Radix Balthica

Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 1251

Adjusting to the extreme

Thermal adaptation in a freshwater gastropod

MAGNUS JOHANSSON

ACTA UNIVERSITATIS UPSALIENSIS ISSN 1651-6214 ISBN 978-91-554-9243-4 UPPSALA urn:nbn:se:uu:diva-248495 2015 Dissertation presented at Uppsala University to be publicly examined in Zootissalen, Villavägen 9, 2tr, Uppsala, Friday, 5 June 2015 at 10:00 for the degree of Doctor of Philosophy. The examination will be conducted in English. Faculty examiner: Prof. Robby Stoks (Laboratory of Aquatic Ecology, Evolution and Conservation, KU Leuven).

Abstract Johansson, M. 2015. Adjusting to the extreme. Thermal adaptation in a freshwater gastropod. Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 1251. 49 pp. Uppsala: Acta Universitatis Upsaliensis. ISBN 978-91-554-9243-4.

Temperature is a ubiquitous force influencing biological processes ranging from cellular responses to life span. The thermal environment for many organisms is predicted to change with globally increasing temperatures and studies conducted in natural systems incorporating various evolutionary forces, such as gene flow, is needed. In my thesis, I investigate how snails (Radix balthica) originating from distinct geothermal environments within Lake Mývatn in northern Iceland have adapted, both genetically and phenotypically, to the respective thermal regime. Locations were classified as either cold, warm or seasonal depending on the average and variance in temperature. A high resolution spatial distribution of genetic variation within Mývatn was obtained using both neutral and outlier AFLPs. In addition, the genetic profile enabled me identify warm origin snails irrespective of geographic location in Iceland. Warm environments were often more stressful than cold or seasonal environments but snails originating from a high temperature location benefited from increased performance elsewhere. Patterns of growth were identical in both common garden and reciprocal transplant experiment; warm origin snails grew faster than both cold and seasonal origin snails. This result is in concordance with quantitative genetics models of thermal adaptation but suggesting cogradient rather than countergradient variation. Although warm origin snails generally had superior performance, survival at cold temperatures (< 12 °C) was reduced. All snails matured at similar size in the common garden experiment but cold origin snails were observed to mature later and lay fewer eggs. Also, snails had a common optimum for growth rate at 20 °C irrespective of thermal origin. This is arguably the reason why snails were observed to have a common thermal preference. Interestingly, warm origin snails had a reduced tolerance to high temperatures compared to cold and seasonal origin snails which did not differ in tolerance. Putatively, natural selection has reduced a putatively unnecessary trait (high temperature tolerance in a stable thermal environment) in favour of higher growth rate and performance in warm habitats. In conclusion, the price of high performance in a warm environment was paid in terms of reduced survival at low temperatures and a potential disadvantage of reduced genetic variability.

Keywords: Radix balthica, Lake Mývatn, geothermal springs, thermal adaptation, isolation by environment, population structure, gene flow, cogradient variation, AFLP, thermal preference, CTmax

Magnus Johansson, Department of Ecology and Genetics, Animal ecology, Norbyvägen 18 D, Uppsala University, SE-752 36 Uppsala, Sweden.

ISSN 1651-6214 ISBN 978-91-554-9243-4 urn:nbn:se:uu:diva-248495 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-248495)

Till Päpplingen och Ärtan

List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals.

I Quintela, M., M.P. Johansson, B.K. Kristjánsson, R. Barreiro and A. Laurila. 2014. AFLPs and mitochondrial haplotypes re- veal local adaptation to extreme environments in a freshwater gastropod. Plos One 9:e101821

II Johansson, M.P., M. Quintela and A. Laurila. 2015. Genetic variation in neutral and outlier markers across thermal environments in the common pond snail (Radix balthica). Manuscript.

III Johansson, M.P., F. Ermold, B.K. Kristjánsson and A. Laurila. 2015. Local adaptation of gastropod life history to contrasting thermal environments in a geothermal lake. Submitted.

IV Johansson, M.P. and A. Laurila. 2015. Similar temperature preference but differing maximum tolerance between contrasting thermal populations of Radix balthica. Manuscript.

Reprints were made with permission from the respective publishers.

Contents

Introduction ...... 9 Concepts of thermal adaptation ...... 9 Temperature and ectotherm ecology ...... 12 Project and aims ...... 13 Methods ...... 15 Lake Mývatn and its thermal environments ...... 15 Southern Iceland locations ...... 16 Radix balthica ...... 19 Outliers in ALFP and mtDNA within Mývatn (I) ...... 20 Genetic structure across Iceland (II) ...... 21 Life history (III) ...... 22 Thermal preference and tolerance (IV) ...... 23 Results and Discussion ...... 24 Outliers in AFLP and mtDNA within Mývatn (I) ...... 24 Genetic structure across Iceland (II) ...... 26 Life history (III) ...... 28 Temperature preference and tolerance (IV) ...... 31 Conclusion ...... 33 English summary ...... 35 Svensk sammanfattning ...... 38 Acknowledgements ...... 41 References ...... 43

Abbreviations

AFLP Amplified Fragment Length Polymorphism CnGV CounterGradient Variation CoGV CoGradient Variation HSP Heat Shock Protein IBD Isolation By Distance IBE Isolation By Environment mtDNA Mitochondrial DNA PCR Polymerase Chain Reaction

Introduction

A fundamental aspect of evolutionary biology is to understand and predict how organisms cope with changes in the environment. Genetically determined differences in performance between individuals are the material that evolution works with through the process of natural selection. Temperature affects most things in living systems from molecular reactions to life history variables including life span (Kingsolver and Huey 2008; Angilletta 2009). As temperature is such a ubiquitous force it has enticed research for a long time and it has been studied over a multitude of biological scales. The field of thermal adaptation of ectotherms, organisms in which internal sources of heat are relatively small compared to external sources, has increased in importance in recent times due to global climate change. The challenge lies in finding broadly applicable patterns of thermal adaptation in midst of the immense variation that exists at all levels from individuals to species.

Concepts of thermal adaptation One of the central tenets in studies of evolutionary adaptation is that over time and with low genetic exchange, populations inhabiting different environments will adapt to local conditions (Endler 1986; Kawecki and Ebert 2004; Hereford 2009; Blanquart et al. 2013; Fig. 1A). However, local adaptation is not inevitable as it may be obstructed by gene flow from surrounding populations or, especially in small populations, confounded by genetic drift (Slatkin 1987; Kirkpatrick and Barton 1997; Lenormand 2002; Bridle and Vines 2007). Populations that are not specialized to particular conditions, often modeled as having a wider performance breadth but with lower maximum performance across an environmental variable, are considered generalists (Fig. 1B). In nature, thermal generalists are more common than specialists (Angilletta 2009; Conover et al. 2009). Traditionally, environments with a large variation in temperature are expected to select for thermal generalists while organisms inhabiting a more constant temperature environments will evolve towards becoming thermal specialists (Janzen 1967; Gilchrist 1995). However, thermal generalists often have both wide thermal breadth and high maximum performance, a so called ‘Jack-of-all-temperatures-is-a-master-of-all’ (Huey and Hertz 1984; Palaima and Spitze 2004; Angilletta 2009). However, this is not a concept adopted exclusively for

Figure 1. When comparing populations from different thermal environments one of four reaction norm patterns could appear. Solid and dashed lines represent cold and warm origin organisms, respectively. A) a shift in Topt (local adaptation), B) shift in thermal breadth (generalist – specialist), C) warm origin organisms performs at a higher level than cold origin organisms (cogradient variation) or D) cold origin organisms perform at higher level than warm origin organisms (countergradient variation). Adopted from Angilletta (2009) with permission of Oxford University Press. thermal generalists as warm adapted genotypes have also been noted to have similar characteristics (Knies et al. 2009). In ‘optimality’ models of thermal adaptation (Lynch and Gabriel 1987; Gilchrist 1995), performance at one temperature is maximized at the expense of performance elsewhere (Fig. 1A). However, general predictions produced from these models have not received wide support from empirical data (reviewed in Angilletta 2009). Instead, most evidence supports quantitative genetic models where genetic and environmentally induced variation affects phenotypes. These models divide thermal generalists into two groups, where the interaction between environment and genotype influence the phenotype in an additive or subtractive manner (Conover and Schultz 1995; Conover et al. 2009). The former causes a shift in performance in which warm origin organisms have higher performance than cold origin organisms (cogradient variation; Fig. 1C), while in the latter the performance pattern is reversed (countergradient variation; Fig. 1D). With countergradient variation (hereaf-

10 ter CnGV), genetic variation is largely hidden by environmental variation under natural conditions and revealed under common garden conditions where cold origin organisms outperform warm origin organisms (e.g. Merilä et al. 2000; Yamahira and Conover 2002; reviewed in Conover et al. 2009). While CnGV is often reported in connection to metabolic traits, cogradient variation (hereafter CoGV) is more common in morphological traits (Conover et al. 2009). However, reports of CoGV and growth rate do exist (e.g. Mitchell and Lampert 2000; Arendt and Reznick 2005; Barton et al. 2014). When performance is plotted across a range of temperatures it constitutes a ‘thermal reaction norm’ or ‘thermal performance curve’ (Angilletta 2009; Fig. 2). Such a curve is characterized by a left-skewed distribution limited by the critical temperature minimum (CTmin) and maximum (CTmax) denoting the thermal breadth. A maximal performance is achieved at a specific temperature (Topt) along the reaction norm. However, Topt can vary substantially between traits of an individual (Huey and Stevenson 1979; Martin and Huey 2008; Angilletta 2009). Hence, the preferred temperature (Tp) is often assumed to reflect an overall performance optimum (Angilletta 2009), but it is likely to vary with time as different traits are prioritized. A common assumption in models of thermal adaptation (e.g. optimality models; Lynch and Gabriel 1987; Gilchrist 1995) is that there is a close as- sociation between Tp and Topt (Angilletta 2009). However, Tp is predicted to lie at a lower temperature than the actual optimum for two reasons: i) organisms that have difficulties maintaining a stable body temperature should

Topt

Tp Performance

Thermal breadth CTmin CTmax

Temperature

Figure 2. A thermal reaction norm with the typical left-skewed distribution. The thermal breadth is limited by the critical minimum (CTmin) and maximum temperatures (CTmax) and the temperature at which performance is optimized (Topt). The preferred temperature (Tp) is usually lower than the Topt. Adopted from Angilletta (2009), originally in Huey & Stevenson (1979) with permission from Oxford Uni- versity Press.

11 buffer against high temperatures as those cause a greater reduction in performance than temperatures below the optimum (Martin and Huey 2008), and ii) when warm adapted organisms outperform cold adapted ones, the preferred temperature should lie below the optimum in order to enhance performance at higher temperatures without reducing performance at lower temperatures (Angilletta 2009; Asbury and Angilletta 2010). The difference between Tp and Topt is expected to increase with environmental temperature variation and degree of thermal specialization (Martin and Huey 2008) as well as with temperature variation within rather than between generations (Asbury and Angilletta 2010).

Temperature and ectotherm ecology Although they do not rely on internal sources of heat, ectotherms have a wide range of behavioral, physiological and morphological adaptations in order to regulate (maintain) body temperature (reviewed in Angilletta 2009). Ectotherms are highly influenced by the thermal environment. However, as water has a high ability to conduct heat, body temperature in small (< 100 g) aquatic ectotherms can be assumed to equal environmental temperature (Reynolds and Casterlin 1979). Such ectotherms are especially sensitive to environmental temperature, and their possibilities for, as an example, behavioral thermoregulation are often limited. Temperature and ectotherm life-history have an intricate relationship and three, more or less general rules, are known. These rules are summarized in the haiku by Kingsolver and Huey (2008):

Bigger is better And hotter makes you smaller Hotter is better

According to Kingsolver and Huey (2008) 80% of ectotherms obey the first two rules (‘Bigger is better’ and ‘Hotter makes you smaller’) meaning that larger individuals have higher fitness and that higher rearing temperature causes a maturation at smaller size. ‘Hotter makes you smaller’ has a coun- terpart in the temperature-size rule stating that in nature, organisms from a cold environment will mature at a larger size than those from a warm environment. A difference in thermal sensitivities for growth and development is one mechanism which would generate this pattern (Kingsolver and Huey 2008). ‘Hotter is better’ implies that warm adapted genotypes with a high Topt also have a higher performance than cold adapted organisms with a lower Topt (Bennett 1987; Kingsolver and Huey 2008; Angilletta 2009). The reason for this is based on a thermodynamic argument (Savage et al. 2004), where biochemical reactions occurs at a faster rate at high temperatures

12 (Hochachka and Somero 2002). Although it may appear illogical at first glance, the solution to the haiku above is to realize that the rules work on separate levels: phenotypic variation, phenotypic plasticity and evolved variation of thermal reaction norms (further described in Kingsolver and Huey 2008). The intimate connection between ectotherm fitness and size is evident in ‘Bigger is better’ and demonstrated in my study organism: the freshwater gastropod Radix balthica (Linnaeus 1758), where size is a better predictor of maturity than age and adult size is positively related to fecundity and offspring size (Dillon 2010). Clearly, size matters and it is both directly and indirectly dependent on temperature as warm environments often have higher amount of soluble salts, nutrients etc. which in turn provide a greater food source allowing for faster growth (Dillon 2010).

Project and aims Many studies on thermal adaptation are done in wide geographical gradients (e.g. Addo-Bediako et al. 2000; Sunday et al. 2011; Nilsson-Örtman et al. 2012). However, it has been hypothesized that different thermal environments can act as barriers to gene flow (Janzen 1967; Ghalambor et al. 2006) and there are indications that migration plays a prominent role in adjusting to climate change (Kirkpatrick and Barton 1997; de Mazancourt et al. 2008). Thus, studies conducted on a smaller scale incorporating the effects of gene flow are important. Iceland provides ample opportunities to study thermal adaptation on both smaller and larger scales as geothermal sites are scattered across the island. Temperature can differ between geothermal sites, and in this respect Lake Mývatn is unique considering the magnitude of the temperature difference across a few kilometers. Here, discrete temporally stable thermal environments ranging from ca. 6 °C to 24 °C are connected by habitats affected by seasonal temperature variation (Einarsson et al. 2004).

In this thesis I studied how geothermal environments affect genetic population structure and thermal evolution of local gastropod populations. The overall aim is to shed light on the role and strength of temperature as the driving force behind divergent natural selection in the face of gene flow. This thesis focuses on the processes shaping thermal adaptation on a micro- geographical level using R. balthica as a study species and geothermal environments in Iceland in general and Lake Mývatn in particular as study systems. In paper I, the aim was to establish if there is population genetic differentiation as well as evidence for natural selection and local adaptation in R. balthica originating from the contrasting thermal environments in Lake

13 Mývatn. I used genetic outlier estimations based on AFLP data and mitochondrial DNA haplotypes. In paper II, I used replicated thermal environments across Iceland and an extensive sampling design within Lake Mývatn to more precisely determine the population genetic structure in the contrasting thermal environments. I also aimed to more thoroughly elucidate patterns of gene flow within Lake Mývatn. In paper III I determined the extent of thermal adaptation in life history traits by conducting common garden and reciprocal transplant experiments with snails originating from contrasting thermal environments in Lake Mývatn. In paper IV my aim was to quantify variation in preferred temperature (overall performance optimum) and maximum temperature tolerance of snails originating from contrasting thermal environments in Lake Mývatn.

14 Methods

Lake Mývatn and its thermal environments Lake Mývatn is located in a geothermally active area in northern Iceland (65°60’N, 17°00’W; Fig. 3). The 37 km2 lake is divided into a southern and northern basin, with the southern basin often subdivided into a western and eastern part based on geology and ecology (Einarsson et al. 2004). It is fed by nutrient rich groundwater along the eastern shore with ca. 44 and 24 % of the intake occurring in the southeast corner and in the northern basin, respectively. The major outlet is the river Laxá situated in the western part of the lake. Lake Mývatn and the surrounding landscape have been the focus of much research due to the unique physical characteristics of the lake (Einarsson et al. 2004; Ives et al. 2008; Hoekman et al. 2011; Karvonen et al. 2013; Dreyer et al. 2015). With a relatively shallow topology (average depth is 2.5 m) and an abundant insect life (Mývatn meaning “midge lake”) makes it an ideal feeding environment for a variety of diving ducks (Einarsson et al. 2004). Indeed, Lake Mývatn is known for its unique mixture of European and North Ameri- can duck species with up to 10 000 individuals from 13 species spending the summer there each year (Gardarsson and Einarsson 1994; Einarsson et al. 2004). The groundwater running in Lake Mývatn passes through differently geothermally heated bedrock creating a mosaic pattern of thermal springs with distinct cold (6 – 7 °C) and warm (22 – 24 °C) temperatures along the eastern shore (Fig. 4). These temperatures are relatively stable over time with only minor variation occurring throughout the year (Einarsson et al. 2004; Fig. 4). The springs are surrounded by areas affected by seasonal temperature variation. With an average ice cover of 190 days per year (Einarsson et al. 2004) and a 10 % chance of having a daily average air temperature above 20 °C in July (Bjornsson and Jonsson 2004), water temperatures at the seasonal sites are often lower than in ‘cold’ springs (Fig. 4). In this thesis I took temperature measurements in two ways: point measurements at the time of the sampling and averages calculated based on the readings from temperature loggers. Three temperature loggers were placed at knee depth along the shore at ten sites in Lake Mývatn and left to record temperature every two hours between mid-May and mid-September in 2011

Figure 3. Maps of the study sites in A) Lake Mývatn (M) and Grӕnavatn (G) and in B) Lake Þingvallavatn (T) and the Hengill (H) stream system (inserted and not to scale). Areas A and B are marked on the outline image of Iceland in the top left corner. Sites marked with squares, triangles and circles represent cold, seasonal and warm environments, respectively. and 2012. Sites were selected to get an even coverage across the lake and to encompass the thermal spring environments.

Southern Iceland locations The large temperature difference occurring within a short distance makes Lake Mývatn unique. However, given that it is situated on the North Atlantic Ridge there are several geothermal areas in Iceland. Warm springs can be found along the ridge that stretches in a Y-shaped formation from southwest to northeast. In paper II, I sampled snails also in Þingvallavatn (64°11’N, 21°09’W; Fig. 3), the largest lake in Iceland. The North Atlantic ridge that goes right through the lake with several cold springs running in the northern area of the lake while the southern end of the lake is a geothermal area. However, only cold and seasonal snails could be sampled here as warm springs are used for geothermal power production. Temperature measurements from locations within Þingvallavatn are based on point measurements from the time of col- lecting snails.

16 For paper II, I also sampled snails in Hengill (64°11’N, 21°30’W; Fig. 3), a geothermal area southwest from Þingvallavatn. This area consists of several streams gathering in a highland valley. In the colder streams (< 10 – 12 °C) the fauna is dominated by chironomids (Chironomidae) and lacks any snails while the situation is reversed in the warmer streams (O'Gorman et al. 2012). Consequently, I could only sample snails in warm streams. Individual streams in the Hengill area have been monitored for a long time providing highly reliable temperature data (O'Gorman et al. 2012).

30 2011 30 2012

25 25

20 20

15 15

10 10 Temperature (°C) Temperature (°C) Temperature 5 5

0 0 May July Sept May July Sept

Figure 4. Temperature readings from a cold (M4), seasonal (M16) and warm (M7) site in 2011 and 2012 shown by solid (6 °C), dotted and solid (23 °C) lines, respectively.

17 Table 1. Summary table of the study sites (arranged from south to north) divided into the four areas Hengill, Þingvallavatn, Grӕnavatn and Mývatn. Locations marked in bold have low temperature variation and are considered temporally stable thermal environments. n refers to the sample size in paper II. Temperatures with standard deviation (± s.d.) are based on long term data while the remaining temperatures are given as point measurements taken when snails were sampled. Category refers to cold (C), seasonal (S) or warm (W) and is based on the average temperature. X/Y coordinates are given in the Icelandic coordinate system ISN93. Area Site n T (°C) Category X Y Hengill H4 26 16.9 W 387399 396785 64°11’N H3 25 22.4 W 387338 396898 21°30’W H2 27 15.2 W 387307 396931

T2 9 9.2 S 391318 409315 Þingvallavtn T1 12 8.8 S 391782 411388 64°11’N T5 15 8 S 400809 417345 21°09’W T4 9 9.7 S 399270 417489 T3 10 6.5 C 397445 418101

Grӕnavatn G2 11 5.5 S 592083 561327 65°54’N 17°00’W G1 12 13 C 592532 561572

M4†* 32 6.4 ± 0.8 C 594896 564603 M3 27 10.3 S 594471 565102 M6 29 16.3 S 593970 565783 M5†* 29 7.1 ± 0.8 C 594494 566203 M17 31 14.4 S 590574 564952 M20 31 10.5 ± 4.3 S 589957 566328 M21 27 9.4 S 588413 569525 M22 28 9.1 S 588950 569773 M23 27 10.1 ± 4.0 S 590658 569773 M16* 29 9.7 ± 3.0 S 594683 567752 Mývatn M15 30 12.5 S 595204 567902 65°60’N 30 7.5 ± 1.0 C 595263 567943 17°00’W M24* M10† 30 9.2 ± 3.1 S 595285 568830 M14* 27 9.0 ± 3.0 S 595004 569952 M13 25 11 S 595634 569573 M11† 26 19 W 595990 570361 M12 27 15.5 W 595961 570321 M19 33 17.5 W 595616 571632 M9†* 29 21.5 ± 2.1 W 595768 571569 M18 29 20.1 W 595742 571434 M7†* 27 23.3± 1.3 W 595659 572224 M8 28 12.2 ± 4.0 S 595827 573297 † Locations used for paper I * Locations used for paper III and, when excluding M5, paper IV

18 Radix balthica Several studies have described the difficulty of assigning individuals to species within the Radix genus as they are infamous for their similarity and plastic variation in morphology (Ellis 1926; Pfenninger et al. 2006; Schniebs et al. 2011; Vinarski and Serbina 2012). Hence, species identification based on plastic morphological traits has led to the description of several Radix species, but with genetic tools these have been reassessed repeatedly. Ac- cording to the current opinion R. ovata, R. peregra (also Lymnea peregra) and R. balthica all belong to the same species named Radix balthica (Bargues et al. 2001; Schniebs et al. 2011). R. balthica (the common pond snail) is a pulmonate snail commonly found throughout central and northern Europe in a variety of freshwater habitats (Pfleger 2000; Pfenninger et al. 2011). The shell is subjected to consid- erable variation but adults have typically a shell length of 11 – 22 mm, light to dark brown color and egg-shaped with 4 – 5 whorls that quickly increase in size (Pfleger 2000; Fig. 5). Pulmonates have a complex and interesting reproductive behavior including hermaphroditism, sperm storage (< 6 months) and multiple paternity and although all pulmonates are capable of selfing they prefer to mate as a male with another snail (Dillon 2010). R. balthica is a direct developing passive disperser relying on water cur- rents and birds as migratory vectors (Frisch et al. 2007; van Leeuwen et al. 2012). Passive dispersers among gastropods have been shown to be capable of moving large distances (Martel and Chia 1991; Frisch et al. 2007) and R. balthica is often one of the first gastropods to arrive to a new habitat (Lakowitz et al. 2008). Potential dispersal vectors between Europe and Ice- land include waterfowl (Frisch et al. 2007; Wada et al. 2011; van Leeuwen et al. 2012) and humans via ships.

Figure 5. Radix balthica sampled from the northern most site in Lake Mývatn (M8). Photo: Magnus Johansson

19 Outliers in ALFP and mtDNA within Mývatn (I) R. balthica from six populations (Table 1) were collected, stored in 70 % ethanol and DNA was extracted from a piece of the mantel using Qiagen DNeasy Blood and Tissue isolation kit following the manufacturer’s proto- col. Both papers I and II utilized AFLP as a method to evaluate genetic differences between sampling locations. AFLP is relatively cheap and easily applied method to non-model organisms for investigating differences at the whole genome level. In short, two restriction enzymes (Tru1I and EcoRI) were used to cut the DNA at sites across the whole genome depending on the sequence (Vos et al. 1995). The resulting fragments were then amplified using PCR based on the specific base pairs following the cut site and each individual’s entire fragment length (‘fingerprint’) was then visualized and scored using computer softwares. 10 % of the samples were replicated and evenly distributed between sampling locations and across PCR plates. The ‘fingerprint’ was checked by eye and the resulting data pruned of loci showing allele frequencies ≤ 5 % and ≥ 95 % to avoid problems related to high noise levels in downstream analyses. The remaining loci were tested in outlier analyses to identify loci more divergent than expected under neutrali- ty and thus potentially candidates for genes under diversifying selection. To reduce the risk of false positives three outlier detection methods were utilized: i) a method using Bayesian statistics (Beaumont and Balding 2004) implemented in the software BayeScan (Foll and Gaggiotti 2008), ii) a method similar to Fdist (Beaumont and Nichols 1996) and implemented for dominant markers in Mcheza (Antao and Beaumont 2011) and iii) a Spatial Analysis Method (SAM) correlating environmental and genetic data without the need for predefined populations (Joost et al. 2007). Geography and environment have a heavy impact on population genetic differentiation (Felsenstein 1976) and patterns of spatial distribution of genetic variation are often described in terms of ‘isolation-by-distance’ (IBD; Wright 1943) and ‘isolation-by-environment’ (IBE; Wang and Summers 2010; reviewed in Wang and Bradburd 2014). These terms are defined as an increase of genetic differences with geographical (IBD) or environmental (IBE) distance. IBD and IBE were investigated by partitioning the data into a ‘neutral’ and a ‘selected’ dataset analyzed separately but with the same methods. Genetic diversity as well as overall and pairwise FST were estimated in AFLP-SURV (Vekemans 2002) and in an AMOVA in GeneAlEx v6.4 (Peakall and Smouse 2006). The ‘neutral’ dataset was also surveyed using two programs that assign individuals to genetic clusters based on Bayesian statistical methods: STRUCTURE v2.3.3 (Pritchard et al. 2000; Falush et al. 2003; Falush et al. 2007; Hubisz et al. 2009) and TESS (Chen et al. 2007; Durand et al. 2009). In addition, genetic structure was assessed by using network analyses and graph theory. In this method a node (sample site) is represented by the aver-

20 age genetic profile within that sample site and connected by edges (gene flow) dependent on the genetic covariance between sites (Dyer and Nason 2004; Minor and Urban 2007). In essence, starting with a full model where all nodes are connected, edges are removed if its absence does not reduce the model’s ability to describe the among population genetic covariance structure until the minimum edge set is obtained (Dyer and Nason 2004). Net- work analyses were performed on both neutral and selected data sets. To further investigate any patterns of IBD and/or IBE more in detail, Mantel tests and Partial Mantel tests were performed as implemented in the software PASSaGE v1 (Rosenberg and Anderson 2011) on both pairwise genetic (i.e. FST) and graph (i.e. minimum genetic covariance) differences against pairwise geographic (IBD) or temperature (IBE) differences. In addition to estimating population genetic structure at nuclear DNA, mitochondrial DNA was also investigated. The presence of population structure in a 643 basepair region of the CO1 gene was tested with a hierarchical AMOVA done on haplotype frequencies and pairwise ΦST was calculated. As in the AFLP analyses, SAM and graph networks were used to investigate any correlation between temperature and mtDNA haplotypes.

Genetic structure across Iceland (II) In paper II, R. balthica was sampled extensively, both within Mývatn and from the other thermal environments in southwestern Iceland (Table 1). Snails were sampled, stored and DNA extracted using the same methods as described above. Seven of the most polymorphic marker combinations used in paper I formed the basis of the AFLP analyses, which were conducted in the same manner as above concerning laboratory protocols, scoring proce- dure and how the “neutral” and “outlier” AFLP data sets were created. Two major datasets were created: one containing all sites (ALL) and one with Mývatn and Grӕnavatn joined (MG; Table 1). From each of these datasets a neutral dataset was created by removing all AFLP loci identified as outliers by either of the programs (here BayeScan and Mcheza). Using the loci joint- ly identified as outliers from the MG dataset, two outlier datasets were created: oALL and oMG. These datasets were the only ones encompassing the entire temperature gradient, i.e. containing both cold and warm thermal environments. Based on the outliers found in the MG dataset, an additional outlier dataset was created including the eight temporally stable sites from all three systems (ST). The different results from STRUCTURE runs on both neutral and outlier datasets were compared. STRUCTURE settings and subsequent methods were identical to those in paper I. Similarly, Mantel and Partial Mantel tests were also used in order to evaluate patterns of IBD and IBE.

21 Life history (III) Juvenile R. balthica from two cold, seasonal and warm locations (Table 1) were collected in 2011 and transported to EBC, Uppsala University, in order to establish laboratory populations. However, during rearing one of the seasonal populations (M14) all individuals died. The snails were divided into two plastic aquaria with 15 L water and kept in a walk-in climate chamber at 16 °C and a 16h light:8h dark rhythm. Water was changed once a week and the snails were fed a food mixture of spinach, Spirulina pacifica and fish flakes ad libitum. The experiment was conducted on F3 snails – egg batches from F2 par- ents were collected weekly and kept separately until hatching. Two weeks after hatching 1 – 4 snails were photographed to measure initial size and subsequently transferred individually to a 125ml plastic cup and kept in one of four temperature treatments: 12, 16, 20 or 24 °C. Additional photographs were taken every two weeks to monitor growth. ImageJ (Rasband 2011) was used to measure size defined as the average difference in shell length from the apex to the furthest distance on the outer shell lip. Once the snail had matured, egg batches were collected and the total number of eggs and the number of hatchlings were recorded. Since the snails were kept individually all estimates on reproduction was based on selfing data. In an additional experiment, survival and growth at 6 °C were evaluated using F2 snails reared in 12°C as initial findings indicated a too high mortality in F3 snails of warm origin reared at 16°C.

To complement the common garden experiment a reciprocal transplant experiment was conducted in May and June 2011. Snails from two cold (M4, M5), two seasonal (M14, M16) and two warm (M7, M9) sites were collected and kept in a common environment (15 °C, 16h:8h) for three weeks at Hólar University College, Iceland. We used 0.5 L PET bottles with holes cut in the sides, the bottom removed and wrapped in a fine mesh net as experimental units. The bottles were attached in styrofoam rafts anchored to the bottom of the lake in the collection sites. The rafts and bottles were placed at each sampling location two weeks before the snails were introduced in the bottles in order for periphyton to grow and provide a food source for the snails. Rafts at one of the seasonal locations (M16) were lost in a storm and thus only one seasonal treatment location could be included in the analyses. Snails were individually marked with nail polish, photographed and two snails were placed in each of the containers (i.e. bottles). Growth in shell length was assessed after 25 days using the same methods as in the common garden experiment.

22 Thermal preference and tolerance (IV) The common method for testing thermal preference of ectotherms in the laboratory is to present the individuals with a thermal gradient and monitor their movements over time (e.g. Gerald and Spezzano 2005; Díaz et al. 2006; Diaz et al. 2011). I measured thermal preference as final preferendum (Reynolds and Casterlin 1979), defined as the temperature at which the snail spent most time, measured as the average temperature during the last four hours in a thermal gradient. I built thermal gradients by milling five 500 x 20 x 20 mm channels separated by 2 mm in an aluminum block. Each channel was filled with 125 ml RSW and the ends of the aluminum block were placed on either a cooling or heating plate. RSW was changed before each trial and the channels washed with detergent to remove any trace of the snails tested previously. After re- filling the channels, the blocks were left on the cooling and heating plates for the thermal gradient to stabilize. The experimental snails were laboratory-born F1 generation originating from two cold, seasonal and warm sites (Table 1) and being reared at 12 °C. The snails were placed at the acclimation temperature in each channel and monitored for six hours. The temperature at each snail was measured three times every 30 min and then averaged. Only data from the last four hours of the trial were used as initial preference can be highly correlated with the acclimation temperature (Reynolds and Casterlin 1979). After six hours the snails were placed individually in a plastic vial with 100 ml RSW and kept at room temperature overnight.

In gastropods, the maximum critical temperature (CTmax) is defined as the temperature at which the snail is no longer capable of remaining attached to the surface (Sandison 1967; White 1983). The day after having been tested in the thermal preference experiment, snails were placed in a 125 ml Falcon tube completely filled with RSW. The tube was placed in a heated water bath at 20 °C. The temperature was raised 1 °C every 15 min until the snails were observed to lose attachment. The Falcon tube was then immediately removed and placed in room temperature. Only snails that had recovered after 24 h were included in statistical analyses.

23 Results and Discussion

Outliers in AFLP and mtDNA within Mývatn (I) In accordance with previous studies associating environmental variables and AFLP markers (Bonin et al. 2006; Galindo et al. 2009; Gray et al. 2014) ca. 2 % of the AFLP loci could be considered outliers and thus putatively under temperature-mediated selection. Variation in neutral AFLPs and uncorrelated mtDNA haplotypes was mainly due to within population variation with significant average FST of 0.12 and of 0.09, respectively. However, outlier AFLPs and putatively temperature correlated haplotypes had much higher FST (0.55, p < 0.0001) and ΦST (0.65, p < 0.0001). Comparing network graphs between the different sets of markers showed contrasting patterns with AFLP loci and mtDNA haplotypes correlated with temperature having gene flow only among similar thermal environments (Fig. 6). While the results suggest local adaptation and genetic structuring in R. balthica across the thermal sites, these conclusions were hampered by the strong correlation between temperature differences and geographic distance given the linear sampling design. Nevertheless, observed patterns in the network analyses indicated that there was very little gene flow between cold and warm sites. In addition, the additional connecting edges within warm habitats suggest that gene flow is more common here than within cold environments (Fig. 6). Three mtDNA haplotypes were found to associate with temperature. Of these three haplotypes, Hap_15 showed increased frequency with decreasing temperature. In contrast, Hap_4 and Hap_6 were almost exclusively present at high temperatures. Enzymatic processes are highly temperature sensitive, and as mtDNA encodes for several subunits in respiratory enzyme complex- es (Blier et al. 2001), it follows that selection can act on the mitochondrial genome. However, since mtDNA is not only influenced by the external environment but is also co-evolving with the nuclear genome (Blier et al. 2001), the functional significance of mtDNA variation remains unclear.

Figure 6. Graphs of the genetic networks for AFLP (upper row) and mtDNA (lower row). The networks have been constructed with the A) eight loci under stabilizing selection, B) 117 neutral loci, C) four loci under directional selection, D) 22 uncorrelated mtDNA haplotypes and E) three haplotypes associated with temperature. Black (He, K), grey (G, V) and white (Hö, B) circles depict warm, intermediate and cold thermal environments, respectively. Sites connected by lines share significant genetic covariation. Solid lines represent proportional genetic and geographical distances, dotted lines indicates a higher genetic than geographic distance (suggesting a barrier to gene flow) and dashed lines suggests long distance migration processes.

Genetic structure across Iceland (II) In general, neutral population genetic structure was more affected by geographic distance (IBD) whereas population structure measured in the outlier loci was more affected by temperature (IBE; Fig. 7A and 7B). By comparing pairwise FST estimates between the neutral and outlier datasets a higher resolution in detecting patterns in the spatial distribution of genetic variation was obtained. In addition, analyses suggested that genetic diversity was reduced with increasing environmental temperature. With K = 2 in STRUCTURE analyses in the neutral ALL dataset, the Hengill area was clearly distinguished from all other sites (Fig. 7A). Howev- er, K = 3 had the highest likelihood and the additional cluster allowed the separation of sites in Þingvallavatn and those in Mývatn and Grӕnavatn (Fig. 7A). The pairwise neutral FST suggested that sites on the western shoreline of Mývatn grouped together. This was also true for the sites with a more exposed location along the eastern shoreline. Five loci were identified as outlier by both BayeScan and Mcheza. Genetic variation in four of these loci was correlated with environmental temperature. STRUCTURE assigned the data on these loci to two clusters (Fig. 7B). Instead of indicating an IBD pattern, the two groups fitted an IBE pattern with sites in one cluster having higher temperatures (Fig. 7B). Forcing a third cluster further increased the resolution and, again, Þingvallavatn sites were separated from Mývatn and Grӕnavatn (Fig. 7B). Pairwise FST estimations gave further insight into the population genetic structure – all warm spring environments grouped together as did the cold area in southeastern corner of Mývatn. Partial Mantel tests within the MG dataset indicated that geographic distance was not correlated with pairwise outlier FST differences after correcting for neutral pairwise FST (Table 2). Instead, outlier FST showed a high correlation with pairwise temperature differences (Table 2). This pattern was even clearer when only stable thermal sites were included in the analyses (Table 2). The low correlation between pairwise temperature and distance, and the higher correlation between pairwise outlier FST and temperature than with spatial distance suggest that the outlier loci were indeed associated with temperature. Use of outlier detection in a large and replicated dataset allowed me to identify the patterns of IBD and IBE at different scales. Reduction of genetic diversity with increasing temperature and correlations between individual gene frequencies and temperature further emphasize the selective pressure by high temperature. Besides, the fact that cold sites could not be separated from sites with seasonal temperature variation suggests that average temperature rather than temperature variation plays a major role in shaping the genetic structure.

H TMG

A) T2 T1 T5 T4 T3 H4 H3 H2 G2 G1 M4 M3 M6 M5 M9 M7 M8 M17 M20 M21 M22 M23 M16 M15 M24 M10 M14 M13 M11 M12 M19 M18

HT G M

WCCC C C W W WCCC C C W W C) D) T3 T3 H3 H3 G2 M4 M5 M9 M7 G2 M4 M5 M9 M7 M24 M24

Figure 7. Cluster analysis results from separate STRUCTURE runs based on: A) 75 neutral loci (K = 2 – 4); B) 5 outlier loci (K = 2 – 3) across all sites (H = Hengill, T = Þingvallavatn, G = Grӕnavatn, M = Mývatn); C) 75 neutral loci from the ALL dataset (K = 2) and D) 5 outlier loci (K = 2) on sites deemed to be temporally stable thermal environments (ST). Sites are mainly arranged from south to north but within Mývatn sites are also grouped from west to east according to pairwise FST values. Bar plots with a low number of clusters (K) are at the top of each subfigure.

Table 2. Mantel’s r values from Mantel and Partial Mantel tests comparing matrices of geographic distance (Geo), temperature (Temp) and genetic distances as pairwise FST obtained from neutral (nFST) and outlier (oFST) AFLPs on the three groups com- prised of all sites (ALL), sites within Mývatn and Grӕnavatn (MG) and sites considered to be temporally stable thermal springs (ST). Significance is estimated after 9999 permutations. Test Matrices ALL MG ST

Geo-nFST 0.63*** 0.41*** 0.65**

Geo-oFST 0.45*** 0.31** 0.39* Mantel NS Temp-nFST 0.23* 0.52*** 0.20 Temp-oFST 0.42*** 0.61*** 0.57*

Geo-nFST (Temp) 0.65*** 0.37*** 0.67** Geo-oFST (Temp) 0.50*** 0.25** 0.52** NS NS Geo-oFST (nFST) 0.18* 0.04 0.07 Partial Mantel NS Temp-nFST (Geo) 0.30** 0.48*** 0.31 Temp-oFST (Geo) 0.48*** 0.59*** 0.65**

Temp-oFST (nFST) 0.37*** 0.41*** 0.56** NS > 0.05. * 0.05 – 0.01, ** 0.01 – 0.001, *** < 0.001

Life history (III) Optimality models of thermal adaptation predict that organisms in temporally stable thermal environments are thermal specialists. However, despite the large difference in temperature between thermal springs there was little evidence for organisms being thermal specialists. However, one indication of thermal specialization was the reduced survival of warm origin snails in the cold environment in the common garden experiment (Fig. 8A). While cold and seasonal origin snails had the highest survival at 6 °C, warm origin snails survived the best in the 16 °C treatment (Fig. 8A). Interestingly, while warm origin snails were not affected, cold and seasonal origin snails had a marked reduction in survival in the reciprocal transplant experiment in the warm environment (Fig. 9A). Survival at cold and seasonal environments was equal irrespective of origin (Fig. 9A). However, the snails in the reciprocal transplant experiment had an initial size corresponding to the plateau in the mortality curve observed in common garden. Thus, initial differences in mortality in the reciprocal transplant could have been missed because of larger starting size. Instead of patterns corresponding to a specialist-generalist evolution, I found that CoGV prevailed in growth and fecundity measures in both the common garden and the reciprocal transplant experiment. This is in line with increasing empirical evidence that optimality models are often poor at pre-

28 dicting patterns of thermal sensitivity (Angilletta 2009). Although most cas- es of CoGV are reported in morphological traits (e.g. body size and shape), there are examples of studies of CoGV in growth rate (Mitchell and Lampert 2000; Arendt and Reznick 2005; Barton et al. 2014). Irrespective of thermal origins, Lake Mývatn R. balthica had similar optimum temperature for growth (20 °C, Fig. 8). This is in agreement with genetic analyses (I, II) as minor gene flow will disrupt the formation of a local optimum at a native temperature in a heterogeneous environment (Day 2000; Angilletta 2009). Further, the fact that the optimum lay closer to the mean temperature of a warm environment than cold or seasonal temperature was additional evidence that the gene flow was relatively low (Day 2000; Angilletta 2009). High gene flow from the ‘poor’ towards the ‘good’ environment would cause a shift in optimum temperature approaching the ‘poor’ environment as an adjustment to bad condition (Day 2000; Angilletta 2009).

100 A

50 Survival (%) Survival 25

612162024 Temperature (C)

0.15 B

0.10

0.05 Growth rate (mm/day) rate Growth

612162024 Temperature (C)

Figure 8. Survival (A) and growth rate (mm/day; B) during the first 84 days in each temperature treatment in the common garden experiment. Cold, seasonal and warm origin populations are symbolized with a square, circle and triangle, respectively. Single points in (B) represent preferred temperatures in IV. All values are mean ± 95 % confidence interval.

29 100 A B

0.15

0.10 50 Survival (%) Survival 0.05 25 Growth rate (mm/day) rate Growth

0 0 Cold Seasonal Warm Cold Seasonal Warm Thermal environments Thermal environments

Figure 9. Mortality (A) and growth (mm/day; B) in the reciprocal transplant experiment. Cold, seasonal and warm origin snails are represented by squares, circles and triangles, respectively. All values are means with 95 % confidence interval.

Warm and seasonal origin snails matured earlier, laid more eggs and had a higher hatching success than cold origin snails. Although warm origin snails seemed to have an advantage especially over cold origin snails, they were limited by the reduced survival at very low temperatures as compared to snails of other thermal origins. In addition, that age and size of maturity did not differ between warm and seasonal origin snails suggests that time constraints in Lake Mývatn were relatively weak as a short growth seasonal should select for a faster maturation. Growth rate is a highly important variable in most ectotherms as roughly 80 % of all ectotherms obey the ‘Bigger is better’ rule where larger individuals have a higher fitness than smaller conspecifics (Kingsolver and Huey 2008), however, high growth rates are costly (reviewed in Dmitriew 2011). Although snails had the same optimum for growth irrespective of thermal origin, warm origin snails had nevertheless a higher growth rate than cold or seasonal origin snails across all temperature treatments (i.e. CoGV). This was true in both the common garden (Fig. 8B) as well as in the reciprocal transplant experiment (Fig. 9B). It appears that the benefit of the high growth rate in the warm environment outweighs the potential costs: e.g. low survival at the low temperature. There is often a time constraint in environments with fast growth rate, but in warm environments this is a weak factor. Instead, size-selective predation (Arendt 1997; Urban 2007) mediated by three-spine stickleback (Gasterosteus aculaeatus), which occurs at high densities at the warm sites and readily preys on snails (Einarsson et al. 2004), can potentially select for increased growth which is further aided by relaxed thermodynamic constraints (‘Hotter is better’).

30 Temperature preference and tolerance (IV) When tested in a stable temperature gradient ranging from 1 to 25 °C, laboratory reared snails of different thermal origins had a common preferred temperature at ca. 17 °C (Fig. 8B). The distribution of average individual preferred temperatures was different from that of a simulated random distribution indicating that the snails had indeed chosen a temperature. Snails in Lake Mývatn had a common Topt for growth at 20 °C irrespective of thermal origin (III). Although different traits can have different Topt, Tp should be adjusted to an overall Topt. That snails have a common Tp is thus in line with the assumption that growth and size are important factors affecting many ectotherm life-history traits (Kingsolver and Huey 2008). A Tp of ca. 17 °C agrees with theory as Tp should be lower than the observed Topt as thermal reaction norms have a greater performance reduction at temperatures above than below Topt (Martin and Huey 2008; Angilletta et al. 2010). Although snails did not vary in preferred temperature they showed a difference in maximum temperature tolerance (Fig. 10). Warm origin snails had lower maximum tolerance than cold and seasonal origin snails (Fig. 10). Lake chub (Couesius plumbeus) originating from temporally stable warm environments have been shown to reduce CTmax compared to those from a varying warm thermal environment (Darveau et al. 2012). Thus, the solution to this seemingly illogical result may lie in the thermal characteristics of the springs. The risk of experiencing temperatures close to the estimated CTmax is limited as temperatures at the thermal springs are i) relatively stable over time, and ii) temperature fluctuations at the warm sites only reduce the water temperature (i.e. cooler surface water coming in during a storm). In addition, warm environments are often associated with other stress factors, e.g. parasite prevalence (Karvonen et al. 2010; Karvonen et al. 2013). Assuming a maintenance cost of CTmax (e.g. heat shock proteins; Sørensen et al. 2003) , it is possible that selection has reduced CTmax in favour of a higher growth rate

100 100 100 Cold Seasonal Warm

75 75 75

50 50 50

CT50 = 38.5 C CT50 = 38.0 C CT50 = 36.7 C

Attachment (%) 25 Attachment (%) 25 Attachment (%) 25 CTmax  = 38.0 C CTmax  = 37.5 C CTmax  = 36.3 C

0 0 0 30 32 34 36 38 40 42 30 32 34 36 38 40 42 30 32 34 36 38 40 42 Temperature (C) Temperature (C) Temperature (C) Figure 10. Percentage of R. balthica remaining attached plotted against temperature. CT50 (dashed line) is estimated based on the fitted polynomial (solid line). Both values of CT50 and CTmax are shown in the figure.

31 and parasite resistance (Zakrzewska et al. 2011; Roze et al. 2013; but see Karl et al. 2013). The available evidence suggests that the thermal breadth is most often regulated by changing the CTmin (reviewed in Addo-Bediako et al. 2000) and that such adaptation can occur rapidly (Barrett et al. 2011; Leal and Gunderson 2012). Hence, further studies focusing on the lower performance limit in snails originating from the thermal springs of Lake Mývatn are needed in order to fully understand how thermal environment, gene flow and performance is linked to the evolution of thermal performance curves.

32 Conclusion

To date there is no unifying theory of thermal adaptation. Nevertheless, there are several aspects of thermal adaptation that are similar across species or even taxa; 80 % of all ectotherms obey the ‘Bigger is better’ and ‘Hotter makes you smaller’ rules stating that a larger size increases fitness and a higher rearing temperature in the laboratory will cause a reduction in size at maturation. The latter rule is complemented by the ‘Temperature-size’ rule stating that ectotherms living in colder environments will mature at a larger size than those in warmer areas. Recent empirical studies and theoretical work have highlighted that the classical optimality models of thermal adaptation, which assume a tight link between environmental temperature and thermal optimum as well as a general trade-off between performance and thermal breadth, do poorly in their ability to explain observed patterns of how organisms perform in different thermal environments. Instead, models incorporating genetic and environmental covariance are receiving more and more attention. Irrespective of the sire of collection in Lake Mývatn, the snails seemed to obey general rules of thermal adaptation mentioned above, and my results also provide evidence that models including genetic and environmental covariance are more successful in explaining observed patterns than general optimality models. However, given the number of traits involved in thermal adaptation it is not surprising that the picture is complex with variation in some traits best explained by one model and variation in another trait by another, e.g. mortality explained by optimality models and growth by cogradient variation. In addition, my results also emphasize that evolutionary models alone are not the way towards a unifying theory of thermal adaptation as ecological aspects and thermal characteristics of the environment can have a great impact on the resulting level of adaptation. For instance, a cogradient pattern of growth rate appears unusual in studies of adaptation (thermal or otherwise) as cold origin organisms released from environmental influence often show a countergradient pattern in metabolic traits. Here, the cogradient pattern in growth rate could potentially be explained by a combination of relaxed thermodynamic constraints, differences in the strength of time constraints between habitats and differing predation pressure between age groups. A reduced tolerance to high temperatures in warm origin snails appears counterintuitive, but can putatively be explained by differences in energetic costs and thermal characteristics between thermal environments.

33 For example, warm origin snails inhabit a highly stressful environment potentially entailing increased costs of, for example, growth and parasite resistance as compared to cold and seasonal thermal environments. Given that the warm habitats are temporally stable and highly deleterious temperatures are never experienced, a trade-off between thermal tolerance and other fitness components may explain the low CTmax. In contrast, this trade-off may not be manifested in cold and seasonal environments and therefore the tolerance to high temperatures remains at a relatively high level. Additional evidence for the stressfulness of warm environments comes from the genetic data. I found a general decrease in genetic variation with increasing temperatures suggesting that specialized genotypes fare the best in warm habitats. Intriguingly, seasonal origin snails appear more similar to cold than warm origin snails in genetic profile, growth rate and maximum thermal tolerance but not in survival and egg production suggesting that the environment does not affect all traits equally. As adaptation to any environment is constrained, and can be hindered by drift and gene flow, any results and conclusions must be viewed when considering these factors. My results suggest that even a relatively low level of gene flow can disrupt local thermal adaptation and create joint optima for a range of traits. However, adding to the complexity is the fact that gene flow does not seem to affect all traits equally as snails had a joint optimum for growth rate and fecundity measures but differed in mortality and thermal tolerance. Thus, in addition to models incorporating genetic and environmental covariance, and not forgetting the conclusions drawn from optimality models, a joint theory of thermal adaptation should include aspects of ‘thermal ecology’ (e.g. differences in predation and parasite prevalence), geography as well as patterns and magnitude of gene flow and – of course – the thermal characteristics of the environment. Disentangling these factors will keep students of thermal evolution occupied for many years in the future.

34 English summary

In one way or another we are all affected by temperature but ectotherms, species relying on external heat sources, are especially sensitive. Since ectotherms make up most of earth’s fauna and global temperatures are increasing at an exceptional rate, the importance of research on thermal adaptation in general and on ectotherms specifically is increasing. There is abundant research on large-scale geographic patterns of thermal adaptation, but thermal adaptation has more seldom been investigated in small-scale natural systems incorporating gene flow, the movement of genes between habitats. In this thesis I present a series of studies on the common pond snail (Ra- dix balthica, Linnaeus 1758) inhabiting contrasting thermal environments in Iceland. Experiments in papers I, III and IV have been conducted on snails originating from Lake Mývatn. Here the groundwater emptying into the lake passes over two distinct geothermal areas – in the south the groundwater is heated to a temperature of ca. 6°C (‘cold’ sites) while in the northern areas the groundwater reaches ca. 24 °C (‘warm’ sites). This creates a mosaic pattern of temporally stable thermal springs and habitats affected by seasonal temperature variation (‘seasonal’ sites). I investigated genetic spatial distribution patterns on both small and large geographic scales. In paper I, I concentrated on finding genetic signatures associated with temperature in snails sampled from contrasting thermal locations within Lake Mývatn. By dividing the genetic dataset into two parts – one in which the genetic differences were more diverged than one would expect by chance and one that reflects the ‘neutral’ data – I found a general separation in genetic markers corresponding to the southern and northern basins of Lake Mývatn and that some elements of both mitochondrial and nuclear DNA associate with either low or high temperatures. In paper II, I was able to decouple the effects of distance and temperature on the genetic patterns using a more widespread sampling design including thermal sites in southwestern Iceland as well as an exhaustive sampling within Lake Mývatn. Besides verifying patterns observed in paper I, I was also able to distinguish other, smaller, distinct areas beside the basins in Lake Mývatn and a more detailed pattern of gene flow. In paper III, I studied several aspects of R. balthica life history – mortality, growth, age and size of maturation and number of offspring – in relation to temperature on laboratory reared snails as well as in a field experiment conducted on wild snails. I found that variation in temperature-

35 mediated survival depended on thermal origin. In the common garden warm origin snails had reduced survival at the low temperature treatment. In the field experiment there was no difference in survival between snails of different thermal origin in the cold and seasonal environments. However, when cold and seasonal origin snails were transplanted to the warm environment they showed increased mortality compared to warm origin snails. Growth rate and size at maturation are affected by temperature and have strong effect on fitness in ectotherms. I found that warm origin snails grew faster than cold or seasonal origin snails both in the laboratory and when transplanted in the field. Seasonal origin snails were often able to match the performance of warm origin snails, and there was no difference in age or size at maturity between the two. This suggests that time constraints, i.e. a short growth sea- son, is not an important factor in seasonal environments. Although the performance level differed, life history of snails of all thermal origins was similar in several aspects. For instance, they all had an optimum for growth at 20 °C, a larger size at maturation increased the number of offspring and a higher rearing temperature reduced the age and size at maturation. In paper IV I investigated the thermal preference and maximum tolerance of snails sampled from the differing thermal environments. Since the traits investigated in paper III appear to have different optima, it is interesting to examine the preferred temperature as it is assumed to be related to the overall performance. In snails from Lake Mývatn the average preferred temperature did not differ between thermal origins (ca. 17 °C). However, maximum temperature tolerance, defined as the temperature at which the snail is no longer able to maintain its grip on a surface, differed between thermal origins. Interestingly, it was the warm origin snails that had a lower tolerance than cold or seasonal origin snails. This was surprising as organisms from a warm environment are expected to have higher tolerance than organisms from colder areas. However, as warm origin snails inhabit an environment often associated with high energetic costs, e.g. high growth rate (III) and higher resistance to parasites, natural selection may have reduced costly temperature tolerance in a low temperature variation habitat. From my results it is clear that living in warm environment can be costly and require a higher degree of specialization than if one originates from a seasonal or cold environment. Additional evidence for this comes from the genetic data in paper II, where I found reduced genetic variation with increasing environmental temperature. Even if genetic variation could be reduced in small populations by chance and/or if inbreeding was increased, the most likely option is a combination of strong selection pressure and barriers to gene flow since there was no difference in time to selfing (III) between seasonal and warm origin snails. At the moment there is no unifying theory of thermal adaptation and the various patterns observed in my experiments highlight the complexity of the problem. Two main conclusions drawn from this thesis are 1) the need for a

36 holistic approach involving both ecology and evolution in models of thermal adaptation and 2) while many aspects of thermal adaptation are likely to be species specific, some of my findings, for example, the effect of temperature mean over variation, are of a general nature and could be applied to a wide range of organisms.

37 Svensk sammanfattning

På ett eller annat sätt påverkas vi alla av den omgivande temperaturen men ektotermer, arter som förlitar sig på externa värmekällor, är särkilt känsliga. Då ektotermer utgör den största delen av jordens fauna och temperaturer ökar globalt i en oöverträffad hastighet, ökar också vikten av forskning på temperaturanpassning generellt och hos ektotermer i synnerhet. Det finns en mängd forskning på storskaliga geografiska temperaturanpassningsmönster, men mer sällan har temperaturanpassning undersökts på en liten skala i na- turliga miljöer som tillåter genflöde, förflyttningen av gener mellan habitat. I den här avhandligen presenterar jag studier gjorda på den ovala damm- snäckan (Radix balthica, Linnaeus 1758) insamlade från skiljda temper- turmiljöer på Island. Experimenten i artikel I, III och IV har gjorts på snäckor insamlade från Mývatn. I denna sjö rinner grundvattnet in från två olika distinkta geotermiska områden – i söder hettas vattnet upp till cirka 6°C (”kall”) medan det i den norra delen värms upp till cirka 24°C (”varm”). Detta skapar ett mosaikmönster av källor där temperaturvariationen är liten bland miljöer påverkade av säsongsmässiga variationer (”säsong”). Jag undersökte genetiska mönster i både liten och stor geografisk skala. I artikel I koncentrerade jag mig på att hitta genetiska signaturer associerade med temperatur i snäckor insamlade från geotermiska källor inom Mývatn. Genom att dela upp datat i två delar – en där de genetiska signaturerna är mer skiljda än vad som kan förväntas av slumpen och en del som avspeglar det ”neutrala” datat – kunde jag särskilja en generell separation inom sjön som korresponderar till de södra och norra delarna och att vissa delar av både det nukleära och mitokondriella DNAt associerar till antingen låga eller höga temperaturer. I artikel II kunde jag sära på effekterna av geografiskt avstånd och temperatur på det genetiska datat genom en mer utbredd in- samling inom Mývatn. Utöver att bekräfta mönster som observerades i artikel I, så kunde jag också urskilja andra, mindre, områden förutom de södra och norra delarna av Mývatn samt ett mycket mer detaljerat genflödesmöns- ter. I artikel III studerade jag olika aspekter av livshistoria – livslängd, ålder och storlek vid första fortplantning och antal avkommor – i relation till temperature på laboratorieuppfödda snäckor samt i ett fältexperiment med vild- födda snäckor. Jag fann att variationen i överlevnad på grund av olika temperaturer beror på ursprungsmiljöns temperatur. I experimentet i laboratoriet hade snäckor från varma områden en reducerad överlevnad i den låga tempe-

38 raturbehandlingen. I fältexperimentet var det ingen skillnad i överlevnad mellan snäckor från olika temperaturmiljöer i kalla och säsongs- behandlingarna. Men när snäckor med kallt och säsongbetonat ursprung flyttades till den varma miljön visade de en ökad mortalitet jämfört med snäckorna som kom från den varma miljön. Tillväxt och storlek vid första fortplantning beror på temperatur och har stor inverkan på fitness hos ektotermer. Jag fann att snäckor från en varm miljö tillväxer snabbare än de andra snäckorna både i laboratoriet och i fält. Prestationen hos snäckor från ett säsongsbetonat område kunde ofta jämföras med den hos de från varma områden, och det var ingen skillnad i ålder eller storlek vid första fortplantning mellan dem. Detta antyder att tidsbegränsningar, d.v.s. en kort tillväxt- säsong, är inte av någon vikt i områden med säsongsvariation i temperatur. Fastän prestationsnivån skiljde så var livshistorievariabler ofta lika mellan snäckor från olika temperaturområden. Till exempel så hade de alla ett optimum för tillväxt vid 20 °C, en större storlek vid första fortplantning ökade antalet avkommor och en högre temperatur vid tillväxt minskade ålder och storlek vid första fortplantning. I artikel IV undersökte jag temperaturpreferensen och den maximala temperaturtoleransen i snäckor från de skiljda geotermiska källorna. Då alla livshistorieaspekter i artikel III kan ha olika optimum är det av stort intresse att kvantifiera den föredragna temperaturen då den anses vara relaterad till den övergripande prestationsförmågan. I snäckor från Mývatn skiljde inte medelpreferensen sig mellan de olika temperaturursprungen (ca. 17 °C). Dock så gjorde den maximala temperaturtoleransen, definierad som den temperatur då snäckorna inte längre klarar av att behålla sitt grepp vid un- derlaget, det. Intressant nog så var det de snäckorna från varma områden som hade den lägsta toleransen. Detta var överraskande då organismer från en varm miljö kan förväntas ha en högre tolerans än organismer från kalla områden. Men då dessa snäckor kommer från en miljö som ofta associeras med högre energikostnader, t.ex. hög tillväxt (III) och högre parasitresistans, kan selektion ha reducerat en kostsam temperaturtolerans i ett stabilt tempe- raturområde. Från mina resultat så är det uppenbart att en varm miljö är kostsam och kräver en högre grad av specialisering än om man kommer från en kall miljö eller en med säsongsvariationer i temperatur. Ytterligare bevis för detta kommer från det genetiska datat i artikel II, där jag fann att den genetiska variationen reducerades med ökande habitattemperatur. Även om den genetiska variationen kan reduceras av slumpen i små populationer och/eller med ökad inavel, så är den mest troliga förklaringen en kombination av starkt selektionstryck på grund av temperatur och minskat genflöde eftersom det inte var någon skillnad i tid till första fortplantning (snäckorna fick bara klona sig; III) mellan snäckor från ett säsongsbetonat område och de från varma områden.

39 För tillfället så finns ingen övergripande teori för temperaturanpassning och de skiljda mönstren som jag observerade i mina experiment belyser proble- mets komplexitet. Två huvudslutsatser från denna avhandling är: 1) att det behövs en helhetssyn som involverar både ekologi och evolution i tempera- turanpassningsmodeller och 2) att även om många aspekter av temperatu- ranspassning sannolikt är artspecifika så är några, till exempel att effekten av en högre medeltemperatur är större än effekten av temperatur-variation, av en mer generell natur och kan appliceras på många organismer.

40 Acknowledgements

First of all, I would like to acknowledge my supervisor Anssi for giving me this opportunity and aiding me throughout the course of my PhD. I have often left our discussions a bit more confused than at the start, but with the help of a few more discussions things have always cleared up. Thanks for your patience and for always having a moment to spare! Secondly, Jacob, the person that initially introduced me to the life at EBC in general and Popbio in particular by allowing me to do an internship, my Master project on the Irish Red grouse and agreeing to be my second supervisor. I am highly in debt to you! Although Anssi and Jacob are my supervisors on paper, you Rike, should also be mentioned in that regard. Without your help in the early (and later) stages of my PhD project I don’t know if this thesis would have existed. I especially think that we did a good job on my first trip to Iceland just one week into the project and with Anssi’s vague directions to follow. Now the memories of working frantically gathering snails until the last minute (liter- ary!) and then driving non-stop over night from Mývatn to Keflavík to catch an early flight home with just moments to spare, are almost pleasant now. And remember the leeches! Two of my dearest friends here at the department are David and Yvonne, with whom I have shared an office with for a long time. Thank you both for providing motivation and inspiration at times of need and for being there after working hours! I hope to continue being a part of your lives for a very long time! A person that has lightened the burden of being a PhD-student is Rado. Thank you so much for your company on field trips to Iceland and for allowing me to sit on your couch at times! There are many people here at EBC that have aided me in various ways during the course of these years and mentioning a few: Björn, David B, Örjan and Arild for numerous statistics and R related questions, Alex for designing the cover among many things, Germán for giving very useful comments on early (and late) drafts, Gunilla and Reija for countless questions regarding the lab, Anders for teaching me to see the light in a new and perhaps more technical way and The Photo Club for teaching me to see the light in a more creative way!, Frank for teaching me about birds, Ingrid for enlightening talks about teaching, Lina for being my partner in crime at Klubban, María Q for a lot of help with the AFLP, María C for having a

41 constant smile and spring in your step around the department, Robert E, Ben, Bruno, Robert D and Pabolo for being additional roomies, Thomas, Tanja, Carro, Jenny and Peter for showing me the ropes early on, Mir- jam, Paolo, Jossan and Mårten for the pub visits and much more! Many, many thanks to everyone!

Tank ju Eryn for corekting me englisch! ju ar verry nise!

Although I have spent most of my time at EBC I could not have done any- thing without the assistance of Árni at the research station at Mývatn. Thank you so much for getting the necessary permits from the land owners, which also should be thanked, and allowing me to use facilities and experiments at the station! I hope to see you and Unnur again in a not so distant future. Doing field work on snails on northern Iceland can sometimes be lonely but making up for that in a most entertaining way are the other researchers and most notably the ‘American crew’. Celebrating 4th of July with you guys, a bonfire, BBQ and countless s’mores is an event I’ll never forget! And thank you Jamie for greeting Rike and me with open arms and giving us the grand tour of the station on our first day. I had an equally enjoyable time with you Bjarni! I’m very grateful for all the help you provided and how you made sure that I felt at home in Hólar. Not the least by introducing me to the beer club and teaching me how to properly drink a beer! Many thanks also to Antoine, Gummi and Skúli! Every time I went to Iceland I stayed at a hostel close to the cathedral in Reykjavik – Our House. Thank you Bedda for being such a friendly face and making sure I would also enjoy Reykjavík on my visits to Iceland! I hope to see you again!

Tack till alla gamla och nya vänner både i Östersund, Uppsala och Delsbo för att ni behandlar mig som en “normal” person de gånger jag är på besök trots att jag forskar på snäckor. Erik och Micke – snart är jag klar med sko- lan och då ska jag skjuta min första älg och bli vuxen! Och syrran, nu kanske vi flyttar tillbaka norrut! Självklart ska också mamma och pappa ha en stor eloge. Tack för att ni alltid har stöttat mig. Jag kan aldrig säga tack nog många gånger!

Maja, tack för att du är du och för att du tycker om mig. Tack för att du har sporrat mig till oanade distanser (42 195m), lagoma höjder (1728 m.ö.h) och vid tillfället omöjliga tider (7h 42m)! Tack för att du har hjälpt mig alla de gånger jag har behövt det!

Tack Agnes för att du sätter perspektiv på tillvaron! Även om du inte vet om det nu har du bidragit mycket till denna avhandling!

42 References

Addo-Bediako, A., S. L. Chown, and K. J. Gaston. 2000. Thermal tolerance, climatic variability and latitude. Proc. R. Soc. B. 267:739-745. Angilletta, M. J. 2009. Thermal adaptation: a theoretical and empirical synthesis. Oxford Univ. Press, Oxford. Angilletta, M. J., R. B. Huey, and M. R. Frazier. 2010. Thermodynamic effects on organismal performance: is hotter better? Physiol. Biochem. Zool. 83:197-206. Antao, T. and M. A. Beaumont. 2011. Mcheza: a workbench to detect selection using dominant markers. Bioinformatics 27:1717-1718. Arendt, J. D. 1997. Adaptive intrinsic growth rates: an integration across taxa. Q. Rev. Biol. 72:149-177. Arendt, J. D. and D. N. Reznick. 2005. Evolution of juvenile growth rates in female guppies (Poecilia reticulata): predator regime or resource level? Proc. R. Soc. B. 272:333-337. Asbury, D. A. and M. J. Angilletta. 2010. Thermodynamic effects on the evolution of performance curves. Am. Nat. 176:E40-E49. Bargues, M. D., M. Vigo, P. Horak, J. Dvorak, R. A. Patzner, J. P. Pointier, M. Jackiewicz, C. Meier-Brook, and S. Mas-Coma. 2001. European Lymnaeidae (Mollusca: Gastropoda), intermediate hosts of trematodiases, based on nuclear ribosomal DNA ITS-2 sequences. Infection, Genetics and Evolution 1:85-107. Barrett, R. D. H., A. Paccard, T. M. Healy, S. Bergek, P. M. Schulte, D. Schluter, and S. M. Rogers. 2011. Rapid evolution of cold tolerance in stickleback. Proc. R. Soc. B. 278:233-238. Barton, M., P. Sunnucks, M. Norgate, N. Murray, and M. Kearney. 2014. Co-gradient variation in growth rate and development time of a broadly distributed butterfly. Plos One 9:e95258. Beaumont, M. A. and D. J. Balding. 2004. Identifying adaptive genetic divergence among populations from genome scans. Mol. Ecol. 13:969- 980. Beaumont, M. A. and R. A. Nichols. 1996. Evaluating loci for use in the genetic analysis of population structure. Proc. R. Soc. B. 263:1619-1626. Bennett, A. F. 1987. Evolution of the control of body temperature: is warmer better? in P. Dejours, L. Bolis, C. R. Taylor, and E. R. Weibel, eds. Comparative physiology: Life in water and on land. Liviana Press, Padova.

43 Bjornsson, H. and T. Jonsson. 2004. Climate and climatic variability at Lake Myvatn. Aquat. Ecol. 38:129-144. Blanquart, F., O. Kaltz, S. L. Nuismer, and S. Gandon. 2013. A practical guide to measuring local adaptation. Ecol. Lett. 16:1195-1205. Blier, P. U., F. Dufresne, and R. S. Burton. 2001. Natural selection and the evolution of mtDNA-encoded peptides: evidence for intergenomic co- adaptation. Trends Genet. 17:400-406. Bonin, A., P. Taberlet, C. Miaud, and F. Pompanon. 2006. Explorative genome scan to detect candidate loci for adaptation along a gradient of altitude in the common frog (Rana temporaria). Mol. Biol. Evol. 23:773- 783. Bridle, J. R. and T. H. Vines. 2007. Limits to evolution at range margins: when and why does adaptation fail? Trends Ecol. Evol. 22:140-147. Chen, C., E. Durand, F. Forbes, and O. Francois. 2007. Bayesian clustering algorithms ascertaining spatial population structure: a new computer program and a comparison study. Mol. Ecol. Notes 7:747-756. Conover, D. and E. T. Schultz. 1995. Phenotypic similarity and the evolutionary significance of countergradient variation. Trends Ecol. Evol. 10:248-252. Conover, D. O., T. A. Duffy, and L. A. Hice. 2009. The covariance between genetic and environmental influences across ecological gradients reassessing the evolutionary significance of countergradient and cogradient variation. Ann. N. Y. Acad. Sci. 1168:100-129. Darveau, C. A., E. B. Taylor, and P. M. Schulte. 2012. Thermal physiology of warm-spring colonists: variation among lake chub (Cyprinidae: Couesius plumbeus) populations. Physiol. Biochem. Zool. 85:607-617. Day, T. 2000. Competition and the effect of spatial resource heterogeneity on evolutionary diversification. Am. Nat. 155:790-803. de Mazancourt, C., E. Johnson, and T. G. Barraclough. 2008. Biodiversity inhibits species’ evolutionary responses to changing environments. Ecol. Lett. 11:380-388. Díaz, F., A. D. Re, Z. Medina, G. Re, G. Valdez, and F. Valenzuela. 2006. Thermal preference and tolerance of green abalone Haliotis fulgens (Philippi, 1845) and pink abalone Haliotis corrugata (Gray, 1828). Aquac. Res. 37:877-884. Diaz, F., A. Salas, A. Denisse Re, M. Gonzalez, and I. Reyes. 2011. Thermal preference and tolerance of Megastrea (Lithopoma) undosa (Wood, 1828; Gastropoda: Turbinidae). J. Therm. Biol. 36:34-37. Dillon, R. T. 2010. The ecology of freshwater molluscs. Cambridge Univ. Press, Cambridge. Dmitriew, C. M. 2011. The evolution of growth trajectories: what limits growth rate? Biol. Rev. 86:97-116.

44 Dreyer, J., P. A. Townsend, J. C. Hook, D. Hoekman, M. J. Vander Zanden, and C. Gratton. 2015. Quantifying aquatic insect deposition from lake to land. Ecology 96:499-509. Durand, E., F. Jay, O. E. Gaggiotti, and O. Francois. 2009. Spatial inference of admixture proportions and secondary contact zones. Mol. Biol. Evol. 26:1963-1973. Dyer, R. J. and J. D. Nason. 2004. Population Graphs: the graph theoretic shape of genetic structure. Mol. Ecol. 13:1713-1727. Einarsson, A., G. Stefansdottir, H. Johannesson, J. S. Olafsson, G. M. Gislason, I. Wakana, G. Gudbergsson, and A. Gardarsson. 2004. The ecology of Lake Myvatn and the River Laxa: variation in space and time. Aquat. Ecol. 38:317-348. Ellis, A. E. 1926. British snails: a guide to non-marine gastropoda of Great Britain and Ireland. Oxford Univ. Press, Oxford. Endler, J. A. 1986. Natural selection in the wild. Princton Univ. Press, New Jersey. Falush, D., M. Stephens, and J. K. Pritchard. 2003. Inference of population structure using multilocus genotype data: linked loci and correlated allele frequencies. Genetics 164:1567-1587. Falush, D., M. Stephens, and J. K. Pritchard. 2007. Inference of population structure using multilocus genotype data: dominant markers and null alleles. Mol. Ecol. Notes 7:574-578. Felsenstein, J. 1976. Theoretical population genetics of variable selection and migration. Annu. Rev. Genet. 10:253-280. Foll, M. and O. Gaggiotti. 2008. A genome-scan method to identify selected loci appropriate for both dominant and codominant markers: a bayesian perspective. Genetics 180:977-993. Frisch, D., A. J. Green, and J. Figuerola. 2007. High dispersal capacity of a broad spectrum of aquatic invertebrates via waterbirds. Aquat Sci 69:568- 574. Galindo, J., P. Moran, and E. Rolan-Alvarez. 2009. Comparing geographical genetic differentiation between candidate and noncandidate loci for adaptation strengthens support for parallel ecological divergence in the marine snail Littorina saxatilis. Mol. Ecol. 18:919-930. Gardarsson, A. and A. Einarsson. 1994. Responses of breeding duck populations to changes in food-supply. Hydrobiologia 279:15-27. Gerald, G. W. and L. C. Spezzano. 2005. The influence of chemical cues and conspecific density on the temperature selection of a freshwater snail (Melanoides tuberculata). J. Therm. Biol. 30:237-245. Ghalambor, C. K., R. B. Huey, P. R. Martin, J. J. Tewksbury, and G. Wang. 2006. Are mountain passes higher in the tropics? Janzen's hypothesis revisited. Integr. Comp. Biol. 46:5-17. Gilchrist, G. W. 1995. Specialists and generalists in changing environments .1. Fitness landscapes of thermal sensitivity. Am. Nat. 146:252-270.

45 Gray, M. M., P. St Amand, N. M. Bello, M. B. Galliart, M. Knapp, K. A. Garrett, T. J. Morgan, S. G. Baer, B. R. Maricle, E. D. Akhunov, and L. C. Johnson. 2014. Ecotypes of an ecologically dominant prairie grass (Andropogon gerardii) exhibit genetic divergence across the US Midwest grasslands' environmental gradient. Mol. Ecol. 23:6011-6028. Hereford, J. 2009. A quantitative survey of local adaptation and fitness trade-offs. Am. Nat. 173:579-588. Hochachka, P. W. and G. S. Somero. 2002. Biochemical adaptation. Oxford University Press, Oxford. Hoekman, D., J. Dreyer, R. D. Jackson, P. A. Townsend, and C. Gratton. 2011. Lake to land subsidies: Experimental addition of aquatic insects increases terrestrial arthropod densities. Ecology 92:2063-2072. Hubisz, M. J., D. Falush, M. Stephens, and J. K. Pritchard. 2009. Inferring weak population structure with the assistance of sample group information. Mol. Ecol. Resour. 9:1322-1332. Huey, R. B. and P. E. Hertz. 1984. Is a jack-of-all-temperatures a master of none. Evolution 38:441-444. Huey, R. B. and R. D. Stevenson. 1979. Integrating thermal physiology and ecology of ectotherms: discussion of approaches. Am. Zool. 19:357-366. Ives, A. R., A. Einarsson, V. A. A. Jansen, and A. Gardarsson. 2008. High- amplitude fluctuations and alternative dynamical states of midges in Lake Myvatn. Nature 452:84-87. Janzen, D. H. 1967. Why mountain passes are higher in tropics. Am. Nat. 101:233-249. Joost, S., A. Bonin, M. W. Bruford, L. Despres, C. Conord, G. Erhardt, and P. Taberlet. 2007. A spatial analysis method (SAM) to detect candidate loci for selection: towards a landscape genomics approach to adaptation. Mol. Ecol. 16:3955-3969. Karl, I., R. Stoks, S. S. Bauerfeind, A. Dierks, K. Franke, and K. Fischer. 2013. No trade-off between growth rate and temperature stress resistance in four insect species. Plos One 8:e62434. Karvonen, A., B. K. Kristjánsson, S. Skúlason, M. Lanki, C. Rellstab, and J. Jokela. 2013. Water temperature, not fish morph, determines parasite infections of sympatric Icelandic threespine sticklebacks (Gasterosteus aculeatus). Ecol Evol 3:1507-1517. Karvonen, A., P. Rintamäki, J. Jokela, and E. T. Valtonen. 2010. Increasing water temperature and disease risks in aquatic systems: climate change increases the risk of some, but not all, diseases. Int. J. Parasitol. 40:1483- 1488. Kawecki, T. J. and D. Ebert. 2004. Conceptual issues in local adaptation. Ecol. Lett. 7:1225-1241. Kingsolver, J. G. and R. B. Huey. 2008. Size, temperature, and fitness: three rules. Evol. Ecol. Res. 10:251-268.

46 Kirkpatrick, M. and N. H. Barton. 1997. Evolution of a species' range. Am. Nat. 150:1-23. Knies, J. L., J. G. Kingsolver, and C. L. Burch. 2009. Hotter is better and broader: thermal sensitivity of fitness in a population of bacteriophages. Am. Nat. 173:419-430. Lakowitz, T., C. Bronmark, and P. Nystrom. 2008. Tuning in to multiple predators: conflicting demands for shell morphology in a freshwater snail. Freshw. Biol. 53:2184-2191. Leal, M. and A. R. Gunderson. 2012. Rapid change in the thermal tolerance of a tropical lizard. Am. Nat. 180:815-822. Lenormand, T. 2002. Gene flow and the limits to natural selection. Trends Ecol. Evol. 17:183-189. Lynch, M. and W. Gabriel. 1987. Environmental tolerance. Am. Nat. 129:283-303. Martel, A. and F. S. Chia. 1991. Drifting and dispersal of small bivalves and gastropods with direct development. J. Exp. Mar. Biol. Ecol. 150:131- 147. Martin, T. L. and R. B. Huey. 2008. Why "suboptimal" is optimal: Jensen's inequality and ectotherm thermal preferences. Am. Nat. 171:E102-E118. Merilä, J., A. Laurila, A. T. Laugen, K. Räsänen, and M. Pahkala. 2000. Plasticity in age and size at metamorphosis in Rana temporaria - comparison of high and low latitude populations. Ecography 23:457-465. Minor, E. S. and D. L. Urban. 2007. Graph theory as a proxy for spatially explicit population models in conservation planning. Ecol. Appl. 17:1771-1782. Mitchell, S. E. and W. Lampert. 2000. Temperature adaptation in a geographically widespread zooplankter, Daphnia magna. J. Evol. Biol. 13:371-382. Nilsson-Örtman, V., R. Stoks, M. De Block, and F. Johansson. 2012. Generalists and specialists along a latitudinal transect: patterns of thermal adaptation in six species of damselflies. Ecology 93:1340-1352. O'Gorman, E. J., D. E. Pichler, G. Adams, J. P. Benstead, H. Cohen, N. Craig, W. F. Cross, B. O. L. Demars, N. Friberg, G. M. Gíslason, R. Gudmundsdóttir, A. Hawczak, J. M. Hood, L. N. Hudson, L. Johansson, M. P. Johansson, J. R. Junker, A. Laurila, J. R. Manson, E. Mavromati, D. Nelson, J. S. Ólafsson, D. M. Perkins, O. L. Petchey, M. Plebani, D. C. Reuman, B. C. Rall, R. Stewart, M. S. A. Thompson, and G. Woodward. 2012. Impacts of warming on the structure and functioning of aquatic communities: individual- to ecosystem-level responses. Adv. Ecol. Res. 47:81-176. Palaima, A. and K. Spitze. 2004. Is a jack-of-all-temperatures a master of none? An experimental test with Daphnia pulicaria (Crustacea : Cladocera). Evol. Ecol. Res. 6:215-225.

47 Peakall, R. O. D. and P. E. Smouse. 2006. genalex 6: genetic analysis in Excel. Population genetic software for teaching and research. Mol. Ecol. Notes 6:288-295. Pfenninger, M., M. Cordellier, and B. Streit. 2006. Comparing the efficacy of morphologic and DNA-based taxonomy in the freshwater gastropod genus Radix (Basommatophora, Pulmonata). BMC Evol. Biol. 6:100. Pfenninger, M., M. Salinger, T. Haun, and B. Feldmeyer. 2011. Factors and processes shaping the population structure and distribution of genetic variation across the species range of the freshwater snail Radix balthica (Pulmonata, Basommatophora). BMC Evol. Biol. 11:135. Pfleger, V. 2000. Molluscs. Bookmart, Leicester. Pritchard, J. K., M. Stephens, and P. Donnelly. 2000. Inference of population structure using multilocus genotype data. Genetics 155:945-959. Rasband, W. S. 2011. ImageJ v1.45s. Reynolds, W. W. and M. E. Casterlin. 1979. Behavioral thermoregulation and the final preferendum paradigm. Am. Zool. 19:211-224. Rosenberg, M. S. and C. D. Anderson. 2011. PASSaGE: Pattern Analysis, Spatial Statistics and Geographic Exegesis. Version 2. Methods Ecol. Evol. 2:229-232. Roze, T., F. Christen, A. Amerand, and G. Claireaux. 2013. Trade-off between thermal sensitivity, hypoxia tolerance and growth in fish. J. Therm. Biol. 38:98-106. Sandison, E. E. 1967. Respiratory response to temperature and temperature tolerance of some intertidal gastropods. J. Exp. Mar. Biol. Ecol. 1:271- 281. Savage, V. M., J. F. Gillooly, J. H. Brown, G. B. West, and E. L. Charnov. 2004. Effects of body size and temperature on population growth. Am. Nat. 163:429-441. Schniebs, K., P. Gloer, M. V. Vinarski, and A. K. Hundsdoerfer. 2011. Intraspecific morphological and genetic variability in Radix balthica (Linnaeus 1758) (Gastropoda: Basommatophora: Lymnaeidae) with mophological comparison to other European Radix species. J. Conchol. 40:657-678. Slatkin, M. 1987. Gene flow and the geographic structure of natural populations. Science 236:787-792. Sunday, J. M., A. E. Bates, and N. K. Dulvy. 2011. Global analysis of thermal tolerance and latitude in ectotherms. Proc. R. Soc. B. 278:1823- 1830. Sørensen, J. G., T. N. Kristensen, and V. Loeschcke. 2003. The evolutionary and ecological role of heat shock proteins. Ecol. Lett. 6:1025-1037. Urban, M. C. 2007. The growth-predation risk trade-off under a growing gape-limited predation threat. Ecology 88:2587-2597. Wada, S., K. Kawakami, and S. Chiba. 2011. Snails can survive passage through a snail's digestive system. J. Biogeogr.

48 van Leeuwen, C. H. A., G. van der Velde, J. M. van Groenendael, and M. Klaassen. 2012. Gut travellers: internal dispersal of aquatic organisms by waterfowl. J. Biogeogr. 39:2031-2040. Wang, I. J. and G. S. Bradburd. 2014. Isolation by environment. Mol. Ecol. 23:5649-5662. Wang, I. J. and K. Summers. 2010. Genetic structure is correlated with phenotypic divergence rather than geographic isolation in the highly polymorphic strawberry poison-dart frog. Mol. Ecol. 19:447-458. Vekemans, X. 2002. AFLP-SURV version 1.0. Distributed by the author. Laboratoire de Génétique et Ecologie Végétale, Université Libre de Bruxelles, Belgium. White, R. L. 1983. Effects of acute temperature change and acclimation temperature on neuromuscular function and lethality in crayfish. Physiol. Zool. 56:174-194. Vinarski, M. V. and E. A. Serbina. 2012. Distribution and quantitative characteristics of common species of pond snails of the subgenera Peregriana and Radix (Mollusca: Gastropoda: Lymnaeidae) in waterbodies of the south of Western Siberia. Inland Water Biol. 5:192- 198. Vos, P., R. Hogers, M. Bleeker, M. Reijans, T. Vandelee, M. Hornes, A. Frijters, J. Pot, J. Peleman, M. Kuiper, and M. Zabeau. 1995. AFLP - a new technique for DNA-fingerprinting. Nucleic Acids Res. 23:4407- 4414. Wright, S. 1943. Isolation by distance. Genetics 28:114-138. Yamahira, K. and D. O. Conover. 2002. Intra- vs. interspecific latitudinal variation in growth: adaptation to temperature or seasonality? Ecology 83:1252-1262. Zakrzewska, A., G. van Eikenhorst, J. E. C. Burggraaff, D. J. Vis, H. Hoefsloot, D. Delneri, S. G. Oliver, S. Brul, and G. J. Smits. 2011. Genome-wide analysis of yeast stress survival and tolerance acquisition to analyze the central trade-off between growth rate and cellular robustness. Mol. Biol. Cell 22:4435-4446.

49 Acta Universitatis Upsaliensis Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 1251 Editor: The Dean of the Faculty of Science and Technology

A doctoral dissertation from the Faculty of Science and Technology, Uppsala University, is usually a summary of a number of papers. A few copies of the complete dissertation are kept at major Swedish research libraries, while the summary alone is distributed internationally through the series Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology. (Prior to January, 2005, the series was published under the title “Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology”.)

ACTA UNIVERSITATIS UPSALIENSIS Distribution: publications.uu.se UPPSALA urn:nbn:se:uu:diva-248495 2015