Physiological and Environmental Processes Influencing Growth Strategies in Amphibian Larvae

List of Papers

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

I Dahl, E., Backström, T., Winberg, S. and Laurila, A. Is growth hormone expression correlated with variation in growth rate along a latitudinal gradient in Rana temporaria. (Submitted manuscript)

II Dahl, E., Orizaola, G., Winberg, S. and Laurila, A. Geographic variation in corticosterone response to chronic predator stress in tadpoles (Manuscript)

III Orizaola, G., Dahl, E., Nicieza, A.G. and Laurila, A. Breeding phenology influences life-history and anti-predator strategies in time-constrained amphibians (Manuscript)

IV Orizaola, G., Dahl, E. and Laurila, A. Compensating for de- layed hatching across consecutive life-history stages in an am- phibian. Oikos 119:980-987

V Dahl, E., Orizaola, G., Nicieza, A.G. and Laurila, A. Seasonal time constraints, compensatory strategies and life history switch points in larval amphibians (Manuscript)

Paper IV is reproduced with permission from the publisher, Nordic Ecologi- cal Society.

Cover illustration by Jonathan Dahl

© Jonathan Dahl

Contents

Introduction...7

Hormonal mechanisms associated to latitudinal gradients in growth (I, II)...9

Influences of breeding phenology on anti-predator strategies and life history (III) ...10

Compensatory strategies in response to environmental stressors (IV, V) 11 Methods ...13

Study species and populations...13

Rearing conditions in the laboratory ...15

Corticosterone manipulation (II) ...16

Temperature and food manipulations (IV, V) ...17

Behaviour (III, V)...17

Morphology (I, II, III, IV, V) ...18

Jumping trials (III, IV, V) ...19

Lipid content (III, V) ...19

Quantitative PCR (I)...20

Thin layer chromatography and Radio immune assay (II) ...21

Results and discussion ...22

Is growth hormone expression correlated with variation in growth rate along a latitudinal gradient in Rana temporaria? (I)...22

Geographic variation in corticosterone response to chronic predator stress in tadpoles (II) ...24

Breeding phenology influences anti-predator strategies in tadpoles across a geographic gradient (III)...26

Compensating for delayed hatching across consecutive life-history stages (IV) ...29

Geographic variation in compensatory growth strategies in response to temperature and food stress (V) ...31

General conclusions ...35

Sammanfattning på svenska...38 Fysiologisk- och miljömässig påverkan av tillväxtstrategier hos

amfibielarver ...38

Acknowledgements...41

References...42

Introduction

Growth rate is a central parameter for an individual’s life history (Roff, 1992; Arendt, 1997; Conover, Duffy & Hice, 2009), and for most animals there is a strong positive association between size and fitness (Roff 1992).

For example, larger individuals have a lower risk of being caught by preda- tors (Arendt 1997) and have a greater reproductive success (Roff, 1992). In theory, the advantages of achieving a large body size should maximize growth and erode genetic variation in growth rate among individuals and populations. Hence, the occurrence of low growth rates is perplexing and suggests that there are costs associated to fast growth leading to evolution of a growth rate where costs and benefits of growth are optimized (Arendt, 1997; Dmitriew, 2011). Perhaps the most immediate cost known to be asso- ciated to fast growth is increased risk of predation, presumably due to higher foraging rates, leading to a fundamental trade-off where the benefits of fast growth are weighed against the risk of being preyed on (Werner & Anholt 1993; Lima 1998; McPeek 2004). However, the costs and benefits of fast growth are likely to vary with the environmental conditions one encounters.

Animals with complex life cycles have distinct larval and adult stages. In seasonal environments, time constraints are a major issue for such animals as there are several environmental factors that can constrain larval development and the timing and conditions of metamorphosis (Wilbur 1980). The length of the growth season is also a major selective factor especially in ectotherms and has lead to that species in higher latitude and altitude habitats have faster growth and developmental rates (see Nylin & Gotthard, 1998; Conover et al., 2009 for reviews). Such variation is adaptive as it enables completion of growth and development under more stringent time constraints. There is also evidence that predator densities are lower at higher latitudes (Schemske et al. 2009) possibly decreasing the costs of high growth rates at high latitudes.

However, growth and developmental rates are very plastic as they are af- fected by various environmental factors, and can even be increased by artifi- cially imposing greater time stress through delayed hatching (Stoks et al.

2006) or after imposing a period of low growth during early development (Metcalfe and Monaghan 2001; Ali et al. 2003). Such compensatory growth provides compelling evidence that animals are normally growing at sub- maximal levels (Metcalfe & Monaghan 2001; Ali et al. 2003).

There are several issues associated to growth strategies that remain to a large

extent unexplored. Few studies have investigated physiological processes

underlying variation in growth (but see Billerbeck et al. 2000; Lindgren &

Laurila, 2005; 2009; Stoks et al. 2006; Arnott et al. 2006 for energetic ex- amples) and hormonal mechanisms are almost completely unexplored (but see e.g., Silverin et al. 1997 and Hau et al. 2010 for field comparisons).

Also, although breeding phenology normally fluctuates from year to year leading to temporal differences in the strength of the time-constrains no stud- ies have investigated the effects that early and late breeding phenology can have on latitudinal differences in growth, development and the expression of behavioural and morphological defences. Further, compensatory strategies following periods of low growth have rarely been compared in species or populations from different environments (but see Schultz et al. 2002; Sogard

& Olla 2002, Alvarez & Metcalfe 2007; De Block et al. 2008) and although many studies have examined compensatory responses in the short-term, there is a lack of studies examining these responses over broader amplitude of life-history stages (e.g. embryo-larva-metamorph transitions; but see e.g., Altwegg 2002; De Block et al. 2008).

Amphibian larvae provide an excellent study system for investigating fac- tors influencing growth strategies. Tadpoles are exposed to a wide array of stressors affecting growth in the aquatic environment, including temperature fluctuations, seasonal time stress and predation risk. Working with amphib- ian larvae, we can examine the effects that environmental factors have on key life-history traits, such as the timing of and size at metamorphosis, which can crucially affect individual fitness (Semlitsch et al. 1988; Altwegg

& Reyer 2003; Scott et al. 2007). Also, many amphibian species are distrib- uted across wide geographic ranges, and show variation in growth and de- velopment along environmental gradients (Morrison & Hero 2003), which makes them suitable for population comparisons. Previous studies have found that the common frog Rana temporaria exhibits a latitudinal pattern with increasing larval growth and development rates towards higher latitudes in Scandinavia (e.g., Merilä et al. 2000; Laugen et al. 2003, Lindgren &

Laurila 2009). High latitude tadpoles are also more active than southern tad- poles, a difference that persists in the presence of predators and leads to higher mortality rates when faced with predators (Laurila et al. 2008). Papers I-III and V in this thesis investigate factors affecting variation in growth and development among R. temporaria populations along this latitudinal gradi- ent.

This thesis consists of three main topics related to growth strategies;

Hormonal mechanisms associated to latitudinal gradients in growth (I, II),

influences of breeding phenology on growth and anti-predator strategies

(III) and compensatory strategies in response to environmental stressors (IV,

V)

Hormonal mechanisms associated to latitudinal gradients in growth (I, II)

The growth hormone, GH, and the stress hormones cortisol and corticoster- one (both referred to as glucocorticoids, GCs), are particularly interesting in the context of adaptive gradients and local adaptation in growth strategies.

High levels of GH are associated to high growth rate (Johnsson & Björnsson, 1994; Huang & Brown, 2000) while elevations in GCs are known to de- crease growth (Barton et al. 1987; Glennmeir & Denver 2002; Hayes & Wu 1995; Belden et al. 2005). Environmental stressors are known to affect both hormones, where GCs increase in response to stress (Sapolsky 2002) and GH generally decreases (e.g., Sheridan 1986; McCormick et al., 1998, but see Terry et al., 1977; Pickering et al., 1991) making them good candidates in proximally causing, or at least being associated to the decrease in growth rates that is often found in conjunction with predator stress (Lima 1998; Pea- cor & Werner 2004; Bolnick et al. 2005). Interestingly, these hormones also affect behaviour. GH manipulated animals are more active, have higher for- aging rates and weaker predator avoidance behaviour (Johnsson et al. 1996;

Abrahams & Sutterlin, 1999), while GCs can increase predator defence be- haviours (Thaker et al. 2009; Kalynchuk et al. 2004) and, when adminis- trated chronically, reduce foraging rates (Gregory & Wood 1999). Hence, GH and GCs may also prove to be of importance in mediating the behav- ioural responses as well as the costs associated to predation stress.

For papers I and II I wanted to investigate if GH or corticosterone (CORT, the main GC in amphibians) play a role in the latitudinal patterns in growth and predator induced effects found in R. temporaria. As a first step, I measured the expression of GH and GHR (growth hormone receptor) (I) and CORT (II) in response to chronic predation stress in tadpoles from eight R.

temporaria populations collected along a 1500 km latitudinal gradient across

Sweden. Two measurements were made for each hormone on laboratory

raised tadpoles in the presence or absence of predatory dragonfly larvae

(Aeshna sp.). The measuring occasions occurred 10 (GH/GHR) or 14

(CORT) days apart. For paper II I also manipulated CORT levels of tadpoles

in a subset of the populations.

I made the following predictions:

(I) 1) High-latitude tadpoles will have higher GH and/or GHR ex- pression, 2) tadpoles raised with predators will have lower GH and/or GHR expression, but 3) the difference will be smaller in the high latitude populations.

(II) 1) Tadpoles will respond to predation risk with elevated CORT response, however, 2) high-latitude tadpoles should show a weaker CORT response to avoid costs associated with elevated CORT lev- els. If not, 3) I expect northern tadpoles to be less sensitive to exter- nally administrated CORT in terms of decreased growth and devel- opmental rates.

Influences of breeding phenology on anti-predator strategies and life history (III)

In this study I investigated the variation of antipredator strategies in R. tem- poraria tadpoles from populations at the extremes of the geographic gradient across Sweden, focusing on the role that early and late breeding phenology can have on life history and the expression of behavioural and morphological defences. As artificially imposing early breeding phenology is very difficult to conduct in this non-model species, I made use of the natural variation in phenology. I conducted studies over two consecutive years with contrasting climatic conditions, which led to differences in breeding phenology and growth season length in the more time-constrained northern populations.

Previous studies have shown that the northern populations invest less in be- havioural antipredator defences than the southern ones, likely as a conse- quence of the more severe time constraints and lower temperature in north- ern latitudes (Laurila et al. 2008). If plasticity is maintained for antipredator traits, it is likely that under more relaxed time constrains, the northern popu- lations will increase their investment in antipredator defences.The main pre- dictions in paper III were:

1) Northern tadpoles will have shorter larval periods and react more weakly

to predator presence than southern tadpoles (i.e. maintain higher activity

rates and less developed morphological defences). 2) Under late breeding in

the north these differences will be stronger, and 3) under more benign condi-

tions in the north (i.e. early breeding), northern tadpoles will increase their

larval period and will be able to improve their predator response.

Compensatory strategies in response to environmental stressors (IV, V)

Environmental conditions experienced early in the ontogeny can have a strong impact on individual fitness and performance later in life. Organisms may counteract the negative effects of poor developmental conditions by showing compensatory responses in growth and development. Immediate and delayed costs may prevent the generalization of compensatory growth strategies and the optimal strategy may depend on environmental conditions.

However, previous studies on compensatory responses have largely ignored the effects that poor embryonic and larval conditions could have during later life stages. I examined the effects poor growing conditions in early life can have in later life history transitions in the moor frog (R. arvalis) (IV) and in R. temporaria (V). In paper IV, I examined the effects of artificially delayed development in early life over two later life history transitions. I investigated the compensatory growth of larvae of R. arvalis. The embryos were exposed to several temperature regimes, and we examined the extent and possible costs of compensatory growth mechanisms during later life stages, both in the aquatic (i.e. larval stage) and terrestrial environments (i.e. juvenile stage). In paper V, I compared populations of R. temporaria at the extremes of the latitudinal gradient in Sweden in their response to reduced growth imposed by two different stressors, low food and low temperatures. The stress treatments were conducted both in the presence and absence of a predator, and I measured possible costs associated to high growth (increased activity in the presence of predators, lipid content and locomotor perform- ance). As it is particularly important to achieve a large size during the short growing season, high latitude populations may be more efficient at compen- sating after period of low growth (Metcalfe et al. 2002), alternatively, high latitude populations are already growing at their maximum rate allowing little possibility for further increases in growth (Sogard & Olla 2002; De Block et al. 2008).

I made the following predictions on compensatory strategies of anuran lar- vae:

(IV) 1) R. arvalis exposed to lower temperature regime as embryos should show compensatory growth and development to complete metamorphosis in the shortest possible time and 2), the response will be more intense the colder the environment. 3) Costs associated with compensatory growth should appear during the ontogeny in the form of lower larval survival, changes in larval or postmetamorphic mor- phology, or reduced juvenile locomotor performance.

(V) 1a) Northern tadpoles will show stronger compensation after a

period of low temperatures or low food, with metamorphic timing

and size being more similar to that of tadpoles growing under con-

stantly favourable conditions. Alternatively, 1b) southern tadpoles

will show stronger compensatory growth because they have a greater

physiological capacity to do so. 2) Compensatory growth will induce

costs in terms of more risk-taking behaviour (higher activity in the

presence of predators), lower lipid levels or impaired locomotor per-

formance after metamorphosis. 3) These costs will be greater in the

populations showing stronger compensatory growth. 4) Due to the

higher cost of increasing growth in the presence of predators, there

will be a weaker increase in activity in predator exposed tadpoles

and, 5) this will be reflected in the degree to which tadpoles are able

to compensate.

Methods

Study species and populations

The common frog, Rana temporaria (I-III, V) (fig 1) is the most widespread amphibian in Europe, occurring from northern Spain to the coast of the Ber- ing Sea (Gasc et al. 1997) and makes an excellent model for studying adap- tation along environmental gradients. R. temporaria breeds in early spring in a variety of freshwater habitats from temporary ponds to shore marshes of large lakes. At high latitudes, it is often the only amphibian species present.

In Sweden, growth season length (defined as yearly number of days with mean temperatures over 5

C) is highly correlated with latitude, and the length of the growth season varies from 217 days in southern Sweden to 117 days in northern Sweden (Odin et al. 1983; Laugen et al. 2003). Adult R.

temporaria are mostly active during the growing season, during the winter they hibernate in water or buried in soil. Under natural conditions much of mortality in this species occurs during the larval stages where natural preda- tors include fish, newts and large aquatic insects. Each female lays a clutch of 500-2000 eggs.

In papers I and II samples were collected from eight populations along a 1500 km north-south gradient in Sweden, in papers III and V samples were collected from four populations located at the extremes of the gradient (Fig.

2, Table 1). I sampled ca. 500 eggs from each of 10 freshly laid clutches from each population.

Figure 1. Tadpole of R. temporaria, the study species in papers I-III and V.

F. Söderman.

In paper IV I used 14 full-sib families of the moor frog, Rana arvalis, ob-

tained from artificial matings of adults collected from populations outside

Skövde, south-western Sweden (Fig. 2, Table 1). R. arvalis is a widespread species of brown frog living in similar habitats as R .temporaria. It co-occurs with R. temporaria in most parts of Sweden except in the north and north- western regions.

Figure 2. Map of Scandinavia showing the studied populations of Rana temporaria and Rana arvalis.

Table 1. Coordinates and species collected from the different populations used. The papers in which the populations are examined are also indicated.

Locality Latitude (

N) Longitude (

E) Species Papers

Tvedöra 55

41’ 13

26’ R. temporaria I, II

Ållskog 55

33’ 13

48’ R. temporaria I, II, III, V

Måryds 55

42’ 13

21’ R. temporaria III, V

Skövde 58

46’ 13

76’ R. arvalis IV

Uppsala 59

51’ 17

28’ R. temporaria I, II

Nordsmyran 60

35’ 17

12’ R. temporaria I, II

Homlsjön 63

58’ 20

25’ R. temporaria I, II

Mjösjö 63

58’ 19

34’ R. temporaria I, II

Jukkasjärvi 67

54’ 21

02’ R. temporaria I, II, III, V

Björkliden 68

41’ 18

68’ R. temporaria I, II, III, V

Karesuando 68

27’ 22

25’ R. temporaria II

Rearing conditions in the laboratory

Papers I-III and V

After collection, eggs were immediately transported to the laboratory in Uppsala. Before the start of the experiments, eggs and tadpoles were evenly distributed in 3L opaque plastic containers. Eggs and tadpoles were kept in a 19 °C temperature controlled room throughout the studies, except during the low temperature treatments (V). Photoperiod was 16 h light: 8 h dark. Water was changed every three days before the start of the experiments and every seven days during the experiments (except in hormone manipulation experi- ment in II where water was changed every four days). To ensure homoge- nous water quality I used reconstituted soft water (RSW: APHA 1985). After hatching, tadpoles were fed ad libitum with finely chopped and lightly boiled spinach. Tadpoles were placed into the experimental treatments once they reached developmental stage 25 (complete gill resorbtion, Gosner 1960).

Tadpoles were haphazardly chosen from each population and divided into groups of 10 (III, V) or 20 (I, II) and placed in opaque plastic containers (38 x 28 x 13 cm) which were used as experimental units. Each container was filled with 10 L of RSW and (except in the CORT manipulation experiment in II) provided with cylindrical transparent mesh-bottom predator cage (di- ameter 11 cm, height 21) hung 2 cm above the container bottom. In predator present treatment the cage was provided with a late-instar larva of Aeshna sp. dragonfly (Fig. 3). Aeshna larvae, which are voracious predators of tad- poles, were collected from ponds near Uppsala.

Figure 3. Predator and no-predator treatments in experiments for papers I-III and V.

Paper IV

In this study it was imperative that the rearing temperatures of the embryos were controlled throughout embryonic development. The experiment was therefore conducted with embryos from artificial crosses in the lab. Adult R.

arvalis were transported to the laboratory in Uppsala, where they were artifi- cially mated within a few days of collection (see Räsänen et al. 2003 for details). A total of 14 full-sib families were used in the experiment. The fer- tilisations were conducted indoors at 15 °C, and the eggs were divided to the experimental treatments within four hours from fertilisation. Ca. 40 eggs from one family were placed in each of 52 1-litre plastic vials. During the experiment, embryos were kept in two distinct constant temperature rooms, set at 4 and 15 °C. Reconstituted soft water was used throughout the study and about half of the water volume was changed every third day. The light rhythm was 17 h light : 7 h dark.

Corticosterone manipulation (II)

This experiment consisted in four treatments, control, vehicle (ethanol only), 100 nM of CORT and 500 nM of CORT. Treatments were begun when the tadpoles had reached Gosner stage 25. There were ten tadpoles per container and ten containers per population in each of the four treatments (in total 80 containers). CORT (Sigma Aldrich, Product No. C2505) was dissolved in ethanol and administered to the water to give water concentrations of 100 or 500 nM CORT. These levels were chosen based on previous studies using exogenous CORT on tadpoles (Belden et al. 2005; Glennmeier & Denver 2002). The volume of ethanol in the water was 0.00001% of total water vol- ume. Vehicle-treated tanks received ethanol only, and control tanks received no treatment. CORT or ethanol was added to the tanks every four days in conjunction with each water change. After 21 days of treatment (and three days after the most recent CORT addition) five tadpoles were taken out from each treatment, rapidly rinsed in water, and frozen for CORT measurements.

Once tadpoles were approaching metamorphosis, the containers were

checked daily. Metamorphosed individuals (Gosner stage 42; emergence of

forelimbs) were removed and weighed to the nearest 0.1 mg, and the length

of larval period was recorded.

Temperature and food manipulations (IV, V)

Paper IV

The experiment consisted of four temperature treatments: constant 15 °C, daily day-night fluctuation between 15 and 4 °C throughout the embryonic development (mimicking normal outdoor conditions), early exposure to con- stant 4 °C, and late exposure to constant 4 °C. The last two treatments mim- icked changes in temperature associated to periods of cold weather shortly after laying and late in embryonic development, respectively. Both condi- tions are normal in nature and could be considered as intermediate tempera- ture treatments. In the early 4 °C treatment, eggs were moved on the second day of the experiment to the 4 °C room (ca. Gosner stage 14), where they stayed until day 6, whereas in the late 4 °C treatment eggs were moved to the 4 °C room on day 6 (ca. Gosner stage 19), staying there until day 10 of the experiment. Duration of the daily exposure to 4 °C in the daily fluctuating treatment was 10 hours (from 23PM to 9AM).

Paper V

Control tadpoles were kept in 19 °C and fed ad libitum (an amount of food that was sufficient for there to be some food left at each feeding) throughout the experiments. In the food experiment containers were assigned to one of four treatment combinations, no predator-control, predator-control, no preda- tor-low food and predator-low food. On days 1-7 all tadpoles were fed ad libitum, on days 8-19 the tadpoles assigned to the low food treatment were fed 1/6 of the ad libitum amount. After day 19, all tadpoles were fed ad libi- tum until metamorphosis. The treatments in the temperature experiment were no predator-control, predator-control, no predator-low temperature and predator-low temperature. All tadpoles were kept in 19 °C for the first 7 days. On days 8-14, tadpoles assigned to the cold treatment were placed in another temperature controlled room, set at 10 °C. At the end of day 14, they were moved back to the 19 °C room. The containers with tadpoles assigned to the constant 19 °C treatment were also moved on days 8 and 14, but put back into the same room, to control for any effect of transportation.

Behaviour (III, V)

Activity was measured by observing each aquarium for 10 seconds and re-

cording the proportion of tadpoles that were moving (e.g. swimming or for-

aging). Activity was scored 6 times per day. In paper III, I recorded the ac-

tivity level of the tadpoles on days 10, 16 and 23 of the experiment. For the

food manipulation experiment (V) activity was measured on day three, nine

and 12 in the low food treatment (10,16 and 19 days into the experiment) and on day one and three after the low food treatment, during the refeeding period (20 and 23 days into the experiment). For the temperature manipula- tion experiment (V) activity measurements were done on day three and seven of the cold treatment (10 and 14 days into the experiment) and on day one, three and eight after the cold treatment (day 15, 17 and 22 into the ex- periment).

Morphology (I-V)

Digitized images of tadpoles (I-IV) and Gosner stage 46 metamorphs (III,V) photographed against a sheet of graphing paper were used to analyze mor- phology. Body length, measured from the snout to the base of the tail, was used as an index of body size in I and II, and total body length (from snout to end of tail) was used as an index of size in Gosner stage 25 tadpoles in IV.

In IV, I also measured tail length, maximum body depth, tail muscle depth and tail fin depth in a subset of more developed larvae (stage 36). In juve- niles (stage 46) used in jumping trails (III-V) I measured the length of the tibia and fibula.

More extensive morphological analysis was done in III where morphology of the tadpoles was analysed using geometric morphometrics. Geometric morphometrics is a powerful tool for analysing and visualising morphologi- cal variation between individuals using digitised landmarks (Rohlf & Marcus 1993; Zelditch et al. 2004). A digital side-view image was taken of each tadpoles and to minimise the effect of non-parallel tail positions, all images were straightened with the straighten plugin in ImageJ (http://rsb.info.nih.gov/ij), and then loaded into MakeFan7 (Sheets 2009) to create a standardised template for digitising the landmarks. The body-shape of tadpoles was captured by digitising 19 landmarks on each individual (Fig.

4), using tpsDig2 software (Rohlf 2008). I conducted landmark-based geo-

metric morphometrics analyses on these digitised landmarks. Landmark

coordinates were imported into tpsRelw (Rohlf 2007) to create relative

warps, the equivalent of principal component scores in a multivariate geo-

metric morphometric analysis. Relative warp scores on tadpole morphology

were estimated using the consensus morphology for each experimental con-

tainer in order to avoid pseudoreplication. Consensus conformations were

estimated using tpsRelw (Rohlf 2007). All the analyses were conducted then

on relative warp scores.

Figure 4. Position of the 19 landmarks used for geometric morphometric analyses in R. temporaria tadpoles in paper III.

Jumping trials (III-V)

After metamorphosis, juveniles were placed in individual 0.8-litre opaque plastic containers filled to a depth of ca. 0.5 cm and provided with a small stone for resting. On the same day as juveniles had reached Gosner stage 46 (complete tail absorption) locomotor performance was examined on half of the metamorphosed individuals by measuring their jumping capacity. This trait was selected as a measure of fitness as it influences survival via effects on food acquisition (Walton 1988) and escape ability from predators (Wass- ersug & Sperry 1977). Jumping was tested at 20 °C. During the jumping trials froglets were individually placed in a small Petri dish contaning food dye positioned in a linear test track (120 x 60) with plastic walls (7 cm) and the floor covered with white paper. Jumping was induced by gently prodding the froglet on the urostyle, and jump lengths were measured from the colour marks left by individuals at the beginning and end of each jump. Froglets were allowed to perform five jumps in two series with a 1-hour resting pe- riod in between. Jumping performance of each individual was defined as the maximum jumping distance from the two measuring series. Using maximum jumping distance rather than the average of several jumps reduced the risk of underestimation of jumping abilities associated to occasional short jumps due to low motivation. After jumping trials frogs were weighed, photo- graphed and anesthesized with MS 222 before preservation in 70% alcohol.

Lipid content (III, V)

After jumping trials the froglets were preserved at -80 °C. For lipid content analyses the froglets were lyophilised, oven-dried at 37 °C overnight (to a constant mass) and weighed to the nearest 0.1 mg with a digital balance (fat dry mass). I extracted the total content of nonpolar lipids by Soxhlet methods using petroleum ether and seven cycles of lipid extraction (ca. 20 min/cycle).

Petroleum ether is highly efficient for the extraction of nonpolar (storage)

lipids, with little removal of polar (structural) lipids (Dobush et al. 1985).

After the extraction, I oven-dried the samples for 24 h and weighed them again (lean dry mass), and the lipid content was calculated as the different between fat and lean dry mass.

Quantitative PCR (I)

RNA was extracted from homogenized tadpoles (tail and gut discarded) us- ing a commercial kit (Qiagen Nordic, Solna, Sweden) according to the manufacturer’s instructions, and the samples were then treated with DNAse (DNA-free, Ambion, Austin, TX, USA) to remove DNA. The quantity and quality of the RNA extractions were checked using spectrophotometric analysis at 260/280 nm and samples judged to be non-qualitative were re- moved. cDNA was then prepared from 5 g of total RNA using Stratascript (Stratagene, AH Diagnostics AB, Stockholm, Sweden) according to the manufacturers instructions. After the cDNA synthesis, the reaction volume of 20 l was diluted to 200 l for subsequent use of 5 l/qPCR reaction.

Primers were designed for GH, GHR and Glyceraldehyde-3-phosphate de- hydrogenase (GAPDH) using R. ridibunda, R. catesbeiana and Xenopus laevis sequences (Table 2). Specificity of all the primers was tested and con- firmed functional for qPCR.

qPCR was performed in a solution of TRIS-HCl (10mM) pH 8.5, KCl

(50nM), MgCl2 (6mM), dNTP (0,2 mM), SYBR Green (1:50 000), using a

Rotor Gene 3000 instrument (Corbett Research, Cambridge, UK). Primer

concentration was 0.2 uM and AmpliTaq Gold (Applied Biosystems, Foster

City, CA, USA) was used at 1.25 U/reaction. Thermocycling parameters

were 95 °C for 15 min, then 40 cycles of 95 °C for 10 s, 58 °C for 15 s and

68 °C for 20 s. After thermocycling, melting point curves were included to

confirm that only one product was formed, if not the data were removed

from further analyses. For each thermocycling, standard curves were per-

formed including at least four serial dilution points of cDNA. The qPCR data

were analysed using Rotor-Gene 6 software. Quantification of the expression

levels of genes requires that expression is compared to that of a gene that is

expressed at a constant level and not affected by experimental conditions. In

this study I normalized the expression of GH and GHR to the amount of

GAPDH in each sample. GAPDH is a housekeeping gene that is commonly

used as internal standards for qPCR (e.g., Peinado et al., 2005). Since it was

only possible to run 72 samples at a time in the qPCR machine, samples

were run block-wise so that variation between runs could be separated from

treatment and/or population differences. Also, some samples were run re-

peatedly on all qPCRs so that blocks could be standardized against each

other.

Thin layer chromatography and Radio immune assay (II)

Whole body CORT content was determined by radioimmunoassay (RIA).

Samples from the two measuring time points (day 2 and 15 of the experi- ment) were assayed separately. Whole body homogenates were used for each assay. Each tadpole was homogenized in 2 ml ethylacetate. For individual recovery determination, 3000 cpm of triated (radioactive) corticosterone (Perkin Elmer Product number: NET399250UC) was added to the homoge- nates. To break emulsion, samples were centrifuged at 2500 rpm and the organic phase was then removed and dried under a stream of nitrogen. The extracts were resuspended in ethylacetate and fractioned by thin layer chro- matography (TLC) to separate CORT from other products, particularly lip- ids. The region of the gel containing the CORT was scraped and the silica collected and extracted with ethyl acetate. The extract was dried and resus- pended in 0.5 ml phosphate buffered saline containing 0.1 % gelatin (PBS- G). Recoveries ranged between 32% and 78% (average 59%).

I used Sigma-Aldrich antiserum (product nr:C 8784), and Perkin Elmer triated corticosterone (product number: NET399250UC). Standards (0-4000 pg/ml) were run in duplicates. For the assay, 400 μl of sample or standard were incubated in 37 °C for three hours with 50 μl triated corticosterone (10K cpm) and 50 μl diluted antiserum (15 pg). Unbound CORT was sepa- rated by adding 100 μl of dextran-coated charcoal and bound CORT was decanted into scintillation vials. Samples were run in 5 assays for each measuring series (day 2 and 15) and were run blockwise, so that each assay contained an equal number of samples from each treatment and population.

One block was excluded from each measuring series due to very low values

(below 0) for all samples (presumably due to unsuccessful extractions), thus

8 out of the 10 replicates were used for the final analyses.

Results and discussion

Is growth hormone expression correlated with variation in growth rate along a latitudinal gradient in Rana temporaria? (I)

In accordance with previous studies on latitudinal variation in growth in R.

temporaria tadpoles (e.g., Merilä et al. 2000; Lindgren & Laurila 2009), size was positively correlated with latitude. However, contrary to my hypothesis there were no significant latitudinal effects on GH or GHR expression in either measuring occasion (Fig. 5), or no correlations between tadpole size and GH or GHR expression. There were some population differences in GH expression at the later measuring point where one mid-latitude population, Mjösjö (Fig. 5c), had significantly lower GH expression than four other populations having the highest expression. These included both the highest and lowest latitude populations as well as the population at the same latitude as Mjösjö, indicating that the lower GH expression was specific for this population. The body size of Mjösjö tadpoles was consistent with the general clinal pattern in growth. There were no significant predator effects on GH or GHR expression.

This study rejects the hypothesis that there would be a straightforward

latitudinal pattern of consistent differences in GH/GHR expression underly-

ing the differences in growth rates or the predator induced decrease in

growth found in R. temporaria. However as hormone expression was only

measured on two occasions (days 7 and 17), I cannot exclude the possibility

that the fast growing R. temporaria tadpoles had higher or more frequent

peaks of GH expression than slower growing tadpoles, and that the two sam-

pling occasions did not capture the differences in GH expression. Although

the GH axis is the most important regulator of growth in vertebrates, the

growth promoting effects of GH interact with other factors, including insu-

line like growth factor (IGF-I), IGF binding proteins (IGFBP) or prolactin

(PRL) (Moriyama et al. 2000). In the light of the present results, it would be

interesting to investigate whether IGF-I or PRL plays a role in the growth

differences along the latitudinal gradient of R. temporaria tadpoles.

Figure 5. Means (± SE) for GH (a, c) and GHR (b, d) expression for tadpoles from

eight populations along the latitudinal gradient across Sweden. Tadpoles were raised

with a predator from day 1 (black triangles), from day 16 (grey triangles) or without

predator (circles). Measurements (qPCR) were made in tadpoles after 7 (a, b), and

17 (c, d) days of exposure. GH and GHR expression is shown relative to GAPDH

expression.

Geographic variation in corticosterone response to chronic predator stress in tadpoles (II)

On day one of the experiment, predator presence influenced whole body CORT content with tadpoles exposed to predators having higher levels.

However, the effect of predators on tadpole CORT levels was stronger in the low-latitude populations, and also basal CORT levels were higher at lower latitudes (Fig. 6a). At day 15, CORT levels were generally higher and there were no significant differences between predator-exposed and control tad- poles in any of the populations, nor any latitudinal patterns in basal CORT levels (Fig. 6b). However, CORT levels were negatively correlated with body mass both on day one and on day 15. Tadpole body length and devel- opmental stage at the end of the experiment increased with latitude and tad- poles raised with predators were less developed than control tadpoles, a dif- ference that was stronger in the south, results again generally agreeing with what has been previously found in R. temporaria.

Figure 6. Mean (± SE) corticosterone content (ng/g) along the latitudinal gradient in tadpoles raised with (solid line, black triangles) and without (dashed line, open cir- cles) predators after 1 (a) and 15 (b) days of exposure.

Artificially elevated CORT levels resulted in whole body CORT levels that

were in the range of CORT levels that are found naturally in tadpoles ex-

posed to environmental stress (Hayes & Wu 1995; Denver 1998; Glennmeier

& Denver 2001; 2002). Elevated CORT levels strongly decreased growth and developmental rates and the higher concentration also decreased sur- vival, but there were no differences between the northern and the southern population (Fig. 7).

Figure 7. Mean (± SE) larval period (a.), mass at metamorphosis (b.), growth rate (mass at metamorphosis/days until metamorphosis) (c.) and mortality (d.) for tad- poles from a northern (open squares) and mid-southern (black squares) Sweden exposed control treatment, ethanol vehicle control, 100 nM CORT or 500 nM CORT.

The differences in predator induced CORT response on day 1 suggest that

there is genetic variation in predator-induced CORT expression along the

latitudinal gradient, and that the weaker CORT responses in high–latitude

populations may be related to the higher adaptive value of rapid growth in

the time-constrained high-latitude environments. The results from the CORT

manipulation experiment emphasized that sustained elevated CORT levels

can be detrimental to larval amphibians, which may explain why tadpoles do

not exhibit sustained elevated CORT levels when under chronic predator

stress. Down regulating CORT responses may be particularly relevant in

larval amphibians, as they often live in confined ponds and have little possi-

bility to recover between predator threat events. Hence, in contrast to mam- mals which continue to exhibit hormonal stress responses when chronically exposed to predators (Boonstra et al. 1998; Blanchard et al. 1998), tadpoles may have evolved a better capacity to down regulate their hormonal stress response under prolonged predation threat.

Breeding phenology influences anti-predator strategies in tadpoles across a geographic gradient (III)

In accordance with previous studies (Laurila et al. 2008), northern R. tempo- raria larvae were more active than southern ones. However, northern tad- poles showed differences in activity rates between the two years by increas- ing their activity throughout the larval period in 2008 (late breeding). Larvae grown with predators developed deeper tail fins and slightly deeper bodies than larvae reared without predators, however, in 2008 northern tadpoles did not react to predation treatment in terms of morphology (Fig. 8).

North P North NP South P South NP

-0.04 -0.02 0.00 0.02 0.04

-0.05 -0.03 -0.01 0.01 0.03 0.05

RW 2 (21.15%)

RW 1 (52.76%) 2008

2008

2009 2009

Figure 8. Effect of predator environment (P= predator, NP= no predator) and geo- graphic area on R. temporaria tadpole morphology. Values are mean relative warps (RW) scores (± SE) for the first two relative warps (RW1 and RW2). The grid plots are thin-plate spline visualizations obtained from tpsRelw (Rolf 2007) of the range of phenotypes found during the study for each RW axis.

Tadpoles differed in growth and developmental rates across the extremes of

the latitudinal gradient, however, while the southern tadpoles did not differ

between the years in terms of larval period or mass at metamorphosis, north-

ern tadpoles prolonged larval period and metamorphosed at a larger size in

2009 (early breeding) (Fig. 9a,b). Growth rate was generally lower in tad- poles reared with predators and growth rates were similar between the areas in 2008, but in 2009 northern tadpoles grew faster than southern ones (Fig.

9c).

The results for northern tadpoles from a late and an early breeding year support my hypothesis that anti-predator strategies and growth and develop- mental rates in R. temporaria tadpoles are affected by breeding phenology.

To my knowledge, this is the first study to show modification of antipredator

strategies in a vertebrate species in response to natural phenological varia-

tion (but see Altwegg 2002 for an experimental study). In our experiment,

photoperiod remained constant for all the populations and treatments during

the study. The mechanisms by which tadpoles are able to adjust their growth

and developmental rates in response to the strength of time constraints in the

absence of photoperiod cues are currently unknown. Studies on birds have

shown that environmental conditions can affect hormone levels of breeding

females, which affect developmental trajectories of their offspring (reviews

in Gil 2003; Groothuis et al. 2005). It seems possible that similar mecha-

nisms can also operate in amphibians. My results indicate that phenology

can play an important role in shaping life histories and highlight the need of

considering time constraints in studies of predator-induced plasticity.

20 25 30 35 40

P NP

La rva l p e rio d ( d ay s )

2008

P NP

2009

0.4 0.5 0.6 0.7

P NP

Ma s s at m e tam orp hosis (g )

2008

P NP

2009

14.0 16.0 18.0 20.0 22.0

P NP

Grow th rate (m g/d ay )

2008

P NP

2009

North South

a)

b)

c)

Figure 9. Effect of predator environment (P= predator, NP= no predator) and geo-

graphic area on R. temporaria larval traits. a) duration of the larval period, b) mass

at metamorphosis and, c) growth rate. Data are shown as means ± SE for two con-

secutive years.

Compensating for delayed hatching across consecutive life-history stages (IV)

The duration of the embryonic period was strongly affected by the embry- onic treatment. Embryos reared at constant 15 °C hatched the earliest, whereas embryos on daily fluctuating cold (DN) treatment hatched the latest (Fig. 10a). Tadpoles exposed to cold temperatures early in embryonic devel- opment (E4) and those exposed later in development (L4) both hatched at intermediate time (Fig. 14a). At Gosner stage 25, larvae differed in total length, larvae from the DN treatment being the smallest and larvae from E4 and L4 being the largest, whereas larvae in constant 15 °C hatched at inter- mediate length (Fig. 10b). At mid larval development (day 23 of the experi- ment) larval mass was the highest for larvae reared in the DN treatment, the lowest for larvae reared at constant 15 °C, and intermediate in E4 and L4 (Fig. 11a). Total developmental time (from fertilization to metamorphosis) was shorter for larvae reared at constant 15 °C than in the other treatments, but larvae in the constant 15 °C treatment had the longest larval period (Fig.

11c). No differences among treatments were detected for mass or body length at metamorphosis and growth rate during the larval period was sig- nificantly affected by the embryonic treatment where DN larvae grew the fastest and the constant 15 °C treated tadpoles the slowest. None of the postmetamorphic traits examined in this study were influenced by the tem- perature conditions experienced by the embryos.

Figure 10. (a) Duration of the embryonic period, and (b) total larval length at Gos- ner stage 25 in R. arvalis exposed to different embryonic temperature treatments. 15, DN, E4 and L4 represent different embryonic conditions (see main text for treatment details). Data shown as mean ± SE.

6 7 8 9 10 11 12 13 14

15 DN E4 L4

Embryonic period (days)

a)

0.98 0.99 1.00 1.01 1.02 1.03 1.04 1.05 1.06 1.07

15 DN E4 L4

Total length Gosner 25 (cm)

b)

Figure 11. (a) Larval mass at day 23, (b) mass at metamorphosis, (c) duration of the larval period, and (d) growth rate during the larval stage in R. arvalis exposed to different embryonic temperature treatments. 15, DN, E4 and L4 represent different embryonic conditions (see main text for treatment details. Data shown as mean ± SE.

The results indicate strong compensatory responses during the larval stage in late hatching larvae. This is similar to results in insects and birds (De Block

& Stoks 2004; 2005; Benowitz-Fredericks & Kitaysky 2005), but this is the first study revealing compensatory responses to delayed hatching in ecto- thermic vertebrates. These compensatory responses were found in the ab- sence of photoperiod cues indicating that the delay in embryonic develop-

0.50 0.55 0.60 0.65 0.70

15 DN E4 L4

Mass at metamorphosis (g)

b)

0.50 0.55 0.60 0.65 0.70

15 DN E4 L4

Mass at day 23 (g)

a)

13.0 14.0 15.0 16.0 17.0

15 DN E4 L4

Growth rate (mg/day)

d)

36 37 38 39 40 41 42

15 DN E4 L4

Larval period (days)

c)

ment was sufficient to initiate the compensatory response in larval growth and development. No apparent costs of compensatory growth were detected in terms of morphology or locomotor performance at the juvenile stage. A potential cost that I did not measure in this study is higher activity rate (Got- thard 2000; Mangel & Munch 2005). Also long-term costs of compensatory growth responses may be detectable only if individuals are again exposed to stressful conditions, such as food limitation or predation risk (e.g. Dmitriew

& Rowe 2007). I found that compensatory responses can be activated as early as at the embryonic stage and extend over several consecutive life his- tory transitions, mitigating the effects of poor conditions experienced early in development.

Geographic variation in compensatory growth strategies in response to temperature and food stress (V)

The 12-day low food treatment and the 7-day low temperature treatment decreased growth, but both low food and low temperature affected the north- ern populations more than the southern ones in terms of decreased delayed growth (Fig. 12). In both experiments, northern tadpoles were generally more active and predator-exposed tadpoles were less active. Tadpoles were more active during the period of low food level, but not during the days fol- lowing this period (Fig. 13a,b). While low food increased activity similarly in northern and southern tadpoles not exposed to predation risk, activity was increased more in the predator-exposed northern tadpoles in the beginning of the starvation period and less in the end of the period (Fig. 13a,b). Tadpoles exposed to 10 °C were less active than those kept in constant 19 °C, but only southern tadpoles were more active than control tadpoles after the low tem- perature period (Fig. 13c,d).

Low food increased mortality in the presence of predators in the southern but not in the northern populations. The combination of predator treatment and low temperature synergistically decreased survival in both northern and southern tadpoles. Low food and cold-treated tadpoles metamorphosed later than control tadpoles (Fig. 12). Predation risk increased developmental time and also weight at metamorphosis in both experiments (Fig. 12). Southern tadpoles exposed to low food had lower mass at metamorphosis than control tadpoles, whereas no such effect was found in the northern tadpoles (Fig.

12a,b). Low temperature treatment did not affect mass at metamorphosis in

tadpoles from either area (Fig. 12c,d). Northern froglets in low food treat-

ment tended to have longer jumps than ad libitum froglets, whereas southern

low food froglets made shorter jumps than ad libitum froglets. Generally,

predator exposure increased jump length in both experiments, but low tem-

perature decreased jump length in the predator exposed froglets. Low-food

froglets had higher lipid content than ad libitum froglets, and predator-reared froglets had lower lipids in both experiments. Low temperature did not affect lipid content.

Figure 12. Mean mass after low food (day 19, a,b) or low temperature (day 15, c,d),

mass at metamorphosis and development time for populations from northern and

southern Sweden exposed for 12 days to low food (a, b, dashed lines, diamonds), 7

days to 10 °C (c, d, dashed lines, diamonds) or ad libitum food and constant 19 °C

(solid lines, squares). Predator treatments are indicated by solid and non-predator

treatments by open symbols.

Figure 13. Mean proportion of tadpoles active during and after a 12 day period of low food (dashed lines, diamonds), and during and after a 7 day period of low tem- perature (dashed lines, diamonds) of tadpoles from southern (a, c) and northern (b, d) Sweden. Solid lines and squares indicate control treatment (ad libitum food level and constant 19 °C). Predator treatments are indicated by solid and non-predator treatments by open symbols.

These results support earlier findings (Räsänen et al. 2002, Capellan &

Nicieza 2007) that R. temporaria tadpoles can compensate in terms of meta- morphic size after a period of reduced growth, but the response varied be- tween populations from different latitudes and depended on the type of stress inducing low growth. The results from the low food experiment support my prediction that high latitude tadpoles have a stronger compensatory response.

However, contrary to my predictions, I found no apparent cost for the

stronger compensatory responses to low food in northern tadpoles. Indeed,

while low food imposed costs in terms of decreased jumping performance

and decreased survival in southern tadpoles, no such costs were apparent in

northern tadpoles. It is possible that such costs would be apparent later in life

(Morgan & Metcalfe 2001). Tadpoles from northern and southern Sweden

were equally able to compensate after a period of low temperature. Tempera-

ture fluctuations may be a more common stressor than starvation and may explain why tadpoles are generally better at compensating after low tempera- ture.

Tadpoles did not increase activity in the presence of predators after the low

growth periods, and predator presence did not affect the ability to compen-

sate suggesting that increased predator exposure is not a major cost for com-

pensating R. temporaria tadpoles. However, southern tadpoles exposed to

low temperatures increased their activity in the absence of predators. These

results highlight that compensatory strategies are largely context dependent

and that future studies investigating the effects of unfavorable growing con-

ditions should carefully consider the population variation, effects of the type

of stressor as well as interactions between different stressors.

General conclusions

Up until now, no studies had compared endocrine processes along adaptive clines under controlled conditions, and this thesis thus provides a first step for understanding hormonal correlates of geographic life history variation in. Even though I only found a short term elevation in CORT that differed between populations from different latitudes, such difference could poten- tially affect growth and developmental trajectories particularly early in de- velopment (Wada & Breuner 2008). The lower CORT response in early development in high latitude tadpoles could suggest that lower hormonal stress sensitivity may be adaptive in fast-growing and -developing organ- isms, and be potentially important in maintaining their fast life histories.

GH/GHR, though known to be associated to high growth (Schulte et al., 1989, Abrahams & Sutterlin, 1999; Huang & Brown, 2000) did not appear to play a significant role in the latitudinal patterns in growth or predator induced effects in my system, however as the hormonal influences on growth are quite complex it may prove interesting to conduct more exhaus- tive studies on factors associated to the GH axis, perhaps also involving manipulation experiments.

This thesis also lends some interesting conclusions on effects of environ-

mental conditions on latitudinal patterns in life history variation. Northern

tadpoles prolonged developmental time when breeding occurred earlier in

the year, allowing them to invest more in both growth and predator avoid-

ance strategies, eroding the latitudinal differences found in the year with late

breeding in the north. This provides compelling evidence that time stress and

fast development pose costs in terms of decreased predator avoidance strate-

gies and growth, and perhaps more interestingly, that the fast life histories of

high time constrained populations are partly plastic and can adapt to varia-

tion in season length. Under the current scenario of global warming (IPCC

2002), organisms thriving in warmer climates will increase in numbers at

higher latitudes leading to higher predation pressure and potentially posing a

threat to prey species in those areas. As global warming will also lead to

longer growth seasons in the north these results suggest that at least some

prey species in the north may be able to counteract the negative effects of

increased predator densities as longer growth seasons will allow an increased

investment in predator avoidance strategies. This study increases our under-

standing of how climate change will affect high latitude ecosystems, and

calls for studies investigating effects of early phenology in other taxa at high latitudes.

My results also raise questions on mechanisms of which larvae are able to respond to increased and decreased time stress. Eggs and larvae were raised in the lab under controlled conditions and without photoperiod cues, yet in III larvae born late in the season were able to respond with faster develop- ment. In IV, delaying hatching caused larvae to increase their developmental rate. In III where larvae responded to late breeding a plausible explanation could be that that information on the timing of egg spawning is transferred maternally. Such mechanisms have only been explored in birds and suggest that hormonal mechanisms are involved (Gil 2003; Groothuis et al. 2005):

my study emphasizes the need for future studies on such mechanisms in amphibians. In IV, increased growth and development may have been in- duced by the delayed embryonic development itself, perhaps by feedback mechanisms where growth and developmental rates are adjusted to reach a target trajectory (reviewed in Ali et al. 2003). However, this more general view of compensatory growth is complicated by studies finding that organ- isms respond to growth delays with increasing growth when they are caused by some stressors but not others (Capellán & Nicieza 2007), suggesting that compensatory mechanisms may have evolved specifically to certain stress- ors. Paper V, though not explicitly measuring growth, also found support for context-specific compensatory strategies, where compensation (in terms of achieving the same metamorphic size as controls) was more effective after low temperatures than after low food. Low temperature is likely to be a more commonly occurring stressor in larval amphibians, which could explain why both northern and southern tadpoles in my study had adapted compensatory responses to low temperatures. The differences found in compensatory abili- ties in between northern and southern tadpoles in response to low food may indicate that northern tadpoles are superior at handling periods of low growth, which may be related to the greater importance of metamorphosing early and achieving a large size in the north, or that low food is a more commonly occurring stressor in the north. V is the first study to show lati- tude related differences in compensatory strategies in animals with complex life cycles, however further studies investigating the mechanisms underlying the latitudinal differences are needed.

Compensatory growth should entail costs to the compensating individual and prevent the generalization of fast growth rates (Metcalfe & Monaghan 2001). There were no apparent costs of compensatory growth in paper IV, though costs would likely be found if more response variables had been measured. In V, a major cost of growth delay was an increase in larval pe- riod, suggesting that even time constrained high latitude tadpoles were more

“willing” to increase larval periods than metamorphose at a smaller size.

This could mean that a smaller metamorphic size is more detrimental than

late metamorphosis. Other costs associated to delayed growth were only

apparent when the period of delayed growth was combined with predator

stress. A transient period of low temperature appeared to be very detrimental

to tadpoles in terms of decreased survival and locomotor performance, when

they were also stressed by predation risk. As most tadpoles are exposed to

predators during their development periods, low temperatures may be a ma-

jor threat to amphibian larvae. This result further highlights the importance

of context when studying the effects of environmental stressors.

Sammanfattning på svenska

Fysiologisk- och miljömässig påverkan av tillväxtstrategier hos amfibielarver

För de flesta organismer är det fördelaktigt att växa snabbt då en större kroppstorlek är starkt förknippad med högre reproduktivitet och ökad livs- längd. De evolutionära fördelarna med att snabbt växa sig stor borde leda till att alla organismer borde växa så snabbt som möjligt och uppnå en storlek som är fysiologiskt maximal. Trots detta existerar onekligen storleksvaria- tion och variation i tillväxthastighet inom arter vilket borde betyda att det även finns nackdelar med att växa snabbt. Den mest uppenbara nackdelen med att växa snabbt är att man måste vara mer aktiv för få i sig den mängd föda som krävs för att upprätthålla hög tillväxt vilket ökar exponering för predatorer och risken att bli uppäten. Detta leder till en avvägning där förde- larna med att investera i snabb tillväxt ställs mot riskerna att bli uppäten. Hur stora dessa för- och nackdelar är varierar i olika miljöer vilket kan förklara variationen i tillväxthastighet bland populationer inom samma art. Det är vanligt att organismer vid högre latituder och altituder har högre tillväxt- och utvecklingshastighet. Hos växelvarma djur beror detta sannolikt på att de måste utvecklas och bli tillräckligt stora för att klara övervintringen efter en kortare tillväxtsäsong, men förmodligen även på att dessa miljöer känne- tecknas av färre predatorer.

Mekanismerna kring klimatrelaterad variation i tillväxthastighet är relativt outforskade och målet med denna avhandling är att undersöka fysiologiska mekanismer som skulle kunna ligga bakom latitudinella tillväxtskillnader samt hur tillväxtstrategier påverkas av yttre faktorer såsom temperaturför- ändringar, födotillgång och varierande säsongslängd. De globala miljöför- ändringarna som för närvarande pågår kan påverka alla dessa faktorer. Kän- nedom om anpassningar till sådana förändringar är därför ett väldigt relevant forskningsämne.

Vanlig groda (Rana temporaria) är en av Europas vanligaste amfibiearter

och förekommer från Spanien i söder till Beringhavets kust i norr och är

därför en passande art för studier kring miljörelaterade skillnader i tillväxt-

strategier. Arten förökar sig tidigt på våren i varierade typer av sötvattensha-

bitat och de är främst aktiva under tillväxtsäsongen, resten av året övervint-

rar de nedgrävda i marken eller i botten på dammar. Längden på tillväxtsä-

songen (definierat som dagar då medeltemperaturen är över 5

C) minskar

med ökande latitud och i Sverige varierar säsongen från 217 dagar i söder till 117 dagar i norr. Vanlig groda uppvisar en latitudgradient i Sverige där nord- liga populationer har genetiskt betingad högre utvecklings- och tillväxthas- tighet än sydligare populationer men är även mer aktiva vilket troligtvis le- der ökad dödlighet hos nordliga yngel i predatorers närvaro.

I min avhandling studerade jag grodyngel från populationer av vanlig gro- da längs en nord-sydlig gradient i Sverige, från Skåne i söder till Lappland i norr. I en av studierna användes yngel från artificiella parningar av åkergro- da (Rana arvalis). Åkergrodan är en närbesläktad art till vanlig groda och förekommer i liknande habitat. Då jag främst var intresserad av att studera genetiskt betingade skillnader utfördes försöken under kontrollerade klimat- förhållanden i labb. På så sätt kunde eventuella populationsskillnader som berodde på effekter från deras olika miljöer i stort sett uteslutas.

I första delen av min avhandling undersökte jag om det fanns skillnader i hormonnivåer mellan populationer av vanlig groda längs gradienten. Två hormon som verkade särskilt intressanta i sammanhanget var tillväxthormo- net, GH, och stresshormonet, kortikosteron. Höga nivåer av GH är associerat till hög tillväxt, medan förhöjda kortikosteronnivåer kan minska tillväxt.

Båda hormon påverkas också av stress; kortikosteron höjs kraftigt till följd av stress, medan GH nivåer kan minska. De stressrelaterade förändringarna i dessa hormon kan orsaka den långsammare tillväxt som kroniskt stressade djur ofta påvisar. Eftersom tillväxthastigheten ofta påverkas mindre av pre- datorstress hos nordliga yngel än hos sydliga var jag intresserad av att under- söka om detta delvis kan bero på att nordliga yngel har en lägre stressre- spons. En lägre stressrespons i norr borde innebära lägre stressinducerade kortikosteronnivåer och mindre påverkade GH uttryck. Från de mätningar som gjordes verkade det dock inte som att GH uttryck påverkades av varken predatornärvaro eller latitud, detta trots att nordliga grodor växte snabbare.

Detta betyder att andra mekanismer och faktorer som interagerar med GH troligtvis spelar en större roll än GH uttryck i skillnaderna i tillväxt hos yng- el av vanlig groda. Däremot hittade jag skillnader i stressinducerade korti- kosteronnivåer; endast sydliga yngel påvisade en respons efter att ha vistats med predatorer i ett dygn men vid den senare mätningen (efter två veckor av predatorstress) påvisade varken nordliga eller sydliga yngel någon kortikos- teronrespons. Bristen på kortikosteronrespons hos nordliga yngel i början av experimentet kan tyda på att yngel från högre latituder har utvecklat en lägre stressrespons för att undvika stressrelaterade fördröjningar i tillväxt då detta kan vara särkilt skadligt för dessa yngel.

I andra delen av min avhandling undersökte jag hur livshistoriekaraktärer

(larvstadielängd, tillväxthastighet och storlek vid metarmorfos) och strategi-

er för att undvika predation påverkas av miljön hos nordliga och sydliga

grodyngel. I en delstudie undersökte jag hur varierande säsongslängd kan

påverka de skillnader vi vanligtvis ser mellan nordliga och sydliga grodyng-

el. Detta genom att jämföra livshistoriekaraktärer och respons till predatorer