This is the accepted version of the article: Feldman, Mariano; Ferrandiz Rovira, Mariona; Espelta Morral, Josep Maria; [et al.]. «Evidence of high individual variability in seed management by scatter- hoarding rodents : does ‘personality’ matter?». Animal behaviour, Vol. 150 (April 2019), p. 167-174. This version is avaible at https://ddd.uab.cat/record/215510 under the terms of the license 1 Evidence of high individual variability in seed 2 management by scatter-hoarding rodents: Does 3 ‘personality’ matter? 4 5 Mariano Feldman1, Mariona Ferrandiz-Rovira1,2, Josep Maria Espelta1 6 & Alberto Muñoz3, * 7 1 8 CREAF, Cerdanyola del Vallès 08193, Catalonia, Spain 2 9 Universitat Autònoma de Barcelona, Cerdanyola del Vallès 08193, Catalonia, Spain 3 10 Departamento de Didáctica de las Ciencias Experimentales, Sociales y Matemáticas. 11 Facultad de Educación. Universidad Complutense de Madrid. Madrid. Spain. 12 13 *Corresponding author: Departamento de Didáctica de las Ciencias Experimentales, 14 Sociales y Matemáticas. Facultad de Educación. Universidad Complutense de Madrid. 15 Avda. Rector Royo Villanova s/n 28040. Madrid. Spain. 16 17 18 Abstract.- The predation and dispersal of seeds by scatter-hoarding animals is one of the most 19 studied processes in the context of animal-plant interactions. The seed management by these 20 animals has been traditionally approached at the population level, assuming that the patterns 21 documented in field are similar among all individuals of the population, and considering the 22 variability within the population as random noise. However, little is known about to what extent 23 this variability responds to different and consistent behavioural displays among individuals. The 24 aim of this study was to analyse the individual variation and consistency in behaviour by 25 scatter-hoarding rodents within a population. As model we used the wood mouse (Apodemus 26 sylvaticus), a key disperser of holm oak acorns (Quercus ilex) that, in turn, suffers a high 27 predation pressure by the common genet (Genetta genetta). In two sets of laboratory 28 experiments, we compared the variance and consistency in behavioural displays and acorn 29 managing generated by the individual differences with that generated by the manipulation of the 30 perceived predation risk using scents of genets. Genet scents reduced the activity (i.e. time out 31 of the refuge) in all rodents, but the differences and the consistency in activity among 32 individuals accounted for the 82.5% of total variance. Also, rodents showed different and 33 consistent displays of stressed or relaxed behaviours. More than 87% of variance in seed 34 managing variables, like dispersal distance and seed size selection, was explained by consistent 35 differences among individuals across scent treatments. The increase of stressed behaviours and 36 decrease of relaxed ones were positively related with the dispersal ability (i.e. longer distances 37 and larger acorns). Our study highlights the importance of considering the individual component 38 of behaviour in scatter-hoarding rodents. This fine-scale level, largely overlooked in the 39 ecological framework, will help to increase our understanding on seed management by scatter- 40 hoarding animals. 41 Keywords: individual variation; personality; predation risk; scatter-hoarding rodent; seed 42 dispersal. 43 1 2 2 44 Plants and animals have coevolved in many complex interactions, ranging from 45 antagonisms, such as herbivory, to mutualisms, such as seed dispersal (Labandeira, 2002). The 46 outcome of these interactions depends on many factors from both sides, and can be approached 47 under different perspectives (e.g. evolutionary, ecological, and behavioural). Seed management 48 by scatter-hoarding animals is one of the most studied processes in the context of animal-plant 49 interactions, because of the dual behaviour displayed by these animals (Herrera & Pellmyr, 50 2002; Vellend et al., 2006; Schupp et al., 2010). On the one hand, they consume seeds, having a 51 negative impact on plant populations, but they can also move and store seeds in certain sites that 52 may favour seed dispersal and seedling recruitment, thus having a positive effect in plant 53 recruitment (Vander Wall, 1990; Herrera, 2002; Gomez et al., 2008; Muñoz & Bonal, 2011; 54 Sunyer et al., 2015). This dual role and its outcome for plants has an important behavioural 55 dimension, as it depends on individual decisions of scatter-hoarding animals (Muñoz & Bonal, 56 2008a, 2008b; Perea et al., 2011; Sunyer et al., 2013, 2015). Understanding the nature of these 57 decisions may help to shed light on seed fate and ultimately into plant populations and 58 community dynamics (Herrera & Pellmyr, 2002; Vellend et al., 2006; Schupp et al., 2010). 59 Moreover, the study of seed dispersal has increased in recent years, probably due to the rising 60 concern about the survival of numerous plant species in the context of global change (Robledo- 61 Arnuncio et al., 2014). 62 Many factors are known to influence animal choices through the decision-making 63 process. For example, the seed choices of scatter-hoarding rodents are influenced by physical 64 characteristics of food, such as seed size and shape (Bonfil, 1998; Gómez, 2004; Preston & 65 Jacobs, 2009; Muñoz et al., 2012; Sunyer et al., 2015). Most studies have documented that 66 dispersers prefer larger seeds because of their higher nutrient content (Jansen et al., 2004; Xiao 67 et al., 2004; Muñoz & Bonal, 2008a), but their choices may also depend on the costs of handling 68 and moving the larger seeds (Kerley & Erasmus, 1991; Muñoz & Bonal, 2008b, Muñoz et al., 69 2012). The cost of handling and carrying seeds can be particularly high for scatter-hoarding 70 rodents, as they are usually prey of a huge variety of terrestrial and aerial predators, so that the 71 perceived predation risk is known to influence their seed preferences and management (Leaver, 3 3 4 72 2004; Sunyer et al., 2013; Navarro-Castilla & Barja, 2014; Heinen-Kay et al., 2016). Also, the 73 perceived risk of pilferage by conspecifics can influence seed management by scatter-hoarding 74 rodents (Hopewell & Leaver, 2008; Steele et al., 2008; Muñoz & Bonal, 2011; Samson & 75 Manser, 2016) 76 The patterns of seed predation and dispersal by scatter-hoarding rodents have been 77 traditionally approached at species or population level (e.g. Hollander & Van der Wall, 2004; 78 Muñoz & Bonal, 2007; Schupp et al., 2010; Perea et al., 2011; Xiao et al., 2013). That is, the 79 responses of rodents to seed characteristics and environmental factors are often assumed to be 80 the same among all individuals within a given species or population. Thus, a lot of studies have 81 focused on the environmental causes of behaviour (i.e. exogenous), such as seed traits, mast 82 seeding, pilferage and predation risk, etc., providing means and deviances for variables of seed 83 management like predation / dispersal rates, seed size, dispersal distances, cache management, 84 for different species and populations worldwide (Muñoz & Bonal, 2007; Gómez et al., 2008; 85 Schupp et al., 2010; Perea et al., 2011; Xiao et al., 2013). These kind of studies assumes the 86 behavioural variability among individuals within the population or species as random noise; 87 however, little is known about to what extent these deviations are random or respond to 88 different and consistent patterns of behaviour among individuals within the same population 89 (i.e. endogenous causes of behaviour). Probably, one of the main reasons of this lack of 90 information has been the methodological difficulties to focus in detail on the behavioural 91 displays and seed management in field at the individual level, because scatter-hoarding rodents 92 are usually small, nocturnal and very sensitive to potential predators. 93 In recent years, the study of animal personality has become a hot topic under an 94 evolutionary and ecological framework (Réale et al., 2007, 2010; Wolf et al., 2007; Stamps & 95 Groothuis, 2010; Wolf & Weissing, 2012; Carter et al., 2013; Moran et al., 2016; Pennisi, 96 2016). The idea that different individuals of the same population show different, but consistent, 97 behavioural patterns is appealing, especially for scatter-hoarding species in the context of 98 animal-plant interactions. To the best of our knowledge, no study has analyzed in detail the 99 variation in individual behaviour and consistency in seed management by scatter-hoarding 5 4 6 100 rodents. Thus, the aim of this study is to analyse the extent of this individual variation and the 101 consistency in behaviour by scatter-hoarding rodents within a population, in order to discuss the 102 role of individual patterns in the observed patterns of seed management (i.e. predation and 103 dispersal of seeds). We specifically assessed how the behavioural variance generated by the 104 individual component was comparatively higher than that generated by an important 105 environmental factor like the perceived predation risk. 106 As a study model we have used a well-documented plant-disperser-predator 107 assemblage: holm oak (Quercus ilex) acorns, which are consumed and dispersed by wood 108 mouse (Apodemus sylvaticus), that in turn is the main prey of the common genet (Genetta 109 genetta). The wood mouse is the most abundant scatter-hoarding rodent in Mediterranean areas 110 of southern Europe, distributed from the Iberian Peninsula to southwestern Asia and the 111 Himalayas, and from northwestern Africa to the south of Scandinavia (Torre et al., 2002; 112 Urgoiti et al., 2018). This nocturnal and small rodent (15-35g) is a prominent consumer and 113 disperser of acorns during the seeding season in autumn, influencing the recruitment dynamics 114 of many oak species (den Ouden et al., 2005; Sunyer et al., 2015), like the holm oak, which is 115 the most widespread oak in the Mediterranean basin (Blondel & Aronson, 1999). This oak 116 shows an extraordinarily variability in acorn size (from less than 1 g to more than 15g, Muñoz 117 & Bonal, 2008b). In Mediterranean areas, the wood mouse is the most abundant prey (up to 118 86.9% of biomass diet) of the common genet, Genetta genetta, a common nocturnal carnivore 119 (Virgós et al., 1999; Torre et al., 2003). In fact, the wood mouse is able to detect and respond to 120 the scents of genets in order to reduce the predation risk when foraging acorns in autumn 121 (Sunyer et al., 2013). Using this oak-rodent-carnivore system, we performed two series of 122 laboratory experiments with a wild population of wood mice: one to characterise and analyse in 123 detail the individual behaviour, and other to assess individual seed management of holm oak 124 acorns. In both sets of experiments, we manipulated experimentally the perceived predation 125 risk, using genet scents, to assess its effects on rodent behaviour. We hypothesised that, if the 126 individual component of behaviour was relevant, rodents would display different, but 7 5 8 127 consistent, individual behaviours and patterns of seed management, even under different levels 128 of perceived predation risk. 129 130 131 METHODS 132 Capture and Maintenance of Rodents 133 We captured 25 adult wood mice (mean weight ± SE: 19.5g ± 3.4, range: 12-26g) in the 134 forest of Can Balasc (Collserola Natural Park; 41° 24′ N, 2° 6′ E, Barcelona, Spain) a natural 135 reserve representative of Mediterranean oak forests dominated by the holm oak (95% of the 136 forest area according to Espelta et al., 2009). In this area, the rodent community is mainly 137 composed of wood mice (99%), which share the habitat with the common genet (Sunyer et al., 138 2013). Rodents were captured from March to April 2017 in a surface of 2 ha using Sherman 139 live-traps (23.5 × 8 × 9 cm; HB Sherman Traps Inc., Tallahassee, Florida USA) baited with a 140 mixture of flour and tuna in oil and a piece of apple for hydration. Additionally, they contained 141 a handful of hydrophobic cotton so that the captured rodents could make a nest to remain 142 protected (Muñoz et al., 2009; Sunyer et al., 2014). 143 Captured rodents were carried out to the laboratories of the Can Balasc field research 144 station in their provisional nests made inside the traps. All captured rodents were then weighed 145 and housed individually indoors in terraria (30 x 30 x 35 cm) filled with a layer of sand of 5 cm 146 deep to provide traction for rodent stepping and facilitate the movement. We also included a 147 refuge (10 x 16 x 8 cm) with a piece of waterproof cotton to mimic their natural nests made in 148 field burrows in order to feel safe and protected. These home-terraria were placed indoors under 149 natural photoperiodicity and no visual contact among individuals to avoid stress. The home- 150 terraria were not open, but had a cover with several slits to provide ventilation, in order to 151 reduce the probability that experimental rodents could detect scents from other rodents. We 152 provided apple and hamster food as food source (Vitacraft Premium Menu), a high quality 9 6 10 153 natural food, made with seeds, cereals, and greens, that offers a complete daily menu for 154 rodents. Rodents were kept in the laboratories on average for 18 days, including habituation and 155 trials (range 15-22). 156 157 Experimental Design 158 After 3 days for habituation to the individual home-terraria, we performed two types of 159 experiments with all captured rodents. Experiment one was conceived to characterise the 160 individual behaviour and experiment two to analyse individual acorn management. 161 162 Characterisation of individual behavior experiments 163 This experiment consisted of three trials per individual in order to analyse the 164 differences in behaviour among individuals and its consistency within each individual. These 165 experiments consisted of recording the activity and detailed behaviour of each individual with 166 nocturnal video cameras in its home-terraria during three nights (i.e. trials) under different scent 167 treatments: (1) a control trial, (2) a predator scent trial and (3) a control after predator scent trial. 168 We attached to the ceiling of each home-terraria three cotton discs (5.5 cm) that were 169 soaked with the stimulus according to the scent treatment: distilled water for the two control 170 trials, and genets’ scent for the predator trial. Genets’ scent was obtained from 10 g of fresh 171 feces collected in latrines located in the study area, which were thawed and mixed with 30 ml of 172 water to obtain an homogeneous mixture which serves rodents as a cue of predator presence 173 (Sunyer et al., 2013). Genet feces are indeed a powerful source of chemical signals for other 174 animals, as the scents secreted by perineal glands are mixed with feces and deposited in latrines, 175 which serve as stations for scent communication, playing a role in territoriality, sexual 176 attraction, warning, etc. (Espirito-Santo et al. 2007). Cottons were placed at 19:00 GMT and we 177 set-up one night-vision cameras for each individual, which recorded continuously during 10h 178 (from 21:00 to 07:00 GMT) under natural photoperiodic conditions. Before each trial, we 11 7 12 179 removed all the remaining food of the home-terrarium (checking especially the nest) except a 180 piece of apple that was placed outside the nest with plastic gloves. 181 Based on the recordings obtained from the three trials, we analysed in detail the 182 behaviour of each experimental rodent. We displayed the video recordings in Microsoft 183 Windows Media Player (Microsoft®) in slow motion (x 0.5) to ensure an accurate measurement 184 of behavioural variables. For each trial, we first measured the ‘total activity’ as the percent of 185 the total time spent out of the refuge, which is usually taken as part of the boldness score of the 186 individuals (Carter et al., 2013; Mamuneas et al., 2015; Yuen et al., 2015). Then, we analysed in 187 detail a sample of 25 min randomly taken from the total activity time spent out of the refuge for 188 each individual and trial. This sample was used to categorise the behaviours displayed by each 189 individual as ‘stressed’ or ‘relaxed’. Stressed behaviours included ‘vigilance and freezing’ 190 (standing still on their back feet or remaining completely immobile), ‘trying to escape’ 191 (jumping, clambering or climbing), and ‘taking and shatter the cotton discs soaked with the 192 stimulus’. Relaxed behaviours included ‘sniffing’, ‘feeding or handling the piece of apple’, 193 ‘self-grooming’ (cleaning itself) and ‘burrowing or digging’ (Apfelbach et al., 2005; Sunyer et 194 al., 2013). Neutral behaviours, such as moving slowly along the terraria, were not included as 195 stressed or relaxed. Once categorised, the proportion of relaxed and stressed behaviors was 196 calculated for each individual and trial. 197 198 Seed Management Experiments 199 The seed management experiments consisted of two consecutive trials per individual 200 conducted in indoor arenas (134 x 92 x 75 cm): in the first experiment water was sprayed along 201 the arena (i.e. control treatment) and the second one the scents of genets were sprayed along the 202 arena (i.e. predator treatment). In each trial, each rodent was provided with 6 holm oak acorns 203 that were classified in 3 categories of size, in order to analyse the effects of acorn size on rodent 204 preferences: 2 small acorns (< 3 g), 2 medium acorns (3-4.5 g) and 2 large acorns (> 4.5 g). 13 8 14 205 Fresh acorns were collected from oaks during January 2017 and maintained at 4ºC until the 206 experiments. In one corner of the arena we placed the own nest of the experimental rodent 207 (taken from the home terrarium) and in the opposite corner the 6 acorns randomly located in a 208 2×3 grid with a distance of 5 cm among them. 209 Two hours before each trial, we removed the remaining food from the nest with plastic 210 gloves, to ensure that experimental acorns were the only source of food available. Rodents and 211 experimental acorns were weighed before and after each trial (to the nearest 0.01 g). Each wood 212 mouse was placed in the arena at 17:00 GMT only with its home refuge in order to leave some 213 time for conditioning. Then, we positioned the acorns and sprayed the stimulus treatment at 214 18:30 GMT. Acorns were manipulated using fresh gloves to avoid effects of human odour cues 215 on rodent choices (Wenny, 2002). After each trial, the arenas were cleaned thoroughly to avoid 216 scent contamination among trials. 217 Digital video cameras with night vision were installed over the arenas and in each trial 218 we video recorded for 13 h (between 19:00 and 08:00 GMT) to monitor the behaviour of each 219 individual during the trial. Experimental acorns were revised at 08:00 GMT, noting the distance 220 moved and whether each acorn had been partially or completely predated. We analysed the 221 video recordings of seed management in Microsoft Windows Media Player (Microsoft®) in 222 slow motion (x 0.5) and also calculated the ‘total activity’ as the percent of the time spent out of 223 the nest with respect to the total time. 224 225 Data Analysis 226 In the behavioural characterisation experiments, we assessed the effects of the predator scent 227 treatment on rodent behaviour using repeated-measures ANOVAs, with total activity (i.e. 228 percentage of time out of the refuge) and the proportion of stressed-relaxed behaviours as 229 dependent variables. We also checked the differences between the two control trials (previous 230 and post-genet scents) to assess whether the predator effects were fixed or not in rodents after 15 9 16 231 the predator scent treatment. We analysed the differences in behaviour among individuals, and 232 the consistency within individuals across treatments, with a model that estimated the 233 components of variance, with the “individual rodent” as random factor and “scent treatment” 234 (i.e. control, predator and control after predator) as fixed factor. These analyses provide the 235 proportion of behavioural variance explained by each factor. We used regression models to 236 analyse the potential effects of rodent weight on the individual variance in behavioural patterns. 237 In the seed managing experiments, we also analysed the effects of ‘individual’ and scent 238 treatment’ using a model that estimated the components of variance, with the “individual 239 rodent” as random factor and “scent treatment” as fixed factor. In this case, the dependent 240 variables were: ‘activity’ (percentage of time spent outside the refuge), ‘dispersal distance’ (∑ 241 distances of seeds moved during the trial), ‘seed size’ (mean weight of the seeds moved during 242 the trial), ‘ratio’ (mean weight of the seeds moved during the trial divided by the weight of the 243 experimental rodent, see Bonal & Muñoz, 2008b), ‘dispersal effort’ (∑weight of seeds moved * 244 distance moved). We used regression models to assess whether the activity of each rodent in the 245 seed managing trials was consistent with the activity recorded in the behaviour characterisation 246 trials, and to what extent seed management was related to the stressed and relaxed behaviours. 247 The comparisons between sexes were not possible given that we just captured 4 females out of 248 the 25 experimental rodents. However, we have previously demonstrated that the sex of wood 249 mice has no effects on seed management (Muñoz & Bonal 2008b). 250 251 Ethical Note 252 In this research, we captured 25 adult wood mice (Apodemus sylvaticus) using Sherman live- 253 traps (23.5 × 8 × 9 cm; HB Sherman Traps Inc., Tallahassee, Florida USA) in the Collserola 254 Natural Park (41° 24′ N, 2° 6′ E, Spain) that were used in the experiments of behaviour. We also 255 captured one pregnant female and four juveniles, but they were immediately released. Traps 256 were baited with a piece of apple and a handful of hydrophobic cotton. These traps do not 257 produce any injury to rodents and allow them to make a nest inside with the piece of cotton 17 10 18 258 provided, which helps rodents to feel safe and relaxed during the capture period (Muñoz et al., 259 2009; Sunyer et al., 2013, 2016). The piece of apple (ca. 10g) ensures enough individual 260 hydration during the capture period, as the piece is never totally consumed. Traps were daily 261 checked at dawn, so that rodents were inside the traps just 6-8 hours at most. No lactating 262 females were caught during the trapping sessions. Capture, handle and maintenance of wood 263 mice authorization was issued by the Generalitat de Catalunya (Departament de Territori i 264 Sostenibilitat; reference SF/156) after approval by the advisory committee of the Collserola 265 Natural Park authorities. All the handling and sampling were done by M.F-R., who holds the 266 EU permit for experimentation with animals authorised by the French Ministry of Agriculture 267 and Fisheries (authorization reference R45GRETAF110). Wood mice capture and management 268 in the laboratory adhered to the ASAB/ABS Guidelines for the Use of Animals in Research. 269 During this research no rodent resulted injured and all were healthy until the experiments 270 finished, when rodents were released at the exact point of capture. 271 272 273 RESULTS 274 Characterisation of Individual Behaviour 275 Predator scent reduced significantly the activity (i.e. proportion of time outside the 276 refuge) of rodents from the control to the predator scent trials (repeated-measures ANOVA: F1,22 277 = 4.83, P = 0.038). Yet, the activity did not differ between the two control trials (before and 278 after the trial of predator scents; repeated-measures ANOVA F1,22 = 0.04, P = 0.85). Hence, 279 experimental rodents adjusted their activity (i.e. time exposed to predators) to the predation risk 280 perceived in each moment, lacking “preventive behavior” after a contact with predator scents. 281 More interestingly, the time expended by rodents out of the refuge was extremely variable 282 among rodents, from 42 min to 5 h and 23 min (out of 9 h of video recordings per individual), 19 11 20 283 and this time was extraordinarily repeatable within the same rodent across different trials (Table 284 1). 285 286 Table 1. Behavioural patterns of experimental rodents in the characterisation trials. % time outside the refuge % stressed displays % relaxed displays (n=23) (n=23) (n=23) Trial Mean ± SE Range Mean ± SE Range Mean ± SE Range Control 25.8 ± 2.7 10.7 – 59.2 23.2 ± 3.5 2.9 – 56.5 46.0 ± 4.8 6.0 – 79.8 Predator scent 22.9 ± 2.3 7.7 – 58.1 24.8 ± 3.2 6.0 – 57.9 44.6 ± 3.8 9.8 – 72.1 Control 2 25.5 ± 2.9 10.7 – 59.3 31.7 ± 4.2 6.3 – 75.7 40.6 ± 3.8 6.9 – 77.0 287 Mean, Standard Error (SE) and range of the behavioural displays (proportion of time expend by 288 experimental individuals out of the refuge and proportion of stressed and relaxed behaviours) 289 displayed during the three behaviour characterisation trials: control, predator scent and control 290 after predator scent (control 2) 291 292 The model for estimating the components of variance showed that the individual 293 differences were highly significant (factor individual: F22,44 = 15.16, P < 0.0001; Figure 1), and 294 explained the 82.5% of total variance in activity out of the refuge, a high variance compared 295 with that generated by the scent treatment (Table 1). These different and fixed patterns of 296 activity among rodents were not related to rodent weight in any of the scent treatments (Control: 297 β = -0.21, F1,21 = 0.95, P = 0.34; Predator: β = -0.21, F1,21 = 0.98, P = 0.33; Control after 298 predator: β = -0.08, F1,21 = 0.14, P = 0.71). 21 12 22 70 Activity out of the refuge (% of total time) 60 50 40 30 20 10 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Individual 299 300 Figure 1. Differences among rodents in the proportion of time expend out of the refuge (of the 301 total recorded) during the behaviour characterisation trials across the three scent treatments. 302 Mean (points) and SE (lines) for the pooled data of the three trials performed by each 303 individual. 304 305 The proportion of relaxed and stressed behaviours did not correlate with the time 306 expended by individuals outside the refuge in any scent treatment (P > 0.3 in all cases). Indeed, 307 although predator scents reduced individual’s activity (i.e. the time exposed out of the refuge), 308 they did not change the type of behavior displayed once individuals decided to leave the shelter, 309 so that the proportion of stressed and relaxed behaviors did not differ significantly between the 310 control and predator experiments (repeated-measures ANOVA: stress F1,22 = 0.36, P = 0.55; 311 relax F1,22 = 0.17, P = 0.68, Table 1). However, the variance in proportion of stressed or relaxed 312 behaviours was mainly explained by individual differences and it was very consistent for each 313 individual across treatments, as in the case of the activity time. The factor individual explained 314 52.3 % of the total variance in stressed behaviors and 42.2% in relaxed behaviors (individual 315 effect: stress: F22,44 = 4.28, P < 0.001; relax F22,44 = 3.19, P < 0.001; Figure 2). The proportion of 23 13 24 316 stressed and relaxed behaviours were inversely correlated (β = -0.75, F1,67 = 88.42, P < 0.0001), 317 and these behaviors did not correlate with rodent size in any of the three scent treatments ( P > 318 0.2 in all models). Thus, the behaviour characterisation trials revealed that rodent behaviour 319 shows a strong individual component, accounting for most behavioural variation in the 320 population of experimental individuals as compared to the variance generated by the scent 321 environment (i.e. perceived predation risk). 80 (a) 70 Proportion of stressed behaviours 60 50 40 30 20 10 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 322 80 (b) 70 Proportion of relaxed behaviours 60 50 40 30 20 10 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Individual 323 324 Figure 2. Differences among rodents in the proportion of different type of behaviour -stressed 325 (a) and relaxed (b)- displayed out of the refuge during activity trials. Mean (squares) and SE 326 (lines) for the pooled data of the three trials performed by each rodent. 25 14 26 327 328 Seed Managing 329 The activity of each individual in the arenas during the seed managing trials (i.e. 330 proportion of time active) was positively correlated with the time out of the refuge measured in 331 the behaviour characterisation trials (i.e. home-terraria; β = 0.52, F1,16 = 6.09, P = 0.025). The 332 variance among individuals in the patterns of seed management and the consistency within each 333 individual between scent treatments were extraordinarily high, so that most variance in seed 334 management was explained by inter-individual differences, and not by the presence of predator 335 scents (Table 2). 336 337 Table 2. Repeatability and individual variability among experimental rodents in seed 338 management. β t d.f. P Var. indiv. Predator effect Activity (1) 0.84 6.63 18 <0.001 87.5 % F1,19= 0.39, P=0.53 Seed size (2) 0.39 2.05 23 0.05 33.0 % F1,24= 0.04, P=0.84 Ratio (3) 0.66 4.24 23 <0.001 66.3 % F1,24= 0.26, P=0.61 (4) Disp. Dist. 0.78 5.78 23 <0.001 76.7 % F1,24= 1.03, P=0.32 Disp. Effort (5) 0.81 6.59 23 <0.001 80.7 % F1,24= 0.74, P=0.39 339 Correlations between the two seed management trials (control vs. predator scent) in the 340 variables (1) activity time in the arenas (2) size of seeds dispersed (3) size of seeds 341 dispersed/rodent size (ratio) (4) dispersal distance and (5) dispersal effort (see data analyses). 342 The proportion of variance explained by the factor “individual” (Var. indiv.) and the predator 343 scent effects are showed. 344 345 The activity time measured in the behaviour characterisation trials was not related to the 346 variables of seed managing, but the proportion of stressed and relaxed behaviours was 347 significantly related to different variables of seed management (Table 3). The increase of 348 stressed behaviours and the decrease of relaxed behaviours were positively related to the 349 dispersal effort (Table 3). This was especially relevant for the dispersal distances of seeds which 350 significantly increased with the proportion of stressed displays and significantly decreased with 351 the proportion of relaxed displays (Table 3). Also, the size of seeds moved decreased 27 15 28 352 significantly with the proportion of relaxed displays so that more relaxed individuals preferred 353 smaller seeds (Table 3). 354 355 Table 3. Relationship between behavioural displays and seed management. Activity time Stressed displays Relaxed displays β t44 P β t44 P β t44 P Dispersal effort -0.05 -0.32 0.74 0.37 2.66 0.010 -0.39 -2.80 <0.01 Disp. Distance -0.06 -0.40 0.69 0.38 2.73 <0.01 -0.32 -2.21 0.03 Seed size 0.11 0.73 0.47 0.17 1.14 0.26 -0.30 -2.06 0.04 356 Relationships between the behavioural displays measured in the behaviour characterisation 357 trials (activity time, proportion of stressed and relaxed displays) and the variables measured in 358 the seed managing trials (Dispersal effort, dispersal distance and size of the seeds moved). 359 360 DISCUSSION 361 We found strong and consistent inter-individual differences in the behavioural patterns of 362 a scatter-hoarding rodent species, so that the differences among individuals account for most 363 behavioural variance compared to the variance generated by the experimental manipulation of 364 environment using scents of genets (i.e. perceived predation risk). We found an effect of the 365 predator scents on rodent activity during the behavioural characterisation trials, suggesting that 366 the wood mouse is able to detect and respond to genet scents by reducing their activity, as has 367 been previously reported for other species (Grabowski & Kimbro, 2005; Verdolin, 2006; Jolles 368 et al., 2015, see Sunyer et al., 2013 for the wood mouse). However, this reduction in activity 369 was negligible in comparison with the consistent variability in the activity among individuals 370 (more than 80% of total variance). These results suggest a prevalence of the endogenous causes 371 of behaviour (i.e. individual) over the exogenous ones (i.e. environmental). The patterns of seed 372 management were also extraordinarily variable among individuals and very repeatable within 373 the same individual across the scent treatment. For example, more than 75% of variation in seed 29 16 30 374 dispersal distances, and more than 80% of variation in dispersal effort (∑weight of seeds moved 375 * distance moved) were explained by consistent differences among individuals. Overall, more 376 than 87% of variance in activity in seed managing experiments was explained by consistent 377 individual differences. Hence, basic foraging decisions like ‘what seed to eat or to move’, or 378 ‘how far to move seeds’, probably have a strong individual component. This component seems 379 to be related with the differences in the proportion of stressed/relaxed behaviours observed 380 among experimental rodents, so that more stressed individuals showed a better ability for 381 moving seeds (with longer dispersal distances and higher dispersal effort) than relaxed ones. 382 This suggests that the level of stress increases the foraging investment and performance of 383 scatter-hoarding rodents at the early stages of seed choice and seed movement, allowing them to 384 hoard larger seeds further away from the seed source to prevent pilferage (Muñoz & Bonal, 385 2011). The level of relax, by contrast, seems to reduce the foraging efficiency by enhancing the 386 movement of seeds at shorter distances, and to increase the movement of smaller seeds, which 387 have a lower nutrient value (Kerley & Erasmus, 1991), and that ultimately represent a lower 388 dispersal effort (Muñoz & Bonal 2008b). These results agree with previous studies suggesting 389 that the stress level may enhance foraging efficiency in small rodents (Chaby et al., 2015). The 390 specific mechanisms by which stress level affects foraging are not clear. One potential 391 explanation is that stressed individuals are probably more familiarised with the effects of the 392 stress response, and so they can function in a high arousal state more easily (Natelson et al., 393 1988), allowing them to increase the foraging effort and performance when facing seeds. 394 The personality of animals is currently an appealing research field under an evolutionary 395 and ecological scenario (Réale et al., 2010; Stamps & Groothuis, 2010; Wolf & Weissing, 2012; 396 Carter et al., 2013; Pennisi 2016) since, in the last decades, a lot of studies have pointed to the 397 importance of considering consistent behaviours through time and conditions at individual level 398 (Verbeek et al., 1996; Hayes & Jenkins, 1997; Koolhaas et al., 1999; Sneddon, 2003; Muñoz & 399 Bonal, 2008a; Korpela et al., 2011). The concept of personality is complex and may be ruled by 400 genetic characteristics, physiological processes, environmental experiences, and their interaction 31 17 32 401 (Bell, 2005; Nussey et al., 2007; Sih & Bell, 2008; Dosmann et al., 2015). Our experiments do 402 not allow disentangling to what extent the individual differences respond to different individual 403 experiences in the field or, by contrast, they have a more innate basis (e.g. genetic, 404 physiological). However, it is known that cognitive processes, such as learning or memory, are 405 exceptionally well developed in scatter-hoarding rodents (Pan et al., 2013), and that they use 406 their own experiences to adjust individual foraging decisions like seed management (Muñoz & 407 Bonal, 2008a; Yi et al., 2016). In scatter-hoarders, these cognitive abilities may be evolutionary 408 advantageous in order to improve the recovery of cached seeds or to avoid cache pilferage 409 (Muñoz & Bonal, 2011; Yi et al., 2016), and might also promote the different strategies and the 410 consistent individual patterns of behaviour we have found in the current study. 411 The study of the patterns of seed predation and dispersal by scatter-hoarding rodents has 412 been traditionally approached from a population scale. That is because experimental designs 413 usually consist of marking and placing seeds in field plots in order to monitor seed fate after 414 predation or dispersal, with no possibility of controlling the number and identity of scatter- 415 hoarders visiting each seed plot. Using such methodological approach, most studies have tested 416 how the patterns of predation and dispersal of a given rodent species or population are 417 influenced by different environmental factors, such as mast seeding (Wang et al., 2017), seed 418 size (Yi & Wang, 2015), seed species (in A. sylvaticus, Sunyer et al., 2014), predation risk 419 (Leaver, 2004; Sunyer et al., 2013) or pilferage risk (Steele et al., 2008; Samson & Manser, 420 2016). The variability in the patterns of seed management not explained by these environmental 421 factors have been usually interpreted as random ‘noise’ (see Wilson, 1998; Dall et al., 2004). 422 However, part of this apparent noise might be explained by variations in behaviour among 423 individuals documented in our study (see Sih et al., 2004; Sih & Bell, 2008; Dingmanse et al., 424 2009; Korpela et al., 2011). This fine-scale level, largely overlooked in the ecological 425 framework, would help to increase our understanding on the spatial and temporal variability of 426 the patterns of seed management by seed-dispersing rodents, and may also increase the 427 explanatory power of the ecological models. For example, the fate of seeds in a given seed plot 33 18 34 428 may change depending on the number and identity of the rodents visiting the seed plot. 429 Tracking individuals in field conditions represents, however, a methodological challenge, 430 because scatter-hoarding rodents are usually small, nocturnal and inhabit underground burrows 431 making very difficult to monitor the individual component of behaviour (see Gu et al., 2017; 432 Lichti et al., 2017). 433 We still know little about the nature of the individual variability and potential 434 personalities of scatter-hoarding rodents, so that further studies are needed to undergo on the 435 role of behavioural genetics, physiology and, for example, the consistency of behavioural 436 patterns across generations, or to what extent the environment can influence the potential 437 personalities. Long-term studies carried out in field conditions can shed light on these issues; 438 although in the case of the wood mouse we have found that individuals show a life-span of just 439 a few months in field conditions (Sunyer et al., 2016). On other hand, the ecological 440 consequences of the individual behaviour are also an appealing research field. Our study 441 suggests that the high individual variability of wood mouse in moving seeds can have 442 consequences for seed dispersal in the holm oak (e.g. influencing the variability of acorn 443 dispersal distances and acorn size selection). However, further field studies carried out at wider 444 spatial scales are needed to determine the current role of individual patterns of scatter-hoarding 445 rodents on seed dispersal and plant recruitment. 446 447 Acknowledgements 448 We are grateful to Rafael Hidalgo, Sara Garcia and Eric Lopez involved in catching and 449 manipulating wood mice and the authorities of Collserola Natural Park for granting us 450 permission to work in Can Balasc. This study was funded by MICINN (Spain) in the project 451 FORASSEMBLY (CGL2015-40558-P). 452 35 19 36 453 REFERENCES 454 Apfelbach, R., Blanchard, C. D., Blanchard, R. J., Hayes, R. A., McGregor, I. S. (2005). The 455 effects of predator odors in mammalian prey species: A review of field and laboratory 456 studies. Neuroscience and Biobehavioral Reviews, 29, 1123-1144 457 Bell, A. M. (2005). Behavioural differences between individuals and two populations of 458 stickleback (Gasterosteus aculeatus). Journal of Evolutionary Biology, 18, 464-473. 459 Blondel, J., & Aronson, J. (1999). Biology and wildlife of the Mediterranean region. Oxford 460 university Press. 461 Bonfil, C., (1998). The effects of seed size, cotyledon reserves, and herbivory on seedling 462 survival and growth in Quercus rugosa and Q. laurina (Fagaceae). American Journal of 463 Botany, 85, 79–87. 464 Carter, A. J., Feeney, W. E., Marshall, H. H., Cowlishaw, G., & Heinsohn, R. (2013). Animal 465 personality: What are behavioural ecologists measuring? Biological Reviews, 88, 465– 466 475. 467 Chaby, L. E., Sheriff, M. J., Hirrlinger, A. M., Braithwaite, V. A. (2015) Does early stress 468 prepare individuals for a stressful future? Stress during adolescence improves foraging 469 under threat. Animal Behaviour, 105, 37-45. 470 Dall, S. R., Houston, A. I., & McNamara, J. M. (2004). The behavioural ecology of personality: 471 consistent individual differences from an adaptive perspective. Ecology Letters, 7, 734- 472 739. 473 Den Ouden, J., Jansen, P.A., Smit, R. (2005). Jays, mice and oaks. Predation and dispersal of 474 Quercus robur and Q. petraea in North-western Europe. In P.M. Forget, J.E. Lambert, 475 P.E. Hulme, & S.B. Vander Wall (Eds.), Seed fate: predation, dispersal and seedling 476 establishment (pp. 223 – 239). CAB International. 37 20 38 477 Dingemanse, N. J., Van der Plas, F., Wright, J., Réale, D., Schrama, M., Roff, D. A., & Barber, 478 I. (2009). Individual experience and evolutionary history of predation affect expression 479 of heritable variation in fish personality and morphology. Proceedings of the Royal 480 Society of London B: Biological Sciences, 276, 1285-1293. 481 Dosmann, A., Brooks, K. C., & Mateo, J. M. (2015). Evidence for a mechanism of phenotypic 482 integration of behaviour and innate immunity in a wild rodent: implications for animal 483 personality and ecological immunology. Animal Behaviour, 101, 179-189. 484 Espelta, J. M., Bonal, R., & Sánchez-Humanes, B. (2009). Pre-dispersal acorn predation in 485 mixed oak forests: Interspecific differences are driven by the interplay among seed 486 phenology, seed size and predator size. Journal of Ecology, 97, 1416–1423. 487 Espirito-Santo, C., Rosalino, L. M., & Santos-Reis, M. (2007). Factors affecting the placement 488 of common genet latrine sites in a mediterranean landscape in Portugal. Journal of 489 Mammalogy, 88, 201-207. 490 Gómez, J. M. (2004). Bigger is not always better: conflicting selective pressures on seed size in 491 Quercus ilex. Evolution, 58, 71-80. 492 Gómez, J. M., Puerta-Piñero, C., & Schupp, E. W. (2008). Effectiveness of rodents as local seed 493 dispersers of holm oaks. Oecologia, 155, 529–537. 494 Grabowski, J. H., & Kimbro, D. L. (2005). Predator avoidance behavior extends trophic 495 cascades to refuge habitats. Ecology, 86, 1312-1319. 496 Gu, H., Zhao, Q., & Zhang, Z. (2017). Does scatter-hoarding of seeds benefit cache owners or 497 pilferers? Integrative Zoology, 12, 477-488 498 Hayes, J. P., & Jenkins, S. H. (1997). Individual variation in mammals. Journal of Mammalogy, 499 78, 274-293. 500 Heinen-Kay, J. L., Schmidt, D. A., Stafford, A. T., Costa, M. T., Peterson, M. N., Kern, E. M. 501 A., & Langerhans, R. B. (2016). Predicting multifarious behavioural divergence in the 502 wild. Animal Behaviour, 121, 3–10. 39 21 40 503 Herrera, C. M. (2002) Seed dispersal by vertebrates. In C. M. Herrera & O. Pellmyr O (Eds.), 504 Plant-animal interactions. An Evolutionary approach (pp. 185-208). Blackwell Science, 505 Oxford. 506 Hollander, J. L., & Vander Wall, S. B. (2004). Effectiveness of six species of rodents as 507 dispersers of singleleaf pinon pine (Pinus monophylla). Oecologia, 138, 57–65. 508 Hopewell, L. J., & Leaver, L. A. (2008) Evidence of Social Influences on Cache‐Making by 509 Grey Squirrels (Sciurus carolinensis). Ethology, 114, 1061-1068. 510 Jansen, P. A., Bongers, F., & Hemerik, L. (2004). Seed mass and mast seeding enhance 511 dispersal by a neotropical scatter-hoarding rodent. Ecological Monographs, 74, 569- 512 589. 513 Jolles, J. W., Boogert, N. J., & Van den Bos, R. (2015). Sex differences in risk- taking and 514 associative learning in rats. The Royal Society Open Science, 2, 150485. 515 Kerley, G. I. H., & Erasmus, T. 1991. What do mice select for in seeds? Oecologia, 86, 261– 516 267. 517 Koolhaas, J. M., Korte, S. M., De Boer, S. F., Van Der Vegt, B. J., Van Reenen, C. G., Hopster, 518 H., & Blokhuis, H. J. (1999). Coping styles in animals: current status in behavior and 519 stress-physiology. Neuroscience & Biobehavioral Reviews, 23, 925-935. 520 Koolhaas, J. M., De Boer, S. F., Coppens, C. M., & Buwalda, B. (2010). Neuroendocrinology of 521 coping styles: towards understanding the biology of individual variation. Frontiers in 522 neuroendocrinology, 31, 307-321. 523 Korpela, K., Sundell, J., & Ylönen, H. (2011). Does personality in small rodents vary depending 524 on population density? Oecologia, 165, 67-77. 525 Labandeira, C. C. (2002). The history of associations between plants and animals. In C. M. 526 Herrera & O. Pellmyr (Eds.), Plant-animal interactions. An Evolutionary approach (pp. 527 26-74). Blackwell Science, Oxford. 41 22 42 528 Leaver, L. A. (2004) Effects of food value, predation risk, and pilferage on the caching 529 decisions of Dipodomys merriami. Behavioral Ecology, 15, 729-734. 530 Lichti, N. I., Steele, M. A., & Swihart, R. K. (2017). Seed fate and decision‐making processes in 531 scatter‐hoarding rodents. Biological Reviews, 92, 474–504. 532 Mamuneas, D., Spence, A. J., Manica, A., & King, A. J. (2015). Bolder stickleback fish make 533 faster decisions, but they are not less accurate. Behavioral Ecology, 26, 91–96. 534 Moran, N. P., Mossop, K. D., Thompson, R. M., & Wong, B. B. M. (2016). Boldness in 535 extreme environments: temperament divergence in a desert-dwelling fish. Animal 536 Behaviour, 122, 125–133. 537 Muñoz, A., & Bonal, R. (2007). Rodents change acorn dispersal behaviour in response to 538 ungulate presence. Oikos, 116, 1631–1638. 539 Muñoz, A., & Bonal, R. (2008a). Seed choice by rodents: learning or inheritance? Behavioral 540 Ecology and Sociobiology, 62, 913-922. 541 Muñoz, A., & Bonal, R. (2008b). Are you strong enough to carry that seed? Seed size/body size 542 ratios influence seed choices by rodents. Animal Behaviour, 76, 709-715. 543 Muñoz, A., Bonal, R., & Díaz, M. (2009). Ungulates, rodents, shrubs: interactions in a diverse 544 Mediterranean ecosystem. Basic and Applied Ecology, 10, 151–160. 545 Muñoz, A., & Bonal, R. (2011). Linking seed dispersal to cache protection strategies. Journal 546 of Ecology, 99, 1016–1025. 547 Muñoz, A., Bonal, R., & Espelta, J. M. (2012). Responses of a scatter-hoarding rodent to seed 548 morphology: Links between seed choices and seed variability. Animal Behaviour, 84, 549 1435–1442. 550 Navarro-Castilla, Á., & Barja, I. (2014). Does predation risk, through moon phase and predator 551 cues, modulate food intake, antipredatory and physiological responses in wood mice 552 (Apodemus sylvaticus)? Behavioral Ecology and Sociobiology, 68, 1505–1512. 43 23 44 553 Natelson, B. H., Ottenweller, J. E., Cook, J. A., Pitman, D., McCarty, R., & Tapp, W. N. (1988). 554 effect of stressor intensity on habituation of the adrenocortical stress response. 555 Physiology & Behavior, 43, 41-46. 556 Nussey, D. H., Wilson, A. J., & Brommer, J. E. (2007). The evolutionary ecology of individual 557 phenotypic plasticity in wild populations. Journal of Evolutionary Biology, 20, 831-844. 558 Pan, Y., Li, M., Yi, X., Zhao, Q., Lieberwirth, C., Wang, Z., & Zhang, Z. (2013). Scatter 559 hoarding and hippocampal cell proliferation in Siberian chipmunks. Neuroscience, 255, 560 76-85. 561 Pennisi, E. (2016). The power of personality. Science, 352, 644-647. 562 Perea, R., Miguel, A. S., & Gil, L. (2011). Acorn dispersal by rodents: The importance of re- 563 dispersal and distance to shelter. Basic and Applied Ecology, 12, 432–439. 564 Preston, S. D. & Jacobs, L. F. (2009). Mechanisms of cache decision making in fox squirrels 565 (Sciurus niger). Journal of Mammalogy, 90, 787-795. 566 Réale, D., Gallant, B. Y., Leblanc, M., & Festa-Bianchet, M. (2000) Consistency of 567 temperament in bighorn ewes and correlates with behaviour and life history. Animal 568 Behaviour, 60, 589–597. 569 Réale, D., Reader, S. M., Sol, D., McDougall, P. T., & Dingemanse, N. J. (2007). Integrating 570 animal temperament within ecology and evolution. Biological Reviews, 82, 291–318. 571 Réale, D., Dingemanse, N. J., Kazem, A. J. N., & Wright, J. (2010). Evolutionary and 572 ecological approaches to the study of personality. Philosophical Transactions of the 573 Royal Society B: Biological Sciences, 365, 3937–3946. 574 Robledo-Arnuncio, J. J., Klein, E. K., Muller-Landau, H. C., & Santamaría, L. (2014). Space, 575 time and complexity in plant dispersal ecology. Movement ecology, 2, 16. 45 24 46 576 Samson, J., & Manser, M. B. (2016) Caching in the presence of competitors: Are Cape ground 577 squirrels (Xerus inauris) sensitive to audience attentiveness? Animal Cognition, 19, 31- 578 38. 579 Schupp, E. W., Jordano, P., & Gómez, J. M. (2010). Seed dispersal effectiveness revisited: a 580 conceptual review. New Phytologist, 188, 333-353. 581 Sih, A., Bell, A., & Johnson, J. C. (2004). Behavioral syndromes: an ecological and 582 evolutionary overview. Trends in Ecology & Evolution, 19, 372-378. 583 Sih, A., & Bell, A. M. (2008). Insights for behavioral ecology from behavioral syndromes. 584 Advances in the Study of Behavior, 38, 227-281. 585 Sikes, R. S., & Gannon, W. L. (2011). Guidelines of the American Society of Mammalogists for 586 the use of wild mammals in research. Journal of Mammalogy, 92, 235-253. 587 Sneddon, L. U. (2003). The bold and the shy: individual differences in rainbow trout. Journal of 588 Fish Biology, 62, 971-975. 589 Stamps, T. G., & Groothuis, G. 2010. Developmental perspectives on personality: implications 590 for ecological and evolutionary studies of individual differences. Philosophical 591 Transactions of the Royal Society B: Biological Sciences, 365, 4029–4041. 592 Michael A. Steele, M. A., Halkin, S. L., Smallwood, P. D., McKenna, T. J., Mitsopoulos, K., & 593 Beam, M. (2008). Cache protection strategies of a scatter-hoarding rodent: Do tree 594 squirrels engage in behavioural deception? Animal Behaviour, 75, 705-714. 595 Sunyer, P., Muñoz, A., Bonal, R., & Espelta, J. M. (2013). The ecology of seed dispersal by 596 small rodents: A role for predator and conspecific scents. Functional Ecology, 27, 597 1313–1321. 598 Sunyer, P., Espelta, J. M., Bonal, R., & Muñoz, A. (2014). Seeding phenology influences wood 599 mouse seed choices: The overlooked role of timing in the foraging decisions by seed- 600 dispersing rodents. Behavioral Ecology and Sociobiology, 68, 1205–1213. 47 25 48 601 Sunyer, P., Boixadera, E., Muñoz, A., Bonal, R., & Espelta, J. M. (2015). The interplay among 602 acorn abundance and rodent behavior drives the spatial pattern of seedling recruitment 603 in mature Mediterranean oak forests. PLoS ONE, 10, 1–13. 604 Sunyer, P., Muñoz, A., Mazerolle, M. J., Bonal, R., & Espelta, J M. (2016). Wood mouse 605 population dynamics: Interplay among seed abundance seasonality, shrub cover and 606 wild boar interference. Mammalian Biology, 81, 372-379. 607 Torre, I., Arrizabalaga, A., & Díaz, M. (2002). Ratón de campo (Apodemus sylvaticus. 608 Linnaeus, 1758). Galemys, 14, 1–26. 609 Torre, I., Ballesteros, T., & Degollada, A. (2003). Cambios en la dieta de la gineta ( Genetta 610 genetta Linnaeus, 1758) con relación a la disponibilidad de micromamíferos: ¿Posible 611 preferencia por el topillo rojo? Galemys, 15, 13-24. 612 Urgoiti, J., Muñoz, A., Espelta, J. M., & Bonal, R. (2018). Distribution and space use of seed- 613 dispersing rodents in central Pyrenees: implications for genetic diversity, conservation 614 and plant recruitment. Integrative Zoology, 13, 307-318. 615 Vander Wall, S. B. (1990). Food hoarding in animals. University of Chicago Press. 616 Vellend, M., Knight, T. M., & Drake, J. M. (2006). Antagonistic effects of seed dispersal and 617 herbivory on plant migration. Ecology Letters, 9, 319-326. 618 Verbeek, M. E., Boon, A., & Drent, P. J. (1996). Exploration, aggressive behaviour and 619 dominance in pair-wise confrontations of juvenile male great tits. Behaviour, 133, 945- 620 963. 621 Verdolin, J. L. (2006). Meta-analysis of foraging and predation risk trade-offs in terrestrial 622 systems. Behavioral Ecology and Sociobiology, 60, 457-464. 623 Virgós, E., Llorente, M., & Cortés, Y. (1999). Geographical variation in genet (Genetta genetta) 624 diet: a literature review. Mammal Review, 29, 119-128. 49 26 50 625 Wang, J., Zhang, B . Xiang H., &, Chen, X., Han, N., & Chang, G. (2017). Effects of mast 626 seeding and rodent abundance on seed predation and dispersal of Quercus aliena 627 (Fagaceae) in Qinling Mountains, Central China. Plant Ecology, 218, 855-865. 628 Wenny, D. G. (2002). Effects of human handling of seeds on seed removal by rodents. The 629 American Midland Naturalist, 147, 404-408. 630 Wilson, D.S. (1998) Adaptive individual differences within single populations. Philosophical 631 Transactions of the Royal Society B: Biological Sciences, 353, 199–205. 632 Wolf, G., van Doorn S., Leimar, O., & Weissing, F. J. (2007). Life-history trade-offs favour the 633 evolution of animal personalities. Nature, 447, 581–584. 634 Wolf, G., & Weissing, F. J. 2012. Animal personalities: consequences for ecology and 635 evolution. Trends in Ecology & Evolution, 27, 452-461. 636 Yi, X., Steele, M. A., Stratford, J. A., Wang, Z., & Yang, Y. (2016). The use of spatial memory 637 for cache management by a scatter-hoarding rodent. Behavioral Ecology and 638 Sociobiology, 70, 1527–1534. 639 Yi, X. & Wang, Z. (2015). Dissecting the roles of seed size and mass in seed dispersal by 640 rodents with different body sizes. Animal Behaviour, 107, 263-267. 641 Xiao, Z. S., Zhang, Z. B., & Wang, Y. S. (2004). Dispersal and germination of big and small 642 nuts of Quercus serrata in a subtropical broad-leaved evergreen forest. Forest Ecology 643 and Management, 195, 141–150. 644 Xiao, Z., Gao, X., & Zhang, Z. (2013). The combined effect of seed perishability and seed size 645 on hoarding decisions by Pére David’s rock squirrels. Behavioral Ecology and 646 Sociobiology, 67, 1067–1075. 647 Yuen, C. H., Pillay, N., Heinrichs, M., Schoepf, I., & Schradin, C. (2015). Personality does not 648 constrain social and behavioural flexibility in African striped mice. Behavioral Ecology 649 and Sociobiology, 69, 1237–1249. 51 27 52