SEARCH

SEARCH BY CITATION

Keywords:

  • behavioural phenotype;
  • cannibalism;
  • Perca fluviatilis ;
  • risk-taking;
  • size-specific predation

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

1. Populations of the same species often display different behaviours, for example, in their response to predators. The question is whether this difference is developed as part of a divergent selection caused by differences in predation pressure, or as a result of phenotypic responses to current environmental conditions.

2. Two populations of Eurasian perch were investigated over a time span of 6 years to see whether risk-taking behaviour in young-of-the-year perch were consistent across cohorts, or if behaviour varied over time with changes in predation regime.

3. Boldness was estimated in aquarium studies by looking at how the fish made trade-offs between foraging in a risky area and staying in shelter. Predation risk of each year and lake was estimated from fishing surveys, using an individual-based model calculating attack rates for cannibalistic perch.

4. The average boldness scores were consistently lower in perch from Fisksjön compared with those in Ängersjön, although the magnitude of the difference varied among years. Variance component analyses showed that differences between lakes in boldness scores only explained 12 per cent of the total variation. Differences between years were contributing at least similarly or more to the total variance, and the variation was higher in Fisksjön than in Ängersjön.

5. The observed risk-taking behaviour of young-of-the-year perch, compared across cohorts, was significantly correlated with the year-specific estimates of cannibalistic attack rates, with lower boldness scores in years with higher predation pressure. In Fisksjön, with significant changes over the years in population structure, the range of both predation risk and boldness scores was wider than in Ängersjön.

6. By following the two perch populations over several years, we have been able to show that the differences in risk-taking behaviour mainly are due to direct phenotypic responses to recent experience of predation risk. Long-term differences in behaviour among perch populations thus reflect consistent differences in predation regime rather than diverging inherent traits.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Predation risk is an important selective agent influencing an animal's trade-off between foraging and hiding (Lima & Dill 1990). A high foraging activity may lead to detection and increased risk of being caught and eaten, while a low activity decreases feeding rate and prolongs the time the individual spends within a predation sensitive size window. Individuals of the same species may differ in their trade-offs between food and safety. This variation can be seen as a shy–bold continuum (Wilson et al. 1994), a component often included in studies of behavioural syndromes and animal personality (Sih et al. 2004; Réale et al. 2007; Toms, Echevarria & Jouandot 2010). A variation in behaviour type is common within animal populations (Sih, Bell & Johnson 2004). Furthermore, average differences in risk-taking behaviour have been found among populations of the same species differing in predation risk (e.g. Magurran et al. 1993; Reale & Festa-Bianchet 2003; Brydges et al. 2008), and the response to predators is sometimes also affected by habitat type (Brydges et al. 2008; Seress et al. 2011), or food availability (Borcherding & Magnhagen 2008; Poulsen et al. 2010).

The degree of boldness may influence the fitness of an individual (Dingemanse & Reale 2005; Wilson, Godin & Ward 2010), and environmental characteristics would determine what behavioural phenotypes that are the most optimal. If boldness is an inherent trait, a divergent selection in environments with different predation pressure is expected. However, in an environment with fluctuating predation rates, the fitness of certain behaviour traits may differ over time (Dall 2004; Dingemanse & Reale 2005). Thus, potential prey animals would benefit from being able to estimate current predation pressure and adjust its behaviour accordingly (Ferrari et al. 2010). Although some components of boldness seem to be heritable (Drent, van Oers & van Noordwijk 2003; Brown, Burgess & Braithwaite 2007; Dingemanse et al. 2009), experience is also an important factor for the development of behaviour phenotypes (e.g. Stamps & Groothuis 2010; Archard & Braithwaite 2011; Hellström & Magnhagen 2011). Accordingly, predation risk is suggested to influence the expression of behaviour traits (Dingemanse et al. 2009).

In this study, we look at long-term consistency of risk-taking behaviour and its connection with predation risk in two populations of Eurasian perch (Perca fluviatilis L.). Trade-offs made between foraging in a risky area and staying in shelter are described for several cohorts of young-of-the-year perch from the lakes Fisksjön and Ängersjön, studied in aquaria over a time span of 6 years (2006–2011, with the exception of 2010). Different aspects of these lakes (life history, morphology, and behaviour) have been studied since 1998, and differences between the two lakes in growth rates, age at maturity and body shape can all be explained as indirect effects of predation pressure (see e.g. Magnhagen & Heibo 2001, 2004; Heibo & Magnhagen 2005; Magnhagen 2006). One of the most severe threats for young perch is that of being cannibalized of one of its larger conspecifics (Treasurer 1989; Persson, Byström & Wahlström 2000). The size relationship between predator and prey determines foraging success of cannibalistic perch (Lundvall et al. 1999; Byström et al. 2003), and the risk of cannibalism depends on the size distribution and density of fish in the population.

The questions we ask here are the following: How consistent are the behavioural phenotypes in a population over time? Are behaviour differences between populations stable over time? Can fluctuations in predation risk be reflected in changing risk-taking behaviour within a population?

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

The study lakes, Fisksjön and Ängersjön, are situated close to the city of Umeå (63°47′N; 20°17′E) in northern Sweden. The two lakes are similar in depth (Ängersjön, 0·9 m mean depth, 3·5 m maximum depth; Fisksjön, 1·9 m, 3·1 m), productivity (Ängersjön, 21 μg L−1 total P; Fisksjön, 18 μg L−1 total P) and amount of littoral vegetation. The fish communities are dominated by perch, pike (Esox lucius), and roach (Rutilus rutilus) in both lakes (Magnhagen & Heibo 2001). Predation risk was estimated from fishing surveys (see below). Behaviour studies for boldness estimates were performed in aquaria, using the same methods during all years, always in September. Some of the data in the current article have also been used in [Magnhagen & Borcherding 2008 (data from 2006); Hellström & Magnhagen 2011 (data from 2007)].

The young-of-the-year perch used in the behaviour studies were caught with a beach seine in the two lakes and transported to Umeå Marine Research Centre, 45 km south of Umeå, where the experiments were performed. Prior to the experiments, the perch were kept in tanks (1 x 1 x 1 m) with continuously running water (17–18 °C). Perch from different lakes were kept separated. They were fed daily with red chironomid larvae. The predators used were older perch, with a body length of 15–24 cm. The young fish from Fisksjön were on average slightly larger than the fish from Ängersjön, and body length varied between years [mean ± SD: 59·7 ± 5·2 mm, = 155 (Fisksjön), 55·7 ± 4·6 mm, = 160 (Ängersjön); F1,305 = 58·0, P < 0·001 (lake); F4,305 = 10·5, P < 0·001 (year); F4,195 = 1·8, P = 0·13 (lake*year)].

Experimental Set-Up

Sixteen experimental aquaria each housed a group of four perch from one of the study lakes. The fish were individually marked with alcian blue on their caudal fin. The aquaria were 170-L (95 × 41 × 44 cm) and had continuously running water (17–18 °C). The light regime was 13L:11D, similar to natural conditions. One-third of each aquarium was used for the predator (a large perch) and the rest for the group of small perch. A plastic net (mesh size 5 mm) was placed between the predator's space and the small perch. During acclimation and between observations, an opaque plastic screen was placed next to the net to prevent the fish habituating to the predator. The water flowed in to the section with the perch group and out through the predator section to minimize olfactory cues before and between observations. The aquaria had gravel on the bottom and artificial vegetation in the predator space and in the third of the space for the perch group that was furthest away from the predator. Before observations started, the perch were acclimated to the aquarium for 3 days and were fed daily with red chironomid larvae in the open area.

The fish were observed for two consecutive rounds, once in the morning and once in the afternoon the same day, to evaluate the behaviour consistency within individuals. Prior to each observation, the opaque screen was moved from its position next to the predator section so that the small perch were enclosed in the half of their section that also contained the vegetation. Chironomid larvae (c. 60–65 larvae) were poured into the aquarium in the open area near the predator section and allowed to sink to the bottom. The opaque screen was then removed, making the large perch visible to the smaller perch through the net and the observations started. The observations lasted for 10 min per aquarium and run. The behaviour of each individual fish was recorded using a computer program, with the choice of three different activities: occurrence in the vegetation, occurrence in the open, and feeding. After each observation, the opaque screen was put back next to the net.

The groups were randomly distributed to the aquaria, and the origin of the group was not known to the observer. All four authors were engaged in the observations during at least 1 year, but no person was doing observations all years.

Boldness Estimates

To get a general measure of boldness, a principal component analyses was performed, using the data for all years on time spent in the open, latency to enter open area, latency to start feeding, and number of changes between the two habitats (open and vegetation) as continuous variables (correlation matrices used). The component explaining most of the variation (PC1) was used as a boldness score for further analyses. To analyse the effect of lake and year on boldness, PC1 was included as response variable in a linear mixed-effects model with lake and year as fixed factors. To avoid pseudoreplication in the analysis, a nested design was created. The repeated measurements within individual were added as random effect at the inner most level. Between individuals, within-groups was added as the next level, and between-groups was added as outer level. As size varied across lakes and years, body length of the tested perch was included as a continuous covariant in the model. Comparisons of the two lakes for each year separately were made by using a priori orthogonal contrasts (Crawley 2007).

To analyse to what extent different factors affected the variance in boldness, another model was performed, using PC1 as response variable and a hierarchical nested design with lake/year/group/individual/run as random effects. From this model, variance component analysis of the random effects was carried out to decompose the variation explained by the different nesting factors (Pinheiro & Bates 2000). Models were also performed separately for each of the two lakes to look at differences in the distribution of variance. Furthermore, best linear unbiased predictors (BLUPs) of the random effect of year, nested within lake, were extracted from the model including both lakes. In a mixed-effects model, fitted values and predictions can be obtained at different levels of nesting (Pinheiro & Bates 2000), and these values thus give estimates of the over-all boldness for each year and population, adjusting for the dependency within lakes.

The mixed-effects analyses were performed using the free software pack R (v 2.11.1) (R Development Core Team 2010) (library nlme v 3.1-102), and Statistica version 10 (StatSoft, Inc. 2011) was used for the PCA.

Predation Risk Estimates

Fishing surveys, using Nordic standard survey nets (Appelberg et al. 1995), were performed in the two lakes in May 2006, 2009 and 2011 and in September 2008, 2009 and 2011 (see Fig. S1, Supporting information). Further, the county administrative board of Västerbotten performed additional surveys in Fisksjön in July 2007 and 2011. The average number and size distribution of fish caught in each net is considered to be proportional to the current population size. Thus, densities and sizes can be compared between lakes and years. We used the results from the surveys to estimate population-specific predation pressure for different sizes of young perch, using an individual-based model calculating attack rates for cannibalistic perch of different sizes (Persson et al. 2004). Predation by pike is not included here, since it would be very small in comparison with cannibalism, because of the much lower density. The estimated daily attack rate for an individual predator depends both on the size of the predator and the size of the victims and indicates the capacity of the predator to attack prey of different sizes. Here, we use the model to calculate, for each perch caught in the survey nets, according to its body length, the number of attacks that this individual would perform daily at different prey sizes (for equation see Appendix S1, Supporting information; see also Persson et al. 2004; Magnhagen 2006). Seven prey length classes were used in the model (10–70 mm total length, TL), corresponding with perch sizes from shortly after hatching and up to the maximum size used in our behaviour study. The sum of the attack rates for all perch caught during one sampling occasion was divided with number of nets used, to get a comparable estimate of predation pressure and thus the potential for predator-prey encounters for each lake and year.

The Effect of Predation Risk on Boldness

The boldness scores specific for each year and lake, estimated as BLUPs for the random factor year nested in lake, were used to test for correlations between behaviour and predation risk, measured as cannibalistic attack rates per CPUE (i.e. per net, see above) for prey sizes 10–70 mm.

In the analyses of the correlations between behaviour and predation risk, the attack rates calculated from the May surveys 2006, 2009 and 2010 were used as estimates for predation risk when the perch were 10–30 mm (i.e. in early summer). For perch of the size range used in the boldness studies (40–60 mm), the risk estimates from September 2008, 2009 and 2011, and the data from Fisksjön July 2007 were used. This way we test whether current risk-taking behaviour reflect past experience of risk, current experience, or both.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Boldness

The principal component analyses gave one dominant component (PC1) explaining 53·4% of the variation (for loadings see Table 1). A high PC1 score indicates that the individual spent a long time in the open area, had a low latency to enter the open and to start feeding. Activity, measured as switches between habitats was also higher than for those with a low PC1 score. Thus, PC1 is giving a good estimate of boldness, with high scores for bold individuals and vice versa.

Table 1. Loadings of different measures of boldness (time in the open, latency to enter open area, latency to start feeding, number of habitat switches), estimated for individual perch in aquarium studies, included in the principal component analyses, for the first principal component (PC1), explaining 53·8% of the variation
LoadingsPC1
  1. Eigenvalue = 2·15.

Open0·76
Latency open−0·87
Latency feeding−0·86
Changes0·24

Perch from Ängersjön showed an overall bolder behaviour than those from Fisksjön (Table 2, Fig. 1). Boldness scores also differed significantly between years. There was no significant interaction between lake and year, with the two lakes varying in the same way across years (Fig. 1). Contrast analyses of differences between the lakes during the separate years showed that PC1 differed significantly between the lakes in 2007 and 2008 (= 0·0001 and 0·003, respectively), marginally in 2006 and 2011 (= 0·09 and 0·07), but did not differ in 2009 (= 0·37). Size did not affect the boldness scores (Table 2).

image

Figure 1. Average boldness score (PC1 ± 95% CI) for young-of-the-year perch in Fisksjön (solid line) and Ängersjön (broken line) tested in aquarium studies during 5 years.

Download figure to PowerPoint

Table 2. Wald statistics for the fixed effects lake and year, and the continuous covariance body length, for PC1 in perch from two populations studied during 5 years, tested in a linear mixed-effects model
 d.f.FP
  1. None of the interaction terms were significant (= 0·22–0·40).

Lake 1·6923·50<0·001
Year 4·6910·43<0·001
Body length 1·2260·770·38

The variance component analyses showed that differences between repeated runs explained almost half of the variation of the boldness scores (Table 3). Individuals within a group showed very little variance, but between-group differences explained 23% of the variation. Variance between years explained 18% and between lakes 12% (Table 3). In the separate models from the two lakes, the variance among the perch from Fisksjön was explained to a higher degree by year (23%) compared with those from Ängersjön (14%) (Table 3).

Table 3. The percentage of total variance of PC1 explained by the nesting factors within individual, between individuals, between-groups, between years, between lakes, and the residual variance of the random effects of linear mixed-effects models, with both lakes included or lakes tested separately
 Between lakesBetween yearsBetween-groupsBetween individualsWithin individualResidual
All12·218·322·6<0·00146·8<0·001
Fisksjön 23·218·7<0·00158·0<0·001
Ängersjön 13·944·80·0837·70·03

Predation Risk

The risk of cannibalism peaks in both lakes at a prey size of around 20–30 mm and falls rapidly with increasing size. In most years, the risk in Fisksjön was higher than in Ängersjön, across most size classes, but in 2009 the situation was the reversed (Fig. 2). The relative attack rates show similar patterns across years in Ängersjön, but in Fisksjön a higher variation is found, with minimum estimated risk in 2009. The graphs in (Fig. 2) show data from different months, because of differences in fishing dates. We compared the repeated fishing surveys done in 2009 and 2011 to see whether catch rates are comparable in samples from different months (Table S1, Supporting information). In Fisksjön, May surveys gave a higher yield than those in September, but the July survey in 2011 was similar to the September survey. Also the estimated attack rates were more similar between July and September than between May and September (Fig. S2, Supporting information). In Ängersjön, the comparisons were inconclusive.

image

Figure 2. Relative cannibalistic attack rates on different lengths (mm) of perch estimated from an individual-based model (Persson et al. 2004), using cannibal size distribution and abundance in fishing surveys carried out 2006–2011 in Fisksjön (solid line) and Ängersjön (broken line, not sampled 2007). Attack rates are presented as proportions of the highest estimate.

Download figure to PowerPoint

Correlation between Boldness and Predation Risk

The adjusted boldness scores (BLUPs), extracted for each year and lake, were not significantly correlated with the estimated attack rates in May at prey sizes 10–30 mm (but with a tendency for a negative correlation at prey size 30 mm) (Table 4). However, the estimated attack rates in September were negatively correlated at prey sizes between 40 and 60 mm (and with a nearly significant correlation for prey size 70 mm) (Table 4, see also Fig. 3 for an example).

image

Figure 3. Correlation between boldness of young-of-the-year perch in aquarium studies, and lake-/year-specific predation pressure on 50 mm long perch in the lakes Fisksjön (squares) and Ängersjön (circles) (rs = −0·93, = 0·0025). Boldness is estimated as best linear predicted values (BLUPs) of the boldness scores (PC1), adjusted for lake effects in a linear mixed-effects model. Size-specific cannibalistic attack rates were estimated from cannibal size distribution and abundance in fishing surveys carried out 2006–2011, using an individual-based model (see Persson et al. 2004).

Download figure to PowerPoint

Table 4. Correlations (Spearman rank) between population boldness score and estimated predation risk at different prey lengths (mm). Boldness scores were estimated as BLUPs for each year and lake (year nested in lake) from a linear mixed-effects model with PC1 as response variable. Predation risks for the prey of different sizes were estimated as cannibal attack rates per CPUE (per net in the test fishing). The May surveys were performed in 2006, 2009 and 2011. The September surveys are from 2008, 2009, 2011 and also include a survey from Fisksjön in July 2007
Survey monthPrey sizenrstP
May1060·03 0·060·957
May206−0·09−0·170·872
May306−0·77−2·420·072
September407−0·93−5·590·0025
September507−0·93−5·590·0025
September607−0·86−3·720·014
September707−0·75−2·530·052

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

The average boldness scores were consistently lower in perch from Fisksjön compared with those in Ängersjön, although the magnitude of the difference varied among years, with the most similar behaviour in 2009. This overall difference in boldness between Fisksjön and Ängersjön was also supported by an earlier study using comparable methods (Magnhagen 2006). These results may suggest that the results are owing to diverging selection in the two perch populations, caused by consistent differences in predation pressure. However, according to the variance components of the nested random factors, the behaviour differences between the lakes explain only 12 per cent of the total variance of the boldness scores. The variation within lakes among years is contributing at least similarly or more to the total variance, and the variation was higher in Fisksjön than in Ängersjön. It is thus important to look at how changes in risk over time influence the trade-off between foraging and antipredator behaviour within a prey population. Common garden designs can reveal the effect of predator exposure on the behaviour of potential prey (Riesch et al. 2009; Hellström & Magnhagen 2011), but the current study is unique in its long-term approach, covering behaviour variation and its connection to population dynamics over several years. Here, risk-taking behaviour of young-of-the-year perch, compared across cohorts, corresponds well with the year-specific estimates of cannibalistic attack rates. Our results corresponded well with a field study that concluded that the timing of habitat shifts of young perch was affected by densities of cannibals that varied between lakes and years (Byström et al. 2003). Another study on whole-lake systems instead found that food availability affected risk-taking behaviour in young rainbow trout (Oncorhynchus mykiss) under the risk of cannibalism (Biro, Post & Parkinson 2003). The young-of-the-year perch were always larger in Fisksjön compared with Ängersjön, suggesting a higher food availability, but no connection with size and boldness were found in the aquarium studies.

In Fisksjön, the range of both predation risk and boldness scores over the years was wider than in Ängersjön. In Ängersjön, the annual catch rates in the fishing surveys were less variable than in Fisksjön where density and size distribution of the perch population changed considerably. In May 1999, Fisksjön contained a typical stunted perch population with a very high density and an average body length of 10 cm (Magnhagen & Heibo 2001). The density has then decreased and the mean size increased over the years with a minimum density in 2009. This is consistent with Persson et al. (2003) who suggested that an interaction between size-dependent cannibalism and intercohort competition in perch leads to shifts between periods when the population is dominated by stunted cannibals and when it is dominated by recruits and larger cannibals. In Ängersjön, the lack of these shifts may be explained by a higher density of large pike (E. lucius), keeping the density of perch cannibals, and thus number of reproducing perch, at a lower level (Magnhagen & Heibo 2001). The pike would be less of a threat for the young-of-the-year perch than the larger cannibalistic perch because of a lower density and lower habitat overlap (Eklöv & Diehl 1994).

This study focussed on general differences between populations rather than on individual personality. Still, the mixed-effects models considered data for each individual, and we can thus draw some conclusions on individual behaviour. Even within a certain predator regime, as estimated for each lake and year, the behaviour of the young perch can be quite variable. The variance component analyses of the random effects on the boldness scores show that a high proportion of the total variance is explained by differences between runs within individuals. This is probably due to habituation to the experimental set-up leading to changes in behaviour with time (see Oosten, Magnhagen & Hemelrijk 2010). Furthermore, individuals within-groups show very similar behaviour and variation between-groups within year and lake is higher than variation within-groups. This is typical for group-living animals that often make consensus decisions (Conradt & Roper 2005; Magnhagen 2012), as in the stickleback that uses public information for foraging and movement decisions (Coolen et al. 2005; Ward et al. 2008).

With the growing number of studies on animal personality and behaviour syndromes, it is clear that behaviour indeed can be a product of natural selection, and it has also often been connected with physiological characteristics (Dingemanse & Reale 2005). Inherited behaviour patterns, compared between individuals or between selected strains, has been found in animals from different taxa, as birds (e.g. great tit, Parus major, van Oers et al. 2004; Alpine swift, Apus melba, Bize, Diaz & Lindström 2012), mammals (house mouse, Mus musculus, Sluyter et al. 1995), and fish (rainbow trout, O. mykiss, Øverli et al. 2002). Lately, though, there has been an increasing interest in the interplay between genetic traits and the plasticity in phenotypic expression of behaviour (Dingemanse et al. 2010). Behaviour is most likely shaped by a combination of innate and environmental factors. The results of the present study suggest that risk-taking behaviour of individual perch is more a response to the surroundings, both in terms of predation regime and social context, than caused by inherited traits. However, that individual boldness has some genetic component cannot be completely ruled out. In several fish species, antipredator behaviour is shown to be inherited, but can still be affected by experience (e.g. European minnow, Phoxinus phoxinus, Magurran 1990; guppy, Poecilia reticulata, Kelley & Magurran 2003; Panamanian bishop, Brachyraphis episcopi, Brown, Burgess & Braithwaite 2007). In the Alpine swift, though, a cross-fostering experiment showed that antipredator behaviour was basically due to genetics rather than to social learning (Bize, Diaz & Lindström 2012). In perch, a common garden experiment, raising perch from the two lakes Fisksjön and Ängersjön in a predator free environment, showed, however, that boldness seemed mainly to be shaped by experience (Hellström & Magnhagen 2011). Perch from both lakes, reared in ponds without predators, were as bold as the wild ones from Ängersjön, while the wild perch from Fisksjön were significantly shyer than all the other types. That study could not show at what time during the ontogeny the behaviour is formed because the perch raised in ponds never experienced any risk (Hellström & Magnhagen 2011). The vulnerability of young perch changes over time and with growth, because piscivores often are gape-limited and predation risk is dependent on the size relationship between predator and prey (Lundvall et al. 1999). Over the growing season, this relationship changes in a complex manner, because both prey and cannibals are increasing in size. In the present study, boldness scores of the tested perch were not correlated with estimated predation risk in the beginning of their life, but only to recent experience of risk. This suggests that risk-taking behaviour must be able to change with changing circumstances. Another study supporting this was comparing young-of-the-year and 1-year-old perch from our two study lakes (Magnhagen & Borcherding 2008). The risk-taking behaviour differed between the two age classes, with the perch from Fisksjön getting bolder with age, and the ones from Ängersjön getting shyer. In both lakes, these results corresponded to changes in predation risk with increasing size (Magnhagen & Borcherding 2008). The study also showed that differences between the lakes did not simply depend on a selective consumption of bold phenotypes at high predator densities but is caused by a flexibility of behaviour in both lakes.

Finally, in the literature, it is possible to find examples of behaviour being mainly influenced by genes, by experience, or by a combination of both. Obviously, species differ in the way genes and environment interact to shape the behaviour phenotype. In a fluctuating environment, the fitness of different fixed phenotypes within a population may vary between years, thus maintaining the genetic variation of behaviour in some species (Dingemanse et al. 2004). In other species, behavioural reaction norms are found, with individual plastic responses to the changing environment (reviewed by Dingemanse et al. 2010). By following the two perch populations over several years, we have been able to show that the consistent differences in risk-taking behaviour mainly were due to direct phenotypic responses to recent experience of predation risk. Variation in predator abundance over years also led to a corresponding change in risk-taking behaviour of potential prey. Differences in behaviour among perch populations clearly reflected consistent differences in predation regime rather than diverging inherent traits. In conclusion, long-term studies are valuable and should be used more often to understand the development of antipredator behaviour in fluctuating environments.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

We thank all the field assistants that helped us with the fish collection and maintenance over the years. We also thank Hasse Fängstam at Västerbotten County Administration Board for the test fishing data from Fisksjön. The study was financially supported by the Swedish Research Council Formas, Carl Tryggers Foundation for Scientific Research, C.F. Lundström Foundation, and German Research Foundation. The experiments in this study comply with the current laws of Sweden and were approved by the local ethics committee of the Swedish National Board for Laboratory Animals (CFN, licence no A94-06 and A94-11).

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
  • Appelberg, M., Berger, H.M., Hesthagen, T., Kleive, E., Kurkilahti, M., Raitaniemi, J. & Rask, M. (1995) Development and intercalibration of methods in Nordic freshwater fish monitoring. Water, air and soil pollution, 85, 401406.
  • Archard, G.A. & Braithwaite, V.A. (2011) Increased exposure to predators increases both exploration and activity level in Brachyrhaphis episcopi. Journal of Fish Biology, 78, 593601.
  • Biro, P.A., Post, J.R. & Parkinson, E.A. (2003) From individuals to populations: Prey fish risk-taking mediates mortality in whole-system experiments. Ecology, 84, 24192431.
  • Bize, P., Diaz, C. & Lindström, J. (2012) Experimental evidence that adult antipredator behaviour is heritable and not influenced by behavioural copying in a wild bird. Proceedings of the Royal Society B: Biological Sciences, 279, 13801388.
  • Borcherding, J. & Magnhagen, C. (2008) Food abundance affects both morphology and behaviour of juvenile perch. Ecology of Freshwater Fish, 17, 207218.
  • Brown, C., Burgess, F. & Braithwaite, V.A. (2007) Heritable and experiential effects on boldness in a tropical poeciliid. Behavioral Ecology and Sociobiology, 62, 237243.
  • Brydges, N.M., Colegrave, N., Heathcote, R.J.P. & Braithwaite, V.A. (2008) Habitat stability and predation pressure affect temperament behaviours in populations of three-spined sticklebacks. Journal of Animal Ecology, 77, 229235.
  • Byström, P., Persson, L., Wahlström, E. & Westman, E. (2003) Size- and density-dependent habitat use in predators: consequences for habitat shifts in young fish. Journal of Animal Ecology, 72, 156168.
  • Conradt, L. & Roper, T.J. (2005) Consensus decision making in animals. Trends in Ecology & Evolution, 20, 449456.
  • Coolen, I., Ward, A.J.W., Hart, P.J.B. & Laland, K.N. (2005) Foraging nine-spined sticklebacks prefer to rely on public information over simpler social cues. Behavioral Ecology, 16, 865870.
  • Crawley, M.J. (2007) The R Book. John Wiley & Sons Ltd, Chichester.
  • Dall, S.R.X. (2004) Behavioural biology: fortune favours bold and shy personalities. Current Biology, 14, R470R472.
  • Dingemanse, N.J. & Reale, D. (2005) Natural selection and animal personality. Behaviour, 142, 11591184.
  • Dingemanse, N.J., Both, C., Drent, P.J. & Tinbergen, J.M. (2004) Fitness consequences of avian personalities in a fluctuating environment. Proceedings of the Royal Society of London Series B-Biological Sciences, 271, 847852.
  • Dingemanse, N.J., Van der Plas, F., Wright, J., Réale, D., Schrama, M., Roff, D.A., Van der Zee, E. & Barber, I. (2009) Individual experience and evolutionary history of predation affect expression of heritable variation in fish personality and morphology. Proceedings of the Royal Society of London Series B-Biological Sciences, 276, 12851293.
  • Dingemanse, N.J., Kazem, A.J.N., Reale, D. & Wright, J. (2010) Behavioural reaction norms: animal personality meets individual plasticity. Trends in Ecology & Evolution, 25, 8189.
  • Drent, P.J., van Oers, K. & van Noordwijk, A.J. (2003) Realized heritability of personalities in the great tit (Parus major). Proceedings of the Royal Society of London Series B-Biological Sciences, 270, 4551.
  • Eklöv, P. & Diehl, S. (1994) Piscivore efficiency and refuging prey – the importance of predator search mode. Oecologia, 98, 344353.
  • Ferrari, M.C.O., Elvidge, C.K., Jackson, C.D., Chivers, D.P. & Brown, G.E. (2010) The responses of prey fish to temporal variation in predation risk: sensory habituation or risk assessment? Behavioral Ecology, 21, 532536.
  • Heibo, E. & Magnhagen, C. (2005) Variation in age and size at maturity in perch (Perca fluviatilis L.), compared across lakes with different predation risk. Ecology of Freshwater Fish, 14, 344351.
  • Hellström, G. & Magnhagen, C. (2011) The influence of experience on risk taking: results from a common-garden experiment on populations of Eurasian perch. Behavioral Ecology and Sociobiology, 65, 19171926.
  • Kelley, J.L. & Magurran, A.E. (2003) Learned predator recognition and antipredator responses in fishes. Fish and Fisheries, 4, 216226.
  • Lima, S.L. & Dill, L.M. (1990) Behavioral decisions made under the risk of predation: a review and prospectus. Canadian Journal of Zoology, 68, 619640.
  • Lundvall, D., Svanbäck, R., Persson, L. & Byström, P. (1999) Size-dependent predation in piscivores: interactions between predator foraging and prey avoidance abilities. Canadian Journal of Fisheries and Aquatic Sciences, 56, 12851292.
  • Magnhagen, C. (2006) Risk-taking behaviour in foraging young-of-the-year perch varies with population size structure. Oecologia, 147, 734743.
  • Magnhagen, C. (2012) Personalities in a crowd – what shapes the behaviour of Eurasian perch and other shoaling fishes? Current Zoology, 58, 3241.
  • Magnhagen, C. & Borcherding, J. (2008) Risk-taking behaviour in foraging perch: does predation pressure influence age-specific boldness? Animal Behaviour, 75, 509517.
  • Magnhagen, C. & Heibo, E. (2001) Gape size allometry in pike reflects variation between lakes in prey availability and relative body depth. Functional Ecology, 15, 754762.
  • Magnhagen, C. & Heibo, E. (2004) Growth in length and in body depth in young-of-the-year Eurasian perch with different predation risk. Journal of Fish Biology, 64, 612624.
  • Magurran, A.E. (1990) The inheritance and development of minnow antipredator behaviour. Animal Behaviour, 39, 834842.
  • Magurran, A.E., Seghers, B.H., Carvalho, G.R. & Shaw, P.W. (1993) Evolution of adaptive variation in antipredator behavior. Marine Behaviour and Physiology, 23, 2944.
  • van Oers, K., Drent, P.J., de Goede, P. & van Noordwijk, A.J. (2004) Realized heritability and repeatability of risk-taking behaviour in relation to avian personalities. Proceedings of the Royal Society of London Series B-Biological Sciences, 271, 6573.
  • Oosten, J.E., Magnhagen, C. & Hemelrijk, C.K. (2010) Boldness by habituation and social interactions: a model. Behavioral Ecology and Sociobiology, 64, 793802.
  • Øverli, Ø., Pottinger, T.G., Carrick, T.R., Øverli, E. & Winberg, S. (2002) Differences in behaviour between rainbow trout selected for high- and low-stress responsiveness. Journal of Experimental Biology, 205, 391395.
  • Persson, L., Byström, P. & Wahlström, E. (2000) Cannibalism and competition in Eurasian perch: population dynamics of an ontogenetic omnivore. Ecology, 81, 10581071.
  • Persson, L., De Roos, A.M., Claessen, D., Byström, P., Lövgren, J., Sjögren, S., Svanbäck, R., Wahlström, E. & Westman, E. (2003) Gigantic cannibals driving a whole-lake trophic cascade. Proceedings of the National Academy of Sciences of the United States of America, 100, 40354039.
  • Persson, L., Claessen, D., DeRoos, A.M., Byström, P., Sjögren, S., Svanbäck, R., Wahlström, E. & Westman, E. (2004) Cannibalism in a size-structured population: energy extraction and control. Ecological Monographs, 74, 135157.
  • Pinheiro, J. & Bates, D. (2000) Mixed-Effects Models in S and S-Plus. Springer-Verlag, New York.
  • Poulsen, S.B., Svendsen, J.C., Jensen, L.F., Schulz, C., Jager-Kleinicke, T. & Schwarten, H. (2010) Effects of food deprivation on refuge use and dispersal in juvenile North Sea houting Coregonus oxyrinchus under experimental conditions. Journal of Fish Biology, 77, 17021708.
  • R Development Core Team (2010). R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, Available at: http://www.R-project.org.
  • Reale, D. & Festa-Bianchet, M. (2003) Predator-induced natural selection on temperament in bighorn ewes. Animal Behaviour, 65, 463470.
  • Réale, D., Reader, S.M., Sol, D., McDougall, P.T. & Dingemanse, N.J. (2007) Integrating animal temperament within ecology and evolution. Biological Reviews, 82, 291318.
  • Riesch, R., Duwe, V., Herrmann, N., Padur, L., Ramm, A., Scharnweber, K., Schulte, M., Schulz-Mirbach, T., Ziege, M. & Plath, M. (2009) Variation along the shy-bold continuum in extremophile fishes (Poecilia mexicana, Poecilia sulphuraria). Behavioral Ecology and Sociobiology, 63, 15151526.
  • Seress, G., Bokony, V., Heszberger, J. & Liker, A. (2011) Response to predation risk in urban and rural house sparrows. Ethology, 117, 896907.
  • Sih, A., Bell, A. & Johnson, J.C. (2004) Behavioral syndromes: an ecological and evolutionary overview. Trends in Ecology & Evolution, 19, 372378.
  • Sih, A., Bell, A.M., Johnson, J.C. & Ziemba, R.E. (2004) Behavioral syndromes: an integrative overview. Quarterly Review of Biology, 79, 241277.
  • Sluyter, F., Bult, A., Lynch, C.B., Vanoortmerssen, G.A. & Koolhaas, J.M. (1995) A comparison between house mouse lines selected for attack latency or nest-building – evidence for a genetic basis of alternative behavioral strategies. Behavior Genetics, 25, 247252.
  • Stamps, J. & Groothuis, T.G.G. (2010) The development of animal personality: relevance, concepts and perspectives. Biological Reviews, 85, 301325.
  • StatSoft, Inc. (2011) STATISTICA (data analysis software system), version 10. www.statsoft.com.
  • Toms, C.N., Echevarria, D.J. & Jouandot, D.J. (2010) A methodological review of personality-related studies in fish: focus on the shy-bold axis of behavior. International Journal of Comparative Psychology, 23, 125.
  • Treasurer, J.W. (1989) Mortality and production of 0 +  perch, Perca fluviatilis L. in two Scottish lakes. Journal of Fish Biology, 34, 913928.
  • Ward, A.J.W., Sumpter, D.J.T., Couzin, L.D., Hart, P.J.B. & Krause, J. (2008) Quorum decision-making facilitates information transfer in fish shoals. Proceedings of the National Academy of Sciences of the United States of America, 105, 69486953.
  • Wilson, D.S., Clark, A.B., Coleman, K. & Dearstyne, T. (1994) Shyness and boldness in humans and other animals. Trends in Ecology and Evolution, 9, 442446.
  • Wilson, A.D.M., Godin, J.G.J. & Ward, A.J.W. (2010) Boldness and reproductive fitness correlates in the Eastern mosquitofish, Gambusia holbrooki. Ethology, 116, 96104.

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

FilenameFormatSizeDescription
jane2007-sup-0001-AppendixS1.docxWord document26KAppendix S1. Equations for calculation of cannibalistic attack rates
jane2007-sup-0002-FigS1.docWord document1081KFig. S1. Size distribution of perch caught in survey nets in (a) Fisksjön and (b) Ängersjön during the years 2006–2011 (with the exception of 2010).
jane2007-sup-0003-FigS2.docWord document136KFig. S2. Graphic comparisons of attack rates in Fisksjön and Ängersjön estimated from fishing surveys May and September 2009 and 2011, and for Fisksjön, July 2011.
jane2007-sup-0004-TableS1.docWord document61KTable S1. Data from fishing surveys in the study lakes 2006–2011.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.