake specific parameters in the available time frame for evolution in Norwegian postglacial lakes. The comparison between Norwegian sticklebacks and the North American benthic-limnetic species pairs was done to describe the extent of diversification, and the degree of similarity, in associations between trophic traits, niche use and trophic position in a continuum of generalist to species-pair stickleback lakes on these two continents.
Study area and sample collection
During August 2006, threespine sticklebacks were sampled from 26 populations (one marine coastal sample, 25 freshwater lakes) from the Island of Andøya and from different islands in the Lofoten archipelago, NW Norway (Fig. 1, Table 1). Marine fish were collected using small mesh-sized gillnets in the Andenes harbour. Freshwater sticklebacks were collected using a standardized set of plexiglass traps set in the shallow (< 2 m depth) littoral. After sampling, fish were stored in 70 % EtOH after being euthanized using an overdose of MS222.
For each lake we estimated lake surface area (A; km2) from maps and used this to estimate the mean depth based on an empirical model of data from 488 Norwegian lakes . First, lake volume (V) was estimated as V = 0.011A1.4. Mean depth (D; m) was then estimated as D = 1000*V/A e.g. . As area and depth parameters are not independent, we used only estimated mean depth for further statistical analyses. Information on the presence of fish predators was obtained from local fishers organizations. Both piscivorous fish and birds may predate on threespine stickleback, and in such impose selection pressures affecting niche use in sticklebacks, but information on birds predating on sticklebacks do not exist in these lakes. Thus, we have used piscivore fish as a general description of predation regime. The piscivore predators in this part of Norway are Atlantic salmon (Salmo salar), brown trout (Salmo trutta), arctic charr (Salvelinus alpinus) and European eel (Anguilla anguilla). Brown trout were present in all lakes except one (Einletvatn Pond) (Table 1).
All lakes in the region were formed following postglacial isostatic uplift. The age of each lake was estimated using the Sea level 3.51 software  (Table 1), and found to vary from 3 100 to 18 700 years before present (ybp). For five lakes situated above the last glacial marine limit, we used glacial isochrones based on the recession of the ice margin  to estimate the upper bounds for the putative timing of stickleback colonization of a lake. As we do not have information regarding exact timing of stickleback colonization, lake age sets an upper time frame for colonization. We assume that marine stickleback retain their ancestral morphological state and assign this population an age of zero. Marine sticklebacks tend to have relatively similar morphology while brackish and freshwater stickleback often differ [21, 36, 37].
In order to visualize the extent of phenotypic diversity in the Norwegian sticklebacks we compared our marine and freshwater populations to the most phenotypically and ecologically divergent sympatric benthic-limnetic species pairs known globally, using Paxton and Priest Lake, British Colombia, Canada . Secondly, this approach visualized how far Norwegian stickleback populations had diverged along the phenotypic-ecological benthic-limnetic axis compared to the North American benthic-limnetic species pairs. Furthermore, another aim was to describe if divergence along this axis in Norwegian freshwater sticklebacks resembled the phenotypic niche divergence seen in the North American benthic-limnetic species pairs. Here, we used head shape as an axis of comparison (see detailed description below). The sticklebacks from Priest and Paxton lakes were collected using minnow traps set in the littoral zone during May-June 2005 . Paxton Lake fish included 37 limnetic morph individuals (24 males and 13 females) and 27 benthic morph individuals (15 males and 12 females), whereas the dataset from Priest Lake comprised 24 limnetic morph individuals (24 males and 0 females) and 40 benthic morph individuals (27 males and 13 females).
Trophic ecology and habitat use: stable isotope analysis
For the Norwegian freshwater sticklebacks we assessed diet and niche use with stable isotope analysis (SIA) of carbon (δ13C) and nitrogen (δ15N) ratios. Stickleback δ13C values provide information on use of prey along the limnetic-littoral (benthic) resource axis with 13C-depleted values reflecting utilization of limnetic prey and 13C-enriched values indicating consumption of littoral (benthic) prey. These isotopes are typically correlated with the long-term diet e.g. . Furthermore, δ15N values provide information on the trophic level of stickleback . No information is available for nitrogen isotope turnover in sticklebacks, but Grey  showed that δ13C data generally reflect the dietary/feeding history during the last 6 months. This has also been found for other temperate fishes .
From each Norwegian freshwater stickleback a piece of the tail muscle was extracted and dried for 24 h at 60 °C, ground, weighed, encapsulated in tin cups and analysed with a stable isotope ratio mass spectrometer following Harrod et al. . Muscle C:N ratios differed among populations (ANOVA: F25, 488 = 61.3, P < 0.0001), indicating variation in lipid concentrations. Thus, the δ13C data were arithmetically lipid-normalised before further analysis . To provide information on isotopic relationships of limnetic and benthic (littoral) resources from each lake, we sampled zooplankton in the limnetic zone using standard plankton nets and littoral (benthic) chironomids using standardized kick sampling with a hand-held dip net . Zooplankton and chironomids were treated as two groups although the species composition within and between lakes may differ. The isotopic endpoints were estimated from the mean δ13C values from littoral/benthic and limnetic prey from each lake, and trophic fractionation in carbon (δ13C) was estimated at 1‰ . We further calculated the trophic position (Tpos) and the proportion of carbon obtained from littoral sources (α) applying formulas 1 and 2 in Post  (also used in ).
The measures we used to compare the trophic position (Tpos) and habitat use (α) of sticklebacks from Norway and North America include corrections for site-specific differences in isotopic baselines e.g. of littoral and open water habitats, allowing direct comparison of results from the two continents. Stable isotopic data (population means and standard deviations) for North American stickleback populations were found in Matthews et al. .
Norwegian marine and freshwater sticklebacks were sexed by internal examination of gonads. Total fish length was measured to the nearest mm, the number of gillrakers on the first right gill arch was counted, and the length of the third gillraker on the lower gill arch was measured using a stereo microscope. These traits, together with head shape, are important for food acquisition in sticklebacks [46–48], and often correlate with stomach content [28, 31], stable isotope values , and foraging efficiency on divergent prey items see [49–51].
A digital image was captured of the lateral side of the head of Norwegian stickleback individuals when immersed in 70 % EtOH using a CANON EOS 350D digital camera with a 90 mm lens (Tamron macro). D. Schluter provided a comparative dataset of pictures for North American sticklebacks. The Norwegian and North American pictures have been gathered by two different persons (K. Østbye and by courtesy of R. Svanbäck). In order to quantify and characterize the variation in head shape among sticklebacks we first applied a geometric morphometric analysis of the pictures using the thin plate spline (Tps) morphometric software TspDig2 . Here, 20 landmarks (Fig. 2) were selected to describe the integrated shape of the head and digitized. All the landmarks were set by one person (K. Østbye). The landmark data were subsequently imported into the software MorphoJ 1.01b [53, 54] where we performed a Procrustes fit and secondly performed a principal component analysis (PCA) saving the most important principal components for our analyses (see below).
First, we tested if a set of trophic traits (head shape - based on the extracted principal components, gillraker number and gillraker length) in Norwegian threespine sticklebacks differed between the marine and freshwater habitats using a general linear mixed model (glmm) with habitat (marine/freshwater), body length and sex as fixed effects and lake as a random factor. This analysis was conducted to visualise potential trophic trait differences between the assumed ancestral marine and the derived freshwater populations.
Secondly, we quantified trophic trait diversification among all the Norwegian freshwater threespine stickleback populations. Head shape was analysed using principal components and the most important PC-axes were extracted. Then, we used glmm to examine how variation in principal component traits, as well as variation in gillraker number and length, was predicted by individual variation in two related habitat use proxies (the proportion of somatic carbon derived from littoral sources (α) and trophic position (Tpos)) and fish characteristics (body length (cm) and sex). Lake identity was used as a random factor in the analysis.
Further, we used glmm to test how trophic position (Tpos) and the proportion of carbon from littoral sources (α) varied among Norwegian freshwater populations, and how these measures were related to lake age, mean depth of lakes, and distance to the sea, using sex and body length as additional predictors. Lake identity was used as a random factor in the analysis.
To visualise diversity in head shape we compared the Norwegian marine and freshwater populations with the benthic-limnetic species pairs in North America using a joint principal component analysis. This analysis produced two informative principal components, where PC1 differentiated the Norwegian and North American stickleback while PC2 differentiated the divergent species pairs in North America (see details in the result section). Thus, we used the polarization along PC2 (hereafter called trophic axis) to evaluate trait differences of the Norwegian stickleback along the “benthic-limnetic phenotypic (associated with niche) axis”.
By using the trophic axis estimates for the Norwegian freshwater sticklebacks we used glmm to test if this axis was correlated with gillraker length and number, mean lake depth, proportion of organic carbon from littoral sources (α), trophic position (Tpos), sex and length (cm), with lake identity as a random factor. The idea was to test if similar trait changes as seen in the benthic-limnetic North American species pairs could be found in solitary Norwegian stickleback populations.
All statistical analyses were conducted using the software package JMP 9.0 .
Freshwater lake characteristics
Most of the Norwegian lakes investigated here were small (mean ± sd: 0.48 ± 0.54 km2), shallow (mean depth: 7.0 ± 3.6 m), and situated at low elevation (23.0 ± 14.2 m above sea level). The distance to the sea ranged between 0.27 and 8.5 km (3.1 ± 2.4 km). However, the lakes covered the whole zone below the post-glacial upper - maximum marine level (Table 1). Most lakes were relatively old for the region (13 372 ± 4 536 years) with a minimum age of 3 100 years and a maximum age of 18 700 years: four lakes were formed less than 10 000 years ago.
Trophic trait differences between marine and freshwater stickleback
In the principal component analysis of the Norwegian marine and freshwater populations we used the first four components explaining 18.4 % (PC1, Eigenvalue = 0.00028), 13.8 % (PC2, 0.00021), 11.7 % (PC3, 0.00018), and 8.6 % (PC4, 0.00013), of the shape variation, respectively. There were significant differences between marine and freshwater stickleback for PC1, PC2 and PC3 (Table 2, Fig. 3). Based on these three PC-axes, the marine sticklebacks had a more upturned mouth, smaller eye lower on the head, larger operculum, and deeper head than freshwater sticklebacks. Gillraker number did not differ between environments. However, gillrakers were significantly longer in marine than freshwater sticklebacks and males had slightly more and longer gillrakers than females (Table 2).
Habitat use and trophic position related to lake characteristics
Stable isotope analysis revealed that Norwegian freshwater sticklebacks obtained on average 76 % (range 11–100 %) of their carbon (α) from littoral sources, having a mean trophic level (Tpos) of 3.47 ± 0.27 (range 3.13–3.74), with large variation among the freshwater stickleback populations (Fig. 4). Only Tpos was significantly and positively associated with total length, and males had a significantly higher trophic position (Tpos) (least square mean (LSM = 3.50) than did females (LSM = 3.45) (Table 3). However, the proportion of littoral carbon (α) in the diet was not related to any of the predictor variables.
Trophic traits, habitat use and trophic position in freshwater
In the principal component analysis contrasting head shape among the Norwegian freshwater stickleback lakes the first four principal components explained 19.1 % (PC1, Eigenvalue = 0.00027), 13.0 % (PC2, 0.00019), 9.6 % (PC3, 0.00014), and 8.1 % (PC4, 0.00011) of the variation. Neither of the principal components were significantly related to any of the two habitat use indices (Tpos, α), but there was a significant effect of sex (Table 4). Gillraker number and length differed between males and females, where males had more and longer gillrakers than females. Gillraker length also increased with an increasing proportion of carbon (α) from littoral sources (Table 4).
Head shape in marine ancestor, generalists and species pairs
In the joint principal component analysis contrasting Norwegian marine and freshwater sticklebacks with North American species pairs, the first two components explained 33.2 % (PC1, Eigenvalue = 0.00096), and 16.3 % (PC2, 0.00047) of the shape variation, respectively. Here, PC1 differentiated the sticklebacks from the two continents, while PC2 mainly differentiated the ecologically divergent species-pairs along the benthic-limnetic shape axis (Fig. 5). Thus, we used PC2 as morphological polarization (hereafter termed “trophic axis”) to test for associated changes in Norwegian freshwater sticklebacks along the same axis as in North American species pairs.
In Norwegian freshwater sticklebacks, head shape on the trophic axis was significantly associated with sex, where males were more limnetic in head shape than females. The trophic axis was also significantly, and positively, associated with body length, but negatively associated with gillraker length and gillraker number (Table 5). Moreover, the mean lake depth was also negatively associated with the trophic axis. Thus, individuals with a “limnetic morph” head shape had more gillrakers as well as longer gillrakers than individuals with a more “benthic morph” head shape. The “limnetic morph” was thus associated with deeper lakes.
Trait transitions between the sea and freshwater environments
In our study, head shape in Norwegian freshwater populations differed significantly from the marine population. In general, marine fish had a more upturned mouth, a smaller eye lower on the head, a larger operculum, and a deeper head when compared with freshwater sticklebacks. The divergent head shape in our marine and freshwater sticklebacks was qualitatively similarly to that found by Voje et al.  who studied 11 marine and 74 freshwater stickleback populations in Norway. This head shape divergence between the two contrasting environments could imply functionally different trophic adaptations to divergent niches, although we do not have comparative stable isotopic data to support our claim. Previous studies comparing marine and freshwater sticklebacks are thus only partly consistent with our data showing a divergent (but contrasting head shape) of marine and freshwater populations in both North America [38, 56, 57] and Europe [37, 58]. We are not aware of studies that have analysed both trophic traits and niche use in marine and freshwater sticklebacks. This issue should be studied to reach a needed contrast of trophic ecology and trophic trait associations.
In our study, gillraker counts were similar in marine and freshwater habitats. This finding is supported by Kristjánsson et al.  who compared marine and freshwater populations of sticklebacks on Iceland and by Berner et al.  comparing marine and freshwater sticklebacks in North America. Overall, only moderate differentiation has been reported for gillraker number in populations from contrasting habitat types [60, 61]. Apparently, the adaptive transition from marine to freshwater habitats may not necessarily involve changes in gillraker number.
The Norwegian marine population had significantly longer gillrakers than freshwater sticklebacks, supporting similar observations from North America  and Europe [59, 61]. This divergence may reflect the more pelagic lifestyle of the marine stickleback . Moreover, Gross & Anderson  observed that inter-raker spacing was smaller in marine populations than in freshwater populations, suggesting a pattern where marine sticklebacks have more and longer rakers and thus a denser gillraker apparatus. Only the length of gill rakers differed in our marine-freshwater comparison. Svanbäck & Schluter  found that solitary generalist populations showed more variation in gillraker length than did either marine ancestors or the benthic-limnetic species pair. The genetic basis for such phenotypic divergence has been found by common garden experiments [22, 38, 62–64], also revealing plasticity for gillraker length . Berner et al.  found that divergence in phenotypic (co)variance structure in marine and freshwater stickleback populations was large and attributable to shifts in variance of gillraker lengths. Thus, gillraker length seems to underlie adaptive changes in divergent environments , and may be free to evolve without strong ties from the ancestral (co)variance structure upon freshwater colonization, but see also .
We have found that Norwegian marine and freshwater stickleback differ significantly in trophic traits (head shape and gill raker length). The degree and rate of phenotypic diversification in these environments could be due to a number of factors. A study by Voje et al.  compared a larger set of Norwegian populations using a model of niche-dependent optima to study trait changes along the spatio-temporal axis in postglacial sticklebacks. In that study, the results suggested that populations diversified in phenotypic directions due to allometric scaling relationships, further implying that evolutionary constraints had shaped phenotypic diversification of freshwater populations. Our Norwegian threespine stickleback populations were sampled from the same regions as in Voje et al. , which may suggest that the same interpretations could partly be valid also for our dataset.
Habitat use and trophic position - lake variables and trophic morphology
The freshwater sticklebacks studied here obtained most of their carbon from littoral sources, suggesting that they largely used benthic and littoral niches. However, at least one population was highly divergent from this pattern, being largely dependent on carbon from the pelagic niche, implying that pelagic foraging specialists occurs in some of these Norwegian lakes. This particular population (Einletvatn pond) had a higher trophic position and smaller α than the other Norwegian populations, suggesting a more pelagic lifestyle. The head shape of this population resembles the head shape of the limnetic morph members of the North American species pairs (Fig. 4), but with a lower degree of shape differentiation.
The associations between habitat use and trophic position in the Norwegian freshwater stickleback was complex. First, we found that habitat use (α) was associated with gillraker length where sticklebacks with longer gillrakers had a larger contribution of carbon from littoral sources than sticklebacks with shorter gillrakers. This is opposite to what was predicted by Matthews et al. . This finding could partly be due to differential niche-adaptation loads with regard to fitness reduction in an alternative niche by having a specific trophic trait adapted to a certain niche. In such, sticklebacks with longer gill rakers could forage effectively also on larger benthic prey sources, whereas sticklebacks with smaller gill rakers would be adapted to the benthic niche and relatively less efficient in the limnetic niche. In such, niche-specific trophic trait adaptations may have different fitness costs in alternative environments such as in the benthic and limnetic niche. Secondly, if we assume that the ancestral marine trophic state is to forage on pelagic resources, and that freshwater lakes have less profitable pelagic niches, we should expect that older stickleback populations in freshwater lakes should prefer to forage in the benthic niche (higher α). However, this seems not to be the case as α was not associated with lake age. This finding imply that Norwegian freshwater sticklebacks, despite their long time in postglacial lakes, behaves as generalists (using both benthic and limnetic niches) and have not had the ecological opportunity or time to specialize to either the benthic or limnetic niche. Here, rewarding food resources in the limnetic niche may only be temporally available during yearly variable summer months – a resource which would be difficult to adapt to. Further, we found an association between trophic position and size; larger sticklebacks had a higher trophic level than smaller sticklebacks, implying a broader niche with age. Moreover, males were foraging at a higher trophic level than females, a finding supporting earlier results .
Changes along the benthic-limnetic trophic axis of divergence
To compare Norwegian freshwater and marine stickleback populations in their morphological adaptation along the commonly discussed benthic-limnetic axis, we compared our data with the ecologically highly specialized benthic-limnetic species pairs in Paxton and Priest lakes. Here, we tested if similar trait associations as seen for these diverged species-pairs were also present in sticklebacks in Norwegian lakes. We found that the Norwegian freshwater populations in general had very low variation along this trophic axis, being intermediate to the North American species pairs (Fig. 5; PC2). However, some Norwegian populations tended to be more “limnetic” and some were more “benthic” compared with the North American species-pairs (Fig. 4; head shape). A similar quantitative differentiation among freshwater lakes in head shape was found by Willaker et al.  in North American stickleback populations. Interestingly, head shape in Norwegian lakes was negatively associated with mean lake depth – fish from deeper lakes had a more “limnetic” head, which could be due to a larger relative pelagic zone in these lakes as compared to the smaller and shallower lakes. In support of this, Willaker et al.  found that more shallow lakes had more “benthic” shaped sticklebacks in North American lakes. In shallow lakes, it is likely that benthic production overwhelms limnetic production [66–68] and thus regulates the availability of divergent niches.
We found, similarly as in the North American species-pairs, more numerous and longer gill rakers in fish with a limnetic head shape. In our study, body size was significantly and positively associated with head shape, further suggesting that larger fish with more benthic head shapes had a more benthic lifestyle. However, neither α nor Tpos was associated with head shape, which implies that the apparent adaptive transitions between the benthic and limnetic niche (as suggested by head shape in species-pairs) do not need to result in differences in trophic trait changes in individuals. This strategy could be due to temporal use of the limnetic zone in Norwegian sticklebacks. Piscivorous predators were present in all but one (Einletvatn Pond) of our Norwegian stickleback lakes, and this could constrain habitat use of the stickelback. In Willaker et al. , the presence of salmonids did not have a significant impact on the shape of sticklebacks, but lakes with pike (Esox lucius) had more “limnetic” shaped sticklebacks. Thus, it is seems likely that different piscivore fish species exerts differential selection pressures in different lake habitats upon threespine stickleback phenotypes.
Evolutionary origin of stickleback morphs in post-glacial lakes
The Norwegian freshwater stickleback utilize carbon sources spanning the whole range from almost 0 % to 100 % littoral origin. When comparing this with the North American solitary populations and species pairs (including Paxton and Priest Lakes) published by Matthews et al. , we see that the solitary Norwegian populations span a wide range of niche use and trophic position compared with the North American populations (see Fig 4). The Norwegian sticklebacks appear overall to be more benthic (mean α = 0.76 ± 0.29, N = 486) than North American species-pairs (0.45 ± 0.31, N = 403) (Fig. 4). Norwegian sticklebacks seem to have a slightly lower trophic position (mean Tpos; 3.47 ± 0.27, N =4 86) than North American sticklebacks (3.60 ± 0.23, N = 403). The wide range in Tpos between Norwegian stickleback populations implies that the populations utilize as diverse ecological niches as the highly diverged benthic-limnetic species pairs in North America. The Norwegian population with the highest trophic level and the smallest α value was a very small pond population (Einletvatn pond) where stickleback was the only fish species present. This population resembles limnetic North American forms (Fig. 4; head shape). These results suggest that the Norwegian sticklebacks have utilized ecologically all the foraging niches typically available to sticklebacks in a given lake (i.e. benthic and limnetic niches), but with only low correlation between trophic trait specialization and their utilized diet sources (Tpos and α).
Of the numerous freshwater lakes studied throughout the Holarctic only a handful of lakes harbour specialized species pairs that have radiated to the full extent into the littoral and pelagic niches, while the major number of stickleback lakes are populated by solitary generalists [1, 7, 8, 19, 21]. Why have so few stickleback populations diversified into these niches following postglacial colonization? A set of hypotheses can be suggested. First, it may be that the occurrence of the benthic-limnetic species pairs in North America, being confined to a geographically restricted area in the straits of Georgia in the Vancouver area, is due to a special set of glacial geological events that have not occurred frequently elsewhere. This hypothesis, supported by genetic studies , suggests that two waves of marine ancestors colonized the lakes at different time periods. In our Norwegian lakes, we do not have data to evaluate these possibilities. Another hypothesis under the double-colonization-scenario is that the ecological opportunity to radiate into the littoral and pelagic habitat is physically possible, but that an extensive ecological radiation is constrained by some factors. One possibility here would be that the piscivorous predators restrict sticklebacks from utilizing the pelagic zone. We have no data at hand to evaluate this hypothesis.
Lakes harbouring benthic-limnetic species pairs have a very specific fish fauna compared to lakes with single generalist populations . Vamosi  found that solitary populations and sympatric benthic-limnetic species pairs had different fish fauna. In lakes with benthic-limnetic species pairs only cutthroat trout (Oncorhynchus clarki) were present, while in 15 of 16 solitary lakes other fish species in addition to cutthroat trout were found. The two most common species in solitary lakes were prickly sculpin (Cottus asper) and rainbow trout (Oncorhynchus mykiss), both potential competitors and predators of threespine sticklebacks. In our Norwegian lakes, piscivorous fish such as brown trout and arctic charr occur frequently, where only one location is missing other fish species (Einletvatn Pond). This pond population showed the highest ecological similarity with the limnetic species pair members in North America (Fig. 4). The occurrence of predators could restrict the postglacial adaptive radiation of sticklebacks.
Another scenario that may explain stickleback diversity in Holarctic lakes is adaptive divergence after colonization of a single stickleback wave at the time of glaciostatic uplift. Here, it is reasonable to expect a longer evolutionary time needed to differentiate the original population into divergent littoral and limnetic niches. In this situation, a driving force of adaptive split into niches may be intraspecific resource competition. The degree of resource competition may be a function of the stickleback population density. However, this scenario is related to the theoretical challenges for initial within-population divergence and assortative mating to develop [1, 2, 70]. This may be a reason for the lack of a large diversity of stickleback morphs, or species, in the Holarctic, because divergence of populations needs assortative mating linked with niche divergence to develop and retain reproductive barriers.
An important issue with regard to stickleback diversity in Holarctic lakes relates to the genetic potential for adaptive tracking of available niches upon postglacial colonization. A study has suggested presence of evolutionary constraints in colonizing sticklebacks where phenotypic diversification largely followed the allometric scaling relationships for the studied traits when comparing both marine and freshwater lake populations in Norway . Our stickleback populations are sampled within the same region as in Voje . North American and Norwegian stickleback populations could harbour different genetic architectures and allometric scaling relationships and be differentially predisposed to radiate into the benthic and limnetic niche in freshwater lakes. However, as analyses such as done by Voje et al.  are not done on the North American sticklebacks we have no means of contrasting the adaptive potential.
To our knowledge, no benthic-limnetic species pair of the threespine stickleback has been found in Europe. However, the small number of sympatric benthic-limnetic stickleback pairs in a small geographical area in North America points to a special geographical occurrence in that particular area leading to the formation of the species pairs (i.e. double-immigration-scenario). If this is the case, we can imagine that the evolution of the North American species pairs was facilitated by the separation time (and adaptation) of the two species pairs before coming into secondary contact in the same lake. In such, expecting that solitary populations in the Holarctic, that likely have evolved by a one-colonization-scenario, should exhibit similar adaptive differentiation along the benthic-limnetic niche continuum as seen in the species pairs seems unrealistic. This is the situation since the North American species pair may have been formed by a different evolutionary scenario initially. Indeed, the few occurrences of benthic-limnetic species pairs as compared to the hundreds of Holarctic stickleback lakes thus imply that the limited extent of adaptive trophic trait evolution along the benthic-limnetic axis is a general phenomena for most of the Holarctic stickleback lakes. Thus, the solitary generalist populations and the species-pairs may represent specializations along the adaptation-speciation continuum, but with different evolutionary initial conditions.
Norwegian sticklebacks utilized the full spectrum of foraging niches, and seem to have the same range in trophic position as North American solitary populations and species pairs. However, Norwegian sticklebacks showed less developed and associated trait adaptations, but, if contrasted along the trophic axis, similar trait changes were seen among continents. The Norwegian sticklebacks can be viewed as solitary generalist where some populations display foraging-related niche specialization. The position along the adaptive trophic axis in Norwegian solitary populations and North American benthic-limnetic species pairs are likely due to two initially different evolutionary scenarios. Here, benthic-limnetic species pairs are likely formed due to a double-immigration-scenario, while most solitary Holarctic stickleback populations likely represent one-immigration-scenarios. The studied Norwegian stickleback lakes were up to 18 000 years old (6 000–18 000 stickleback generations assuming a 3–1 year life cycle). However, lake age showed no association with ecological divergence (Tpos, α). This postglacial timeframe of putative evolution is similar to the North American stickleback lakes. The mechanisms behind the origin of stickleback morphs with different initial starting conditions (i.e. double and one colonization scenarios) should be contrasted in future studies.
Permit for sampling of stickleback individuals were given by the Norwegian Directorate for nature management (permit no 10/06). Other permits are not required.
Consent to publish
Availability of data and materials
The dataset) supporting the conclusions of this article are available in the Dryad Digital Repository: http://dx.doi.org/10.5061/dryad.qp512.
Schluter D. The ecology of adaptive radiation. 1st ed. Oxford: Oxford University Press; 2000.
Gavrilets S. Fitness landscapes and the origin of species. 1st ed. Princeton, New Jersey, USA: Princeton University Press; 2004.
Kahilainen K, Østbye K. Morphological differentiation and resource polymorphism in three sympatric whitefish Coregonus lavaretus (L.) forms in a subarctic lake. J Fish Biol. 2006;68:63–79.
Kahilainen KK, Siwertsson A, Gjelland KØ, Knudsen R, Bøhn T, Amundsen P-A. The role of gill raker number variability in adaptive radiation of coregonid fish. Evol Ecol. 2011;25:573–88.
Hendry AP, Bolnick DI, Berner D, Peichel CL. Along the speciation continuum in stickleback. J Fish Biol. 2009;75:2000–36.
Møller JJ, Holmeslet B. SeaLevel Change Ver. 3.51. University of Tromsø; 1998.
Wootton RJ. A Functional Biology of Sticklebacks. 1st ed. London: Croom Helm; 1984.
Bell MA, Foster SA. The evolutionary biology of the threespine stickleback. 1st ed. Oxford: Oxford University Press; 1994.
Colosimo PF, Hoseman KE, Balabhadra S, Villareal Jr G, 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;307:1928–33.
Berner D, Stutz W, Bolnick DI. Foraging trait (co)variances in stickleback evolve deterministically and do not predict trajectories of adaptive diversification. Evolution. 2010;64:2265–77.
Jones FC, Chan YF, Schmutz J, Grimwood J, Brady SD, Southwick AM, Absher DM, Myers RM, Reimchen TE, Deagle BE, Schluter D, Kingsley DM. A Genome-wide SNP genotyping array reveals patterns of global and repeated species-pair divergence in sticklebacks. Curr Biol. 2012;22:83–90.
Barret RDH, Paccard A, Healy TM, Bergek S, Schulte PM, Schluter D, Rogers SM. Rapid evolution of cold tolerance in stickleback. Proc Royal Soc, Series B. 2011;278:233–8.
DeFaveri J, Shikano T, Shimada Y, Goto A, Merilä J. Global analysis of genes involved in freshwater adaptation in threespine sticklebacks (Gasterosteus aculeatus). Evolution. 2011;65:1800–7.
Klepaker TO. Morphological changes in a marine population of threespined stickleback, Gasterosteus aculeatus, recently isolated in fresh water. Can J of Zool. 1993;71:1251–8.
von Hippel FA, Weigner H. Sympatric anadromous-resident pairs of threespine stickleback species in young lakes and streams at Bering Glacier, Alaska. Behaviour. 2004;141:1441–64.
Kristjánsson BK. Rapid morphological changes in threespine stickleback, Gasterosteus aculeatus, in freshwater. Env Biol Fish. 2005;74:357–63.
Bell MA, Aguirre WE, Buck NJ. Twelve years of contemporary armor evolution in a threespine stickleback population. Evolution. 2004;58:814–24.
Le Rouzic A, Østbye K, Klepaker TO, Hansen TF, Bernatchez L, Schluter D, Vøllestad LA. Strong and consistent natural selection associated with armor reduction in sticklebacks. Mol Ecol. 2011;20:2483–93.
Voje KL, Mazzarella AB, Hansen TF, Østbye K, Klepaker T, Bass A, Herland A, Bærum KM, Gregersen F, Vøllestad LA. Adaptation and constraint in a stickleback radiation. J Evol Biol. 2013;26:2396–414.
Matthews B, Marchinko KB, Bolnick DI, Mazumder A. Specialization of trophic position and habitat use by sticklebacks in an adaptive radiation. Ecology. 2010;91:1025–34.
Williams DD, Delbeek JC. Biology of the threespine stickleback, Gasterosteus aculeatus and the blackspotted stickleback, G. wheatlandi, during their marine pelagic phase in the Bay of Fundy, Canada. Env Biol Fish. 1989;24:33–41.
Larson GL. Social behavior and feeding ability of two phenotypes of Gasterosteus aculeatus in relation to their spatial and trophic segregation in a temperate lake. Can J Zool. 1976;54:107–21.
Bell MA, Andrews CA. Evolutionary consequences of postglacial colonization of freshwater by primatively anadromous fishes. In: Streit B, Stadler T, Lively CM, editors. Evolutionary ecology of freshwater animals. Boston, MA: Birkhauser Verlag; 1997. p. 323–63.
Willacker JJ, von Hippel FA, Wilton PR, Walton KM. Classification of threespine stickleback along the benthic–limnetic axis. Biol J Linn Soc. 2010;101:595–608.
Vamosi SM. The presence of other fish species affects speciation in threespine sticklebacks. Evol Ecol Res. 2003;5:717–30.
Schluter D. Adaptive radiation in sticklebacks: trade-offs in feeding performance and growth. Ecology. 1995;76:82–90.
Gow JL, Peichel CL, Taylor EB. Ecological selection against hybrids in natural populations of sympatric threespine stickleback. J Evol Biol. 2007;20:2173–80.
Bolnick DI, Caldera EJ, Matthews B. Evidence for asymmetric migration load in a pair of ecologically divergent stickleback populations. Biol J Linn Soc. 2008;94:273–87.
Snowberg LK, Bolnick DI. Assortative mating by diet in a phenotypically unimodal but ecologically variable population of stickleback. Am Nat. 2008;172:733–9.
Bolnick DI, Lau OL. Predictable patterns of disruptive selection in stickleback in postglacial lakes. Am Nat. 2008;172:1–11.
Bolnick DI. Can intraspecific competition generate disruptive selection? An experimental test in natural populations of sticklebacks. Evolution. 2004;58:608–18.
Doucette LI, Skuálason S, Snorrason S. Risk of predation as a promoting factor of species divergence in threespine sticklebacks (Gasterosteus aculeatus L.). Biol J Linn Soc. 2004;82:189–203.
Larsen S, Andersen T, Hessen DO. Predicting organic carbon in lakes from climate drivers and catchment properties. Global Biogeochem Cycles. 2011;25, GB3007.
Hutchinson GE. Concluding remarks, Cold Spring Harbor Symposium. Quant Biol. 1957;22:415–27.
Andersen BG, Karlsen M. Nasjonalatlas for Norge. Glasialkronologi - isfrontens tilbaketrekning. Statens kartverk. (National atlas for Norway. Glacialchronology - the recession of the ice margin. Statens kartverk). (in Norwegian); 1986.
Aguirre WE, Ellis KE, Kusenda M, Bell MA. Phenotypic variation and sexual dimorphism in anadromous threespine stickleback: implications for postglacial adaptive radiation. Biol J Linn Soc. 2008;95:465–78.
Leinonen T, Cano JM, Makinen H, Merila J. Contrasting patterns of body shape and neutral genetic divergence in marine and lake populations of threespine sticklebacks. J Evol Biol. 2006;19:1803–12.
Svanbäck R, Schluter D. Niche specialization influences adaptive phenotypic plasticity in threespine stickleback. Am Nat. 2012;180:50–9.
Ravinet M, Prodöhl PA, Harrod C. Parallel and nonparallel ecological, morphological and genetic divergence in lake-stream stickleback from a single catchment. J Evol Biol. 2012;26:186–204.
Post DM. Using stable isotopes to estimate trophic position: models, methods, and assumptions. Ecology. 2002;83:703–18.
Grey J. Trophic fractionation and the effects of diet switch on the carbon stable isotopic “signatures” of pelagic consumers. Verh Int Verein Limnol. 2000;27:3187–91.
Pinnegar JK, Polunin NVC. Differential fractionation of d13C and d15N among fish tissues: implications for the study of trophic interactions. Funct Ecol. 1999;13:25–231.
Harrod C, Grey J, McCarthy TK, Morrissey M. Stable isotope analyses provide new insights into ecological plasticity in a mixohaline population of European eel. Oecologia. 2005;144:673–83.
Kiljunen M, Grey J, Sinisalo T, Harrod C, Immonen H, Jones RI. A revised model for lipid-normalizing d13C values from aquatic organisms, with implications for isotope mixing models. J Appl Ecol. 2006;43:1213–22.
Brittain JE, Lillehammer A. The fauna of the exposed zone of Øvre Heimdalsvatn: Methods, sampling stations and general results. Holarctic Ecol. 1978;1:221–8.
Larson GL, McIntire CD. Food habits of different phenotypes of threespine stickleback in Paxton Lake, British Columbia. Trans Am Fish Soc. 1993;122:543–9.
Berner D, Adams DC, Grandchamp A-C, Hendry AP. Natural selection drives patterns of lake-stream divergence in stickleback foraging morphology. J Evol Biol. 2008;21:1653–65.
Berner D, Grandchamp A-C, Hendry AP. Variable progress toward ecological speciation in parapatry: stickleback across eight lake-stream transitions. Evolution. 2009;63:1740–53.
Ibrahim AA, Huntingford FA. Laboratory and field studies of the effect of predation risk on foraging in three-spined sticklebacks (Gasterosteus aculeatus). Behaviour. 1989;109:46–57.
Robinson BW. Trade offs in habitat-specific foraging efficiency and the nascent adaptive divergence of sticklebacks in lakes. Behaviour. 2000;137:865–88.
Schluter D. Frequency dependent natural selection during character displacement in sticklebacks. Evolution. 2003;57:1142–50.
Rohlf FJ. TspDig2. http://life.bio.sunysb.edu/morph/. 2007.
Klingenberg CP. MorphoJ. UK: Faculty of Life Sciences, University of Manchester; 2008. http://www.flywings.org.uk/MorphoJ_page.htm.
Klingenberg CP. MorphoJ: an integrated software package for geometric morphometrics. Mol Ecol Res. 2011;11:353–7.
SAS Institute Inc. Using JMP 9. Cary, NC, USA: SAS Institute Inc; 2010.
Aguirre WE. Microgeographical diversification of threespine stickleback: body shape-habitat correlations in a small, ecologically diverse Alaskan drainage. Biol J Linn Soc. 2009;98:139–51.
Walker JA, Bell MA. Net evolutionary trajecories of body shape evolution within a microgeographic radiation of threespine sticklebacks (Gasterosteus aculeatus). J Zool. 2000;252:293–302.
Bjærke O, Østbye K, Lampe HM, Vøllestad LA. Covariation in shape and foraging behaviour in lateral plate morphs in the three-spined stickleback. Ecol Freshw Fish. 2010;19:249–56.
Kristjánsson BK, Skúlason S, Noakes DLG. Morphological segregation of Icelandic threespine stickleback (Gasterosteus aculeatus L). Biol J Linn Soc. 2002;76:247–57.
Hagen DW, Gilbertson LG. Geographic variation and environmental selection in Gasterosteus aculeatus in Pacific Northwest, America. Evolution. 1972;26:32–43.
Gross HP, Anderson JM. Geographic variation in the gillrakers and diet of European sticklebacks, Gasterosteus aculeatus. Copeia. 1984;1:87–97.
Day T, Pritchard J, Schluter D. A comparison of two sticklebacks. Evolution. 1994;48:1723–34.
Day T, McPhail JD. The effect of behavioural and morphological plasticity on foraging efficiency in the threespine stickleback (Gasterosteus sp.). Oecologia. 1996;108:380–8.
Hatfield T. Genetic divergence in adaptive characters between sympatric species of stickleback. Am Nat. 1997;149:1009–29.
Reimchen TE, Nosil P. Variable predation regimes predict the evolution of sexual dimorphism in a population of threespine stickleback. Evolution. 2004;58:1274–81.
Vadeboncoeur Y, Vander Zanden MJ, Lodge DM. Putting the lake back together: reintegrating benthic pathways into lake food webs. BioSci. 2002;52:44–54.
Sierszen ME, McDonald ME, Jensen DA. Benthos as the basis for Arctic lake food webs. Aquat Ecol. 2003;37:437–45.
Vander Zanden MJ, Chandra S, Park S-K, Vadeboncoeur Y, Goldman CR. Efficiencies of benthic and pelagic trophic pathways in a subalpine lake. Can J Fish Aquat Sci. 2006;63:2608–20.
Taylor EB, McPhail JD. Historical contingency and ecological determinism interact to prime speciation in sticklebacks, Gasterosteus. Proc Royal Soc B. 2000;267:2375–84.
Bernatchez L. Ecological theory of adaptive radiation: an empirical assessment from coregonine fishes (Salmoniformes). In: Hendry AP, Stearns SC, editors. Evolution illuminated: salmon and their relatives. Oxford, UK: Oxford University Press; 2004. p. 175–207.
We wish to thank J. Møller at Tromsø University for help with lake age estimates, K. M. Bærum for help with photo and body shape analysis, and A. Taugbøl for comments on an earlier version of the manuscript. Thanks to two referees that helped improve the quality of the manuscript. We thank H. Buhtz and A. Möller for help in the laboratory.
This study was financially supported by the Research Council of Norway (grant no 170755/V20 to LAV). CH thanks Diethard Tautz, Winfried Lampert and the Max Planck Society for financial support.
The authors declare that they have no competing interest.
KØ, CH, LAV, TK and DS conceived the study. KØ, FG and MS planned the study design and performed the fieldwork. KØ conducted the shape and morphometric analyses. CH conducted the stable isotope analyses. KØ and LAV conducted the statistical analyses. KØ wrote the manuscript with major contribution from LAV. All the authors read and contributed to manuscript writing, and approved the final version of the manuscript.
About this article
- Adaptive radiation
- Natural selection
- Ecological niche
- Stable isotopic analysis
- Isostatic uplift
- Pleistocene ice sheet
- Benthic-limnetic stickleback species pairs