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

Learn More →

Exploring links between personality traits and their social and non-social environments in wild poison frogs

Exploring links between personality traits and their social and non-social environments in wild... An animal’s behavioral phenotype comprises several traits, which are hierarchically structured in functional units. This is manifested in measured behaviors often being correlated, partly reflecting the need of a coordinated functional response. Unfortunately, we still have limited understanding whether consistent differences in animal behaviors are due to underlying physiological constraints or a result of plastic adaptation to their current environment. Therefore, characterizing the spatial distribution of behaviors can provide important insights into causes and consequences of behavioral variation. In the present study, we quantified behaviors in a wild, free-ranging population of the Neotropical frog Allobates femoralis. We investigated how these behaviors were linked to the frogs’ natural and social environment and quantified the extent to which these behav - iors consistently differed among individuals (i.e., animal personality). We assessed levels of aggressiveness, exploration, and boldness by measuring several underlying behaviors expressed in a set of experimental assays, and found evidence for consistent among-individual differences along these axes. Contrary to our expectation, there was no relationship between individual behaviors and their natural environment, but we found a plastic response of males to changes in female density, which might reflect how individuals cope with their socio-ecological environment. Significance statement How are behavioral phenotypes distributed across space? Here, we studied an entire free-ranging population of poison frogs, and investigated if the personality traits aggressiveness, exploration, and boldness are linked to the frogs’ natural or social environment. We found that behavioral traits were non-randomly distributed across the population, suggesting that the spatial arrangement of behavioral traits reflects how individuals cope with their complex natural and social environment. Keywords Behavioral variation · Animal personality · Poison frogs · Non-random distribution · Environment Communicated by A. Taylor Baugh. * Mélissa Peignier Department of Biology and Ecology, University melissa.peignier@unibe.ch of Montpellier, Montpellier, France 1 6 Department of Evolutionary Biology, University of Vienna, Division of Behavioural Ecology, Institute of Ecology Vienna, Austria and Evolution, University of Bern, Wohlenstrasse 50a, CH-3032 Hinterkappelen, Bern, Switzerland iES Landau, Institute for Environmental Sciences, University of Koblenz-Landau, Mainz, Germany Messerli Research Institute, University of Veterinary Medicine Vienna, Vienna, Austria Central Research Laboratories, Natural History Museum Vienna, Vienna, Austria Department of Behavioral and Cognitive Biology, University of Vienna, Vienna, Austria Institute of Ecology and Evolution, School of Biological Sciences, University of Edinburgh, Edinburgh, UK Department of Biology, Centre for Biodiversity Dynamics, Norwegian University of Science and Technology, University of Music and Performing Arts Graz, Graz, Austria Trondheim, Norway Vol.:(0123456789) 1 3 Behavioral Ecology and Sociobiology ( 2 0 2 2) 76:93 93 Page2of14 succeed in establishing a territory in high-density patches, Introduction where they have access to more mates but face elevated intra- sexual competition. In turn, the link between aggressiveness Behavioral variation is ubiquitous in nature. Behaviors may and population density could also be caused by individual vary considerably among species; and within species, we find plasticity allowing individuals to match their aggressiveness variation both among and within individuals. Behavioral vari- to competition levels. Regardless of the mechanisms linking ation at all of these levels plays a major role in reproduction behavior and environment, identifying non-random spatial dis- and survival, for example, affecting the risk of being detected tribution of behaviors provides important insights into their and caught by predators, likelihood of dispersal, foraging effi - function and their role in allowing organisms to cope with ciency, and/or the attractiveness to mating partners (Pigliucci environmental variation. Long-term studies of the distribution 2005; Réale and Dingemanse 2010; Sih et al. 2012). This of behaviors in the wild are thus a necessary first step towards variation can occur in the form of consistent among-individ- understanding the mechanisms underlying the non-random dis- ual differences in behavior, referred to as animal personality, tribution of behavioral traits (Archard and Braithwaite 2010). which has been documented in many animal taxa (Bell 2005; Amphibians, and in particular Neotropical poison frogs Dochtermann and Jenkins 2007; Tremmel and Müller 2013; (Dendrobatidae, sensu AmphibiaWeb 2022), are great models Zidar et al. 2017; Goursot et al. 2019). Animal personality, to study behavioral variation across their environment. Many however, does not preclude the existence of individual plas- species show territoriality or site fidelity that facilitates repeated ticity (Dingemanse et al. 2010), and we need to take into measurements in wild individuals (Wells 2007; Kelleher et al. consideration both sources of variation when thinking about 2018), which can then be linked to local environmental parame- adaptive significance of behavioral variation in animals. ters. Many species also exhibit elaborate courtship behavior, ter- Five main personality traits are generally characterized in restrial oviposition, and obligatory tadpole transport of hatched animals along the following axes: active/passive, aggressive/ larvae to aquatic sites (e.g., Crump 1972; Roithmair 1994; Pröhl docile, bold/shy, exploratory/stationary, and sociable/non- 2005; Pašukonis et al. 2013; Rojas and Pašukonis 2019; Yang sociable (Réale et al. 2007). Personality traits can be seen as et al. 2019; Souza et al. 2021), offering ideal prerequisites for latent variables that affect multiple quantifiable behaviors of an within- and between-individual variation in behavior to arise. In organism in certain contexts (Pigliucci 2003; Araya-Ajoy and the present study, we aim to quantify how male-male aggression, Dingemanse 2014). For instance, aggressiveness can be seen exploratory, and anti-predator behaviors are expressed as func- as an unobservable (i.e., latent) variable that affects several tional units that can be described as latent variables reflecting observable behaviors during agonistic encounters, which can three personality axes (referred to by one dimension of the axis be assessed and quantified in an experimental context (e.g., as “aggression,” “exploration,” and “boldness”). We also aim speed of territorial reaction, number of attacks). Situations at identifying how behaviors relate to the individual’s natural where personality traits are non-randomly distributed across and social environment using a free-ranging population of the space are typically referred to as “phenotype by environment Neotropical poison frog Allobates femoralis (Dendrobatidae). correlation,” where environment refers to both the natural and Based on the natural history of A. femoralis, we make mul- social surroundings of the individual (Conover and Schultz tiple predictions. In males, we expect to find consistent among- 1995; Dingemanse and Araya-Ajoy 2015). For example, individual differences in boldness and aggression. During the anemones (Condylactis gigantea) living in areas with higher reproductive season, males produce loud advertisement calls seagrass were found to be shyer than those living in more open to warn intruders and to attract females (Ringler et al. 2011; areas (Hensley et al. 2012) and western bluebirds (Sialia mexi- Chaloupka et al. 2022). Differences in personality traits across cana) modify their aggression to match the level of aggressive males might be related to differences in the trade-off they face behavior of their mate (Duckworth and Kruuk 2009). between calling to secure mating and exposure to predators. Since personality traits can vary simultaneously within and As females do not display territoriality but merely perch fidel - among individuals, associations between behavior and envi- ity (Fischer et al. 2020) and are never observed in any other ronment can originate from multiple processes, such as a non- aggressive interaction, they are generally considered non- random distribution of behaviors and/or phenotypic plasticity aggressive; we have no clear expectation if and how varying (Sprau and Dingemanse 2017). A non-random distribution degrees of boldness could be maintained in females. However, of behaviors might be a direct effect of certain phenotypes we expect to find consistent individual differences in explora - (caused by genes and permanent environmental effects) show - tion in both sexes, because males transport tadpoles to natural ing a preference for certain environments. Alternatively, selec- pools located up to 200 m away from their territory and females tive pressures induced by the heterogeneity of the environment commute to male territories within 20 m distance of their perch can also maintain or generate individual differences in behavior for mating (Ringler et al. 2012, 2013, 2018; Fischer et al. 2020). within a population (Dingemanse et al. 2004). For instance, Finally, we expect males’ personality traits to be non-randomly in territorial species, only highly aggressive individuals may 1 3 Behavioral Ecology and Sociobiology ( 2 0 2 2) 76:93 Page3of14 93 distributed across space because individual males face distinct to evoke territorial defense behavior in focal males (Fig. 1; challenges based on their social and/or natural environment. On Ursprung et al. 2009). During agonistic encounters, male A. the one hand, we expect more aggressive and bolder individuals femoralis display typical responses consisting of an orien- to occupy areas of low complexity (e.g., with sparse vegetation tation of their head/body, jumps towards the intruder, and and few ground structures) where they are easier to spot for sometimes direct attack (i.e., wrestling) (Hödl 1987; Nar- females, while more passive or shyer individuals should occupy ins et al. 2003). We expect that in the context of territorial areas of higher complexity with more places to hide. On the defense, the personality trait “aggression” affects the latency other hand, we expect bolder and more aggressive individuals until the first head-body orientation, the latency to the first to occupy territories in areas with a higher population density jump, the probability to jump during moments when the than those occupied by more passive or shyer individuals. intruder does not call, and the speed to approach the intruder (Chaloupka et al. 2022). Aggressive territorial behavior was only assessed in males, as females do not defend territories. Methods We used a music player with integrated loudspeaker (MUVO 2c, Creative, Singapore), centered on top of a black Study population and area PVC disc (radius = 15 cm), 2 m from and facing the calling focal male (Fig. 1). We established the 2 m distance between We conducted our study in a population of A. femoralis on a the focal frog and the loudspeaker with a laser rangefinder lowland rainforest river island of approximately 5 ha. The island (DLE 50, Bosch, Stuttgart, Germany). The experimenter is situated in the “Les Nouragues” nature reserve in French remained a further 1 m behind the loudspeaker. We gave a Guiana (4°02′ N, 52°41′W; Bongers et  al.  2001), near the 30-s acclimation period to the frog, which was enough for “Saut Pararé” field camp of the CNRS Nouragues Ecological the frog to return to normal behavior and calling, before Research Station. The population was introduced in 2012 and randomly presenting one of seven synthetic calls. The calls has reached a stable size of approximately 150 adult individu- varied in their inter-note and inter-call intervals to avoid als (see Ringler et al. 2014). The experiments took place from habituation over the course of the experiment. Each syn- February to April of 2019, coinciding with the breeding season. thetic call featured the spectral and temporal parameters of a We surveyed the population every day during its most active nearby free-ranging population of A. femoralis (Narins et al. hours from 0900 to 1800 h, aiming to sample all adult males 2003; Gasser et al. 2009). Each of the playbacks lasted for and females on the island in the course of the study period. 5 min and was presented from the original WAV-files (16-bit, We caught frogs using a transparent plastic bag to mini- 44.1 kHz) using the same volume settings across all trials. mize stress and direct contact, thereby limiting the potential Using a digital voice recorder (ICD-PX333, Sony, Tokyo, influence of handling on behaviors. We identified all frogs via Japan), we commented on the behavior of the focal male, not- digital pictures of their distinct ventral patterns and with the ing its first head-body orientation, its jumps, and its arrival help of the pattern matching software Wild-ID (Bolger et al. at the speaker (i.e., touching the disc with at least one part of 2012), and sexed them by the presence (males) or absence his body). The trial ended when the male entered the 15-cm (females) of vocal sacs. Ventral patterns clearly differ among perimeter or when the playback stopped after 5 min. Trials individuals and are consistent across their adult lifespan, pro- where a focal male began to call were stopped and excluded viding a reliable method of identification. Information on the age of individuals was available from a concurrent long- term monitoring project of the island’s population since its origins in 2012. We recorded the locations of all frogs on a detailed digital map (Ringler et al. 2016) using the mobile GIS software ArcPad 10 (ESRI, Redlands, CA, USA) on rugged Win10 tablets (CAT T20, Bullitt Group, Reading, UK), and further handled the data in ArcGIS 10.6 (ESRI). We determined body size of all adults (snout urostyle length; SUL) from dorsal photographs in front of a reference scale using the software Image J 1.52a (Rasband 1997–2021). Quantification of aggressive behaviors Aggression is often measured via “an individual’s agonistic reaction towards a conspecific” (Réale et al. 2007). To assess Fig. 1 Picture of the territorial defense trial, presenting the speaker individual levels of aggression, we used acoustic playbacks and a male on top of the perimeter 1 3 Behavioral Ecology and Sociobiology ( 2 0 2 2) 76:93 93 Page4of14 from further analysis, as this was most likely the result of previous amphibian personality studies (Kelleher et  al. the speaker being positioned outside the defended area of 2018). Although counterintuitive at first, the number of the focal male’s territory (cf. Ringler et al. 2011). Vegetation jumps and the area covered are both important aspects of density may cause the sound to be reflected and/or distorted, a frog’s exploratory tendency. Indeed, two individuals per- so we measured the received sound pressure level (“SPL” forming an equal number of jumps can visit more or fewer in dB) of the signal after each trial using a sound pressure distinct areas of the box depending on their respective aver- meter (SL-100, Voltcraft, Hirschau, Germany). In all cases, age jump lengths. Individuals were tested at the same loca- the received SPL was above the threshold of 56 dB to elicit tion they were found. Males were tested immediately after a behavioral response (Hödl 1983). The minimum duration the territorial defense test, and females were caught at their between two consecutive tests with the same individual was encounter location and put straight inside the NET (Fig. 2). 24 h (on average tests were 12.11 ± 7.92 SD days apart). Although this method prevents us from comparing levels of Allobates femoralis typically lives for 1.5  years (Ringler exploratory tendency and boldness between sexes, it was et al. 2009; Ursprung et al. 2011a, b) and males only display chosen to keep handling as minimal as possible on males. aggressive reactions to territory intruders during the repro- The NET setup consisted of a cooler box (50 × 25 × 29 cm, ductive season. In our study, we aimed to test individuals hereafter “Novel Environment”), with a 10-cm PVC tube with similar among- and within-individual inter-test intervals attached on one side of the box (hereafter “shelter”). An to cover several weeks of the reproductive season. Therefore, opaque sliding door separated the Novel Environment from an average interval of 12 days between tests represents an the shelter. In the lid of the box, we installed a wide-angle effective compromise across the available time and life span video camera (Hero Black 5, GoPro, San Mateo, CA, USA) of the species. and two elongated, battery powered LED lights (LUMIstixx, We blinded audio file names prior to analysis and coding to avoid observer bias. From the dictaphone recordings of the territorial defense trials, we extracted the latency until the first head-body orientation (in s), the latency until the first jump (in s), the speed to reach the speaker (in cm/s), and if the male jumped (1) or not (0) during inter-bout-intervals (hereafter “inter-bout jumps,” cf. Ursprung et al. 2009) using the software VLC (VideoLAN 2021). Males that did not reach the speaker were coded a censored speed of 0 cm/s. Following the same reasoning, males that did not perform a head-body orientation or a jump were given a censored latency of 300 s (corresponding to the total duration of the experiment). In total, we conducted 163 valid territorial defense trials with 51 males (mean ± SD = 3.20 ± 1.31 rep- etitions per individual). Quantification of exploratory tendency and boldness To assess levels of exploratory tendency and boldness in male and female A. femoralis, we used a Novel Environment Test (NET), an approach that has been commonly used in personality studies (Carter et al. 2013; Kelleher et al. 2018). “Boldness” (or the corollary “shyness”) is defined as “an individual’s reaction to any risky situation,” and “explora- Fig. 2 Schematic (a) and picture (b) of the Novel Environment Test. tion” as “an individual’s reaction to a new situation” (Réale The cooler box measured 50 × 25 × 29  cm, with a 10  cm PVC tube et  al. 2007). Therefore, we expected the personality trait attached on one side of the box. A sliding door separated the shelter “boldness” to affect the latency to go out of a dark shel - from the box. Two LED tubes and a Hero Black 5 Go pro camera ter into a bright (novel) environment and the probability were attached to the lid of the box. Three solid PVC tubes (10  cm height, 5 cm diameter) were placed inside the box as visual obstacles to enter the novel environment, and the personality trait at randomized positions every day. A grid was drawn on the floor “exploration” to affect the distance travelled, the number of the cooler box to help randomize the position of the obstacles. A of jumps performed, and the area covered in the novel envi- mesh net was placed in the cooler box at 20 cm height to prevent the ronment. Those behaviors were chosen in accordance with frog from jumping on a wall outside of the camera range 1 3 Behavioral Ecology and Sociobiology ( 2 0 2 2) 76:93 Page5of14 93 Osram/Ledvance, Garching, Germany) for homogeneous Territory size, habitat complexity, and number illumination. We set the camera to “superview” mode to of neighbors ensure full visibility of the Novel Environment. We installed a mesh net in the Novel Environment at 20 cm height to To find out the distribution patterns of behaviors with respect prevent the frog from jumping on a wall outside of the cam- to a male’s natural and social environment, we determined era range. We used three solid PVC tubes (10 cm height, the territory size and the local complexity of the habitat, as 5 cm diameter) as visual obstacles, to motivate the frogs well as the number of female and male neighbors for each to explore the entire Novel Environment. The positions of male. This was only studied in males, as in A. femoralis the obstacles were changed daily and their positions were males acquire and defend non-overlapping territories, while determined using a random number generator. female display only site fidelity (Ringler et al. 2009, 2011 At the beginning of each trial, we placed all individuals 2012; Fischer et al. 2020). We used Dirichlet tessellation in in the shelter for 10 min to allow them to acclimatize. The ArcGIS to approximate male territories as Voronoi polygons shelter remained accessible throughout the trials to encour- (Voronoi 1908) on a day-to-day basis (for more information, age natural behaviors within the Novel Environment (Carter see Supplementary Materials). For the daily territory estima- et al. 2013; Kelleher et al. 2018), as the individual was free tion, we applied a roving-window approach, using all capture to remain inside or return to the shelter at any time. After points of the last 5 days a male was seen, including the focal this acclimation period, we switched on the lights and the day, but excluding all points where a male was likely found camera, closed the lid of the Novel Environment, opened outside its territory (i.e., all capture points linked to tadpole the sliding door between shelter and Novel Environment, transport). To ensure that the sizes of territories located at and filmed for 15 min. The minimum duration between two the population periphery were not over-estimated (Ringler consecutive tests with the same individual was 24 h (on aver- et al. 2009), we included the vertices of the island outline age tests were 11.36 ± 8.43 SD days apart). into the Dirichlet tessellation procedure, to establish a buffer We blinded video file names prior to analysis and cod - from the edge of the island until halfway to the outermost ing to avoid observer bias. To analyze the video recordings capture points of peripheral males. To ensure that estimates obtained during the NET, we used the software TRACKER at the beginning and end of the season were not impacted by (Brown 2019) to correct for distances that were distorted individuals not being captured, we used the POPAN formula by the camera wide-angle lens. Using the coding software (Schwarz and Arnason 1996) in program MARK (White BORIS (Friard et al. 2016), we assessed the latency from the and Burnham 1999; White 2020) to correct our population opening of the sliding door until frogs left the shelter (in s) estimates at these points in the breeding season (for more and the number of jumps performed inside the Novel Envi- detailed information, see Supplementary Materials). ronment. For individuals who stayed inside the shelter dur- After conducting the Dirichlet tessellation, we dissolved ing the entire experiment, the time spent in the shelter was the resulting Voronoi polygons based on male identity to censored to a value of 900 s (corresponding to the total dura- obtain “Voronoi territories” for each male and then calcu- tion of the experiment). We also coded the decision to enter lated territory sizes in ArcGIS. We also counted the direct the Novel Environment (1) or not (0) as a binomial response. male neighbors for each focal male from the daily Voronoi Using the automated tracking software TOXTRAC (Rod- territories. To obtain the number of female neighbors, we riguez et al. 2018), we measured the distance travelled (in calculated female centroid points (mean center) across the pixels) as well as the area visited inside the Novel Environ- entire season. Based on the typical distances females com- ment. For the latter variable, the software divided the floor mute for mating (Ringler et al. 2012; Fischer et al. 2020), we of the Novel Environment into a 9 × 8 grid and automatically constructed 20-m buffer circles and counted the number of counted the number of distinct squares a frog visited during contained centroid points of male Voronoi territories. a trial. As individuals varied in the time they spent inside To determine the complexity of the habitat for each indi- the Novel Environment, we standardized the data collected vidual male, we took four photographs (camera in automatic by only considering movements performed during the first (P) mode, focal length 50 mm, no flash, jpg images) from 2 min after leaving the shelter. On average, individuals spent each of the cardinal directions (0°, 90°, 180°, and 270° 624.21 s inside the Novel Environment (± 303.13 SD). This towards magnetic north) of a 100 cm × 100 cm red fabric. timeframe allowed for meaningful levels of movement to The fabric was placed at 3.5 m from the centroid point of occur but maximized the available data. Using this criterion, each male’s territory at forest floor level as this represents 21 out of 259 tests had to be excluded from the exploration half the radius of the typical male territory (cf. Ringler et al. analysis, leaving a total of 238 valid NET trials with 52 2011). We positioned the camera 20 cm above the forest males and 35 females (mean ± SD = 2.74 ± 1.33 repetitions floor, corresponding to the perch location of a male dur - per individual). Only very few individuals did not leave the ing territorial advertisement calling. We calculated habitat shelter in all repetitions. complexity as the average percentage of red fabric that was 1 3 Behavioral Ecology and Sociobiology ( 2 0 2 2) 76:93 93 Page6of14 covered by vegetation on the four pictures, with increasing time spent in the shelter). In addition, Bayesian generalized coverage of the fabric representing more complex habitat. linear mixed models were used to confirm that there were For that, we cropped each image in Paint.net v4.1.6 (Brew- no impacts from the fact that SEM analysis combines within ster 2019) to show only the lower half of the red fabric as and between individual effects (for details see Supplemen - this would represent the most visually relevant part of the tary Material). habitat for the frogs. Finally, we counted the number of pixels of visible fabric to calculate the percentage of fabric covered by vegetation. To facilitate this process, we aimed at Personality traits increasing the differences between fabric and vegetation by setting the image hue to 180 and decreasing the luminance To investigate if A. femoralis exhibits personality (i.e., and contrast to − 20. If males changed their territory location between-individual variation in behaviors), we assessed the during the study period, analysis of habitat complexity was repeatability (R) of all measured behaviors using the “rpt” performed in the territory in which most experimental trials function in the rptR package (Stoffel et al. 2017). Latency had been conducted. until the first head-body orientation and until the first jump had to be log transformed to achieve normal distribution. Statistical analysis We used the function “transformTukey” to apply a constant transformation on the speed to reach the speaker, the dis- We conducted all statistical analyses in R v3.6.0 (R Core tance travelled, and the time spent in the shelter. Repeat- Team 2020) using the integrated development environment ability was estimated based on the models applied to the RStudio v1.3.1093 (RStudio Team 2019). transformed data. We estimated repeatability from models fitted with a Gaussian error distribution for the latency until the first head-body orientation and until the first jump, the Structure of behaviors speed to reach the speaker, the time spent in the shelter, and the distance travelled. We estimated repeatability from mod- To determine how the measured behaviors are structured els fitted with a Poisson error distribution for the number of into functional units (i.e., aggression, exploration and bold- jumps and the number of areas, and from models fitted with ness), we investigated the phenotypic covariance structure a binary distribution for the inter-bout jumps and the prob- among different measurements during the behavioral tests. ability to enter the box. For all models, ID was included as To infer a latent variable describing aggression from behav- a random effect. iors measured during the territorial defense test (e.g., speed to reach the speaker, latency until the first head-body orien - tation, latency until the first jump, inter-bout jumps), using Distribution of behaviors across the natural the SEM package, we applied structural equation model- and social environment ling to the phenotypic covariance matrix derived from the means of each behavior for each individual, in order to We used a bivariate approach to study how behaviors meas- avoid pseudo-replication (Fox et al. 2020). Models were ured during the territorial defense trial and the NET corre- compared, to determine the most parsimonious model (for late with variation in the habitat complexity, territory size, details see Supplementary Fig. 1), based on differences in and number of male and female neighbors at the among- and Akaike’s information criterion (AIC) values, with small val- within-individuals level using Bayesian generalized linear ues indicating higher parsimony and a ΔAIC ≥ 2 indicating mixed models (Hadfield 2010). In the analyses, the latency significant differences (Burnham and Anderson 2002). We until the first jump, distance travelled, and time spent in expected the best model to have a latent variable explaining the shelter were chosen as they best represented the latent the covariance among the four measurements. We applied variables of aggression, exploration, and boldness, respec- the same technique to determine whether two latent vari- tively (see the “Structure of behaviors” section and Fig. 3). ables describing boldness and exploration can be inferred Although the probability to go inside the box was slightly from the behaviors measured during the NET (i.e., distance more correlated to boldness than “time spent in shelter,” we travelled, number of jumps, number of areas, time spent decided to use the latter variable to avoid using a binomial in the shelter, and probability to enter the box; for details variable in the model. Models were built with the trans- see, Supplementary Fig. 2). We expected the best model to formed data (see the “Personality traits” section). have two correlated latent variables, one that would explain To investigate the among-individual covariance between explorative behaviors (e.g., distance travelled, number of behaviors and environment variables, we divided each of the jumps, number of areas) and one that would explain bold- environmental variables by their mean value and added all of ness related behaviors (e.g., probability to enter the box, them as response variables (see Houslay and Wilson 2017). 1 3 Behavioral Ecology and Sociobiology ( 2 0 2 2) 76:93 Page7of14 93 Table 1 Repeatability (R) and confidence intervals (CI) of the behav - iors measured during the territorial defense test and the Novel Envi- ronment Test (NET) Test Variable R 95% CI Territorial Latency until head-body orienta- 0.17 [0.02; 0.34] defense tion test Latency until first jump 0.24 [0.07; 0.40] Speed to reach the perimeter 0.37 [0.19; 0.53] Inter-bout jumps 0.20 [0.001; 0.37] NET Probability to go in the box 0.44 [0.13; 0.89] Time spent in the shelter 0.30 [0.15; 0.44] Distance travelled in the box 0.36 [0.21; 0.49] Number of jumps 0.45 [0.27; 0.62] Number of areas visited 0.48 [0.30; 0.62] with univariate linear mixed models corrected for multiple comparisons, which led to the same biological conclusions. Results Personality traits Among the measures taken from the territorial defense test, all were considered repeatable and ranged in repeatability from 0.17 to 0.37 (Table 1; Fig. 4a). The best SEM model Fig. 3 Path diagrams of the best structure equation models (SEMs) supports a latent variable explaining the covariance of the (based on difference in Akaike’s information criterion (AIC) values) explaining the covariance structure among four behaviors assessed four behavioral responses measured in a territorial defense during an aggressivity test (a), and five behaviors assessed during a test (Fig. 3a; Supplementary Fig. 1b). Together, these results Novel Environment Test (b). “HBO” refers to the latency until the suggest that A. femoralis exhibits a personality trait “aggres- first head-body orientation. Squares represent the variances of the dif - siveness” encompassing the latency until the first head-body ferent behaviors explained by the SEM structure (R ). Numbers asso- ciated with arrows are standardized factor loadings which represent orientation and until the first jump towards a calling intruder, how behavioral responses are predicted to change based on changes the speed to reach the intruder, and the probability to jump to the latent variable. Number in brackets represent variances of during inter-bout. However, the residual variances were high residuals or error variances (e) associated to each behavior. All sim- (Fig. 3a; Supplementary Fig. 1b) and another model encom- ulated models can be found in the Supplementary material (Supple- mentary Fig. 1, 2) passing two latent variables had an AIC score close to the best model (∆AIC = 1.7; Supplementary Fig. 1 g). Similarly, we measured individual behavioral responses We included age (i.e., as a binomial trait: newly encoun- in a NET, where we found significant repeatabilities in all tered adults vs. recaptures from previous years) as a trait- behavioral measurements (Table 1; Fig. 4b, c). Based on specific fixed effects to control for effects of age on all our the comparison of the SEM, the best model supported the response variables. To calculate the variance due to differ - existence of a latent variable explaining the covariance pat- ences among individuals and the covariance between meas- terns of three behavioral measurements (distance travelled, ured behaviors, we fitted an unstructured covariance matrix number of jumps, and number of areas). It also supports that for the grouping variable ID. We then used the posterior another latent variable can be derived from the covariance of distributions to estimate the among- and within-individual two behavioral measurements (time spent inside the shelter correlations and covariances between each of the behavior and probability to enter the box). The two latent variables measured and the environment. We assumed statistical sig- are correlated (Fig. 3b; Supplementary Fig. 2 h). Together, nificance if the 95% credible intervals did not overlap 0 and these results suggest that A. femoralis males and females performed the same model verification as previously (see exhibit a personality trait “exploration” encompassing the the “Structure of behaviors” section). We also fitted these distance travelled, the number of jumps performed, and the relationships between behaviors and environment variables area covered in a new environment. The results also suggest 1 3 Behavioral Ecology and Sociobiology ( 2 0 2 2) 76:93 93 Page8of14 Fig. 4 Range of variation in the three behaviors that best repre- using a log (a) or constant (b and c) transformation. Boxes indicate sented the latent variables of aggression, exploration and boldness. the inter quartile range, with the central line depicting the median and The latency until the first jump ( a), the time spent in the shelter (b), the whiskers extending to 1.5*IQR. Dots represent the results of each and the distance travelled in the Novel Environment Test (c) are pre- trial. Males are ordered by their median (represented as a horizontal sented for individual males. All variables have been transformed bold line) the existence of the personality trait “boldness,” encompass- by vegetation in the measure of habitat complexity ing the latency to leave a safe place and enter a new environ- (SD = 9.31). Males occupied territories of 669.31 m on ment and the probability to enter a new environment. average (SD = 440.99). Males had on average 7 female For both the behaviors measured in the territorial defense and 5 male neighbors (mean of individual means) across test and in the NET, the phenotypic covariances in the SEM the season (absolute range females: 1–15, absolute range were mostly driven by the within-individual variances. The males: 1–10). Throughout the season, the average vari- results of the Bayesian models also show that the direc- ation in the number of both female and male neighbors tion of the among- and within-individual covariances was was 2 for individual territory holders (range females: 0–7; similar (Supplementary Tables 2, 3). Therefore, doing the range males: 0–6). SEM using only the among- or within-individual covari- There was no relation between the aggressive responses ance matrix would have resulted in a similar interpretation of a male during a territorial defense test and its social or of the results. natural environment. However, there was a significant rela - tion between the level of exploration and boldness of a male Behavior and the characteristics of an individual and the number of female neighbors at the within-individual and its environment level (Supplementary Table 4). This suggests that individu- als respond plastically to their social environment, increas- Next, we investigated the correlation between the behav- ing their level of exploration and boldness when the number iors measured during the territorial defense trial and of females around increases. The spatial setup of the territo- habitat complexity, territory size, and the number of ries on the day with the most individuals present at the same male and female neighbors. On average, habitats were time (07th of March 2019) is presented in Fig. 5. The aver- relatively complex with 82.77% of the fabric covered age response of each individual in terms of latency until the 1 3 Behavioral Ecology and Sociobiology ( 2 0 2 2) 76:93 Page9of14 93 Fig. 5 Maps showing the spatial distribution of individual performance of male frogs on the island in the behavio- ral essays. The maps use the Voronoi territories of 7 March 2019 when the most individuals were present at the same time on the island. The maps show the mean value, calculated over all respective trials of (a) the individuals’ latency to jump in the territorial defense trial, (b) the time spent in the shelter during the Novel Environment Test (NET), and (c) the distance travelled in the NET. All color ramps have 20 equal intervals across the full range of the respective value; darker colors represent shorter latency in (a), shorter time spent in the shelter in (b), and longer distances in (c). Black squares indicate the 14 artificial pools that were in place on the island since 2018; thin gray lines show 50 cm elevation isoclines; the blue area shows the river Arataye. The territories of two males that were not tested in the behavioral assays are shown with a hatched white pattern first jump, time spent in the shelter, and distance travelled to an individual’s natural and social environment in a wild, in the box is also represented. free-ranging, entire population of A. femoralis. Personality traits Discussion Our analysis showed that the repeatability of the variables In the present study, we investigated the structure of poten- measured in the territorial defense trial (0.17 to 0.37) was tial personality traits in A. femoralis and determined how in the lower range of the repeatability found in most stud- behaviors related to aggressiveness, exploration, and bold- ies (with overlapping confidence intervals; Bell et al. 2009; ness are structured into functional units (i.e., personality mean = 0.37, 95% confidence limits = 0.35–0.38) but con- traits). We also investigated how individual behaviors relate sistent with the average findings in other personality studies 1 3 Behavioral Ecology and Sociobiology ( 2 0 2 2) 76:93 93 Page10of14 on amphibians (Brodin et al. 2013; Maes et al. 2013; Gif- Future studies are needed to identify the link between call- ford et al. 2014; González-Bernal et al. 2014). Moreover, ing activity and individual reproductive success, in order to we found that a latent variable explains the covariance of investigate possible trade-offs in A. femoralis. the four behaviors measured during a territorial defense test Finally, we found that the distance travelled, the num- (i.e., the latency until the first head-body orientation and ber of jumps performed, and the area covered in a new until the first jump towards a calling intruder, the speed to environment were highly repeatable. Moreover, we found reach the intruder and the probability to jump during the that a latent variable explained some of the covariance inter-bout interval). Together, these results suggest that A. among the three behavioral measurements. Together, femoralis males exhibit a personality trait “aggressiveness.” these results suggest that A. femoralis males and females However, a great portion of the speed and the inter-bout exhibit a personality trait “exploration.” Exploration jumps measured during the territorial defense trial were behavior is especially relevant for dispersal and resource explained by exogenous factors. Together with the existence acquisition (Dingemanse et al. 2003; Gruber et al. 2017). of another model with similar support, these results sug- In A. femoralis, males rely on spatial memory to find gest that the behavioral measures capture different aspects water bodies (Pašukonis et al. 2016) and distribute their of aggressiveness. In this model (Supplementary Fig. 1 g), tadpoles across multiple sites to decrease risks of losing a latent variable explains the covariance of the latencies entire clutches due to desiccation or predation (Erich et al. until the first head-body orientation and until the first jump. 2015; Ringler et al. 2018). Therefore, being more explora- Another latent variable explains the covariance between the tive might enable males to find more sites for tadpole speed to reach the speaker and the inter-bout jumps. We deposition or a better territory to settle, while females interpret these two latent variables as reactivity and offen - put their explorative behaviors at use when looking for a siveness, respectively. Males of many amphibian species mate and, more rarely, for tadpole transport (Ringler et al. defend and fight over territories using acoustic and visual 2015). A highly explorative individual will, however, be displays (Hödl and Amezquita 2001; Toledo et al. 2007). For more conspicuous to predators or at risk of losing its ter- instance, in the Bornean rock-skipper frog (Staurois lato- ritory during periods of absence. palmatus), males perform foot-flagging and advertisement calls to defend their territories against conspecific intruders Distribution of behaviors across the natural (Preininger et al. 2009). Our results show that more aggres- and social environment sive males react fiercer towards conspecific territory intrud - ers. This is of particular ecological relevance, because terri- Contrary to our expectation, we did not find a relationship tory possession is the most important prerequisite for male between individual behaviors and their natural environment reproductive success in this species (Ursprung et al. 2011a). (i.e., habitat complexity and territory size). However, we However, high levels of aggression might also come with a cannot exclude that other habitat characteristics that we did cost, if more aggressive individuals engage more often in not measure, such as quantity and quality of the leaf lit- energetically costly and potentially harmful fights. ter, tree species, or canopy cover, are linked to individual We also found that the probability to enter a novel envi- behavior. Likewise, we did not find any relationship between ronment and the time it took to do so were highly repeatable, individual behavior and territory size, as more aggressive and a latent variable explained a relatively large part of the and bolder individuals settled in territories of varying sizes covariance among the two variables. Together, these results (Fig. 5a, b). This is in line with the previous finding that in suggest the existence of the personality trait “boldness” in A. A. femoralis, only the possession of a territory is important femoralis males and females. Boldness relates to the reaction for reproductive success, but not territory size (Ursprung of an individual to a predator, a novel object, or a conspecific et al. 2011a), and as a consequence, males cannot increase and is relevant in many contexts such as predator avoidance, their reproductive success by extending territorial space. As feeding, or mating (Réale et al. 2007). In A. femoralis, a gen- already suggested in a previous study, variation in territory erally high propensity to take risks might be reflected not sizes is probably strongly dependent on characteristics of only in the response to a predation threat, but also in how the natural environment (Ringler et al. 2017) rather than a prominently the advertisement call is presented. Thus, higher consequence of varying levels of aggression in A. femoralis risk taking could be reflected in males calling at higher rates, males. higher amplitude, over longer durations, from more exposed More explorative individuals were not necessarily locations, or moving/turning more during calling. In A. sub- located near the artificial pools on the island (Fig.  5c), folionidificans , a closely related species, calling activity has although water bodies are of critical importance for indi- been found to be positively related to reproductive success vidual fitness, given that they are obligate for tadpole (Souza et al. 2021). Being bold, however, can also be costly development after hatching. Previous studies have shown as it might lead to more frequent encounters with predators. that males distribute their tadpoles across multiple water 1 3 Behavioral Ecology and Sociobiology ( 2 0 2 2) 76:93 Page11of14 93 bodies located outside their territories (Ringler et  al. Conclusion 2013, 2018; Erich et  al. 2015; Beck et  al. 2017). This suggests that the ability of an individual to find water We studied the structure of personality traits and the dis- bodies once it has settled in a territory is critical, while tribution of behaviors across the environment in an entire the location of the territory is not. Finding water bodies free-ranging population of the poison frog A. femoralis by in A. femoralis is strongly related to olfactory sensing measuring several behaviors in situ in different contexts. (Serrano-Rojas and Pašukonis 2021) and spatial memory We found that A. femoralis indeed exhibits animal person- might help in revisiting such sites once they have been ality along the aggressiveness, exploration, and boldness encountered (Pašukonis et al. 2016; Beck et al. 2017). axes. Furthermore, we found non-random distribution of Taken together, our results suggest that A. femoralis behaviors across the animals’ social environment, which males establish their territories independent of large- may allow individuals to cope with their complex socio- scale resource distribution and the wider structure of the ecological environment. While amphibians have been largely habitat. And likewise, the characteristics of their natural overlooked in animal personality research, this study is one environment apparently are not associated with individual of the first comprehensive study of animal personality in differences in behavior. amphibian in the wild. By providing a detailed description We also did not find support for a relationship between of how behavioral measurements are structured in functional aggressiveness and the social environment. We initially units that allow individuals to cope with their socio-eco- expected more aggressive individuals to occupy territories logical environment, it broadens our understanding of the in high-density areas, where they would be more likely to functional role of behavior in frogs and offers a first step find mating partners but also face elevated male-male com - towards understanding the mechanisms that play a role in the petition. However, with this trade-off between female avail - emergence and maintenance of behavioral variation. ability and intra-sexual competition, males might have equal Supplementary Information The online version contains supplemen- reproductive outcomes regardless of their level of aggres- tary material available at https://doi. or g/10. 1007/ s00265- 022- 03202-9 . siveness and independent of the overall density, as long as they manage to establish a territory at all. Future studies Acknowledgements The authors thank B. Furdi for helping in pro- should investigate how the interplay between aggressiveness cessing the data and R. Fouchier for providing Fig. 1. The authors also and population density affects reproductive success. thank E. Haeler and C. Eliasch for their assistance during field work, and B. Szabo for comments on earlier versions of this manuscript. We Our results show that exploration- and boldness-related are grateful to the staff of CNRS Guyane for logistic support in French behaviors were positively linked to the number of females in Guiana. We thank the Nouragues research field station (managed by the vicinity to male territories. Since significant correlations CNRS) which benefits from “Investissement d’Avenir” grants man - were only found at the within-individual level, this suggests aged by Agence Nationale de la Recherche (AnaEE France ANR-11- INBS-0001; Labex CEBA ANR-10-LABX-25-01). This work has been that the link between exploration or boldness and the social done on the Natural Reserve of Nouragues. We thank the editors and environment is mainly driven by individual plasticity. This two anonymous reviewers for valuable comments on this and previous indicates that males, who mostly have stable territories and versions of the manuscript. move little, increase their overall levels of exploration and boldness when the number of females around their terri- Author contribution Conceptualization: MP, YA, ER; Methodology: MP, LB, SC, KD, CL, YA, PW, MR, ER; Formal Analysis: MP, LB, tory increases. There are currently no studies showing any MR; Investigation: MP, SC, KD, CL, YA, MR, ER; Resources: ER; direct mechanisms how males could assess the presence and Data curation: MP, MR, CL; Writing—original draft: MP; Writing— number of nearby females. We suggest that future studies review and editing: MP, YA, PW, MR, ER; Supervision: MP, MR, ER; should investigate secondary or indirect cues, such as dis- Project administration: MR, ER; Funding acquisition: ER. tribution or density of feeding sites, that might determine Funding Open access funding provided by University of Bern. This female density and that might allow males to identify areas work was funded through a Standalone Grant (FWF P-31518) from with more females. However, because the among-individual the Austrian Science Fund (FWF). MR was supported by an Erwin effects have broad credible intervals that are not centered Schrödinger Grant (FWF J-3868) and a Standalone Grant (FWF around zero, we cannot rule out the possibility of non-ran- P-33728) from the Austrian Science Fund (FWF). dom settlement. Data availability The datasets generated during and/or analyzed during Our study does not claim or identify any causal relation- the current study are available in the Open Science Framework reposi- ship between behaviors or personality traits and their natural tory: https:// doi. org/ 10. 17605/ OSF. IO/ AJV5X. and/or social environment. Still, the identification of such relationship in a natural free-ranging population of animals Declarations provides a first step towards understanding the mechanisms underlying the distribution of behaviors across space (cf. Ethics approval This study was approved by the scientific committee Archard and Braithwaite 2010). of the “Nouragues Ecological Research Station” and the ethics and 1 3 Behavioral Ecology and Sociobiology ( 2 0 2 2) 76:93 93 Page12of14 animal welfare committee of the University of Veterinary Medicine common frog (Rana temporaria). Behav Ecol Sociobiol 67:135– Vienna in accordance with Good Scientific Practice (GSP) guidelines 143. https:// doi. org/ 10. 1007/ s00265- 012- 1433-1 and national legislation. Sampling for this study was conducted in Burnham KP, Anderson DR (2002) Model selection and multimodel strict accordance with current French and EU law and followed the inference: a practical information-theoretic approach, 2nd edn. (ASAB 2020) guidelines on the treatment of study animals. Work- Springer, New York ing and tissue sampling permissions were provided by the CNRS Carter AJ, Feeney WE, Marshall HH, Cowlishaw G, Heinsohn R Guyane (“Centre National de la Recherche Scientifique Guyane”), by (2013) Animal personality: what are behavioural ecologists meas- the “Ministère de la transition écologique et solidaire” (permit num- uring? Biol Rev 88:465–475. https:// doi. org/ 10. 1111/ brv. 12007 ber: TREL2002508S/303), and the Secretariat of the Convention on Chaloupka S, Peignier M, Stückler S, Araya-Ajoy Y, Walsh P, Ringler Biological Diversity (APA declaration: TREL1734890A/34). M, Ringler E (2022) Repeatable territorial aggression in a Neo- tropical poison frog. Front Ecol Evol 10:881387. https:// doi. org/ Competing interests The authors declare no competing interests.10. 3389/ fevo. 2022. 881387 Conover DO, Schultz ET (1995) Phenotypic similarity and the evo- lutionary significance of countergradient variation. Trends Ecol Open Access This article is licensed under a Creative Commons Attri- Evol 10:248–252. https:// doi. or g/ 10. 1016/ S0169- 5347(00) bution 4.0 International License, which permits use, sharing, adapta- 89081-3 tion, distribution and reproduction in any medium or format, as long Crump ML (1972) Territoriality and mating behavior in Dendrobates as you give appropriate credit to the original author(s) and the source, granuliferus (Anura: Dendrobatidae). Herpetologica 28:195–198 provide a link to the Creative Commons licence, and indicate if changes Dingemanse NJ, Araya-Ajoy YG (2015) Interacting personalities: were made. The images or other third party material in this article are behavioural ecology meets quantitative genetics. Trends Ecol included in the article’s Creative Commons licence, unless indicated Evol 30:88–97 otherwise in a credit line to the material. If material is not included in Dingemanse NJ, Both C, van Noordwijk AJ, Rutten AL, Drent PJ the article’s Creative Commons licence and your intended use is not (2003) Natal dispersal and personalities in great tits (Parus permitted by statutory regulation or exceeds the permitted use, you will major). Proc R Soc Lond B 270:741–747. https://doi. or g/10. 1098/ need to obtain permission directly from the copyright holder. To view a rspb. 2002. 2300 copy of this licence, visit http://cr eativ ecommons. or g/licen ses/ b y/4.0/ . Dingemanse NJ, Both C, Drent PJ, Tinbergen JM (2004) Fitness conse- quences of avian personalities in a fluctuating environment. Proc R Soc Lond B 271:847–852 Dingemanse NJ, Kazem AJN, Réale D, Wright J (2010) Behavioural References reaction norms: animal personality meets individual plasticity. Trends Ecol Evol 25:81–89. https:// doi. org/ 10. 1016/j. tree. 2009. 07. 013 AmphibiaWeb (2022) AmphibiaWeb: information on amphibian biol- Dochtermann NA, Jenkins SH (2007) Behavioural syndromes in Mer- ogy and conservation riam’s kangaroo rats (Dipodomys merriami): a test of competing Araya-Ajoy YG, Dingemanse NJ (2014) Characterizing behavioural hypotheses. Proc R Soc Lond B 274:2343–2349. https:// doi. org/ ‘characters’: an evolutionary framework. Proc R Soc Lond B 10. 1098/ rspb. 2007. 0622 281:20132645. https:// doi. org/ 10. 1098/ rspb. 2013. 2645 Duckworth RA, Kruuk LEB (2009) Evolution of genetic integration Archard GA, Braithwaite VA (2010) The importance of wild popu- between dispersal and colonization ability in a bird. Evolution lations in studies of animal temperament. J Zool 281:149–160. 63:968–977. https:// doi. org/ 10. 1111/j. 1558- 5646. 2009. 00625.x https:// doi. org/ 10. 1111/j. 1469- 7998. 2010. 00714.x Erich M, Ringler M, Hödl W, Ringler E (2015) Brood-partitioning ASAB (2020) Guidelines for the treatment of animals in behavioural behaviour in unpredictable environments: hedging the bets? research and teaching. Anim Behav 159:I–XI. https:// doi. org/ 10. Behav Ecol Sociobiol 69:1011–1017. https:// doi. org/ 10. 1007/ 1016/j. anbeh av. 2019. 11. 002 s00265- 015- 1913-1 Beck KB, Loretto M-C, Ringler M, Hödl W, Pašukonis A (2017) Rely- Fischer M-T, Ringler M, Ringler E, Pašukonis A (2020) Reproductive ing on known or exploring for new? Movement patterns and repro- behavior drives female space use in a sedentary Neotropical frog. ductive resource use in a tadpole-transporting frog. PeerJ 5:e3745. PeerJ 8:e8920. https:// doi. org/ 10. 7717/ peerj. 8920 https:// doi. org/ 10. 7717/ peerj. 3745 Fox J, Nie Z, Byrnes J, Culbertson M, DebRoy S, Friendly M, Goodrich Bell AM (2005) Behavioural differences between individuals and two B, Jones RH, Kramer A, Monette G (2020) Package ‘sem’: Struc- populations of stickleback (Gasterosteus aculeatus). J Evol Biol tural Equation Models, https://CRAN .R-pr oject. or g/pac kag e=sem 18:464–473. https:// doi. org/ 10. 1111/j. 1420- 9101. 2004. 00817.x Friard O, Gamba M, Fitzjohn R (2016) BORIS: a free, versatile open- Bell AM, Hankison SJ, Laskowski KL (2009) The repeatability of source event-logging software for video/audio coding and live behaviour: a meta-analysis. Anim Behav 77:771–783. https://doi. observations. Methods Ecol Evol 7:1325–1330. https:// doi. org/ org/ 10. 1016/j. anbeh av. 2008. 12. 022 10. 1111/ 2041- 210X. 12584 Bolger DT, Morrison TA, Vance B, Lee D, Farid H (2012) A computer- Gasser H, Amézquita Torres A, Hödl W (2009) Who is calling? assisted system for photographic mark–recapture analysis. Meth- Intraspecific call variation in the aromobatid frog Allobates femo- ods Ecol Evol 3:813–822. https:// doi. org/ 10. 1111/j. 2041- 210X. ralis. Ethology 115:596–607. https://doi. or g/10. 1111/j. 1439- 0310. 2012. 00212.x 2009. 01639.x Bongers F, Charles-Dominique P, Forget P-M, Théry M (eds) (2001) Gifford ME, Clay TA, Careau V (2014) Individual (co)variation in Nouragues: dynamics and plant-animal interactions in a neo- standard metabolic rate, feeding rate, and exploratory behavior tropical rainforest. Monographiae biologicae, v. 80. Kluwer in wild-caught semiaquatic salamanders. Physiol Biochem Zool Academic Publishers, Dordrecht 87:384–396. https:// doi. org/ 10. 1086/ 675974 Brewster R (2019) paint.net, https:// www. getpa int. net/ González-Bernal E, Brown GP, Shine R (2014) Invasive cane toads: Brown D (2019) Tracker: Video analysis and modeling tool for phys- social facilitation depends upon an individual’s personality. PLoS ics education. Open Source Physics, Davidson. https:// physl ets. ONE 9:e102880. https:// doi. org/ 10. 1371/ journ al. pone. 01028 80 org/ track er/ Goursot C, Düpjan S, Kanitz E, Tuchscherer A, Puppe B, Leliveld Brodin T, Lind MI, Wiberg MK, Johansson F (2013) Personality LMC (2019) Assessing animal individuality: links between trait differences between mainland and island populations in the 1 3 Behavioral Ecology and Sociobiology ( 2 0 2 2) 76:93 Page13of14 93 personality and laterality in pigs. Curr Zool 65:541–551. https:// Réale D, Dingemanse NJ (2010) Personality and individual social doi. org/ 10. 1093/ cz/ zoy071 specialisation. In: Székely T, Moore AJ, Komdeur J (eds) Social Gruber J, Brown G, Whiting MJ, Shine R (2017) Is the behavioural behaviour: genes, ecology and evolution. Cambridge University divergence between range-core and range-edge populations of Press, Cambridge, pp 417–441 cane toads (Rhinella marina) due to evolutionary change or devel- Réale D, Reader SM, Sol D, McDougall PT, Dingemanse NJ (2007) opmental plasticity? R Soc Open Sci 4:170789. https://doi. or g/10. Integrating animal temperament within ecology and evolution. 1098/ rsos. 170789 Biol Rev 82:291–318. https:// doi. or g/ 10. 1111/j. 1469- 185X. Hadfield JD (2010) MCMC Methods for multi-response generalized 2007. 00010.x linear mixed models the MCMCglmmR package. J Stat Soft Ringler M, Ursprung E, Hödl W (2009) Site fidelity and patterns of 33:1–22 short- and long-term movement in the brilliant-thighed poison Hensley NM, Cook TC, Lang M, Petelle MB, Blumstein DT (2012) frog Allobates femoralis (Aromobatidae). Behav Ecol Sociobiol Personality and habitat segregation in giant sea anemones (Con- 63:1281–1293. https:// doi. org/ 10. 1007/ s00265- 009- 0793-7 dylactis gigantea). J Exp Mar Biol Ecol 426–427:1–4. https://doi. Ringler M, Ringler E, Magaña Mendoza D, Hödl W (2011) Intru- org/ 10. 1016/j. jembe. 2012. 05. 011 sion experiments to measure territory size: development of the Hödl W (1983) Phyllobates femoralis (Dendrobatidae): Rufverhalten method, tests through simulations, and application in the frog und akustische Orientierung der Männchen (Freilandaufnahmen). Allobates femoralis. PLoS ONE 6:e25844. https:// doi. org/ 10. Wissenschaftlicher Film, Bundesstaatliche Hauptstelle für Wis-1371/ journ al. pone. 00258 44 senschaftliche Kinematographie, Wien Ringler E, Ringler M, Jehle R, Hödl W (2012) The female perspec- Hödl W (1987) Dendrobates femoralis (Dendrobatidae): a handy fellow tive of mating in A femoralis, a territorial frog with paternal for frog bioacoustics. In: van Gelder J. J., Strijbosch H, Bergers care a spatial and genetic analysis. PLoS ONE 7:e40237. https:// P (eds) Proceedings of the 4th Ordinary General meeting of the doi. org/ 10. 1371/ journ al. pone. 00402 37 Societas Europaea Herpetologica, Nijmegen, Netherlands Ringler E, Pašukonis A, Hödl W, Ringler M (2013) Tadpole transport Hödl W, Amezquita A (2001) Visual signaling in anuran amphibians. logistics in a Neotropical poison frog: indications for strategic In: Ryan MJ (ed) Anuran communication. Smithsonian Institution planning and adaptive plasticity in anuran parental care. Front Press, Washington, DC, pp 121–141 Zool 10:67. https:// doi. org/ 10. 1186/ 1742- 9994- 10- 67 Houslay TM, Wilson AJ (2017) Avoiding the misuse of BLUP in Ringler E, Mangione R, Ringler M (2014) Where have all the tad- behavioural ecology. Behav Ecol 28:948–952. https:// doi. org/ 10. poles gone? Individual genetic tracking of amphibian larvae 1093/ beheco/ arx023 until adulthood. Mol Ecol Resour 15:737–746. https:// doi. org/ Kelleher SR, Silla AJ, Byrne PG (2018) Animal personality and behav-10. 1111/ 1755- 0998. 12345 ioral syndromes in amphibians: a review of the evidence, experi- Ringler E, Pašukonis A, Fitch WT, Huber L, Hödl W, Ringler M mental approaches, and implications for conservation. Behav Ecol (2015) Flexible compensation of uniparental care: female poi- Sociobiol 72:79. https:// doi. org/ 10. 1007/ s00265- 018- 2493-7 son frogs take over when males disappear. Behav Ecol 26:1219– Maes J, van Damme R, Matthysen E (2013) Individual and among- 1225. https:// doi. org/ 10. 1093/ beheco/ arv069 population variation in dispersal-related traits in Natterjack toads. Ringler M, Mangione R, Pašukonis A et al (2016) High-resolution Behav Ecol 24:521–531. https:// doi. org/ 10. 1093/ beheco/ ars193 forest mapping for behavioural studies in the Nature Reserve Narins PM, Hödl W, Grabul DS (2003) Bimodal signal requisite for ‘Les Nouragues’, French Guiana. J Maps 12:26–32. https://doi. agonistic behavior in a dart-poison frog, Epipedobates femoralis. org/ 10. 1080/ 17445 647. 2014. 972995 P Natl Acad Sci USA 100:577–580. https://d oi.or g/1 0.1 073/p nas. Ringler M, Szipl G, Hödl W, Khil L, Kofler B, Lonauer M, Provin 02371 65100 C, Ringler E (2017) Acoustic ranging in poison frogs—it is not Pašukonis A, Ringler M, Brandl HB, Mangione R, Ringler E, Hödl W about signal amplitude alone. Behav Ecol Sociobiol 71:114. (2013) The homing frog: high homing performance in a territorial https:// doi. org/ 10. 1007/ s00265- 017- 2340-2 dendrobatid frog Allobates femoralis (Dendrobatidae). Ethology Ringler E, Szipl G, Harrigan RJ, Bartl-Binder P, Mangione R, Ring- 119:762–768. https:// doi. org/ 10. 1111/ eth. 12116 ler M (2018) Hierarchical decision-making balances current and Pašukonis A, Trenkwalder K, Ringler M, Ringler E, Mangione R, future reproductive success. Mol Ecol 27:2289–2301. https:// Steininger J, Warrington I, Hödl W (2016) The significance of doi. org/ 10. 1111/ mec. 14583 spatial memory for water finding in a tadpole-transporting frog. Rodriguez A, Zhang H, Klaminder J, Brodin T, Andersson PL, Anim Behav 116:89–98. https:// doi. org/ 10. 1016/j. anbeh av. 2016. Andersson M (2018) ToxTrac: a fast and robust software for 02. 023 tracking organisms. Methods Ecol Evol 9:460–464. https://d oi. Pigliucci M (2003) Phenotypic integration: studying the ecology and org/ 10. 1111/ 2041- 210X. 12874 evolution of complex phenotypes. Ecol Lett 6:265–272. https:// RStudio Team (2019) RStudio: integrated development for R. doi. org/ 10. 1046/j. 1461- 0248. 2003. 00428.x RStudio, Inc., Boston, http:// www. rstud io. com Pigliucci M (2005) Evolution of phenotypic plasticity: where are we Roithmair ME (1994) Male territoriality and female mate selection going now? Trends Ecol Evol 20:481–486. https:// doi. org/ 10. in the dart-poison frog Epipedobates trivattatus (Dendrobatidae, 1016/j. tree. 2005. 06. 001 Anura). Copeia 1994:107–115. https:// doi. org/ 10. 2307/ 14466 76 Preininger D, Böckle M, Hödl W (2009) Communication in noisy Rojas B, Pašukonis A (2019) From habitat use to social behavior: nat- environments II: visual signaling behavior of male foot-flagging ural history of a voiceless poison frog. Dendrobates Tinctorius frogs Staurois latopalmatus. Herpetologica 65:166–173. https:// Peerj 7:e7648. https:// doi. org/ 10. 7717/ peerj. 7648 doi. org/ 10. 1655/ 08- 037R.1 Schwarz CJ, Arnason AN (1996) A general methodology for the analy- Pröhl H (2005) Territorial behavior in dendrobatid frogs. J Herpetol sis of capture-recapture experiments in open populations. Biom- 39:354–365. https:// doi. org/ 10. 1670/ 162- 04A.1 etrics 52:860–873. https:// doi. org/ 10. 2307/ 25330 48 Rasband WS (1997–2021) ImageJ. U. S. National Institutes of Serrano-Rojas SJ, Pašukonis A (2021) Tadpole-transporting frogs use Health, Bethesda, http:// imagej. nih. gov/ ij/ stagnant water odor to find pools in the rainforest. J Exp Biol. R Core Team (2020) R: a language and environment for statistical https:// doi. org/ 10. 1242/ jeb. 243122 computing. R Foundation for Statistical Computing, Vienna. Sih A, Cote J, Evans M, Fogarty S, Pruitt J (2012) Ecological implica- http:// www.r- proje ct. org tions of behavioural syndromes. Ecol Lett 15:278–289. https:// doi. org/ 10. 1111/j. 1461- 0248. 2011. 01731.x 1 3 Behavioral Ecology and Sociobiology ( 2 0 2 2) 76:93 93 Page14of14 Souza JRD, Kaefer IL, Lima AP (2021) Calling activity determines Ursprung E, Ringler M, Jehle R, Hödl W (2011b) Toe regeneration male mating success in a territorial frog with parental care. Ethol- in the neotropical frog Allobates femoralis. Herpetol J 21:83–86 ogy 127:359–365. https:// doi. org/ 10. 1111/ eth. 13135 VideoLAN (2021) VLC media player. VideoLAN, Paris. https://www . Sprau P, Dingemanse NJ (2017) An approach to distinguish between video lan. org/ vlc/ plasticity and non-random distributions of behavioral types along Voronoi G (1908) Nouvelles applications des paramètres continus à la urban gradients in a wild passerine bird. Front Ecol Evol 5:92. théorie des formes quadratiques. Premier mémoire. Sur quelques https:// doi. org/ 10. 3389/ fevo. 2017. 00092 propriétés des formes quadratiques positives parfaites. J Reine Stoffel MA, Nakagawa S, Schielzeth H (2017) rptR: Repeatability esti - Angew Math 1908:97–102. https:// doi. or g/ 10. 1515/ cr ll. 1908. mation and variance decomposition by generalized linear mixed-133. 97 effects models. Methods Ecol Evol 8:1639–1644. https:// doi. org/ Wells KD (2007) The ecology and behavior of amphibians. University 10. 1111/ 2041- 210X. 12797 of Chicago Press, Chicago Toledo LF, Araújo OGS, Guimarães LD, Lingnau R, Haddad CFB White GC (2020) MARK. Colorado State University, Fort Collins (2007) Visual and acoustic signaling in three species of Brazil-http:// phidot. org/ softw are/ mark/ index. html ian nocturnal tree frogs (Anura Hylidae). Phyllomedusa 6:61–68. White GC, Burnham KP (1999) Program MARK: survival estimation https:// doi. org/ 10. 11606/ issn. 2316- 9079. v6i1p 61- 68 from populations of marked animals. Bird Study 46:S120–S139. Tremmel M, Müller C (2013) Insect personality depends on environ-https:// doi. org/ 10. 1080/ 00063 65990 94772 39 mental conditions. Behav Ecol 24:386–392. https:// doi. org/ 10. Yang Y, Blomenkamp S, Dugas MB, Richards-Zawacki CL, Pröhl 1093/ beheco/ ars175 H (2019) Mate choice versus mate preference: inferences about Ursprung E, Ringler M, Hödl W (2009) Phonotactic approach pattern color-assortative mating differ between field and lab assays of in the neotropical frog Allobates femoralis: a spatial and temporal poison frog behavior. Am Nat 193:598–607. https:// doi. org/ 10. analysis. Behaviour 146:153–170. https:// doi. org/ 10. 1163/ 15685 1086/ 702249 3909X 410711 Zidar J, Balogh A, Favati A, Jensen P, Leimar O, Løvlie H (2017) A Ursprung E, Ringler M, Jehle R, Hödl W (2011a) Strong male/male comparison of animal personality and coping styles in the red competition allows for nonchoosy females: high levels of polygy- junglefowl. Anim Behav 130:209–220. https:// doi. org/ 10. 1016/j. nandry in a territorial frog with paternal care. Mol Ecol 20:1759–anbeh av. 2017. 06. 024 1771. https:// doi. org/ 10. 1111/j. 1365- 294X. 2011. 05056.x Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. 1 3 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Behavioral Ecology and Sociobiology Springer Journals

Exploring links between personality traits and their social and non-social environments in wild poison frogs

Loading next page...
 
/lp/springer-journals/exploring-links-between-personality-traits-and-their-social-and-non-gqzDi62L3P

References (86)

Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2022
ISSN
0340-5443
eISSN
1432-0762
DOI
10.1007/s00265-022-03202-9
Publisher site
See Article on Publisher Site

Abstract

An animal’s behavioral phenotype comprises several traits, which are hierarchically structured in functional units. This is manifested in measured behaviors often being correlated, partly reflecting the need of a coordinated functional response. Unfortunately, we still have limited understanding whether consistent differences in animal behaviors are due to underlying physiological constraints or a result of plastic adaptation to their current environment. Therefore, characterizing the spatial distribution of behaviors can provide important insights into causes and consequences of behavioral variation. In the present study, we quantified behaviors in a wild, free-ranging population of the Neotropical frog Allobates femoralis. We investigated how these behaviors were linked to the frogs’ natural and social environment and quantified the extent to which these behav - iors consistently differed among individuals (i.e., animal personality). We assessed levels of aggressiveness, exploration, and boldness by measuring several underlying behaviors expressed in a set of experimental assays, and found evidence for consistent among-individual differences along these axes. Contrary to our expectation, there was no relationship between individual behaviors and their natural environment, but we found a plastic response of males to changes in female density, which might reflect how individuals cope with their socio-ecological environment. Significance statement How are behavioral phenotypes distributed across space? Here, we studied an entire free-ranging population of poison frogs, and investigated if the personality traits aggressiveness, exploration, and boldness are linked to the frogs’ natural or social environment. We found that behavioral traits were non-randomly distributed across the population, suggesting that the spatial arrangement of behavioral traits reflects how individuals cope with their complex natural and social environment. Keywords Behavioral variation · Animal personality · Poison frogs · Non-random distribution · Environment Communicated by A. Taylor Baugh. * Mélissa Peignier Department of Biology and Ecology, University melissa.peignier@unibe.ch of Montpellier, Montpellier, France 1 6 Department of Evolutionary Biology, University of Vienna, Division of Behavioural Ecology, Institute of Ecology Vienna, Austria and Evolution, University of Bern, Wohlenstrasse 50a, CH-3032 Hinterkappelen, Bern, Switzerland iES Landau, Institute for Environmental Sciences, University of Koblenz-Landau, Mainz, Germany Messerli Research Institute, University of Veterinary Medicine Vienna, Vienna, Austria Central Research Laboratories, Natural History Museum Vienna, Vienna, Austria Department of Behavioral and Cognitive Biology, University of Vienna, Vienna, Austria Institute of Ecology and Evolution, School of Biological Sciences, University of Edinburgh, Edinburgh, UK Department of Biology, Centre for Biodiversity Dynamics, Norwegian University of Science and Technology, University of Music and Performing Arts Graz, Graz, Austria Trondheim, Norway Vol.:(0123456789) 1 3 Behavioral Ecology and Sociobiology ( 2 0 2 2) 76:93 93 Page2of14 succeed in establishing a territory in high-density patches, Introduction where they have access to more mates but face elevated intra- sexual competition. In turn, the link between aggressiveness Behavioral variation is ubiquitous in nature. Behaviors may and population density could also be caused by individual vary considerably among species; and within species, we find plasticity allowing individuals to match their aggressiveness variation both among and within individuals. Behavioral vari- to competition levels. Regardless of the mechanisms linking ation at all of these levels plays a major role in reproduction behavior and environment, identifying non-random spatial dis- and survival, for example, affecting the risk of being detected tribution of behaviors provides important insights into their and caught by predators, likelihood of dispersal, foraging effi - function and their role in allowing organisms to cope with ciency, and/or the attractiveness to mating partners (Pigliucci environmental variation. Long-term studies of the distribution 2005; Réale and Dingemanse 2010; Sih et al. 2012). This of behaviors in the wild are thus a necessary first step towards variation can occur in the form of consistent among-individ- understanding the mechanisms underlying the non-random dis- ual differences in behavior, referred to as animal personality, tribution of behavioral traits (Archard and Braithwaite 2010). which has been documented in many animal taxa (Bell 2005; Amphibians, and in particular Neotropical poison frogs Dochtermann and Jenkins 2007; Tremmel and Müller 2013; (Dendrobatidae, sensu AmphibiaWeb 2022), are great models Zidar et al. 2017; Goursot et al. 2019). Animal personality, to study behavioral variation across their environment. Many however, does not preclude the existence of individual plas- species show territoriality or site fidelity that facilitates repeated ticity (Dingemanse et al. 2010), and we need to take into measurements in wild individuals (Wells 2007; Kelleher et al. consideration both sources of variation when thinking about 2018), which can then be linked to local environmental parame- adaptive significance of behavioral variation in animals. ters. Many species also exhibit elaborate courtship behavior, ter- Five main personality traits are generally characterized in restrial oviposition, and obligatory tadpole transport of hatched animals along the following axes: active/passive, aggressive/ larvae to aquatic sites (e.g., Crump 1972; Roithmair 1994; Pröhl docile, bold/shy, exploratory/stationary, and sociable/non- 2005; Pašukonis et al. 2013; Rojas and Pašukonis 2019; Yang sociable (Réale et al. 2007). Personality traits can be seen as et al. 2019; Souza et al. 2021), offering ideal prerequisites for latent variables that affect multiple quantifiable behaviors of an within- and between-individual variation in behavior to arise. In organism in certain contexts (Pigliucci 2003; Araya-Ajoy and the present study, we aim to quantify how male-male aggression, Dingemanse 2014). For instance, aggressiveness can be seen exploratory, and anti-predator behaviors are expressed as func- as an unobservable (i.e., latent) variable that affects several tional units that can be described as latent variables reflecting observable behaviors during agonistic encounters, which can three personality axes (referred to by one dimension of the axis be assessed and quantified in an experimental context (e.g., as “aggression,” “exploration,” and “boldness”). We also aim speed of territorial reaction, number of attacks). Situations at identifying how behaviors relate to the individual’s natural where personality traits are non-randomly distributed across and social environment using a free-ranging population of the space are typically referred to as “phenotype by environment Neotropical poison frog Allobates femoralis (Dendrobatidae). correlation,” where environment refers to both the natural and Based on the natural history of A. femoralis, we make mul- social surroundings of the individual (Conover and Schultz tiple predictions. In males, we expect to find consistent among- 1995; Dingemanse and Araya-Ajoy 2015). For example, individual differences in boldness and aggression. During the anemones (Condylactis gigantea) living in areas with higher reproductive season, males produce loud advertisement calls seagrass were found to be shyer than those living in more open to warn intruders and to attract females (Ringler et al. 2011; areas (Hensley et al. 2012) and western bluebirds (Sialia mexi- Chaloupka et al. 2022). Differences in personality traits across cana) modify their aggression to match the level of aggressive males might be related to differences in the trade-off they face behavior of their mate (Duckworth and Kruuk 2009). between calling to secure mating and exposure to predators. Since personality traits can vary simultaneously within and As females do not display territoriality but merely perch fidel - among individuals, associations between behavior and envi- ity (Fischer et al. 2020) and are never observed in any other ronment can originate from multiple processes, such as a non- aggressive interaction, they are generally considered non- random distribution of behaviors and/or phenotypic plasticity aggressive; we have no clear expectation if and how varying (Sprau and Dingemanse 2017). A non-random distribution degrees of boldness could be maintained in females. However, of behaviors might be a direct effect of certain phenotypes we expect to find consistent individual differences in explora - (caused by genes and permanent environmental effects) show - tion in both sexes, because males transport tadpoles to natural ing a preference for certain environments. Alternatively, selec- pools located up to 200 m away from their territory and females tive pressures induced by the heterogeneity of the environment commute to male territories within 20 m distance of their perch can also maintain or generate individual differences in behavior for mating (Ringler et al. 2012, 2013, 2018; Fischer et al. 2020). within a population (Dingemanse et al. 2004). For instance, Finally, we expect males’ personality traits to be non-randomly in territorial species, only highly aggressive individuals may 1 3 Behavioral Ecology and Sociobiology ( 2 0 2 2) 76:93 Page3of14 93 distributed across space because individual males face distinct to evoke territorial defense behavior in focal males (Fig. 1; challenges based on their social and/or natural environment. On Ursprung et al. 2009). During agonistic encounters, male A. the one hand, we expect more aggressive and bolder individuals femoralis display typical responses consisting of an orien- to occupy areas of low complexity (e.g., with sparse vegetation tation of their head/body, jumps towards the intruder, and and few ground structures) where they are easier to spot for sometimes direct attack (i.e., wrestling) (Hödl 1987; Nar- females, while more passive or shyer individuals should occupy ins et al. 2003). We expect that in the context of territorial areas of higher complexity with more places to hide. On the defense, the personality trait “aggression” affects the latency other hand, we expect bolder and more aggressive individuals until the first head-body orientation, the latency to the first to occupy territories in areas with a higher population density jump, the probability to jump during moments when the than those occupied by more passive or shyer individuals. intruder does not call, and the speed to approach the intruder (Chaloupka et al. 2022). Aggressive territorial behavior was only assessed in males, as females do not defend territories. Methods We used a music player with integrated loudspeaker (MUVO 2c, Creative, Singapore), centered on top of a black Study population and area PVC disc (radius = 15 cm), 2 m from and facing the calling focal male (Fig. 1). We established the 2 m distance between We conducted our study in a population of A. femoralis on a the focal frog and the loudspeaker with a laser rangefinder lowland rainforest river island of approximately 5 ha. The island (DLE 50, Bosch, Stuttgart, Germany). The experimenter is situated in the “Les Nouragues” nature reserve in French remained a further 1 m behind the loudspeaker. We gave a Guiana (4°02′ N, 52°41′W; Bongers et  al.  2001), near the 30-s acclimation period to the frog, which was enough for “Saut Pararé” field camp of the CNRS Nouragues Ecological the frog to return to normal behavior and calling, before Research Station. The population was introduced in 2012 and randomly presenting one of seven synthetic calls. The calls has reached a stable size of approximately 150 adult individu- varied in their inter-note and inter-call intervals to avoid als (see Ringler et al. 2014). The experiments took place from habituation over the course of the experiment. Each syn- February to April of 2019, coinciding with the breeding season. thetic call featured the spectral and temporal parameters of a We surveyed the population every day during its most active nearby free-ranging population of A. femoralis (Narins et al. hours from 0900 to 1800 h, aiming to sample all adult males 2003; Gasser et al. 2009). Each of the playbacks lasted for and females on the island in the course of the study period. 5 min and was presented from the original WAV-files (16-bit, We caught frogs using a transparent plastic bag to mini- 44.1 kHz) using the same volume settings across all trials. mize stress and direct contact, thereby limiting the potential Using a digital voice recorder (ICD-PX333, Sony, Tokyo, influence of handling on behaviors. We identified all frogs via Japan), we commented on the behavior of the focal male, not- digital pictures of their distinct ventral patterns and with the ing its first head-body orientation, its jumps, and its arrival help of the pattern matching software Wild-ID (Bolger et al. at the speaker (i.e., touching the disc with at least one part of 2012), and sexed them by the presence (males) or absence his body). The trial ended when the male entered the 15-cm (females) of vocal sacs. Ventral patterns clearly differ among perimeter or when the playback stopped after 5 min. Trials individuals and are consistent across their adult lifespan, pro- where a focal male began to call were stopped and excluded viding a reliable method of identification. Information on the age of individuals was available from a concurrent long- term monitoring project of the island’s population since its origins in 2012. We recorded the locations of all frogs on a detailed digital map (Ringler et al. 2016) using the mobile GIS software ArcPad 10 (ESRI, Redlands, CA, USA) on rugged Win10 tablets (CAT T20, Bullitt Group, Reading, UK), and further handled the data in ArcGIS 10.6 (ESRI). We determined body size of all adults (snout urostyle length; SUL) from dorsal photographs in front of a reference scale using the software Image J 1.52a (Rasband 1997–2021). Quantification of aggressive behaviors Aggression is often measured via “an individual’s agonistic reaction towards a conspecific” (Réale et al. 2007). To assess Fig. 1 Picture of the territorial defense trial, presenting the speaker individual levels of aggression, we used acoustic playbacks and a male on top of the perimeter 1 3 Behavioral Ecology and Sociobiology ( 2 0 2 2) 76:93 93 Page4of14 from further analysis, as this was most likely the result of previous amphibian personality studies (Kelleher et  al. the speaker being positioned outside the defended area of 2018). Although counterintuitive at first, the number of the focal male’s territory (cf. Ringler et al. 2011). Vegetation jumps and the area covered are both important aspects of density may cause the sound to be reflected and/or distorted, a frog’s exploratory tendency. Indeed, two individuals per- so we measured the received sound pressure level (“SPL” forming an equal number of jumps can visit more or fewer in dB) of the signal after each trial using a sound pressure distinct areas of the box depending on their respective aver- meter (SL-100, Voltcraft, Hirschau, Germany). In all cases, age jump lengths. Individuals were tested at the same loca- the received SPL was above the threshold of 56 dB to elicit tion they were found. Males were tested immediately after a behavioral response (Hödl 1983). The minimum duration the territorial defense test, and females were caught at their between two consecutive tests with the same individual was encounter location and put straight inside the NET (Fig. 2). 24 h (on average tests were 12.11 ± 7.92 SD days apart). Although this method prevents us from comparing levels of Allobates femoralis typically lives for 1.5  years (Ringler exploratory tendency and boldness between sexes, it was et al. 2009; Ursprung et al. 2011a, b) and males only display chosen to keep handling as minimal as possible on males. aggressive reactions to territory intruders during the repro- The NET setup consisted of a cooler box (50 × 25 × 29 cm, ductive season. In our study, we aimed to test individuals hereafter “Novel Environment”), with a 10-cm PVC tube with similar among- and within-individual inter-test intervals attached on one side of the box (hereafter “shelter”). An to cover several weeks of the reproductive season. Therefore, opaque sliding door separated the Novel Environment from an average interval of 12 days between tests represents an the shelter. In the lid of the box, we installed a wide-angle effective compromise across the available time and life span video camera (Hero Black 5, GoPro, San Mateo, CA, USA) of the species. and two elongated, battery powered LED lights (LUMIstixx, We blinded audio file names prior to analysis and coding to avoid observer bias. From the dictaphone recordings of the territorial defense trials, we extracted the latency until the first head-body orientation (in s), the latency until the first jump (in s), the speed to reach the speaker (in cm/s), and if the male jumped (1) or not (0) during inter-bout-intervals (hereafter “inter-bout jumps,” cf. Ursprung et al. 2009) using the software VLC (VideoLAN 2021). Males that did not reach the speaker were coded a censored speed of 0 cm/s. Following the same reasoning, males that did not perform a head-body orientation or a jump were given a censored latency of 300 s (corresponding to the total duration of the experiment). In total, we conducted 163 valid territorial defense trials with 51 males (mean ± SD = 3.20 ± 1.31 rep- etitions per individual). Quantification of exploratory tendency and boldness To assess levels of exploratory tendency and boldness in male and female A. femoralis, we used a Novel Environment Test (NET), an approach that has been commonly used in personality studies (Carter et al. 2013; Kelleher et al. 2018). “Boldness” (or the corollary “shyness”) is defined as “an individual’s reaction to any risky situation,” and “explora- Fig. 2 Schematic (a) and picture (b) of the Novel Environment Test. tion” as “an individual’s reaction to a new situation” (Réale The cooler box measured 50 × 25 × 29  cm, with a 10  cm PVC tube et  al. 2007). Therefore, we expected the personality trait attached on one side of the box. A sliding door separated the shelter “boldness” to affect the latency to go out of a dark shel - from the box. Two LED tubes and a Hero Black 5 Go pro camera ter into a bright (novel) environment and the probability were attached to the lid of the box. Three solid PVC tubes (10  cm height, 5 cm diameter) were placed inside the box as visual obstacles to enter the novel environment, and the personality trait at randomized positions every day. A grid was drawn on the floor “exploration” to affect the distance travelled, the number of the cooler box to help randomize the position of the obstacles. A of jumps performed, and the area covered in the novel envi- mesh net was placed in the cooler box at 20 cm height to prevent the ronment. Those behaviors were chosen in accordance with frog from jumping on a wall outside of the camera range 1 3 Behavioral Ecology and Sociobiology ( 2 0 2 2) 76:93 Page5of14 93 Osram/Ledvance, Garching, Germany) for homogeneous Territory size, habitat complexity, and number illumination. We set the camera to “superview” mode to of neighbors ensure full visibility of the Novel Environment. We installed a mesh net in the Novel Environment at 20 cm height to To find out the distribution patterns of behaviors with respect prevent the frog from jumping on a wall outside of the cam- to a male’s natural and social environment, we determined era range. We used three solid PVC tubes (10 cm height, the territory size and the local complexity of the habitat, as 5 cm diameter) as visual obstacles, to motivate the frogs well as the number of female and male neighbors for each to explore the entire Novel Environment. The positions of male. This was only studied in males, as in A. femoralis the obstacles were changed daily and their positions were males acquire and defend non-overlapping territories, while determined using a random number generator. female display only site fidelity (Ringler et al. 2009, 2011 At the beginning of each trial, we placed all individuals 2012; Fischer et al. 2020). We used Dirichlet tessellation in in the shelter for 10 min to allow them to acclimatize. The ArcGIS to approximate male territories as Voronoi polygons shelter remained accessible throughout the trials to encour- (Voronoi 1908) on a day-to-day basis (for more information, age natural behaviors within the Novel Environment (Carter see Supplementary Materials). For the daily territory estima- et al. 2013; Kelleher et al. 2018), as the individual was free tion, we applied a roving-window approach, using all capture to remain inside or return to the shelter at any time. After points of the last 5 days a male was seen, including the focal this acclimation period, we switched on the lights and the day, but excluding all points where a male was likely found camera, closed the lid of the Novel Environment, opened outside its territory (i.e., all capture points linked to tadpole the sliding door between shelter and Novel Environment, transport). To ensure that the sizes of territories located at and filmed for 15 min. The minimum duration between two the population periphery were not over-estimated (Ringler consecutive tests with the same individual was 24 h (on aver- et al. 2009), we included the vertices of the island outline age tests were 11.36 ± 8.43 SD days apart). into the Dirichlet tessellation procedure, to establish a buffer We blinded video file names prior to analysis and cod - from the edge of the island until halfway to the outermost ing to avoid observer bias. To analyze the video recordings capture points of peripheral males. To ensure that estimates obtained during the NET, we used the software TRACKER at the beginning and end of the season were not impacted by (Brown 2019) to correct for distances that were distorted individuals not being captured, we used the POPAN formula by the camera wide-angle lens. Using the coding software (Schwarz and Arnason 1996) in program MARK (White BORIS (Friard et al. 2016), we assessed the latency from the and Burnham 1999; White 2020) to correct our population opening of the sliding door until frogs left the shelter (in s) estimates at these points in the breeding season (for more and the number of jumps performed inside the Novel Envi- detailed information, see Supplementary Materials). ronment. For individuals who stayed inside the shelter dur- After conducting the Dirichlet tessellation, we dissolved ing the entire experiment, the time spent in the shelter was the resulting Voronoi polygons based on male identity to censored to a value of 900 s (corresponding to the total dura- obtain “Voronoi territories” for each male and then calcu- tion of the experiment). We also coded the decision to enter lated territory sizes in ArcGIS. We also counted the direct the Novel Environment (1) or not (0) as a binomial response. male neighbors for each focal male from the daily Voronoi Using the automated tracking software TOXTRAC (Rod- territories. To obtain the number of female neighbors, we riguez et al. 2018), we measured the distance travelled (in calculated female centroid points (mean center) across the pixels) as well as the area visited inside the Novel Environ- entire season. Based on the typical distances females com- ment. For the latter variable, the software divided the floor mute for mating (Ringler et al. 2012; Fischer et al. 2020), we of the Novel Environment into a 9 × 8 grid and automatically constructed 20-m buffer circles and counted the number of counted the number of distinct squares a frog visited during contained centroid points of male Voronoi territories. a trial. As individuals varied in the time they spent inside To determine the complexity of the habitat for each indi- the Novel Environment, we standardized the data collected vidual male, we took four photographs (camera in automatic by only considering movements performed during the first (P) mode, focal length 50 mm, no flash, jpg images) from 2 min after leaving the shelter. On average, individuals spent each of the cardinal directions (0°, 90°, 180°, and 270° 624.21 s inside the Novel Environment (± 303.13 SD). This towards magnetic north) of a 100 cm × 100 cm red fabric. timeframe allowed for meaningful levels of movement to The fabric was placed at 3.5 m from the centroid point of occur but maximized the available data. Using this criterion, each male’s territory at forest floor level as this represents 21 out of 259 tests had to be excluded from the exploration half the radius of the typical male territory (cf. Ringler et al. analysis, leaving a total of 238 valid NET trials with 52 2011). We positioned the camera 20 cm above the forest males and 35 females (mean ± SD = 2.74 ± 1.33 repetitions floor, corresponding to the perch location of a male dur - per individual). Only very few individuals did not leave the ing territorial advertisement calling. We calculated habitat shelter in all repetitions. complexity as the average percentage of red fabric that was 1 3 Behavioral Ecology and Sociobiology ( 2 0 2 2) 76:93 93 Page6of14 covered by vegetation on the four pictures, with increasing time spent in the shelter). In addition, Bayesian generalized coverage of the fabric representing more complex habitat. linear mixed models were used to confirm that there were For that, we cropped each image in Paint.net v4.1.6 (Brew- no impacts from the fact that SEM analysis combines within ster 2019) to show only the lower half of the red fabric as and between individual effects (for details see Supplemen - this would represent the most visually relevant part of the tary Material). habitat for the frogs. Finally, we counted the number of pixels of visible fabric to calculate the percentage of fabric covered by vegetation. To facilitate this process, we aimed at Personality traits increasing the differences between fabric and vegetation by setting the image hue to 180 and decreasing the luminance To investigate if A. femoralis exhibits personality (i.e., and contrast to − 20. If males changed their territory location between-individual variation in behaviors), we assessed the during the study period, analysis of habitat complexity was repeatability (R) of all measured behaviors using the “rpt” performed in the territory in which most experimental trials function in the rptR package (Stoffel et al. 2017). Latency had been conducted. until the first head-body orientation and until the first jump had to be log transformed to achieve normal distribution. Statistical analysis We used the function “transformTukey” to apply a constant transformation on the speed to reach the speaker, the dis- We conducted all statistical analyses in R v3.6.0 (R Core tance travelled, and the time spent in the shelter. Repeat- Team 2020) using the integrated development environment ability was estimated based on the models applied to the RStudio v1.3.1093 (RStudio Team 2019). transformed data. We estimated repeatability from models fitted with a Gaussian error distribution for the latency until the first head-body orientation and until the first jump, the Structure of behaviors speed to reach the speaker, the time spent in the shelter, and the distance travelled. We estimated repeatability from mod- To determine how the measured behaviors are structured els fitted with a Poisson error distribution for the number of into functional units (i.e., aggression, exploration and bold- jumps and the number of areas, and from models fitted with ness), we investigated the phenotypic covariance structure a binary distribution for the inter-bout jumps and the prob- among different measurements during the behavioral tests. ability to enter the box. For all models, ID was included as To infer a latent variable describing aggression from behav- a random effect. iors measured during the territorial defense test (e.g., speed to reach the speaker, latency until the first head-body orien - tation, latency until the first jump, inter-bout jumps), using Distribution of behaviors across the natural the SEM package, we applied structural equation model- and social environment ling to the phenotypic covariance matrix derived from the means of each behavior for each individual, in order to We used a bivariate approach to study how behaviors meas- avoid pseudo-replication (Fox et al. 2020). Models were ured during the territorial defense trial and the NET corre- compared, to determine the most parsimonious model (for late with variation in the habitat complexity, territory size, details see Supplementary Fig. 1), based on differences in and number of male and female neighbors at the among- and Akaike’s information criterion (AIC) values, with small val- within-individuals level using Bayesian generalized linear ues indicating higher parsimony and a ΔAIC ≥ 2 indicating mixed models (Hadfield 2010). In the analyses, the latency significant differences (Burnham and Anderson 2002). We until the first jump, distance travelled, and time spent in expected the best model to have a latent variable explaining the shelter were chosen as they best represented the latent the covariance among the four measurements. We applied variables of aggression, exploration, and boldness, respec- the same technique to determine whether two latent vari- tively (see the “Structure of behaviors” section and Fig. 3). ables describing boldness and exploration can be inferred Although the probability to go inside the box was slightly from the behaviors measured during the NET (i.e., distance more correlated to boldness than “time spent in shelter,” we travelled, number of jumps, number of areas, time spent decided to use the latter variable to avoid using a binomial in the shelter, and probability to enter the box; for details variable in the model. Models were built with the trans- see, Supplementary Fig. 2). We expected the best model to formed data (see the “Personality traits” section). have two correlated latent variables, one that would explain To investigate the among-individual covariance between explorative behaviors (e.g., distance travelled, number of behaviors and environment variables, we divided each of the jumps, number of areas) and one that would explain bold- environmental variables by their mean value and added all of ness related behaviors (e.g., probability to enter the box, them as response variables (see Houslay and Wilson 2017). 1 3 Behavioral Ecology and Sociobiology ( 2 0 2 2) 76:93 Page7of14 93 Table 1 Repeatability (R) and confidence intervals (CI) of the behav - iors measured during the territorial defense test and the Novel Envi- ronment Test (NET) Test Variable R 95% CI Territorial Latency until head-body orienta- 0.17 [0.02; 0.34] defense tion test Latency until first jump 0.24 [0.07; 0.40] Speed to reach the perimeter 0.37 [0.19; 0.53] Inter-bout jumps 0.20 [0.001; 0.37] NET Probability to go in the box 0.44 [0.13; 0.89] Time spent in the shelter 0.30 [0.15; 0.44] Distance travelled in the box 0.36 [0.21; 0.49] Number of jumps 0.45 [0.27; 0.62] Number of areas visited 0.48 [0.30; 0.62] with univariate linear mixed models corrected for multiple comparisons, which led to the same biological conclusions. Results Personality traits Among the measures taken from the territorial defense test, all were considered repeatable and ranged in repeatability from 0.17 to 0.37 (Table 1; Fig. 4a). The best SEM model Fig. 3 Path diagrams of the best structure equation models (SEMs) supports a latent variable explaining the covariance of the (based on difference in Akaike’s information criterion (AIC) values) explaining the covariance structure among four behaviors assessed four behavioral responses measured in a territorial defense during an aggressivity test (a), and five behaviors assessed during a test (Fig. 3a; Supplementary Fig. 1b). Together, these results Novel Environment Test (b). “HBO” refers to the latency until the suggest that A. femoralis exhibits a personality trait “aggres- first head-body orientation. Squares represent the variances of the dif - siveness” encompassing the latency until the first head-body ferent behaviors explained by the SEM structure (R ). Numbers asso- ciated with arrows are standardized factor loadings which represent orientation and until the first jump towards a calling intruder, how behavioral responses are predicted to change based on changes the speed to reach the intruder, and the probability to jump to the latent variable. Number in brackets represent variances of during inter-bout. However, the residual variances were high residuals or error variances (e) associated to each behavior. All sim- (Fig. 3a; Supplementary Fig. 1b) and another model encom- ulated models can be found in the Supplementary material (Supple- mentary Fig. 1, 2) passing two latent variables had an AIC score close to the best model (∆AIC = 1.7; Supplementary Fig. 1 g). Similarly, we measured individual behavioral responses We included age (i.e., as a binomial trait: newly encoun- in a NET, where we found significant repeatabilities in all tered adults vs. recaptures from previous years) as a trait- behavioral measurements (Table 1; Fig. 4b, c). Based on specific fixed effects to control for effects of age on all our the comparison of the SEM, the best model supported the response variables. To calculate the variance due to differ - existence of a latent variable explaining the covariance pat- ences among individuals and the covariance between meas- terns of three behavioral measurements (distance travelled, ured behaviors, we fitted an unstructured covariance matrix number of jumps, and number of areas). It also supports that for the grouping variable ID. We then used the posterior another latent variable can be derived from the covariance of distributions to estimate the among- and within-individual two behavioral measurements (time spent inside the shelter correlations and covariances between each of the behavior and probability to enter the box). The two latent variables measured and the environment. We assumed statistical sig- are correlated (Fig. 3b; Supplementary Fig. 2 h). Together, nificance if the 95% credible intervals did not overlap 0 and these results suggest that A. femoralis males and females performed the same model verification as previously (see exhibit a personality trait “exploration” encompassing the the “Structure of behaviors” section). We also fitted these distance travelled, the number of jumps performed, and the relationships between behaviors and environment variables area covered in a new environment. The results also suggest 1 3 Behavioral Ecology and Sociobiology ( 2 0 2 2) 76:93 93 Page8of14 Fig. 4 Range of variation in the three behaviors that best repre- using a log (a) or constant (b and c) transformation. Boxes indicate sented the latent variables of aggression, exploration and boldness. the inter quartile range, with the central line depicting the median and The latency until the first jump ( a), the time spent in the shelter (b), the whiskers extending to 1.5*IQR. Dots represent the results of each and the distance travelled in the Novel Environment Test (c) are pre- trial. Males are ordered by their median (represented as a horizontal sented for individual males. All variables have been transformed bold line) the existence of the personality trait “boldness,” encompass- by vegetation in the measure of habitat complexity ing the latency to leave a safe place and enter a new environ- (SD = 9.31). Males occupied territories of 669.31 m on ment and the probability to enter a new environment. average (SD = 440.99). Males had on average 7 female For both the behaviors measured in the territorial defense and 5 male neighbors (mean of individual means) across test and in the NET, the phenotypic covariances in the SEM the season (absolute range females: 1–15, absolute range were mostly driven by the within-individual variances. The males: 1–10). Throughout the season, the average vari- results of the Bayesian models also show that the direc- ation in the number of both female and male neighbors tion of the among- and within-individual covariances was was 2 for individual territory holders (range females: 0–7; similar (Supplementary Tables 2, 3). Therefore, doing the range males: 0–6). SEM using only the among- or within-individual covari- There was no relation between the aggressive responses ance matrix would have resulted in a similar interpretation of a male during a territorial defense test and its social or of the results. natural environment. However, there was a significant rela - tion between the level of exploration and boldness of a male Behavior and the characteristics of an individual and the number of female neighbors at the within-individual and its environment level (Supplementary Table 4). This suggests that individu- als respond plastically to their social environment, increas- Next, we investigated the correlation between the behav- ing their level of exploration and boldness when the number iors measured during the territorial defense trial and of females around increases. The spatial setup of the territo- habitat complexity, territory size, and the number of ries on the day with the most individuals present at the same male and female neighbors. On average, habitats were time (07th of March 2019) is presented in Fig. 5. The aver- relatively complex with 82.77% of the fabric covered age response of each individual in terms of latency until the 1 3 Behavioral Ecology and Sociobiology ( 2 0 2 2) 76:93 Page9of14 93 Fig. 5 Maps showing the spatial distribution of individual performance of male frogs on the island in the behavio- ral essays. The maps use the Voronoi territories of 7 March 2019 when the most individuals were present at the same time on the island. The maps show the mean value, calculated over all respective trials of (a) the individuals’ latency to jump in the territorial defense trial, (b) the time spent in the shelter during the Novel Environment Test (NET), and (c) the distance travelled in the NET. All color ramps have 20 equal intervals across the full range of the respective value; darker colors represent shorter latency in (a), shorter time spent in the shelter in (b), and longer distances in (c). Black squares indicate the 14 artificial pools that were in place on the island since 2018; thin gray lines show 50 cm elevation isoclines; the blue area shows the river Arataye. The territories of two males that were not tested in the behavioral assays are shown with a hatched white pattern first jump, time spent in the shelter, and distance travelled to an individual’s natural and social environment in a wild, in the box is also represented. free-ranging, entire population of A. femoralis. Personality traits Discussion Our analysis showed that the repeatability of the variables In the present study, we investigated the structure of poten- measured in the territorial defense trial (0.17 to 0.37) was tial personality traits in A. femoralis and determined how in the lower range of the repeatability found in most stud- behaviors related to aggressiveness, exploration, and bold- ies (with overlapping confidence intervals; Bell et al. 2009; ness are structured into functional units (i.e., personality mean = 0.37, 95% confidence limits = 0.35–0.38) but con- traits). We also investigated how individual behaviors relate sistent with the average findings in other personality studies 1 3 Behavioral Ecology and Sociobiology ( 2 0 2 2) 76:93 93 Page10of14 on amphibians (Brodin et al. 2013; Maes et al. 2013; Gif- Future studies are needed to identify the link between call- ford et al. 2014; González-Bernal et al. 2014). Moreover, ing activity and individual reproductive success, in order to we found that a latent variable explains the covariance of investigate possible trade-offs in A. femoralis. the four behaviors measured during a territorial defense test Finally, we found that the distance travelled, the num- (i.e., the latency until the first head-body orientation and ber of jumps performed, and the area covered in a new until the first jump towards a calling intruder, the speed to environment were highly repeatable. Moreover, we found reach the intruder and the probability to jump during the that a latent variable explained some of the covariance inter-bout interval). Together, these results suggest that A. among the three behavioral measurements. Together, femoralis males exhibit a personality trait “aggressiveness.” these results suggest that A. femoralis males and females However, a great portion of the speed and the inter-bout exhibit a personality trait “exploration.” Exploration jumps measured during the territorial defense trial were behavior is especially relevant for dispersal and resource explained by exogenous factors. Together with the existence acquisition (Dingemanse et al. 2003; Gruber et al. 2017). of another model with similar support, these results sug- In A. femoralis, males rely on spatial memory to find gest that the behavioral measures capture different aspects water bodies (Pašukonis et al. 2016) and distribute their of aggressiveness. In this model (Supplementary Fig. 1 g), tadpoles across multiple sites to decrease risks of losing a latent variable explains the covariance of the latencies entire clutches due to desiccation or predation (Erich et al. until the first head-body orientation and until the first jump. 2015; Ringler et al. 2018). Therefore, being more explora- Another latent variable explains the covariance between the tive might enable males to find more sites for tadpole speed to reach the speaker and the inter-bout jumps. We deposition or a better territory to settle, while females interpret these two latent variables as reactivity and offen - put their explorative behaviors at use when looking for a siveness, respectively. Males of many amphibian species mate and, more rarely, for tadpole transport (Ringler et al. defend and fight over territories using acoustic and visual 2015). A highly explorative individual will, however, be displays (Hödl and Amezquita 2001; Toledo et al. 2007). For more conspicuous to predators or at risk of losing its ter- instance, in the Bornean rock-skipper frog (Staurois lato- ritory during periods of absence. palmatus), males perform foot-flagging and advertisement calls to defend their territories against conspecific intruders Distribution of behaviors across the natural (Preininger et al. 2009). Our results show that more aggres- and social environment sive males react fiercer towards conspecific territory intrud - ers. This is of particular ecological relevance, because terri- Contrary to our expectation, we did not find a relationship tory possession is the most important prerequisite for male between individual behaviors and their natural environment reproductive success in this species (Ursprung et al. 2011a). (i.e., habitat complexity and territory size). However, we However, high levels of aggression might also come with a cannot exclude that other habitat characteristics that we did cost, if more aggressive individuals engage more often in not measure, such as quantity and quality of the leaf lit- energetically costly and potentially harmful fights. ter, tree species, or canopy cover, are linked to individual We also found that the probability to enter a novel envi- behavior. Likewise, we did not find any relationship between ronment and the time it took to do so were highly repeatable, individual behavior and territory size, as more aggressive and a latent variable explained a relatively large part of the and bolder individuals settled in territories of varying sizes covariance among the two variables. Together, these results (Fig. 5a, b). This is in line with the previous finding that in suggest the existence of the personality trait “boldness” in A. A. femoralis, only the possession of a territory is important femoralis males and females. Boldness relates to the reaction for reproductive success, but not territory size (Ursprung of an individual to a predator, a novel object, or a conspecific et al. 2011a), and as a consequence, males cannot increase and is relevant in many contexts such as predator avoidance, their reproductive success by extending territorial space. As feeding, or mating (Réale et al. 2007). In A. femoralis, a gen- already suggested in a previous study, variation in territory erally high propensity to take risks might be reflected not sizes is probably strongly dependent on characteristics of only in the response to a predation threat, but also in how the natural environment (Ringler et al. 2017) rather than a prominently the advertisement call is presented. Thus, higher consequence of varying levels of aggression in A. femoralis risk taking could be reflected in males calling at higher rates, males. higher amplitude, over longer durations, from more exposed More explorative individuals were not necessarily locations, or moving/turning more during calling. In A. sub- located near the artificial pools on the island (Fig.  5c), folionidificans , a closely related species, calling activity has although water bodies are of critical importance for indi- been found to be positively related to reproductive success vidual fitness, given that they are obligate for tadpole (Souza et al. 2021). Being bold, however, can also be costly development after hatching. Previous studies have shown as it might lead to more frequent encounters with predators. that males distribute their tadpoles across multiple water 1 3 Behavioral Ecology and Sociobiology ( 2 0 2 2) 76:93 Page11of14 93 bodies located outside their territories (Ringler et  al. Conclusion 2013, 2018; Erich et  al. 2015; Beck et  al. 2017). This suggests that the ability of an individual to find water We studied the structure of personality traits and the dis- bodies once it has settled in a territory is critical, while tribution of behaviors across the environment in an entire the location of the territory is not. Finding water bodies free-ranging population of the poison frog A. femoralis by in A. femoralis is strongly related to olfactory sensing measuring several behaviors in situ in different contexts. (Serrano-Rojas and Pašukonis 2021) and spatial memory We found that A. femoralis indeed exhibits animal person- might help in revisiting such sites once they have been ality along the aggressiveness, exploration, and boldness encountered (Pašukonis et al. 2016; Beck et al. 2017). axes. Furthermore, we found non-random distribution of Taken together, our results suggest that A. femoralis behaviors across the animals’ social environment, which males establish their territories independent of large- may allow individuals to cope with their complex socio- scale resource distribution and the wider structure of the ecological environment. While amphibians have been largely habitat. And likewise, the characteristics of their natural overlooked in animal personality research, this study is one environment apparently are not associated with individual of the first comprehensive study of animal personality in differences in behavior. amphibian in the wild. By providing a detailed description We also did not find support for a relationship between of how behavioral measurements are structured in functional aggressiveness and the social environment. We initially units that allow individuals to cope with their socio-eco- expected more aggressive individuals to occupy territories logical environment, it broadens our understanding of the in high-density areas, where they would be more likely to functional role of behavior in frogs and offers a first step find mating partners but also face elevated male-male com - towards understanding the mechanisms that play a role in the petition. However, with this trade-off between female avail - emergence and maintenance of behavioral variation. ability and intra-sexual competition, males might have equal Supplementary Information The online version contains supplemen- reproductive outcomes regardless of their level of aggres- tary material available at https://doi. or g/10. 1007/ s00265- 022- 03202-9 . siveness and independent of the overall density, as long as they manage to establish a territory at all. Future studies Acknowledgements The authors thank B. Furdi for helping in pro- should investigate how the interplay between aggressiveness cessing the data and R. Fouchier for providing Fig. 1. The authors also and population density affects reproductive success. thank E. Haeler and C. Eliasch for their assistance during field work, and B. Szabo for comments on earlier versions of this manuscript. We Our results show that exploration- and boldness-related are grateful to the staff of CNRS Guyane for logistic support in French behaviors were positively linked to the number of females in Guiana. We thank the Nouragues research field station (managed by the vicinity to male territories. Since significant correlations CNRS) which benefits from “Investissement d’Avenir” grants man - were only found at the within-individual level, this suggests aged by Agence Nationale de la Recherche (AnaEE France ANR-11- INBS-0001; Labex CEBA ANR-10-LABX-25-01). This work has been that the link between exploration or boldness and the social done on the Natural Reserve of Nouragues. We thank the editors and environment is mainly driven by individual plasticity. This two anonymous reviewers for valuable comments on this and previous indicates that males, who mostly have stable territories and versions of the manuscript. move little, increase their overall levels of exploration and boldness when the number of females around their terri- Author contribution Conceptualization: MP, YA, ER; Methodology: MP, LB, SC, KD, CL, YA, PW, MR, ER; Formal Analysis: MP, LB, tory increases. There are currently no studies showing any MR; Investigation: MP, SC, KD, CL, YA, MR, ER; Resources: ER; direct mechanisms how males could assess the presence and Data curation: MP, MR, CL; Writing—original draft: MP; Writing— number of nearby females. We suggest that future studies review and editing: MP, YA, PW, MR, ER; Supervision: MP, MR, ER; should investigate secondary or indirect cues, such as dis- Project administration: MR, ER; Funding acquisition: ER. tribution or density of feeding sites, that might determine Funding Open access funding provided by University of Bern. This female density and that might allow males to identify areas work was funded through a Standalone Grant (FWF P-31518) from with more females. However, because the among-individual the Austrian Science Fund (FWF). MR was supported by an Erwin effects have broad credible intervals that are not centered Schrödinger Grant (FWF J-3868) and a Standalone Grant (FWF around zero, we cannot rule out the possibility of non-ran- P-33728) from the Austrian Science Fund (FWF). dom settlement. Data availability The datasets generated during and/or analyzed during Our study does not claim or identify any causal relation- the current study are available in the Open Science Framework reposi- ship between behaviors or personality traits and their natural tory: https:// doi. org/ 10. 17605/ OSF. IO/ AJV5X. and/or social environment. Still, the identification of such relationship in a natural free-ranging population of animals Declarations provides a first step towards understanding the mechanisms underlying the distribution of behaviors across space (cf. Ethics approval This study was approved by the scientific committee Archard and Braithwaite 2010). of the “Nouragues Ecological Research Station” and the ethics and 1 3 Behavioral Ecology and Sociobiology ( 2 0 2 2) 76:93 93 Page12of14 animal welfare committee of the University of Veterinary Medicine common frog (Rana temporaria). Behav Ecol Sociobiol 67:135– Vienna in accordance with Good Scientific Practice (GSP) guidelines 143. https:// doi. org/ 10. 1007/ s00265- 012- 1433-1 and national legislation. Sampling for this study was conducted in Burnham KP, Anderson DR (2002) Model selection and multimodel strict accordance with current French and EU law and followed the inference: a practical information-theoretic approach, 2nd edn. (ASAB 2020) guidelines on the treatment of study animals. Work- Springer, New York ing and tissue sampling permissions were provided by the CNRS Carter AJ, Feeney WE, Marshall HH, Cowlishaw G, Heinsohn R Guyane (“Centre National de la Recherche Scientifique Guyane”), by (2013) Animal personality: what are behavioural ecologists meas- the “Ministère de la transition écologique et solidaire” (permit num- uring? Biol Rev 88:465–475. https:// doi. org/ 10. 1111/ brv. 12007 ber: TREL2002508S/303), and the Secretariat of the Convention on Chaloupka S, Peignier M, Stückler S, Araya-Ajoy Y, Walsh P, Ringler Biological Diversity (APA declaration: TREL1734890A/34). M, Ringler E (2022) Repeatable territorial aggression in a Neo- tropical poison frog. Front Ecol Evol 10:881387. https:// doi. org/ Competing interests The authors declare no competing interests.10. 3389/ fevo. 2022. 881387 Conover DO, Schultz ET (1995) Phenotypic similarity and the evo- lutionary significance of countergradient variation. Trends Ecol Open Access This article is licensed under a Creative Commons Attri- Evol 10:248–252. https:// doi. or g/ 10. 1016/ S0169- 5347(00) bution 4.0 International License, which permits use, sharing, adapta- 89081-3 tion, distribution and reproduction in any medium or format, as long Crump ML (1972) Territoriality and mating behavior in Dendrobates as you give appropriate credit to the original author(s) and the source, granuliferus (Anura: Dendrobatidae). Herpetologica 28:195–198 provide a link to the Creative Commons licence, and indicate if changes Dingemanse NJ, Araya-Ajoy YG (2015) Interacting personalities: were made. The images or other third party material in this article are behavioural ecology meets quantitative genetics. Trends Ecol included in the article’s Creative Commons licence, unless indicated Evol 30:88–97 otherwise in a credit line to the material. If material is not included in Dingemanse NJ, Both C, van Noordwijk AJ, Rutten AL, Drent PJ the article’s Creative Commons licence and your intended use is not (2003) Natal dispersal and personalities in great tits (Parus permitted by statutory regulation or exceeds the permitted use, you will major). Proc R Soc Lond B 270:741–747. https://doi. or g/10. 1098/ need to obtain permission directly from the copyright holder. To view a rspb. 2002. 2300 copy of this licence, visit http://cr eativ ecommons. or g/licen ses/ b y/4.0/ . Dingemanse NJ, Both C, Drent PJ, Tinbergen JM (2004) Fitness conse- quences of avian personalities in a fluctuating environment. Proc R Soc Lond B 271:847–852 Dingemanse NJ, Kazem AJN, Réale D, Wright J (2010) Behavioural References reaction norms: animal personality meets individual plasticity. Trends Ecol Evol 25:81–89. https:// doi. org/ 10. 1016/j. tree. 2009. 07. 013 AmphibiaWeb (2022) AmphibiaWeb: information on amphibian biol- Dochtermann NA, Jenkins SH (2007) Behavioural syndromes in Mer- ogy and conservation riam’s kangaroo rats (Dipodomys merriami): a test of competing Araya-Ajoy YG, Dingemanse NJ (2014) Characterizing behavioural hypotheses. Proc R Soc Lond B 274:2343–2349. https:// doi. org/ ‘characters’: an evolutionary framework. Proc R Soc Lond B 10. 1098/ rspb. 2007. 0622 281:20132645. https:// doi. org/ 10. 1098/ rspb. 2013. 2645 Duckworth RA, Kruuk LEB (2009) Evolution of genetic integration Archard GA, Braithwaite VA (2010) The importance of wild popu- between dispersal and colonization ability in a bird. Evolution lations in studies of animal temperament. J Zool 281:149–160. 63:968–977. https:// doi. org/ 10. 1111/j. 1558- 5646. 2009. 00625.x https:// doi. org/ 10. 1111/j. 1469- 7998. 2010. 00714.x Erich M, Ringler M, Hödl W, Ringler E (2015) Brood-partitioning ASAB (2020) Guidelines for the treatment of animals in behavioural behaviour in unpredictable environments: hedging the bets? research and teaching. Anim Behav 159:I–XI. https:// doi. org/ 10. Behav Ecol Sociobiol 69:1011–1017. https:// doi. org/ 10. 1007/ 1016/j. anbeh av. 2019. 11. 002 s00265- 015- 1913-1 Beck KB, Loretto M-C, Ringler M, Hödl W, Pašukonis A (2017) Rely- Fischer M-T, Ringler M, Ringler E, Pašukonis A (2020) Reproductive ing on known or exploring for new? Movement patterns and repro- behavior drives female space use in a sedentary Neotropical frog. ductive resource use in a tadpole-transporting frog. PeerJ 5:e3745. PeerJ 8:e8920. https:// doi. org/ 10. 7717/ peerj. 8920 https:// doi. org/ 10. 7717/ peerj. 3745 Fox J, Nie Z, Byrnes J, Culbertson M, DebRoy S, Friendly M, Goodrich Bell AM (2005) Behavioural differences between individuals and two B, Jones RH, Kramer A, Monette G (2020) Package ‘sem’: Struc- populations of stickleback (Gasterosteus aculeatus). J Evol Biol tural Equation Models, https://CRAN .R-pr oject. or g/pac kag e=sem 18:464–473. https:// doi. org/ 10. 1111/j. 1420- 9101. 2004. 00817.x Friard O, Gamba M, Fitzjohn R (2016) BORIS: a free, versatile open- Bell AM, Hankison SJ, Laskowski KL (2009) The repeatability of source event-logging software for video/audio coding and live behaviour: a meta-analysis. Anim Behav 77:771–783. https://doi. observations. Methods Ecol Evol 7:1325–1330. https:// doi. org/ org/ 10. 1016/j. anbeh av. 2008. 12. 022 10. 1111/ 2041- 210X. 12584 Bolger DT, Morrison TA, Vance B, Lee D, Farid H (2012) A computer- Gasser H, Amézquita Torres A, Hödl W (2009) Who is calling? assisted system for photographic mark–recapture analysis. Meth- Intraspecific call variation in the aromobatid frog Allobates femo- ods Ecol Evol 3:813–822. https:// doi. org/ 10. 1111/j. 2041- 210X. ralis. Ethology 115:596–607. https://doi. or g/10. 1111/j. 1439- 0310. 2012. 00212.x 2009. 01639.x Bongers F, Charles-Dominique P, Forget P-M, Théry M (eds) (2001) Gifford ME, Clay TA, Careau V (2014) Individual (co)variation in Nouragues: dynamics and plant-animal interactions in a neo- standard metabolic rate, feeding rate, and exploratory behavior tropical rainforest. Monographiae biologicae, v. 80. Kluwer in wild-caught semiaquatic salamanders. Physiol Biochem Zool Academic Publishers, Dordrecht 87:384–396. https:// doi. org/ 10. 1086/ 675974 Brewster R (2019) paint.net, https:// www. getpa int. net/ González-Bernal E, Brown GP, Shine R (2014) Invasive cane toads: Brown D (2019) Tracker: Video analysis and modeling tool for phys- social facilitation depends upon an individual’s personality. PLoS ics education. Open Source Physics, Davidson. https:// physl ets. ONE 9:e102880. https:// doi. org/ 10. 1371/ journ al. pone. 01028 80 org/ track er/ Goursot C, Düpjan S, Kanitz E, Tuchscherer A, Puppe B, Leliveld Brodin T, Lind MI, Wiberg MK, Johansson F (2013) Personality LMC (2019) Assessing animal individuality: links between trait differences between mainland and island populations in the 1 3 Behavioral Ecology and Sociobiology ( 2 0 2 2) 76:93 Page13of14 93 personality and laterality in pigs. Curr Zool 65:541–551. https:// Réale D, Dingemanse NJ (2010) Personality and individual social doi. org/ 10. 1093/ cz/ zoy071 specialisation. In: Székely T, Moore AJ, Komdeur J (eds) Social Gruber J, Brown G, Whiting MJ, Shine R (2017) Is the behavioural behaviour: genes, ecology and evolution. Cambridge University divergence between range-core and range-edge populations of Press, Cambridge, pp 417–441 cane toads (Rhinella marina) due to evolutionary change or devel- Réale D, Reader SM, Sol D, McDougall PT, Dingemanse NJ (2007) opmental plasticity? R Soc Open Sci 4:170789. https://doi. or g/10. Integrating animal temperament within ecology and evolution. 1098/ rsos. 170789 Biol Rev 82:291–318. https:// doi. or g/ 10. 1111/j. 1469- 185X. Hadfield JD (2010) MCMC Methods for multi-response generalized 2007. 00010.x linear mixed models the MCMCglmmR package. J Stat Soft Ringler M, Ursprung E, Hödl W (2009) Site fidelity and patterns of 33:1–22 short- and long-term movement in the brilliant-thighed poison Hensley NM, Cook TC, Lang M, Petelle MB, Blumstein DT (2012) frog Allobates femoralis (Aromobatidae). Behav Ecol Sociobiol Personality and habitat segregation in giant sea anemones (Con- 63:1281–1293. https:// doi. org/ 10. 1007/ s00265- 009- 0793-7 dylactis gigantea). J Exp Mar Biol Ecol 426–427:1–4. https://doi. Ringler M, Ringler E, Magaña Mendoza D, Hödl W (2011) Intru- org/ 10. 1016/j. jembe. 2012. 05. 011 sion experiments to measure territory size: development of the Hödl W (1983) Phyllobates femoralis (Dendrobatidae): Rufverhalten method, tests through simulations, and application in the frog und akustische Orientierung der Männchen (Freilandaufnahmen). Allobates femoralis. PLoS ONE 6:e25844. https:// doi. org/ 10. Wissenschaftlicher Film, Bundesstaatliche Hauptstelle für Wis-1371/ journ al. pone. 00258 44 senschaftliche Kinematographie, Wien Ringler E, Ringler M, Jehle R, Hödl W (2012) The female perspec- Hödl W (1987) Dendrobates femoralis (Dendrobatidae): a handy fellow tive of mating in A femoralis, a territorial frog with paternal for frog bioacoustics. In: van Gelder J. J., Strijbosch H, Bergers care a spatial and genetic analysis. PLoS ONE 7:e40237. https:// P (eds) Proceedings of the 4th Ordinary General meeting of the doi. org/ 10. 1371/ journ al. pone. 00402 37 Societas Europaea Herpetologica, Nijmegen, Netherlands Ringler E, Pašukonis A, Hödl W, Ringler M (2013) Tadpole transport Hödl W, Amezquita A (2001) Visual signaling in anuran amphibians. logistics in a Neotropical poison frog: indications for strategic In: Ryan MJ (ed) Anuran communication. Smithsonian Institution planning and adaptive plasticity in anuran parental care. Front Press, Washington, DC, pp 121–141 Zool 10:67. https:// doi. org/ 10. 1186/ 1742- 9994- 10- 67 Houslay TM, Wilson AJ (2017) Avoiding the misuse of BLUP in Ringler E, Mangione R, Ringler M (2014) Where have all the tad- behavioural ecology. Behav Ecol 28:948–952. https:// doi. org/ 10. poles gone? Individual genetic tracking of amphibian larvae 1093/ beheco/ arx023 until adulthood. Mol Ecol Resour 15:737–746. https:// doi. org/ Kelleher SR, Silla AJ, Byrne PG (2018) Animal personality and behav-10. 1111/ 1755- 0998. 12345 ioral syndromes in amphibians: a review of the evidence, experi- Ringler E, Pašukonis A, Fitch WT, Huber L, Hödl W, Ringler M mental approaches, and implications for conservation. Behav Ecol (2015) Flexible compensation of uniparental care: female poi- Sociobiol 72:79. https:// doi. org/ 10. 1007/ s00265- 018- 2493-7 son frogs take over when males disappear. Behav Ecol 26:1219– Maes J, van Damme R, Matthysen E (2013) Individual and among- 1225. https:// doi. org/ 10. 1093/ beheco/ arv069 population variation in dispersal-related traits in Natterjack toads. Ringler M, Mangione R, Pašukonis A et al (2016) High-resolution Behav Ecol 24:521–531. https:// doi. org/ 10. 1093/ beheco/ ars193 forest mapping for behavioural studies in the Nature Reserve Narins PM, Hödl W, Grabul DS (2003) Bimodal signal requisite for ‘Les Nouragues’, French Guiana. J Maps 12:26–32. https://doi. agonistic behavior in a dart-poison frog, Epipedobates femoralis. org/ 10. 1080/ 17445 647. 2014. 972995 P Natl Acad Sci USA 100:577–580. https://d oi.or g/1 0.1 073/p nas. Ringler M, Szipl G, Hödl W, Khil L, Kofler B, Lonauer M, Provin 02371 65100 C, Ringler E (2017) Acoustic ranging in poison frogs—it is not Pašukonis A, Ringler M, Brandl HB, Mangione R, Ringler E, Hödl W about signal amplitude alone. Behav Ecol Sociobiol 71:114. (2013) The homing frog: high homing performance in a territorial https:// doi. org/ 10. 1007/ s00265- 017- 2340-2 dendrobatid frog Allobates femoralis (Dendrobatidae). Ethology Ringler E, Szipl G, Harrigan RJ, Bartl-Binder P, Mangione R, Ring- 119:762–768. https:// doi. org/ 10. 1111/ eth. 12116 ler M (2018) Hierarchical decision-making balances current and Pašukonis A, Trenkwalder K, Ringler M, Ringler E, Mangione R, future reproductive success. Mol Ecol 27:2289–2301. https:// Steininger J, Warrington I, Hödl W (2016) The significance of doi. org/ 10. 1111/ mec. 14583 spatial memory for water finding in a tadpole-transporting frog. Rodriguez A, Zhang H, Klaminder J, Brodin T, Andersson PL, Anim Behav 116:89–98. https:// doi. org/ 10. 1016/j. anbeh av. 2016. Andersson M (2018) ToxTrac: a fast and robust software for 02. 023 tracking organisms. Methods Ecol Evol 9:460–464. https://d oi. Pigliucci M (2003) Phenotypic integration: studying the ecology and org/ 10. 1111/ 2041- 210X. 12874 evolution of complex phenotypes. Ecol Lett 6:265–272. https:// RStudio Team (2019) RStudio: integrated development for R. doi. org/ 10. 1046/j. 1461- 0248. 2003. 00428.x RStudio, Inc., Boston, http:// www. rstud io. com Pigliucci M (2005) Evolution of phenotypic plasticity: where are we Roithmair ME (1994) Male territoriality and female mate selection going now? Trends Ecol Evol 20:481–486. https:// doi. org/ 10. in the dart-poison frog Epipedobates trivattatus (Dendrobatidae, 1016/j. tree. 2005. 06. 001 Anura). Copeia 1994:107–115. https:// doi. org/ 10. 2307/ 14466 76 Preininger D, Böckle M, Hödl W (2009) Communication in noisy Rojas B, Pašukonis A (2019) From habitat use to social behavior: nat- environments II: visual signaling behavior of male foot-flagging ural history of a voiceless poison frog. Dendrobates Tinctorius frogs Staurois latopalmatus. Herpetologica 65:166–173. https:// Peerj 7:e7648. https:// doi. org/ 10. 7717/ peerj. 7648 doi. org/ 10. 1655/ 08- 037R.1 Schwarz CJ, Arnason AN (1996) A general methodology for the analy- Pröhl H (2005) Territorial behavior in dendrobatid frogs. J Herpetol sis of capture-recapture experiments in open populations. Biom- 39:354–365. https:// doi. org/ 10. 1670/ 162- 04A.1 etrics 52:860–873. https:// doi. org/ 10. 2307/ 25330 48 Rasband WS (1997–2021) ImageJ. U. S. National Institutes of Serrano-Rojas SJ, Pašukonis A (2021) Tadpole-transporting frogs use Health, Bethesda, http:// imagej. nih. gov/ ij/ stagnant water odor to find pools in the rainforest. J Exp Biol. R Core Team (2020) R: a language and environment for statistical https:// doi. org/ 10. 1242/ jeb. 243122 computing. R Foundation for Statistical Computing, Vienna. Sih A, Cote J, Evans M, Fogarty S, Pruitt J (2012) Ecological implica- http:// www.r- proje ct. org tions of behavioural syndromes. Ecol Lett 15:278–289. https:// doi. org/ 10. 1111/j. 1461- 0248. 2011. 01731.x 1 3 Behavioral Ecology and Sociobiology ( 2 0 2 2) 76:93 93 Page14of14 Souza JRD, Kaefer IL, Lima AP (2021) Calling activity determines Ursprung E, Ringler M, Jehle R, Hödl W (2011b) Toe regeneration male mating success in a territorial frog with parental care. Ethol- in the neotropical frog Allobates femoralis. Herpetol J 21:83–86 ogy 127:359–365. https:// doi. org/ 10. 1111/ eth. 13135 VideoLAN (2021) VLC media player. VideoLAN, Paris. https://www . Sprau P, Dingemanse NJ (2017) An approach to distinguish between video lan. org/ vlc/ plasticity and non-random distributions of behavioral types along Voronoi G (1908) Nouvelles applications des paramètres continus à la urban gradients in a wild passerine bird. Front Ecol Evol 5:92. théorie des formes quadratiques. Premier mémoire. Sur quelques https:// doi. org/ 10. 3389/ fevo. 2017. 00092 propriétés des formes quadratiques positives parfaites. J Reine Stoffel MA, Nakagawa S, Schielzeth H (2017) rptR: Repeatability esti - Angew Math 1908:97–102. https:// doi. or g/ 10. 1515/ cr ll. 1908. mation and variance decomposition by generalized linear mixed-133. 97 effects models. Methods Ecol Evol 8:1639–1644. https:// doi. org/ Wells KD (2007) The ecology and behavior of amphibians. University 10. 1111/ 2041- 210X. 12797 of Chicago Press, Chicago Toledo LF, Araújo OGS, Guimarães LD, Lingnau R, Haddad CFB White GC (2020) MARK. Colorado State University, Fort Collins (2007) Visual and acoustic signaling in three species of Brazil-http:// phidot. org/ softw are/ mark/ index. html ian nocturnal tree frogs (Anura Hylidae). Phyllomedusa 6:61–68. White GC, Burnham KP (1999) Program MARK: survival estimation https:// doi. org/ 10. 11606/ issn. 2316- 9079. v6i1p 61- 68 from populations of marked animals. Bird Study 46:S120–S139. Tremmel M, Müller C (2013) Insect personality depends on environ-https:// doi. org/ 10. 1080/ 00063 65990 94772 39 mental conditions. Behav Ecol 24:386–392. https:// doi. org/ 10. Yang Y, Blomenkamp S, Dugas MB, Richards-Zawacki CL, Pröhl 1093/ beheco/ ars175 H (2019) Mate choice versus mate preference: inferences about Ursprung E, Ringler M, Hödl W (2009) Phonotactic approach pattern color-assortative mating differ between field and lab assays of in the neotropical frog Allobates femoralis: a spatial and temporal poison frog behavior. Am Nat 193:598–607. https:// doi. org/ 10. analysis. Behaviour 146:153–170. https:// doi. org/ 10. 1163/ 15685 1086/ 702249 3909X 410711 Zidar J, Balogh A, Favati A, Jensen P, Leimar O, Løvlie H (2017) A Ursprung E, Ringler M, Jehle R, Hödl W (2011a) Strong male/male comparison of animal personality and coping styles in the red competition allows for nonchoosy females: high levels of polygy- junglefowl. Anim Behav 130:209–220. https:// doi. org/ 10. 1016/j. nandry in a territorial frog with paternal care. Mol Ecol 20:1759–anbeh av. 2017. 06. 024 1771. https:// doi. org/ 10. 1111/j. 1365- 294X. 2011. 05056.x Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. 1 3

Journal

Behavioral Ecology and SociobiologySpringer Journals

Published: Jul 1, 2022

Keywords: Behavioral variation; Animal personality; Poison frogs; Non-random distribution; Environment

There are no references for this article.