- Research article
- Open Access
Ongoing niche differentiation under high gene flow in a polymorphic brackish water threespine stickleback (Gasterosteus aculeatus) population
© The Author(s). 2018
- Received: 29 July 2017
- Accepted: 22 January 2018
- Published: 5 February 2018
Marine threespine sticklebacks colonized and adapted to brackish and freshwater environments since the last Pleistocene glacial. Throughout the Holarctic, three lateral plate morphs are observed; the low, partial and completely plated morph. We test if the three plate morphs in the brackish water Lake Engervann, Norway, differ in body size, trophic morphology (gill raker number and length), niche (stable isotopes; δ15N, δ13C, and parasites (Theristina gasterostei, Trematoda spp.)), genetic structure (microsatellites) and the lateral-plate encoding Stn382 (Ectodysplasin) gene. We examine differences temporally (autumn 2006/spring 2007) and spatially (upper/lower sections of the lake – reflecting low versus high salinity).
All morphs belonged to one gene pool. The complete morph was larger than the low plated, with the partial morph intermediate. The number of lateral plates ranged 8–71, with means of 64.2 for complete, 40.3 for partial, and 14.9 for low plated morph. Stickleback δ15N was higher in the lower lake section, while δ13C was higher in the upper section. Stickleback isotopic values were greater in autumn. The low plated morph had larger variances in δ15N and δ13C than the other morphs. Sticklebacks in the upper section had more T. gasterostei than in the lower section which had more Trematoda spp. Sticklebacks had less T. gasterostei, but more Trematoda spp. in autumn than spring. Sticklebacks with few and short rakers had more T. gasterostei, while sticklebacks with longer rakers had more Trematoda. spp. Stickleback with higher δ15N values had more T. gasterostei, while sticklebacks with higher δ15N and δ13C values had more Trematoda spp. The low plated morph had fewer Trematoda spp. than other morphs.
Trait-ecology associations may imply that the three lateral plate morphs in the brackish water lagoon of Lake Engervann are experiencing ongoing divergent selection for niche and migratory life history strategies under high gene flow. As such, the brackish water zone may generally act as a generator of genomic diversity to be selected upon in the different environments where threespine sticklebacks can live.
- Gill raker
- Natural selection
- Stable isotope analyses
- Theristina gasterostei
- Trematoda spp
Within most populations some individuals tend to be better able to disperse and successfully colonize new environments than others. Are such individuals genetically pre-adapted, environmentally cued, or simply a random draw of the population? This is a relevant question for the threespine stickleback, Gasterosteus aculeatus, a highly adaptable euryhaline species commonly observed in salt water, brackish water and fresh water throughout the Holarctic [1–3]. Although originally marine, parallel freshwater colonization has occurred following the last Pleistocene deglaciation [1, 4–6]. Sticklebacks have been studied in great detail in freshwater systems, while less is known regarding the very early steps preceding freshwater colonization. As such, studying sticklebacks in brackish water where gene flow from both marine ancestors and freshwater populations occur may increase our understanding of how divergent multifarious adaptive processes act under gene flow.
Threespine stickleback are morphologically diverse, for example there is extensive variation in the number of lateral plates. Here, three nominal lateral plate morphs are recognized [7, 8]; (1) a completely plated morph with a full cover of lateral plates along the body flank most commonly associated with salt and brackish water, (2) a partially plated morph with a reduced lateral plate cover along the body flank, but with a fully or partly developed keel on the tail, mostly in brackish water and fresh water, and (3) a low plated morph with only a few anterior lateral plates along the body lacking a keel, dominating in fresh water. In addition, a rare fourth morph lacking all lateral plates is only found in a few freshwater lakes . Strong directional selection for loss of lateral plates has occurred during freshwater colonization, causing shifts in mean phenotype within a few generations . As such, there appears to be a directional pattern of decreased number of lateral plates linked with salinity regimes.
At the Holarctic scale, lateral plate reduction in threespine stickleback has occurred repeatedly and independently as freshwater rivers and lakes were colonized after the last glacial . Rapid adaptive loss of armor plates appears to be due to selection on the ancestral pool of standing genetic variation in the ocean [5, 11]. Genetic studies have shown that Ectodysplasin (Eda) on chromosome IV is a major bi-allelic locus for lateral plate development, with low (aa), partial (Aa) and completely plated morphs (AA). Additional QTLs also appear to play a role in determining lateral plate number within morphs . The frequency of the Eda low allele (a) has been estimated for marine and anadromous threespine stickleback populations and range between 0.0–19.2% (see overview of studies presented in Bell et al. ). After colonization, an increase in the low plated morph implies strong selection on the Eda gene and/or other traits in the haplotype block [10–16]. Based on experimental crosses, and field surveys, the Eda locus can explain up to 70% of the variation in lateral plate number and is associated with lateral plate morphs in geographically diverse populations [5, 13, 14, 16–19].
Since stickleback plate morphs are associated with different salinity environments, an important question is whether the different plate morphs have other specific traits that potentially “pre-adapt” them to e.g. freshwater colonization. An experiment showed that offspring of partial- and low plated sticklebacks grew equally well as completely plated morphs when raised in salt water, but outcompeted the complete morph in fresh water . In a common garden experiment using replicate semi-natural ponds in fresh water, the “low plate” a allele individuals had a higher juvenile growth rate and a higher overwinter survival while the “complete plate” A allele individuals caught up in size at sexual maturation . Barret et al.  further showed that the low plated fish explored both salt and freshwater habitats regardless of being acclimated to salt or fresh water, while the completely plated fish preferred the environment they were acclimatized in. Furthermore, Greenwood et al.  showed that Eda partly explained schooling behaviour in freshwater benthic and marine pelagic threespine stickleback. A recent study by Robertson et al.  found an association between the Eda haplotype block and the relative expression of transcripts of immune system genes. In that study, Eda genotypes in F2 from an experimental cross of one freshwater lowplated and one marine completely plated fish where exposed to fresh and marine water. In that experiment, the aa low plated morphs had a lower growth rate than the faster growing AA completely plated fish in both water conditions, The lowplated morph also had a higher parasite load. These results imply that sticklebacks with a lowplated aa genotype may have a different and more explorative foraging behaviour, as well as potentially also different immune system genes compared to the two other Eda genotypes. Thus, Eda genotypes may be in linkage with genes underlying other adaptive traits important for colonization.
The efficiency of natural selection under gene flow remains an intriguing question in evolutionary biology (e.g. ). Under low levels of gene flow, populations may diverge due to local selection pressures, while under higher levels of gene flow from populations adapted to divergent environments local adaptations may be swamped by gene flow . Many studies on sticklebacks have targeted freshwater populations and analyzed adaptive diversification in various habitat types (limnetic-benthic / lake-stream) (e.g. [1, 24]. In comparison, we know less about behaviour, life history, ecology and genetic structure in the brackish water zone sticklebacks (but see [16, 17, 26–31]. In brackish water, immigrants from fresh water and salt water may generate a zone of high genotypic and phenotypic diversity. Hybridization and genomic introgression between freshwater and saltwater adapted sticklebacks (as well as among plate morphs) could produce novel adaptive genetic combinations in this zone. Thus, sticklebacks of a specific genomic variant for a successful freshwater colonization may exist in brackish water being more prone to colonize fresh water when new opportunities arise. Depending upon the geographical area, postglacial lakes likely went through a temporal increasing isolation from the ocean due to isostatic rebound resulting in a variable timeframe of brackish water influence. Studying brackish water zones may be important for our general understanding of post-glacial colonization success of stickleback in the different salinity environments.
The main aim of this study was to describe niche occupation and associated morphological traits in three sympatric lateral plate morphs of threespine stickleback in the brackish water Lake Engervann, Norway. Here, we tested if morphs differed in body size, trophic morphology (gillraker number and raker length), and niche occupation (stable isotopes and parasite load). Microsatellites were used to test if morphs belonged to different genetic populations. We further tested for associations in niche occupation and trophic morphology in the three stickleback morphs to indirectly search for trait combinations related to the three underlying Eda genotypes. Studying the threespine stickleback in the brackish water zone is important for our understanding of how adaptation may proceed in such zone of high gene flow, as well as how the brackish water population may act as a facilitator of rapid adaptive radiation by likely harboring genomic combinations from divergent environments.
The upper section of Lake Engervann receives more fresh water than the lower section due to the inflowing fresh water from River Øverlandselva, while the lower section receives more salt water due to tidal influence (Fig. 1). There is large spatial and temporal variation in salinity, with conductivity varying between 60 and 3000 mS/m between sites and days in Lake Engervann reflecting the incoming fresh water from River Øverlandselva and incoming tide near the outlet of the lake (see ). In general, freshwater conductivity is rarely higher than 20 mS/m and seawater conductivity is commonly ca 5200 mS/m [34, 35]. In such, Lake Engervann can be evaluated as a typical brackish water area in Norway.
Stickleback sample collection
Body length measures (minimum-mean-maximum) in the 12 sample groups (sex pooled) of the three lateral plate threespine stickleback morphs in Lake Engervann based on sampling locations and dates
Lateral plate morph
Body length (cm) min - mean - max
N fish for Stn382/μsat
Completely plated (CPM)
4.60–5.04 – 5.45
N fish total = 76:
4.55–5.08 – 5.70
N males = 38
4.70–5.22 – 5.70
N females = 38
4.75–5.27 – 6.75
Partially plated (PPM)
3.90–4.58 – 5.30
N fish total = 84:
4.15–4.77 – 5.35
N males = 42
4.25–4.84 – 6.45
N females = 42
4.50–4.99 – 5.90
Low plated (LPM)
3.35–4.24 – 5.50
N fish total = 76:
3.45–4.21 – 5.35
N males = 36
2.90–3.72 – 4.80
N females = 40
3.50–4.07 – 4.90
Stable isotope analyses
Long-term diet and niche preferences, as a measure of the niche occupation were estimated using carbon (δ13C) and nitrogen (δ15N) stable isotope ratios in individual sticklebacks. Although data are not available for nitrogen isotope turnover in threespine sticklebacks, Grey  showed that δ13C in threespine stickleback muscle generally reflect a dietary history of approximately six months, a time span comparable to other north temperate fish species . Stable isotopes reflect what the consumer has assimilated through its diet consumption in different habitats [40–42]. Here, δ13C reveal information on prey use along the saltwater-freshwater resource axis (depleted δ13C values reflect utilization of freshwater localized prey . In contrast, δ15N values reveal information on trophic levels .
For analysis, a piece of the tail muscle was dissected out, dried for 24 h at 60 °C, ground, weighed, encapsulated in tin cups and analyzed in a stable isotope ratio mass spectrometer after methods in Harrod et al. . Since C:N ratios differed among the three lateral plate morphs (ANOVA: F2, 235 = 41.85, P < 0.001), muscle lipid content also likely varies. Therefore all δ13C data was arithmetically lipid-normalized before further analysis .
Potential prey items were sampled using a dipnet and a sieve in the upper, middle and lower sections of Lake Engervann (Fig. 1). In a more marine influenced environment, benthic invertebrates along the shoreline of Kalvøya Island were sampled with a dip net and the pelagic habitat nearby the Steilene Island was sampled with a plankton net. Prey items were analyzed for stable isotopes. For each prey item, five or more specimens were pooled precluding a quantification of the variation within the species.
The putative stickleback prey items were; Asellus aquaticus, Gammarus duebeni, Copepoda spp., Chironomidae spp., Palaemon adspersus, Neomysis integer, Pandalina profunda, Nudibranchia spp., Polychaeta spp., Planorboidea spp., “unidentified snail spp.”, Mytilus edulis, Balanus balanoides, and Chaegnata spp. No zooplankton species were found in Lake Engervann despite several sampling efforts. The stable isotopic values for putative prey items are given in Additional file 1: Table S1.
Parasites as ecological markers
Two ecto-parasites were additionally used as ecological markers to reflect long-term habitat- and diet, reflecting niche occupation. The choice of parasite species to be analyzed was based on their easy visual recognition and also their robustness with regard to sticklebacks being removed from storage in EtOH. Thus, we did not analyse Gyrodactylus spp. as the fish had been dried several times. First, the number of the crustacean copepod Thersitina gasterostei (family Ergasilidae) [Pagenstecher, 1861] was counted on the inner side of the operculum and on the muscle tissue close to the branchial arch on both sides of the fish. The sum of the copepods on both sides of the fish was used in analyses. T. gasterostei is a parasite on Holarctic euryhaline fishes , often on gasterosteids [45–47], and particularly on threespine stickleback .
The second parasite was an unidentified metacercaria of Trematoda spp. encysted only on the pectoral, dorsal, anal and caudal fins. The number of Trematoda spp. metacercaria cysts were counted on both pectoral fins, the dorsal, anal and caudal fins and then summed. To clarify the identity we used three primer-pairs to amplify a 1410 bp region (combined into one sequence) of the rDNA ITS gene  partly covering the ribosomal gene clusters 18S, ITS1, 5.8S, ITS2, 28S. PCR was performed using PuReTaq ready-to-go PCR beads (GE Healthcare), 1 mMol− 1 of each primer, and 5 ml of the extracted DNA in a 50-ml reaction volume. The thermal cycling was: 94 °C for 3 min; 35 cycles of 94 °C for 30 s, 55 °C for 47 s, and 72 °C for 1 min; and a final extension at 72 °C for 5 min. PCR products were purified by 10× diluted exoSAP-IT (USB). Cycle sequencing, using the same primers as in the PCR, was performed in 10-ml reactions using 2-ml BigDye terminator cycle sequencing ready kit (Applied Biosystems), 2 ml 5× sequencing buffer, 10 pmol primer, and 3-ml cleaned PCR product. One cyst from each of eight sticklebacks was sequenced. Data analyses were done in Sequencher 5.0 (Gene Codes Corporation, Ann Arbor, Michigan, USA) and aligned with the Crustal W algorithm in MEGA 6.0  using default settings. Sequences were compared in BLAST (blast.ncbi.nlm.nih.gov/Blast.cgi) using a reduced set of 871 bp to increase number of comparisons to four sequences. Evolutionary analyses were conducted in MEGA6 with evolutionary history inferred using the Minimum Evolution (ME) method  with 1000 bootstraps . Evolutionary distances were computed using the Maximum Composite Likelihood method  and are in units of number of base substitutions per site. The ME tree was searched using the Close-Neighbor-Interchange (CNI) algorithm  at a search level of 1. The Neighbor-joining algorithm  was used to generate the initial tree. Codon positions included were 1st + 2nd + 3rd + Noncoding. All positions with gaps and missing data were eliminated.
Population genetic analysis
DNA was extracted from pectoral fins using standard proteinase K phenol chloroform protocol . A set of 25 microsatellites, where 12 loci were a priori suggested to be neutral and 13 loci to be QTLs, were analyzed. Specific information about all the primers, PRC run conditions, binning and applied laboratory methods are reported in Le Rouzic et al.  and Klepaker et al. . Two of the markers, Stn381 and Stn382, are situated in two introns of the Ectodysplasin gene (Eda) that is a major determinant of lateral plate morphs partitioning the completely plated, partially plated and the low plated morph [5, 13]. We analyzed individuals in the upper and lower part of Lake Engervann in spring 2007 (Table 1) as this should be sufficient for testing for potential population genetic structure. Genetic analyses were performed on all genotyped stickleback in Lake Engervann together assuming that all plate morphs belonged to a single population.
Microsatellites were first screened in MICRO-CHECKER 2.2.3  to evaluate presence of stutter, allelic drop-out, homozygote excess and null-alleles. Four loci (Stn152, Stn180, Stn211, Stn271) showed homozygote excess/null alleles in two or three comparisons and were removed from all the further analyses.
To test if microsatellites were neutral or candidates for either directional or balancing selection, all 21 loci were run in LOSITAN [59, 60] under the stepwise mutation model (SMM) and the infinite alleles model (IAM). Here, we used 100,000 simulation with the “Neutral mean Fst” and “Force mean Fst” options when analyzing data separately for IAM and SMM models. For all the simulations, the two microsatellites Stn381 and Stn382 emerged as candidates for directional selection and was thus removed before analyzing the rest of the markers collectively. None of the remaining 19 microsatellites showed any signals of selection and were therefore inferred and used as neutral markers in population genetic structure analysis.
Genotypic linkage disequilibrium and deviations from Hardy-Weinberg equilibrium (HWE) were analyzed by the exact (probability) test estimated in GENEPOP 4.0.10 . The results showed that a total of 4 loci differed from HWE, and only 3 after Bonferroni corrections. To be conservative, we thus removed the three loci (Stn178, Stn180 and Gac2111) from further population genetic analysis.
To test for population genetic structure, we used Bayesian population clustering in STRUCTURE 2.3.3.  with the 17 neutral microsatellites. We used an admixture model, correlated gene frequencies, 500,000 burn-in steps and 700,000 MCMC iterations, with K set to vary between 1 and 3 clusters with 5 replicates for each K. In addition, we conducted similar STRUCTURE analyses with the same burn-in steps and MCMC iterations, but now using the locprior option comparing (i) upper versus lower lake section regardless of morph (K:1–3, 5 replicates), (ii) three plate morphs (K:1–5, 5 replicates), (iii) or contrasting the morphs x location (K:1–6), 5 replicates. These additional analyses were performed with all loci (including loci deviating from HWE), but excluding Stn381 and Stn382 (being Eda linked loci), and null allele loci. The results were evaluated based on Evanno et al.  and Pritchard et al.  also using STRUCTURE-HARVESTER v0.6.93 . We also ran a DAPC analysis (a principal component analyses) using adegent (http://adegenet.r-forge.r-project.org/).
Differences in body length was tested using an ANOVA with the three lateral plate morphs and the two sampling dates (autumn 2006 or spring 2007) as predictor variables, grouping the upper and lower sampling sites in the lake together to increase statistical power. A post hoc Tukey HSD test was here used to test if lateral plate morphs differed significantly from each other in body length in the two time periods.
Using an ANCOVA we quantified the differences in lateral plate number among morphs while correcting for body length. The aim here was not to test if morphs were different, as we already have grouped them based on coverage of lateral plates – but rather to quantify the number of lateral plates in the three morphs. Here, we used the summed number of plates on both sides grouping all samples together in each morph.
As the two loci Stn381 and Stn382 provided similar information, given that they are both indel markers of Eda, we focused only on Stn382 in subsequent analyses. This marker consistently gave two alleles and three genotypes in contrast to Stn381 which may have three alleles and more genotypes [13, 17]. Association between Stn382 genotypes AA, Aa and aa and the lateral plate morph categories (complete, partial and low plated) were tested using a contingency analysis. Due to the observed strong association between plate morph categories and Stn382-genotypes (see below) we only used plate morph category in further analyses.
We used general linear models to test if the number of frontal gill rakers and the 3rd gill-raker length on the first right lower gill arch were associated with the three lateral plate morphs (complete, partial or low plated), sex (male or female), or body length.
Using general linear models we tested if stable isotopic signatures of nitrogen (δ15N) and carbon (δ13C) were associated with sampling area (upper or lower part), sampling date (Autumn 2006 or Spring 2007), sex (male or female), lateral plate morphs (complete, partial or low plated) and body length. To evaluate if lateral plate morphs had different range in diet items (based on values of δ15N and δ13C) we tested if stable isotope variance differed among the three lateral plate morphs using the Levene’s test on the residuals from a correlation of body length vs stable isotopes.
Further, using parasite taxa as markers to reflect the ecological niche of lateral plate morphs, we estimated their prevalence (i.e. % of individuals in the population being infected) and intensity (i.e. mean value of fish being infected) in the whole Lake Engervann population, as well as separately for each of the three lateral plate morphs.
We tested variation in parasite infection (each taxon separately) using a generalized linear model with a Poisson distribution and log link, with the following predictors; sampling area (upper or lower part), sampling date (autumn 2006 or spring 2007), sex (male or female), lateral plate morphs (complete, partial or low plated), trophic traits (gill-raker number, 3rd gill-raker length), isotopes (δ15N, δ13C) and total body length.
All the statistical analyses were performed using the software JMP 9.0 .
All the four performed STRUCTURE analyses and the additional DAPC analysis suggested that the lateral plate morphs of threespine stickleback in Lake Engervann belonged to one single gene pool (K = 1). The result is reported in the Additional file 2: Table S2 and Additional file 3: Figure S3, while the data for genetic markers is given in Additional file 4: Table S3 and Additional file 5: Table S4.
Lateral plate morphs and Eda marker Stn382
The total number of lateral plates on both sides of the fish summed ranged between 8 and 71 (Fig. 3). Here, the completely plated morph had a mean (± SD) number of lateral plates of 64.2 ± 2.1, the partially plated morph a mean of 40.3 ± 10.3, and the lowplated morph a mean of 14.9 ± 5.9.
Associations among the three lateral plate stickleback morphs in Lake Engervann and the Stn382 - Ectodysplasin genotypes. - denotes lack of observations of the specified Stn382 - Ectodysplasin genotype in the complete- and lowplated morphs, respectively
Lateral plate morph
Ectodysplasin genotypes Stn382
N fish Stn382
% occurrence within morph
Plate number min-max
Plate number mean values
Plate number standard deviation
Completely plated (CPM)
N fish total = 76
Partially plated (PPM)
N fish total = 84
Low plated (LPM)
N fish total = 76
General linear mixed models on frontal gill raker counts on first right gill arc and on the 3rd gill-raker length on first right lower gill arch with predictor variables in the Lake Engervann threespine sticklebacks. - denote non-reported values
Estimate ± se
Number of frontal gill rakers on the first right gill arch
20.637 ± 0.978
R2 = 0.02
Lateral plate morph - CPM*
0.319 ± 0.172
N = 236
Lateral plate morph - LPM*
−0.353 ± 0.189
P = 0.324
− 0.014 ± 0.098
Df = 4
−0.167 ± 0.208
3rd gill-raker length on the first right lower gill arch
−0.029 ± 0.082
R2 = 0.656
Lateral plate morph - CPM#
0.011 ± 0.014
N = 236
Lateral plate morph - LPM#
− 0.010 ± 0.016
P < 0.001
−0.056 ± 0.008
Df = 4
0.243 ± 0.017
Description of stable isotope values in putative prey items and sticklebacks
The variation in stable isotope values of sticklebacks in Lake Engervann (all data combined) ranged from − 22.5 to − 13.0 ‰ for lipid-corrected δ13C and between 7.0 and 14.3 ‰ for δ15N (Fig. 4). The variance was 2.5 for δ13C and 1.7 for δ15N in all stickleback. This minimum-maximum range in stable isotope values of sticklebacks is visualized by the thin black line in the convex hull plot of their putative prey items in Fig. 4. The mean values of δ13C and δ15N were similar in morphs. Mean δ13C values (mean ± SD) for each morph were; completely plated: − 17.1 ± 1.3, partially plated: − 17.1 ± 1.3 and the low plated morph: − 17.8 ± 1.9. Mean δ15N values were more variable; completely plated: 12.0 ± 1.0, partially plated: 11.6 ± 1.2 and the low plated morph: 10.9 ± 1.5.
Predictors of stable isotope values in sticklebacks
General linear mixed models on niche/diet preferences by the use of stable isotopic responses of δ15N and δ13C with predictor variables for the threespined stickleback in Lake Engervann. - denote non-reported values
Estimate ± se
Stable isotopic values of nitrogen (δ15N)
9.380 ± 0.772
R2 = 0.21
0.238 ± 0.077
N = 236
−0.210 ± 0.078
P < 0.001
−0.086 ± 0.077
Df = 6
Lateral plate morph - CPM*
0.258 ± 0.135
Lateral plate morph - LPM*
−0.313 ± 0.149
0.470 ± 0.164
Stable isotopic values of carbon (δ13C)
−20.86 ± 0.943
R2 = 0.20
0.193 ± 0.095
N = 236
−0.474 ± 0.095
P < 0.001
0.001 ± 0.094
Df = 6
Lateral plate morph - CPM#
−0.117 ± 0.165
Lateral plate morph - LPM#
− 0.105 ± 0.182
0.762 ± 0.201
A Levene test (F = 4.84, N = 236, df = 233, 2, P = 0.009) showed that the variance of δ13C was larger in the low plated morph (variance: 3.15, N = 76) than in the partially plated (1.77, N = 84) and the completely plated morph (1.89, N = 76). The same pattern was seen for δ15N (F = 9.40, N = 236, df = 233, 2, P < 0.001), where the low plated morph (variance: 2.11, N = 76) had a significantly larger variance than the partially plated (1.27, N = 84) and the completely plated morph (1.01, N = 76). The partially plated morph and completely plated morph did not differ in δ13C and δ15N variance.
Parasite infection in sticklebacks
The prevalence and mean intensity of infection was quite similar among the morphs. In the completely plated morph (combining time and locations) 5 out 76 individuals lacked T. gasterostei giving a prevalence of 93.4% and mean intensity of 7.1 (±4.8). For the partially plated morph (combining time and locations), 5 of 84 individuals lacked T. gasterostei giving a prevalence of 92.9% and a mean intensity of 6.6 (±7.1). In the lowplated morph (combining time and locations), 2 out of 71 individuals lacked T. gasterostei thus giving a prevalence of 97.2% and a mean intensity of 4.8 (± 3.1).
General linear mixed models of counts of the parasitic copepod T. Gasterostei and counts of Trematoda spp. cysts with predictor variables for the threespined stickleback in Lake Engervann. - denote non-reported values
Estimate ± se
Number of T. gasterostei (Poisson distribution, log link)
3.524 ± 0.724
χ2 = 231.92
−0.078 ± 0.028
N = 236
0.055 ± 0.030
P < 0.001
0.072 ± 0.031
Df = 10
Lateral plate morph - CPM*
0.030 ± 0.046
Lateral plate morph - LPM*
−0.032 ± 0.053
Gill raker number
−0.073 ± 0.019
3rd gill-raker length
−0.676 ± 0.220
−0.193 ± 0.024
0.033 ± 0.021
0.668 ± 0.082
Number of Trematoda spp. (Poisson distribution, log link)
−0.458 ± 0.583
χ2 = 784.95
0.114 ± 0.024
N = 236
−0.116 ± 0.025
P < 0.001
−0.063 ± 0.026
Df = 10
Lateral plate morph - CPM#
0.214 ± 0.037
Lateral plate morph - LPM#
−0.266 ± 0.048
Gill raker number
0.009 ± 0.015
3rd gill-raker length
0.566 ± 0.173
0.190 ± 0.026
0.137 ± 0.018
0.365 ± 0.073
All the eight sequenced cysts collected from the sticklebacks were identical (Genbank accession number KY620038).The sequence clustered most closely with the heterophyid trematode causing the black-spot disease Cryptocotyle lingua (Genbank Accession Number: KJ641524  with 956 of 1006 bp matched (97% query coverage, identification 100%). The next best hit was Cryptocotyle lingua (Genbank Accession Number: KJ641518; ) with 950 of 1006 bp matched (96% query coverage). As our sequence had a bootstrap value of 100% grouping out from C. lingua (and the outgroup of Pygidiopsis genata (AY245710 , our metacercaria was evaluated as being significantly different from C. lingua (Additional file 7: Figure S2). The taxonomy is currently unknown.
The summed distribution of Trematoda spp. on the pectoral, pelvic, dorsal, anal and caudal fins (combining plate morphs, time and locations) approached a Poisson distribution (Fig. 5 upper panel) (see also Additional file 6: Figure S1. The number of Trematoda spp. ranged from 0 to 65 (mean ± SD; 8.42 ± 9.25) (Fig. 3), where 29 of 236 fish lacked Trematoda spp. giving a prevalence of 89.9% and a mean intensity of 3.5 (±1.4).
Prevalence and mean infection intensity varied among morphs. In the completely plated morph (combining time and locations) only 1 of 76 fish lacked Trematoda spp. giving a prevalence of 98.7% and a mean intensity of 12.8 (±11.2). For the partially plated morph, 5 of 84 fish lacked Trematoda spp. giving a prevalence of 94.0% and a mean intensity of 12.8 (±8.2), while in the lowplated morph 22 out 71 fish lacked Trematoda spp. giving a prevalence of 69.0% with a mean intensity of 5.4 (±5.0).
The number of Trematoda spp. was significantly associated with a set of variables (Table 5). Here, sticklebacks in the lower part of the lake had more Trematoda than in the upper part. More Trematoda spp. were recorded in Autumn 2006 than in Spring 2007. Males had more Trematoda spp. than females. The lowplated morph had fewer Trematoda spp. than the other two morphs. Fish with larger gill raker length had more Trematoda sp. than fish with smaller raker length. Furthermore, stickleback with larger values of both δ15N and δ13C had more Trematoda. spp. cysts. Finally, fish with larger body length had more Trematoda spp. cysts than smaller sticklebacks (Table 5).
One panmictic gene pool comprising all the three sympatric lateral plate morphs
The three morphs of threespine stickleback in Lake Engervann belonged to one single population revealed as one genepool. Conversely, we observed marked between morph differences in body length and lateral plate counts indicating limited gene flow (or divergent selection) among morphs for such traits. As such, it may still be that the underlying genes for heritable traits have a lower degree of gene flow among plate morphs if associated with divergent adaptive traits, behaviour, niche preferences or life history. Such heterogeneous genomic differentiation was observed in Baltic Sea sticklebacks by DeFaveri et al.  and Guo et al. . Based on the Eda results, it seems that the plate morphs in Lake Engervann are more or less genotypically distinct while yet sharing the same neutral genepool. This could suggest that morphs may differ in traits, behavior and life history due to associations with the Eda haplotype block, or alternatively that traits are free to be exposed to natural selection and evolve without genetic constraints if residing on different chromosomes. Natural selection may fine-tune important adaptive traits in the EDA haplotype block while still allowing for neutral genes to flow among the three lateral plate morphs.
Associations among Eda genotypes, lateral plate counts and lateral plate morphs
The range in plate number observed in Lake Engervann sticklebacks (both sides summed; 8–71) covers most of the variation seen in populations of the threespine stickleback in the Holarctic [7, 69–71]. The three Eda Stn382-genotypes AA, Aa and aa were clearly associated with the three plate morphs where the AA genotype (89.5%) dominated in the completely plated morph, the Aa genotype (59.5%) in the partially plated morph and the aa genotype (73.7%) in the low plated morph. Similar patterns of Eda markers (Stn380, Stn381, Stn382), plate morphs and plate number have been found in other studies, although with variation in relative percentages of genotypes in morphs [5, 13, 16–18, 72, 73]. An interesting comparison is the threespine stickleback in Lough Furnace, Burrishole Catchment, Western Ireland where a small bodied freshwater stationary form enters brackish water (8.3–29.7 ppt) area living together with a larger brackish water resident form and an even larger anadromous form of stickleback . The Irish plate morphs are comparable in lateral plate number and body size to the three plate morphs in Lake Engervann, where similarities are also evident in the association between lateral plate morphs and the Eda marker genotypes at Stn382. The repeated association between lateral plate morphs and salinity environments from ancestral marine to derived freshwater habitats suggests adaptive loss of lateral plates [2, 3, 6, 7]. In support, genetic studies in sticklebacks document that Eda is under divergent selection in contrasting environments [5, 14] where also other traits in the Eda haplotype block could be under linkage-disequilibrium-based hitchiking and pleiotropic selection. The proportion of plate morphs varies in the Holarctic, where similar patterns as seen in Lake Engervann are found in brackish water and young freshwater lakes along the whole Norwegian coast .
Ecological niche diversity of plate morphs based on stable isotopes and parasites
The ecological niche of an individual comprises the temporally framed accumulated habitat use and food preferences, which can also change during ontogeny through niche shifts. We did not perform quantitative sampling aiming at estimating the density of putative prey species for sticklebacks, but we evaluated that the most available invertebrate species would be the most likely prey for threespine stickleback in Lake Engervann based on an earlier survey in Lake Engervann by Halvorsen et al.  (Additional file 8: Table S5). In that study, the most common species were Chironomidae, Oligochatea and Gastropoda (Hydrobia ulvae). We did not analyze stomach content in the sticklebacks in Lake Engervann, but other studies have found that a diversity of prey species are consumed by threespine stickleback in brackish water environments spanning benthic-littoral zone animals such as molluscs, ostracods, chironomida, cladocerans, ciliates, orthocladiinae, and copepods in the pelagic zone [74–79].
Stable isotope signatures
In general, the patterns in the stable isotope data suggest that sticklebacks were foraging in the middle-lower part of Lake Engervann and more marine environments, rather than in the upper lake area (Fig. 4). The total range in δ13C and δ15N values in the sticklebacks ranged − 22.5 to − 13.0 ‰ and 7.0 to 14.3 ‰, respectively. Based on one prey item, Chironomidae spp. δ13C and δ15N values showed large variation from upper-middle and lower parts of Lake Engervann from − 31.2 to − 26.0 to − 20.9 ‰ and 4.5 to 7.1 to 8.6 ‰, respectively (see other putative prey items in Additional file 1: Table S1). Despite the impression that the three lateral plate morphs in Lake Engervann had generally similar diets based on stable isotopes when visualized in the biplot of δ13C versus δ15N, we observed interesting statistical differences. First, we found that the low plated morph had larger variance in δ15N and δ13C than the two other morphs. This finding may suggest that the low plated morph is more explorative in its habitat use and prey preferences than individuals in the two other morphs. Secondly, there was a spatial component in niche use in Lake Engervann as sticklebacks in the lower lake section were more enriched for 15N and 13C than in the upper section. This pattern may reflect freshwater stationary feeding behavior on a simple food chain in the upper part of the lake and a migratory foraging behavior on a more complex food chain with higher diversity of prey at different trophic layers in the lower part of the lake - with foraging migrations into marine environments. Also, it may reflect a difference in the relative amount of benthic v pelagic food consumed in each of the zones. Alternatively, it may reflect different salinity-isotope influences in the prey items and stationary sticklebacks in the two areas in Lake Engervann. There was also a seasonal component showing that sticklebacks were more enriched in 13C and 15N in the Autumn than Spring, likely reflecting differences due to prey production in the two seasons and/or accumulated diet diversity increasing over time. In another study by Nordström et al.  studying threespine stickleback in the Bothnian Bay, Finland, threespine stickleback isotope mean values varied around − 21 for δ13C and 10 for δ15N in a brackish water salinity regime (i.e. 5.2–6.0). In a previous study of 25 North Norwegian threespine stickleback populations we found that δ13C ranged from − 30.5 to − 13.2 (mean ± SD; − 22.3 ± 3.6) and that δ15N ranged from 5.0 to 15.3 (8.9 ± 1.8) (baseline data used in Østbye et al. ). Ravinet et al.  studied two stickleback clades in Japan and found that the upstream populations had a mean δ13C of − 19.4 ‰ and δ15N of 11.9 ‰. Other populations closer to marine environments had means of − 15.8 to − 18.0 ‰ for δ13C and 13.9 to 12.8 ‰ for δ15N. The total mean range in that study was − 15.1 to − 18.0 for δ13C and 11.9 to 14.0 for δ15N, with individual variation of ca − 35.0 to − 15.0 for δ13C and 6 to 15 for δ15N. It was found that the Japan Sea clade were less adapted to fresh water having a higher trophic position utilizing more brackish - marine areas than the Pacific clade. The pattern and range in δ13C (− 22.5 to − 13.0) and δ15N (7.0 to 14.3) in Lake Engervann resembles these comparative studies implying that the lateral plate morphs in Lake Engervann utilize fresh, brackish and marine areas.
Prevalence and intensity of T. Gasterostei
It is not fully understood if T. gasterostei is stenohaline, preferring brackish water habitats or if it is euryhaline and able to handle a large range in salinity. A number of studies have described T. gasterostei from threespine stickleback and other stickleback species in the Holarctic [79, 83–96]. Based on these studies it appears that this parasite can be found in a salinity range of 0.5–32, but are mostly found in brackish water and rarely in fresh water. Thus, it seems reasonable to infer that this parasite is largely distributed and transmitted in brackish water. The prevalence and infection intensity in these studies suggested that infection levels were consistently higher in brackish water than in fresh water and marine environments.
The prevalence of T. gasterostei was similar and high for all three lateral plate morphs in Lake Engervann, with values of 93.4%, 94.0%, 96.1%, for complete, partial and low plated morphs, respectively. The mean intensity was highest in the complete morph (7.1), lower in the partial morph (6.6) and lowest in the low plated morph (4.8), with large standard deviations. Studies of Baltic Sea threespine sticklebacks close to Poland by Morozinska-Gogol [97–99] revealed a prevalence range of 7.3–100% and a mean intensity range of 2.9–113 of T. gasterostei. Prevalence differed among the three Baltic Sea stickleback lateral plate morphs with 69.6%, 67.6% and 44.7% reported for the complete, partial and low plated morphs, respectively. These values are different from our results which are lower in prevalence, but close to the mean intensity range in the Baltic Sea plate morphs (a mean intensity of 6.2 observed in Lake Engervann). Valdez  found in a study of Alaskan threespine stickleback that 41% of the low plated morph compared to 21% of the partially plated morph (few completely plated fish caught) were infected. In contrast, Walkey et al.  observed that T. gasterostei was most common in the marine form (completely plated morph) in England. In Canadian threespine stickleback, Peddle  found a prevalence of T. gasterostei of 8% in full salinity (32.1) seawater. A comparison of T. gasterostei among studies should consider that this species has a seasonal distribution as seen in Mecklenburg area in the Baltic Sea, where the intensity of T. gasterostei was higher in June and July . We found a lower infection of T. gasterostei in Autumn than in Spring which may reflect a seasonal influence due to the life history of the parasite. However, in Poulin and FitzGerald  the highest infection of T. gasterostei was found in three stickleback species in September–November in a salt marsh in Quebec, Canada. Donoghue  report that the highest prevalence and intensity of T. gasterostei on ninespine stickleback during the year was in November and May–June with low prevalence in March. Based on these different studies, in different stickleback species, it is not easy to see a clear temporal pattern in infection-intensity dynamics. Our result in Lake Engervann seems to be within the range of values in other studies, but are somehow distinctive in their high prevalences and relatively low mean intensities.
In the statistical analyses we found that sticklebacks sampled in the upper part of the lake had more T. gasterostei, and that sticklebacks with higher δ15N values had more T. gasterostei. Further, larger-bodied sticklebacks had more T. gasterostei than smaller fish, and females had more T. gasterostei than males. Sticklebacks with fewer and shorter gill rakers were more infected with T. gasterostei. These findings may collectively imply divergent foraging modes and life history adaptations towards divergent habitats within Lake Engervann as well as towards marine habitats.
Prevalence and intensity of Trematoda spp.
With regard to our observed Trematoda spp., the sequence suggested a genetic relationship to the trematode Cryptocotyle lingua which is responsible for the “black spot disease”. However, in contrast to the black spot disease which only occur on the dermal surface on the body - our cysts were only encysted in the fins and less pigmented. Also, our Trematoda spp. cysts appear slightly larger than blackspot cysts from other localities in Norway (visual evaluation; no statistical analysis). Our Trematoda spp. could potentially be e.g. Cryptocotyle concavum as this species have been described from stickleback before  and this taxon has larger cysts than its sister taxon C. lingua . In addition, the common goby appears regularly in Lake Engervann which is a reported second intermediate host after the first intermediate host being the snail Hydrobia stagnalis and the final host being Larus ridibundis for C. concavum (more species of gulls can be final hosts; see Zander et al. ). Unfortunately, C. concavum has no sequences entered in Genbank. It may be that our Trematoda spp. has a transmission route that resembles Cryptocotyle lingua which is being transmitted through the intermediate host periwinkles (Littorina littorina) and its main host Larus spp.  or resembles C. concavum transmitted via H. stagnalis snails  to its main host Larus spp. Littorina littorina is more marine than Hydrobia ulvae, the dominant gastropod in Lake Engervann. Thus, Trematoda spp. may be transmitted via second intermediate host of H. ulvae in Lake Engervann or that sticklebacks with high infection levels are foraging in more marine environments than in Lake Engervann. Support comes from Möller  who found in a salinity-temperature and survival experiment of different stages of C. lingua that a salinity below 4 resulted in 50% of eggs developing and that salinities above 8 increased the living time of free swimming cercaria. It was also shown that this parasite preferred lower temperatures that were associated with increased survival time. Lake Engervann is shallow, resulting in higher mean temperatures, as well as more marked temperature fluctuations, relative to more thermally-stable marine environments. As such, it seems likely that transmission of our Trematoda spp. occurs more in marine-like habitats.
We observed that sticklebacks in the lower area had more Trematoda spp., and a higher infection of Trematoda spp. in Autumn than Spring, which fits with seasonal foraging migrations into the marine environment. We have observed that in the middle of summer often there is a very low catch of threespine stickleback in Lake Engervann which could support such a seasonal migration pattern. Furthermore, the finding that the low plated morph had fewer Trematoda spp. than the other morphs, and that sticklebacks with larger gill raker length had more Trematoda spp., may also support a seasonal migration scenario that may differ among morphs. This pattern may fit with theoretical expectations of a divergent trait-associated benthic-pelagic foraging mode assuming that T. gasterostei is transmitted in littoral-benthic areas in brackish water, while Trematoda spp. is transmitted in more marine influenced areas where sticklebacks may also forage more on pelagic zooplankton.
Body length patterns in the stickleback lateral plate morphs in Lake Engervann
The three lateral plate morphs in Lake Engervann differed significantly in body length with the low plated morph being smallest and the completely plated morph being largest, with the partially plated morph in between. Ravinet et al.  found that the three plate morphs of the threespine stickleback that coexisted in the Lough Furnace, Burrishole Catchment, Western Ireland were genetically segregated. Here, a small bodied freshwater stationary form seemed to enter (or being washed in to the lake passively) brackish water areas where a larger brackish water resident form and an even larger anadromous form of stickleback occurs. In contrast, the three Lake Engervann plate morphs belonged to one genetic population. Furthermore, our low plated morph is more genetically related to the partial and complete morph in the lagoon than to an upstream low plated morph in the Sandvikselva River (unpublished results).
Threespine sticklebacks in Lake Engervann live in a brackish water environment, but we do not know if they are truly stationary in the lake or if they are conducting migrations into the sea or to fresh water. However, our stable isotopic data suggest sticklebacks use both brackish water and marine areas for feeding. In a common garden experiment, Marchinko & Schluter  observed that offspring of reduced lateral plate morphs (partial and low) grew equally well as the offspring of the completely plated morph if raised in salt water, but that the former group grew better in fresh water. We are aware of no laboratory experiment looking at specific growth patterns of the completely plated, partially plated and low plated morphs raised under brackish water conditions. In another set of laboratory rearing experiments of stationary and anadromous threespine sticklebacks in the Navarro River, USA, Snyder & Dingle  and Snyder  suggested a genetic basis for life history variation implying different migratory lifestyles had evolved as reflected in genetic variation for size and growth. We expect that the three lateral plate morphs in Lake Engervann should have attained the same body size if they lived in the same environment and if they had the same niche, when assuming no genetic influence on growth by the three Eda genotypes. An interesting study in that regard was conducted by Robertson et al.  who found the aa low plated fish had a lower growth rate than the faster growing AA completely plated fish in an experiment exposing three Eda genotypes to natural marine and freshwater sites, which could support our results. Bowles et al.  found a genetic basis for body size variation between anadromous and derived lacustrine populations of threespine sticklebacks in southwest Alaska when crossing and raising offspring in water of salinity 4–6. The authors concluded that body size is a heritable trait that could be plastic as well as influenced by natural- and sexual selection. Thus, our results suggest that different body lengths of the lateral plate morphs may reflect either phenotypic plasticity in divergent niches and/or that growth, foraging and life histories are contingent upon genetic architecture associated with the Eda block itself.
Ongoing divergent niche differentiation with high gene flow in the lagoon?
The utilization of a diversity of habitats, and thus niches, could render individuals and populations exposed to natural selection in divergent salinity environments, where selection introduce and trade-off conflicting trait adaptations. In the brackish water zone, the threespine stickleback may consist of such diversity in phenotypes and physiological traits due to hybridization among fresh water, brackish water and marine sticklebacks. Thus, adapting to the brackish water zone itself may be difficult due to high gene flow for the sticklebacks that are prone to be more brackish water resident. In our study, we argue for finding signs of ongoing divergent niche adaptation among the three lateral plate morphs based on ecologically related measures such as stable isotope values and parasites. Further, the ecological preferences may somehow be associated with the plate morphs through the Stn382 link to the Eda haplotype block since sticklebacks belonged to one panmictic gene pool. The higher stable isotope variance seen in the low plated fish in Lake Engervann may imply a more explorative behaviour - which may fit with the salinity preference experiment by Barrett et al. (2009b). The plate morphs in Lake Engervann also have different foraging efficiency on pelagic Daphnia and benthic Chironomidae . In an experiment on standard metabolic rate (SMR), as a cost for osmoregulation in salinity environments, we found that sticklebacks were able to move among salinity environments without short-term metabolic costs, irrespective of environment of origin from marine, brackish or fresh water, with no differences in SMR found among the three lateral plate morphs in Lake Engervann . This suggests that these three plate morphs may be equally successful in different salinity environments, at least on a short time scale. This physiological ability thus seems to have been evolutionarily conserved within each plate morph. Further support for a multi-adaptive-trait association with the Eda block can be found in Greenwood et al. , who documented that schooling behavior was associated with Eda in benthic low plated and pelagic completely plated morphs. Also, the study by Robertson et al.  adds another dimension to the adaptive diversification of Eda genotypes and plate morphs as genotypes were found to differ in immune system genes associated with parasite load. The presence of sticklebacks from marine to brackish and freshwater environments intuitively suggest highly divergent selection pressures where one should expect adaptations to evolve. An experimental study by Peeke & Morgan  found differences in the response to behavioural stimuli associated with aggression, courtship and feeding in sticklebacks in marine, estuarine and upstream freshwater river in California, USA. This points to adaptive behavioural modifications along the marine-freshwater transect. Barrett et al.  observed in an experiment that cold tolerance could evolve rapidly in threespine stickleback from marine to fresh water, which could be different in colder brackish water lagoons than in warmer marine environments. Interestingly, DeFaveri and Merilä , De Faveri et al.  and Guo et al.  observed signatures of adaptive genetic differentiation associated with salinity and temperature gradients in the brackish water Baltic Sea threespine stickleback. These authors suggested that the same adaptive processes could occur in brackish water as in the marine-freshwater environment transects. Thus, the more stationary threespine sticklebacks in Lake Engervann could experience more variable temperature selection pressures than the more migratory-marine individuals. Further, being stationary in brackish water may also prime a temperature selection pressure for cold tolerance that could be benign when colonizing fresh water. Jones et al.  found a repeated pattern documenting that three haplotype blocks were important in differentiating marine and freshwater sticklebacks on the Holarctic scale, comprising genes associated with mucus production, salinity tolerance and lateral plate plates. Moreover, De Faveri et al.  found that physiological adaptation in freshwater-marine stickleback populations can follow alternative routes for genomic adaptation. A set of studies show that both de novo evolution and selection on standing genetic variation are involved in the marine-brackish water-freshwater stickleback adaptation [5, 6, 68, 109, 110]. Further research should aim at investigate the broad-spectra of physiology, and its plasticity, in different plate morphs to reveal their relative contribution upon adaptive radiation potential in different environments.
The association of morphs, Eda, body length and ecology suggest that the three lateral plate morphs in the brackish water Lake Engervann are experiencing ongoing divergent selection for niche and migratory life history strategies even under high gene flow level as all the three plate morphs belonged to the same gene pool. The brackish water zone may as such act as a generator in a continuous adaptive iterative process where the plate morphs/Eda genotypes derive adaptive traits, resulting in novel functional genomic diversity to be selected upon in the different environments.
We thank Eivind Østbye for field assistance and scientific discussions, and referees that improved the quality of the manuscript. Laboratory assistance with regard to genetic methods was kindly supplied by Vicky Albert, Guillaume Côté, Lucie Pappillon, Christian Landry, Catherine Potvin and Robert St-Laurent at Louis Bernatchez population genetic lab, University of Laval, Quebec, Canada. We thank S. Rogers, D. Schluter and C. Peichel for help with microsatellites. CH thanks Diethard Tautz, Winfried Lampert and the Max Planck Society for financial support. We thank H. Buhtz and A. Möller for help in the stable isotopic laboratory.
This study was financially supported by the Research Council of Norway (grant no 170755/V20 to LAV).
Availability of data and materials
The dataset and additional analyses supporting the conclusions are available here: Additional file 1: Table S1, Additional file 2: Table S2, Additional file 4: Table S3, Additional file 5: Table S4, Additional file 8: Table S5, Additional file 9: Table S6 and Additional file 3: Figure S3, Additional file 6: Figure S1, Additional file 7: Figure S2.
KØ, LB and LAV conceived the study. KØ and CH planned the study design. KØ performed the majority of the field work. KØ and AT collected prey items for stable isotope analyses. KØ performed morphometric analyses and most of the genetic analyses with support of the laboratory personel of LB. AT performed parts of the Stn382 genetic laboratory analysis. RAP performed genetic analyses of Trematoda. CH and MR conducted the stable isotope analyses. KØ and LAV performed statistical analyses. KØ wrote the manuscript with major contributions from AT and LAV. All authors read and contributed to manuscript writing, and approved the final manuscript.
Sampling was conducted under a fishing permit (Fisketillatelse 10/2016; 2006/-SNO-1/TW) by the Norwegian Directorate for Nature Management. Sticklebacks were euthanized using an overdose of MS222. Care was taken to minimize suffering of fish.
Consent for publication
The authors declare that they have no competing interest.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
- Bell MA, Foster SA. The evolutionary biology of the threespine stickleback. 1st ed. Oxford: Oxford University Press; 1994.Google Scholar
- Klepaker TO. Postglacial evolution in lateral plate morphs in Norwegian freshwater populations of the three-spine stickleback (Gasterosteus aculeatus). Can J Zool. 1995;73:898–906.View ArticleGoogle Scholar
- Klepaker TO. Lateral plate polymorphism in marine and estuarine populations of the threespine stickleback (Gasterosteus aculeatus) along the coast of Norway. Copeia. 1996;4:832–8.View ArticleGoogle Scholar
- Taylor EB, McPhail JD. Historical contingency and ecological determinism interact to prime speciation in sticklebacks, Gasterosteus. Proc Royal Soc B. 2000;267:2375–84.View ArticleGoogle Scholar
- Colosimo PF, Hoseman KE, Balabhadra S, Villareal G Jr, Dickson M, Grimwood J, Schmutz J, Myers RM, Schluter D, Kingsley DM. Widespread parallel evolution in sticklebacks by repeated fixation of ectodysplasin alleles. Science. 2005;307:1928–33.PubMedView ArticleGoogle Scholar
- 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. Current Biol. 2012;22:83–90.View ArticleGoogle Scholar
- Hagen DW, Gilbertson LG. Geographic variation and environmental selection in Gasterosteus aculeatus in Pacific Northwest. Am. Evol. 1972;26:32–43.View ArticleGoogle Scholar
- Reimchen TE. Structural relationships between spines and lateral plates in threespine stickleback (Gasterosteus aculeatus). Evolution. 1983;37:931–46.PubMedGoogle Scholar
- Spence R, Wootton RJ, Barber I, Przybylski M, Smith C. Ecological causes of morphological evolution in the three-spined stickleback. Ecol Evol. 2013;3:1717–26.PubMedPubMed CentralView ArticleGoogle Scholar
- Bell MA, Aguirre WE, Buck NJ. Twelve years of contemporary armor evolution in a threespine stickleback population. Evolution. 2004;58:814–24.PubMedView ArticleGoogle Scholar
- Barrett RDH, Rogers SM, Schluter D. Natural selection on a major armor gene in threespine sticklebacks. Science. 2008;322:255–7.PubMedView ArticleGoogle Scholar
- Bell MA, Gangavalli AK, Bewick A, Aguirre WE. Frequency of Ectodysplasin alleles and limited introgression between sympatric threespine stickleback populations. Environ Biol Fish. 2010;89:189–98.View ArticleGoogle Scholar
- 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.PubMedView ArticleGoogle Scholar
- Raeymaekers JAM, Konijnendijk N, Larmuseau MHD, Hellemans B, De Meester L, Volckaert FAM. A gene with major phenotypic effects as a target for selection vs. homogenezing gene flow. Mol Ecol. 2014;23:162–81.PubMedView ArticleGoogle Scholar
- Rennison DJ, Heilbron K, Barrett RDH, Schluter D. Discriminating selection on lateral plate phenotype and its underlying gene, Ectodysplasin, in threespine stickleback. Am. Nat. 2015;185:150–6.PubMedView ArticleGoogle Scholar
- Ravinet M, Hynes R, Poole R, Cross TF, McGinnity P, Harrod C, Prodöhl PA. Where the lake meets the sea: strong reproductive isolation is associated with adaptive divergence between lake resident and anadromous three-spined sticklebacks. Plos One. 2015; doi:10.1371/journal.pone.0122825.
- Raeymaekers JAM, Van Houdt JKJ, Larmuseau MHD, Geldof S, Volckaert FAM. Divergent selection as revealed by PST and QTL-based FST in three-spined stickleback (Gasterosteus aculeatus) populations along a coastal-inland gradient. Mol Ecol. 2007;16:891–905.PubMedView ArticleGoogle Scholar
- Kitano J, Bolnick DI, Beauchamp DA, Mazur MM, Mori S, Nakano T, Peichel C. Reverse evolution of armor plates in the threespine stickleback. Curr Biol. 2008;18:769–74.PubMedView ArticleGoogle Scholar
- Barrett RD, Rogers SM, Schluter D. Environment specific pleiotropy facilitates divergence at the Ectodysplasin locus in threespine stickleback. Evolution. 2009a;63:2831–7.PubMedView ArticleGoogle Scholar
- Marchinko KB, Schluter D. Parallel evolution by correlated response: lateral plate reduction in threespine stickleback. Evolution. 2007;61:1084–90.PubMedView ArticleGoogle Scholar
- Barrett RDH, Vines TH, Bystriansky JS, Schulte PM. Should I stay or should I go? The Ectodysplasin locus is associated with behavioural differences in threespine stickleback. Biol Lett. 2009b;5:788–91.PubMedPubMed CentralView ArticleGoogle Scholar
- Greenwood AK, Mills MG, Wark AR, Archambeault SL, Peichel CL. Evolution of schooling behavior in threespine sticklebacks is shaped by the Eda gene. Genetics. 2016;203:677–81.PubMedPubMed CentralView ArticleGoogle Scholar
- Robertson S, Bradley JE, MacColl ADC. Eda haplotypes in three-spined stickleback are associated with variation in immune gene expression. Nature Sci Rep. 2017;7:42677.View ArticleGoogle Scholar
- Hendry AP, Taylor EB. How much of the variation in adaptive divergence can be explained by gene flow? An evaluation using lake/stream stickleback pairs. Evolution. 2004;58:2319–31.PubMedView ArticleGoogle Scholar
- Hendry AP. Selection against migrants contributes to the rapid evolution of ecologically dependent reproductive isolation. Evol Ecol Res. 2004;6:1219–36.Google Scholar
- Snyder RJ, Dingle H. Adaptive, genetically based differences in life history between estuary and freshwater threespine stickleback (Gasterosteus aculeatus L.). Can J Zool. 1989;67:2448–54.View ArticleGoogle Scholar
- Peeke HVS, Morgan LE. Behavioral, differentiation of adjacent marine and fluvial populations of threespine stickleback in California: A laboratory study. Behaviour. 2000;137:1011–27.View ArticleGoogle Scholar
- Webster MM, Atton N, Hart PJB, Ward AJW. Habitat-specific morphological variation among threespine sticklebacks (Gasterosteus aculeatus) within a drainage basin. Plos One. 2011;6:e21060.PubMedPubMed CentralView ArticleGoogle Scholar
- McCairns RJS, Bernatchez L. Plasticity and heritability of morphological variation within and between parapatric stickleback demes. J Evol Biol. 2012;25:1097–112.PubMedView ArticleGoogle Scholar
- Guo B, DeFaveri J, Sotelo G, Nair A, Merilä J. Population genomic evidence for adaptive differentiation in Baltic Sea three-spined sticklebacks. BMC Biol. 2015;13:19.PubMedPubMed CentralView ArticleGoogle Scholar
- Konijnendijk N, Shikano T, Daneels S, Volckaert FAM, Raeymaekers JAM. Signatures of selection in the three-spined stickleback along a small-scale brackish water – freshwater transition zone. Ecol Evol. 2015;5:4174–86.PubMedPubMed CentralView ArticleGoogle Scholar
- Blindheim T, Løvdal I, Olsen KM. Naturfaglige registreringer og vurderinger i forbindelse med utbygging av nytt dobbeltspor Sandvika-Lysaker, Bærum kommune. - Siste Sjanse-Rapport 2005-1: 53 s + vedlegg (In Norwegian).Google Scholar
- Halvorsen G, Often A, Svalastog D. Engervannet og Øverlandselva – statusrapport. NINA Minirapport. 2005;2005:136. (In Norwegian)Google Scholar
- Kjensmo J. Electrolytes in Norwegian lakes. Schweiz Z Hydrol. 1966;28:29–42.Google Scholar
- Vennerød K. Vassdragsundersøkelser. En metodebok i limnologi. Oslo: Universitetsforlaget; 1984. p. 283. (In Norwegian)Google Scholar
- Breder CM. Design for a fry trap. Zoologica. 1960;45:155–9.Google Scholar
- Banbura J. Lateral plate number development in the complete morph of the three-spined stickleback, Gasterosteus aculeatus L. Zool Scripta. 1989;18:157–9.View ArticleGoogle Scholar
- Grey J. Trophic fractionation and the effects of diet switch on the carbon stable isotope 'signatures' of pelagic consumers. Verh Int Vere Theor Angew Limn. 2000;27:3187–91.Google Scholar
- 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.View ArticleGoogle Scholar
- Post DM. Using stable isotopes to estimate trophic position: models, methods, and assumptions. Ecology. 2002;1:703–18.View ArticleGoogle Scholar
- 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.PubMedView ArticleGoogle Scholar
- Harrod C, Mallela J, Kahilainen K. Phenotype-environment correlations in a putative whitefish adaptive radiation. J Anim Ecol. 2010;79:1057–68.PubMedView ArticleGoogle Scholar
- 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.View ArticleGoogle Scholar
- Yamaguti S. Parasitic copepoda and branchiura of fishes. New York: Interscience; 1963.Google Scholar
- Reichenback-Klinke H, Elkan E. The principal diseases of lower vertebrates. London, New York: Academic press; 1965.Google Scholar
- van Duijn C Jr. Diseases of fishes. 2nd ed. Charles C. Thomas, publisher: Springfield, Illinois; 1967.Google Scholar
- Poulin R, FitzGerald GJ. The potential of parasitism in the structuring of a salt marsh stickleback community. Can J Zool. 1987;65:2793–8.View ArticleGoogle Scholar
- Zander CD, Kocoglu Ö, Skroblies M, Strobach U. Parasite populations and communities from the shallow littoral of the Orther Bight (Fehmarn, SW Baltic Sea). Parasitol Res. 2002;88:734–44.PubMedView ArticleGoogle Scholar
- Zietara MS, Huyse T, Lumme J, Volkaert FA. Deep divergence among subgenera of Gyrodactylus inferred from rDNA ITS region. Parasitology. 2002;124:39–52.PubMedView ArticleGoogle Scholar
- Tamura K, Stecher G, Peterson D, Filipski A, Kumar S. MEGA6: Molecular Evolutionary Genetics Analysis version 6.0. Mol Biol Evol. 2013;30:2725–9.PubMedPubMed CentralView ArticleGoogle Scholar
- Rzhetsky A, Nei M. A simple method for estimating and testing minimum evolution trees. Mol Biol Evol. 1992;9:945–67.Google Scholar
- Felsenstein J. Confidence limits on phylogenies: an approach using the bootstrap. Evolution. 1985;39:783–91.PubMedView ArticleGoogle Scholar
- Tamura K, Nei M, Kumar S. Prospects for inferring very large phylogenies by using the neighbor-joining method. PNAS. 2004;101:11030–5.PubMedPubMed CentralView ArticleGoogle Scholar
- Nei M, Kumar S. Molecular Evolution and Phylogenetics. New York: Oxford University Press; 2000.Google Scholar
- Saitou N, Nei M. The neighbor-joining method: A new method for reconstructing phylogenetic trees. Mol Biol Evol. 1987;4:406–25.PubMedGoogle Scholar
- Sambrook J, Fritch EF, Maniatis T. Molecular Cloning: A Laboratory Manual. 2nd ed. New York: Cold Spring Harbor Laboratory Press; 1989.Google Scholar
- Klepaker TO, Østbye K, Bernatchez L, Vøllestad LA. Spatio - temporal patterns in pelvic reduction in threespine stickleback in Lake Storvatnet. Evol Ecol Res. 2012;14:169–91.Google Scholar
- van Oosterhout C, Hutchinson WF, Wills DPM, Shipley P. MICRO-CHECKER: software for identifying and correcting genotyping errors in microsatellite data. Mol Ecol Notes. 2004;4:535–8.View ArticleGoogle Scholar
- Beaumont MA, Nichols RA. Evaluating loci for use in the genetic analysis of population structure. Proc R Soc Lond B. 1996;363:1619–26.View ArticleGoogle Scholar
- Antao T, Lopes A, Lopes RJ, Beja-Pereira A, Luikart G. LOSITAN: a workbench to detect molecular adaptation based on a Fst-outlier method. BMC Bioinformatics. 2008;9:323.PubMedPubMed CentralView ArticleGoogle Scholar
- Raymond M, Rousset F. GENEPOP (version 1.2): population genetics software for exact test and ecumenism. J Hered. 1995;86:248–9.View ArticleGoogle Scholar
- Pritchard JK, Stephens M, Donnelly P. Inference of population structure using multilocus genotype data. Genetics. 2000;155:945–59.PubMedPubMed CentralGoogle Scholar
- Evano G, Regnaut S, Goudet J. Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Mol Ecol. 2005;14:2611–20.View ArticleGoogle Scholar
- Earl DA, von Holdt BM. STRUCTURE HARVESTER: a website and program for visualizing STRUCTURE output and implementing the Evano method. Conserv Genet Res. 2012;4:359–61.View ArticleGoogle Scholar
- SAS Institute Inc. Using JMP 9. SAS Institute Inc., Cary, N.C. 2010.Google Scholar
- Borges JN, Skov J, Bahlol QZM, Møller OS, Kania PW, Santos CP, Buchmann K. Viability of Cryptocotyle lingua metacercariae from Atlantic cod (Gadus morhua) after exposure to freezing and heating in the temperature range from -80 oCto 100 oC. Food Control. 2015;50:371–7.View ArticleGoogle Scholar
- Dzikowski R, Levy MG, Poore MF, Flowers JR, Paperna I. Use of rDNA polymorphism for identification of Heterophyidae infecting freshwater fish. Dis Aquat Org. 2004;59:35–41.PubMedView ArticleGoogle Scholar
- DeFaveri J, Jonsson PR, Merilä J. Heterogeneous genomic differentiation in marine threespine stickleback: Adaptation along an environmental gradient. Evolution. 2013;67:2530–46.PubMedView ArticleGoogle Scholar
- Gross HP. Natural selection by predtors on the defensive apparatus of the three-spined stickleback, Gasterosteus aculeatus L. Can J Zool. 1978;56:398–413.View ArticleGoogle Scholar
- Klepaker TO, Østbye K. Pelvic anti-predator armour reduction in Norwegian populations of the threespine stickleback: a rare phenomenon with adaptive implications? J. Zool. 2008;276:81–8.View ArticleGoogle Scholar
- Lucek K, Roy D, Bezault E, Sivasundar A, Seehausen O. Hybridization between distant lineages increases adaptive variation during a biological invasion: stickleback in Switzerland. Mol Ecol. 2010;19:3995–4011.PubMedView ArticleGoogle Scholar
- Lucek K, Haesler MP, Sivasundar A. When phenotypes do not match genotypes – unexpected phenotypic diversity and potential environmental constraints in Icelandic stickleback. J. Heredity. 2012;103:579. ess021View ArticleGoogle Scholar
- Taugbøl A, Arntsen T, Østbye K, Vøllestad LA. Small changes in gene expression of targeted osmoregulatory genes when exposing marine and freshwater threespine stickleback (Gasterosteus aculeatus) to abrubt salinity transfers. PLoS One. 2014;9:e106894.PubMedPubMed CentralView ArticleGoogle Scholar
- Blegvad H. On the food of fish in the Danish waters within the Skaw. RepDanish Biol Sta. 1917;24:19–72.Google Scholar
- Thorman S, Wiederholm A-M. Seasonal occurrence and food resource use of an assemblage of nearshore fish species in the Bothnian Sea, Sweden. Mar Ecol Prog Series. 1983;10:223–9.View ArticleGoogle Scholar
- Delbeek JC, Williams DD. Food resource partitioning between sympatric populations of brackishwater sticklebacks. J Anim Ecol. 1987;56:949–67.View ArticleGoogle Scholar
- Sánchez-Gonzáles S, Ruiz-Campos G, Contreras-Balderas S. Feeding ecology and habitat of the threespine stickleback, Gasterosteus aculeatus microcephalus, in a remnant population of northwestern Baja California, Mexico. Ecol Freshw Fish. 2001;10:191–7.View ArticleGoogle Scholar
- Demchuk A, Ivanov M, Ivanova T, Polyakova N, Mas-Martì E, Lajus D. Feeding patterns in seagrass beds of three-spined stickleback Gasterosteus aculeatus juveniles at different growth stages. J Mar Biol Ass of UK. 2015;95:1635–4.View ArticleGoogle Scholar
- Rybkina EV, Demchuk LDL, Ivanova TS, Ivanov MV, Galaktionov KV. Dynamics of parasite community during early ontogenesis of marine threespine stickleback, Gasterosteus aculeatus. Evol Ecol res. 2016;17:335–54.Google Scholar
- Nordström MC, Lindblad P, Aarnio K, Bonsdorff E. A neighbour is a neighbour? Consumer diversity, trophic function, and spatial variability in benthic food webs. J Exp Marine Biol and Ecol. 2010;391:101–11.View ArticleGoogle Scholar
- Østbye K, Harrod C, Gregersen F, Klepaker T, Schulz M, Schluter D, Vøllestad LA. 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 Evol Biol. 2016;16:102.PubMedPubMed CentralView ArticleGoogle Scholar
- Ravinet M, Takeuchi N, Kume M, Mori S, Kitano J. Comparative analysis of Japanese three-spined stickleback clades reveals the Pacific Ocean lineage has adapted to freshwater environments while the Japan Sea has not. Plos ONE. 2014;9:e112404. https://doi.org/10.1371/journal.pone.0112404.
- Scott T, Scott A. Brittish parasitic copepoda. 1. London: Ray Society; 1913.Google Scholar
- Shulman SS, Shulman-Albova RE. Parasites of fishes of the White Sea. Akad. Nsauk. S.S.S.R: Moskow; 1953. [In Russian]Google Scholar
- Isakov LS, Shulman SS. The tolerance of some ectoparasites of sticklebacks to changes in salinity. Fish Res Bd Can Transl Series. 1966;780: 1-13.Google Scholar
- Bykovskaya-Pavlovskaya IE, Gusev AV, Dubinina NA, Izyumova TS, Smirnova IL, Sokolovskaya GA, Shtein GA, Shulman SS, Epshtein VM. Key to parasites of freshwater fish of the U.S.S.R. Jerusalem: Israel Program for Scientific Translations. 1964;Part I:180–218.Google Scholar
- Threlfall W. A mass die-off of three-spined sticklebacks (Gasterosteus aculeatus L.) caused by parasites. Can J Zool. 1968;46:105–6.PubMedView ArticleGoogle Scholar
- Hanek G, Threlfall W. Thersitina gasterostei (Pagenstecher, 1861) (Copepod: Ergasilidae) from Gasterosteus wheatlandi Putnam 1867. Can J Zool. 1969;47:627–9.View ArticleGoogle Scholar
- Walkey M, Lewis DB, Dartnall HCJ. Observations on the host-parasite relationship of Thersitina gasterostei (Crustacea: Copepoda). J Zool, London. 1970;162:371–81.View ArticleGoogle Scholar
- Valdez RA. Two parasites of threespine stickleback from Amchitka, Aleution Islands. Alaska. Trans Am Fish Soc. 1974;103:632–5.View ArticleGoogle Scholar
- Donoghue S. Thersitina gasterostei (PAGENSTECHER, 1861) (Copepods: Ergasilidae) infecting the stickleback Pungitus pungitus L. at Chalk Marshes, Gravesend, Kent. Ann. Parasitol. Hum Comp. 1986;6:673–82.View ArticleGoogle Scholar
- Levsen A. Parasites of the three-spined stickleback Gasterosteus aculeatus from Norway, with emphasis on trochodinid ciliates. Fauna. 1992;45:40–8. [In Norwegian]Google Scholar
- Zander CD, Reimer LW, Barz K. Parasite communities of the Salzhaff (Northwest Mecklenburg, Baltic Sea). I. Structure and dynamics of communities of littoral fish, especially small-sized fish. Parasitol Res. 1999;85:356–72.PubMedView ArticleGoogle Scholar
- Peddle JC. Biodiversity and community ecology of the parasites of the three-spine stickleback, Gasterosteus aculeatus, in the southern Gulf of St. Lawrence. The university of New Brunswick: Master thesis; 2004.Google Scholar
- Savoie VL. Taxonomy and ecology of Ergasilus sp. and Thersitina gasterostei (Copepoda) parasitizing gasterosteiforms along the coast of the Atlantic Canadian Provinces. Master of Science in Applied Sciences, Biology Department, Saint Mary`s University, Halifax, Nova Scotia. 2004.Google Scholar
- Zander CD. Parasite diversity of sticklebacks from the Baltic Sea. Parasit Res. 2007;100:287–97.View ArticleGoogle Scholar
- Morozinska-Gogol J. Dynamics of select parasite infestation of the three-spined stickleback in dependence on the place of catching in the southern Baltic. Baltic Coastal Zone. 1999;3:77–88.Google Scholar
- Morozinska-Gogol J. Correlation between morphological forms of the three-spined stickleback and parasites infestation in the Baltic Sea. Baltic Coastal Zone. 2000;4:87–94.Google Scholar
- Morozinska-Gogol J. Changes in the parasite communities as one of the potential causes of decline in abundance of the three-spined sticklebacks in the Puck Bay. Oceanologia. 2015;57:280–7.View ArticleGoogle Scholar
- Wootton DM. The life history of Cryptocotyle concavum (Creplin, 1825) Fishchoeder, 1903 (Trematoda: Heterophyidae). J Parasitology. 1957;43:271–9.View ArticleGoogle Scholar
- ElMayas H, Kearn GC. In vitro excystment of the metacercaria of Cryptocotyle concavum from the common goby Pomatoschistus microps. J Helmintology. 1995;69:285–97.View ArticleGoogle Scholar
- Zander CD, Kollra H-G, Antholz B, Meyer W, Westphal D. Small-sized euryhaline fish as intermediate hosts of the digenetic trematoda Cryptocotyle concavum. Helgoländer Meeresunters. 1984;37:433–43.View ArticleGoogle Scholar
- Möller H. The effects of salinity and temperature on the development and survival of fish parasites. J Fish Biol. 1978;12:311–23.View ArticleGoogle Scholar
- Snyder RJ. Migration and life histories of the threespine stickleback: evidence for adaptive variation in growth rate between populations. Env Biol Fish. 1991;31:381–8.View ArticleGoogle Scholar
- Bowles E, Johnston RA, Vanderswan SL, Rogers SM. Genetic basis for body size variation between an anadromous and two derived lacustrine populations of threespine stickleback Gasterosteus aculeatus in southwest Alaska. Current Zoology. 2016;62:71–8.View ArticleGoogle Scholar
- 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.View ArticleGoogle Scholar
- Grøtan K, Østbye K, Taugbøl A, Vøllestad LA. No short-term effect of Salinity on oxygen consumption in threespine stickleback (Gasterosteus aculeatus) from fresh, brackish, and salt water. Can J Zool. 2012;90:1386–93.View ArticleGoogle Scholar
- Barret RDH, Paccard A, Healy TM, Bergek S, Schulte PM, Schluter D, Rogers SM. Rapid evolution of cold tolerance in stickleback. Proc Royal Soc B. 2010; doi:10.1098/rspb.2010.0923.
- DeFaveri J, Merilä J. Evidence for adaptive phenotypic differentiation in Baltic Sea sticklebacks. J Evol Biol. 2013;26:1700–15.PubMedView ArticleGoogle Scholar
- 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.PubMedView ArticleGoogle Scholar