Fixed behavioural plasticity in response to predation risk in the three-spined stickleback

Animal Behaviour 112 (2016) 147e152

Contents lists available at ScienceDirect

Animal Behaviour journal homepage: www.elsevier.com/locate/anbehav

Fixed behavioural plasticity in response to predation risk in the three-spined stickleback Sin-Yeon Kim* Departamento de Ecoloxía e Bioloxía Animal, Universidade de Vigo, Vigo, Spain

a r t i c l e i n f o Article history: Received 7 July 2015 Initial acceptance 4 August 2015 Final acceptance 26 October 2015 Available online MS. number: 15-00587R Keywords: behavioural plasticity individual variation personality predator-prey interaction stickleback

I experimentally tested the repeatability and plasticity of two antipredator behaviours, shoaling and risk taking, in a sample of 443 juvenile three-spined sticklebacks, Gasterosteus aculeatus. I quantified between-individual variation in these behaviours as well as behavioural changes over time in two groups of sticklebacks that were either exposed or not exposed to simulated predation pressure. Shoaling and risk taking were repeatable within individuals in both experimental and control fish. Individual willingness to shoal increased over time in both experimental and control groups, but there was no evidence that shoaling changed in response to predation risk. Risk taking also showed temporal changes: sticklebacks exposed to simulated predation risk became increasingly fearful, unlike the control fish, suggesting that this behaviour is plastic. There was, however, no evidence of between-individual variation in the behavioural changes over time in either the control or experimental condition, suggesting that behavioural plasticity is a fixed response in the individuals of this population. © 2015 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.

Phenotypic plasticity is the ability of a single genotype to produce more than one phenotype in response to environmental conditions (Pigliucci, 2001; Scheiner, 1993). Behavioural plasticity, in particular, may be linked to fitness, for example by adjusting foraging behaviour to the level of predation risk or competition or by changing mating behaviour according to sociosexual environments (Bell & Sih, 2007; Han & Brooks, 2013; Krebs & Davies, 1997; Laskowski & Bell, 2013). Recent studies have shown that individuals from the same population may differ in the level of behavioural plasticity (Dingemanse & Wolf, 2013). Withinpopulation variation in behavioural plasticity can have important consequences for animal populations by increasing or decreasing individual differences in behavioural strategies or by affecting the consistency of the behaviours in different environmental contexts (Nussey, Wilson, & Brommer, 2007). Predators play an important role in the evolutionary process of shaping behavioural patterns. Predator-mediated directional selection on behaviour (Bell & Sih, 2007; Huntingford, Wright, & Tierney, 1994; Lima, 1998; Wolf, van Doorn, Leimar, & Weissing, 2007) can reduce behavioural variation in a population under a constant level of predation pressure. On the other hand, short-term

* Correspondence: S.-Y. Kim, Departamento de Ecoloxía e Bioloxía Animal, Universidade de Vigo, Vigo, 36310, Spain. E-mail address: [email protected].

effects of predation pressure on the phenotypic expression of behaviours (Dingemanse, Barber, Wright, & Brommer, 2012; Relyea, 2005) may be important for the viability of individuals in a heterogeneous environment in which predation risk varies over time (Pigliucci, 2001; Roff, 1997). Adaptive plasticity in antipredator behaviour could reduce mortality when predators are present, but maximize fitness gains via increased feeding rate in other circumstances (Luttbeg & Sih, 2010; Stamps, 2007). To improve our understanding of how predation pressure influences prey behaviour, it is necessary to test whether individuals vary in the pattern of behavioural change in response to predation risk (Dingemanse et al., 2012). The behaviour of individuals can vary in multiple ways as a function of personality and plasticity. For instance, the average level of antipredator behaviour, which represents personality, may vary between individuals (I: individual variation). Individuals may also vary in the level of their environment-behaviour gradient representing behavioural plasticity (I)E: individual)environment interaction). If a simple behavioural rule is favoured according to predation risk (Houston & McNamara, 1999), selection should erode withinpopulation variation in behavioural plasticity (Dingemanse, ale, & Wright, 2010). Individual and genetic variation in Kazem, Re behavioural plasticity can be maintained if selection depends on the frequency of different types of behaviour within a population (Wolf, van Doorn, & Weissing, 2011) or if selection fluctuates (Sasaki & Ellner, 1997). The variation may also be maintained when

http://dx.doi.org/10.1016/j.anbehav.2015.12.004 0003-3472/© 2015 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.

148

S.-Y. Kim / Animal Behaviour 112 (2016) 147e152

the behavioural plasticity reflects alternative strategies with comparable average performance over evolutionary time (Oliveira, Taborsky, & Brockmann, 2008). I experimentally tested behavioural plasticity in response to predation risk in the three-spined stickleback, Gasterosteus aculeatus. Juvenile sticklebacks used in this study were born in captivity, but originated from a natural population preyed on by piscivores. Predation pressure could thus have shaped the behavioural plasticity in this population during their evolutionary history. I studied the expression of behavioural phenotypes in full-sib and half-sib families in different environments. Juvenile fish from different genetic families were exposed to two environments that differed in predation risk. Experimental fish were exposed to simulated predation risk and control fish were kept in the predatorabsent condition. During the experiment, changes over time in social responses to conspecifics (i.e. shoaling) and willingness to take risks for foraging (i.e. risk taking) were observed. By comparing behavioural changes over time in the control and experimental fish, I tested whether these behaviours are plastic in relation to predation risk. I also tested whether the behavioural plasticity varies between individuals by testing behavioural reaction norms based on temporal changes in behaviour within each treatment. METHODS Study Population and Breeding Design Sexually immature three-spined sticklebacks were captured with hand nets from the Rio Ulla, Galicia (Spain), in February 2013 (for a map see the Supplementary material). Once mature, 16 males and 16 females (of the 70 fish originally collected) were used for breeding. The breeding design and fish husbandry of adults and juveniles are fully described in a previous paper (Kim & Velando, 2015). Each fish bred twice with two different mates, producing 32 full-sib families of the F1 generation, during AprileMay 2013. Thus, each full-sib family had a maternal and a paternal half-sib family. At age 40 days, fry from each full-sib family were divided among two (N ¼ 7 families) or four (N ¼ 25 families) ‘growth tanks’ (N ¼ 114 tanks; 24  16.5 cm and 17.5 cm high), depending on the brood size. Each tank housed 11 or 12 juvenile fish. The tanks were connected to closed water systems equipped with the combined continuous function of a mechanical filter, a circulation pump and a flow-through water-cooling device. Juvenile fish were fed to satiation twice daily until 5 months old then once a day. They were fed on a progressive diet of newly hatched Artemia from hatching to 3 months old and a commercial pelleted diet (Gemma Micro, Skretting, Norway) from 2 months old onwards. The natural photoperiod was simulated by programmed illumination. Experimental Protocol The experiment was carried out during SeptembereNovember with 448 juvenile sticklebacks around 5 months old (143e160 days) from 31 full-sib families (and 112 different growth tanks) in seven weekly experimental sessions. One family from the stock was excluded from this experiment to match the number of fish across all different experimental tanks and weeks. Age effects on behaviours were not significant in preliminary analyses; therefore age was not included in further analyses. Prior to each weekly experimental session, I created four experimental and four control tanks (33  18 cm and 19 cm high). Each tank contained eight sticklebacks from four or five different full-sib families (see also Kim & Velando, 2015). Four individuals were randomly selected from each growth tank; two individuals were then allocated to two

experimental tanks and the other two to two control tanks. Before allocation to a tank, individuals were weighed and permanently marked with colour elastomer tags (Northwest Marine Technologies, Shaw Island, WA, U.S.A.) under a low dose of benzocaine anaesthetic. Each individual was marked with a coloured tag on either the anterior or posterior dorsal of both lateral sides to allow rapid identification of the eight different individuals in the same tank. Body weight did not differ between the experimental and control groups (mean ± SE; experimental: 0.307 ± 0.005 g, N ¼ 224; control: 0.308 ± 0.005 g, N ¼ 224; Student's t test: t446 ¼ 0.086, P ¼ 0.931). Each experimental tank contained a sponge filter, an artificial plant and a transparent food cup to which bloodworms were added as food once a day. The front wall of the tanks was transparent to enable observation. The other walls were opaque. Large opaque dividers were inserted between the tanks to prevent interference from different experimental treatments. During the acclimatization period of 6 days in the experimental tank, the fish were accustomed to feeding on bloodworms from the food cup. After acclimatization, two different behaviours were recorded in all fish (day 0; shoaling was recorded between 0900 and 1200 hours and risk taking at 1500e1700 hours); the experimental treatment began immediately after the behavioural tests (the first predator attack simulation was performed at 1800 hours on day 0). The chemical and visual simulation treatments consisted of adding 20 ml of water from an aquarium holding brown trout, Salmo trutta, before introducing a model trout (13 cm long) into the tank and chasing the sticklebacks for 10 min with this model. I ensured that all the sticklebacks in the tank were chased during each treatment. The control tanks were treated by adding the same amount of clean water and omitting the visual stimulus. The treatments were executed repeatedly at randomly chosen times of day between 0900 and 1800 hours. Each experimental tank was subjected to 12 treatments (120 min) throughout days 0e4. One treatment on day 0, four treatments each on day 1 and day 2, two treatments on day 3 and one treatment on day 4 were scheduled. Behavioural Observations Behavioural observations were made repeatedly for all individual sticklebacks before the treatment began (day 0), during the treatment (day 3) and after the treatment (day 4). Sticklebacks that died during the experiment were excluded from the analyses (experimental: N ¼ 3 individuals; control: N ¼ 2 individuals). A total of 2658 observations made on 443 individuals (two behaviours  three repeated measures) were used for statistical analyses. The tests for shoaling and risk taking are fully described in a previous paper (Kim & Velando, 2015). In summary, shoaling was tested for each individual in an observation tank, which contained three unfamiliar conspecifics of similar size. A focal fish was allowed to swim between the acclimatization and conspecific zones (16 cm distance); the time taken to reach the conspecific zone was measured up to 180 s. This test assesses the individual's willingness to join the conspecific group. Individuals were returned to their experimental tanks after this test. At least 3 h after the shoaling test, I assessed individual willingness to forage under predation risk simulated by a model avian predator (the grey heron, Ardea cinerea) in the experimental tanks (see also Bell, 2005). I attached a dummy head of a grey heron over the experimental tank and then added bloodworms to the food cup. When at least one fish took a bite of food, an attack was simulated by quickly releasing the predator's head. I observed individual behaviours for 300 s while the predator model was still present above the tank and recorded the time taken since the attack for each individual to take the first bite of food. Risk taking was measured simultaneously in all

S.-Y. Kim / Animal Behaviour 112 (2016) 147e152

individuals in the same tank. Thus, nonindependence of samples, for example due to interference and cues from the others and their behaviour, may be an important environmental component of variation in this measure (Magnhagen, 2007). This environmental effect was taken into account in my statistical analyses. Ethical Note Wild sticklebacks were sampled with the permission of the Xunta de Galicia (021/2013). The three-spined stickleback is abundant at the site of capture. Six juvenile brown trout were provided by the Carballedo fish farm, where wild trout are reared specifically for the purpose of restoring natural populations in Galician rivers. All methods involving this experiment were approved by the Animal Ethics Committee of the University of Vigo (17/12). The F1 stickleback stock was reared for other studies, and only part of the stock was used in this study. After the elastomer tags were fitted, each fish was retained in an individual tank for a few minutes until it recovered from the anaesthesia. All fish recovered well without any ill effects or mortality caused by anaesthesia or tagging. As exposure to simulated predation risk can be stressful for fish, I tried to maximize animal welfare by providing an artificial plant as refuge in all fish tanks and by minimizing sample size and the duration of exposure to simulated predation risk. Testing individual or genetic variation in animal personality or behavioural plasticity across different environments usually requires many more samples and repeated measures than a simple comparison of average behaviours across different experimental groups. I minimized the number of samples used in this study by equally allocating genetically related individuals (full- and halfsibs) in both experimental groups, thereby reducing sampling bias due to different genotypes being exposed to different conditions, and by using only three repeated measures of behaviours, which is the minimum number required for successful testing of an individual)time interaction in a random regression model (for a similar approach, see also Dingemanse et al., 2012). At the end of the experiment, the control fish were returned to the same growth tanks from which they were sampled. The experimental fish were killed by an overdose of benzocaine anaesthetic for a study of the relationship between physiological and behavioural traits (results not reported here).

149

0.2857 and 0.7143. By using RR models fitted separately to the experimental and control groups, I estimated the betweenindividual variation in average behaviour (elevations), the between-individual variation in behavioural change with time (slopes) and the covariance between elevation and slope for each behavioural trait. Thus, I fitted an RR model for the behaviour B of individual i at standardized observation day d as follows: Bi,d ¼ m þ weekF þ observation dayF þ gti þ eti þ f(indxi,d) þ 3i,d. The coefficient m is the overall mean trait expression (intercept). The terms weekF and observation dayF are factorial fixed effects that denote the time of experimental session (a seven-level categorical variable) and observation day (a three-level categorical variable), which were included for a full description of any nonlinearity. The random term f(indxi,d) was the RR function of order x of the individual-specific variation over the continuous variable d. I compared the model fits with different polynomial functions to examine how the effect differed between constant (zero-order, x ¼ 0) and linear (first-order, x ¼ 1) forms of the reaction norms. Zero-order function assesses whether there is any substantial variation in the average behaviour between individuals (repeatability); first-order function assesses variation in behavioural change over time. Growth tank- and experimental tankspecific common environment variance components (gti and eti) and heterogeneous residual error (3 i) were also included in the model as random terms. Use of a heterogeneous error variance structure (3 i,d) that allows residual variance to vary across different days of observation did not produce a better fit than a model with a homogeneous error (3 i) in the analysis of shoaling (experimental: c22 ¼ 5.18, P ¼ 0.075; control: c22 ¼ 0.20, P ¼ 0.905) and risk taking (experimental: c22 ¼ 1.16, P ¼ 0.560; control: c22 ¼ 0.36, P ¼ 0.835). Nevertheless, heterogeneous residual variances were assumed in all RR models to avoid forcing any residual errors into the other components and inflating estimates of I)T (Westneat, Wright, & Dingemanse, 2015). The significance of fixed terms was evaluated on the basis of Wald F statistics. Significance of (co)variance components was assessed by calculating the log likelihood ratio and testing against a chi-square distribution with degrees of freedom equal to the difference in degrees of freedom between the two hierarchical models compared (Pinheiro & Bates, 2000).

Statistical Analyses RESULTS All restricted maximum likelihood (REML) models were fitted using ASReml v 3.0 (Gilmour, Gogel, Cullis, & Thompson, 2008). I first tested whether shoaling and risk taking change in response to the presence of a predator at the population level by comparing temporal changes in behaviour between the experimental and control groups. I explored the effects of the experimental treatment, time (observation day: day 0, 3 and 4) and their interaction on behaviour in an REML linear mixed-effect model (LME), including individual identity, growth tank and experimental tank as random effects and weekly experimental session as an additional fixed effect. In addition, I used a random regression (RR) model to examine whether temporal changes in behaviour during the experiment varied between individuals (individual)time interaction, I)T; for a similar approach, see Brommer, 2013; Dingemanse et al., 2012). In this model, the individual values were modelled as a linear random effect function over the continuous variable, observation day. ASReml calculates the variances of RR slopes specific to the covariate scaling (Schaeffer, 2004). Thus, observation day was standardized to a scale ranging from 1 and a mean of 0 (d in the model described below). Observation days 0, 3 and 4 were scaled to 1,

The levels of shoaling and risk taking changed over time during the experimental period for both experimental and control groups (Fig. 1). Linear mixed-effect model analysis showed that over the 4 days the time it took for the individuals to shoal decreased (LME: standardized coefficient, b ¼ 0.063, 95% confidence interval for b, CI ¼ 0.110, 0.016; F1, 885 ¼ 6.57, P ¼ 0.011). There was a significant negative relationship between week of experiment and time taken to approach conspecifics (b ¼ 0.103, CI ¼ 0.176, 0.031; F1,50 ¼ 7.66, P ¼ 0.008). The patterns of temporal change in shoaling, however, did not differ between the experimental and control groups (treatment: F1, 48.9 ¼ 0.23, P ¼ 0.637; treatment)day: F1, 884 ¼ 1.12, P ¼ 0.292). There was a significant interaction between observation day and treatment in risk taking (observation day: b ¼ 0.035, CI ¼ 0.086, 0.016; F1, 884 ¼ 0.08, P ¼ 0.777; treatment: b ¼ 0.140, CI ¼ 0.044, 0.324; F1, 49.6 ¼ 2.23, P ¼ 0.142; treatment) day: b ¼ 0.079, CI ¼ 0.006, 0.152; F1, 884 ¼ 4.60, P ¼ 0.033). The control individuals took progressively less time to take a bite of food over time, but the experimental individuals became fearful and took more time. There was no significant effect of week (F1,55.7 ¼ 0.34, P ¼ 0.561).

S.-Y. Kim / Animal Behaviour 112 (2016) 147e152

Shoaling (s)

Unsociable

150

240

ind1 ¼ 387.1 ± 356.2, Covind0,ind1 ¼ 213.2 ± 267.5; experimental: ind0 ¼ 613.0 ± 243.7, ind1 ¼ 187.3 ± 334.7, Covind0,ind1 ¼ 214.3 ± 241.7). In other words, the behavioural reaction norm slope did not differ significantly between individuals (i.e. nonsignificant I)T) for shoaling. The VI value was also significant for risk taking during foraging (Table 1; control: r ¼ 0.443, CI ¼ 0.294, 0.592; experimental: r ¼ 0.403, CI ¼ 0.254, 0.552). There was no statistical support for I) T in the behavioural plasticity in either control or experimental group (Table 1; RR estimates from the model including I)T term: control: ind0 ¼ 6234.2 ± 1258.5, ind1 ¼ 564.1 ± 982.3, Covind0,ind1 ¼ 228.8 ± 741.5; experimental: ind0 ¼ 5594.2 ± 1118.4, ind1 ¼ 1524.9 ± 891.2, Covind0,ind1 ¼ 850.4 ± 697.8). Nevertheless, the visualized reaction norms of risk taking based on the best linear unbiased predictor (BLUP) values suggest that the level of behavioural plasticity was related to the initial level of boldness in the experimental group (Fig. 2b). Initially bold individuals (i.e. those that took less time to feed under predation risk at the onset of the experimental treatment) tended to increase in fearfulness more rapidly than the others (Fig. 2b).

230

DISCUSSION

90 (a)

Control

85

Experimental

80 75 70

Risk taking (s)

Fearful

Sociable

65 60 55

0

1

3

2

4

250 (b)

220

Bold

210 200

0

1

3

2 Day

4

Figure 1. Mean ± SE time spent (a) shoaling and (b) risk taking according to observation day from the onset of the experimental treatment.

Results of the RR analyses of shoaling showed significant between-individual variation (VI) in average behaviour in both control and experimental groups (Table 1; control: repeatability, r ¼ 0.115, 95% confidence interval for r, CI ¼ 0.001, 0.229; experimental: r ¼ 0.164, CI ¼ 0.052, 0.276). However, including individualspecific slopes (ind1) in addition to individual-specific elevations (ind0) did not significantly improve the model fit in either the control or the experimental group (Table 1, Fig. 2a; RR estimates from the model including the I)T term: control: ind0 ¼ 608.7 ± 285.4,

My results show that the level of shoaling increased over time probably because individuals became familiar with the repeated shoaling test; however, there was no evidence for plasticity in response to predation risk. The difference in temporal change in risk taking between the sticklebacks exposed to predation pressure and those under no such stress demonstrates that the fish adjusted their level of foraging activity to the current level of predation risk in the environment. Both shoaling and risk taking were highly repeatable within individuals. Nevertheless, there was no statistical evidence for between-individual variation in the behavioural changes over time in either control or experimental fish, suggesting that the behavioural plasticity was a fixed response in the individual members of this population. In natural populations, shoaling may be beneficial for individuals because joining social groups can dilute the individual risk of predation in the presence of predators, but in other situations individuals can avoid competition for resources by being alone (Hoare, Couzin, Godin, & Krause, 2004; Krause & Ruxton, 2002). Although individual differences in behavioural plasticity may be beneficial for individuals to manage environmental uncertainty (Briffa, 2013; Mathot, Wright, Kempenaers, & Dingemanse, 2012), the temporal pattern of shoaling was not flexible across different individuals or environments.

Table 1 Variance components (estimate ± SE, significance given in parentheses) from the best fit RR models of shoaling and risk taking in juvenile three-spined sticklebacks from the control and experimental groups Random regression model variances of the best fit models

Test for I)T (ind1)

VR0

VR3

VR4

VGT

VET

VI (ind0)

c2

P

Shoaling Control Experimental Risk taking Control

3164.6±349.0 3016.9±332.3

2551.3±296.8 1755.5±222.3

2811.8±316.8 2641.2±296.3

145.2±149.5 (P¼0.327) 257.7±148.7 (P¼0.077)

89.3±77.7 (P¼0.159) 0 (P¼1)

416.2±206.0 (P¼0.019) 490.4±179.8 (P<0.001)

1.18 1.00

0.554 0.607

6401.5±817.6

5193.6±718.3

6056.2±787.9

957.9±890.3 (P¼0.277)

720.7±510.6 (P¼0.068)

5971.6±1147.5 (P<0.001)

0.4

0.819

Experimental

6443.5±799.7

4953.8±677.7

5742.5±739.1

772.4±774.1 (P¼0.327)

662.7±447.0 (P¼0.055)

4937.0±1012.1 (P<0.001)

2.84

0.242

VR0, VR3 and VR4: residual variances on observation days 0, 3 and 4, respectively. VGT and VET: growth tank-specific and experimental tank-specific common environment variances, respectively. VI: individual variance. The significance of I)T was assessed on the basis of the log likelihood ratio between the best-fit model and the higher rank model, including the RR function f(ind1i,d).

S.-Y. Kim / Animal Behaviour 112 (2016) 147e152

140

(a)

160

Control

151

Experimental

140

120

120 100 100 80 80

BLUP(ind)

60

60

40 0

350 (b)

3

4

40 0

350

Control

300

300

250

250

200

200

150

150

100

100

50

50

0 0

3

3

4

3

4

Experimental

0 4 0 Observation day

Figure 2. Plots of the reaction norms for (a) shoaling and (b) risk taking of sticklebacks since the onset of the experimental treatment for the control and experimental groups. The behavioural reaction norms are based on the fixed effects and the best linear unbiased predictor (BLUP) values obtained from the model including the RR function f(ind1i,d) in Table 1. Each line represents the reaction norm of a single individual (control: N ¼ 221; experimental: N ¼ 222).

Despite the substantial degree of variation in the level of risk taking between individuals, fish exposed to predation pressure during the experimental period showed a fixed behavioural response manifested as a decreasing willingness to forage under risk. In other words, the level of risk taking of most individuals changed in a similar direction and at a similar rate. Risk taking (or boldness) is an important component of animal personality that influences extrinsic mortality and intrinsic condition through behavioural mechanisms whereby individuals avoid predation and secure resources (Bell & Sih, 2007; Briffa, Rundle, & Fryer, 2008). Therefore, selection pressure over evolutionary history may have imposed a single optimal behavioural rule in this population, thereby eroding differences in behavioural plasticity between individuals (Houston & McNamara, 1999). Behavioural plasticity in response to predation risk was present but did not vary between individuals in this population. The study findings show that the current predation pressure may determine the average level of risk taking; the individual differences in the level of risk taking should be maintained across environments with different levels of predation pressure. A previous paper proposed that feedback between individual states and the state-dependent behaviour should have profound effects on between-individual differences in personality (Wolf & Weissing, 2010). Although positive feedback explains the emergence and maintenance of between-individual differences in behaviour, negative-feedback mechanisms may erode the differences (Dingemanse & Wolf,

2013). The observed plasticity may be a fixed response for this population to optimize individual-specific behaviour patterns according to their state in a given environment (Dall, Houston, & McNamara, 2004; Luttbeg & Sih, 2010). Future studies integrating individual state, personality, behavioural plasticity and fitness are therefore warranted.

Acknowledgments  C. Noguera and the anonymous I thank Alberto Velando, Jose referees who provided constructive and helpful comments. I am also grateful to Niels Dingemanse and Jon Brommer for advice on the study design and statistical analysis and to Pablo Caballero, Jorge Dominguez, Daniel Sabucedo, Antonio Gonzales and Andrea Tato for helps with the experiment. Finance was provided by the Spanish Ministerio de Economía y Competitividad (CGL201240229-C02-02 and CGL2014-60291-JIN) and the Xunta de Galicia (2012-305).

Supplementary Material Supplementary material associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.anbehav. 2015.12.004.

152

S.-Y. Kim / Animal Behaviour 112 (2016) 147e152

References Bell, A. M. (2005). Behavioural differences between individuals and two populations of stickleback (Gasterosteus aculeatus). Journal of Evolutionary Biology, 18, 464e473. Bell, A. M., & Sih, A. (2007). Exposure to predation generates personality in threespined sticklebacks (Gasterosteus aculeatus). Ecology Letters, 10, 828e834. Briffa, M. (2013). Plastic proteans: reduced predictability in the face of predation risk in hermit crabs. Biology Letters, 9, 20130592. Briffa, M., Rundle, S. D., & Fryer, A. (2008). Comparing the strength of behavioural plasticity and consistency across situations: animal personalities in the hermit crab Pagurus bernhardus. Proceedings of the Royal Society B: Biological Sciences, 275, 1305e1311. Brommer, J. E. (2013). Variation in plasticity of personality traits implies that the ranking of personality measures changes between environmental contexts: calculating the cross-environmental correlation. Behavioral Ecology and Sociobiology, 67, 1709e1718. Dall, S. R., Houston, A. I., & McNamara, J. M. (2004). The behavioural ecology of personality: consistent individual differences from an adaptive perspective. Ecology Letters, 7, 734e739. Dingemanse, N. J., Barber, I., Wright, J., & Brommer, J. E. (2012). Quantitative genetics of behavioural reaction norms: genetic correlations between personality and behavioural plasticity vary across stickleback populations. Journal of Evolutionary Biology, 25, 485e496. ale, D., & Wright, W. (2010). Behavioural reDingemanse, N. J., Kazem, A. J. N., Re action norms: animal personality meets individual plasticity. Trends in Ecology & Evolution, 25, 81e89. Dingemanse, N. J., & Wolf, M. (2013). Between-individual differences in behavioural plasticity within populations: causes and consequences. Animal Behaviour, 85, 1031e1039. Gilmour, A. R., Gogel, B. J., Cullis, B. R., & Thompson, R. (2008). ASReml user guide (release 3.0). Hemel Hempstead, U.K.: VSN International Ltd. Han, C. S., & Brooks, R. C. (2013). Evolution of individual variation in behaviour and behavioural plasticity under scramble competition. Animal Behaviour, 86, 435e442. Hoare, D. J., Couzin, I. D., Godin, J.-G. J., & Krause, J. (2004). Context-dependent group size choice in fish. Animal Behaviour, 67, 155e164. Houston, A. I., & McNamara, J. M. (1999). Models of adaptive behaviour: An approach based on state. Cambridge, U.K.: Cambridge University Press. Huntingford, F. A., Wright, P. J., & Tierney, J. F. (1994). Adaptive variation and antipredator behaviour in threespine stickleback. In M. A. Bell, & S. A. Foster (Eds.), The evolutionary biology of the threespine stickleback (pp. 277e295). Oxford, U.K.: Oxford University Press. Kim, S.-Y., & Velando, A. (2015). Phenotypic integration between anti-predator behaviour and camouflage pattern in juvenile sticklebacks. Evolution, 69, 830e838. Krause, J., & Ruxton, G. D. (2002). Living in groups. Oxford, U.K.: Oxford University Press.

Krebs, J. R., & Davies, N. B. (1997). Behavioural ecology: An evolutionary approach. Sunderland, MA: Sinauer. Laskowski, K. L., & Bell, A. M. (2013). Competition avoidance drives individual differences in response to a changing food resource in sticklebacks. Ecology Letters, 16, 746e753. Lima, S. L. (1998). Stress and decision-making under the risk of predation: recent developments from behavioral, reproductive, and ecological perspectives. Advances in the Study of Behavior, 27, 215e290. Luttbeg, B., & Sih, A. (2010). Risk, resources and state-dependent adaptive behavioural syndromes. Philosophical Transactions of the Royal Society B: Biological Sciences, 365, 3977e3990. Magnhagen, C. (2007). Social influence on the correlation between behaviours in young-of-the-year perch. Behavioral Ecology and Sociobiology, 61, 525e531. Mathot, K. J., Wright, J., Kempenaers, B., & Dingemanse, N. J. (2012). Adaptive strategies for managing uncertainty may explain personality-related differences in behavioural plasticity. Oikos, 121, 1009e1020. Nussey, D. H., Wilson, A. J., & Brommer, J. E. (2007). The evolutionary ecology of individual phenotypic plasticity in wild populations. Journal of Evolutionary Biology, 20, 831e844. Oliveira, R., Taborsky, M., & Brockmann, H. J. (2008). Alternative reproductive tactics. Cambridge, U.K.: Cambridge University Press. Pigliucci, M. (2001). Phenotypic plasticity: Beyond nature and nurture. Baltimore, MA: Johns Hopkins University Press. Pinheiro, J. C., & Bates, D. (2000). Mixed-effects models in S and S-Plus. New York, NY: Springer Verlag. Relyea, R. A. (2005). The heritability of inducible defenses in tadpoles. Journal of Evolutionary Biology, 18, 856e866. Roff, D. A. (1997). Evolutionary quantitative genetics. New York, NY: Chapman and Hall. Sasaki, A., & Ellner, S. (1997). Quantitative genetic variance maintained by fluctuating selection with overlapping generations: variance components and covariances. Evolution, 51, 682e696. Schaeffer, L. R. (2004). Application of random regression models in animal breeding. Livestock Production Science, 86, 35e45. Scheiner, S. M. (1993). Genetics and evolution of phenotypic plasticity. Annual Review of Ecology and Systematics, 24, 35e68. Stamps, J. A. (2007). Growthemortality tradeoffs and ‘personality traits’ in animals. Ecology Letters, 10, 355e363. Westneat, D. F., Wright, J., & Dingemanse, N. J. (2015). The biology hidden inside residual within-individual phenotypic variation. Biological Reviews, 90, 729e743. Wolf, M., van Doorn, G. S., Leimar, O., & Weissing, F. J. (2007). Life-history trade-offs favour the evolution of animal personalities. Nature, 447, 581e584. Wolf, M., van Doorn, G. S., & Weissing, F. J. (2011). On the coevolution of social responsiveness and behavioural consistency. Proceedings of the Royal Society B: Biological Sciences, 278, 440e448. Wolf, M., & Weissing, F. J. (2010). An explanatory framework for adaptive personality differences. Philosophical Transactions of the Royal Society B: Biological Sciences, 365, 3959e3968.