Abstract Where organisms encounter novel conditions during range expansion, behavioral changes suited to the new habitat can enhance survival. Behavioral changes that mitigate predation risk are particularly important for the persistence of range-edge populations, especially where plastic responses outpace genetic adaptation. We use a climate-driven spatial mismatch between the arboreal mangrove tree crab (Aratus pisonii) and its primary mangrove habitat to evaluate differences in predator avoidance behavior between populations in range-center mangroves and adjacent range-edge salt marshes. We expected that differences in vegetation stature and diameter mediate changes in Aratus behavior. We combined crab and vegetation surveys with tethering experiments and in situ behavioral trials to determine habitat-specific predation risk and predator avoidance via evasion and autotomy. Tethering trials revealed that predation risk was always greater from aquatic sources than terrestrial sources and that aquatic risk was enhanced in marsh habitat. Vegetation structural form constrained Aratus predator avoidance during in situ behavioral assays: in mangroves, Aratus escaped upward into the canopy, but short-statured marsh grass restricted evasion to downward movement towards the higher risk aquatic environment. Given this restricted evasion route, Aratus in salt marshes were less likely to evade and showed more evidence of secondary escape via leg dropping. Shifting predator avoidance behavior away from a fleeing escape strategy may ameliorate the fitness costs of reduced escape opportunities for Aratus in novel marsh habitat along the range edge. Similar changes in behavior to match local habitat conditions could be integral to the persistence of many range-edge populations that encounter novel habitats. INTRODUCTION Biotic interactions between organisms are mediated by the structural characteristics of their surrounding habitat (Moermond 1979; Robinson and Holmes 1982; Belgrad and Griffen 2017). In particular, predator–prey interactions depend on how habitat structural features conceal, separate, or shelter prey from predators (Camp et al. 2012; Ceradini and Chalfoun 2017). Accordingly, differences between habitat types in substrate and refuge quality can produce shifts in species’ behavior. Behavioral plasticity or adaptation that addresses habitat-specific risks can enhance survival and can increase population persistence, especially in novel conditions (Agrawal 2001; Canestrelli et al. 2016). Responsiveness to novel conditions could be increasingly important for the persistence of range-edge populations that are encountering new habitats during climate change-induced range shifts (Hickling et al. 2006; Chen et al. 2011). Both habitat-forming foundation species and their inhabitants are shifting ranges with climate change (Cavanaugh et al. 2014; Riley et al. 2014). Where climate change affects community members differently (Lurgi et al. 2012), species interactions can become spatially and temporally mismatched (Schweiger et al. 2008; Yang and Rudolf 2010). When interactions become mismatched, expanding species that outpace foundational species often face novel habitats that present shifting selection pressures. Indeed, relative to range-center populations, edge populations of expanding species can differ in morphology, life-history traits, dispersal, and aggression (Hudina et al. 2014; Chuang and Peterson 2016). Where the range edge is characterized by novel habitat—such as a different foundation species—behavioral changes should be especially prevalent because plastic behavior changes outpace genetic adaptation (Robinson and Dukas 1999). However, it is unclear how habitat structure can affect predator avoidance behaviors of expanding species. In particular, we expect behavior to reflect habitat-specific risk and avoidance opportunities, such that individuals in novel range-edge habitat will shift to behaviors that enhance survival and, thus, the persistence of frontier populations (Pierce et al. 2017). We investigated vegetation-mediated shifts in predator avoidance behavior for a species exhibiting a mismatch with its primary habitat. With climate change, Aratus pisonii (hereafter: Aratus), an arboreal omnivorous mangrove tree crab, has out-paced the northward movement of mangroves along the Atlantic coast of Florida (Cavanaugh et al. 2014; Osland et al. 2016). Aratus populations are now established in temperate salt marshes more than 150 km north of the most northern mangrove (Riley et al. 2014). Mangroves and salt marshes are both coastal wetland environments, but the 2 habitats are formed by morphologically distinct plant foundation species. On the Florida Atlantic coast, salt marshes are dominated by low-stature herbaceous vegetation (<1 m tall) composed primarily of the marsh cordgrass Spartina alterniflora. In contrast, mangrove forests—dominated by the red mangrove Rhizophora mangle and the black mangrove Avicennia germinans—have significantly greater aboveground biomass (3–7 m tall) composed of woody trunks and branches, leafy canopies, and extensive aerial root systems (Friess et al. 2012; Simpson et al. 2017). We hypothesized that Aratus populations in salt marsh (range edge) and mangrove (range center) habitats differ in predator avoidance behaviors based on the distinct structural characteristics of these 2 qualitatively different habitats. Aratus are vulnerable to predation from both aquatic (e.g. fish, swimming crabs) and terrestrial sources (e.g. birds, mammals, terrestrial crabs) (Wilson 1989). Although Aratus are primarily arboreal, they will leap into the water to avoid perceived terrestrial or aerial threats, increasing their exposure to aquatic risk and creating a behavioral trade-off between minimizing terrestrial and aquatic predation (Yeager et al. 2016). We expected mangrove and marsh vegetation structure to mediate Aratus predator-avoidance behavior in 2 ways. First, as an arboreal species, crab size relative to vegetation substrate (i.e. leaf or stem) diameter should determine Aratus concealment; crabs with carapace widths greater than their associated substrate are less well concealed, with exposure increasing predation risk. Thus, the narrower leaf and stem substrates of marsh cordgrass should decrease crab concealment, leading to higher probability of evasion in marsh populations (Camp et al. 2012), as crabs perceive greater risk. A greater evasion response with lower concealment would indicate that Aratus perceive and respond to increased risk from aerial predators in the poor cover of salt marsh vegetation. Second, the stature of vegetation—especially the extent of emergent growth—should determine potential routes of evasion (i.e. fleeing) from attacking predators. Contrary to expectations based on concealment, if predation risk is greatest from aquatic predators (Yeager et al. 2016; Riley and Griffen 2017), then limited upward escape due to the short stature of marsh cordgrass could reduce overall evasion in marsh populations. With lower evasion, we expected evidence of higher predator encounter and secondary escape behavior via limb autotomy (i.e. leg dropping). Autotomy has higher fitness consequences than evasion due to the debilitating effects of missing limbs and energetic cost of limb regeneration; therefore, we expected higher autotomy only where vegetation structure prohibited the effective use of evasion as the primary escape behavior (Lindsay 2010; Lavalli and Spanier 2015). Thus, shifts in predator avoidance behavior should optimize survival in response to differences in vegetation structure. To test habitat-specific differences in predation risk and avoidance behavior, we 1) surveyed crab sizes and vegetation substrate diameters to characterize relative scaling, 2) measured the prevalence of predation-avoidance behaviors in a field behavior assay, and 3) performed a field tethering experiment to determine predation probability from aerial and aquatic sources in range-center mangrove and novel range-edge salt marsh habitat. Together, we use these studies to characterize differences in predator-avoidance behavior among Aratus populations in distinctive vegetation types to examine how structural habitat attributes can drive changes in behavior as species ranges expand into novel environments. METHODS Study system To examine differences in Aratus predator avoidance behavior in mangrove and salt marsh habitats, we first identified representative salt marsh and mangrove sites with distinct vegetation types (Figure 1). We chose 2 pure mangrove sites located at St. Lucie Inlet, FL (27.16°N, 80.17°W) and Sebastian Inlet, FL (27.85°N, 80.45°W) that were dominated by the red mangrove Rhizophora mangle, with some black mangrove Avicennia germinans interspersed. To represent pure salt marsh, we chose 2 sites, one in Pancho Creek by St. Augustine Inlet (29.94°N, 81.32°W) and another in Nassau Sound, FL (30.51°N, 81.46°W), that were both dominated by marsh cordgrass Spartina alterniflora. To maintain relative uniformity in hydrological conditions, all sites were located within 6 km of an inlet, and we conducted all surveys and experiments on creek banks with monospecific vegetation bordering deep channels. We characterized habitat and Aratus attributes and performed behavioral assays at all 4 sites and performed the relative predation risk tethering experiment at Pancho Creek and St. Lucie Inlet sites. Figure 1 View largeDownload slide Study sites and species ranges. Aratus behavior relative to vegetation attributes was studied at 4 sites (black circles): 2 at the range center dominated by mangroves and 2 at the range edge dominated by marsh. Mangrove distribution (dark gray) after Osland et al (2013). Aratus Northern range limit (open circle) after Riley et al (2014). Figure 1 View largeDownload slide Study sites and species ranges. Aratus behavior relative to vegetation attributes was studied at 4 sites (black circles): 2 at the range center dominated by mangroves and 2 at the range edge dominated by marsh. Mangrove distribution (dark gray) after Osland et al (2013). Aratus Northern range limit (open circle) after Riley et al (2014). Aratus and habitat attributes To characterize Aratus populations relative to their primary and novel habitats, we conducted surveys of Aratus and their associated vegetation substrates in mangrove and salt marsh-dominated wetlands during September 2015. We surveyed Aratus at high tide, when they are most accessible due to submerged benthos. We hand-captured ~100 Aratus at each site, except for Sebastian Inlet, where we were only able to capture 50 individuals. For each crab, we recorded carapace width (CW; the widest point behind the eyestalks), number of missing legs, and the diameter of their vegetation substrate. Number of missing legs was recorded as a metric of close predator encounter and autotomy (Lavalli and Spanier 2015). Predator avoidance behavior To evaluate differences in behavior across habitats, we measured the prevalence of predation-avoidance strategies. First, to assess evasive behavior, we performed a behavioral assay on an additional 75–85 crabs per site during high tide when the benthos was inaccessible to crabs. We observed and carefully approached each crab and then gently, but vigorously, tapped the substrate immediately adjacent to their perch (not above or below) 3 times with an extended ruler from 50 to 150 cm away to evaluate crab response. The tapping visibly alarmed the crabs but was gentle enough to prevent physically dislodging them, stimulating representative and measurable variation in threat response behavior. We classified the absence (0) or presence (1) of an evasion response: crabs that did not evade often sheltered in place in a hunched posture, while evading crabs dropped off vegetation or rapidly ran along their substrate. When evasion occurred, we also recorded the direction of evasion as movement downward toward water or upward away from water. Downward evasion behaviors included dropping rapidly downward along the substrate and leaping off the substrate into the water; upward evasion was characterized by fleeing laterally or away from the water along the substrate. Given the variation in evasion behavior, we were unable to consistently capture individual crabs following each evasion assay and thus did not measure sex, carapace width, or limb loss for crabs in the behavioral assay. We did record whether crabs were categorically larger or smaller than their substrates. Evasion behavior among Aratus in each habitat was then compared with the prevalence of missing limbs—evidence of autotomy, a secondary predation-avoidance strategy—among individuals measured in that habitat during the initial survey (see above). Predation probability and relative risk To assess predation risk in each habitat type, we performed Aratus tethering experiments in July and August of 2016. Aerial predation was enumerated at low tide (i.e. aerial exposure), when Aratus are exposed to terrestrial and aerial predators including birds, terrestrial crabs, and small mammals. Aquatic predation was enumerated at high tide (i.e. aquatic exposure), when Aratus are exposed to aquatic predators, including fish and swimming crabs. The experiment was conducted at the Pancho Creek salt marsh site in July, and the St. Lucie Inlet mangrove site in August within a week of a new or full moon to encompass maximum variation in exposure (i.e. spring tides). At each site, we haphazardly captured 100 Aratus. In the lab, we affixed a 25-cm tether of 10-lb test fishing line to each crab by a noose and cyanoacrylate glue (Superglue). We assigned half of the tethered crabs randomly to each exposure treatment (n = ~50 crabs per treatment per habitat type). We held tethered crabs in damp coolers and deployed them within 24 h of initial capture. For each treatment, we deployed ~50 tethered crabs by attaching tethers directly to vegetation. Crabs were set >1 m apart along continuous stretches of vegetation. Crabs were deployed for the 3 h surrounding a focal tide (e.g. for aerial exposure, crabs were deployed 1.5 h before low tide and retrieved 1.5 h after low tide). In the aerial predation treatment, we placed tethered crabs on the base of vegetation, fully extended the tether, and attached it to the vegetation (~25 cm above the sediment) to ensure that crabs were exclusively subject to aerial predation. In the aquatic predation treatment, we attached tethered crabs at or just below the water line at the time of deployment (1.5 h before high tide) to ensure that crabs were constrained to 25 cm above and below the water line. This placement ensured that crabs in the aquatic exposure treatment were fully submerged for the same amount of time (~0.5 h at peak high tide) at both sites and were predominantly subject to aquatic predation. Estimates of aquatic predation are conservative, because individuals were not submerged throughout the trial; however, the design should provide realistic estimates of aquatic predation because Aratus actively avoid prolonged submersion but were constrained to the water surface, where they remain at risk from actively foraging fish (Yeager et al. 2016). Drowning is unlikely to have affected the aquatic exposure treatment, as crabs remained alive and active for at least 3 h when fully submersed in water during controlled lab trials (n = 10). Upon tether retrieval, we scored each crab as present or absent. Tether retention in the lab was 100% over more than 12 h (n = 10), so we considered missing crabs to be reliable markers of predation (Johnston and Lipcius 2012). Autotomy was not evaluated on tethered individuals, because handling during tether deployment and retrieval is likely to inflate leg drop. Data analysis We conducted all analyses in R version 3.3.2 (R Core Team 2015). To analyze Aratus condition and habitat use, we used linear models to predict crab carapace width, vegetation substrate diameter, and individual carapace width:substrate diameter ratios as a function of overall vegetation type (mangrove vs. salt marsh) combined across sites. For the carapace width:substrate diameter ratio, values <1 represented crabs narrower than their substrates (i.e. more concealed) and values >1 represented crabs broader than their substrates (i.e. less concealed). Data were log-transformed where appropriate to meet model assumptions of normality and homoscedascity. To analyze differences in evasion behavior between habitats, we first tested the probability of an evasive response (response versus no response to stimulus) across individual crabs with a generalized linear model fit with a binomial distribution, using habitat type (mangrove, salt marsh) as a fixed effect. A random effect of site did not account for significant variability or change model inference and thus is omitted. To more specifically test evasion relative to concealment, we also analyzed evasion probability using a binomial GLM with crab size relative to substrate (qualitatively recorded as broader or narrower than their substrates) as a fixed effect. An interaction of size with habitat type was also tested. For crabs that did respond to stimulus, we analyzed the direction of evasion (down vs. up) in each habitat with a repeated G-Test (chi-square framework). Aratus in marsh never escaped upward, so to use a chi-square analysis framework we added a single incidence of upward evasion to crab responses in marsh. Expected values for evasion in each direction were calculated as 1/2 (2 potential directions) of the total number of responsive crabs per habitat type. To analyze the frequency of crabs missing limbs in each habitat during the behavioral surveys, we used a generalized linear model fit with a binomial distribution to compare the probability that a crab is missing at least one leg, using habitat type (mangrove, salt marsh) as a fixed effect. Among crabs that were missing at least one leg, we then compared the number of legs missing with a generalized linear model fit with a Poisson distribution and habitat type as a fixed effect. To avoid habitat-specific tethering artifacts, we used generalized linear models fitted with a binomial distribution to analyze predation probability by tidal exposure treatment (i.e. aerial at low tide or aquatic at high tide) separately for each habitat (Peterson and Black 1994; Riley and Griffen 2017). To visualize the relative influence of aerial and aquatic predation risk, we combined aerial and aquatic predation probabilities into an aerial:aquatic predation probability where >1 indicates relatively higher aerial risk and <1 indicates relatively higher aquatic risk. RESULTS Characterization of Aratus and vegetation attributes Crab carapace widths were significantly smaller in salt marshes than in mangroves (10.01 ± 0.15 mm CW and 17.50 ± 0.36 mm CW, respectively, mean ± SE; F1, 360 = 409.6, P < 0.0001; Figure 2a). However, the range of carapace widths for crabs collected in mangroves encompasses all crab sizes observed in salt marsh vegetation. We observed a comparable pattern in substrate diameters, which were significantly smaller on average in salt marshes (6.48 ± 0.14 mm, mean ± SE) compared with mangroves (34.57 ± 1.22 mm, mean ± SE; F1, 516 = 1628, P < 0.0001; Figure 2b). Furthermore, combining the carapace width and substrate diameter data into individual carapace width:substrate diameter ratios revealed that crab size relative to substrate differed between vegetation types (F1, 360 = 388.4, P < 0.0001; Figure 2c). On average, crabs were larger (>1) than their substrates in salt marsh habitat, but smaller (<1) than their substrates in mangrove habitat. Figure 2 View largeDownload slide Differences in size-scaling between crabs and vegetation. Widths of (a) Aratus carapaces and (b) their perches (i.e. vegetation substrate, stem, or leaf), and (c) carapace width:substrate diameter ratio of crabs collected from mangrove (n = 152) and salt marsh habitat (n = 210) measured in field surveys. A 1:1 ratio means that crab carapace width and substrate diameter are equivalent. Asterisks indicate statistical significance at α = 0.05. Photo credits: Tom Murray and CAJ. Figure 2 View largeDownload slide Differences in size-scaling between crabs and vegetation. Widths of (a) Aratus carapaces and (b) their perches (i.e. vegetation substrate, stem, or leaf), and (c) carapace width:substrate diameter ratio of crabs collected from mangrove (n = 152) and salt marsh habitat (n = 210) measured in field surveys. A 1:1 ratio means that crab carapace width and substrate diameter are equivalent. Asterisks indicate statistical significance at α = 0.05. Photo credits: Tom Murray and CAJ. Predator avoidance behavior Probability of evasion was significantly lower in marsh (0.44) than in mangroves (0.58; binomial GLM, Likelihood Ratio Test χ2 = 6.1, df = 1, P = 0.013; n =159 per habitat; Figure 3a). There was no significant difference in probability of evasion (0.5) between crabs that were qualitatively categorized as “broader than” (n = 145) or “narrower than” (n = 121) their substrates (binomial GLM, LRT = 0.0553, df = 1, P = 0.814); this pattern remained true when habitat type was included as an interaction term. Among crabs that evaded, the direction of evasion (i.e. movement upward away from water or downward toward water) differed significantly between habitat types (heterogeneity G-test by habitat: df = 1, P < 0.0001). In mangroves, Aratus preferentially evaded upward (n = 74 up, n = 18 down; individual G-test for direction of evasion: df = 1, P < 0.0001). In salt marsh, Aratus exclusively evaded downward (n = 70; individual G-test for direction of evasion: df = 1, P < 0.0001, Figure 3b). Figure 3 View largeDownload slide Evasion behavior in mangrove and salt marsh habitats. (a) Probability and (b) direction of evasion in mangrove and marsh habitat determined by in situ behavior assays (no response: n = 67 mangrove, n = 89 marsh). In mangrove habitats, sufficient aboveground substrate allowed crabs to move either up or down to avoid predators. In salt marsh habitat, limited aboveground substrate constrained crabs to downward movement into the aquatic environment to avoid predators. (c) Number (mean ± SE) of missing legs per crab—a measure of autotomy—documented in field surveys. Asterisks indicate statistical significance at α = 0.05. Illustration components credit: J. Thomas, T. Saxby, K. Kraeer, and L. Van Essen-Fishman; Integration & Application Network, University of Maryland Center for Environmental Science. Figure 3 View largeDownload slide Evasion behavior in mangrove and salt marsh habitats. (a) Probability and (b) direction of evasion in mangrove and marsh habitat determined by in situ behavior assays (no response: n = 67 mangrove, n = 89 marsh). In mangrove habitats, sufficient aboveground substrate allowed crabs to move either up or down to avoid predators. In salt marsh habitat, limited aboveground substrate constrained crabs to downward movement into the aquatic environment to avoid predators. (c) Number (mean ± SE) of missing legs per crab—a measure of autotomy—documented in field surveys. Asterisks indicate statistical significance at α = 0.05. Illustration components credit: J. Thomas, T. Saxby, K. Kraeer, and L. Van Essen-Fishman; Integration & Application Network, University of Maryland Center for Environmental Science. Among surveyed Aratus, the probability of a crab having dropped at least one leg (≥1 leg missing) was no different in salt marsh (0.30) and mangrove (0.26) habitat (binomial GLM, LRT = 0.47, df = 1, P = 0.49). However, among those missing at least one leg, the average number of legs missing was nearly twice as high in marsh (2.9 ± 0.16 SE) as in mangroves (1.5 ± 0.12 SE; Poisson GLM, LRT = 5.22, df = 1, P = 0.02; Figure 3c). With the greater number of legs dropped, there was also significantly more variance in the number of legs dropped by crabs in salt marshes than in mangroves (F = 2.3823, P < 0.0001, ratio of variances = 2.38; Supplementary Figure S1). Preliminary analyses also reveal some influence of sex and carapace width on probability and number of legs dropped (Appendix 1). Predation probability and relative risk Aquatic risk exceeded aerial risk in both salt marshes (aerial:aquatic predation risk ratio = 0.16) and mangroves (0.5) (Figure 4). In novel salt marsh habitat, aquatic predation probability (i.e. during high tide exposure) was significantly higher (0.24) than aerial predation probability (0.04) (binomial GLM, LRT = 9.31, df = 1, P = 0.00228). In mangrove habitats, predation probability did not differ significantly between aquatic (0.14) and aerial (0.07) exposure (binomial GLM, LRT = 1.28, df = 1, P = 0.2575). Figure 4 View largeDownload slide Aerial and Aquatic Predation Risk. Using predation probabilities determined from tethered survival experiments, the ratio of aerial:aquatic predation probability provides the relative risk from above and below. Values >1 indicate higher aerial risk and <1 indicate higher aquatic risk. Asterisks indicate significant at α = 0.05 within-habitat differences between aerial and aquatic predation risk. Illustrations from Phylopic.com, with credit given to Rebecca Groome (https://creativecommons.org/licenses/by/3.0/) for bird silhouette. Figure 4 View largeDownload slide Aerial and Aquatic Predation Risk. Using predation probabilities determined from tethered survival experiments, the ratio of aerial:aquatic predation probability provides the relative risk from above and below. Values >1 indicate higher aerial risk and <1 indicate higher aquatic risk. Asterisks indicate significant at α = 0.05 within-habitat differences between aerial and aquatic predation risk. Illustrations from Phylopic.com, with credit given to Rebecca Groome (https://creativecommons.org/licenses/by/3.0/) for bird silhouette. DISCUSSION In range-center mangrove habitat, 3 out of 4 Aratus evaded upward into the complex aboveground vegetation structure provided by mangrove branches and canopies, effectively avoiding the greater predation risk associated with water. In contrast, in the range-edge salt marsh habitat, the dominant short-statured grass restricted Aratus escape to downward evasion (Figure 3b), resulting in a lower probability of evasion but a higher observed frequency of limb autotomy (Supplementary Figure S1). Given that downward evasion toward water exposes Aratus to higher predation risk, the increase in autotomy in marsh populations likely represents a switch to the secondary predator avoidance strategy where vegetation substrate has made the primary avoidance strategy less effective. The behavior shift likely optimizes survival among populations in novel habitat along the range edge by minimizing contact with water, which is associated with the highest mortality risk. Previous studies report that Aratus face higher predation from aquatic sources compared with aerial or terrestrial sources (Wilson 1989; Yeager et al. 2016). However, prior experimental work suggests that Aratus mortality increases further when crabs evade downward into water to avoid aerial predation risk and are eaten by aquatic predators (Yeager et al. 2016). Accordingly, we expected that crab behavioral choices could be mediated by relative, rather than absolute, predation risk, such that the ratio of aquatic to aerial predation risk represents the trade-off that Aratus navigate when choosing whether and when to evade, shaping the risk of subsequently having to drop limbs (autotomize). Here, aerial-to-aquatic predation probability ratios indicated a relatively higher risk from aquatic than aerial sources in both habitats, but the relative risk of aquatic predation was 3-fold stronger in marsh (0.5 compared with 0.16 in mangroves; Figure 4). Thus, predation conditions should select for water avoidance in both habitats, but the strength of selection for this behavior is likely stronger in salt marsh habitats. Though predation was low in both habitats during our tethering study, predation probability during aquatic exposure exceeded that from aerial exposure in salt marsh but not mangrove habitat (Figure 4). We chose to tether crabs in situ to evaluate representative predation probability for the local population. Thus, we tethered smaller crabs in salt marsh than in mangroves (Figure 2a). The relatively smaller individuals tethered in salt marshes could have led to higher predation rates during aquatic exposure in marsh by making Aratus available to a broader suite of potentially gape-limited predators (i.e. fish). However, large Aratus transported from mangrove habitat are even more likely than small crabs to be depredated in salt marsh habitat (Riley and Griffen 2017). Thus, the higher observed aquatic predation probability in salt marsh habitat is likely due to an expanded predator suite or decreased concealment, rather than to the Aratus size distribution. The differences that we observed in predator avoidance behavior between range-center and range-edge Aratus populations is consistent with previous work showing that Aratus life history traits and foraging behavior differ markedly in salt marsh habitats (range edge) compared with mangrove habitats (range center). Although Aratus in novel salt marsh habitats are reproductive and persistent across years (Riley and Griffen 2017), they are also smaller and have lower fecundity and offspring quality compared with populations from primary mangrove habitats (Riley and Griffen 2017). Aratus also exhibit greater site fidelity in mangrove habitats than in salt marsh habitats (Cannizzo and Griffen 2016). Vegetation structural disparities can change the abiotic environment and thereby affect these physiological and behavioral traits; for example, salt marsh vegetation creates less shade than mangroves, leading to higher microhabitat and Aratus body temperatures that alter foraging and thermoregulatory behavior (Cannizzo et al. 2018). Thus, underlying vegetation structural attributes that place selection pressures on key life-history traits also mediate behaviors that affect the survival and persistence of Aratus populations across their expanding range. Differences in vegetation structure that alter concealment or visibility can also change predator avoidance behavior (Camp et al. 2012). Reduced concealment leads to enhanced vigilance in both terrestrial (Beauchamp 2010; Embar et al. 2011) and aquatic systems (Laurel and Brown 2006). In our study, salt marsh substrates were narrower than mangrove substrates (Figure 2b), and although Aratus were also smaller in salt marsh habitats compared with mangrove habitats (Riley and Griffen 2017), on average, salt marsh Aratus remained larger than their substrates (Figure 2c). In contrast, mangrove Aratus were smaller than their substrates (Figure 2c). Thus, Aratus in salt marshes are less concealed than those in mangroves, which was expected to increase predation risk and avoidance behavior in marsh habitat. Behavioral change with perceived predation risk has been demonstrated in other crab species (Belgrad and Griffen 2017), but Aratus behavior in this study did not correspond with differences in concealment. Instead, we observed that Aratus were less likely to evade in marsh than in mangroves and that probability of evasion was unrelated to carapace to substrate scaling (Figure 3). These data suggest that the novel salt marsh habitat offered lower concealment as well as fewer opportunities to evade and still avoid the risky water interface. Although Aratus were less likely to evade (primary avoidance strategy) in salt marshes, individuals in salt marshes had greater numbers of missing legs. The probability of missing a limb did not differ by habitat, but crabs in salt marshes were missing nearly twice as many limbs on average as those in mangroves (Figure 3). More limb loss likely indicates an increased use of autotomy as a secondary avoidance strategy where evasion options are limited, but it could also be explained by higher rates of aquatic predator encounter due to increased downward evasion. In many crab species, limbs can also be lost due to intraspecific fighting. We did not record the location (e.g. chelipeds vs. walking legs) of missing limbs and thus could not evaluate potential leg loss from intraspecific fighting; however, limb loss due to aggression is rare in Aratus (Warner 1970). Though loss of a single limb may have minimal fitness costs, costs increase with loss of additional limbs, suggesting that Aratus in salt marshes incur relatively higher bioenergetic costs from limb loss and regeneration (Lindsay 2010; Maginnis et al. 2014; Lavalli and Spanier 2015). The higher average number of missing limbs among Aratus in salt marshes suggests that there is an increased fitness cost to evasion in salt marshes that makes otherwise costly autotomy relatively more favorable. In our study, differences in vegetation structure between mangrove (range center) and salt marsh (range edge) habitats altered Aratus predator avoidance behavior by controlling possible evasion routes. The physical limitations imposed by salt marsh vegetation structure caused crabs to shift behavioral strategies from predator evasion to autotomy to optimize survival in a novel habitat. Future work, including a common garden experiment to distinguish whether behavioral changes originate from plasticity or microevolution, will be needed to deepen our understanding of how behavioral shifts among frontier populations arise and shape expanding population persistence. CONCLUSION A climate-driven spatial mismatch between Aratus and its primary mangrove habitat provided an opportunity to better understand how structural vegetation attributes mediate predator avoidance behaviors in novel risk landscapes. As climate change leads to expanded species distributions, mismatches between interacting organisms will increase contact with novel environments where organisms experience new abiotic and biotic conditions. Range expansion that puts organisms in contact with new habitats magnifies the potential that alternative behavior strategies will be necessary to navigate novel habitat forms and associated changes in risk. Behavioral changes that minimize novel risk sources will likely undergo strong selection, as these behaviors are essential to survival and persistence of populations in novel habitats (Canestrelli et al. 2016; Siepielski and Beaulieu 2017). Indeed, behavioral shifts may be particularly important for frontier population persistence if plastic changes in behavior outpace genetic changes in physiology and morphology. Future studies should include considerations of emerging behavior that not only promotes range expansion (Hudina et al. 2014; Chuang and Peterson 2016) but also ensures local survival among range-edge populations. SUPPLEMENTARY MATERIAL Supplementary data are available at Behavioral Ecology online. FUNDING This work was completed without direct funding. Data accessibility: Analyses reported in this article can be reproduced using the data provided by Johnston and Smith (2018). This study was conceptualized in part thanks to early discussions with M. Riley and L. Yeager. We appreciate the assistance of J. Beauvais, E. Dark, M. Nathan, and W. Scheffel in the field and the logistical support provided by T. Osborne, the staff at the Smithsonian Marine Station in Fort Pierce, FL and Little Talbot Island State Park, FL. We are grateful to D. Adams, A. Brown, A. Gehman, L. Haram, A. Majewska, C. Phillips, V. Schutte, and E. Tielens for providing constructive feedback on early drafts of this paper. REFERENCES Agrawal AA . 2001 . Phenotypic plasticity in the interactions and evolution of species . Science . 294 : 321 – 326 . Google Scholar CrossRef Search ADS PubMed Beauchamp G . 2010 . Relationship between distance to cover, vigilance and group size in staging flocks of semipalmated sandpipers . Ethology . 116 : 645 – 652 . Belgrad BA , Griffen BD . 2017 . Habitat quality mediates personality through differences in social context . Oecologia . 184 : 431 – 440 . Google Scholar CrossRef Search ADS PubMed Camp MJ , Rachlow JL , Woods BA , Johnson TR , Shipley LA . 2012 . When to run and when to hide: the influence of concealment, visibility, and proximity to refugia on perceptions of risk . Ethology . 118 : 1010 – 1017 . Google Scholar CrossRef Search ADS Canestrelli D , Bisconti R , Carere C . 2016 . Bolder takes all? The behavioral dimension of biogeography . Trends Ecol Evol . 31 : 35 – 43 . Google Scholar CrossRef Search ADS PubMed Cannizzo ZJ , Dixon SR , Griffen BD . 2018 . An anthropogenic habitat within a suboptimal colonized ecosystem provides improved conditions for a range-shifting species . Ecol Evol . 8 : 1521 – 1533 . Google Scholar CrossRef Search ADS PubMed Cannizzo ZJ , Griffen BD . 2016 . Changes in spatial behaviour patterns by mangrove tree crabs following climate-induced range shift into novel habitat . Anim Behav . 121 : 79 – 86 . Google Scholar CrossRef Search ADS Cavanaugh KC , Kellner JR , Forde AJ , Gruner DS , Parker JD , Rodriguez W , Feller IC . 2014 . Poleward expansion of mangroves is a threshold response to decreased frequency of extreme cold events . Proc Natl Acad Sci USA . 111 : 723 – 727 . Google Scholar CrossRef Search ADS PubMed Ceradini JP , Chalfoun AD . 2017 . When perception reflects reality: non-native grass invasion alters small mammal risk landscapes and survival . Ecol Evol . 7 : 1823 – 1835 . Google Scholar CrossRef Search ADS PubMed Chen IC , Hill JK , Ohlemüller R , Roy DB , Thomas CD . 2011 . Rapid range shifts of species associated with high levels of climate warming . Science . 333 : 1024 – 1026 . Google Scholar CrossRef Search ADS PubMed Chuang A , Peterson CR . 2016 . Expanding population edges: theories, traits, and trade-offs . Glob Chang Biol . 22 : 494 – 512 . Google Scholar CrossRef Search ADS PubMed Embar K , Kotler BP , Mukherjee S . 2011 . Risk management in optimal foragers: the effect of sightlines and predator type on patch use, time allocation, and vigilance in gerbils . Oikos . 120 : 1657 – 1666 . Google Scholar CrossRef Search ADS Friess DA , Krauss KW , Horstman EM , Balke T , Bouma TJ , Galli D , Webb EL . 2012 . Are all intertidal wetlands naturally created equal? Bottlenecks, thresholds and knowledge gaps to mangrove and saltmarsh ecosystems . Biol Rev Camb Philos Soc . 87 : 346 – 366 . Google Scholar CrossRef Search ADS PubMed Hickling R , Roy DB , Hill JK , Fox R , Thomas CD . 2006 . The distributions of a wide range of taxonomic groups are expanding polewards . Glob Change Biol . 12 : 450 – 455 . Google Scholar CrossRef Search ADS Hudina S , Hock K , Žganec K . 2014 . The role of aggression in range expansion and biological invasions . Curr Zool . 60 : 401 – 409 . Google Scholar CrossRef Search ADS Johnston CA , Lipcius R . 2012 . Exotic macroalga Gracilaria vermiculophylla provides superior nursery habitat for native blue crab in Chesapeake Bay . Mar Ecol Prog Ser . 467 : 137 – 146 . Google Scholar CrossRef Search ADS Johnston CA , Smith RS . 2018 . Data from: vegetation structure mediates a shift in predator avoidance behavior in a range-edge population . Dryad Digital Repository . http://dx.doi.org/10.5061/dryad.1cp758h. Laurel BJ , Brown JA . 2006 . Influence of cruising and ambush predators on 3-dimensional habitat use in age 0 juvenile Atlantic cod Gadus morhua . J Exp Mar Biol Ecol . 329 : 34 – 46 . Google Scholar CrossRef Search ADS Lavalli K , Spanier E . 2015 . Predator adapations of Decapods . In: Thiel M , Watling L , editors. Lifestyles and feeding biology . New York, NY : Oxford University Press . p. 190 – 228 . Lindsay SM . 2010 . Frequency of injury and the ecology of regeneration in marine benthic invertebrates . Integr Comp Biol . 50 : 479 – 493 . Google Scholar CrossRef Search ADS PubMed Lurgi M , López BC , Montoya JM . 2012 . Novel communities from climate change . Philos Trans R Soc Lond B Biol Sci . 367 : 2913 – 2922 . Google Scholar CrossRef Search ADS PubMed Maginnis TL , Niederhausen M , Bates KS , White-Toney TB . 2014 . Patterns of autotomy and regeneration in Hemigrapsus nudus . Mar Freshw Behav Physiol . 47 : 135 – 146 . Google Scholar CrossRef Search ADS Moermond TC . 1979 . Habitat constraints on the behavior, morphology, and community structure of Anolis lizards . Ecology . 60 : 152 – 164 . Google Scholar CrossRef Search ADS Osland MJ , Enwright N , Day RH , Doyle TW . 2013 . Winter climate change and coastal wetland foundation species: salt marshes vs. mangrove forests in the southeastern United States . Glob Chang Biol . 19 : 1482 – 1494 . Google Scholar CrossRef Search ADS PubMed Osland MJ , Feher LC , Griffith KT , Cavanaugh KC , Enwright NM , Day RH , Stagg CL , Krauss KW , Howard RJ , Grace JB et al. 2016 . Climatic controls on the global distribution, abundance, and species richness of mangrove forests . Ecol Monogr . 87 : 341 – 359 . doi: https://doi.org/10.1002/ecm.1248 Google Scholar CrossRef Search ADS Peterson CH , Black R . 1994 . An experimentalist’s challenge: when artifacts of intervention interact with treatments . Mar Ecol Prog Ser . 111 : 289 – 297 . Google Scholar CrossRef Search ADS Pierce AA , Gutierrez R , Rice AM , Pfennig KS . 2017 . Genetic variation during range expansion: effects of habitat novelty and hybridization . Proc R Soc B . 284 : 20170007 . Google Scholar CrossRef Search ADS PubMed R Core Team . 2015 . R: a language and environment for statistical computing . Vienna (Austria) : R Foundation for Statistical Computing . Riley ME , Griffen BD . 2017 . Habitat-specific differences alter traditional biogeographic patterns of life history in a climate-change induced range expansion . PLoS One . 12 : e0176263 . Google Scholar CrossRef Search ADS PubMed Riley ME , Johnston CA , Feller IC , Griffen BD . 2014 . Range expansion of Aratus pisonii (mangrove tree crab) into novel vegetative habitats . Southeast Nat . 13 : N43 – N48 . Google Scholar CrossRef Search ADS Robinson BW , Dukas R . 1999 . The influence of phenotypic modifications on evolution: the Baldwin effect and modern perspectives . Oikos . 85 : 582 – 589 . Google Scholar CrossRef Search ADS Robinson SK , Holmes RT . 1982 . Foraging behavior of forest birds: the relationships among search tactics, diet, and habitat structure . Ecology . 63 : 1918 – 1931 . Google Scholar CrossRef Search ADS Schweiger O , Settele J , Kudrna O , Klotz S , Kühn I . 2008 . Climate change can cause spatial mismatch of trophically interacting species . Ecology . 89 : 3472 – 3479 . Google Scholar CrossRef Search ADS PubMed Siepielski AM , Beaulieu JM . 2017 . Adaptive evolution to novel predators facilitates the evolution of damselfly species range shifts . Evolution . 71 : 974 – 984 . Google Scholar CrossRef Search ADS PubMed Simpson LT , Osborne TZ , Duckett LJ , Feller IC . 2017 . Carbon storages along a climate induced coastal wetland gradient . Wetlands . 37 (6) : 1 – 13 . Warner GF . 1970 . Behaviour of two species of Grapsid crab during intraspecific encounters . Behaviour . 36 : 9 – 19 . Google Scholar CrossRef Search ADS Wilson KA . 1989 . Ecology of mangrove crabs: predation, physical factors and refuges . Bull Mar Sci . 44 : 263 – 273 . Yang LH , Rudolf VH . 2010 . Phenology, ontogeny and the effects of climate change on the timing of species interactions . Ecol Lett . 13 : 1 – 10 . Google Scholar CrossRef Search ADS PubMed Yeager LA , Stoner EW , Peters JR , Layman CA . 2016 . A terrestrial-aquatic food web subsidy is potentially mediated by multiple predator effects on an arboreal crab . J Exp Mar Biol Ecol . 475 : 73 – 79 . Google Scholar CrossRef Search ADS © The Author(s) 2018. Published by Oxford University Press on behalf of the International Society for Behavioral Ecology. All rights reserved. For permissions, please e-mail: firstname.lastname@example.org This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)
Behavioral Ecology – Oxford University Press
Published: Jun 2, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera