R E P O R T Predation risk alters prey caching behavior and resultant tree establishment patterns Alex D. Potash | Robert A. McCleery Department of Wildlife Ecology and Conservation, University of Florida, Gainesville, Florida, USA Correspondence Alex D. Potash Email: apotash@ufl.edu Handling Editor: Burt P. Kotler Abstract The drivers of plant dispersal and distribution are poorly understood. For nut-bearing trees, interactions between seeds and caching granivores can have a strong influence on population dynamics as unrecovered caches often recruit new trees. Granivore-seed interactions may be constrained by predators as fear alters caching and foraging behavior. We investigated how predation risk affects caching and foraging behavior in fox squirrels ( Sciurus niger ) and subse- quent recruitment patterns of American turkey oak ( Quercus laevis ) acorns. We marked and tracked acorns to locate fox squirrel cache and consumption points where we assessed fox squirrel perceived predation risk, measured how caching influenced acorn survival and dispersal, and surveyed juvenile oak distributions. Fox squirrels perceived greater risk at caches than consumption points, suggesting predation risk influences caching decisions and cache distribution. Buried acorns survived at higher rates and only buried acorns germinated, though survival did not vary with dispersal distance. Juvenile oaks and cache locations showed similar hump-shaped patterns of establishment away from nearest potential source trees. We conclude that predation risk indirectly shapes plant establishment by altering granivore behavior. K E Y W O R D S caching, fear, predation, predator – prey, seed dispersal, seed predation, squirrel INTRODUCTION Plants form the foundation of most terrestrial ecosystems. The factors that shape plant establishment, the transition of a dispersed seed to a viable individual, are therefore fundamental to understanding terrestrial ecosystems. For nut-bearing trees, whose large seeds are mostly dispersed by animals (G omez et al., 2019), establishment patterns (i.e., distribution of juvenile trees) may reflect animal behaviors that dictate where dispersers deposit seeds. Many seed-dispersing animals are prey species, and their behaviors are influenced by predators (Brown, 1999). Accordingly, the distributions of nut-bearing trees, such as oaks ( Quercus sp.), may reflect predator – prey interac- tions that shape prey behavior. Across a gradient of perceived predation risk, changes in prey behavior may alter important ecological processes (Palmer et al., 2022), potentially impacting the Received: 31 December 2025 Revised: 16 July 2026 Accepted: 24 August 2026 DOI: 10.1002/ecy.70570 © 2026 The Ecological Society of America. All rights reserved, including rights for text and data mining and training of artificial intelligence technologies or similar technologies. Ecology. 2026;107:e70570. https://onlinelibrary.wiley.com/r/ecy 1 of 8 https://doi.org/10.1002/ecy.70570 establishment patterns of nut-producing trees. Nut- bearing trees often rely on granivores for seed dispersal, but granivores also reduce seed survival via seed preda- tion (Steele et al., 2005, 2020). Thus, factors shaping a granivore ’ s decision to disperse or depredate a seed likely influence the distribution patterns of nut-bearing trees. Granivores prefer foraging in areas with protective cover from predators (Brown, 1999). Accordingly, patterns of seed consumption by granivores may reflect perceived predation from vegetation structure and may further influence where granivores disperse seeds, ultimately shaping patterns of tree establishment. By moving seeds away from their sources, granivores can alter seeds ’ probability of survival (Nathan & Casagrandi, 2004; Peterson & Parker, 2016). Increased dispersal distances may improve survival by reducing density-dependent mortality from pathogens and seed predators (i.e., the Janzen-Connell Hypothesis [Janzen, 1970, Connell, 1971]). Dispersal-survival relationships are also likely to be altered by granivore caching, the burying of seeds for later use, particularly among scatterhoarders, which bury seeds in numerous shallow caches (Steele et al., 2020). Buried seeds are protected from threats like fire, desiccation, and some predators (Greenberg et al., 2012; Woods & Elliott, 2004). Despite high cache recovery rates by granivores (Stapanian & Smith, 1984), some seeds remain unrecovered, and caches provide favorable conditions for germination (Peterson & Parker, 2016; but see Sipes et al., 2013). Perceived predation risk may also drive granivore caching behavior. To reduce pilferage by conspecifics, some scatterhoarders cache seeds where they perceive increased predation risk (Steele et al., 2014, 2015). Scatterhoarders do not need to cache in the riskiest loca- tions; any relative increase in predation risk is sufficient to reduce potential pilferage from naïve conspecifics and benefit the caching individual that knows the location of the cache. This risk-driven caching behavior may increase seed dispersal distances and improve survival, influencing establishment patterns. However, the extent to which fear-driven changes in granivore behavior shape seed dispersal and subsequent tree recruitment remains unclear (Green et al., 2022). If perceived predation risk systematically alters dispersal distances, it could have cascading effects on forest composition and regeneration by shifting where new seedlings establish. We used a series of field experiments and surveys to determine if perceived predation risk influences plant establishment patterns by altering animal interactions with seeds. We used a model scatterhoarding granivore, the fox squirrel ( Sciurus niger ), to study its impact on establishment patterns of the American turkey oak ( Quercus laevis ; henceforth turkey oak). Specifically, we ask the following questions: (1) Does perceived predation risk influence granivore seed consumption and caching? (2) Does perceived predation risk affect the distance at which granivores disperse and cache seeds from their source? (3) Does seed burial (i.e., caching) improve seed survival and germination? (4) Do observed establishment patterns correspond to the survival and dispersal con- straints associated with granivore behaviors? We predicted that: (1) fox squirrels would cache acorns in more open areas where they perceive higher predation risk while consuming acorns in safer locations with more cover (Brown, 1999; Steele et al., 2014); (2) higher per- ceived predation risk would increase seed dispersal dis- tances by driving fox squirrels to cache acorns farther from turkey oaks, particularly in areas with less protec- tive cover (Potash et al., 2019); (3) buried acorns would have greater survival and germination rates than surface seeds due to reduced exposure to seed predators and envi- ronmental stressors (Greenberg et al., 2012; Woods & Elliott, 2004); (4) establishment patterns would align with the predicted survival and dispersal constraints from granivore behavior, with juvenile oak distribution reflecting where granivores were most likely to cache seeds (Steele et al., 2005). METHODS Study species The fox squirrel is a scatterhoarding granivore found throughout much of eastern North America (Koprowski, 1994). Due to their large body size, fox squirrels primarily forage on the ground. In the southeastern United States, fox squirrels are preyed on by terrestrial and avian preda- tors, including coyotes ( Canis latrans ), bobcats ( Lynx rufus ), and red-tailed hawks ( Buteo jamaicensis ). They perceive the greatest risk in areas with less overhead cover (Potash et al., 2019, 2023). We selected the turkey oak because of its reliance on animal dispersal, and its low density in longleaf pine ( Pinus palustris ) savannas, allowing us to better simulate acorn dispersal from an individual tree. Turkey oaks pro- duce acorns that mature and drop September – December and are consumed by numerous animals (Appendix S1). Fox squirrels cache acorns 1 – 2 cm below ground, poten- tially selecting cache points in areas with increased pre- dation risk to combat pilferage (Bartlow et al., 2018; Steele et al., 2014). Unrecovered caches may have increased germination rates and are potential establish- ment locations for oaks (Haas & Heske, 2005). 2 of 8 POTASH and M C CLEERY 19399170, 2026, 10, Downloaded from https://esajournals.onlinelibrary.wiley.com/doi/10.1002/ecy.70570 by University Of Florida, Wiley Online Library on [29/10/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Study site We conducted research at the Ordway Swisher Biological Station (Ordway), a 3850-ha research facility located near Melrose, FL. Ordway was dominated by upland sandhill open-pine ecosystems, primarily comprised of longleaf pine savanna. Turkey oaks occurred sporadically thro- ughout the longleaf pine matrix and have been suppressed by frequent, low-intensity fires. Land man- agers at Ordway use prescribed burns on a 2-year rotation to prevent hardwood encroachment, with most pre- scribed fires occurring between January and March. Field methods To investigate if predation risk altered acorn dispersal distance and survival, ultimately influencing oak estab- lishment pattern, we (1) identified where fox squirrels cached and consumed acorns, (2) assessed fox squirrel perceived predation risk at cache and consumption points, (3) measured the influence of caching on acorn survival and dispersal, and (4) surveyed juvenile oak dis- tributions in unmanipulated areas. All field methods were approved by the University of Florida Institutional Animal Care and Use Committee (IACUC #202011137). Cache and consumption point identification Between October 2021 and March 2022, we established eight research sites at adult ( ≥ 10 cm diameter at breast height [dbh]) turkey oaks, >5 m away from the nearest conspecific. We selected a 10 cm dbh cutoff because oaks of this size are more resistant to topkill by fire (Dey et al., 2008) and are therefore the trees most likely to contribute to establishment. One meter from the base of each tree in a random direction, we cleared a 1-m 2 area of bare ground (henceforth: station). We mounted a short-focus wildlife camera (Spartan SR2-BK; Spartan Camera, Johns Creek, Georgia, USA) on each tree to identify the species that removed each acorn from the station (Appendix S1: Figure S1A). At each station, we placed 15 marked acorns in a 3 × 5 pattern, separating each acorn from its neighbors by 5 – 10 cm (Appendix S1: Figure S1B). To each acorn, we affixed 10 cm of steel wire with 5 cm of flagging with the acorn ID. Every day we systematically searched for acorns up to 30 m away, the maximum distance fox squirrels moved acorns during our pilot study, and monitored cached acorns for recaching (Bartlow et al., 2018). We replenished stations with 15 new acorns between one and six times, depending on removal rates, and defined each replenishment as a Deployment. We classified recovered acorns (points) as consumed, cached (i.e. buried), or moved but remaining on the surface. We identified consumption points based on shell fragments and considered an acorn cached if it was not visible from directly above. At each point, we surveyed environmental characteristics associated with caching and foraging including canopy cover, visual obstruction, distance to the nearest adult tree (tree distance), distance to the nearest adult turkey oak (oak distance), and the distance moved from the station (Appendix S1). Perceived risk at cache and consumption sites In February 2022, we used a giving-up-density (GUD) approach (Brown, 1988) to measure perceived predation risk at cache and consumption points and paired each point with a random point (Appendix S1). By quantifying quitting harvest rates and cumulative foraging costs, GUD measures perceived risk as the amount of rem- aining food at a foraging patch, with more food indicat- ing greater perceived predation risk. Impacts of caching on acorn survival and germination Between October 2022 and March 2023, we assessed how caching altered acorn survival by establishing 16 plots at adult turkey oak trees, >5 m from the nearest adult con- specific. Using the turkey oak as the plot center, we cre- ated five evenly spaced (i.e., 72 apart) 10-m-long transects. Along each transect, we buried one acorn 1 – 2 cm below ground at 1, 5, and 10 m to simulate a fox squirrel cache and placed a second acorn on the surface. At three of the plots, we only deployed buried acorns to maximize our understanding of the influence of caching on survival and germination given our limited acorn sup- ply. Some acorns placed at 5 and 10 m were closer to neighboring adult turkey oaks than to the oak at the cen- ter of the plot, and we therefore measured the distance from each acorn to the nearest adult turkey oak to assess the effects of putative dispersal distance on acorn survival and germination probability. We visited each plot weekly for 12 weeks and recorded the status of each acorn as surviving, failed, or germinated. All acorns were classified as surviving until we observed evidence of germination or failure. We clas- sified an acorn as germinating if we observed a protrud- ing radical (Branco et al., 2002; Appendix S1: Figure S4). ECOLOGY 3 of 8 19399170, 2026, 10, Downloaded from https://esajournals.onlinelibrary.wiley.com/doi/10.1002/ecy.70570 by University Of Florida, Wiley Online Library on [29/10/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Surveying establishment pattern In February 2024, we measured turkey oak establishment patterns by surveying juvenile oaks at 12 randomly selected locations. Unlike our other sampling strategies, which established locations at adult turkey oaks, we selected random locations independent of turkey oaks to prevent biasing our sampling. At each point, we walked five equally spaced 15-m-long transects radiating out- wards, identifying all turkey oak juveniles ( ≤ 2 m tall), within 10 cm of either side of the transect. We measured the distance from the stem of each turkey oak juvenile to the nearest adult turkey oak, which we used as a measure for putative dispersal distance. Statistical methods To identify the factors that best explained fox squirrel caching and consumption behavior, we limited our data to only acorns cached and consumed by fox squirrels and used mixed logistic regression to evaluate the response variable Result (1 = cached, 0 = consumed or moved) against each of the environmental covariates we mea- sured. We fit candidate models that included each predic- tor variable by itself and a null model. All models included Deployment nested within Site as a random effect to account for spatiotemporal autocorrelation among groups of acorns deployed at each station. Using the parameter from our most parsimonious model, we used a resource selection function approach to assess whether cache and consumption sites differed from avail- able sites (Appendix S1). We used Akaike ’ s information criteria for small sample size (AICc) to identify the most parsimonious models and tested parameter significance using a Wald test. To identify the establishment pattern based on cache distribution, we used kernel density esti- mates (KDE) to calculate smoothed probability distribu- tions with 95% CI of cache and juvenile oak distributions relative to the log-transformed distance to the nearest adult turkey oak (Appendix S1: Figure S2). To test differences in perceived predation risk between cache and consumption points, we compared GUDs using a mixed-effects analysis of variance (ANOVA). We transformed the GUD data with a Tukey ’ s Ladder of Powers to improve normality assumptions. We fit the ANOVA with the GUD as a response variable and the binary variable Type (1 = Cached, 0 = Not Cached) as the only fixed effect. We included the point ID as a random effect to account for repeated measures at the same point over several days and restricted our data to points where we recorded GUD for 3 days. If we found a significant difference in GUD at cache and consumption points, we assessed the environmental correlates of GUD by separately fitting cached and consumed points to the same set of candidate models we used to assess cache point selection. To examine the influence of burial and putative dis- persal distance on acorn survival, we used mixed logistic regression with the response variable survival (1 = sur- vived, 0 = not survived). We assessed acorn survival based on Depth (1 = buried, 0 = surface) and the scaled log-transformed distance to the nearest adult turkey oak. We also fit additive (Depth + Oak Distance), interaction (Depth × Oak Distance), and null models. We included plot as a random effect to account for unmeasured spatial auto- correlation between acorns from the same plot and ranked models by AICc to identify the most parsimonious. We fit models using the glmmTMB package (Brooks et al., 2017) implemented in program R (R Core Team, 2024). If depth, but not oak distance, influenced acorn sur- vival, we further estimated the weekly survival of buried and surface acorns using a Kaplan – Meier survival analysis, a nonparametric method for comparing categorical survival curves (George et al., 2014; Kaplan & Meier, 1958). We right-censored acorns that germinated and survived the duration of the study. We excluded acorns that failed because of nonvertebrate sources because of unequal detec- tion of invertebrate-caused mortalities between surface and buried acorns. We used a log-rank test to assess the signifi- cance of differences in buried and surface acorn survival (Bland & Altman, 2004). We then assessed whether germination probability differed between surface and buried acorns using a Fisher ’ s Exact Test because of the low number of germinants. Fisher ’ s Exact Test is preferred over the Chi-square test for categorical data comparisons when sample sizes are small (Choi et al., 2015). If we found that only buried or surface acorns germinated, we removed data points from the group without germinants before fitting a mixed logistic regression model to identify whether the odds of germination (1 = germinated, 0 = not germinated) were driven by dispersal distance and included plot as a random effect. We estimated the spatial establishment pattern as a function of distance to the nearest adult turkey oak using the KDE for cache points. We constrained the KDE to positive values by log-transforming the dispersal distance and calculated the KDE iteratively over several orders of magnitude of smoothing parameters (h) to identify the range of possible density estimates (Appendix S1: Figure S3). To determine if the observed establishment pattern of juvenile oaks could be derived from the distri- bution of caches, we conducted a two-sided nonparamet- ric Kolmogorov – Smirnov test to assess whether they originate from the same distribution. 4 of 8 POTASH and M C CLEERY 19399170, 2026, 10, Downloaded from https://esajournals.onlinelibrary.wiley.com/doi/10.1002/ecy.70570 by University Of Florida, Wiley Online Library on [29/10/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License R E S U L T S Caching and consumption points We deployed a total of 371 marked acorns at eight sta- tions. Marks failed (flagging recovered but not the acorn) on 123 acorns, and 14 acorns were removed due to holes, rot, or other damage, resulting in 234 usable data points. Of these, 18 were cached by fox squirrels and one was cached by a southern flying squirrel ( Glaucomys volans ). Ten other acorns were moved but not buried by coyotes ( n = 3), fox squirrels ( n = 2), and unidentified rodents ( n = 5). The remaining 205 acorns were consumed (Appendix S1: Table S1); 137 (66.8%) by fox squirrels, 111 (81%) of which were consumed at the station. The only competitive model (Appendix S1: Table S2) explaining caching behavior included Oak Distance as the only predictor ( b β = 4.36, 95% CI = 2.69, 6.03), indicat- ing that odds of caching increased with distance away from the nearest adult oak tree. Fox squirrels selected cache points, but not consumptions points, farther from turkey oaks relative to available habitat (Appendix S1). GUD We collected 78 GUD observations at used sites ( N Cached = 45, N Consumed = 33) and 38 at random sites ( N Cached = 20, N Consumed = 18). Mean GUD was higher, indicating greater perceived risk, at cache points (5.67 g, SD = 1.42) than at consumption points (4.66 g, SD = 1.58; b β = − 21.95, 95% CI = − 42.7, − 1.19, Figure 1A). Mean GUD at used cache points was also sig- nificantly greater than at random points (5.01 g, SD = 1.85), indicating that fox squirrels selected cache points where they perceived greater risk (Appendix S1). Acorn survival and germination For the acorn survival experiment, we deployed 433 acorns (194 surface, 239 buried) across 16 sites. In total, 413 acorns experienced a mortality event, including 165 (40%) due to non-vertebrate agents. Excluding non- vertebrate-caused mortalities, median survival for buried acorns (68 days, 95% CI = 56, 74) was 22 days longer ( χ 2 = 15.3, df = 1, p < 0.001; Figure 1B) than for surface acorns (46 days, 95% CI = 38, 57). The most parsimoni- ous model included only the predictor variable depth ( b β = 1.23, 95% CI = 0.11, 2.34; Appendix S1: Table S3), indicating that buried acorns were 3.41 times more likely to survive than acorns on the surface. The only other model ranked better than the null also included oak dis- tance ( b β = 0.15, 95% CI = − 0.29, 0.58), though it was not an informative parameter as the CI included 0. Of the 20 acorns that survived through the conclusion of the experiment, 16 (80%) were buried, and 14 of these germinated compared to no surface acorns (Fisher ’ s Exact p < 0.001). Since only buried acorns germinated, we removed surface acorns from our model investigating the effects of putative dispersal distance on the odds of germination, finding no effect ( p = 0.37). Establishment surveys We surveyed 193 juvenile turkey oaks located 0.98 – 32 m from the nearest adult. The mean distance (8.26 m; 95% CI: 7.36 – 9.16 m) exceeded the median (6.42 m), indicat- ing a positively skewed distribution similar to that of caches relative to the nearest adult turkey oak (Figure 1C). However, the Kolmogorov – Smirnov test showed that distributions of juvenile oaks and caches dif- fered ( D = 0.23, p < 0.001), with caches having a shorter mean (5.11 m, 95% CI = 3.34 m, 6.89 m) and median (3.18 m) distance to the nearest adult oak. D I S C U S S I O N Perceived predation risk affected plant establishment by constraining granivore caching (Figure 1). As predicted, per- ceived risk was higher at cache compared with consumption points, and squirrels cached farther from adult oaks. Caching as a means of burying acorns improved their sur- vival and germination probability. Thus, fear-driven caching constrained oak establishment patterns by shaping seed dis- persal, survival, and germination (Steele et al., 2015). While our conclusions are based on a limited number of caching events, the patterns we observed are consistent with foraging behavior theory under predation risk. Fear of predation influenced whether squirrels cached or con- sumed acorns (Figure 1D). Consistent with optimal forag- ing theory under predation risk (Brown, 1999), squirrels consumed most acorns under protective canopy where they perceived less risk (Figure 1E) and cached acorns in riskier locations (Figure 1F). The importance of distance to the nearest oak rather than the nearest tree likely reflects fox squirrel use of oak trees as refugia more than longleaf pines, which offer little overhead cover (Conner & Godbois, 2003). Although caching farther from source oaks may place acorns in locations convenient for later recovery, higher GUDs at cache sites than at consumption sites sug- gest that predation risk also influences cache-site selection. ECOLOGY 5 of 8 19399170, 2026, 10, Downloaded from https://esajournals.onlinelibrary.wiley.com/doi/10.1002/ecy.70570 by University Of Florida, Wiley Online Library on [29/10/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Average cache distance (5.11 m) equaled the approxi- mate radius of a turkey oak crown, clustering acorns at the canopy edge (Figure 1F). At this distance, squirrels likely experienced increased predation risk from avian predators due to reduced canopy cover (Potash et al., 2019). Although the canopy edge provides some overhead cover, any increase in predation risk relative to the base of the tree, where canopy cover is greater, should be suffi- cient to deter conspecific pilferers, who lack precise knowledge of cache locations and must engage in risky searching behavior to find it (Brown, 1999). Furthermore, burial, inherent in caching, was critical to acorn survival and germination (Figure 1G). Although juvenile tree dis- tributions differed from cache distributions, both increased with distance from adult turkey oaks, suggesting fox squir- rel caching shaped vegetation patterns in this longleaf pine ecosystem (Figure 1H). Juvenile oak distribution was more positively skewed than cache distribution, with greater maximum dispersal distances observed. Variation in pilfering behavior may F I G U R E 1 (A) Fox squirrels perceive greater risk at cache points than consumption points, indicated by higher giving-up densities (mean and 95% CI; gray points show individual observations). (B) Buried seeds have higher survival than surface seeds, and (C) similar distributions of fox squirrel caches and turkey oak juveniles (kernel density estimates ±95% CI) relative to adult oaks are consistent with unrecovered caches contributing to establishment patterns (hash marks on top and bottom show individual observations). Fox squirrels (D) acquiring viable acorns (brown acorn) either (E) consume (seed mortality indicated by black acorn) them under protective canopy cover or (F) cache them farther from the source oak and under more open canopies, which are cues of increased avian predation risk. Acorn survival is independent of distance, but (G) buried seeds have higher survival than surface seeds, and unrecovered caches (H) may germinate and for the basis for establishment. Figure created in BioRender (Potash, 2026; https://BioRender.com/dimr1cu). 6 of 8 POTASH and M C CLEERY 19399170, 2026, 10, Downloaded from https://esajournals.onlinelibrary.wiley.com/doi/10.1002/ecy.70570 by University Of Florida, Wiley Online Library on [29/10/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License have contributed to observed differences between cache and juvenile distributions, though little is known about pilfering in this system. High mark failure rates (36%) likely biased our data by limiting detection of rare, distant caches (Appendix S1). Nonmammalian scatterhoarders, such as corvids, can also disperse acorns farther than fox squirrels (Vander Wall, 2001), but we rarely observed them interacting with marked acorns. Our assessment of putative establishment was limited because we did not distinguish seed-derived individuals from resprouts; thus, distance to the nearest adult oak was a coarse proxy of dispersal. Moreover, the larger juvenile turkey oaks may have been filtered by selection pressures such as herbiv- ory (Martínez et al., 2013), which tends to be more intense near adult trees. If juveniles close to adults expe- rience higher mortality, surviving individuals would be biased toward greater distances from adults. Finally, in developed and rapidly developing regions, including the southeastern United States (Potash et al., 2026), anthro- pogenic changes in land use and predator communities may alter granivores ’ landscape of fear, impacting emer- gent establishment patterns. By linking predation risk and prey behavior to plant establishment patterns, this study enhances our under- standing how fear of predators can shape terrestrial eco- systems. Faunal communities are increasingly recognized for their influence on ecosystem functions (e.g., Schmitz et al., 2018), and our findings highlight the importance of predator – prey interactions in these processes, providing insight into the mechanisms that underpin global ecosystems. A C K N O WL E D G M E N T S Support was provided by the University of Florida IFAS and Ordway Swisher Biological Station. L. Anderson cre- ated the Red-tailed hawk in Figure 1, available via PhyloPic. Figure 1 was produced using BioRender. We thank B. Kotler and anonymous reviewers, whose com- ments improved this manuscript. We thank the volun- teers that aided in data collection. C O N F L I C T O F I N T E R E S T S T A T E M E N T The authors declare no conflicts of interest. D A T A A V A I L A B I L I T Y S T A T E M E N T Data and code (Potash & McCleery, 2026) are available in Figshare at https://doi.org/10.6084/m9.figshare.33289692.v2. O R C I D Alex D. Potash https://orcid.org/0000-0002-6343-9390 Robert A. McCleery https://orcid.org/0000-0001-7018- 005X RE FER EN CES Bartlow, A. W., N. I. Lichti, R. Curtis, R. K. Swihart, and M. A. Steele. 2018. “ Re-Caching of Acorns by Rodents: Cache Management in Eastern Deciduous Forests of North America. ” Acta Oecologica 92: 117 – 122. Bland, J. M., and D. G. Altman. 2004. “ The Logrank Test. ” BMJ 328: 1073. Branco, M., C. Branco, H. Merouani, and M. H. Almeida. 2002. “ Germination Success, Survival and Seedling Vigour of Quercus suber Acorns in Relation to Insect Damage. ” Forest Ecology and Management 166: 159 – 164. Brooks, M. E., K. Kristensen, K. J. van Benthem, A. Magnusson, C. W. Berg, A. Nielsen, H. J. Skaug, M. Mächler, and B. M. Bolker. 2017. “ glmmTMB Balances Speed and Flexibility among Packages for Zero-Inflated Generalized Linear Mixed Modeling. ” R Journal 9: 378 – 400. Brown, J. S. 1988. “ Patch Use as an Indication of Habitat Preference, Predation Risk, and Competition. ” Behavioral Ecology and Sociobiology 22: 37 – 47. Brown, J. S. 1999. “ Vigilance, Patch Use and Habitat Selection: Foraging under Predation Risk. ” Evolutionary Ecology Research 1: 49 – 71. Choi, L., J. D. Blume, and W. D. Dupont. 2015. “ Elucidating the Foundations of Statistical Inference with 2 × 2 Tables. ” PLoS One 10: e0121263. Connell, J. 1971. “ On the Role of the Natural Enemies in Preventing Competitive Exclusion in some Marine Animals and in Rain Forest Trees. ” In Dynamics of Populations 298 – 312. Wageningen: Centre for Agricultural Publishing and Documentation. Conner, L. M., and I. A. Godbois. 2003. “ Habitat Associated with Daytime Refugia of Fox Squirrels in a Longleaf Pine Forest. ” The American Midland Naturalist 150: 123 – 29. Dey, D. C., D. Jacobs, K. McNabb, G. Miller, V. Baldwin, and G. Foster. 2008. “ Artificial Regeneration of Major oak (Quercus) Species in the Eastern United States — A Review of the Literature. ” Forest Science 54: 77 – 106. George, B., S. Seals, and I. Aban. 2014. “ Survival Analysis and Regression Models. ” Journal of Nuclear Cardiology 21: 686 – 694. G omez, J. M., E. W. Schupp, and P. Jordano. 2019. “ Synzoochory: The Ecological and Evolutionary Relevance of a Dual Interaction. ” Biological Reviews 94: 874 – 902. Green, A. J., C. Baltzinger, and ́ A. Lovas-Kiss. 2022. “ Plant Dispersal Syndromes are Unreliable, Especially for Predicting Zoochory and Long-Distance Dispersal. ” Oikos 2022: e08327. Greenberg, C. H., T. L. Keyser, S. J. Zarnoch, K. Connor, D. M. Simon, and G. S. Warburton. 2012. “ Acorn Viability Following Prescribed Fire in Upland Hardwood Forests. ” Forest Ecology and Management 275: 79 – 86. Haas, J. P., and E. J. Heske. 2005. “ Experimental Study of the Effects of Mammalian Acorn Predators on Red Oak Acorn Survival and Germination. ” Journal of Mammalogy 86: 1015 – 21. Janzen, D. H. 1970. “ Herbivores and the Number of Tree Species in Tropical Forests. ” The American Naturalist 104: 501 – 528. Kaplan, E. L., and P. Meier. 1958. “ Nonparametric Estimation from Incomplete Observations. ” Journal of the American Statistical Association 53: 457 – 481. Koprowski, J. L. 1994. “ Sciurus niger ” Mammalian Species 479: 1 – 9. ECOLOGY 7 of 8 19399170, 2026, 10, Downloaded from https://esajournals.onlinelibrary.wiley.com/doi/10.1002/ecy.70570 by University Of Florida, Wiley Online Library on [29/10/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Martínez, I., F. Gonz alez Taboada, T. Wiegand, and J. R. Obeso. 2013. “ Spatial Patterns of Seedling-Adult Associations in a Temperate Forest Community. ” Forest Ecology and Management 296: 74 – 80. Nathan, R., and R. Casagrandi. 2004. “ A Simple Mechanistic Model of Seed Dispersal, Predation and Plant Establishment: Janzen-Connell and beyond. ” Journal of Ecology 92: 733 – 746. Palmer, M. S., K. M. Gaynor, J. A. Becker, J. O. Abraham, M. A. Mumma, and R. M. Pringle. 2022. “ Dynamic Landscapes of Fear: Understanding Spatiotemporal Risk. ” Trends in Ecology & Evolution 37: 911 – 925. Peterson, N. B., and V. T. Parker. 2016. “ Dispersal by Rodent Caching Increases Seed Survival in Multiple Ways in Canopy-Fire Ecosystems. ” Ecology and Evolution 6: 4298 – 4306. Potash, A. 2026. “ Figure 1. Created in BioRender. ” https:// BioRender.com/dimr1cu. Potash, A. D., L. M. Conner, M. Clinchy, L. Y. Zanette, and R. A. McCleery. 2023. “ Prey Species Increase Activity in Refugia Free of Terrestrial Predators. ” Oecologia 201: 661 – 671. Potash, A. D., L. M. Conner, M. Lashley, M. Cherry, A. J. Edelman, E. Garrison, D. U. Greene, et al. 2026. “ Biogeography and Anthropogenic Development Shape Mesopredator Distributions in a Rapidly Developing Region. ” Journal of Biogeography 53: e70147. Potash, A. D., L. M. Conner, and R. A. McCleery. 2019. “ Vertical and Horizontal Vegetation Cover Synergistically Shape Prey Behaviour. ” Animal Behaviour 152: 39 – 44. Potash, A. D., and R. A. McCleery. 2026. “ Data and Code From: Predation Risk Alters Prey Caching Behavior and Resultant Tree Establishment Patterns. ” [Dataset]. Figshare. https://doi. org/10.6084/m9.figshare.33289692.v2. R Core Team. 2024. R: A Language and Environment for Statistical Computing . Vienna, Austria: R Core Team. Schmitz, O. J., C. C. Wilmers, S. J. Leroux, C. E. Doughty, T. B. Atwood, M. Galetti, A. B. Davies, and S. J. Goetz. 2018. “ Animals and the Zoogeochemistry of the Carbon Cycle. ” Science 362: 12. Sipes, A. R., N. I. Lichti, and R. K. Swihart. 2013. “ Acorn Germination is Not Enhanced Near Cache Sites Relative to Random Locations. ” Canadian Journal of Zoology 91: 529 – 532. Stapanian, M. A., and C. C. Smith. 1984. “ Density-Dependent Survival of Scatterhoarded Nuts: An Experimental Approach. ” Ecology 65: 1387 – 96. Steele, M., G. Rompre, J. Stratford, H. Zhang, M. Suchocki, and S. Marino. 2015. “ Scatterhoarding Rodents Favor Higher Predation Risks for Cache Sites: The Potential for Predators to Influence the Seed Dispersal Process. ” Integrative Zoology 10: 257 – 266. Steele, M., L. A. Wauters, and K. W. Larsen. 2005. “ Selection, Predation, and Dispersal of Seeds by Tree Squirrels in Temperate and Boreal Forests: Are Tree Squirrels Keystone Granivores? ” In Seed Fate: Predation, Dispersal, and Seedling Establishment , edited by P.-M. Forget, J. F. Lambert, P. E. Hulme, and S. B. Vander Wall, 205 – 222. Wallingford: CAB International. Steele, M. A., T. A. Contreras, L. Z. Hadj-Chikh, S. J. Agosta, P. D. Smallwood, and C. N. Tomlinson. 2014. “ Do Scatter Hoarders Trade Off Increased Predation Risks for Lower Rates of Cache Pilferage? ” Behavioral Ecology 25: 206 – 215. Steele, M. A., X. Yi, and M. V. Mazzamuto. 2020. “ Squirrel-Seed Interactions: The Evolutionary Strategies and Impact of Squirrels as Both Seed Predators and Seed Dispersers. ” Frontiers in Ecology and Evolution 8: 1 – 11. Vander Wall, S. B. 2001. “ The Evolutionary Ecology of Nut Dispersal. ” Botanical Review 67: 74 – 117. Woods, K., and S. Elliott. 2004. “ Direct Seeding for Forest Restoration on Abandoned Agricultural Land in Northern Thailand. ” Journal of Tropical Forest Science 16: 248 – 259. SU PP O R TI N G I N F O RMA TI O N Additional supporting information can be found online in the Supporting Information section at the end of this article. How to cite this article: Potash, Alex D., and Robert A. McCleery. 2026. “ Predation Risk Alters Prey Caching Behavior and Resultant Tree Establishment Patterns. ” Ecology 107(10): e70570. https://doi.org/10.1002/ecy.70570 8 of 8 POTASH and M C CLEERY 19399170, 2026, 10, Downloaded from https://esajournals.onlinelibrary.wiley.com/doi/10.1002/ecy.70570 by University Of Florida, Wiley Online Library on [29/10/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License