Forest Ecology and Management 502 (2021) 119713 Available online 11 October 2021 0378-1127/© 2021 Elsevier B.V. All rights reserved. Facultative seed predators drive community-level indirect effects of mast seeding C. Moriah Boggess a , David S. Mason b , * , Heather D. Alexander c , Bronson K. Strickland a , Marcus A. Lashley b a Department of Wildlife, Fisheries, and Aquaculture, Mississippi State University, MS 39762, United States b Department of Wildlife Ecology and Conservation, University of Florida, Gainesville, FL 32611, United States c School of Forestry and Wildlife Sciences, Auburn University, Auburn, AL 36849, United States A R T I C L E I N F O Keywords: Resource pulse Janzen-Connell hypothesis White-tailed deer Odocoileus virginianus Acorn addition Oak masting A B S T R A C T Mast seeding is a resource pulse that is thought to be a strategy in plants to satiate obligate seed predators. Behavioral responses of facultative consumers receive less attention in mast seeding studies despite evidence that they may be more important to community-level indirect effects. We designed an acorn addition experiment to test the hypothesis that mast seeding in oaks ( Quercus spp.) generates indirect effects on the plant community by generating a behavioral response in white-tailed deer ( Odocoileus virginianus ) — the largest herbivore in the ecosystem and a facultative seed predator. Deer activity at masting trees increased substantially for months following acorn addition. Deer may require multiple dietary sources to satisfy energetic and nutritional demands and are consumers of oaks at multiple life stages. As such, acorn addition generated opposite density-dependent effects on seeds and seedlings. In acorn addition plots, we documented increased seed survival but decreased growth of oak seedlings relative to a common competitor. Importantly, acorn additions also resulted in different local plant assemblages than those at non-masting parent trees, which suggests that beta diversity may increase with localized variation in herbivory pressure. Overall, our results indicate that oaks may promote coexistence in the plant community by modulating spatiotemporal variation in herbivory pressure. These results highlight the important role of facultative consumers in linking resource pulses to the broader community. 1. Introduction A resource pulse is a low-frequency, short-duration increase in resource availability concentrated in space and time (Ostfeld and Keesing 2000, Yang et al. 2008). Although relatively rare, resource pulses can have unique behavioral consequences for individual con- sumers, stimulate numerical responses at the population level, and generate indirect effects at the community level (Yang et al. 2008). Resource pulses commonly beget bottom-up effects, which are followed by delayed top-down effects (Ostfeld and Keesing 2000). These effects can be indirect and arise from various mechanisms but often result from numerical or behavioral responses of facultative consumers or their facultative predators because these organisms interact with a large portion of the community and can diet switch when the resource is depleted (see Yang et al. 2008). Mast seeding is a resource pulse resulting from a reproductive strategy employed by many perennial plants. In this strategy, the seed production cycle is synchronized within populations and highly variable from year to year (Crone and Rapp 2014, Kelly 1994). As an important wildlife food source, years of high yield in mast-seeding plants generate effects similar to other types of resource pulses (Ostfeld and Keesing 2000, Yang et al. 2008). For example, mast seeding in oaks ( Quercus spp.) and southern beeches ( Nothofagus spp.) causes a numerical in- crease in many small mammals that attract avian and mammalian predators, strengthening apparent competition between those small mammals and alternative prey of their predators (e.g., Elliott et al. 1996; King 1983; Murphy and Dowding 1995; O ’ Donnell and Phillipson 1996; O ’ Donnell 1996; Schmidt 2003; Schmidt and Ostfeld 2003, 2008). In those examples, the observed indirect effects were a result of diet switching by a facultative consumer. Oaks have a high potential to generate indirect effects that structure communities because their seeds (i.e., acorns) are an important food for a wide array of taxa that are facultative consumers of acorns (McShea et al. 2007). Most studies on mast seeding focus on the evolutionary significance * Corresponding author. E-mail address: masond@ufl.edu (D.S. Mason). Contents lists available at ScienceDirect Forest Ecology and Management journal homepage: www.elsevier.com/locate/foreco https://doi.org/10.1016/j.foreco.2021.119713 Received 9 August 2021; Received in revised form 14 September 2021; Accepted 18 September 2021 Forest Ecology and Management 502 (2021) 119713 2 of this phenomenon rather than characterizing community-level indirect effects. The predator satiation hypothesis (PSH) is the most widely tested hypothesis providing an evolutionary explanation of mast seeding (e.g., Bogdziewicz et al. 2018a,b; Greenberg and Zarnoch 2018; Kelly 1994). The PSH posits that synchronous conspecific seeding cycles starve specialized seed predators during poor production years and then overwhelm those predators during masting events, leading to positive density-dependent seed survival (Janzen 1971, Silvertown 1980). Research on oak masting ecology has primarily explored how mecha- nisms associated with mast seeding affect oak fitness through this pos- itive density-dependent mechanism (Bogdziewicz et al. 2018a,b; Crawley and Long 1995; Greenberg and Zarnoch 2018; Sork 1993). At some threshold, which may be caused by handling time (i.e., time taken for a consumer to handle a food item) or satiety, increased acorn density causes deaccelerating intake by seed predators, and the linear rela- tionship plateaus in a type II or III functional response (Fletcher et al. 2010). However, the same functional response driven by seed density on facultative seed predators that are simultaneously herbivores can cause negative effects on seedlings near the parent tree (Janzen, 1970; Con- nell, 1971; Zhang et al., 2020). That seed predator populations crash after years of more mast pro- duction is fundamental to the PSH but does not fit well populations of generalist facultative seed predators. White-tailed deer ( Odocoileus vir- ginianus ; hereafter deer) abundance is relatively independent of masting cycles. One of the most important vertebrate predators of acorns in North America (Steiner 1996), deer have fall diets primarily composed of acorns during good mast years (Harlow et al. 1975, McCullough 1985, McShea and Schwede 1993, Steiner 1996). To endure periods of poor mast production, deer readily diet switch or change behavior when acorns are scarce (Harlow et al. 1975, McCullough 1985). For example, McShea and Schwede (1993) reported that deer expanded and shifted their home ranges to oak-hickory ( Carya spp.) forests in fall during years when mast production was high but favored other food types in poor acorn production years. Even with readily digestible, high-energy forage or mast available, deer may select other food items depending on nutritional needs (Dykes et al., 2020; Provenza et al., 2003). As such, deer likely have a strong effect on oak fitness and the broader plant community by consuming acorns and seedlings of many species, espe- cially because deer remain abundant during poor mast years (Harlow et al., 1975; Lashley et al., 2011; Steiner 1996). We designed an acorn addition experiment to test a series of hy- potheses regarding oak masting effects on consumer activity, predator satiation, and local plant communities. First, we used camera traps to monitor deer activity to test the hypothesis that acorn additions near parent trees would cause a shift in consumer space use. Second, we marked acorns and tracked their fate to test the hypothesis that masting generates positive density-dependent seed survival. Third, we trans- planted oak and blackgum ( Nyssa sylvatica ) seedlings near parent trees with and without acorn additions to test the hypothesis that masting (seed density) would generate either negative growth in oak seedlings or strengthen deer-mediated apparent competition with blackgum (i.e., oak mast seeding attracts deer, a shared predator, that suppresses rela- tive growth of competing blackgum seedlings). We selected blackgum because this species is a preferred deer browse and a common compet- itor of oak seedlings in eastern forests (Iverson et al. 2008, Lashley et al. 2011, Nowacki and Abrams 2008). Finally, we characterized the local plant community near parent trees with and without acorn additions to test the hypothesis that mast seeding would elevate local herbivory pressure, thus providing a spatiotemporal niche opportunity for her- bivory tolerant plants and causing understory plant communities to diverge in composition. 2. Materials and methods 2.1. Study site The experiment was conducted on Spirit Hill Farm, a 492 ha property in Tate and Marshall counties in northwestern Mississippi, USA. This region is characterized by the Loess Hills, with soils consisting of pri- marily wind-blown material deposited east of the Mississippi Alluvial Valley containing silt and silt loam. Average annual precipitation and temperature are 145.9 cm and 15.1 ◦ C, respectively; average low and high temperatures are 7.9 and 22.3 ◦ C, respectively (U.S. Climate Data, 2020). The property includes ~140 ha of temperate mixed upland hardwood forest with an overstory ( > 20 cm diameter at breast height; DBH) basal area of 21.3 m 2 /ha composed primarily of white oak ( Quercus alba ), Shumard oak ( Quercus shumardii ), black oak ( Quercus velutina ), southern red oak ( Quercus falcata ), post oak ( Quercus stellata ), and hickory ( Carya spp.). The midstory (10 – 20 cm DBH) is composed primarily of sweetgum ( Liquidambar styraciflua ), blackgum, red maple ( Acer rubrum ), and winged elm ( Ulmus alata ). Forest stands are separated by mixed bahiagrass ( Paspalum notatum ) pastures and row crop agriculture. 2.2. Experimental design In September 2018, 50 southern red oak trees with DBH ’ s ranging from 29 to 90 cm were selected as parent trees for masting plots within four upland hardwood stands that were 300 – 615 m apart. Each tree was paired with another individual tree similar in DBH and growing within 50 m, making 25 parent tree pairs. To estimate the average area of natural seedfall in these forest stands, we measured crown diameters of 10 random southern red oaks (not including selected parent trees) in the same forest at a random azimuth and again at a 90 ◦ angle to the original measurement. Circular plots equal in diameter to the estimated stand mean tree crown diameter (11 m diameter) were established under the 50 parent trees, with one edge of each plot bordering the main stem and the opposite edge of the plot extending out beyond the crown dripline (Fig. 1). To determine which tree of the parent tree pair was to receive acorn additions, rapid mast surveys were conducted in October 2018 to estimate the relative abundance of acorns in the total acorn-producing tree crown above each parent tree ’ s plot (Koenig et al. 1994). Because parent trees were in upland hardwood stands with a prominent oak Fig. 1. Experimental plots were 11 m in diameter, based off average southern red oak crown diameter in the forest. Plots were placed directly below parent trees with the main stem of the parent southern red oak ( Quercus falcata ) on the plot edge. Two blackgum ( Nyssa sylvatica ) and two Shumard oak ( Quercus shumardii ) seedlings were planted within the dripline portion of the plot. A camera trap was placed on the parent tree overlooking the plot, and a pin flag was placed 5 m in front of the camera; deer detections were recorded within 5 m to standardize area observed by cameras. C.M. Boggess et al. Forest Ecology and Management 502 (2021) 119713 3 component, the total acorn-producing crown over each plot often included non-parent oak trees. Based on these surveys, the parent tree in each pair with less relative acorn abundance was assigned to be the control (i.e., no acorn addition). Mast production is associated with genetic variation rather than microsite conditions, so we assumed that the distribution of control oaks was random in regards to microsite variations that could affect understory assemblages and animal activity (Brooke et al., 2019; Lashley et al., 2009). In October 2018 two blackgum (basal diameter = 6.7 ± 0.1 mm, stem length = 763.5 ± 25.6 mm) and two Shumard oak (basal diameter = 5.6 ± 0.1 mm, stem length = 433.5 ± 13.1 mm) 16 month old greenhouse-reared seedlings were transplanted into each plot. Seedlings were purchased from Mossy Oak Nativ Nurseries in West Point, Mis- sissippi, and had been reared in 1200 cm 3 containers underneath 30% shade cloth, receiving 20 minutes of mist daily to maintain soil moisture. Seedlings were planted in random spatial order and configurations 1 m apart and marked with wire loop tags with unique identification numbers. One week after seedlings were planted, a camera trap (Bush- nell Trophy Cam HD, manufactured in Thailand) on a 1 minute photo delay and normal PIR sensor mode was deployed 0.5 m up the main stem of each parent tree and parallel to the ground to monitor deer use of plots throughout the dormant season, prior to acorn germination. Pin flags were placed 5 m in front of cameras to establish a set camera trap detection area to standardize area monitored by cameras in each plot and ensure that deer within the area would trigger cameras. All natu- rally established saplings were removed between cameras and the 5 m pin flags to standardize detection opportunities for each camera. Seventy-five thousand Shumard oak acorns were collected across the Mississippi State University campus and Starkville, Mississippi, USA from mid-October to mid-November 2018 and stored in burlap bags in refrigerated storage at 4 ◦ C until use. Average acorns per liter ( x = 132) was measured by counting acorns per liter from 10 different cohorts of acorns collected; this estimate was also later used to approximate the number of acorns added to plots. Percent viable acorns (72%) was estimated by sampling 10 random samples of 10 acorns from 10 collected cohorts of acorns (n = 1000) using the acorn viability in- spection method described by Morina et al. (2017) to remove nonviable acorns. Over the first three weeks of November 2018, 3000 acorns were distributed under each of the selected acorn addition parent trees to achieve a density of 31.5 acorns per crown m 2 , typical of average mast crops by red oak species (Auchmoody et al. 1993). Acorns were haphazardly distributed by hand every seven days while standing at plot center. Five hundred acorns were added the first week, two thousand the second week, and five hundred the third, to simulate a typical distri- bution of seedfall (Burns et al. 1954). Control plots were also visited for a similar amount of time on acorn addition days to avoid any biases among plots generated by human visitation. To monitor acorn removal rates, 30 marked acorns were distributed in a 2 m 2 area in the center of all plots on the first day of acorn additions. Acorns were marked by painting a black dot on the abscission point beneath the cupule using a permanent marker prior to distribution, a standard technique for monitoring acorn depredation (Brooke et al. 2019). 2.3. Data collection Two-minute standardized searches were conducted in each plot in late March 2019 to recover marked acorns. Total stem lengths of planted seedlings were measured in March and again in September. To calculate stem length growth for each seedling species while accounting for variation in initial individual seedling stem lengths, the difference in stem length between March and September was divided by initial March stem length. This value was averaged between the two seedlings by species for each plot. The blackgum average was subtracted from the oak average in each plot to calculate the effect of acorn addition on relative growth of oak and blackgum. Time, date, animal location, species, sex, age, posture, and feeding status were recorded for all camera trap im- ages. The intensity of deer use was estimated for each plot by summing the number of deer observed feeding in images within 5 m of the camera during the experimental period (03 November 2018 – 29 March 2019, 146 days). Additionally, during August 2019, point-intercept vegetation sur- veys were conducted in all plots. Surveys consisted of two 10-m long vegetation transects situated at random perpendicular azimuths in each plot. Using a 1.5 m Robel pole, all plants intersecting the pole at 0.5 m intervals along each transect were recorded by species. Each detection was further classified by the following growth forms: forbs, trees/ shrubs, vines, and graminoids (rushes, sedges, grasses), and total de- tections of each growth form were summed for each plot. 2.4. Data analysis All analyses were conducted in R version 3.5.1 and R studio version 1.2.5001 (R Development Core Team 2017). The MASS package was used to conduct generalized linear models with a negative binomial distribution to examine the relationship between acorn addition and four response variables: total deer feeding detections, raccoon ( Procyon lotor ) detections, eastern grey squirrel ( Sciurus carolinensis ) detections, and all seed predator detections other than deer (Venables and Ripley 2002). Negative binomial distributions were used because of over- dispersion in the data, which had no upper limit. The lme4 package was used to create a binomial generalized linear model with tree pair as a random effect to compare apparent acorn survival between treatments using the total number of marked acorns recovered in each plot (Bates et al. 2015). While acorn detection during timed searches was likely not 100%, detection probability should be constant between treatments, yielding an accurate index of acorn survival for both treatments. A linear mixed model was created using the lme4 package with tree pair as a random effect to test treatment effects on relative stem length growth for oak, blackgum, and oak-blackgum (Bates et al. 2015). Eleven seedlings (eight blackgum and three oak) were excluded from the analysis because they were girdled by eastern cottontail rabbits ( Syl- vilagus floridanus ) at the root collar and experienced considerably more stem growth than those not girdled by rabbits. We analyzed differences in seedling survival using a binary logistic model including treatment, species, and the interaction term as predictor variables. In cases where seedlings of both species survived, growth rate was averaged between the seedlings. If only one seedling survived, the relative growth rate for that individual was used. At some plots, no seedlings survived for a single species, and at one plot no seedlings of either species survived. As a result, we dropped 9 pairs, leaving 16 pairs for comparison. Generalized linear models were constructed using the base stats package with a Poisson distribution to test for treatment effects on coverage by different growth forms in each plot (R Development Core Team 2017), and the mvabund package was used to run generalized linear models with a Poisson distribution to test for differences in plant communities between treatments (Wang et al. 2019). After developing the models, significance was assessed using an anova (stats package) to generate an analysis of deviance table with adjusted p values (R Development Core Team 2017). Finally, the vegan package was used to create species accumulation curves with rarefaction (Oksanen et al. 2019). The curves were then plotted with standard deviation using the ggplot2 package (Wickham 2016). 3. Results In support of our hypothesis that acorn additions would cause a change in behavior of consumers, deer use was 441% greater in acorn addition than control plots (z = 7.202, p < 0.001, Fig. 2A and B). We did not observe a relationship between treatment and detections of raccoon (z = 1.079, p = 0.281), eastern grey squirrel (z = 1.120, p = 0.263), and all non-deer seed predators combined (z = 0.712, p = 0.477): bluejay C.M. Boggess et al. Forest Ecology and Management 502 (2021) 119713 4 ( Cyanocitta cristata ), eastern grey squirrel, mice ( Peromyscus spp.), southern flying squirrel ( Glaucomys volans ), fox squirrel ( Sciurus niger ), Virginia opossum ( Didelphis virginiana ), raccoon, wild turkey ( Meleagris gallopavo ). In support of our hypothesis that acorn additions would cause positive density-dependent seed survival at the seed life stage, apparent survival of marked acorns was greater in acorn addition (3% recovered) than in control plots (0.26% recovered; z = 3.388, p < 0.001). From the 200 seedlings planted, the number of surviving blackgum seedlings (n = 86) was 1.4 times greater than the number of surviving oak seedlings (n = 63). However, the difference in survival was not explained by acorn additions. In fact, between treatments, survival was nearly equal (n = 75 for treatment, n = 74 for control). Our binary lo- gistic regression for seedling survival was significant ( χ 2 = 14.67349, p = 0.002). The model indicated that species influenced seedling survival ( χ 2 = 12.38, p < 0.001), but treatment ( χ 2 = 0.50, p = 0.480) and the interaction term did not ( χ 2 = 1.20, p = 0.272). Relative growth of oak seedlings was not affected by acorn treat- ment, but the mean tended to be lower in acorn addition compared to control plots (t = 1.323, p = 0.203). Although relative growth of blackgum seedlings was not affected by acorn treatment, the mean tended to be higher in acorn addition plots (t = 1.026, p = 0.311). When considering the collective effects of acorn addition on blackgum and oak seedlings relative to one another, our data indicate that acorn addition tended to weaken oak growth relative to blackgum (t = 2.038, p = 0.057, Fig. 3). We detected 63 plant species in plant community surveys: 20 trees/ shrubs, 9 vines, 26 forbs, and 8 graminoids. Control plots contained 52 species, of which 21 were unique to treatment; acorn addition plots contained 42 species, including 11 unique to treatment. In support of our hypothesis (Fig. 2 C), community composition differed in acorn addition from control plots (LRT = 122.6, p = 0.036). This pattern appeared to be driven by changes in the detections of forbs (z = 2.373, p = 0.017) and graminoids (z = 3.747, p < 0.001), with control plots having 59% more forb detections and 31% more detections of grami- noids than acorn addition plots. We did not observe any effect of acorn addition on detections of vines (z = 0.173, p = 0.863) or trees/shrubs (z = 1.241, p = 0.215). As a result of these local changes in the plant community, the overall accumulation of species (i.e., beta diversity) across treatments was greater than with either treatment alone (Fig. 2D). 4. Discussion Our data support the idea that facultative consumers play an essen- tial role in linking resource pulses to the broader plant community. Consumer interactions are important to the structure and function of food webs and have been implicated as mediators promoting coexis- tence mechanisms (e.g., Hanley and Sykes 2009, Ishii and Shimada 2012). While the underlying interactions vary among ecosystems, by interacting with a large portion of the community, facultative consumers may mediate the indirect effects of resource pulses on other trophic levels commonly (Yang et al. 2008). In our case, oak mast seeding may promote coexistence by generating spatiotemporal variation in deer herbivory that broadly affects plant diversity in upland oak-hickory hardwood forests. The bottom-up effect of acorn additions in our study caused a behavioral response by deer, driving indirect effects on the community. Fig. 2. (A) Conceptual depiction of deer use as a function of masting. (B) Average deer feeding detections by week of year for both treatments. (C) Conceptual map of plant cover based off masting and deer activity. (D) Species accumulation curves for understory vegetation transects at control plots, feeder plots, and both combined. White-tailed deer activity is concentrated around masting trees, leading to increased weekly detections (B). Likely resulting from the increased herbivory pressure or biopedturbation associated with increased deer activity, graminoid and forb detections are lessened beneath masting trees (C). These changes in composition create heterogeneity, which increases predicted understory richness in accumulation curves (D). C.M. Boggess et al. Forest Ecology and Management 502 (2021) 119713 5 This behavioral response was comparable to that described by McShea and Schwede (1993), who observed increased intensity of deer use in an oak-hickory forest during mast years. We considered two mechanisms by which deer could affect plant community assemblages. The first mechanism is through biopedturbation, a soil disturbance caused by organisms that may promote heterogeneity in plant communities via patchiness of soil conditions (Whitford and Kay 1999). This bio- pedturbation could have arisen from increased deer use in treatment plots disturbing leaf litter and soil through trampling and likely would increase the abundance of annual plants germinating from the seed bank. The second mechanism is a top-down effect produced via selective herbivory by deer, which can alter plant community composition (Russell et al., 2001; Westlake et al., 2020) and would primarily affect perennial plants persisting through winter when masting elevated deer use. Although our experiment was not designed to distinguish between the relative influence of biopedturbation and herbivory on understory assemblages, our data may indicate that acorn additions strengthened the top-down effect of deer herbivory on plant communities because plant community changes were driven primarily by perennial plants rather than annual plants. Although our mast production did not affect seedling survival, we documented increased relative growth rate of blackgum compared to oaks which may have been associated with increased herbivory pres- sure. Increased herbivory over long periods could shift relative competitiveness between browse tolerant and intolerant seedlings (e.g., oaks are less tolerant of herbivory than other seedlings) ultimately leading to community changes over time. Moreover, the reduction in species richness we observed in acorn addition plots was consistent with Abrams and Johnson (2012), who reported a significant increase in species richness when deer were excluded from plots in southeastern Pennsylvania. Usually changes in species richness are driven by the ef- fects of herbivory on tree seedlings, shrubs, and vines in other studies focused on deer (Abrams and Johnson 2012, Gill and Beardall 2001), but we only observed changes in perennial graminoids and forbs. Importantly, differences we observed could have stemmed from the short term effects of our acorn addition experiment, or be a long term consequence of heightened herbivory around oaks that tend to consis- tently produce the majority of acorns (Brooke et al., 2019; Lashley et al., 2009). Similarly, Barrere et al. (2020) reported increases in the magnitude of oak mast crops, and the acorn percentage in red deer ( Cervus elaphus ) diet shifted from coniferous browse toward grasses. If masting affected diet selection similarly in our study, the patterns in the plant community we observed may be in part explained by diet shifts in the herbivore. The functional response of deer also has direct implications for oak propagule success through acorn depredation and herbivory on seed- lings. Despite the significant behavioral response by deer, we still observed a positive density-dependent relationship between acorn density and acorn survival. Positive density-dependence of seeds, an important element of PSH (Boucher 1985), is well supported in several families of perennial mast seeding plants across temperate and tropical forests (e.g., Boucher 1981, Crawley and Long 1995, Kelly and Sullivan 1997, Zwolak et al. 2016). Also, numerous empirical studies indicate predator satiation as a valuable strategy for acorn survival to varying degrees depending on conspecific tree density, hatching rates of acorn specialist insects, oak density, and consistency of acorn production (Bogdziewicz et al. 2018b, Crawley and Long 1995). We also observed a tendency for acorn addition to decrease growth of oak seedlings relative to blackgum, which was likely associated with increased activity and subsequently herbivory pressure near parent trees with abundant acorns. Negative density-dependence of plant propagules was proposed by Janzen (1970) and Connell (1971) in tropical forests as a mechanism to maintain tree species diversity. Our results agree with other studies indicating that evidence for the Janzen-Connell hypothesis is stronger for seedlings than for seeds (Comita et al. 2014, Connell 1971, Hyatt et al. 2003, Johnson et al. 2012). However, when an animal consumes both the seed and seedling life stages, mast seeding may simultaneously decrease per capita seed predation while increasing herbivory pressure on conspecific seedlings. Put differently, mast seeding may engender positive density-dependent effects for seeds, but the resulting aggrega- tion of herbivore activity associated with seed density may generate negative density-dependent effects for conspecific seedlings. Observed direct and indirect effects in our study have important implications for the conservation of oak ecosystems. Oaks are declining in dominance across eastern North America because (1) of a bottleneck in the regeneration cycle at the seedling stage (Abrams 1992, Aldrich et al. 2005, Alexander et al. 2021, Fei et al. 2011, McShea et al. 2007), and (2) fire exclusion creates a feedback loop that decreases the competitiveness of oaks seedlings with non-oak hardwood seedlings (i. e., mesophication, Alexander et al. 2021, Nowacki and Abrams 2008). The positive density-dependence we observed in acorn survival may be important for providing viable acorns for spring germination, but our seedling data suggest conspecific heterogeneity in seed production among masting oaks (Brooke et al. 2019) is also important for oak regeneration because trees with little mast production may be refugia from deer herbivory. Overall, the resource pulse created by masting creates fine-scale spatiotemporal variation in the strength of community interactions. Increased herbivory associated with masting decreases understory richness beneath oaks. However, this spatiotemporal varia- tion in herbivory increases overall understory richness when consid- ering both masting trees and non-masting trees, likely because of differences in browse tolerance across plant species. How those in- teractions culminate to affect oak recruitment success deserves further exploration. CRediT authorship contribution statement C. Moriah Boggess: Conceptualization, Methodology, Investigation, Writing – original draft, Writing – review & editing, Formal analysis, Data curation. David S. Mason: Visualization, Writing – review & Fig. 3. Relative growth of seedlings represents a negative effect of acorn ad- ditions on the competitive relationship of Shumard oak ( Quercus shumardii ) compared to blackgum ( Nyssa sylvatica ). Relative growth was calculated by first dividing stem length change over summer by initial stem length for all seedlings and calculating an average among seedlings of each species in all plots; the relative stem length growth of oak was then subtracted by relative stem length growth of blackgum. The difference in relative growth rate between Shumard oak and blackgum was greater in control plots than acorn addition plots, indicating negative effects on oak seedling growth generated by acorn addition. C.M. Boggess et al. Forest Ecology and Management 502 (2021) 119713 6 editing, Formal analysis. Heather D. Alexander: Supervision, Meth- odology, Writing – review & editing. Bronson K. Strickland: Supervi- sion, Formal analysis, Writing – review & editing. Marcus A. Lashley: Conceptualization, Methodology, Supervision, Writing – review & editing, Data curation, Project administration, Resources, Funding acquisition. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements This project was also supported by McIntire-Stennis Project 1024883. Special thanks to B. and S. Bowen at Spirit Hill Farm for providing our study area and wonderful research support. Thanks to L. Resop, C. Baruzzi, A. Dykstra, and Mossy Oak Nativ Nurseries for logistical support. References Abrams, M.D., 1992. Fire and the development of oak forests. Bioscience 42 (5), 346 – 353. Abrams, M.D., Johnson, S.E., 2012. Long-term impacts of deer exclosures on mixed-oak forest composition at the Valley Forge National Historical Park, Pennsylvania, USA. J. Torrey Bot. Soc. 139, 167 – 180. Aldrich, P.R., Parker, G.R., Romero-severson, J., Michler, C.H., 2005. Confirmation of oak recruitment failure in Indiana old-growth forest: 75 Years of Data. Forest Sci. 51, 406 – 416. Alexander, H.D., Siegert, C.M., Brewer, J.S., Kreye, J.K., Lashley, M.A., McDaniel, J.K., Paulson, A.K., Renninger, H.J., Varner, J.M., 2021. Mesophication and loss of oak landscapes: evidence, knowledge gaps, and future research. Bioscience 71, 531542. Auchmoody, L.R., Smith, H.C., Walters, R.C., 1993. Acorn production in northern red oak stands in northwestern Pennsylvania. Forest Service research paper (Final) (No. PB- 94-142445/XAB; FSRP-NE–680). U.S. Department of Agriculture, Forest Service, Northeastern Forest Experiment Station, Radnor, USA. Barrere, J., Boulanger, V., Collet, C., Walker, E., Siat, V., Henry, L., Saïd, S., 2020. How does oak mast seeding affect the feeding behavior of sympatric red and roe deer? Basic Appl. Ecol. 47, 83 – 94. Bates, D., Maechler, M., Bolker, B., Walker, S., 2015. Fitting linear mixed-effects models using lme4. J. Stat. Softw. 67, 1 – 48. Bogdziewicz, M., Espelta, J.M., Mu ̃ noz, A., Aparicio, J.M., Bonal, R., 2018a. Effectiveness of predator satiation in masting oaks is negatively affected by conspecific density. Oecologia 186 (4), 983 – 993. Bogdziewicz, M., Marino, S., Bonal, R., Zwolak, R., Steele, M.A., 2018b. Rapid aggregative and reproductive responses of weevils to masting of North American oaks counteract predator satiation. Ecology 99 (11), 2575 – 2582. Boucher, D.H., 1981. Seed predation by mammals and forest dominance by Quercus oleoides , a tropical lowland oak. Oecologia 49 (3), 409 – 414. Boucher, D.H., 1985. Lotka-volterra models of mutualism and positive density- dependence. Ecol. Model. 27 (3-4), 251 – 270. Brooke, J.M., Basinger, P.S., Birckhead, J.L., Lashley, M.A., McCord, J.M., Nanney, J.S., Harper, C.A., 2019. Effects of fertilization and crown release on white oak ( Quercus alba ) masting and acorn quality. For. Ecol. Manage. 433, 305 – 312. Burns, P.Y., Christisen, D.M., Nichols, J.M., 1954. Acorn production in the Missouri Ozarks. University of Missouri, College of Agriculture, Agricultural Experiment Station. Comita, L.S., Queenborough, S.A., Murphy, S.J., Eck, J.L., Xu, K., Krishnadas, M., Beckman, N., Zhu, Y., 2014. Testing predictions of the Janzen-Connell hypothesis: A meta-analysis of experimental evidence for distance- and density-dependent seed and seedling survival. J. Ecol. 102, 845 – 856. Connell, J.H., 1971. On the role of natural enemies in preventing competitive exclusion in some marine animals and in rain forest trees. Dyn. Populat. 5, 298 – 312. Crawley, M.J., Long, C.R., 1995. Alternate bearing, predator satiation and seedling recruitment in Quercus robur L. J. Ecol. 83 (4), 683. https://doi.org/10.2307/ 2261636. Crone, E.E., Rapp, J.M., 2014. Resource depletion, pollen coupling, and the ecology of mast seeding. Ann. N. Y. Acad. Sci. 1322, 21 – 34. Dykes, J.L., Strickland, B.K., Demarais, S., Reynolds, D.B., Lashley, M.A., 2020. Diet selection of white-tailed deer supports the nutrient balance hypothesis. Behav. Process. 179, 104196. https://doi.org/10.1016/j.beproc.2020.104196. Elliott, G.P., Dilks, P.J., O ’ Donnell, C.F.J., 1996. The ecology of yellow-crowned parakeets (Cyanoramphus auriceps) in Nothofagus forest in Fiordland, New Zealand. N. Z. J. Zool. 23 (3), 249 – 265. Fei, S., Kong, N., Steiner, K.C., Moser, W.K., Steiner, E.B., 2011. Change in oak abundance in the eastern United States from 1980 to 2008. For. Ecol. Manage. 262 (8), 1370 – 1377. Fletcher, Q.E., Boutin, S., Lane, J.E., LaMontagne, J.M., McAdam, A.G., Krebs, C.J., Humphries, M.M., 2010. The functional response of a hoarding seed predator to mast seeding. Ecology 91 (9), 2673 – 2683. Gill, R.M., Beardall, V., 2001. The impact of deer on woodlands: The effects of browsing and seed dispersal on vegetation structure and composition. Forestry 74, 209 – 218. Greenberg, C.H., Zarnoch, S.J., 2018. A test of the predator satiation hypothesis, acorn predator size, and acorn preference. Can. J. For. Res. 48 (2), 237 – 245. Hanley, M.E., Sykes, R.J., 2009. Impacts of seedling herbivory on plant competition and implications for species coexistence. Ann. Bot. 103, 1347 – 1353. Harlow, R.F., Whelan, J.B., Crawford, H.S., Skeen, J.E., 1975. Deer foods during years of oak mast abundance and scarcity. J. Wildl. Manag. 39 (2), 330. https://doi.org/ 10.2307/3799910. Hyatt, L.A., Rosenberg, M.S., Howard, T.G., Bole, G., Fang, W., Anastasia, J., Brown, K., Grella, R., Hinman, K., Kurdziel, J.P., Gurevitch, J., 2003. The distance dependence prediction of the Janzen-Connell hypothesis: a meta-analysis. Oikos 103 (3), 590 – 602. Ishii, Y., Shimada, M., 2012. Learning predator promotes coexistence of prey species in host-parasitoid systems. PNAS 109 (13), 5116 – 5120. Iverson, L.R., Hutchinson, T.F., Prasad, A.M., Peters, M.P., 2008. Thinning, fire, and oak regeneration across a heterogeneous landscape in the eastern U.S.: 7-year results. For. Ecol. Manage. 255, 3035 – 3050. Janzen, D.H., 1970. Herbivores and the number of tree species in tropical forests. Am. Nat. 104 (940), 501 – 528. Janzen, D.H., 1971. Seed predation by animals. Annu. Rev. Ecol. Evol. Syst. 2 (1), 465 – 492. Johnson, D.J., Beaulieu, W.T., Bever, J.D., Clay, K., 2012. Conspecific negative density dependence and forest diversity. Science 336 (6083), 904 – 907. Kelly, D., 1994. The evolutionary ecology of mast seeding. Trends Ecol. Evol. 9 (12), 465 – 470. Kelly, D., Sullivan, J.J., 1997. Quantifying the benefits of mast seeding on predator satiation and wind pollination in Chionochloa pallens (Poaceae). Oikos 78 (1), 143. https://doi.org/10.2307/3545810. King, C.M., 1983. The relationships between beech ( Nothofagus Sp.) seedfall and populations of mice ( Mus musculus ), and the demographic and dietary responses of stoats ( Mustela erminea ), in three New Zealand forests. J. Anim. Ecol. 52, 141 – 166. Koenig, W.D., Knops, J.M.H., Carmen, W.J., Stanback, M.T., Mumme, R.L., 1994. Estimating acorn crops using visual surveys. Can. J. For. Res. 24 (10), 2105 – 2112. Lashley, M.A., McCord, J.M., Greenberg, C.H., Harper, C.A., 2009. Masting characteristics of white oak: implications for management. In: Proceedings of the Annual Conference of the Southeast Association Fish and Wildlife Agencies, 63, pp. 21 – 26. Lashley, M.A., Harper, C.A., Bates, G.E., Keyser, P.D., 2011. Forage availability for white- tailed deer following silvicultural tr