Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Living on the edge: morphological and behavioral adaptations to a marginal high‐elevation habitat in an arboreal mammal

Living on the edge: morphological and behavioral adaptations to a marginal high‐elevation habitat... INTRODUCTIONLocal environmental conditions, and their temporal changes, can exert selective pressures on animals, resulting in phenotypic variation among populations living in different habitats (Kawecki & Ebert 2004). These local phenotypic adaptations often refer to physiological or morphological traits subject to natural selection, but also to behavioral traits, that are assumed to be more flexible with animals capable of adjusting their behavior in response to fluctuations in environmental conditions (Bolnick et al. 2003; Holtmann et al. 2017). However, both theoretical and empirical studies on animal personality, defined as behavioral differences among individuals of a species that are consistent over time and across situations (e.g. Réale et al. 2007; Carter et al. 2013), suggest that behavioral plasticity may be restrained by an individual's personality type, and that animals select habitats that best fit their personality (Dall et al. 2004; Sih et al. 2004; Holtmann et al. 2017). In other words, personality traits may reduce an individual's ability to cope with certain environmental conditions and, consequently, result in non‐random distribution of behavioral phenotypes across different habitats (Cote et al. 2010; Holtmann et al. 2017; Wauters et al. 2019).For mammals living in marginal habitats on the edge of a species’ distribution range, environmental conditions can be extreme and exert strong selective pressure on individuals, different from those acting in more central, higher‐quality habitats, characterized by more abundant resources and, in some cases, less extreme weather conditions (Endler 1977; Rodrigues et al. 2010). This could result in habitat‐dependent phenotypic differences in body size and/or personality for mammals living in heterogeneous environments (Wauters et al. 2007; Hewes et al. 2017; Brehm & Mortelliti 2021; Boone et al. 2022). Also, phenotypic differences between animals from marginal and high‐quality areas may be due to individuals showing higher behavioral plasticity in populations living at the distribution edge (Dingemanse et al. 2010).Depending heavily on tree seeds as their main food resource throughout most of the year, arboreal squirrels in temperate regions typically occur in resource‐limited producer–consumer systems (Boutin & Larsen 1993; Koprowski 1994a; Boutin et al. 2006; Wauters et al. 2008). In heterogeneous landscapes, spatio‐temporal variation in tree seed abundance varies greatly, creating a mosaic of habitats that differ in their quality for tree squirrel populations in terms of overall food availability as well as temporal fluctuations in the size of tree‐seed crops (Koprowski 1994a,b; Salmaso et al. 2009; Di Pierro et al. 2011; Tranquillo et al. 2022). Moreover, morphological and/or behavioral differences and variation in personality traits are related to space and/or habitat use (Wauters et al. 2007, 2021; Boon et al. 2008) and to variation in survival and/or reproductive success (McAdam et al. 2002; Wauters et al. 2007; Santicchia et al. 2018). Therefore, tree squirrels are particularly interesting models to study habitat‐related phenotypic differences in body size and/or personality for mammals living in heterogeneous environments.Here, we used the Eurasian red squirrel (Sciurus vulgaris) to investigate if phenotypic variation differs between individuals from a population living in a marginal low‐quality habitat on the edge of the elevational distribution and animals occurring in moderate to high‐quality conifer forests at similar or lower elevations in the Italian Alps. As explained above, habitat quality refers to average seed‐crop size and annual fluctuations in seed production of conifers, the main food resource for red squirrels in Alpine forests (Wauters et al. 2008; Salmaso et al. 2009; Rodrigues et al. 2010; Tranquillo et al. 2022 for detailed measurements of food abundance in these study sites).Eurasian red squirrels occupying extreme, high‐elevation habitat differ from conspecifics in higher quality habitats at lower elevations in population dynamics (Mari et al. 2008; Rodrigues et al. 2010) and space use (Romeo et al. 2010). Space use differs between high‐elevation marginal habitat of dwarf mountain pine (Pinus mugo) and high‐quality Scots pine (Pinus sylvestris) forest at lower‐elevations (Romeo et al. 2010). In fact, home ranges and core areas are larger in the marginal habitat, with more pronounced sex differences in core area size (Romeo et al. 2010). In contrast, patterns of core area overlap and intrasexual territoriality are similar as in other populations from various habitat types (Wauters & Dhondt 1992; Lurz et al. 2000; Wauters et al. 2001; Di Pierro et al. 2011). However, low densities in the edge population (Rodrigues et al. 2010) result in higher female by males overlap in spring–summer, when males expand their home ranges during the mating season to increase access to estrus females (Romeo et al. 2010). Moreover, squirrels’ spacing behavior responds more strongly to variation in food availability and/or density fluctuations in marginal than in high‐quality habitats (Romeo et al. 2010; Wauters et al. 2021), suggesting apparent behavioral plasticity to respond to the ecological constraints in marginal habitats.We measured morphological (body mass and hind foot length) and behavioral (exploration, activity, activity‐exploration and social tendency) traits of red squirrels. Hind foot length is strongly correlated with total head and body length and thus is considered a reliable proxy for body size (Wauters & Dhondt 1989a). Behavioral traits have been shown to determine an animal's personality (Mazzamuto et al. 2019; Wauters et al. 2019; Santicchia et al. 2020a, 2021).In particular, we predicted: (i) Differences in body size and mass between the habitats, with smaller squirrels of lower body mass in the marginal, poorer‐quality habitat, where (based on earlier studies of habitat‐related differences in foot length and body mass of red squirrels, Wauters et al. 2007) smaller squirrels may be advantaged since they have lower daily energy‐requirements than larger and heavier individuals. Alternatively, squirrels may be smaller and weigh less in the poorer‐quality habitat because of lower food intake. (ii) Differences in the expression of behavioral traits, with squirrels in the edge population having higher average levels of activity and exploration than in high‐quality habitats, since active‐explorative personality types should be better adapted at exploring low‐abundance food resources (see also Spiegel et al. 2017). (iii) A more pronounced advantage of higher behavioral plasticity in edge than in high‐quality habitats, resulting in higher within‐individual variance of behavioral traits in the edge population (e.g. Dingemanse & Wolf 2013).MATERIALS AND METHODSStudy areas, trapping, and handlingAs part of a long‐term research project on spatio‐temporal variation in producer‐consumer dynamics in pulsed resource systems, tree‐seed abundance, the population dynamics, life‐history, personality traits and space use of red squirrels were studied in 3 populations occupying different types of conifer forests (Wauters et al. 2007, 2008; Salmaso et al. 2009; Rodrigues et al. 2010; Santicchia et al. 2021; Tranquillo et al. 2022). These study areas, located in Lombardy, Northern Italy, are characterized by subalpine mixed conifer forests that differ in the dominant tree species. Cancano (CAN; 46°33ʹN, 10°15ʹE, elevation from 1940 m to 1970 m a.s.l.) extends over 54 ha in a high‐elevation valley, and it is part of a monospecific forest of dwarf mountain pine, Pinus mugo, whose seeds are the only major food resource for red squirrels (Rodrigues et al. 2010; Romeo et al. 2010). Bormio (BOR; 46°27ʹN, 10°30ʹE, elevation from 1950 m to 2130 m a.s.l.; size 92 ha) is dominated by Arolla pine, Pinus cembra (73.2%), with larch, Larix decidua (17.8%) and scarce Norway spruce, Picea abies (8.0%), and dead trees (1.0%) (Wauters et al. 2007; Santicchia et al. 2021). Valfurva (VAL; 46°27ʹN, 10°31ʹE, elevation from 1650 m to 1870 m a.s.l.) is located on an east‐north‐east slope in Valfurva valley and covers 78 ha. It is characterized by Norway spruce (88.9%) with a lower presence of Arolla pine (6.2%), dead trees (3.0%), and larch (1.9%) (Wauters et al. 2007; Santicchia et al. 2021).Based on elevation and habitat quality (average seed‐crop size and annual fluctuations in seed production, see also Table 1), Cancano is a marginal edge‐habitat, Bormio is a high‐elevation habitat of moderate habitat‐quality with harsh winter conditions, and Valfurva is a high‐quality habitat, typical for the dominant subalpine conifer forest‐type of the central Italian‐Alps (Wauters et al. 2007; Rodrigues et al. 2010).1TableRed squirrel density estimated using MNA and food abundance expressed in energy‐content of conifer seeds (MJ·ha−1, mean ± SE), in the 3 study areas per yearCancano (54 ha) (0.35 traps·ha−1) (n = 68; ID = 34)Bormio (92 ha) (0.22 traps·ha−1) (n = 55; ID = 36)Valfurva (77 ha) (0.26 traps·ha−1) (n = 226; ID = 92)YearRed squirrel density (ind·ha−1)Food abundance (MJ·ha−1)Red squirrel density (ind·ha−1)Food abundance (MJ·ha−1)Red squirrel density (ind·ha−1)Food abundance (MJ·ha−1)20160.17128 ± 130.152301 ± 2990.402863 ± 32820170.20294 ± 350.131605 ± 2080.491361 ± 10320180.20243 ± 580.111939 ± 2160.381559 ± 27020190.07234 ± 320.101273 ± 1490.35170 ± 1920200.11496 ± 840.255404 ± 59720210.09108 ± 230.1947 ± 14Size of the study areas (ha) and trap density (traps·ha−1) between brackets. Arena test sample sizes (n) and number of different animals (ID) between brackets, for each study areas.Population dynamics and life‐history traits were studied by Capture‐Mark‐Recapture (CMR), and food abundance was estimated by annual cone counts in Cancano (2005–2021), Bormio (2002–2019), and Valfurva (2002–2021). Squirrel density was calculated by dividing the minimum number of animals known to be alive (MNA) by study area size (ha) (Wauters et al. 2008). Details of CMR methods and conifer seed‐crop size are given elsewhere (Wauters et al. 2008; Salmaso et al. 2009; Rodrigues et al. 2010). Briefly, CMR sessions were carried out each year in May–June (spring–summer) and in September–October (autumn), using live traps (Tomahawk trap model 202, Tomahawk Live Trap Co., Tomahawk, Wisconsin, USA) placed with baling wire on tree trunks at breast height or on the ground (Wauters et al. 2008; Tranquillo et al. 2022). We pre‐baited traps with hazelnuts once per week for about 1 month prior to trapping; trapping sessions lasted 4 to 5 days, and we checked traps 2 to 3 times per day to reduce the time a squirrel was confined in the trap. A trapped squirrel was flushed from the trap in a zipper‐tube handling bag, and, at first capture, individually marked with a numbered metal ear tag (Monel 1005 1L1 National Band and Tag Co, Newport, Kentucky, USA). Sex and reproductive condition were determined based on external genitalia (details in Wauters et al. 2007; Tranquillo et al. 2022). Males were classified based on testes size and position (testes abdominal, testes scrotal); females as non‐breeding, post‐estrus (pregnant), or lactating. At each capture, an animal was weighed to the nearest 5 g with a spring balance (Pesola AG, Baar, Switzerland), and the length of the right hind foot (claws excluded) was measured (0.5 mm precision) with a thin wing ruler.Personality quantification: arena testWe studied the personality of red squirrels performing the open field test (OFT; 4 min), to estimate an animal's activity and exploration in a novel environment, followed by the mirror image stimulation test (MIS; 3 min) to determine the degree of sociability or avoidance (social tendency), aggressiveness, and being alert toward a conspecific, as well as the tendency for expressing behaviors that define a combined activity‐exploration trait (Mazzamuto et al. 2019; Wauters et al. 2019; Santicchia et al. 2020a, 2022). Both tests were carried out during the same capture event in a portable arena, placed within 20 m from the trap location, where a previously identified squirrel was released by opening a sliding door and allowing the animal to move from the handling bag into the arena (Mazzamuto et al. 2019; Santicchia et al. 2021, 2022). At the end of the experiment, the squirrel was released and the arena was cleaned with 90% ethylic alcohol. More details on the arena test are given in previous studies (Santicchia et al. 2021, 2022). The time that an individual spent in each behavioral state (behaviors defined in Table S1, Supporting Information) was calculated using the CowLog 3.0.2 software (Hänninen & Pastell 2009). Next, we applied the expert‐based method, defined and validated by Mazzamuto et al. (2019), to obtain new scores for 7 personality‐linked variables (Ethogram description and personality traits in Table S1, Supporting Information).To check the assumptions of repeatability of the traits, squirrels were tested multiple times in different CMR sessions to have repeated measures (individuals tested multiple times: n = 78, mean ± SE = 2.63 ± 0.10, range = 2–9). Nevertheless, to increase the power of the test, squirrels with only one arena test were also included in the adjusted repeatability estimates (Martin et al. 2011; Santicchia et al. 2022).Body size and body mass analysesA total of 193 individuals (114 males and 79 females) were sampled to assess the differences in morphological traits (see also Results, Differences in body size and body mass). Squirrels were classified as juveniles or adults based on external genitalia and body mass (<225 g; Wauters et al. 2007 and references therein), and juveniles were excluded from the analyses to avoid any bias in relation to our predictions.We explored the effects of study area and sex on hind foot length with a general linear model (GLM). Mean foot length of each individual was used as dependent variable (n = 193 different squirrels), adding study area and sex as categorical variables and the study area by sex interaction.Next, we investigated whether study area affected variation in body mass. Since body mass fluctuates seasonally and with reproductive condition, and since effects of reproductive condition differ between the sexes (Wauters & Lens 1995; Wauters et al. 2007), we used Linear Mixed Models (LMM), with multiple body mass measurements for individual squirrels, for each sex separately. Body mass was used as dependent variable, hind foot length as explanatory continuous variable, and study area, season, and reproductive condition as fixed effects. We further included all 2‐factor interactions with study area in the full model, and added squirrel identity as a random intercept to account for repeated measures over time. A model with compound symmetry correlation structure best fitted the data (based on Schwarz's Bayesian Information Criteria). We tested interactions and fixed effects using Type‐III SS and performed model selection using a backward procedure based on P‐values (eliminating parameters with partial P > 0.15). Degrees of freedom and standard errors of F‐ and t‐tests were obtained using the Kenward–Rogers method (Verbeke & Molenberghs 2000; Wauters et al. 2007), and pairwise comparisons were based on differences between least square means (DLSM).Personality analysesA total of 162 squirrels (100 males and 62 females) were behaviorally tested (see also Results, Behavioral differences). To account for possible confounding effect due to modification in personality with developmental changes (Cabrera et al. 2021; see also Discussion for details), we excluded juvenile individuals from this dataset.Since personality traits are defined as behaviors consistent over time and under different contexts (Réale et al. 2007; Carter et al. 2013) we first calculated the adjusted repeatability (R) of each expert‐based group (see Table S2, Supporting Information). Repeatability was estimated using a LMM through the function rpt, with the R package rptR (Stoffel et al. 2017). We performed a model for each square‐root‐transformed value of expert‐based group (activity, exploration and immobility from OFT and sociability, avoidance, alert and activity‐exploration from MIS) as dependent variable. Likewise, we estimated the adjusted repeatability also for social tendency, a variable that combines avoidance and sociability scores (Pearson correlation coefficient sociability‐avoidance: r = −0.87; n = 349; P < 0.0001) and measures the tendency to react more or less social or amenable to the mirror image (Social tendency = log[(sociability score + (1 − avoidance score))/2], Santicchia et al. 2020a).Each model had squirrel identity (ID) as random intercept and sex, arena test order (first test was coded as 1, subsequent tests were coded as 0), year (from 2016 to 2021), and number of days since the previous arena test as fixed effects. As in previous studies (Santicchia et al. 2020a,b, 2021, 2022), we included in the models arena test order and number of days since the previous arena test, to account for habituation and temporal proximity between 2 arena tests (Dingemanse et al. 2012; Mazzamuto et al. 2019). The likelihood ratio test (LRT) was used to test for significance of the random intercept of each model (Martin & Réale 2008). Full repeatability model outputs are provided in Table S2, Supporting Information.We then investigated whether behavioral traits expression changed between study areas using a Bayesian generalized linear mixed effects model based on a Markov Chain Monte Carlo algorithm with the R package MCMCglmm version 2.32 (Hadfield 2010). This model included 4 behavioral traits (activity and exploration from the OFT, activity‐exploration and social tendency from MIS) as dependent variable, that were previously centered and scaled [(xi – mean x)/SD x] with a Gaussian residual error distribution. Squirrel identity (ID) was added as random effect since repeated observations were present, and to estimate both among‐individual and within‐individual variation of the dependent variables. Study area, sex, body mass (centered and scaled), season (spring–summer [May–June–July]; autumn [September–October]) and arena test order were included as fixed effects. Posterior distribution of the model was based on 1 050 000 iterations with a burn‐in of 50 000 iterations and thinning of 40, such that 25 000 iterations were used to obtain point estimates and 95% credibility intervals (CIs). We used an inverse‐gamma uninformative prior, for the random effect and the residual variation (Wilson et al. 2010). We applied the Gelman–Rubin statistic (Gelman & Rubin 1992) and Geweke diagnostic (Geweke 1992) to confirm model consistency and convergence. Full model output is provided in Table S3, Supporting Information.In a final step, to obtain the among‐individual and within‐individual variances and the estimates of the correlations among the behavioral traits for each study area, we performed other 3 MCMCglmm models using 3 restricted dataset (one for each study area), in which we included the same dependent variables, fixed effects and random effect as in the main model. This approach allowed us to estimate the differences between study areas in the among‐ and within‐individual variances (Royauté & Dochtermann 2021; Santicchia et al. 2022). Statistical analyses were carried out using the software R (R Core Team 2020).Ethical noteTrapping, marking and handling procedures and arena test experiments complied with the Guidelines for the treatment of animals in behavioral research and teaching (Animal Behaviour, 2020, 159, I–XI; https://doi.org/10.1016/j.anbehav.2019.11.002). Legal requirements according to the Italian Wildlife Protection and Hunting Law L.N. 157 from 1992 and fieldwork was approved by authorization decrees n. 11190 of 29/11/2013, n. 9523 of 15/10/2014, and n. 198 of 13/01/2017 from Direzione Generale Agricoltura, Regione Lombardia; and the permission Protocol N. 414 of 28/02/2014 of the Stelvio National Park.RESULTSDifferences in body size and body massHind foot length in red squirrels (n = 193) varied between 54 and 61 mm (overall mean ± SD = 57.6 ± 1.1 mm). Foot length differed between study areas and there was a significant study area × sex interaction (Area effect F2,187 = 11.24; P < 0.0001; Sex effect F1,187 = 1.27; P = 0.26; area × sex F2,187 = 4.88; P = 0.009). Female squirrels in Valfurva were slightly larger than in Bormio (Table 2a, DLSM P = 0.049, all other pairwise comparisons P > 0.30). Males in Valfurva were larger than in Bormio (DLSM P = 0.041) while males in Cancano were the smallest (DLSM VAL‐CAN P < 0.0001; BOR‐CAN P = 0.0008; Table 2a). Consequently, there was no size difference between the sexes in the populations in Valfurva and Bormio, while in Cancano females were significantly larger than males (Table 2a: DLSM P = 0.004).2Table(a) Mean right hind foot length (±SD) of red squirrels in 3 populations in the Italian Alps; data by sex and for the sexes combined; sample size is given between brackets. (b) Mean body mass (±SD) of male and female red squirrels in 3 populations in the Italian Alps; sample size is given between bracketsFemalesMalesBoth sexesa) Foot lengthValfurva(46) 57.8 ± 1.1(60) 58.0 ± 1.1(106) 57.9 ± 1.1Bormio(17) 57.3 ± 0.9(31) 57.5 ± 1.0(48) 57.4 ± 0.9Cancano(16) 57.5 ± 0.9(23) 56.5 ± 1.0(39) 56.9 ± 1.1b) Body massValfurva(140) 315 ± 27(275) 305 ± 18(415) 308 ± 22Bormio(25) 299 ± 26(74) 290 ± 19(99) 293 ± 21Cancano(50) 311 ± 31(100) 285 ± 16(150) 294 ± 25Body mass of subadult and adult red squirrels (n = 664, with repeated measures of 193 different animals) varied between 235 and 380 g (overall mean ± SD = 303 ± 24 g). For females we obtained 215 measures of 79 individuals (full LMM reported in Table S4, Supporting Information). In the final model, female body mass increased with foot length (estimate ± SE = 25.2 ± 4.7; t79 = 5.38; P < 0.0001), but the slope differed between study areas (foot length × study area F2,92 = 4.37; P = 0.015). The increase in body mass with foot length was stronger in the marginal habitat Cancano than in the high‐quality habitat Valfurva (difference in slope 15.2 ± 5.2; t82 = 2.91; P = 0.005; Fig. 1). Overall, females weighed slightly more in Valfurva than in Bormio, with Cancano intermediate (Table 2b, area effect F2,92 = 4.38; P = 0.015; Fig. 1), and lactating females had a higher body mass than animals in post‐estrus or non‐breeding (F2,191 = 25.9; P < 0.0001, DLSM all P < 0.01).1FigureRelationship between foot length and body mass of female red squirrels in Bormio (solid line, triangles), Cancano (dotted line, crosses), and Valfurva (dashed line, circles). The lines represent the predicted relationships, shaded areas represent the 95% confidence intervals, and symbols represent observed values.For male red squirrels, we used 449 body mass measures of 114 individuals (full LMM reported in Table S4, Supporting Information). In the final model, male body mass increased with foot length (estimate ± SE = 7.4 ± 1.3; t128 = 5.78; P < 0.0001; Fig. 2) and on average males were heavier in Valfurva than in the other 2 areas (Table 2b; area effect F2,100 = 6.32; P = 0.003; DLSM VAL > BOR P = 0.005; VAL > CAN P = 0.005; BOR and CAN P = 0.77). Males in breeding condition weighed more than non‐breeding males (estimate ± SE = 9.6 ± 2.7; t426 = 3.57; P = 0.0004), and body mass increased from spring–summer to autumn (estimate ± SE = 9.5 ± 2.7; t421 = 3.57; P = 0.0005).2FigureRelationship between foot length and body mass of male red squirrels. The solid line represents the predicted relationship, shaded areas represent the 95% confidence intervals, and diamonds represent observed values.Behavioral differencesWe estimated the adjusted repeatability of the 7 expert‐based traits and of social tendency, using a total of 349 arena tests on 162 different red squirrels (100 males and 62 females). Immobility, activity, and activity‐exploration had moderate repeatability (R range: 0.28 to 0.37; see also Bell et al. 2009), while exploration and social tendency had low repeatability (R = 0.14 and 0.15, respectively; Table S2, Supporting Information). The repeatability of sociability, avoidance, and alert was low and not significant (Table S2, Supporting Information). Immobility was strongly and negatively correlated with activity (r = −0.93; n = 349; P < 0.0001); therefore, this trait was not included in the multivariate models.We found several significant differences in personality traits between study areas. Squirrels had higher activity, activity‐exploration and social tendency in Cancano than in Bormio and Valfurva (Table 3; Fig. 3). Squirrels living in Cancano were also significantly more explorative than conspecifics living in Valfurva (Table 3; Fig. 3). There was no significant difference in activity, exploration, activity‐exploration or social tendency between Bormio and Valfurva (Table 3; Fig. 3). Male squirrels have a lower active‐explorative score than female squirrels (β = −0.29, 95% CI = −0.56 to −0.01, pMCMC = 0.041; Table S3, Supporting Information), and heavier animals were more explorative than individuals with lower body mass (β = 0.12, 95% CI = 0.01 to 0.23, pMCMC = 0.028). In spring–summer, red squirrels were significantly more active (β = 0.29, 95% CI = 0.12 to 0.47, pMCMC = 0.001), explorative (β = 0.58, 95% CI = 0.39 to 0.78, pMCMC < 0.0001), and also active‐explorative (β = 0.41, 95% CI = 0.21 to 0.60, pMCMC < 0.0001) than in autumn. Finally, activity and exploration were higher during the first arena test than during subsequent ones (activity: β = 0.75, 95% CI = 0.56 to 0.94, pMCMC < 0.0001; exploration: β = 0.52, 95% CI = 0.32 to 0.72, pMCMC < 0.0001), contrary to social tendency, that was lower during the first experiment (β = −0.33, 95% CI = −0.54 to −0.11, pMCMC = 0.003).3TablePosterior means and 95% credibility intervals of the comparisons between the 3 study areas (BOR, Bormio; CAN, Cancano; VAL, Valfurva) for each behavioral trait, obtained from the posterior distribution of the full MCMCglmm modelDifferencesDependent variableCAN–BORVAL–BORCAN–VALOFTActivity0.62 (0.24 to 1.01)0.27 (−0.05 to 0.59)0.35 (0.03 to 0.66)Exploration0.19 (−0.16 to 0.55)−0.27 (−0.57 to 0.03)0.46 (0.19 to 0.74)MISActivity‐exploration0.55 (0.15 to 0.94)0.10 (−0.24 to 0.44)0.45 (0.12 to 0.77)Social tendency0.51 (0.11 to 0.90)0.17 (−0.15 to 0.51)0.34 (0.03 to 0.65)Significant results in bold.3FigureComparisons of the behavioral scores between the 3 study areas (Cancano, dark grey; Bormio, grey; Valfurva, light grey). (a) Activity, exploration, and activity‐exploration square root transformed. (b) Social tendency true values. Boxplots show median (solid horizontal line), mean (black diamond), and 1st (25%) and 3rd (75%) quartiles.Among‐ and within‐individual variances and behavioral syndromeMost of the posterior estimates of the correlations among the personality traits were significant (Fig. 4). There were positive correlations of activity with activity‐exploration, and with social tendency; of exploration with activity‐exploration, and with social tendency; and between activity‐exploration and social tendency (Fig. 4). This suggests a behavioral syndrome (a close association of different behavioral traits; Sih et al. 2004, 2012; Réale et al. 2007) whereby more active squirrels being also more explorative and more social.4FigureCorrelations (estimate ± 95% credibility intervals) between the dependent variables derived from the full MCMCglmm model. Significant results (0 not included in the 95% CIs).Comparing among‐ and within‐individual variances between the study areas (Table S5, Supporting Information), we found that the among‐individual variance of exploration was higher in Cancano than in Valfurva (Table 4), suggesting more phenotypic variation in exploration behavior in red squirrels living in the marginal study area. In contrast, the within‐individual variances of exploration were lower in Cancano than in Bormio and Valfurva (Table 4), suggesting a higher plasticity of this behavioral trait in Bormio and in Valfurva than in Cancano, the marginal habitat.4TableMean differences ± SD (95% credibility intervals) in between‐individual variances (a) and within‐individual variances (b) of the dependent variables in the study areas (BOR, Bormio; CAN, Cancano; VAL, Valfurva) based on the 3 MCMCglmm modelsDifferencesDependent variableBOR–CANVAL–CANBOR–VALa) Between‐individualActivity0.11 ± 0.22 (−0.28 to 0.59)0.16 ± 0.15 (−0.17 to 0.44)−0.05 ± 0.21 (−0.39 to 0.42)Exploration−0.19 ± 0.37 (−0.89 to 0.62)−0.49 ± 0.25 (−1.06 to −0.11)0.30 ± 0.29 (−0.04 to 1.03)Activity‐exploration0.15 ± 0.50 (−0.72 to 1.22)−0.34 ± 0.28 (−0.96 to 0.11)0.49 ± 0.43 (−0.16 to 1.47)Social tendency0.04 ± 0.28 (−0.49 to 0.66)−0.08 ± 0.19 (−0.52 to 0.22)0.12 ± 0.24 (−0.24 to 0.70)b) Within‐individualActivity−0.07 ± 0.23 (−0.51 to 0.40)−0.07 ± 0.16 (−0.42 to 0.21)0.003 ± 0.20 (−0.32 to 0.44)Exploration0.67 ± 0.31 (0.14 to 1.35)0.36 ± 0.14 (0.04 to 0.60)0.31 ± 0.29 (−0.17 to 0.97)Activity‐exploration0.46 ± 0.35 (−0.11 to 1.27)0.29 ± 0.16 (−0.08 to 0.57)0.17 ± 0.34 (−0.33 to 0.96)Social tendency−0.0001 ± 0.25 (−0.47 to 0.52)0.27 ± 0.18 (−0.12 to 0.60)−0.27 ± 0.23 (−0.66 to 0.23)Significant results in bold.For estimates of the between‐individual and the within‐individual variances by study area, see Table S5, Supporting Information.DISCUSSIONIn this study, we investigated whether there were morphological and/or behavioral differences between 3 populations of Eurasian red squirrels in alpine study areas. Our results demonstrate that squirrels in the edge habitat of their altitudinal distribution showed both morphological and behavioral adaptations to the extreme weather and food conditions. First, we found that males in the marginal habitat were smaller and weighed less than in the high‐quality Norway spruce dominated forest. In contrast, females living in the edge habitat did not differ in body size and body mass from females living in the other 2 study areas. Secondly, squirrels living in the edge habitat had higher scores of behavioral traits, confirming our hypothesis that more active‐explorative types are advantaged in low‐quality habitats. Finally, and contrary to our predictions, animals living in the marginal habitat did not have a higher behavioral plasticity, but their within‐individual variance of exploration was significantly lower than in the other 2 study areas.Morphological differencesBetween‐habitat variation in body mass and body size, measured by hind‐foot length, revealed in this study confirmed earlier findings based on 6 populations (Wauters et al. 2007), 2 of which were also included here (Valfurva and Bormio). As in the previous study, there was no size difference between the sexes in Valfurva and Bormio, but in the marginal habitat (Cancano) hind foot length of females was on average 1 mm longer than males. Moreover, among males, the body size – body mass relationship did not differ between the study areas and males at Cancano weighed on average 20 g less than in the high‐quality habitat Valfurva. Among females, patterns were more complicated: body mass increased more strongly with size (hind foot length) in the marginal than in the high‐quality habitat and, consequently, there was little difference in the body mass of adult females between the study areas; on average only 4 g between Cancano and Valfurva. These sex‐specific patterns differ from previous findings on this and other tree squirrel species over most of their range (Nash & Seaman 1977; Boutin & Larsen 1993; Koprowski 1994a,b; Wauters & Dhondt 1989a; Lurz & Lloyd 2000), and suggest differential selective pressures for male and female red squirrels in the lowest‐quality marginal habitat Cancano than in neighboring populations living in the same geographic area (see also Hoffman & Merilä 1999; McAdam et al. 2002; McAdam & Boutin 2003). Hence, our data suggest that in the marginal, high‐elevation mountain‐pine habitat, smaller and lighter males have an advantage over larger, heavier ones. In mammals, animals of smaller size/body mass can meet daily energy requirements with a lower food‐intake than bigger individuals. In Cancano, red squirrels feed mainly on the small seeds from cones of mountain pine from late summer to early spring and on its buds and male flowers in spring–early summer (Romeo et al. 2010). Species‐specific energy content of all seeds contained in a cone of mountain pine is only 6.0 kJ against the 51 and 121 kJ, respectively, of a Norway spruce cone and an Arolla pine cone, the main food resources in the other 2 study areas. Thus, smaller size among male red squirrels in the marginal habitat could be an adaptation to a reduced availability of high‐energy food resources characterized by relatively long handling times (unpublished data; see also data on Scots pine in Wauters et al. 1992).However, this trend was not observed among female red squirrels. We believe this is due to the differential energy investment in reproduction between the sexes. Reaching a high body mass is essential for female red squirrels to enhance their reproductive success. Indeed, the probability to enter estrus (Wauters & Dhondt 1989b), lactation success (Wauters & Dhondt 1995; Wauters & Lens 1995; Wauters et al. 2007), offspring weaning mass and early survival (Wauters et al. 1993), and individual variation in lifetime reproductive success (Wauters & Dhondt 1995) are all strongly and positively related to body mass. Also in this study, lactating females weighed more than non‐breeding ones. Moreover, in the marginal habitat, persistency of adults is low and most females will have only a single reproductive event in their lifetime (Rodrigues et al. 2010); hence having a high body mass should aid females to maximize reproductive output in terms of litter size and early offspring survival (Mari et al. 2008).Finally, heavier males tend to be more successful in mating than males of lower body mass, which should favor larger and heavier animals (Wauters et al. 1990). However, these observations were carried out in high‐density populations where up to 4 males competed for access to the estrus female. In Cancano, densities were so low that, probably only 1 or 2 males took part at a mating chase, which could reduce sexual selection for larger males. Finally, these sex‐specific differences between habitats argue against the hypothesis that squirrels are smaller and weigh less in the poorer‐quality habitat because of having an overall lower food intake.Personality traits and habitat selectionRed squirrels in the marginal edge habitat were more active, active‐explorative and social than in the other 2 higher‐quality areas. Moreover, animals in Cancano were also more explorative than those in Valfurva but not than those living in the Arolla pine forest of Bormio, which is characterized by a higher overall tree‐seed abundance than the mountain pine forest of Cancano, but also by a higher patchiness (forest heterogeneity) (Table 1; see also Salmaso et al. 2009). Hence, individuals that behave more explorative might have an advantage in finding food‐rich patches and optimize home range use (Wauters et al. 2021), explaining why the expression of exploration was similar in Bormio as in the marginal habitat of Cancano. Several studies on wild animals found a non‐random distribution of individuals’ personality traits between heterogeneous habitats (Cerqueira et al. 2016; Holtmann et al. 2017; Brehm & Mortelliti 2021; Rabdeau et al. 2021). For example, in dunnocks (Prunella modularis), personality has been observed having an impact on habitat choice, bolder individuals settled in areas where human disturbance is high, contrary to shier ones that prefer sites with less human disturbance (Holtmann et al. 2017). As suggested in these studies, and in agreement with the “personality‐matching habitat choice” hypothesis (Edelaar et al. 2008), an individual settles in the habitat that is the most suitable for its personality and that best matches the animal's capacity to use this habitat (Edelaar et al. 2008; Holtmann et al. 2017). These authors also suggest that behavioral adjustments over time contribute little to the personality‐habitat relationship (Edelaar et al. 2008; Holtmann et al. 2017). This could explain why, in contrast with our prediction, we did not find a higher plasticity (within‐individual variance) of the personality traits in the marginal habitat compared to the other ones. In our study system, we believe that increased activity and activity‐exploration, but also high exploration scores, result in higher success in acquiring resources, and thus enhance the probability to settle in marginal areas of the distribution range, characterized by poorer habitat quality (Biro & Stamps 2008; Wauters et al. 2021). Similarly, a study by Boone and colleagues (2022) investigating the interaction between the personality of another scatter‐hoarding mammal (deer mice, Peromyscus maniculatus), and seed (food) availability, found that more active/explorative mice choose lower‐quality seeds when food availability was reduced by 50%, in contrast with less active/explorative individuals that continued to select high‐quality seeds both at high and low food availability (Boone et al. 2022). Hence, this study suggests that more active and explorative individuals usually tend to use a wider variety of food resources than less active/explorative ones (Toscano et al. 2016; Boone et al. 2022). Therefore, the presence of more active, explorative and active‐explorative individuals in the edge habitat may also be due to the fact that these personality types can more easily switch and adapt to other, lower‐quality food resources (in our case the seeds of mountain pine cones but also male flowers and buds of mountain pine, rather than the more energetic Norway spruce and Arolla pine seeds) than less active/explorative ones.Together with the personality traits related to habitat‐ and space use, we also found a higher social tendency in the marginal habitat than in other study areas. This could be due to the behavioral syndrome we described (see also Santicchia et al. 2022). In other words, if higher activity and exploration may benefit animals in the marginal habitat and both are associated with higher social‐tendency in a behavioral syndrome, then the squirrels in the marginal habitat will express more social behavior during the MIS test, without there being a selective advantage of behaving more social.Alternative hypotheses: early experience or behavioral plasticityAn alternative explanation of the differences in personality between the study areas may be due to the shaping of squirrels’ behavior by the different pressures that characterized each habitat. In fact, changes in personality may occur throughout the individual's life (Stamps & Groothuis 2010), leading to a weakening of the personality‐matching habitat choice hypothesis. In our study, we did not specifically examine this behavioral plasticity throughout an individual's life, but in general, considerable modifications in personality seem to happen after drastic developmental changes (e.g. sexual maturation) (Cabrera et al. 2021). To account for this, we removed juveniles from our analyses and we performed multiple arena tests in different sessions (over several months to several years) of the same individual to test the repeatability of the trait over time. Moreover, it is reported that an animal's experience during early development could have an effect on its personality as adult (Stamps & Groothuis 2010). However, the majority of red squirrels disperse when between 4 and 10 months old, while some disperse as adults (>1‐year old), over 1 up to 5–10 km, and philopatry is rare (Wauters & Dhondt 1993; Wauters et al. 1994, 2011; Di Pierro et al. 2010). There is genetic evidence that dispersal occurs between 2 of our study sites (Di Pierro et al. 2010) and that also in the marginal site most of the newly caught animals are immigrants (our unpublished data). Hence, the animals monitored here arrived in a given study site coming from several directions and distances, guaranteeing differences in the early experiences of animals settling in that area. Therefore, we think that behavioral plasticity or early life experience did not have a great effect on the differences in personality that we found between the different habitats.A further point to consider is the possibility that squirrels rapidly change their behavior (rapid behavioral change; e.g. within a few days from the settlement in the new area) in response to the environmental pressures (Holtmann et al. 2017). For this study, we did not examine this rapid behavioral plasticity, due to the difficulties to observe personality in the first days of establishment in the new area. However, a deep investigation of behavior before and after the settlement is necessary to examine this alternative.Other factors affecting the expression of personality traitsFemale red squirrels tended to have a higher score for the activity‐exploration trait than males. This could be related to the higher energy investment by females in reproduction (see Morphological differences), producing an advantage in resource acquisition for more active‐explorative individuals. In line with this and with previous studies on red squirrels, heavier squirrels were more explorative than smaller ones (Santicchia et al. 2021, 2022). This association is found also in other tree squirrel species (Tamias striatus: Martin & Réale 2008; Sciurus carolinensis: Santicchia et al. 2019, 2020b).We also observed a seasonal effect on behavioral traits: squirrels in spring–summer were more active, explorative and active‐explorative than in autumn. Animals may show some degree of behavioral plasticity among seasons to cope with predictable abiotic and biotic environmental conditions (e.g. changes in photoperiod, rainfall, temperature, tree‐seed availability), especially at high latitudes and altitudes (Zhang & Buck 2022; see also Beaman et al. 2016). However, only few studies investigated seasonal fluctuations in behavior in free‐living Sciurids (Arctic ground squirrels, Urocitellus parryii: Chmura et al. 2020; Abert's squirrels, Sciurus aberti: Zhang & Buck 2022). In our studied species, the seasonal variation in trait expression could be linked to tree phenology; new fresh cones with mature seeds are available from the end of summer and during autumn and winter (Romeo et al. 2010; Zong et al. 2010). Therefore, augmented activity and exploration in spring–early summer might be linked to higher foraging needs when feeding on low‐energy food sources such as male flowers, buds, or cached seeds that need long searching times (Zong et al. 2010).In conclusion, the evidence from this study suggests that personality has an important role in habitat selection with consequences on animals’ morphology. This study was carried out in red squirrels populations in the altitudinal limit of their distribution range, but the same study system could be replicated in different contexts with different environmental stressors (Wauters et al. 2019; Santicchia et al. 2022) to better understand an individual's choice of habitat (e.g. differences among urban‐rural gradient). Such future studies should also measure fitness components of individuals (survival and/or reproductive success) which would allow to investigate whether natural selection favors different personalities in relation to habitat type (e.g. Santicchia et al. 2018). Also, the habitat‐related patterns in animal personality and the comparisons of among‐ and within‐individual variance in personality traits documented here, contributes to our still limited knowledge about behavioral plasticity differences between populations occurring in different habitats in vertebrates (Dingemanse & Wolf 2013; Hewes et al. 2017). Particularly, our findings that only for exploration, and not for the other behavioral traits, the within‐individual variance differed among populations strengthens the hypothesis that it is primarily the animal's personality that determined the individual's choice of habitat, and that behavioral plasticity was not strongly associated with habitat selection.ACKNOWLEDGMENTSThe authors thank the Stelvio National Park for allowing fieldwork in the 3 study areas. The authors are also grateful to the field technicians and thesis students for helping with data collection. This study is part of the ASPER (Alpine Squirrel Population Ecology Research) project (paper #36).Open Access Funding provided by Universita degli Studi dell'Insubria within the CRUI‐CARE Agreement.REFERENCESBeaman JE, White CR, Seebacher F (2016). Evolution of plasticity: Mechanistic link between development and reversible acclimation. Trends in Ecology & Evolution 31, 237–49.Bell AM, Hankison SJ, Laskowski KL (2009). The repeatability of behaviour: A meta‐analysis. Animal Behaviour 77, 771–83.Biro PA, Stamps JA (2008). Are animal personality traits linked to life‐history productivity? Trends in Ecology & Evolution 23, 361–8.Bolnick DI, Svanbäck R, Fordyce JA et al. (2003). The ecology of individuals: Incidence and implications of individual specialization. The American Naturalist 161, 1–28.Boon AK, Réale D, Boutin S (2008). Personality, habitat use, and their consequences for survival in North American red squirrels Tamiasciurus hudsonicus. Oikos 117, 1321–8.Boone SR, Brehm AM, Mortelliti A (2022). Seed predation and dispersal by small mammals in a landscape of fear: effects of personality, predation risk and land‐use change. Oikos 2022, e08232.Boutin S, Larsen K (1993). Does food availability affect growth and survival of males and females differently in a promiscuous small mammal, Tamiasciurus hudsonicus. Journal of Animal Ecology 62, 364–70.Boutin S, Wauters LA, McAdam AG, Humphries MM, Tosi G, Dhondt AA (2006). Anticipatory reproduction and population growth in seed predators. Science 314, 1928–30.Brehm AM, Mortelliti A (2021). Land‐use change alters associations between personality and microhabitat selection. Ecological Applications 31, e02443.Cabrera D, Nilsson JR, Griffen BD (2021). The development of animal personality across ontogeny: A cross‐species review. Animal Behaviour 173, 137–44.Carter AJ, Feeney WE, Marshall HH, Cowlishaw G, Heinsohn R (2013). Animal personality: What are behavioural ecologists measuring? Biological Reviews 88, 465–75.Cerqueira M, Rey S, Silva T, Featherstone Z, Crumlish M, MacKenzie S (2016). Thermal preference predicts animal personality in Nile tilapia Oreochromis niloticus. Journal of animal ecology 85, 1389–400.Chmura HE, Zhang VY, Wilbur SM, Barnes BM, Buck CL, Williams CT (2020). Plasticity and repeatability of activity patterns in free‐living Arctic ground squirrels. Animal Behaviour 169, 81–91.Cote J, Clobert J, Brodin T, Fogarty S, Sih A (2010). Personality‐dependent dispersal: Characterization, ontogeny and consequences for spatially structured populations. Philosophical Transactions of the Royal Society B 365, 4065–76.Dall SRX, Houston AI, McNamara JM (2004). The behavioural ecology of personality: Consistent individual differences from an adaptive perspective. Ecology Letters 7, 734–9.Di Pierro E, Bertolino S, Martinoli A, Preatoni D, Tosi G, Wauters LA (2010). Estimating offspring production using capture‐mark‐recapture and genetic methods in red squirrels. Ecological Research 25, 395–402.Di Pierro E, Ghisla A, Wauters LA et al. (2011). The effects of seed availability on habitat use by a specialist seed predator. European Journal of Wildlife Research 57, 585–95.Dingemanse NJ, Kazem AJN, Réale D, Wright J (2010). Behavioural reaction norms: Animal personality meets individual plasticity. Trends in Ecology & Evolution 25, 81–9.Dingemanse NJ, Bouwman KM, van de Pol M et al. (2012). Variation in personality and behavioural plasticity across four populations of the great tit Parus major. Journal of Animal Ecology 81, 116–26.Dingemanse NJ, Wolf M (2013). Between‐individual differences in behavioural plasticity within populations: causes and consequences. Animal Behaviour 85, 1031–9.Edelaar P, Siepielski AM, Clobert J (2008). Matching habitat choice causes directed gene flow: A neglected dimension in evolution and ecology. Evolution 62, 2462–72.Endler JA (1977). Geographical Variation, Speciation and Clines. Princeton University Press, Princeton, NJ.Gelman A, Rubin DB (1992). Inference from iterative simulation using multiple sequences. Statistical Science 7, 457–72.Geweke J (1992). Evaluating the accuracy of sampling‐based approaches to the calculation of posterior moments. In: Bernardo JM, Berger AP, Dawid AP, Smith AFM, eds. Bayesian Statistics. Oxford University Press, Oxford, pp. 169–93.Hadfield JD (2010). MCMC methods for multi‐response generalized linear mixed models: The MCMCglmm R package. Journal of Statistical Software 33, 1–22.Hänninen L, Pastell M (2009). CowLog: Open‐source software for coding behaviors from digital video. Behavior Research Methods 41, 472–6.Hewes ME, Delventhal K, Chaves‐Campos J (2017). Behavioral plasticity and consistency in the naked‐footed mouse (Peromyscus nudipes) with habitat disturbance. Journal of Ethology 35, 279–92.Hoffman AA, Merilä J (1999). Heritable variation and evolution under favourable and unfavourable conditions. Trends in Ecology and Evolution 12, 460–1.Holtmann B, Santos ESA, Lara CE, Nakagawa S (2017). Personality‐matching habitat choice, rather than behavioural plasticity, is a likely driver of a phenotype–environment covariance. Proceedings of the Royal Society B 284, 20170943.Kawecki TJ, Ebert D (2004). Conceptual issues in local adaptation. Ecology Letters 7, 1225–41.Koprowski J (1994a). Sciurus niger. Mammalian Species 479, 1–9.Koprowski J (1994b). Sciurus carolinensis. Mammalian Species 480, 1–9.Lurz PWW, Garson PJ, Wauters LA (2000). Effects of temporal and spatial variations in food supply on the space and habitat use of red squirrels, Sciurus vulgaris L. Journal of Zoology 251, 167–78.Lurz PWW, Lloyd AJ (2000). Body weights in grey and red squirrels: Do seasonal weight increases occur in conifer woodland? Journal of Zoology 252, 539–43.Mari V, Martini S, Romeo C et al. (2008). Record litter size in the Eurasian red squirrel (Sciurus vulgaris). Hystrix, the Italian Journal of Mammalogy 19, 61–5.Martin JGA, Réale D (2008). Animal temperament and human disturbance: Implications for the response of wildlife to tourism. Behavioural Processes 77, 66–72.Martin JGA, Nussey DH, Wilson AJ, Réale D (2011). Measuring individual differences in reaction norms in field and experimental studies: A power analysis of random regression models: Power analysis of random regression models. Methods in Ecology and Evolution 2, 362–74.Mazzamuto MV, Cremonesi G, Santicchia F, Preatoni D, Martinoli A, Wauters LA (2019). Rodents in the arena: A critical evaluation of methods measuring personality traits. Ethology Ecology & Evolution 31, 38–58.McAdam AG, Boutin S, Réale D, Berteaux D (2002). Maternal effects and the potential for evolution in a natural population of animals. Evolution 56, 846–51.McAdam AG, Boutin S (2003). Effects of food abundance on genetic and maternal variation in the growth rate of juvenile red squirrels. Journal of Evolutionary Biology 16, 1249–56.Nash DJ, Seaman RN (1977). Sciurus aberti. Mammalian Species 80, 1–5.R Core Team (2020). R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. Available from URL: https://www.R‐project.org/Rabdeau J, Arroyo B, Mougeot F, Badenhausser I, Bretagnolle V, Monceau K (2021). Do human infrastructures shape nest distribution in the landscape depending on individual personality in a farmland bird of prey? Journal of Animal Ecology 90, 2848–58.Réale D, Reader SM, Sol D, McDougall PT, Dingemanse NJ (2007). Integrating animal temperament within ecology and evolution. Biological Reviews 82, 291–318.Rodrigues D, Wauters LA, Romeo C et al. (2010). Living on the edge: Can Eurasian red squirrels (Sciurus vulgaris) persist in extreme high‐elevation habitats? Arctic, Antarctic, and Alpine Research 42, 106–12.Romeo C, Wauters LA, Preatoni D, Tosi G, Martinoli A (2010). Living on the edge: Space use of Eurasian red squirrels in marginal high‐elevation habitat. Acta Oecologica 36, 604–10.Royauté R, Dochtermann NA (2021). Comparing ecological and evolutionary variability within datasets. Behavioral Ecology and Sociobiology 75, 127.Salmaso F, Molinari A, Di Pierro E et al. (2009). Estimating and comparing food availability for tree‐seed predators in typical pulsed‐resource systems: alpine conifer forests. Plant Biosystems 143, 258–67.Santicchia F, Gagnaison C, Bisi F et al. (2018). Habitat‐dependent effects of personality on survival and reproduction in red squirrels. Behavioral Ecology and Sociobiology 72, 134.Santicchia F, Romeo C, Ferrari N et al. (2019). The price of being bold? Relationship between personality and endoparasitic infection in a tree squirrel. Mammalian Biology 97, 1–8.Santicchia F, Wauters LA, Piscitelli AP et al. (2020a). Spillover of an alien parasite reduces expression of costly behaviour in native host species. Journal of Animal Ecology 89, 1559–69.Santicchia F, Wauters LA, Dantzer B et al. (2020b). Relationships between personality traits and the physiological stress response in a wild mammal. Current Zoology 66, 197–204.Santicchia F, Van Dongen S, Martinoli A, Preatoni D, Wauters LA (2021). Measuring personality traits in Eurasian red squirrels: A critical comparison of different methods. Ethology 127, 187–201.Santicchia F, Wauters LA, Tranquillo C et al. (2022). Invasive alien species as an environmental stressor and its effects on coping style in a native competitor, the Eurasian red squirrel. Hormones and Behavior 140, 105127.Sih A, Bell A, Johnson JC (2004). Behavioral syndromes: An ecological and evolutionary overview. Trends in Ecology & Evolution 19, 372–8.Sih A, Cote J, Evans M, Fogarty S, Pruitt J (2012). Ecological implications of behavioural syndromes. Ecology Letters 15, 278–89.Spiegel O, Leu ST, Bull CM, Sih A (2017). What's your move? Movement as a link between personality and spatial dynamics in animal populations. Ecology Letters 20, 3–18.Stamps JA, Groothuis TGG (2010). Developmental perspectives on personality: Implications for ecological and evolutionary studies of individual differences. Philosophical Transactions of the Royal Society B: Biological Sciences 365, 4029–41.Stoffel MA, Nakagawa S, Schielzeth H (2017). rptR: Repeatability estimation and variance decomposition by generalized linear mixed‐effects models. Methods in Ecology and Evolution 8, 1639–44.Toscano BJ, Gownaris NJ, Heerhartz SM, Monaco CJ (2016). Personality, foraging behavior and specialization: integrating behavioral and food web ecology at the individual level. Oecologia 182, 55–69.Tranquillo C, Villa F, Wauters LA et al. (2022). Physiological stress and spatio‐temporal fluctuations of food abundance and population density in Eurasian red squirrels. Hystrix, the Italian Journal of Mammalogy 33, 26–32.Verbeke G, Molenberghs G (2000). Linear Mixed Models for Longitudinal Data. Springer, New York.Wauters LA, Dhondt AA (1989a). Variation in length and body weight of the red squirrel in two different habitats. Journal of Zoology 217, 93–106.Wauters LA, Dhondt AA (1989b). Body weight, longevity and reproductive success in red squirrels (Sciurus vulgaris). Journal of Animal Ecology 58, 637–51.Wauters LA, Dhondt AA, De Vos R (1990). Factors affecting male mating success in red squirrels (Sciurus vulgaris). Ethology Ecology & Evolution 2, 195–204.Wauters LA, Dhondt AA (1992). Spacing behaviour of red squirrels, Sciurus vulgaris: Variation between habitats and the sexes. Animal Behaviour 43, 297–311.Wauters LA, Swinnen C, Dhondt AA (1992). Activity budget and foraging behaviour of red squirrels (Sciurus vulgaris) in coniferous and deciduous habitats. Journal of Zoology 227, 71–86.Wauters LA, Bijnens L, Dhondt AA (1993). Body mass at weaning and juvenile recruitment in the red squirrel. Journal of Animal Ecology 62, 280–6.Wauters LA, Dhondt AA (1993). Immigration patterns and success in red squirrels. Behavioral Ecology and Sociobiology 33, 159–67.Wauters LA, Matthysen E, Dhondt AA (1994). Survival and lifetime reproductive success in dispersing and resident red squirrels. Behavioral Ecology and Sociobiology 34, 197–201.Wauters LA, Dhondt AA (1995). Components of lifetime reproductive success of female Eurasian red squirrels. Oikos 72, 402–10.Wauters LA, Lens L (1995). Effects of food abundance and density on red squirrel (Sciurus vulgaris) reproduction. Ecology 76, 2460–9.Wauters LA, Gurnell J, Preatoni D, Tosi G (2001). Effects of spatial variation in food availability on spacing behaviour and demography of Eurasian red squirrels. Ecography 24, 525–38.Wauters LA, Vermeulen M, Van Dongen S et al. (2007) Effects of spatio‐temporal variation in food supply on red squirrel Sciurus vulgaris body size and body mass and its consequences for some fitness components. Ecography 30, 51–65.Wauters LA, Githiru M, Bertolino S, Molinari A, Tosi G, Lens L (2008). Demography of alpine red squirrel populations in relation to fluctuations in seed crop size. Ecography 31, 104–14.Wauters LA, Preatoni D, Martinoli A, Verbeylen G, Matthysen E (2011). No sex bias in natal dispersal of Eurasian red squirrels. Mammalian Biology 76, 369–72.Wauters LA, Mazzamuto MV, Santicchia F, Van Dongen S, Preatoni DG, Martinoli A (2019). Interspecific competition affects the expression of personality‐traits in natural populations. Scientific Reports 9, 11189.Wauters LA, Mazzamuto MV, Santicchia F et al. (2021). Personality traits, sex and food abundance shape space use in an arboreal mammal. Oecologia 196, 65–76.Wilson AJ, Réale D, Clements MN et al. (2010). An ecologist's guide to the animal model. Journal of Animal Ecology 79, 13–26.Zhang VY, Buck CL (2022). Seasonal patterns in behavior and glucocorticoid secretion of a specialist Holarctic tree squirrel (Sciurus aberti). Journal of Comparative Physiology B 192, 541–59.Zong C, Wauters LA, Van Dongen S et al. (2010). Annual variation in predation and dispersal of Arolla pine (Pinus cembra L.) seeds by Eurasian red squirrels and other seed‐eaters. Forest Ecology and Management 260, 587–94. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Integrative Zoology Wiley

Living on the edge: morphological and behavioral adaptations to a marginal high‐elevation habitat in an arboreal mammal

Loading next page...
 
/lp/wiley/living-on-the-edge-morphological-and-behavioral-adaptations-to-a-YdLmw1p0U6

References (51)

Publisher
Wiley
Copyright
© 2023 International Society of Zoological Sciences, Institute of Zoology/Chinese Academy of Sciences and John Wiley & Sons Australia, Ltd.
ISSN
1749-4868
eISSN
1749-4877
DOI
10.1111/1749-4877.12679
Publisher site
See Article on Publisher Site

Abstract

INTRODUCTIONLocal environmental conditions, and their temporal changes, can exert selective pressures on animals, resulting in phenotypic variation among populations living in different habitats (Kawecki & Ebert 2004). These local phenotypic adaptations often refer to physiological or morphological traits subject to natural selection, but also to behavioral traits, that are assumed to be more flexible with animals capable of adjusting their behavior in response to fluctuations in environmental conditions (Bolnick et al. 2003; Holtmann et al. 2017). However, both theoretical and empirical studies on animal personality, defined as behavioral differences among individuals of a species that are consistent over time and across situations (e.g. Réale et al. 2007; Carter et al. 2013), suggest that behavioral plasticity may be restrained by an individual's personality type, and that animals select habitats that best fit their personality (Dall et al. 2004; Sih et al. 2004; Holtmann et al. 2017). In other words, personality traits may reduce an individual's ability to cope with certain environmental conditions and, consequently, result in non‐random distribution of behavioral phenotypes across different habitats (Cote et al. 2010; Holtmann et al. 2017; Wauters et al. 2019).For mammals living in marginal habitats on the edge of a species’ distribution range, environmental conditions can be extreme and exert strong selective pressure on individuals, different from those acting in more central, higher‐quality habitats, characterized by more abundant resources and, in some cases, less extreme weather conditions (Endler 1977; Rodrigues et al. 2010). This could result in habitat‐dependent phenotypic differences in body size and/or personality for mammals living in heterogeneous environments (Wauters et al. 2007; Hewes et al. 2017; Brehm & Mortelliti 2021; Boone et al. 2022). Also, phenotypic differences between animals from marginal and high‐quality areas may be due to individuals showing higher behavioral plasticity in populations living at the distribution edge (Dingemanse et al. 2010).Depending heavily on tree seeds as their main food resource throughout most of the year, arboreal squirrels in temperate regions typically occur in resource‐limited producer–consumer systems (Boutin & Larsen 1993; Koprowski 1994a; Boutin et al. 2006; Wauters et al. 2008). In heterogeneous landscapes, spatio‐temporal variation in tree seed abundance varies greatly, creating a mosaic of habitats that differ in their quality for tree squirrel populations in terms of overall food availability as well as temporal fluctuations in the size of tree‐seed crops (Koprowski 1994a,b; Salmaso et al. 2009; Di Pierro et al. 2011; Tranquillo et al. 2022). Moreover, morphological and/or behavioral differences and variation in personality traits are related to space and/or habitat use (Wauters et al. 2007, 2021; Boon et al. 2008) and to variation in survival and/or reproductive success (McAdam et al. 2002; Wauters et al. 2007; Santicchia et al. 2018). Therefore, tree squirrels are particularly interesting models to study habitat‐related phenotypic differences in body size and/or personality for mammals living in heterogeneous environments.Here, we used the Eurasian red squirrel (Sciurus vulgaris) to investigate if phenotypic variation differs between individuals from a population living in a marginal low‐quality habitat on the edge of the elevational distribution and animals occurring in moderate to high‐quality conifer forests at similar or lower elevations in the Italian Alps. As explained above, habitat quality refers to average seed‐crop size and annual fluctuations in seed production of conifers, the main food resource for red squirrels in Alpine forests (Wauters et al. 2008; Salmaso et al. 2009; Rodrigues et al. 2010; Tranquillo et al. 2022 for detailed measurements of food abundance in these study sites).Eurasian red squirrels occupying extreme, high‐elevation habitat differ from conspecifics in higher quality habitats at lower elevations in population dynamics (Mari et al. 2008; Rodrigues et al. 2010) and space use (Romeo et al. 2010). Space use differs between high‐elevation marginal habitat of dwarf mountain pine (Pinus mugo) and high‐quality Scots pine (Pinus sylvestris) forest at lower‐elevations (Romeo et al. 2010). In fact, home ranges and core areas are larger in the marginal habitat, with more pronounced sex differences in core area size (Romeo et al. 2010). In contrast, patterns of core area overlap and intrasexual territoriality are similar as in other populations from various habitat types (Wauters & Dhondt 1992; Lurz et al. 2000; Wauters et al. 2001; Di Pierro et al. 2011). However, low densities in the edge population (Rodrigues et al. 2010) result in higher female by males overlap in spring–summer, when males expand their home ranges during the mating season to increase access to estrus females (Romeo et al. 2010). Moreover, squirrels’ spacing behavior responds more strongly to variation in food availability and/or density fluctuations in marginal than in high‐quality habitats (Romeo et al. 2010; Wauters et al. 2021), suggesting apparent behavioral plasticity to respond to the ecological constraints in marginal habitats.We measured morphological (body mass and hind foot length) and behavioral (exploration, activity, activity‐exploration and social tendency) traits of red squirrels. Hind foot length is strongly correlated with total head and body length and thus is considered a reliable proxy for body size (Wauters & Dhondt 1989a). Behavioral traits have been shown to determine an animal's personality (Mazzamuto et al. 2019; Wauters et al. 2019; Santicchia et al. 2020a, 2021).In particular, we predicted: (i) Differences in body size and mass between the habitats, with smaller squirrels of lower body mass in the marginal, poorer‐quality habitat, where (based on earlier studies of habitat‐related differences in foot length and body mass of red squirrels, Wauters et al. 2007) smaller squirrels may be advantaged since they have lower daily energy‐requirements than larger and heavier individuals. Alternatively, squirrels may be smaller and weigh less in the poorer‐quality habitat because of lower food intake. (ii) Differences in the expression of behavioral traits, with squirrels in the edge population having higher average levels of activity and exploration than in high‐quality habitats, since active‐explorative personality types should be better adapted at exploring low‐abundance food resources (see also Spiegel et al. 2017). (iii) A more pronounced advantage of higher behavioral plasticity in edge than in high‐quality habitats, resulting in higher within‐individual variance of behavioral traits in the edge population (e.g. Dingemanse & Wolf 2013).MATERIALS AND METHODSStudy areas, trapping, and handlingAs part of a long‐term research project on spatio‐temporal variation in producer‐consumer dynamics in pulsed resource systems, tree‐seed abundance, the population dynamics, life‐history, personality traits and space use of red squirrels were studied in 3 populations occupying different types of conifer forests (Wauters et al. 2007, 2008; Salmaso et al. 2009; Rodrigues et al. 2010; Santicchia et al. 2021; Tranquillo et al. 2022). These study areas, located in Lombardy, Northern Italy, are characterized by subalpine mixed conifer forests that differ in the dominant tree species. Cancano (CAN; 46°33ʹN, 10°15ʹE, elevation from 1940 m to 1970 m a.s.l.) extends over 54 ha in a high‐elevation valley, and it is part of a monospecific forest of dwarf mountain pine, Pinus mugo, whose seeds are the only major food resource for red squirrels (Rodrigues et al. 2010; Romeo et al. 2010). Bormio (BOR; 46°27ʹN, 10°30ʹE, elevation from 1950 m to 2130 m a.s.l.; size 92 ha) is dominated by Arolla pine, Pinus cembra (73.2%), with larch, Larix decidua (17.8%) and scarce Norway spruce, Picea abies (8.0%), and dead trees (1.0%) (Wauters et al. 2007; Santicchia et al. 2021). Valfurva (VAL; 46°27ʹN, 10°31ʹE, elevation from 1650 m to 1870 m a.s.l.) is located on an east‐north‐east slope in Valfurva valley and covers 78 ha. It is characterized by Norway spruce (88.9%) with a lower presence of Arolla pine (6.2%), dead trees (3.0%), and larch (1.9%) (Wauters et al. 2007; Santicchia et al. 2021).Based on elevation and habitat quality (average seed‐crop size and annual fluctuations in seed production, see also Table 1), Cancano is a marginal edge‐habitat, Bormio is a high‐elevation habitat of moderate habitat‐quality with harsh winter conditions, and Valfurva is a high‐quality habitat, typical for the dominant subalpine conifer forest‐type of the central Italian‐Alps (Wauters et al. 2007; Rodrigues et al. 2010).1TableRed squirrel density estimated using MNA and food abundance expressed in energy‐content of conifer seeds (MJ·ha−1, mean ± SE), in the 3 study areas per yearCancano (54 ha) (0.35 traps·ha−1) (n = 68; ID = 34)Bormio (92 ha) (0.22 traps·ha−1) (n = 55; ID = 36)Valfurva (77 ha) (0.26 traps·ha−1) (n = 226; ID = 92)YearRed squirrel density (ind·ha−1)Food abundance (MJ·ha−1)Red squirrel density (ind·ha−1)Food abundance (MJ·ha−1)Red squirrel density (ind·ha−1)Food abundance (MJ·ha−1)20160.17128 ± 130.152301 ± 2990.402863 ± 32820170.20294 ± 350.131605 ± 2080.491361 ± 10320180.20243 ± 580.111939 ± 2160.381559 ± 27020190.07234 ± 320.101273 ± 1490.35170 ± 1920200.11496 ± 840.255404 ± 59720210.09108 ± 230.1947 ± 14Size of the study areas (ha) and trap density (traps·ha−1) between brackets. Arena test sample sizes (n) and number of different animals (ID) between brackets, for each study areas.Population dynamics and life‐history traits were studied by Capture‐Mark‐Recapture (CMR), and food abundance was estimated by annual cone counts in Cancano (2005–2021), Bormio (2002–2019), and Valfurva (2002–2021). Squirrel density was calculated by dividing the minimum number of animals known to be alive (MNA) by study area size (ha) (Wauters et al. 2008). Details of CMR methods and conifer seed‐crop size are given elsewhere (Wauters et al. 2008; Salmaso et al. 2009; Rodrigues et al. 2010). Briefly, CMR sessions were carried out each year in May–June (spring–summer) and in September–October (autumn), using live traps (Tomahawk trap model 202, Tomahawk Live Trap Co., Tomahawk, Wisconsin, USA) placed with baling wire on tree trunks at breast height or on the ground (Wauters et al. 2008; Tranquillo et al. 2022). We pre‐baited traps with hazelnuts once per week for about 1 month prior to trapping; trapping sessions lasted 4 to 5 days, and we checked traps 2 to 3 times per day to reduce the time a squirrel was confined in the trap. A trapped squirrel was flushed from the trap in a zipper‐tube handling bag, and, at first capture, individually marked with a numbered metal ear tag (Monel 1005 1L1 National Band and Tag Co, Newport, Kentucky, USA). Sex and reproductive condition were determined based on external genitalia (details in Wauters et al. 2007; Tranquillo et al. 2022). Males were classified based on testes size and position (testes abdominal, testes scrotal); females as non‐breeding, post‐estrus (pregnant), or lactating. At each capture, an animal was weighed to the nearest 5 g with a spring balance (Pesola AG, Baar, Switzerland), and the length of the right hind foot (claws excluded) was measured (0.5 mm precision) with a thin wing ruler.Personality quantification: arena testWe studied the personality of red squirrels performing the open field test (OFT; 4 min), to estimate an animal's activity and exploration in a novel environment, followed by the mirror image stimulation test (MIS; 3 min) to determine the degree of sociability or avoidance (social tendency), aggressiveness, and being alert toward a conspecific, as well as the tendency for expressing behaviors that define a combined activity‐exploration trait (Mazzamuto et al. 2019; Wauters et al. 2019; Santicchia et al. 2020a, 2022). Both tests were carried out during the same capture event in a portable arena, placed within 20 m from the trap location, where a previously identified squirrel was released by opening a sliding door and allowing the animal to move from the handling bag into the arena (Mazzamuto et al. 2019; Santicchia et al. 2021, 2022). At the end of the experiment, the squirrel was released and the arena was cleaned with 90% ethylic alcohol. More details on the arena test are given in previous studies (Santicchia et al. 2021, 2022). The time that an individual spent in each behavioral state (behaviors defined in Table S1, Supporting Information) was calculated using the CowLog 3.0.2 software (Hänninen & Pastell 2009). Next, we applied the expert‐based method, defined and validated by Mazzamuto et al. (2019), to obtain new scores for 7 personality‐linked variables (Ethogram description and personality traits in Table S1, Supporting Information).To check the assumptions of repeatability of the traits, squirrels were tested multiple times in different CMR sessions to have repeated measures (individuals tested multiple times: n = 78, mean ± SE = 2.63 ± 0.10, range = 2–9). Nevertheless, to increase the power of the test, squirrels with only one arena test were also included in the adjusted repeatability estimates (Martin et al. 2011; Santicchia et al. 2022).Body size and body mass analysesA total of 193 individuals (114 males and 79 females) were sampled to assess the differences in morphological traits (see also Results, Differences in body size and body mass). Squirrels were classified as juveniles or adults based on external genitalia and body mass (<225 g; Wauters et al. 2007 and references therein), and juveniles were excluded from the analyses to avoid any bias in relation to our predictions.We explored the effects of study area and sex on hind foot length with a general linear model (GLM). Mean foot length of each individual was used as dependent variable (n = 193 different squirrels), adding study area and sex as categorical variables and the study area by sex interaction.Next, we investigated whether study area affected variation in body mass. Since body mass fluctuates seasonally and with reproductive condition, and since effects of reproductive condition differ between the sexes (Wauters & Lens 1995; Wauters et al. 2007), we used Linear Mixed Models (LMM), with multiple body mass measurements for individual squirrels, for each sex separately. Body mass was used as dependent variable, hind foot length as explanatory continuous variable, and study area, season, and reproductive condition as fixed effects. We further included all 2‐factor interactions with study area in the full model, and added squirrel identity as a random intercept to account for repeated measures over time. A model with compound symmetry correlation structure best fitted the data (based on Schwarz's Bayesian Information Criteria). We tested interactions and fixed effects using Type‐III SS and performed model selection using a backward procedure based on P‐values (eliminating parameters with partial P > 0.15). Degrees of freedom and standard errors of F‐ and t‐tests were obtained using the Kenward–Rogers method (Verbeke & Molenberghs 2000; Wauters et al. 2007), and pairwise comparisons were based on differences between least square means (DLSM).Personality analysesA total of 162 squirrels (100 males and 62 females) were behaviorally tested (see also Results, Behavioral differences). To account for possible confounding effect due to modification in personality with developmental changes (Cabrera et al. 2021; see also Discussion for details), we excluded juvenile individuals from this dataset.Since personality traits are defined as behaviors consistent over time and under different contexts (Réale et al. 2007; Carter et al. 2013) we first calculated the adjusted repeatability (R) of each expert‐based group (see Table S2, Supporting Information). Repeatability was estimated using a LMM through the function rpt, with the R package rptR (Stoffel et al. 2017). We performed a model for each square‐root‐transformed value of expert‐based group (activity, exploration and immobility from OFT and sociability, avoidance, alert and activity‐exploration from MIS) as dependent variable. Likewise, we estimated the adjusted repeatability also for social tendency, a variable that combines avoidance and sociability scores (Pearson correlation coefficient sociability‐avoidance: r = −0.87; n = 349; P < 0.0001) and measures the tendency to react more or less social or amenable to the mirror image (Social tendency = log[(sociability score + (1 − avoidance score))/2], Santicchia et al. 2020a).Each model had squirrel identity (ID) as random intercept and sex, arena test order (first test was coded as 1, subsequent tests were coded as 0), year (from 2016 to 2021), and number of days since the previous arena test as fixed effects. As in previous studies (Santicchia et al. 2020a,b, 2021, 2022), we included in the models arena test order and number of days since the previous arena test, to account for habituation and temporal proximity between 2 arena tests (Dingemanse et al. 2012; Mazzamuto et al. 2019). The likelihood ratio test (LRT) was used to test for significance of the random intercept of each model (Martin & Réale 2008). Full repeatability model outputs are provided in Table S2, Supporting Information.We then investigated whether behavioral traits expression changed between study areas using a Bayesian generalized linear mixed effects model based on a Markov Chain Monte Carlo algorithm with the R package MCMCglmm version 2.32 (Hadfield 2010). This model included 4 behavioral traits (activity and exploration from the OFT, activity‐exploration and social tendency from MIS) as dependent variable, that were previously centered and scaled [(xi – mean x)/SD x] with a Gaussian residual error distribution. Squirrel identity (ID) was added as random effect since repeated observations were present, and to estimate both among‐individual and within‐individual variation of the dependent variables. Study area, sex, body mass (centered and scaled), season (spring–summer [May–June–July]; autumn [September–October]) and arena test order were included as fixed effects. Posterior distribution of the model was based on 1 050 000 iterations with a burn‐in of 50 000 iterations and thinning of 40, such that 25 000 iterations were used to obtain point estimates and 95% credibility intervals (CIs). We used an inverse‐gamma uninformative prior, for the random effect and the residual variation (Wilson et al. 2010). We applied the Gelman–Rubin statistic (Gelman & Rubin 1992) and Geweke diagnostic (Geweke 1992) to confirm model consistency and convergence. Full model output is provided in Table S3, Supporting Information.In a final step, to obtain the among‐individual and within‐individual variances and the estimates of the correlations among the behavioral traits for each study area, we performed other 3 MCMCglmm models using 3 restricted dataset (one for each study area), in which we included the same dependent variables, fixed effects and random effect as in the main model. This approach allowed us to estimate the differences between study areas in the among‐ and within‐individual variances (Royauté & Dochtermann 2021; Santicchia et al. 2022). Statistical analyses were carried out using the software R (R Core Team 2020).Ethical noteTrapping, marking and handling procedures and arena test experiments complied with the Guidelines for the treatment of animals in behavioral research and teaching (Animal Behaviour, 2020, 159, I–XI; https://doi.org/10.1016/j.anbehav.2019.11.002). Legal requirements according to the Italian Wildlife Protection and Hunting Law L.N. 157 from 1992 and fieldwork was approved by authorization decrees n. 11190 of 29/11/2013, n. 9523 of 15/10/2014, and n. 198 of 13/01/2017 from Direzione Generale Agricoltura, Regione Lombardia; and the permission Protocol N. 414 of 28/02/2014 of the Stelvio National Park.RESULTSDifferences in body size and body massHind foot length in red squirrels (n = 193) varied between 54 and 61 mm (overall mean ± SD = 57.6 ± 1.1 mm). Foot length differed between study areas and there was a significant study area × sex interaction (Area effect F2,187 = 11.24; P < 0.0001; Sex effect F1,187 = 1.27; P = 0.26; area × sex F2,187 = 4.88; P = 0.009). Female squirrels in Valfurva were slightly larger than in Bormio (Table 2a, DLSM P = 0.049, all other pairwise comparisons P > 0.30). Males in Valfurva were larger than in Bormio (DLSM P = 0.041) while males in Cancano were the smallest (DLSM VAL‐CAN P < 0.0001; BOR‐CAN P = 0.0008; Table 2a). Consequently, there was no size difference between the sexes in the populations in Valfurva and Bormio, while in Cancano females were significantly larger than males (Table 2a: DLSM P = 0.004).2Table(a) Mean right hind foot length (±SD) of red squirrels in 3 populations in the Italian Alps; data by sex and for the sexes combined; sample size is given between brackets. (b) Mean body mass (±SD) of male and female red squirrels in 3 populations in the Italian Alps; sample size is given between bracketsFemalesMalesBoth sexesa) Foot lengthValfurva(46) 57.8 ± 1.1(60) 58.0 ± 1.1(106) 57.9 ± 1.1Bormio(17) 57.3 ± 0.9(31) 57.5 ± 1.0(48) 57.4 ± 0.9Cancano(16) 57.5 ± 0.9(23) 56.5 ± 1.0(39) 56.9 ± 1.1b) Body massValfurva(140) 315 ± 27(275) 305 ± 18(415) 308 ± 22Bormio(25) 299 ± 26(74) 290 ± 19(99) 293 ± 21Cancano(50) 311 ± 31(100) 285 ± 16(150) 294 ± 25Body mass of subadult and adult red squirrels (n = 664, with repeated measures of 193 different animals) varied between 235 and 380 g (overall mean ± SD = 303 ± 24 g). For females we obtained 215 measures of 79 individuals (full LMM reported in Table S4, Supporting Information). In the final model, female body mass increased with foot length (estimate ± SE = 25.2 ± 4.7; t79 = 5.38; P < 0.0001), but the slope differed between study areas (foot length × study area F2,92 = 4.37; P = 0.015). The increase in body mass with foot length was stronger in the marginal habitat Cancano than in the high‐quality habitat Valfurva (difference in slope 15.2 ± 5.2; t82 = 2.91; P = 0.005; Fig. 1). Overall, females weighed slightly more in Valfurva than in Bormio, with Cancano intermediate (Table 2b, area effect F2,92 = 4.38; P = 0.015; Fig. 1), and lactating females had a higher body mass than animals in post‐estrus or non‐breeding (F2,191 = 25.9; P < 0.0001, DLSM all P < 0.01).1FigureRelationship between foot length and body mass of female red squirrels in Bormio (solid line, triangles), Cancano (dotted line, crosses), and Valfurva (dashed line, circles). The lines represent the predicted relationships, shaded areas represent the 95% confidence intervals, and symbols represent observed values.For male red squirrels, we used 449 body mass measures of 114 individuals (full LMM reported in Table S4, Supporting Information). In the final model, male body mass increased with foot length (estimate ± SE = 7.4 ± 1.3; t128 = 5.78; P < 0.0001; Fig. 2) and on average males were heavier in Valfurva than in the other 2 areas (Table 2b; area effect F2,100 = 6.32; P = 0.003; DLSM VAL > BOR P = 0.005; VAL > CAN P = 0.005; BOR and CAN P = 0.77). Males in breeding condition weighed more than non‐breeding males (estimate ± SE = 9.6 ± 2.7; t426 = 3.57; P = 0.0004), and body mass increased from spring–summer to autumn (estimate ± SE = 9.5 ± 2.7; t421 = 3.57; P = 0.0005).2FigureRelationship between foot length and body mass of male red squirrels. The solid line represents the predicted relationship, shaded areas represent the 95% confidence intervals, and diamonds represent observed values.Behavioral differencesWe estimated the adjusted repeatability of the 7 expert‐based traits and of social tendency, using a total of 349 arena tests on 162 different red squirrels (100 males and 62 females). Immobility, activity, and activity‐exploration had moderate repeatability (R range: 0.28 to 0.37; see also Bell et al. 2009), while exploration and social tendency had low repeatability (R = 0.14 and 0.15, respectively; Table S2, Supporting Information). The repeatability of sociability, avoidance, and alert was low and not significant (Table S2, Supporting Information). Immobility was strongly and negatively correlated with activity (r = −0.93; n = 349; P < 0.0001); therefore, this trait was not included in the multivariate models.We found several significant differences in personality traits between study areas. Squirrels had higher activity, activity‐exploration and social tendency in Cancano than in Bormio and Valfurva (Table 3; Fig. 3). Squirrels living in Cancano were also significantly more explorative than conspecifics living in Valfurva (Table 3; Fig. 3). There was no significant difference in activity, exploration, activity‐exploration or social tendency between Bormio and Valfurva (Table 3; Fig. 3). Male squirrels have a lower active‐explorative score than female squirrels (β = −0.29, 95% CI = −0.56 to −0.01, pMCMC = 0.041; Table S3, Supporting Information), and heavier animals were more explorative than individuals with lower body mass (β = 0.12, 95% CI = 0.01 to 0.23, pMCMC = 0.028). In spring–summer, red squirrels were significantly more active (β = 0.29, 95% CI = 0.12 to 0.47, pMCMC = 0.001), explorative (β = 0.58, 95% CI = 0.39 to 0.78, pMCMC < 0.0001), and also active‐explorative (β = 0.41, 95% CI = 0.21 to 0.60, pMCMC < 0.0001) than in autumn. Finally, activity and exploration were higher during the first arena test than during subsequent ones (activity: β = 0.75, 95% CI = 0.56 to 0.94, pMCMC < 0.0001; exploration: β = 0.52, 95% CI = 0.32 to 0.72, pMCMC < 0.0001), contrary to social tendency, that was lower during the first experiment (β = −0.33, 95% CI = −0.54 to −0.11, pMCMC = 0.003).3TablePosterior means and 95% credibility intervals of the comparisons between the 3 study areas (BOR, Bormio; CAN, Cancano; VAL, Valfurva) for each behavioral trait, obtained from the posterior distribution of the full MCMCglmm modelDifferencesDependent variableCAN–BORVAL–BORCAN–VALOFTActivity0.62 (0.24 to 1.01)0.27 (−0.05 to 0.59)0.35 (0.03 to 0.66)Exploration0.19 (−0.16 to 0.55)−0.27 (−0.57 to 0.03)0.46 (0.19 to 0.74)MISActivity‐exploration0.55 (0.15 to 0.94)0.10 (−0.24 to 0.44)0.45 (0.12 to 0.77)Social tendency0.51 (0.11 to 0.90)0.17 (−0.15 to 0.51)0.34 (0.03 to 0.65)Significant results in bold.3FigureComparisons of the behavioral scores between the 3 study areas (Cancano, dark grey; Bormio, grey; Valfurva, light grey). (a) Activity, exploration, and activity‐exploration square root transformed. (b) Social tendency true values. Boxplots show median (solid horizontal line), mean (black diamond), and 1st (25%) and 3rd (75%) quartiles.Among‐ and within‐individual variances and behavioral syndromeMost of the posterior estimates of the correlations among the personality traits were significant (Fig. 4). There were positive correlations of activity with activity‐exploration, and with social tendency; of exploration with activity‐exploration, and with social tendency; and between activity‐exploration and social tendency (Fig. 4). This suggests a behavioral syndrome (a close association of different behavioral traits; Sih et al. 2004, 2012; Réale et al. 2007) whereby more active squirrels being also more explorative and more social.4FigureCorrelations (estimate ± 95% credibility intervals) between the dependent variables derived from the full MCMCglmm model. Significant results (0 not included in the 95% CIs).Comparing among‐ and within‐individual variances between the study areas (Table S5, Supporting Information), we found that the among‐individual variance of exploration was higher in Cancano than in Valfurva (Table 4), suggesting more phenotypic variation in exploration behavior in red squirrels living in the marginal study area. In contrast, the within‐individual variances of exploration were lower in Cancano than in Bormio and Valfurva (Table 4), suggesting a higher plasticity of this behavioral trait in Bormio and in Valfurva than in Cancano, the marginal habitat.4TableMean differences ± SD (95% credibility intervals) in between‐individual variances (a) and within‐individual variances (b) of the dependent variables in the study areas (BOR, Bormio; CAN, Cancano; VAL, Valfurva) based on the 3 MCMCglmm modelsDifferencesDependent variableBOR–CANVAL–CANBOR–VALa) Between‐individualActivity0.11 ± 0.22 (−0.28 to 0.59)0.16 ± 0.15 (−0.17 to 0.44)−0.05 ± 0.21 (−0.39 to 0.42)Exploration−0.19 ± 0.37 (−0.89 to 0.62)−0.49 ± 0.25 (−1.06 to −0.11)0.30 ± 0.29 (−0.04 to 1.03)Activity‐exploration0.15 ± 0.50 (−0.72 to 1.22)−0.34 ± 0.28 (−0.96 to 0.11)0.49 ± 0.43 (−0.16 to 1.47)Social tendency0.04 ± 0.28 (−0.49 to 0.66)−0.08 ± 0.19 (−0.52 to 0.22)0.12 ± 0.24 (−0.24 to 0.70)b) Within‐individualActivity−0.07 ± 0.23 (−0.51 to 0.40)−0.07 ± 0.16 (−0.42 to 0.21)0.003 ± 0.20 (−0.32 to 0.44)Exploration0.67 ± 0.31 (0.14 to 1.35)0.36 ± 0.14 (0.04 to 0.60)0.31 ± 0.29 (−0.17 to 0.97)Activity‐exploration0.46 ± 0.35 (−0.11 to 1.27)0.29 ± 0.16 (−0.08 to 0.57)0.17 ± 0.34 (−0.33 to 0.96)Social tendency−0.0001 ± 0.25 (−0.47 to 0.52)0.27 ± 0.18 (−0.12 to 0.60)−0.27 ± 0.23 (−0.66 to 0.23)Significant results in bold.For estimates of the between‐individual and the within‐individual variances by study area, see Table S5, Supporting Information.DISCUSSIONIn this study, we investigated whether there were morphological and/or behavioral differences between 3 populations of Eurasian red squirrels in alpine study areas. Our results demonstrate that squirrels in the edge habitat of their altitudinal distribution showed both morphological and behavioral adaptations to the extreme weather and food conditions. First, we found that males in the marginal habitat were smaller and weighed less than in the high‐quality Norway spruce dominated forest. In contrast, females living in the edge habitat did not differ in body size and body mass from females living in the other 2 study areas. Secondly, squirrels living in the edge habitat had higher scores of behavioral traits, confirming our hypothesis that more active‐explorative types are advantaged in low‐quality habitats. Finally, and contrary to our predictions, animals living in the marginal habitat did not have a higher behavioral plasticity, but their within‐individual variance of exploration was significantly lower than in the other 2 study areas.Morphological differencesBetween‐habitat variation in body mass and body size, measured by hind‐foot length, revealed in this study confirmed earlier findings based on 6 populations (Wauters et al. 2007), 2 of which were also included here (Valfurva and Bormio). As in the previous study, there was no size difference between the sexes in Valfurva and Bormio, but in the marginal habitat (Cancano) hind foot length of females was on average 1 mm longer than males. Moreover, among males, the body size – body mass relationship did not differ between the study areas and males at Cancano weighed on average 20 g less than in the high‐quality habitat Valfurva. Among females, patterns were more complicated: body mass increased more strongly with size (hind foot length) in the marginal than in the high‐quality habitat and, consequently, there was little difference in the body mass of adult females between the study areas; on average only 4 g between Cancano and Valfurva. These sex‐specific patterns differ from previous findings on this and other tree squirrel species over most of their range (Nash & Seaman 1977; Boutin & Larsen 1993; Koprowski 1994a,b; Wauters & Dhondt 1989a; Lurz & Lloyd 2000), and suggest differential selective pressures for male and female red squirrels in the lowest‐quality marginal habitat Cancano than in neighboring populations living in the same geographic area (see also Hoffman & Merilä 1999; McAdam et al. 2002; McAdam & Boutin 2003). Hence, our data suggest that in the marginal, high‐elevation mountain‐pine habitat, smaller and lighter males have an advantage over larger, heavier ones. In mammals, animals of smaller size/body mass can meet daily energy requirements with a lower food‐intake than bigger individuals. In Cancano, red squirrels feed mainly on the small seeds from cones of mountain pine from late summer to early spring and on its buds and male flowers in spring–early summer (Romeo et al. 2010). Species‐specific energy content of all seeds contained in a cone of mountain pine is only 6.0 kJ against the 51 and 121 kJ, respectively, of a Norway spruce cone and an Arolla pine cone, the main food resources in the other 2 study areas. Thus, smaller size among male red squirrels in the marginal habitat could be an adaptation to a reduced availability of high‐energy food resources characterized by relatively long handling times (unpublished data; see also data on Scots pine in Wauters et al. 1992).However, this trend was not observed among female red squirrels. We believe this is due to the differential energy investment in reproduction between the sexes. Reaching a high body mass is essential for female red squirrels to enhance their reproductive success. Indeed, the probability to enter estrus (Wauters & Dhondt 1989b), lactation success (Wauters & Dhondt 1995; Wauters & Lens 1995; Wauters et al. 2007), offspring weaning mass and early survival (Wauters et al. 1993), and individual variation in lifetime reproductive success (Wauters & Dhondt 1995) are all strongly and positively related to body mass. Also in this study, lactating females weighed more than non‐breeding ones. Moreover, in the marginal habitat, persistency of adults is low and most females will have only a single reproductive event in their lifetime (Rodrigues et al. 2010); hence having a high body mass should aid females to maximize reproductive output in terms of litter size and early offspring survival (Mari et al. 2008).Finally, heavier males tend to be more successful in mating than males of lower body mass, which should favor larger and heavier animals (Wauters et al. 1990). However, these observations were carried out in high‐density populations where up to 4 males competed for access to the estrus female. In Cancano, densities were so low that, probably only 1 or 2 males took part at a mating chase, which could reduce sexual selection for larger males. Finally, these sex‐specific differences between habitats argue against the hypothesis that squirrels are smaller and weigh less in the poorer‐quality habitat because of having an overall lower food intake.Personality traits and habitat selectionRed squirrels in the marginal edge habitat were more active, active‐explorative and social than in the other 2 higher‐quality areas. Moreover, animals in Cancano were also more explorative than those in Valfurva but not than those living in the Arolla pine forest of Bormio, which is characterized by a higher overall tree‐seed abundance than the mountain pine forest of Cancano, but also by a higher patchiness (forest heterogeneity) (Table 1; see also Salmaso et al. 2009). Hence, individuals that behave more explorative might have an advantage in finding food‐rich patches and optimize home range use (Wauters et al. 2021), explaining why the expression of exploration was similar in Bormio as in the marginal habitat of Cancano. Several studies on wild animals found a non‐random distribution of individuals’ personality traits between heterogeneous habitats (Cerqueira et al. 2016; Holtmann et al. 2017; Brehm & Mortelliti 2021; Rabdeau et al. 2021). For example, in dunnocks (Prunella modularis), personality has been observed having an impact on habitat choice, bolder individuals settled in areas where human disturbance is high, contrary to shier ones that prefer sites with less human disturbance (Holtmann et al. 2017). As suggested in these studies, and in agreement with the “personality‐matching habitat choice” hypothesis (Edelaar et al. 2008), an individual settles in the habitat that is the most suitable for its personality and that best matches the animal's capacity to use this habitat (Edelaar et al. 2008; Holtmann et al. 2017). These authors also suggest that behavioral adjustments over time contribute little to the personality‐habitat relationship (Edelaar et al. 2008; Holtmann et al. 2017). This could explain why, in contrast with our prediction, we did not find a higher plasticity (within‐individual variance) of the personality traits in the marginal habitat compared to the other ones. In our study system, we believe that increased activity and activity‐exploration, but also high exploration scores, result in higher success in acquiring resources, and thus enhance the probability to settle in marginal areas of the distribution range, characterized by poorer habitat quality (Biro & Stamps 2008; Wauters et al. 2021). Similarly, a study by Boone and colleagues (2022) investigating the interaction between the personality of another scatter‐hoarding mammal (deer mice, Peromyscus maniculatus), and seed (food) availability, found that more active/explorative mice choose lower‐quality seeds when food availability was reduced by 50%, in contrast with less active/explorative individuals that continued to select high‐quality seeds both at high and low food availability (Boone et al. 2022). Hence, this study suggests that more active and explorative individuals usually tend to use a wider variety of food resources than less active/explorative ones (Toscano et al. 2016; Boone et al. 2022). Therefore, the presence of more active, explorative and active‐explorative individuals in the edge habitat may also be due to the fact that these personality types can more easily switch and adapt to other, lower‐quality food resources (in our case the seeds of mountain pine cones but also male flowers and buds of mountain pine, rather than the more energetic Norway spruce and Arolla pine seeds) than less active/explorative ones.Together with the personality traits related to habitat‐ and space use, we also found a higher social tendency in the marginal habitat than in other study areas. This could be due to the behavioral syndrome we described (see also Santicchia et al. 2022). In other words, if higher activity and exploration may benefit animals in the marginal habitat and both are associated with higher social‐tendency in a behavioral syndrome, then the squirrels in the marginal habitat will express more social behavior during the MIS test, without there being a selective advantage of behaving more social.Alternative hypotheses: early experience or behavioral plasticityAn alternative explanation of the differences in personality between the study areas may be due to the shaping of squirrels’ behavior by the different pressures that characterized each habitat. In fact, changes in personality may occur throughout the individual's life (Stamps & Groothuis 2010), leading to a weakening of the personality‐matching habitat choice hypothesis. In our study, we did not specifically examine this behavioral plasticity throughout an individual's life, but in general, considerable modifications in personality seem to happen after drastic developmental changes (e.g. sexual maturation) (Cabrera et al. 2021). To account for this, we removed juveniles from our analyses and we performed multiple arena tests in different sessions (over several months to several years) of the same individual to test the repeatability of the trait over time. Moreover, it is reported that an animal's experience during early development could have an effect on its personality as adult (Stamps & Groothuis 2010). However, the majority of red squirrels disperse when between 4 and 10 months old, while some disperse as adults (>1‐year old), over 1 up to 5–10 km, and philopatry is rare (Wauters & Dhondt 1993; Wauters et al. 1994, 2011; Di Pierro et al. 2010). There is genetic evidence that dispersal occurs between 2 of our study sites (Di Pierro et al. 2010) and that also in the marginal site most of the newly caught animals are immigrants (our unpublished data). Hence, the animals monitored here arrived in a given study site coming from several directions and distances, guaranteeing differences in the early experiences of animals settling in that area. Therefore, we think that behavioral plasticity or early life experience did not have a great effect on the differences in personality that we found between the different habitats.A further point to consider is the possibility that squirrels rapidly change their behavior (rapid behavioral change; e.g. within a few days from the settlement in the new area) in response to the environmental pressures (Holtmann et al. 2017). For this study, we did not examine this rapid behavioral plasticity, due to the difficulties to observe personality in the first days of establishment in the new area. However, a deep investigation of behavior before and after the settlement is necessary to examine this alternative.Other factors affecting the expression of personality traitsFemale red squirrels tended to have a higher score for the activity‐exploration trait than males. This could be related to the higher energy investment by females in reproduction (see Morphological differences), producing an advantage in resource acquisition for more active‐explorative individuals. In line with this and with previous studies on red squirrels, heavier squirrels were more explorative than smaller ones (Santicchia et al. 2021, 2022). This association is found also in other tree squirrel species (Tamias striatus: Martin & Réale 2008; Sciurus carolinensis: Santicchia et al. 2019, 2020b).We also observed a seasonal effect on behavioral traits: squirrels in spring–summer were more active, explorative and active‐explorative than in autumn. Animals may show some degree of behavioral plasticity among seasons to cope with predictable abiotic and biotic environmental conditions (e.g. changes in photoperiod, rainfall, temperature, tree‐seed availability), especially at high latitudes and altitudes (Zhang & Buck 2022; see also Beaman et al. 2016). However, only few studies investigated seasonal fluctuations in behavior in free‐living Sciurids (Arctic ground squirrels, Urocitellus parryii: Chmura et al. 2020; Abert's squirrels, Sciurus aberti: Zhang & Buck 2022). In our studied species, the seasonal variation in trait expression could be linked to tree phenology; new fresh cones with mature seeds are available from the end of summer and during autumn and winter (Romeo et al. 2010; Zong et al. 2010). Therefore, augmented activity and exploration in spring–early summer might be linked to higher foraging needs when feeding on low‐energy food sources such as male flowers, buds, or cached seeds that need long searching times (Zong et al. 2010).In conclusion, the evidence from this study suggests that personality has an important role in habitat selection with consequences on animals’ morphology. This study was carried out in red squirrels populations in the altitudinal limit of their distribution range, but the same study system could be replicated in different contexts with different environmental stressors (Wauters et al. 2019; Santicchia et al. 2022) to better understand an individual's choice of habitat (e.g. differences among urban‐rural gradient). Such future studies should also measure fitness components of individuals (survival and/or reproductive success) which would allow to investigate whether natural selection favors different personalities in relation to habitat type (e.g. Santicchia et al. 2018). Also, the habitat‐related patterns in animal personality and the comparisons of among‐ and within‐individual variance in personality traits documented here, contributes to our still limited knowledge about behavioral plasticity differences between populations occurring in different habitats in vertebrates (Dingemanse & Wolf 2013; Hewes et al. 2017). Particularly, our findings that only for exploration, and not for the other behavioral traits, the within‐individual variance differed among populations strengthens the hypothesis that it is primarily the animal's personality that determined the individual's choice of habitat, and that behavioral plasticity was not strongly associated with habitat selection.ACKNOWLEDGMENTSThe authors thank the Stelvio National Park for allowing fieldwork in the 3 study areas. The authors are also grateful to the field technicians and thesis students for helping with data collection. This study is part of the ASPER (Alpine Squirrel Population Ecology Research) project (paper #36).Open Access Funding provided by Universita degli Studi dell'Insubria within the CRUI‐CARE Agreement.REFERENCESBeaman JE, White CR, Seebacher F (2016). Evolution of plasticity: Mechanistic link between development and reversible acclimation. Trends in Ecology & Evolution 31, 237–49.Bell AM, Hankison SJ, Laskowski KL (2009). The repeatability of behaviour: A meta‐analysis. Animal Behaviour 77, 771–83.Biro PA, Stamps JA (2008). Are animal personality traits linked to life‐history productivity? Trends in Ecology & Evolution 23, 361–8.Bolnick DI, Svanbäck R, Fordyce JA et al. (2003). The ecology of individuals: Incidence and implications of individual specialization. The American Naturalist 161, 1–28.Boon AK, Réale D, Boutin S (2008). Personality, habitat use, and their consequences for survival in North American red squirrels Tamiasciurus hudsonicus. Oikos 117, 1321–8.Boone SR, Brehm AM, Mortelliti A (2022). Seed predation and dispersal by small mammals in a landscape of fear: effects of personality, predation risk and land‐use change. Oikos 2022, e08232.Boutin S, Larsen K (1993). Does food availability affect growth and survival of males and females differently in a promiscuous small mammal, Tamiasciurus hudsonicus. Journal of Animal Ecology 62, 364–70.Boutin S, Wauters LA, McAdam AG, Humphries MM, Tosi G, Dhondt AA (2006). Anticipatory reproduction and population growth in seed predators. Science 314, 1928–30.Brehm AM, Mortelliti A (2021). Land‐use change alters associations between personality and microhabitat selection. Ecological Applications 31, e02443.Cabrera D, Nilsson JR, Griffen BD (2021). The development of animal personality across ontogeny: A cross‐species review. Animal Behaviour 173, 137–44.Carter AJ, Feeney WE, Marshall HH, Cowlishaw G, Heinsohn R (2013). Animal personality: What are behavioural ecologists measuring? Biological Reviews 88, 465–75.Cerqueira M, Rey S, Silva T, Featherstone Z, Crumlish M, MacKenzie S (2016). Thermal preference predicts animal personality in Nile tilapia Oreochromis niloticus. Journal of animal ecology 85, 1389–400.Chmura HE, Zhang VY, Wilbur SM, Barnes BM, Buck CL, Williams CT (2020). Plasticity and repeatability of activity patterns in free‐living Arctic ground squirrels. Animal Behaviour 169, 81–91.Cote J, Clobert J, Brodin T, Fogarty S, Sih A (2010). Personality‐dependent dispersal: Characterization, ontogeny and consequences for spatially structured populations. Philosophical Transactions of the Royal Society B 365, 4065–76.Dall SRX, Houston AI, McNamara JM (2004). The behavioural ecology of personality: Consistent individual differences from an adaptive perspective. Ecology Letters 7, 734–9.Di Pierro E, Bertolino S, Martinoli A, Preatoni D, Tosi G, Wauters LA (2010). Estimating offspring production using capture‐mark‐recapture and genetic methods in red squirrels. Ecological Research 25, 395–402.Di Pierro E, Ghisla A, Wauters LA et al. (2011). The effects of seed availability on habitat use by a specialist seed predator. European Journal of Wildlife Research 57, 585–95.Dingemanse NJ, Kazem AJN, Réale D, Wright J (2010). Behavioural reaction norms: Animal personality meets individual plasticity. Trends in Ecology & Evolution 25, 81–9.Dingemanse NJ, Bouwman KM, van de Pol M et al. (2012). Variation in personality and behavioural plasticity across four populations of the great tit Parus major. Journal of Animal Ecology 81, 116–26.Dingemanse NJ, Wolf M (2013). Between‐individual differences in behavioural plasticity within populations: causes and consequences. Animal Behaviour 85, 1031–9.Edelaar P, Siepielski AM, Clobert J (2008). Matching habitat choice causes directed gene flow: A neglected dimension in evolution and ecology. Evolution 62, 2462–72.Endler JA (1977). Geographical Variation, Speciation and Clines. Princeton University Press, Princeton, NJ.Gelman A, Rubin DB (1992). Inference from iterative simulation using multiple sequences. Statistical Science 7, 457–72.Geweke J (1992). Evaluating the accuracy of sampling‐based approaches to the calculation of posterior moments. In: Bernardo JM, Berger AP, Dawid AP, Smith AFM, eds. Bayesian Statistics. Oxford University Press, Oxford, pp. 169–93.Hadfield JD (2010). MCMC methods for multi‐response generalized linear mixed models: The MCMCglmm R package. Journal of Statistical Software 33, 1–22.Hänninen L, Pastell M (2009). CowLog: Open‐source software for coding behaviors from digital video. Behavior Research Methods 41, 472–6.Hewes ME, Delventhal K, Chaves‐Campos J (2017). Behavioral plasticity and consistency in the naked‐footed mouse (Peromyscus nudipes) with habitat disturbance. Journal of Ethology 35, 279–92.Hoffman AA, Merilä J (1999). Heritable variation and evolution under favourable and unfavourable conditions. Trends in Ecology and Evolution 12, 460–1.Holtmann B, Santos ESA, Lara CE, Nakagawa S (2017). Personality‐matching habitat choice, rather than behavioural plasticity, is a likely driver of a phenotype–environment covariance. Proceedings of the Royal Society B 284, 20170943.Kawecki TJ, Ebert D (2004). Conceptual issues in local adaptation. Ecology Letters 7, 1225–41.Koprowski J (1994a). Sciurus niger. Mammalian Species 479, 1–9.Koprowski J (1994b). Sciurus carolinensis. Mammalian Species 480, 1–9.Lurz PWW, Garson PJ, Wauters LA (2000). Effects of temporal and spatial variations in food supply on the space and habitat use of red squirrels, Sciurus vulgaris L. Journal of Zoology 251, 167–78.Lurz PWW, Lloyd AJ (2000). Body weights in grey and red squirrels: Do seasonal weight increases occur in conifer woodland? Journal of Zoology 252, 539–43.Mari V, Martini S, Romeo C et al. (2008). Record litter size in the Eurasian red squirrel (Sciurus vulgaris). Hystrix, the Italian Journal of Mammalogy 19, 61–5.Martin JGA, Réale D (2008). Animal temperament and human disturbance: Implications for the response of wildlife to tourism. Behavioural Processes 77, 66–72.Martin JGA, Nussey DH, Wilson AJ, Réale D (2011). Measuring individual differences in reaction norms in field and experimental studies: A power analysis of random regression models: Power analysis of random regression models. Methods in Ecology and Evolution 2, 362–74.Mazzamuto MV, Cremonesi G, Santicchia F, Preatoni D, Martinoli A, Wauters LA (2019). Rodents in the arena: A critical evaluation of methods measuring personality traits. Ethology Ecology & Evolution 31, 38–58.McAdam AG, Boutin S, Réale D, Berteaux D (2002). Maternal effects and the potential for evolution in a natural population of animals. Evolution 56, 846–51.McAdam AG, Boutin S (2003). Effects of food abundance on genetic and maternal variation in the growth rate of juvenile red squirrels. Journal of Evolutionary Biology 16, 1249–56.Nash DJ, Seaman RN (1977). Sciurus aberti. Mammalian Species 80, 1–5.R Core Team (2020). R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. Available from URL: https://www.R‐project.org/Rabdeau J, Arroyo B, Mougeot F, Badenhausser I, Bretagnolle V, Monceau K (2021). Do human infrastructures shape nest distribution in the landscape depending on individual personality in a farmland bird of prey? Journal of Animal Ecology 90, 2848–58.Réale D, Reader SM, Sol D, McDougall PT, Dingemanse NJ (2007). Integrating animal temperament within ecology and evolution. Biological Reviews 82, 291–318.Rodrigues D, Wauters LA, Romeo C et al. (2010). Living on the edge: Can Eurasian red squirrels (Sciurus vulgaris) persist in extreme high‐elevation habitats? Arctic, Antarctic, and Alpine Research 42, 106–12.Romeo C, Wauters LA, Preatoni D, Tosi G, Martinoli A (2010). Living on the edge: Space use of Eurasian red squirrels in marginal high‐elevation habitat. Acta Oecologica 36, 604–10.Royauté R, Dochtermann NA (2021). Comparing ecological and evolutionary variability within datasets. Behavioral Ecology and Sociobiology 75, 127.Salmaso F, Molinari A, Di Pierro E et al. (2009). Estimating and comparing food availability for tree‐seed predators in typical pulsed‐resource systems: alpine conifer forests. Plant Biosystems 143, 258–67.Santicchia F, Gagnaison C, Bisi F et al. (2018). Habitat‐dependent effects of personality on survival and reproduction in red squirrels. Behavioral Ecology and Sociobiology 72, 134.Santicchia F, Romeo C, Ferrari N et al. (2019). The price of being bold? Relationship between personality and endoparasitic infection in a tree squirrel. Mammalian Biology 97, 1–8.Santicchia F, Wauters LA, Piscitelli AP et al. (2020a). Spillover of an alien parasite reduces expression of costly behaviour in native host species. Journal of Animal Ecology 89, 1559–69.Santicchia F, Wauters LA, Dantzer B et al. (2020b). Relationships between personality traits and the physiological stress response in a wild mammal. Current Zoology 66, 197–204.Santicchia F, Van Dongen S, Martinoli A, Preatoni D, Wauters LA (2021). Measuring personality traits in Eurasian red squirrels: A critical comparison of different methods. Ethology 127, 187–201.Santicchia F, Wauters LA, Tranquillo C et al. (2022). Invasive alien species as an environmental stressor and its effects on coping style in a native competitor, the Eurasian red squirrel. Hormones and Behavior 140, 105127.Sih A, Bell A, Johnson JC (2004). Behavioral syndromes: An ecological and evolutionary overview. Trends in Ecology & Evolution 19, 372–8.Sih A, Cote J, Evans M, Fogarty S, Pruitt J (2012). Ecological implications of behavioural syndromes. Ecology Letters 15, 278–89.Spiegel O, Leu ST, Bull CM, Sih A (2017). What's your move? Movement as a link between personality and spatial dynamics in animal populations. Ecology Letters 20, 3–18.Stamps JA, Groothuis TGG (2010). Developmental perspectives on personality: Implications for ecological and evolutionary studies of individual differences. Philosophical Transactions of the Royal Society B: Biological Sciences 365, 4029–41.Stoffel MA, Nakagawa S, Schielzeth H (2017). rptR: Repeatability estimation and variance decomposition by generalized linear mixed‐effects models. Methods in Ecology and Evolution 8, 1639–44.Toscano BJ, Gownaris NJ, Heerhartz SM, Monaco CJ (2016). Personality, foraging behavior and specialization: integrating behavioral and food web ecology at the individual level. Oecologia 182, 55–69.Tranquillo C, Villa F, Wauters LA et al. (2022). Physiological stress and spatio‐temporal fluctuations of food abundance and population density in Eurasian red squirrels. Hystrix, the Italian Journal of Mammalogy 33, 26–32.Verbeke G, Molenberghs G (2000). Linear Mixed Models for Longitudinal Data. Springer, New York.Wauters LA, Dhondt AA (1989a). Variation in length and body weight of the red squirrel in two different habitats. Journal of Zoology 217, 93–106.Wauters LA, Dhondt AA (1989b). Body weight, longevity and reproductive success in red squirrels (Sciurus vulgaris). Journal of Animal Ecology 58, 637–51.Wauters LA, Dhondt AA, De Vos R (1990). Factors affecting male mating success in red squirrels (Sciurus vulgaris). Ethology Ecology & Evolution 2, 195–204.Wauters LA, Dhondt AA (1992). Spacing behaviour of red squirrels, Sciurus vulgaris: Variation between habitats and the sexes. Animal Behaviour 43, 297–311.Wauters LA, Swinnen C, Dhondt AA (1992). Activity budget and foraging behaviour of red squirrels (Sciurus vulgaris) in coniferous and deciduous habitats. Journal of Zoology 227, 71–86.Wauters LA, Bijnens L, Dhondt AA (1993). Body mass at weaning and juvenile recruitment in the red squirrel. Journal of Animal Ecology 62, 280–6.Wauters LA, Dhondt AA (1993). Immigration patterns and success in red squirrels. Behavioral Ecology and Sociobiology 33, 159–67.Wauters LA, Matthysen E, Dhondt AA (1994). Survival and lifetime reproductive success in dispersing and resident red squirrels. Behavioral Ecology and Sociobiology 34, 197–201.Wauters LA, Dhondt AA (1995). Components of lifetime reproductive success of female Eurasian red squirrels. Oikos 72, 402–10.Wauters LA, Lens L (1995). Effects of food abundance and density on red squirrel (Sciurus vulgaris) reproduction. Ecology 76, 2460–9.Wauters LA, Gurnell J, Preatoni D, Tosi G (2001). Effects of spatial variation in food availability on spacing behaviour and demography of Eurasian red squirrels. Ecography 24, 525–38.Wauters LA, Vermeulen M, Van Dongen S et al. (2007) Effects of spatio‐temporal variation in food supply on red squirrel Sciurus vulgaris body size and body mass and its consequences for some fitness components. Ecography 30, 51–65.Wauters LA, Githiru M, Bertolino S, Molinari A, Tosi G, Lens L (2008). Demography of alpine red squirrel populations in relation to fluctuations in seed crop size. Ecography 31, 104–14.Wauters LA, Preatoni D, Martinoli A, Verbeylen G, Matthysen E (2011). No sex bias in natal dispersal of Eurasian red squirrels. Mammalian Biology 76, 369–72.Wauters LA, Mazzamuto MV, Santicchia F, Van Dongen S, Preatoni DG, Martinoli A (2019). Interspecific competition affects the expression of personality‐traits in natural populations. Scientific Reports 9, 11189.Wauters LA, Mazzamuto MV, Santicchia F et al. (2021). Personality traits, sex and food abundance shape space use in an arboreal mammal. Oecologia 196, 65–76.Wilson AJ, Réale D, Clements MN et al. (2010). An ecologist's guide to the animal model. Journal of Animal Ecology 79, 13–26.Zhang VY, Buck CL (2022). Seasonal patterns in behavior and glucocorticoid secretion of a specialist Holarctic tree squirrel (Sciurus aberti). Journal of Comparative Physiology B 192, 541–59.Zong C, Wauters LA, Van Dongen S et al. (2010). Annual variation in predation and dispersal of Arolla pine (Pinus cembra L.) seeds by Eurasian red squirrels and other seed‐eaters. Forest Ecology and Management 260, 587–94.

Journal

Integrative ZoologyWiley

Published: Jul 1, 2023

Keywords: behavioral plasticity; marginal edge‐habitat; morphology; personality; Sciurus vulgaris

There are no references for this article.