1 Environmental Biology of Fishes December 2015, Volume 99, Issue 1, Pages 105-115 http://dx.doi.org/10.1007/s10641-015-0459-4 http://archimer.ifremer.fr/doc/00296/40697/ © Springer Science+Business Media Dordrecht 2015 Achimer http://archimer.ifremer.fr First insight into personality traits in Northern pike (Esox lucius) larvae: a basis for behavioural studies of early life stages Pasquet Alain 1 2 1, * 1 2 1 1 , Sebastian Anthony , Begout Marie-Laure , Ledore Yannick , Teletchea Fabrice , 1 Fontaine Pascal CNRS, UR AFPA, USC INRA 340, University of Lorraine, 54506, Vandoeuvre les Nancy, France Ifremer, place Gaby Coll, BP 7, 17137, La Rochelle, L’Houmeau, France * Corresponding author : Alain Pasquet, email address : alain.pasquet@univ-lorraine.fr Abstract : Personality traits have been studied for some decades in fish species. Yet, most often, studies focused on juveniles or adults. Thus, very few studies tried to demonstrate that traits could also be found in fish larvae. In this study, we aimed at identifying personality traits in Northern pike (Exos lucius) larvae. Twenty first-feeding larvae aged 21 days post hatch (16.1 +/− 0.4 mm in total length, mean +/− SD) were used to establish personality traits with two tests: a maze and a novel object. These tests are generally used for evaluating the activity and exploration of specimens as well as their activity and boldness, respectively. The same Northern pike twenty larvae were challenged in the two tests. Their performances were measured by their activity, their exploratory behaviour and the time spent in the different arms of the maze or near the novel object. Then, we used principal component analysis (PCA) and a hierarchical ascendant classification (HAC) for analysis of each data set separately. Finally, we used PCA reduction for the maze test data to analyse the relationship between a synthetic behavioural index (PCA1) and morphometric variables. Within each test, larvae could be divided in two sub groups, which exhibited different behavioural traits, qualified as bold (n = 7 for the maze test and n = 13 for the novel object test) or shy (n = 9 for the maze test and n = 11 for the novel object test). Nevertheless, in both tests, there was a continuum of boldness/shyness. Besides, some larvae were classified differently between the two tests but 40 % of the larvae showed cross context consistency and could be qualified as bold and/or proactive individuals. This study showed that it is possible to identify personality traits of very young fish larvae of a freshwater fish species. Keywords : Behavioural tests, Activity, Exploration, Boldness/shyness continuum, Freshwater species Please note that this is an author-produced PDF of an article accepted for publication following peer review. The definitive publisher-authenticated version is available on the publisher Web site. 42 Introduction 43 For more than twenty years, many studies focusing on behavioural variations have been conducted, 44 and a new concept has emerged: animal personality, which is unique to each individual (Réale et al. 45 2007; Toms et al. 2010; Conrad et al. 2011; Nakayama et al. 2012). Personality can be defined as a 46 collection of behavioural traits, which are consistent both over time and in different environmental 47 conditions (Gosling 2001; Dall et al. 2004; Castanheira et al. 2013). Nevertheless, this does not 48 imply that individual behavioural traits do not evolve over time, but rather that the combination of 49 them is rather constant whatever the period of life. Likewise, even though personality of fish is 50 consistent (e.g., a “shy fish” will remain “shy”), it can be shaped differently according to 51 environmental factors, such as water temperature (Biro et al. 2010), predation pressure (Brown et al. 52 2005; Bell and Sih 2007), food predictability (Chapman et al. 2010) or life history traits (Biro and 53 Stamps 2008; Adriaenssens and Johnsson 2011). From these studies, personality can be considered 54 as an individual variable; it is a resultant behavioural characteristic, which could be correlated with 55 other individual variables (morphology, genetic origin) and compared between different 56 environmental situations. 57 Classically, five main behavioural traits are considered to establish behavioural personality 58 in fish: activity, boldness, exploration, aggressiveness and sociability (Réale, et al. 2007). For fish, 59 activity is defined by several variables linked to fish movements: type and characteristics of 60 swimming (speed, angular speed) and distances covered (Norton et al. 2011). Second, boldness (or 61 its opposite shyness) (e.g. Dyer et al. 2009) is defined as the propensity to take risks by exploring 62 new environments and consequently taking more risks of being predated. Boldness can be tested 63 with a lot of devices in different situations. For example, fishes can be placed in front of new types 64 of food (Höjesjö et al. 2011) or new objects (Norton et al. 2011). The performances, such as the 65 latency time for taking new food or for exploring the new object, are then measured. Boldness can 3 66 also be evaluated by the reactions of fish facing a novel environment (Norton et al. 2011); in this 67 case special devices are used such as mazes. Besides, boldness can also be measured under stressful 68 conditions, such as in face of a predator or a lure of predator (Moretz et al. 2007). Third, exploration 69 is the ability of a fish to discover a new environment. It is tested by using different devices (mazes, 70 open field). Fourth, aggressiveness is defined as the involvement in fights or aggressive interactions 71 with congeners (e.g., Norton et al. 2011). Introducing two fish in the same space (or separated by a 72 glass) can be used for testing aggressiveness; interactions between fishes may be characterized by 73 aggressive behaviours, such as chases, attacks or bites (Biro et al. 2010); it can also be done using a 74 mirror (Gerlai et al. 2000). Fifth, sociability corresponds to the interactions of an individual within a 75 social group and its relationships with its conspecifics (e.g. Cote et al. 2010). Inter-individual 76 distances and interactions are often used as proxies for sociability (Cote et al. 2010). 77 Most studies on fish personality have been performed either on juveniles (Westerberg et al. 78 2004) or adults (Dahlbom et al. 2011), whereas very few studies focused on early life stages. 79 Sundström et al. (2004) realized one of the few complete studies on larvae, yet in this case, the trout 80 Salmo trutta fry have a relative large size at hatching (more than 15 mm). In general, the tiny size 81 and the fragility of larvae are two reasons explaining why there are so few studies focusing on the 82 personality traits of these stages. However, such knowledge at these early life stages may be 83 important both for understanding the dynamic of recruitment of fish population in wild conditions 84 and the management of farmed fish by identifying particular individuals that could be strongly 85 aggressive or potentially cannibals under rearing conditions. 86 The Northern pike Esox lucius, is a solitary species at adult stage, which inhabits lakes, 87 ponds and also rivers with weak current, in Eurasia and North America (Kottelat and Freyhof 88 2007). The larvae are about 9 mm at hatching (Teletchea et al. 2009; Parlier and Corolla 2013; 89 Trabelsi et al. 2013). The diet of the larvae is based mainly on zooplankton, and both juveniles and 4 90 adults are chiefly piscivorous (Kottelat and Freyhof 2007). Northern pike is a diurnal predator using 91 vision as the main sense to localize its prey; cannibalism in larvae can appear within few days after 92 first feeding and not only as juveniles (one month after hatching) (Giles et al. 1986, Kucharczyk et 93 al. 1997). There is only one study on the existence of behavioural syndromes defined as a suite of 94 correlated personality traits (Sih et al. 2004) in pike: they found that personality existed in young- 95 of-the year pikes (47-74mm fork length) across three contexts (activity in presence of a competitor, 96 exploration of a novel environment and boldness under predation risk) but there were no temporally 97 personality traits (Nyqvist et al. 2013). Until now, most of the explanations of cannibalism in fish 98 larvae are based on size heterogeneity whatever the origin of this heterogeneity (growth 99 heterogeneity, environmental factors (quantity of food, density of larvae, luminosity, water 100 turbidity) (for review see Baras 2012)). If these studies furnished a lot of information on the 101 physiological origins of cannibalism, they did not answer to the question: why does a larva become 102 cannibal? The study of the larva personality could be an answer to this question. 103 The goals of the present study were to develop a method for sorting pike larvae according to 104 their personality traits. This classification could be used for ranking fish larvae according to their 105 behavioural performances. Such a classification requires (1) identifying different behavioural 106 responses, (2) testing the variability of these behaviours, and (3) dividing the studied population 107 into groups separated by their behavioural traits. 108 109 Material and methods 110 Rearing conditions 111 Eggs were obtained from wild adults caught in ponds located in the Lindre Centre (Moselle, France, 112 (L: 48.799, l: 6.747). Sperm of four males was pooled for fertilizing oocytes of one female, so 113 tested fish could be siblings or half-siblings. Eggs were incubated at 11°C. After hatching, larvae 5 114 were maintained in small nets (15x15x10cm) (n=15; 80-100 individuals per net) in a large tank at a 115 temperature between 11°C and 13°C. larvae were collected at the beginning of the hatching period 116 and there was no age difference between them. During the period of larval rearing, a photoperiod of 117 12 hours (day light from 8:00 to 20:00 h) was applied; the luminosity in the room was set at 118 100 Lux. The air temperature of the rearing room was set at 13°C (± 2°C). Water pH was stabilized 119 at 7.4. Food (Artemia nauplii) was distributed ad libitum from 9:00 to 16:30 hours starting at 10 120 days post hatching (dph), that was the beginning of the exogenous feeding period, and was 121 completed during five days with food granulates (100 µm, Biomar, Brande (Danemark), chemical 122 composition: NFE 14%, protein 58%, lipids 12%, Vitamin A (lU Kg-1) 17500 , DHA (g kg-1) 11, 123 EPA (g kg-1) 9). After five days of co-feeding, a larva received only granulates. 124 125 Tests 126 Three personality traits were studied: activity, exploration and boldness. The behavioural 127 performances of fish were evaluated using two devices: the maze and the novel object. For the two 128 tests, we used a luminous table with a low light (less than 100 Lux); the different devices (maze and 129 small aquariums for the novel object test) had a translucent bottom and all behaviours were 130 recorded via a camera (Sony Handycam DCR-SR72E) set just above the device. For each test, the 131 device was filled with water from the hatching room to keep the same water quality. In order to 132 avoid differences in temperature between the water of the device and those of the hatchery, we 133 maintained the temperature of the test room at 13°C. During the tests, the water temperature was 134 monitored in all devices. For both tests, the same twenty larvae were used. Larvae were between 18 135 and 28 dph. We tested two larvae per day; for each larva, the order of the test was randomly 136 distributed. Each larva was thus evaluated in the two tests during the same day; so we had to take 137 the larva from one device after the first test and place it into an individual holding rack in the 138 hatching room before the second test was conducted. 6 139 140 Maze test 141 142 We used a plus maze (Fig.1A); the size of each arm (16x7x3 cm) was defined in order to fit 143 with the size and the swimming abilities of larvae. At the extremity of one arm, we put a small box 144 (9x7x3 cm) as a start box, for acclimating the larva to its new environment. This start box was 145 separated from the first arm by a removable wall, which we removed after five minutes, which was 146 typically (Ninkovic and Bally-Cuif 2006) used in such experiments. The larvae were allowed to 147 swim freely in the maze during 30 min. For video analyses, we divided the maze into five zones 148 (the four arms and the square part in the middle). 149 150 151 Novel object test 152 To evaluate this test, we made two small aquaria (19.5x5x4.5 cm) stuck together by an 153 opaque common wall: each aquarium was divided into two unequal compartments by an opaque 154 divider (fig. 1B). This device allowed us testing two fish at the same time. Before the beginning of 155 the test, we introduced a steel nut in the separated novel object compartment. At the beginning of 156 the test, each larva was put into each aquarium and the removable divider was maintained in place. 157 After 15 min, the divider was removed and the larva was free to explore the novel object in their 158 environment during 30 min. For video analyses, the aquaria were divided into two equal zones (Fig. 159 1B). 160 161 Video data collection 162 163 We used the software “The Observer XT” (Noldus, version 10.0) to analyse the data. From 164 preliminary observations, we had defined the relevant behaviours (see below) exhibited by the 7 165 larvae. We encoded all behaviours (variables of interest, Table 1), and measured the time laps 166 between them. For each test, we chose to analyse six periods of two minutes separated by a period 167 of three minutes for video of the twenty larvae. By comparing such sampled analysis with the 168 analysis of an entire video recording, we found that this method allowed us reducing time without 169 losing behavioural information. 170 171 Behaviours 172 In the two tests, we considered fish activity by measuring the time spent swimming or immobile 173 (Table 1). Swimming was subdivided into rapid swimming (RS) or slow swimming (SS). RS was 174 defined as instantaneous speed more than one larva body length s-1 (BL s-1) and SS less than one BL 175 s-1. Immobility (IM) was defined by the fact that the fish did not move from its position (the head 176 did not change its position in space but the body may present some undulations) and measured by 177 its duration (s). The second set of variables dealt with exploration of the space and time spent in the 178 different zones of the devices in the two tests. In the maze, we measured the latency for the larva to 179 go out from the start box (LA), the number of arms visited (NA), the total number of zones visited 180 (NZ1) and the time spent in the first arm of the maze added to the time spent in the start box (TT). 181 In the novel object test, we counted the number of zones visited (NZ2) by the larva and the time 182 spent at a distance of less than one BL-1 from the object (TO). 183 184 Morphological data 185 Just after running the tests, each fresh larva was weighted (Sartorius Scale, accuracy 0.1 mg) and 186 stored individually in formalin (4%). Thereafter, they were placed under a binocular microscope 187 (OPTIKA microscope, SZP-10), connected to a camera (MICROVISON Instruments, Lw1235C- 188 GT1). This system allowed us obtaining photographs. From the photographs, we measured two 8 189 morphological variables (software: Archimed, MICROVISON Instruments 6.0.14): the total length 190 of the body from the extremity of the head to the end of the caudal fin and the eye diameter. We 191 used the total length of the larva, because it is a characteristic directly linked to growth and to 192 muscular abilities. The eye diameter could be a relevant parameter of the perception of the 193 environment and demonstrated as a good correlate of personality in fish (e.g.: bigger eye fish 194 showed bolder than smaller eye one, Rey et al. 2013) even though we have no information on the 195 maturation of the visual system of the northern pike larvae. 196 197 Data analyses 198 199 Data on the behavioural variables of interest were summed for the six periods of two minutes 200 leading to twelve minutes of observation per fish in each test. The normality for all the data was 201 tested with the Shapiro-Wilk method. Mean, standard deviation (SD), and coefficient of variation 202 (CV in percentage) were calculated for each variable. 203 The analysis was conducted in three steps. First, we characterized the personality traits of 204 fish in each of the two tests, second we compared the personality traits of fish between the two tests 205 (cross context consistency analysis) and third we evaluated the possible links between these traits 206 and morphological traits for the maze test. We analysed both kinds of data (behavioural and 207 morphological) by using a principal component analysis (PCA). In the PCA, only principal 208 components with Eigen values greater than 1 were considered for further analyses. . For each PCA, 209 we gave the percentage of variance of each variable on the first axis (e.g.; the NZ1 represented 15% 210 of the variance expressed by the first axis, PC1). For each test separately, a hierarchical ascending 211 classification (HCA) was performed on the scores of the larvae along the axis 1 of the PCA (PC1 212 scores of PCA reduction). This analysis was done to identify groups of individuals displaying 9 213 similar personality traits. The level of group discrimination was considered as the level 1 when the 214 dissimilarity was higher (method used is the smaller distance between two individuals). After each 215 HCA, the data on variables of interest were compared between the two groups with non-parametric 216 tests (Mann-Whitney U test) due to the low number of individuals in each group. We compared the 217 different positions of the individuals after the HCA between the two tests and determined the 218 proportion of individuals, which were or not in the same group. We used these two statistical 219 methods (PCA and HCA) because they are complementary and present different powers with non- 220 transformed data. For the maze test, the correlations between the scores of the larvae along the axis 221 1 of the PCA and the morphological data were analysed using Pearson regression. All the statistical 222 analyses were performed with Statistica (version10). 223 224 Results 225 226 Morphological and behavioural data 227 228 The morphological data and the mass weakly differed between the 20 larvae studied (Table 2). The 229 behavioural variables showed a large variability between individuals and the CV varied from 53% 230 to 206% (Table 2). The morphological variables showed less variation; their CV varied from 10% 231 to 43% (Table 2). 232 233 Activity and Exploration of pike larvae (maze test) 234 235 The first two axes of the PCA represented around 60% of the total variability (fig. 2A, F1=40.25%, 236 F2=19.02%), thus they were the only ones considered further. On the first axis (Fig. 2A), 237 morphological variables were associated positively with behavioural variables, which corresponded 10 238 to exploratory behaviours (NA, SS); they were opposite to non-exploratory behaviours (LA, and 239 TT). For example, the number of arms visited represented 16% of the total variance and the latency 240 to go out of the start zone 11%. There was a great variability in the individual contribution to the 241 total variability (from 0% to 17%). This implies that larvae exhibited a great variation in their 242 behavioural responses in the device. The second axis represented more the morphological variables 243 (body length (12%), eye diameter (11%) of the total variance) opposed to two behavioural variables 244 (NA (11%) and NZ1 (15%) of the total variance). 245 Two main groups were found in the hierarchical ascending clustering (Fig. 2B) (Variance 246 within class=49.8% and between classes=50.2%). The limit of the dissimilarity between the two 247 groups was clearly established. The first group gathered thirteen individuals (G1) and the second 248 group seven (G2). The larvae of G2 were more active (Mann-Whitney U test for all comparisons, 249 NG1=13, NG2=7): they visited more arms (U=8.5, p=0.002), spent more time slow swimming (U 250 =21.0, p=0.057), stayed less time in the first zone (U=8.5, p=0.002), and had a shorter latency time 251 (U=8.3, p=0.003). These individuals had a higher body length (U=14.0, p=0.014) and a greater eye 252 diameter (U=17.0, p=0.025). 253 254 Activity and boldness of pike larvae (novel object test) 255 256 The first two axes of the PCA represented around 66 % of the total variability (F1=38.64%, 257 F2=27.85%, fig. 3A). Exploratory behaviours (SS and NZ2) were associated with the first axis and 258 as well as morphological variables with IM and opposed to RS (Fig.3A). Activity variable (TO) was 259 associated with the second axis (Fig. 3A). Therefore, it appeared that the variables were strongly 260 divided by their nature on the first axis; exploration variables (SS represented 32% and NZ2 32% of 11 261 the variance), and morphological data (body length represented 23% of the total variance, eye 262 diameter 20%), 263 264 The results of the hierarchical ascending clustering showed a clear division of the larvae into 265 two groups (Fig. 3B) (Variance within class=53.8% and between classes=46.2%). The limit of the 266 dissimilarity between the two groups was clearly established: a group of nine (G1) and a group of 267 eleven larvae (G2). The larvae of G2 were more active (Mann-Whitney U test for all comparisons, 268 NG1=11, NG2=9): they spent more time slow swimming (U=18.0, p=0.003), tended to spent more 269 time rapid swimming (U=26.0, p=0.074), they spent more time near the novel object (U=3.0, 270 p<0.001), and visited a higher number of zones (U=6.0, p=0.001) with less immobility (U=42 271 p=0.57). There was no difference in morphological variables between the two groups (Total length 272 of the body U=46 p=0.79, eye diameter U=48, p=0.91). 273 274 Cross context analysis and relationships between behavioural, morphological data and age 275 276 Among the twenty larvae tested in the two situations, eight fish gathered in the same group of more 277 active, more exploratory and bolder individuals. Twelve fish differed in their responses depending 278 on the test. There was no morphological difference between fish that were or not in the same group. 279 For this correlation analysis, we used only the data of the maze test, because the PCA showed more 280 divergent results on the first axis (F1) for the measured traits. We found no significant correlation 281 between the PC1 scores of each larva and body length (r2=0.05, t=0.98, p=0.34), eye diameter 282 (r2=0.03, t=0.69, p=0.50), and body mass (r2=0.02, t=0.57, p=0.57). There was also, no significant 283 correlation with the age of the larvae at the moment of the test (r2=0.05, t=0.15, p=0.88). 284 12 285 Discussion 286 One of the most important results of the present study is that different behavioural traits were 287 observed in very young fish larvae using a maze and a novel object tests. These tests allowed us to 288 discriminate different groups of fish. This implies that personality traits appeared very early in the 289 life of the fish, and personality could be an inheritable character (Wright et al. 2003). For instance, a 290 simple gene mutation linked to growth factors may modify fish personality in zebra fish (Danio 291 rerio) (Norton et al. 2011). Other works suggested that a genetic basis of behavioural traits related 292 to personality could exist. For instance, the study of behavioural syndromes of 12 populations of 293 three-spined stickleback Gasterosteus aculeatus, showed that some syndromes existed in fish reared 294 in ponds where their piscivorous predators were present (Dingemanse et al. 2007), or boldness 295 varied between four wild-caught populations of zebra fish (Wright et al. 2007). For the syndrome 296 aggression-boldness there were differences across but not within populations of Zebra fish (Martins 297 and Bhat 2014), further isogenic clones clones developed different behavioural syndromes (Millot 298 et al. 2014). Domestication may also act as a selection process for personality traits (activity, 299 boldness) (Moretz et al. 2007). As we had no information on the genetic relationships between the 300 pike larvae we tested, we cannot exclude that some observed differences in personality could be the 301 consequences of genetic differences. 302 303 The behavioural tests 304 305 In our study, the device for each test had been chosen and designed to evaluate a particular 306 behavioural trait. It is why there could be some differences between individual responses to the test. 307 For example, Bell (2005) found a correlation between three behavioural traits in one population of 308 sticklebacks but not in another, tested using different devices. The maze gives us information on the 309 abilities of fish to explore a new environment, and the novel object is used for testing the stress and 13 310 the fear facing a new element in the fish surroundings (Frost et al. 2007). According to the 311 definition of the five axes of personality and their clustering in two groups proactive vs. reactive 312 (Koolhaas et al. 1999), animals that are more aggressive will also be bolder and more explorative in 313 a novel environment (Huntingford 1976). Boldness measured as the amount of time spent close to a 314 novel object, gave more reproducible results (Wright et al. 2003) and could be more adapted to the 315 study of larva. Exploration was quantified as the amount of time needed to explore the arms of a 316 maze. In this latter device, exploratory behaviour was measured as the time taken to explore and 317 swim in each arm. Thus, this device seemed more dependent on the activity and consequently on 318 the physical strength of the larvae. Consequently, the two devices (maze and novel object) did not 319 measure the same behavioural trait linked to the characteristics of the larva development. 320 321 Personality and Northern pike biology 322 323 Even if there exists a genetic background and there are some probabilities that the fish we 324 tested were not genetically equivalent, it is known that fish personality also varies according to 325 environmental factors. For instance, an increase in water temperature modifies activity, boldness 326 and aggressiveness of two damselfishes Pomacentrus moluccensis and P. bankanensis (Biro et al. 327 2010). Other environmental factors such as predation pressure (Brown et al. 2005), food 328 predictability (Chapman et al. 2010) or period of activity (Millot et al. 2009) may also influence the 329 expression of personality traits and especially boldness. Personality varies also with the social 330 context (Budaev and Brown 2011). In Eurasian perch Perca fluaviatilis, if shy fish are grouped, 331 they foraged during a longer period of time in open area (without vegetation) than when they were 332 with bold individuals (Magnhagen and Staffan 2005). Social similarity has been demonstrated in a 333 social-learning context in which sticklebacks copied the choices of demonstrator fish 14 334 proportionately more as the number of demonstrator increased (Pike and Laland 2010). The 335 experiments that we realized with northern pike larvae were done under standardized conditions for 336 temperature and light, so there was no change in the environmental conditions between the two 337 tests. Moreover all fish larvae were reared under the same social conditions, by groups of 338 individuals of the same age. The only difference introduced between them is the size and the mass 339 of the larva, which might be linked to personality (Brown and Braithwaite 2004). Our results 340 showed that in the maze test, the more active and exploratory larvae were also longer with a bigger 341 eye. For the latter, it was also observed to vary in relation to personality in sticklebacks particularly 342 with risk-taking behaviour but not with shoaling behaviour (Kim and Velando 2015). 343 Our results showed a clear separation between the two main groups of larvae, the bold and 344 shy ones; in the two tests that did not evaluate the same components of the larva personality, there 345 were some similarities (40% of fish were in the same groups) between the two rankings. Larvae 346 may be ranked also, along a bold–shy continuum (Réale et al. 2007). The degree of boldness (or 347 shyness) is determined by a trade-off between foraging gains (and/or mating opportunities) and the 348 associated risks (Wilson et al. 1993; 1994). Hence, variation in boldness is driven by the balance of 349 costs and benefits, and could be affected by metabolic rate (Krause et al. 1998), food deprivation 350 (Godin and Smith 1988) and the perception of predation risk (Coleman and Wilson 1998). Our 351 larvae were reared in the same conditions, but we cannot be sure that the individuals have the same 352 abilities to react to environmental constraints such as the introduction of exogenous food. We noted 353 that there were differences between the tests and that the majority of the fish changed of groups and 354 seemed to show different behavioural traits between the two tests. First, we found that 40% of the 355 individuals did not change groups and it was a good result compared to previous studies showing 356 less than 20 % of behavioural cross context consistency (Ferrari et al. 2014). Second, the different 357 tests did not target the same behavioural traits and larvae may differ in their behavioural responses 15 358 to different environmental constraints. This does not mean that they present different personality but 359 that, for the same behavioural trait, they showed graduate responses. This variability may be 360 correlated to age (the larvae tested at the beginning of the experiments were ten days younger than 361 the latest ones) or to small differences in development variables for individuals of the same age. 362 In conclusion, our study put forward the possibility to discriminate personality traits in 363 young fish larvae, such as the Northern pike. Such results open new opportunities for testing 364 behavioural abilities of very young fish. The existence of behavioural syndromes or personality in 365 fish larvae could be used in different situations: during the domestication process, with the selection 366 of fish with particular morphological and physiological traits, for resolving bottlenecks in 367 aquaculture (i.e. cannibalism between young larvae may limit the farming success of species of 368 economical interest; cannibalism could be linked to particular fish personality and this trait could be 369 selected as morphological or physiological traits), or for predicting the invasive abilities of new 370 species (Zhao and Feng 2015). The personality traits vary between individuals in a bold-shy 371 continuum, with a clear separation between bold and shy larvae, and in our study, most of the fish 372 changed their status between the two tests. Nevertheless, the differences are sufficient to rank the 373 larvae along a gradient of boldness-shyness. This information could be used to sort out and to select 374 fish and to study their behaviours during their development. In this case, each individual should be 375 marked individually, i.e. with a micro-tag (Ferrari et al. 2014) and followed during its development. 376 377 Acknowledgments This study was realized under the supervision of the University of Lorraine, 378 which paid for a Ms grant of Anthony Sebastian, and the experiments were conducted in the 379 building of the Museum-Aquarium of Nancy (MAN). We thank all the staff of this establishment, 380 who facilitated the study. 381 16 382 Compliance with Ethical Standards 383 Conflict of interest: the authors declare that they have no conflict of interest. 384 Ethical approval: All fish treatments and procedures used in this study were in accordance with the 385 general guidelines of the Council of European Communities (1986, No. 86/609/CEE) and the 386 French Animal Care Guidelines (Animal approval No. C54-547-18). 387 388 389 References 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 Adriaenssens B, Johnsson JI (2011) Shy trout grow faster: exploring links between personality and fitness-related traits in the wild. Behav Ecol 22, 135–143 Baras E (2012) Cannibalism in fish larvae: What have we learned? In: Larval Fish Aquaculture (Qin JG, ed.). Nova Science Publishers, New York, pp. 1-37 Bell AM (2005) Behavioural differences between individuals and two populations of stickleback(Gasterosteus aculeatus). J Evol Biol 18, 464–473 Bell AM, Sih A (2007) Exposure to predation generates personality in three-spined sticklebacks (Gasterosteus aculeatus). Ecol Lett 10, 828–834 Biro PA, Stamps JA (2008) Are animal personality traits linked to life-history productivity? Tends Ecol Evol 23, 361–368 Biro PA, Beckmann C, Stamps JA (2010) Small within-day increases in temperature affects boldness and alters personality in coral reef fish. Proc R Soc B Biol Sci 277, 71–77 Brown C, Braithwaite V (2004) Size matters: a test of boldness in eight populations of the poeciliid Brachyraphis episcopi. Anim Behav 68, 1325–1329 Brown C, Jones F, Braithwaite V (2005) In situ examination of boldness-shyness traits in the tropical poeciliid, Brachyraphis episcopi. Anim Behav 70, 1003–1009. 17 407 408 409 410 Bry C, Basset E, Rognon X, Bonamy F (1992) Analysis of sibling cannibalism among pike, Exox lucius juveniles reared under semi-natural conditions. Env Biol Fish 35, 75-84. Budaev S, Brown C (2011) Personality Traits and Behaviour. In: Fish Cognition and Behaviour. (Brown C, Laland K, Krause J, eds). Wiley-Blackwell, Oxford, pp 135–165 411 Castanheira MF, Herrera M, Costas B, Conceicao LEC, Martins CIM (2013) Can we predict 412 personality in fish? Searching for consistency over time and across contexts. PLOS ONE, 413 e.0062037 414 415 416 417 Chapman BB, Morrell LJ, Krause J (2010) Unpredictability in food supply during early life influences boldness in fish. Behav Ecol 21, 501–506 Coleman K, Wilson DS (1998) Shyness and boldness in pumpkinseed sunfish: individual differences are context-specific. Anim Behav 56, 927–936 418 Conrad JL, Weinersmith KL, Brodin T, Saltz JB, Sih A (2011) Behavioural syndromes in 419 fishes: a review with implications for ecology and fisheries management. J Fish Biol 78, 420 395–435 421 Cote J, Fogarty S, Weinersmith K, Brodin T, Sih A (2010) Personality traits and dispersal 422 tendency in the invasive mosquitofish (Gambusia affinis). Proc Roy Soc B Biol Sci 277, 423 1571–1579 424 425 426 427 Dahlbom SJ, Lagman D, Lundstedt-Enkel K, Sundström LF, Winberg S (2011) Boldness predicts social status in zebrafish (Danio rerio). PLOS ONE 6: e23565 Dall SRX, Houston AI, Mc Namara JM (2004) The behavioural ecology of personality: consistent individual differences from an adaptive perspective. Ecol Lett 7: 734-739 428 Dingemanse NJ, Wright J, Kazem AJN, Hickling r, Dawnay N (2007) Behavioural syndromes 429 differ predictability between 12 populations of three-spined stickleback. J Anim Ecol 76, 430 1128-1138 18 431 432 Dyer JRG, Croft DP, Morrell LJ, Krause J (2009) Shoal composition determines foraging success in the guppy. Behav Ecol 20, 165–171 433 Ferrari S, Benhaïm D, Colchen T, Chatain B, Begout ML(2014) First link between self-feeding 434 behaviour and personality traits in European seabass, Dicentrarchus labrax. Appl Anim 435 Behav Sci (In press) doi: 101016/j.applanim.2014.09.019 436 Frost AJ, Winrow-Giffen A, Ashley PJ, Sneddon LU (2007) Plasticity in animal personality 437 traits: does prior experience alter the degree of boldness? Proc Roy Soc B Biol Sci 274, 438 333–339 439 440 Gerlai R, Lahav M, Guo S, Rosenthal A (2000) Drinks like a fish: zebra fish (Danio rerio) as a behavior genetic model to study alcohol effects. Pharma Biochem Behav 67, 773–782 441 Goding JGJ, Smith SA (1988) A fitness cost of foraging in the guppy. Nature 333, 69-71 442 Gosling SD (2001) From mice to men: what can we learn about personality from animal 443 research? Psycho Bull 127, 45–86 444 Höjesjö J, Adriaenssens B, Bohlin T, Jönsson C, Hellström I, Johnsson JI (2011) 445 Behavioural syndromes in juvenile brown trout (Salmo trutta); life history, family 446 variation and performance in the wild. Behavioral Ecology Sociobiology 65, 1801–1810 447 Huntingford FA (1976) The relationship between anti-predator behaviour and aggression 448 among conspecifics in the three-spined stickleback, Gasterosteus aculeatus. Anim 449 Behav 24, 245–260 450 451 Kim SY, Velando A (2015) Phenotypic integration between anti-predator behaviour and camouflage pattern in juvenile sticklebacks. Evolution doi: 10.1111/evo.12600 452 Koolhaas JM, Korte SM, De Boer SF, Van Der Vegt BJ, Van Reenen CG, Hopster H, De 453 Jong IC, Ruis MAW, Blokhuis HJ (1999) Coping styles in animals: current status in, 454 behavior and stressphysiology. Neurosci Biobehav Rev 23, 925‐935 19 455 456 Kottelat Mc, Freyhof J (2007) Handbook of European freshwater fishes. Publications Kottelat, Cornol, Switzerland. 646 pp 457 Krause J, Loader SP, McDermott J, Ruxton GD (1998) Refuge use by fish as a function of 458 body-length-related metabolic expenditure and predation risk. Proc Roy Soc B Biol Sci 459 265, 2373–2379 460 461 462 463 Magnhagen C, Staffan F (2005) Is boldness affected by group composition in young-of-the-year perch (Perca fluviatilis)? Behavioral Ecology Sociobiology 57, 295–303 Martins ET, Bhat A (2014) Population level personalities in Zebra fish: aggression-boldness across but not within populations. Behavioural Ecology 25, 368-373 464 Millot S, Pean S, Labbe L, Kemeis T, Quillet E, Dupont-Nivet M, Begout ML (2014) Assessment 465 of genetic variability of fish personality traits using rainbow trout isogenis lines. Behavior 466 genetics 44, 383-393 467 Millot S, Bégout ML, Chatain B (2009) Risk-taking behaviour variation over time in sea bass 468 Dicentrarchus labrax: effects of day-night alternation, fish phenotypic characteristics and 469 selection for growth. J Fish Biol 75, 1733–1749 470 471 472 473 474 475 476 477 Moretz JA, Martins EP, Robison BD (2007) The effects of early and adult social environment on zebrafish (Danio rerio) behavior. Env Biol Fish 80, 91–101 Moretz JA, Martins EP, Robison BD (2007) Behavioral syndromes and the evolution of correlated behavior in zebrafish? Behav Ecol 556-562 Nakayama S, Harcourt JL, Johnstone RA, Manica A (2012) Initiative, personality and leadership in pairs of foraging fish. PLOS ONE 7: e36606 Ninkovic J, Bally-Cuif L (2006) The zebrafish as a model system for assessing the reinforcing properties of drugs of abuse. Methods, 3é, 262–274 20 478 Norton WHJ, Stumpenhorst K, Faus-Kessler T, Folchert A, Rohner, N, Harris MP, Callebert J, 479 Bally-Cuif L (2011) Modulation of Fgfr1a signaling in zebrafish reveals a genetic basis for 480 the aggression-boldness syndrome. J Neurosci 31, 13796–13807 481 482 483 484 485 486 487 488 Nyqvist M, Gozlan RE, Cucherousset J, Britton JR (2013) Absence of a context-general behavioural syndrome in a solitary predator. Ethology 119, 156-166 Parlier E, Corolla JP (2013) In: Doris, Esox lucius (Linnaeus, 1758), http://doris.ffessm.fr/fiche2.aspj?fiche_numero=336 Pike TW, Laland KN (2010) Conformist learning in nine-spined sticklebacks’ foraging decisions. Biol Lett 6, 466-468 Réale D, Reader SM, Sol D, McDougall PT, Dingemanse NJ (2007) Integrating animal temperament within ecology and evolution. Biol Rev Camb Philos Soc 82, 291–318 489 Rey S, Beltana S, Vargas R, Roher N, Mackenzie S (2013) Combining animal personalities with t 490 ranscriptomics resolves individual variation within a wild-type zebrah fish population and 491 identifies underpinning molecular differences in brain function. Molecular Ecology 22, 6100- 492 6115 493 494 Sih A, Bell A, Johnson JC (2004) Behavioural syndromes: an ecological and evolurionary overview. TRENDS Ecol and Evol 19, 372-378 495 Sundström LF, Petersson E, Höjesjö J, Johnsson JI, Järvi T (2004) Hatchery selection promotes 496 boldness in newly hatched brown trout (Salmo trutta): implications for dominance. Behav 497 Ecol 15, 192–198 498 Teletchea F, Fostier A, Kamler E, Gardeur JN, Le Bail PY, Jalabert B, Fontaine P (2009) 499 Comparative analysis of reproductive traits in 65 freshwater fish species: application to the 500 domestication of new fish species. Fish Biol Fisheries 19, 403-430 21 501 502 Toms CN, Echevarria DJ, Jouandot DJ (2010) A methodological review of personality-related studies in fish: focus on the shy-bold axis of behavior. Int J Comp Psychol 23, 1–25 503 Trabelsi A, Gardeur JN, Telechea F, Brun-Bellut J, Fontaine P (2013) Hatching time effect on the 504 intra-spawning larval morphology and growth in the Northern pike (Esox Lucius L.). 505 Aquaculture Research 44, 657-666 506 Westerberg M, Staffan F, Magnhagen C (2004) Influence of predation risk on individual 507 competitive ability and growth in Eurasian perch, Perca fluviatilis. Anim Behav 67, 508 273–279 509 Wilson DS, Coleman K, Clark AB, Biederman L (1993) Shy-bold continuum in pumpkinseed 510 sunfish (Lepomis gibbosus): an ecological study of a psychological trait. J Comp Psycho 511 107, 250–260 512 513 Wilson DS, Clark AB, Coleman K, Dearstyne T (1994) Shyness and boldness in human and other animals. Trends Ecol Evol 9, 442–445 514 Wright D, Rimmer LB, Pritchard VL, Butlin RK, Krause J (2003) Inter and intra-population 515 variation in shoaling and boldness in the zebrafish (Danio rerio). Naturwissenschaften 90, 516 374-377 517 Zhao DP, Feng PSH (2015) Temperature increases impacts personality traits in aquatic non-native 518 species: implications for biological invasions under climate change. Cur Zool (on line) 519 22 519 Figures legends 520 521 Figure 1: Scheme of the two devices used for the behavioural tests. Figure 1A represents the maze 522 with four arms (A1, A2, A3 and A4), the central zone (CZ) and the start zone (SZ). Figure 523 1b represents two devices (D1 and D2) used for the novel object test with the two different 524 zones: near (ZN) or far (ZF) the novel object. 525 526 Figure 2: PCA (2A) and HAC (2B) realized with seven behavioural variables and integrating two 527 morphological variables for the maze test. In the PCA, ED: Eye Diameter and BL: Body 528 length. For theother indexes see the mean in Table 1. For the HCA, the X-axis gives the ID 529 number of each larva and each has the same ID number in the two tests (see also figure 3B). 530 531 Figure 3: PCA (3A) and AHC (3B) realized with seven behavioural variables and integrating two 532 morphological variables for the novel object test. (For indexes see Figure 2 legend and Table 533 1) 534 23 534 Table 1: Variables of interest used to describe activity, exploration and boldness in the maze and 535 novel object test by Northen pike larvae. 536 537 538 Rapid swimming (RS) This variable is defined as the movement of the fish larva in less than 539 second on a distance equal to the length of the body. It stops when the fish is immobile. 540 Slow Swimming (SS) The contrary as above. We measured the time during which the larva moved 541 more than a body length during one second. 542 Immobility (IM) As the time of observation was limited (30 minutes), immobility was considered as 543 a correlative variable of swimming; when fish did not swim. We considered immobility as the time 544 spent by a larva to swim (undulations of the body) but without any displacement of the body. 545 Duration of exploratory behaviour (TT) It is the time during which, in the maze test, the larva went 546 out of the start zone and explored only the first arm; this duration comprised also returns in the start 547 zone. 548 Latency to go out the start zone (LA) In the maze test, it was the time that an individual used to go 549 out the habituation zone and venture in the first arm. Number of arms (NA): in the maze, it was the 550 total number of arms that the larva visited during the period of observation. 551 Number of zones (NZ) In novel object test, it was the total number of zones that the larva visited 552 during the period of observation. It could concern the labyrinth (NZ1) or novel object (NZ2). 553 Time spent near the novel object (TO) It was the total time during which the larva was in contact or 554 at less than one centimetre from the novel object. On the opposite, for from the novel object (FO). 555 556 557 558 559 24 one 559 Table 2: Morphological and behavioural data used in the two tests (Mean, Standard deviation and 560 coefficient of variation (CV) (n= 20 individuals). 561 562 563 Variables Mean SD 564 Total body length (mm) 16.13 0.37 10 565 Eye diameter (mm) 1.23 0.03 10 566 Body Mass (mg) 19.4 1.84 43 CV (%) 567 568 Behaviour in SS (s) 107.4 12.8 53 569 Maze test RS (s) 3.9 1.6 187 570 IM (s) 25.0 3.3 60 571 TT (s) 343.9 55.3 72 572 LA (s) 164.3 43.6 118 573 NZ1 6 1 56 574 NA 2 0 65 24.7 61 3.8 206 575 576 Behaviour in SS (s) 179.6 577 Novel object test RS (s) 8.2 578 IM (s) 20.5 3.7 80 579 TO (s) 218.6 30.4 62 580 NZ 2 8 2 99 581 582 583 584 25 584 585 586 587 588 589 590 Pasquet et al. Figure 1A 591 592 26 592 593 594 595 596 597 Pasquet et al. Figure 1B 598 599 27 599 600 601 602 603 Figure 2 A) 604 605 606 607 608 609 610 611 612 613 614 28 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 B) 633 634 635 636 637 638 639 29 639 640 641 642 643 644 Figure 3 A) 645 646 647 648 649 650 651 652 653 654 655 656 657 658 30 659 660 661 662 663 664 665 666 667 668 B) 669 670 31