Effects of whitefish speciation on piscivorous birds

Master thesis, 30 ECTS

Effects of whitefish speciation

on piscivorous birds

A dietary study of piscivorous birds in

central and northern Sweden

Abstract

The ecological communities we observe today are a product of the bidirectional interactions between ecological and evolutionary processes. Although the effects of ecological processes on population divergence and speciation have been studied extensively, far less is known about the effects of divergence and speciation on ecological dynamics. This is especially true for effects of ecological speciation processes on higher trophic levels. In this thesis I focus on how divergence in the European whitefish (Coregonus lavaretus) affects a guild of six piscivorous bird species. Previous studies have indicated that population densities of these species are higher on lakes with polymorphic whitefish than on lakes with monomorphic whitefish. Here I test the hypothesis that the high densities of piscivores is a response to the rich food resource provided by dwarf whitefish ecotypes, which are of suitable size and occur in very high abundance.

To test this hypothesis I analyzed fecal samples from piscivorous birds in lakes with polymorphic whitefish, using samples from lakes with monomorphic whitefish as controls. With the method of ddPCR (digital droplet Polymerase Chain Reaction) the amount of DNA from different prey fish species in the droppings of six fish-eating birds was quantified and converted to proportional abundances. The results shows that the proportion of whitefish in the diet of the entire fish-eating guild was significantly higher in lakes with polymorphic whitefish (44%) than in lakes with monomorphic whitefish (18%). Species-level analyses showed that this result also holds for both black-throated loon (Gavia artica) and red-throated loon (Gavia stellata). Common merganser (Mergus merganser), red-red-throated merganser (Mergus serrator) and terns (Sterna paradisaea and Sterna hirundo) did not show any difference between the two lake categories.

Thus, my study supports the idea that the evolution of small-sized whitefish ecotypes provides a profitable food source for piscivorous birds. However, the finding that only some species of piscivorous bird populations seem to rely heavily on dwarfed whitefish as food suggests that also some other aspect of the speciation process may favor these species. Thus, more studies are needed to further assess what effects polymorphic whitefish have on piscivorous bird populations.

Table of content

1 Introduction

1.1. Background

1.2. Aim

2 Method

2.1. Study sites

2.2. Sampling

2.3. Extraction of samples

2.4. Primers

2.5. ddPCR-analysis

2.6. Statistical analysis

3 Results

4 Discussion

5 Conclusion

6 Acknowledgments

7 References

8 Appendix

8.1. Studied lakes with and without pike

1 Introduction

1.1. Background

The complex interaction between ecological processes such as colonization and extinction, and evolutionary processes such as speciation and adaptation, have formed the ecological communities we observe today (Myers and Giller 1988). The interaction is bidirectional: Ecological interactions between species create the selection pressures that direct evolution, and evolutionary processes shape species traits that determine the outcome of ecological interactions (Koene, Crotti, Elmer and Adams 2019). That evolution, in the form of

speciation processes, affects the ecological community’s function is apparent, since it forms the species who interact (Fowler and MacMahon 1982, Post and Palkovacs 2009). Yet, few studies have quantified the effects of recent speciation processes on ecological processes. Aquatic freshwater species found in lakes have adapted to local conditions since the withdrawal of the last glaciation, around 10 000 years ago, when large areas were either submerged or covered by glaciers. As the glaciers retreated, aquatic species have sporadically migrated through barriers and obstacles, hence colonizing freshly formed lakes (Bernatchez and Wilson 1998, Fraser and Bernatchez 2005, Wilson and Hebert 1998). This event allowed species to enter new niches and subsequently to diversify in sympatric conditions (Fraser and Bernatchez 2005). It is known that some aquatic freshwater fish species have formed

genetically distinct ecotypes during Holocene (Raeymaekers, Maes, Audenaert and Volckaert 2005, Häkli et al. 2018). How the evolution of these ecotypes, who represent an early stage in a speciation process, affects ecological interactions remains relatively unknown. An exception is a study by Harmon et al. (2009) of predation effects of two stickleback-ecotypes on lower trophic levels. This experimental study showed that the evolution of ecotypes influenced properties such as lower-level species composition (especially the pelagic prey community) and the composition of DOC (Dissolved Organic Carbon). This would suggest the ecotypes act as ecosystem engineers. However, scarcely any light has been shed on how top predators are affected by lower trophic levels.

One of the few examples on how predators are affected by the evolution of ecotype concerns divergence in the European whitefish (Coregonus lavaretus). A study of lakes in northern Finland by Thomas et al. (2017) showed that polymorphic whitefish lakes had higher

abundance of brown trout (Salmo trutta) than monomorphic whitefish lakes, supposedly due to an enlarged trophic niche for the brown trout. Swedish studies of predatory fish show that the evolution of whitefish ecotypes, in particular dwarfed ecotypes, leads to a shift from benthivory to piscivory, causing predators such as perch and brown trout to target small sized whitefish ecotypes to a greater extent (Karlberg 2017, Jansson 2017). This shift leads to faster growth and higher abundance of large-bodied individuals of these fish species.

An intriguing aspect of this process is that this divergence in the European whitefish is driven by the presence of another predator, the northern pike (Esox lucius) (Öhlund et al 2020). Thus we have here an example of facilitation, rather than competition, between the different predators. The whitefish ecotypes, which differ in body size structure and habitat use, are often referred to as benthic giants and pelagic dwarfs. Due to predation pressure from the pike, the ecotypes have evolved different survival strategies as a trade-off between safety and foraging. The strategy of the dwarf ecotype is to avoid pike predation by foraging zooplankton in the pelagic area (Öhlund et al. 2020). Their small body size makes them efficient

planktivores and they often occur in great quantities. However, dwelling in open waters does not come without risk.

their prey, yet it is suggested that they substantially reduce survival and growth rate in fish populations (Beckmann, Biro and Post 2006). Some piscivorous birds appear to switch to terrestrial prey when there is a drop in pelagic prey (Hebert et al., 2008). When foraging, the choice of lakes for black-throated loon (Gavia arctica) and especially red throated loon (Gavia stellata), have shown to be influenced by the presence of salmonids such as several

Coregonus species, specifically the vendace (Coregonus albula) (Eriksson and Paltto 2006).

In addition, Eriksson and Sundberg (1991) observed that red-throated loons tended to occur in lakes where species such as perch and roach were the most abundant. A study of

oligotrophic lakes in Canada by Kerekes et al. (1994) indicated that breeding success was positively correlated with lake area for a range of piscivorous birds, including the common loon (Gavia immer) and the common merganser (Mergus merganser). The common merganser has also been suggested to cause great mortality to salmonids (Wood and Hand, 1985). Loons have shown to have a low fecundity (Jackson 2003) contrary to the common merganser species who have a high fecundity (Jorgensen and Fath 2008). On the other hand, loons are highly efficient divers that are specialized in capturing prey under water (Clifton and Biewener 2018) and once reaching adulthood the survival rate is high (Eriksson 2010). Previous studies have indicated that population densities of piscivorous birds are higher on lakes with polymorphic whitefish than on lakes with monomorphic whitefish (Öhlund, unpublished). A hypothesis is that these birds are favored by the presence of dwarf whitefish. These fish often appear in extremely large numbers in lakes with polymorphic whitefish, and their small size, 10-20 cm, should make them suitable prey for these birds. One way to test this hypothesis is to investigate the abundance of dwarfed whitefish in the diets of

piscivorous birds on lakes with pike, that have polymorphic whitefish, and lakes without pike, where whitefish is monomorphic and rather large. Further understanding in this question might provide a deeper insight into the complex interaction between evolutionary divergence and ecology.

1.2. Aim

The aim for my thesis is to compare the diet of piscivorous birds living on lakes with polymorphic and monomorphic whitefish, respectively. Specifically, I hypothesize that the diet of piscivorous birds is dominated by whitefish in pike lakes where whitefish is

polymorphic.

2 Method

2.1. Study sites

The collection of data took place during four years (2017 - 2020) ranging from august - mid September in 36 different lakes around northern and central Sweden (Appendix 1). The lakes were selected primarily based on the presence of pike and polymorphic whitefish (22 lakes) or monomorphic whitefish and absence of pike (14 lakes), and secondarily on logistics and accessibility of the lakes (Figure 1).

2.2. Sampling

Droppings from the following bird species were collected: black-throated loon (Gavia

arctica), throated loon (Gavia stellata), common merganser (Mergus merganser),

when resting on the lake; By confusing the bird with a powerful headlamp, we could get close enough to catch them with a hand net. Caught birds were then put in

individual boxes where they typically defecated within a few minutes. All boxes were set up with a clean plastic bag in the interior. One additional method when collecting droppings from the merganser species was to spook birds resting on islets and skerries. This typically made the birds defecate and sampling could be made. Caught birds were weighed, photographed, occasionally measured, and then released. Droppings were collected with a plastic spoon or a swab and transferred to a marked sample tube. An environmental control was taken in the exact same way but without touching the dropping. All samples were stored on ice in a cooling box and eventually in a freezer. Out of all 146 individual based samples collected, I took part in the collection of 13 samples in the year 2020.

2.3. Extraction of samples

Samples were thawed and given a temporary identification number to assure that the person analyzing the samples were unaware of where the sample had been taken. Fecal samples were divided into three categories (1-3) based on volume. For fecal samples that had a volume between 0-200 μl (category 1), the following concentrations were used: 170 μl Buffer ATL (Qiagen), 20 μl Proteinase K 600 U/ml (Qiagen) and 10 μl DTT 1 mol (Qiagen). If the fecal samples attained a volume between 200-400 μl (category 2) or 400-600 μl (category 3), the concentrations were increased proportionally. Swab samples attained a volume of 200-400 μl (category 2). The samples were vortexed for 5 minutes and incubated at 56 C° overnight, then centrifuged the following day at 8,000 x g (14,000 rpm) for 1 minute. 200 μl of the supernatant was transferred into a 1,5 ml microcentrifuge and 200 μl Buffer AE was added into the tube, ending with a final elution volume of 200 μl. After the elution process instructions for DNeasy Blood & Tissue Kit (Qiagen) were followed, continuing from step two. Lastly, instructions from the OneStep PCR Inhibitor removal kit (Zymo Research) were followed to ensure a cleaner product with no inhibitors interfering enzymatic reactions. An extraction blanc control was generated for every extraction session.

2.4. Primers

Species specific primers were used when quantifying DNA-concentrations for the following species: European whitefish (Coregonus lavaretus), brown trout (Salmo trutta), Arctic char (Salvelinus alpinus), perch (Perca fluviatilis), roach (Rutilus rutilus), burbot (Lota lota) and European minnow (Phoxinus phoxinus) (Appendix 2). The primers are mitochondrial and were manually designed by Olajos (unpublished), following general guidelines specified in Taberlet, Bonin, Zinger, and Coissac (2018).

2.5. ddPCR-analysis

For ddPCR-analysis I prepared a PCR-mixture containing the following concentrations: 1 x QX200 ddPCR EvaGreen Supermix Supergreen, 0.1 μm forward primer, 0.1 μm reverse primer, and deionized and sterilized water. The solution was mixed and 20 μl was put into a 0.1 ml PCR tube. All primers and DNA were centrifuged before use. For each PCR tube, 3 μl of the sample were added, leaving the per sample reaction volume at 23 μl. At least one positive and one negative control was added with 3 μl of target DNA respective 3 μl of water. This was done for each primer per PCR session. An eight-strip lid was put on the PCR tubes, vortexed and centrifuged. The lids were carefully removed and an amount of 20 μl was simultaneously transferred to a DG8 cartridge (Bio-Rad). All the wells were filled with 70 μl of QX200 Droplet Generation Oil (Bio-Rad) each and sealed with a DG8 Gasket (Bio-Rad). The cartridge was loaded into the QX200 Droplet Generator (Bio-Rad) in order to form droplets. The droplets were carefully transferred to designated rows in the Semi-Skirted 96-Well PCR Plate (Bio-Rad), then sealed in a PX1 PCR Plate Sealer (Bio-Rad) at 175 C° for four seconds with an PCR plate Heat-Seal pierceable (Bio-Rad). For PCR, a T100 Thermal Cycler (Bio-Rad) was used. During PCR, a seven-step protocol was set accordingly: C°95 for 10 minutes (step 1), then 95 C° for 30 seconds (step 2) followed by 60 C° for 1 minute (step 3). Step 2 and 3 was repeated 34 times (step 4). Continuing at 4 C° for 5 minutes (step 5), followed by 90C° for 5 minutes (step 6) and finally ending with an infinite hold at 4 C° (step 7). Note, all the primers were run at 60 C° during step 3, except for the European minnow which was run at 56.5 C°. Lastly, the sealed PCR was loaded into a QX200 Droplet Reader (Bi0-Rad), set to read the droplets in ABS (Absolute quantification) for final results.

2.6. Statistical analysis

QuantaSoft v.1.4.7 was used to read and interpret the PCR results. A lower limit of 10500 generated droplets was set and samples which indicated a hit but did not have enough droplets were rerun. Samples that showed ambiguous results or proved difficult to assess were diluted by a ratio of 1/10 and 1/100. An amount of 5 or less positives was considered a machine error or simply too low concentration to account as a hit. The concentration of the target DNA in a sample were calculated according Eq. 1.

Eq. 1

where Nneg is the number of negative droplets, N is the total number of droplets and Vdroplet is the analyzed volume (0.85 nl). The concentrations were then decoded to respective species and lakes. All concentrations were rounded to two decimals as standard in QuantaSoft, then converted to proportions and analyzed through statistical software R studios V4.0.3 (R core Team 2020). All samples not indicating hits on any designated primers were removed. The ddPCR results was used to calculate relative abundances for the mentioned fish species within the bird species’ diets. Beta regression was used to test if relative abundances in the diet of the birds differed between polymorphic and monomorphic whitefish lakes (Cribari-Neto 2010). Beta regression is suitable for data that are recorded as proportions. Since beta regression cannot handle proportions that are zero or one, I replaced zeroes with the

smallest non-zero value observed, and observation of one were replaced by the largest value observed that was smaller than one. A significance level of P<0.05 was set throughout the analysis. Note, samples from terns could often not be assigned to species (Arctic tern and common tern). Thus, I present results for the two species combined.

3 Results

The relative abundance of whitefish in the diet of the entire fish-eating guild was higher in pike lakes with polymorphic whitefish than in pike-free lakes with monomorphic whitefish (44% and 18%, respectively, beta regression, N = 36, p = 0.004, pseudo-R2 _{= 0.2246) (Figure}

“pike-free lakes”. The total diet in pike lakes consisted mainly of whitefish, followed by perch, minnow, roach and least burbot, only making up 5% (Figure 2). No char was to be found in pike lakes. In pike-free lakes, the most common to least common prey was: perch, minnow, whitefish, brown trout, char, burbot and roach (Figure 3).

Figure 2. Multi-pie chart illustrating the average proportion of different prey species in the diets of piscivorous bird species, in pike lakes. The circles represent the different bird species, and the colors represent different fish species

Analyzes of the different bird species showed that the relative abundance of whitefish in the diet was higher for black-throated loon in pike lakes (73%) than pike-free lakes (20%) (beta regression, N = 36, p = 0.0139) (Figure 4, Table 1). For red-throated loon this effect was even stronger: the relative abundance of whitefish was 98% in pike lakes and 20% in pike-free lakes (beta regression, N = 8, p < 0.0001) (Figure 4, Table 1). The same tests were non-significant between lake categories for common merganser (beta regression, N = 19, p = 0.2578) and red-breasted merganser (beta regression, N=12, p = 0.658), as well as for terns (beta regression, N = 14, p = 0.322) (Table 1).

Figure 4. Bars represent average proportion of whitefish in the diet of piscivorous birds in lakes with and without pike. Error bars represent ±1 standard error.

Shifting focus from the main hypotheses, I found that black-throated loon had significantly higher proportions of perch in their diet in pike-less lakes (44%) than in pike lakes (16%) (beta regression, N = 12, p = 0.0429) (Figure 5, Table 1). This was also true for red-throated loon (46% in pike-free lakes and 0% in pike lakes, beta regression, N = 8, p = 0.0417) (Figure 5, Table 1). All other comparisons were non-significant (p = 0.112-0.98) (Table 1).

Figure 5. Bars represent average proportion of perch in the diet of piscivorous birds in lakes with and without pike. Error bars represent ±1 standard error.

The common merganser had higher abundance of European minnow in their diets in pike-less lakes (33%) (beta regression, N = 19, p = 0.0348) compared to pike lakes (3%) (figure 6, Table 1). All other comparisons were non-significant (p =0.3142-0.7878) (Table 1). However, the relative abundances of European minnow were relatively high for all bird species on

non-0 0,2 0,4 0,6 0,8 1 1,2 Black-throated loon Red-throated loon Common merganser Red-breasted merganser Terns P rop orti on i n di et Piscivorous birds

Whitefish - with or without pike

Pike No pike 0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 Black-throated loon Red-throated loon Common merganser Red-breasted merganser Terns Pro por ti on in di et Piscivorous birds

Perch - with or without pike

pike lakes, except for terns which had a more equal representation of this prey species, around (~24%) (Figure 6).

Figure 6. Bars represent average proportion of minnow in diet of piscivorous birds, in lakes with pike and without pike. Error bars represent ±1 standard error.

Table 1. Beta regression analysis of piscivorous bird diets in lakes with and without pike. The independent variable was pike presence or absence and response variables were the relative abundances of the different fish species in the diets. N refers to the number of lakes where informative samples were obtained.

Variable Bird species N Pseudo-R² P-value

Whitefish Whole guild 36 0.2246 0.004 **

Whitefish Black-throated loon 12 0.4244 0.0139 * Red-throated loon 8 0.6535 5.8e-06 *** Common merganser 19 0.1071 0.2578 Red-breasted merganser 12 0.0314 0.658

Terns 14 0.0797 0.322

Perch Black-throated loon 12 0.2823 0.0429 * Red-throated loon 8 0.5517 0.0417 * Common merganser 19 0.1272 0.112 Red-breasted merganser 12 0.0302 0.575

Terns 14 9.897e-05 0.98

Minnow Black-throated loon 12 0.2365 0.3142 Red-throated loon 8 0.0840 0.6573 Common merganser 19 0.2661 0.0348 * Red-breasted merganser 12 0.0076 0.7861

Terns 14 0.0058 0.7878

17 samples were excluded from the analyses due to no hit of any designated primer. No contamination in environmental control blanks, extraction blanks or NTC (negative template controls) were found.

4 Discussion

The primary aim of this thesis was to test the idea that the high abundance of piscivorous bird species on lakes with polymorphic whitefish reflects a high availability of whitefish prey in these lakes. The hypothesis was supported when tested at the level of the whole guild of fish-eating birds as whitefish was the dominating prey in pike lakes, but not in pike-free

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 Black-throated loon Red-throated loon Common merganser Red-breasted merganser Terns Pro por ti on in di et Piscivorous birds

Minnow - With or without pike

lakes. Analyzes at the level of individual bird species was less clear, only two species, the black-throated loon and the red-throated loon appeared to be influenced by the polymorphic whitefish species. For the other bird taxa, the mergansers and the terns, no effect of whitefish divergence could be detected. However, the variance within treatments groups was rather large, with many records of 0% or 100% of whitefish in the diet, suggesting that the statistical power may have been too low to detect less pronounced effects of whitefish divergence.

Still, it was not unexpected the two loon species had higher intake of whitefish in pike lakes, and that the effect was particularly strong for the red-throated loon. This is in line with the finding that the red-throated loon has showed tendency to choose foraging habitat in lakes inhabited by small bodied Coregonus species, especially the vendace (Eriksson and Paltto 2006). In addition, the red-throated loon almost exclusively feed their chicks with fish, whereas the black-throated loon also includes insects when feeding their offspring (Eriksson and Sundberg 1991). This would suggest that the red-throated loon may spend more time foraging in the pelagic area where dwarf whitefish are particularly abundant (Öhlund et al. 2020). Neither the mergansers nor the terns displayed any significant difference between pike and non-pike lakes in the intake of whitefish prey. Since dwarf whitefish spend much of their time in deeper waters, this may reflect that mergansers, and especially the terns, do not move as agile or deep as the loons, which have been reported to dive as deep as 60 meters (Schorger 1947). This limitation may be less important early in the season when young-of-year whitefish dwell in shallow waters along the shoreline (Öhlund et al. 2020). Most of our merganser samples were taken in August/September, and it is therefore possible that mergansers feed more on whitefish earlier in the season. Also, there is a probability that the total diet of mergansers and terns includes unknown species, diluting the chance of

consuming a specific prey species, such as the whitefish.

Although the main hypothesis concerns the role of dwarfed whitefish as a rich food resource, there may be alternative explanations for the increase in birds density on pike lakes. A previous study has shown that pike lakes have a higher abundance of littoral zoobenthos, possibly because pike predation reduce the abundance of benthivorous fish species such as roach perch and whitefish (Stenman 2014). Also, the studies of Jansson (2017) and Karlberg (2017) suggests that brown trout and perch switch from insectivory to piscivory, which would also relieve insects from predation pressure in pike present lakes. As the piscivorous bird species we study also feed on insect prey when they are abundant, it is possible that the availability of insect prey also contributes to the high density of fish-eating birds on pike lakes. Thus, I propose that future studies of this hypothesis should also consider the potential role of insect prey.

There was a higher proportion of perch in the diet of loons in pike-free lakes in contrast to the pike lakes. With the presence of pike, loons seem to shift focus from perch, possibly foraging on other species not included in the analysis. The diet of the fish-eating guild on pike-free lakes was dominated by perch and minnow, and for the common merganser there was a significantly higher representation of minnow in lakes without pike. While not

significant there was a proportionally lower representation of minnow in lakes with pike for all bird species except for the terns (figure 4). A first suggestion would be that predation by the pike limits the availability of minnows for loons and mergansers. As for the terns, who obtained a more equal proportion between the lake categories, perhaps a different yet more successful predation technique was used. Terns are known to strike from air (Cramp 1985), possibly providing a better overview, whereas mergansers and loons dive directly from the surface (Clifton and Biewener 2018, Kålås, Heggberget, Bjørn and Reitan 1993). While only a small fraction of trout was present in the diet found in pike present lakes, a moderate

The birds species included in the analysis are migratory (Svensson, Zetterström and

Mullarney 2009), and perhaps do not always forage in the same place as they breed, such as the throated loon (Eriksson and Johansson 1997). Both the common tern and the red-throated loon has been reported to forage as far as 10 km from nesting sites (Eriksson and Johansson 1997, Becker, Frank and Wagener 2008). This implicates that the observed diet may also include fish caught on other lakes than the one where the bird was caught.

However, almost all the recorded prey belonged to species found in the lake where the bird was caught.

Most of the data used in this thesis was sampled during summer or early fall. An interesting thought would be to further explore if season may affect the bird species intake of whitefish. Many migratory birds breed in spring (Bowers et al., 2016) and may forage differently depending on prey abundance and preference of their offspring (Hebert et al., 2008). The hatching of whitefish occurs during spring and vary with temperature (Urpanen 2005). Perhaps this can cause variation in whitefish abundance since temperature can shift between the years. Yet, during the spring period it might be difficult to perform sampling of loon and merganser since it is not dark enough.

Interpreting the outcome of ddPCR-analysis involves some complex trade-offs. Fecal

samples that gave a very low amount of positive droplets (1-5 droplets) was accounted as “no hit”. This was done to ensure that no machine error would pass into the analysis, since the machine sometimes can produce false-positive droplets. A problem arises when potentially positive samples are diluted (1/10 or 1/100), which means that the concentrations of DNA will decrease ten- or hundredfold. If any of these samples ended up as five droplets or less, a considerable concentration might be overlooked. Moreover, if a false negative (overlooked) sample were excluded when calculating the fractions, other standalone low hit

concentrations within the sample may appear as 100%, causing deviations in the result.

5 Conclusion

Studies on how speciation affects higher trophic levels are not too common, yet just as important. This study, in combination with other studies showing that the abundance of fish-eating birds are high on lakes with polymorphic whitefish (Öhlund unpublished.), provide some support for the idea that whitefish divergence induced by pike favors, not only other predatory fish species (Jansson 2017, Karlberg 2017), but also piscivorous birds. The finding that not all bird species increased their use of whitefish prey on lakes with polymorphic whitefish may indicate that there are other not yet studied advantages associated with whitefish divergence. Thus, more studies are needed to further assess what effects polymorphic whitefish have on piscivorous bird populations.

6 Acknowledgments

Design of study and sampling methods comply with the current laws of Sweden and were approved by the local ethics committee of the Swedish National Board for Laboratory Animals in Umeå. (CFN, license no. A-20-17 to Gunnar Öhlund).

7 References

Becker, P., Frank, D., and Wagener, M. 2008. Luxury in freshwater and stress at sea? The foraging of the Common Tern (Sterna hirundo). Ibis, 139(2), 264-269. doi:

10.1111/j.1474-919x.1997.tb04624.x

Beckmann, C., Biro, P. and Post, J. 2006. Asymmetric impact of piscivorous birds on size-structured fish populations. Canadian Journal of Zoology, 84(11), pp.1584-1593.

Bernatchez, L. and Wilson, C., 1998. Comparative phylogeography of Nearctic and Palearctic fishes. Molecular Ecology, 7(4), pp.431-452.

Bowers, E., Grindstaff, J., Soukup, S., Drilling, N., Eckerle, K., Sakaluk, S., and Thompson, C. 2016. Spring temperatures influence selection on breeding date and the potential for phenological mismatch in a migratory bird. Ecology, 97(10), 2880-2891. doi:

10.1002/ecy.1516

Clifton, G., and Biewener, A. 2018. Foot-propelled swimming kinematics and turning

strategies in common loons. Journal of Experimental Biology, 221(19), jeb168831–. doi: 10.1242/jeb.168831

Cramp, S. 1985. Handbook of the birds of Europe, the Middle East, and North Africa: the birds of the western Palearctic. Vol. 4, Terns to woodpeckers. Oxford Univ. Press. Cribari-Neto F,. and Zeileis A. 2010. Beta Regression in R. Journal of Statistical

Software, 34(2), 1–24. http://www.jstatsoft.org/v34/i02/.

Eriksson, M.O.G., and Johansson, I. 1997. Smålommen Gavia stellata i sydvästra Sverige – beståndsutveckling och häckningsframgång. Ornis Svecica 7: 1–10.

Eriksson, M.O.G., and Sundberg, P. 1991. The choice of fishing lakes by the Red-throated Diver Gavia stellata and Black-throated Diver G. arctica during the breeding season in south-west Sweden. Bird Study, 38(2), 135–144.

Eriksson, M.O.G., and Paltto, H. 2010. Smålommens Gavia stellata häckningsframgång i relation till vattenkemi och fiskbeståndens sammansättning i olika fiskevatten. Ornis

Svecica 16: 211-231.

Eriksson, M.O.G. 2010. Storlommen och smålommen i Sverige - populationsstatus, hotbild och förvaltning. - Sveriges Ornitologiska Förening, Stockholm och Svenska LOM-föreningen / Projekt LOM, Göteborg.

Fowler, C. and MacMahon, J. 1982. Selective Extinction and Speciation: Their Influence on the Structure and Functioning of Communities and Ecosystems. The American

Naturalist, 119(4), pp.480-498.

Fraser, D. and Bernatchez, L. 2005. Allopatric origins of sympatric brook charr populations: colonization history and admixture. Molecular Ecology, 14(5), pp.1497-1509.

Harmon, L., Matthews, B., Des Roches, S., Chase, J., Shurin, J. and Schluter, D. 2009. Evolutionary diversification in stickleback affects ecosystem functioning. Nature, 458(7242), pp.1167-1170.

Hebert, C., Weseloh, D., Idrissi, A., Arts, M., O'Gorman, R., and Gorman, O. et al. 2008. Restoring Piscivorous Fish Populations in the Laurentian Great Lakes Causes Seabird Dietary Change. Ecology, 89(4), 891-897. doi: 10.1890/07-1603.1

Häkli, K., Østbye, K., Kahilainen, K., Amundsen, P. and Praebel, K. 2018. Diversifying selection drives parallel evolution of gill raker number and body size along the speciation continuum of European whitefish. Ecology and Evolution, 8(5), pp.2617-2631.

Jackson, D. 2003. Between-lake differences in the diet and provisioning behaviour of Black-throated Divers Gavia arctica breeding in Scotland. Ibis, 145(1), 30-44. doi:

10.1046/j.1474-919x.2003.00119.x

Jansson, J. 2017. The influence of northern pike on the diet of brown trout. Bachelor's thesis. Umeå university.

Jorgensen, S., and Fath, B. 2008. Encyclopedia of ecology (pp. 3113-3122). Amsterdam:

Elsevier.

Kerekes, J., Tordon, R., Kerekes, J., Nieuwburg, A., Risk, L. 1994. Fish-eating bird abundance in oligotrophic lakes in Kejimkujik National Park, Nova Scotia, Canada.

Hydrobiologia, 279-280(1), 57–61. doi: 10.1007/BF00027840

Koene, J., Crotti, M., Elmer, K. and Adams, C. 2019. Differential selection pressures result in a rapid divergence of donor and refuge populations of a high conservation value

freshwater fish Coregonus lavaretus (L.). Evolutionary Ecology, 33(4), pp.533-548. Kålås, J., Heggberget, T., Bjørn, P., and Reitan, O. 1993. Feeding behavior and diet of

goosanders (Mergus merganser) in relation to salmonid seaward migration. Aquatic

Living Resources, 6(1), 31-38. doi: 10.1051/alr:1993003

Myers, A. and Giller, P. 1988. Analytical Biogeography - An Integrated Approach to the Study of Animal and Plant Distributions. Dordrecht: Springer Netherlands, pp.147-163.

https://link.springer.com/book/10.1007%2F978-94-009-1199-4. Olajos, F. [Unpublished].

Post, D. and Palkovacs, E. 2009. Eco-evolutionary feedbacks in community and ecosystem ecology: interactions between the ecological theatre and the evolutionary

play. Philosophical Transactions of the Royal Society B: Biological Sciences, 364(1523), pp.1629-1640.

Raeymaekers , J., Maes, G., Audenaert, E. and Volckaert, F. 2005. Detecting Holocene divergence in the anadromous-freshwater three-spined stickleback (Gasterosteus

aculeatus) system - Holocene Divergence In Three-spined Stickleback. Molecular Ecology, 14(4), pp.1001-1014.

R Core Team. 2020. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. http://www.R-project.org/

Schorger, A. 1947. The deep diving of the loon and old-squaw and its mechanism.

Wilson Bull. 59, 151-159.

Steinmetz, J., Kohler, S., and Soluk, D. 2003. Birds Are Overlooked Top Predators In Aquatic Food Webs. Ecology, 84(5), 1324-1328. doi:

10.1890/0012-9658(2003)084[1324:baotpi]2.0.co;2

Stenman, M. 2014. Effects of northern pike (Esox lucius) and speciation in whitefish (Coregonus lavaretus) on abundance of beach near benthos. Bachelor's thesis. Umeå university.

Svensson, L., Zetterström, D., and Mullarney, K. 2009. Fågelguiden (2nd ed.). Stockholm:

Bonnier Fakta.

Taberlet, P., Bonin, A., Zinger, L., and Coissac, E. 2018. Environmental DNA: For biodiversity research and monitoring. Oxford: Oxford University Press. doi:10.1093/oso/9780198767220.001.0001

Thomas, S., Harrod, C., Hayden, B., Malinen, T., & Kahilainen K. 2017. Ecological speciation in a generalist consumer expands the trophic niche of a dominant predator. Scientific

Reports, 7(1). doi: 10.1038/s41598-017-08263-9

Urpanen, O., H. Huuskonen, T. J. Marjomäki and J. Karjalainen. 2005. Growth and size-selective mortality of vendace (Coregonus albula (L.)) and whitefish (C. lavaretus L.) larvae. Boreal Environment Research 10: 225–238.

Wilson, C. and Hebert, P. 1998. Phylogeography and postglacial dispersal of lake trout (Salvelinus namaycush) in North America. Canadian Journal of Fisheries and Aquatic

Sciences, 55(4), pp.1010-1024.

Wood, C., and Hand, C. 1985. Food-searching behaviour of the common merganser (Mergus

merganser) I: Functional responses to prey and predator density. Canadian Journal Of Zoology, 63(6), 1260-1270. doi: 10.1139/z85-189

Öhlund, G., Bodin, M., Nilsson, K., Öhlund, S., Mobley, K., Hudson, A., Peedu, M.,

Brännström, Å., Bartels, P., Præbel, K., Hein, C., Johansson, P. and Englund, G. 2020. Ecological speciation in European whitefish is driven by a large‐gaped

8 Appendix

8.1. Studied lakes with and without pike

Table S1. A list of the lakes where the fecal samples were obtained.

Lake Pike presence Altitude Area

Max depth Coordinates (RT90) (masl) (km²) (m) X Y Bastansjön No 453 374 20 1518650 7251770 Bodsjön Yes 275 342 15 1476970 6967110 Bölessjön Yes 376 208 18 1451070 6979910 Börjesjön Yes 293 374 40 1479500 6985170 Dikasjön Yes 416 643 15.5 1511090 7234650 Fullsjön No 372 180 20 1506990 7036400 Gräsvattnet No 378 228 33 1452060 7128550 Gunnarvattnet No 388 339 37.5 1418520 7110620 Gysen Yes 396 1190 40 1430050 7060030 Gäuta No 438 3163 58 1486800 7277820 Hotagen Yes 313 4541 75 1443950 7076210 Häggsjön No 503 117 16 1374610 7045500 Hökvattnet Yes 391 348 22 1452520 7086590 Idsjön Yes 261 919 40 1495260 6967470 Ismunden Yes 281 1235 60 1469310 7004330 Laisan Yes 478 2753 66.2 1563200 7322000 Locknesjön Yes 328 2658 57.4 1456020 6979110 Långvattnet Yes 460 1764 20 1534610 7231000 Lännässjön Yes 437 1931 40 1414190 6947800 Näkten Yes 325 8309 47 1437200 6978530 Paktajaure No 403 173 36 1614140 7596910 Pauträsket Yes 422 1004 30 1584350 7188750 Råssjön Yes 384 485 29 1446940 6950940 Stor-Arasjön Yes 542 713 21.5 1585960 7167170 Stor-Laisan No 446 1306 28 1478600 7293200

Storsjön Ljungdalen Yes 564 2768 27 1369390 6966330

Stor-Skirsjön Yes 272 882 30 1513640 7106600 Storvattnet No 338 269 48 1432507 7148649 Torringen No 399 677 29 1503850 6948370 Tuvattnet No 415 175 40 1445533 7085349 Uddjaur Yes 420 24915 31 1602210 7306910 Valsjön Yes 331 919 72 1421890 7102420 Volgsjön Yes 334 2173 17 1543630 7160860 Västansjön No 446 430 25 1470800 7293110

Yttre & Västra Jippmokksjön No 439 323 30 1528730 7243300

8.2. Gene sequences for the different fish species