Northeastern Naturalist 466 S.E. Cilles, G. Coy, C.R. Stieha, J.J. Cox, P.H. Crowley, and D.S. Maehr 22001166 NORTHEASTERN NATURALIST 2V3(o4l). :2436,6 N–4o8. 14 A Comparison of Seed Predation, Seed Dispersal, and Seedling Herbivory in Oak and Hickory: Species with Contrasting Regenerating Abilities in a Bluegrass Savanna–Woodland Habitat Sara E. Cilles1, Garnett Coy1, Christopher R. Stieha1,2,*, John J. Cox3, Philip H. Crowley1, and David S. Maehr3,† Abstract - Quercus (oak) regeneration failure threatens many forest and savanna communities worldwide, where preservation of vegetation structure and composition depends on acorns germinating and surviving into adulthood. However, predation on the acorns and browsing of seedlings limits oak regeneration. To better understand the effects of these 2 mechanisms on oak recruitment in the endangered Bluegrass savanna–woodland of Kentucky, we compared seed predation and herbivory on Quercus muehlenbergii (Chinquapin Oak) with Carya laciniosa (Shellbark Hickory), a successfully regenerating tree species. Compared to hickory nuts, acorns were predated more, cached less, and dispersed shorter distances. Neither the distribution of the seedlings under the parent canopy nor browse damage differed between the 2 species. Our results suggest that seed-predation prevents regeneration of oaks in this endangered community. Introduction The widespread failure of Quercus (oak) to regenerate threatens to alter many communities (Abrams 2003, Beck 1992, Cho and Boerner 1991, Hanberry et al. 2014, Lorimer et al. 1994, Plieninger et al. 2010) through shifts in the dominant tree species and changes to the composition of the canopy (Asbjornsen et al. 2005). Although similar shifts are occurring in communities all over the world, from the Dongling Mountains of China (Sun et al. 2004) to the Inner Bluegrass of Kentucky (Bryant et al. 1980), the causes of oak regeneration failure may differ among communities (Lorimer 1992). Of the many factors influencing oak regeneration, seed predation and browsing may be the dominant mechanisms in systems with few oak seedlings in the understory (Lorimer 1992). Understanding oak regeneration failure in these systems requires an understanding of the effects of seed predators and browsers on the seed and seedling stages, especially in comparison to a species that appears to be successfully regenerating. Seed-eating animals have both negative and positive effects on oak regeneration (Pulido et al. 2013). In oak systems around the world (Gomez et al. 2003, Haas and Heske 2005, Johnson et al. 1989, Li and Zhang 2003, Sun et al. 2004, Takahashi 1Department of Biology, University of Kentucky, 101 Morgan Building, Lexington, KY 40506-0225. 2Department of Biology, Case Western Reserve University, Cleveland, OH 44106. 3Department of Forestry, University of Kentucky, 102 TP Cooper Building, Lexington, KY 40506-0073. †Deceased. *Corresponding author - stieha@case.edu. Manuscript Editor: Thomas Philbrick Northeastern Naturalist Vol. 23, No. 4 S.E. Cilles, G. Coy, C.R. Stieha, J.J. Cox, P.H. Crowley, and D.S. Maehr 2016 467 et al. 2006), seed predators often consume nearly all of the available acorns in any given year (Lorimer 1992), but many seed-eating animals also cache seeds, which could positively influence oak regeneration by moving seeds into areas conducive to seed survival, germination, and growth (Perea et al. 2011, Sork 1983a, Stapanian and Smith 1984, Vander Wall 1990; but see Sork 1983b). The caching process decreases seed density (Steele and Koprowski 2001), protects seeds from seed predators that are attracted to areas of high seed-density (Sork 1983a), and decreases the likelihood that the seeds will be found (Borchert et al. 1989, Hulme 1994; but see Gomez et al. 2003, Johnson et al. 1997, McCarthy 1994). Quantification of seed consumption and caching can help determine both the positive and negative effects of seed-eating animals. Caching patterns by seed predators reflect (1) perishability, an important component of which is dormancy (Hadj-Chikh et al. 1996, Smallwood et al. 2001, Steele et al. 2004) and (2) nutritional value. Many seeds of species such as Carya (hickories; Graney 1990, Schlesinger 1990), Juglans (walnuts; Williams 1990), and oaks of the red oak subgenus (Erythrobalanus, Schopmeyer 1974) remain dormant through the fall and winter and germinate in the spring, whereas acorns of the white oak subgenus (Leucobalanus), such as Quercus muehlenbergii Engelm. (Chinquapin Oak), germinate in the autumn (Barnett 1977, Steele et al. 2004). Although tree squirrels can prevent germination (Wood 1938) by excising the embryo of the acorn before caching it (Fox 1982, Zhang et al. 2014), squirrels usually choose to cache dormant nuts rather than non-dormant ones (Hadj-Chikh et al. 1996, Smallwood et al. 2001, Steele et al. 2004). In addition to differences in dormancy, nuts also differ in their nutritional content. Squirrels get proportionately less energy from hickory nuts than from acorns because the hard shell of hickory nuts requires more handling time (Lewis 1982), but hickory nuts have more digestible protein, crude fat, and digestible energy than do acorns (Abrahamson and Abrahamson 1989, Lewis 1982). Despite the higher handling time, squirrels have been found to consume proportionately more hickory nuts than acorns (Lewis 1982). Due to these differences in dormancy and nutritional value, hickory nuts may be more valuable than acorns as a food source for tree squirrels. More-valuable items are cached farther from the source tree to decrease the chance of cache theft, and less-preferred items are usually cached closer to the source, often under the canopy of the parent tree (Stapanian and Smith 1984). Increased seed survival via caching can positively affect recruitment, but does not guarantee survival to adulthood. Besides seed predation, herbivory can also contribute to oak regeneration failure by leading to direct mortality or arrested development of tree seedlings. Browsing has been shown to dramatically reduce the recruitment of oak species (Marquis et al. 1976, Russell and Fowler 2004), but in some cases, even unpalatable species, such as the hickories, are heavily browsed (McCarthy 1994). Overabundance of herbivores in an area can quickly eliminate the most palatable species from communities at local to regional spatial scales (Alverson et al. 1988, Ross et al. 1970), inhibit regeneration (Lorimer 1992, Ross et al. 1970), and lead to changes in species composition within ecological communities (McEwan et al. 2011, Ripple and Beschta 2007, Smith et al. 2003). Northeastern Naturalist 468 S.E. Cilles, G. Coy, C.R. Stieha, J.J. Cox, P.H. Crowley, and D.S. Maehr 2016 Vol. 23, No. 4 The globally endangered Bluegrass savanna–woodland community was historically dominant, but now comprises less than 1% of the Inner Bluegrass physiographic region of central Kentucky (Jones 2005). The structure of this community is characterized by a mosaic of savanna, woodland, and herbaceous plant-dominated open areas. These savanna–woodland communities support various mixtures of dominant or co-dominant deciduous hardwood tree species including Fraxinus quadralangulata Michx. (Blue Ash), Fraxinus americana L. (White Ash), Quercus macrocarpa Michx. (Bur Oak), Chinquapin Oak, Quercus shumardii Buckley (Shumard Oak), and Carya laciniosa (Michx. f.) G. Don (Shellbark Hickory) (Bryant et al. 1980). Chinquapin Oak recruitment has been very minimal within this system since at least 1980 (Cilles 2008, Wharton and Barbour 1991), especially when compared with the other nut-bearing species, such as Shellbark Hickory, and Juglans nigra L. (Black Walnut). Failure of seedling establishment and recruitment by dominant canopy tree species is considered to be perhaps the greatest threat to this community (Bryant et al. 1980); therefore, its management requires a better understanding of these phenomena. We posited that oak regeneration failure in the Bluegrass savanna–woodland is due to seed predation and herbivory of seedlings. To understand oak regeneration failure in this system, we compared seed predation and browsing of Chinquapin Oak to that of Shellbark Hickory, a species that has experienced good recruitment in the savanna in the past 30 years (J. Cox, pers. observ.). Acorns have a thinner, softer shell that is penetrable by more species than hickory nuts; thus we predicted that acorns would have a greater number of predators than hickory nuts. Hickory nuts remain dormant over winter, making them a potentially more valuable longterm food source than Chinquapin acorns, which can begin germination in fall and may spoil more quickly. Therefore, we predicted that hickory nuts were more likely to be cached than acorns, and would tend to be cached farther from the parent tree than acorns. Further, we predicted that differences in cache distance would influence spatial distribution of seedlings; Chinquapin Oak seedlings would be found closer to the parent tree compared to hickory seedlings. Finally, given that Chinquapin Oak seedlings are relatively more palatable to small mammals and Odocoileus virginianus Zimmermann (White-tailed Deer) (Hannah 1987, Van Dersal 1940) as compared to Shellbark Hickory seedlings (Graney 1990), we hypothesized that we would observe greater browsing of seedlings of the former than the latter. Our work on oak regeneration failure in the Bluegrass savanna–woodland provides insight into oak regeneration failure in other systems, especially in systems with few oak seedlings in the understory (Lorimer 1992). Field-site Description The largest remnant of the Bluegrass savanna–woodland ecosystem is Griffith Woods, in Harrison County, KY (38°19'N, 84°21'W; Wharton and Barbour 1991), which consists of a small savanna (~24 ha) with characteristic Blue Ash, White Ash, and mature nut-bearing trees (3.5 Chinquapin Oaks per ha, 3.9 Shellbark Northeastern Naturalist Vol. 23, No. 4 S.E. Cilles, G. Coy, C.R. Stieha, J.J. Cox, P.H. Crowley, and D.S. Maehr 2016 469 Hickories per ha, 3.7 Black Walnuts per ha, and 3 Blue Ash Trees per ha; Cilles 2008). We determined that the average diameter at breast height (dbh) was 104.1 ± 5.1 cm for Chinquapin Oak (n = 22) and 20.9 ± 4.0 cm for Shellbark Hickory trees in a 5.2-ha area in the center of the savanna (see sapling distribution and herbivory survey in Methods). The climate of the Inner Bluegrass region is warm-temperate and humid, with an average rainfall of 112 cm (44 in) and an average temperature of 13 °C (55 °F) (Wharton and Barbour 1991). Soil in this region is deep, fertile, well-drained, and derived from phosphatic limestone. Sinkholes, springs, and caves are common throughout the rolling, karst landscape of the Inner Bluegrass region (Bryant et al. 1980, Wharton and Barbour 1991). Methods Foraging-preference experiment The fate of a nut depends on whether a seed predator or seed disperser finds it; thus, we first needed to identify the species taking nuts at our site. In late November 2006, we placed a dish of 10 Chinquapin Oak acorns and a dish of 10 Shellbark Hickory nuts side by side on the ground 2 m from the trunk of each of 10 trees. We randomly selected 5 Chinquapin Oak and 5 Shellbark Hickory trees from the large trees (>30 cm dbh) present and installed a digital remote camera (Trail Sentry, Bushnell Corporation, Overland Park, KS) at each one set to record 15 sec of video whenever its motion sensor was activated over a 10-d period. We noted the species responsible for each predation or dispersal event. Nut-fate experiment with visitor identification To quantify the effects of each documented animal species on the predation and dispersal of nuts, we tallied visitations to our plots and tracked nuts taken from the plot. We set up a 0.5 m x 0.5 m plot 2 m from the bases of the trees used in the foraging-preference experiment. We arranged 18 acorns and 18 hickory nuts in an alternating pattern in a 6 x 6 grid in each plot. To each nut, we superglued the end of a 30-m length of polyester thread wound around a plastic sewing-machine bobbin (distance based on results from Li and Zhang 2003; technique from Dennis 2003, Forget and Wenny 2005). Three-inch roofing nails with round plastic caps secured each nut–bobbin complex to the grid and allowed the bobbins to rotate freely. We aimed a digital remote-camera at each plot to record seed-predation and dispersal events. Cameras were set up on 30 October 2006 and checked daily for 2 d. The majority of nuts had been removed after the first 48 h; thus, after the initial 2 d, we checked the cameras approximately weekly until mid-January. We followed the threads until we found their ends and recorded the distance from the plots to the end of each thread. If we found a nut fragment on the end of the thread, we recorded the fate of that nut as “eaten”. If the nut was buried in soil, under leaves, or in a tuft of grass, the fate was recorded as “cached.” If no nut or fragment was attached to the end of the thread, the fate of the nut was recorded as “removed”. We also noted whether cached acorns had excised embryos. We mapped cache locations using a GPS unit with sub-meter accuracy (GeoXT, Trimble Navigation Limited, Northeastern Naturalist 470 S.E. Cilles, G. Coy, C.R. Stieha, J.J. Cox, P.H. Crowley, and D.S. Maehr 2016 Vol. 23, No. 4 Sunnyvale, CA) and checked them approximately weekly to determine if and when the nuts were recovered. To improve the normality of our data, we square-root transformed the paired difference between the number of hickory nuts and the number of acorns cached at each plot, and performed a paired t-test. We also employed a paired t-test to compare the average caching distance of acorns versus hickory nuts at each plot. We found no difference in results between plots at Chinquapin Oak trees and plots at Shellbark Hickory trees; thus we pooled the data. All statistical tests were conducted in SAS 9.1 (SAS Institute, Inc. 2003) with α = 0.05. Nut-fate experiment with caching To better quantify the fate of the nuts (Alverson and Diaz 1989), we tracked nuts using a combination of 2 methods: (1) the thread-and-bobbin method described previously and (2) magnetic location. To use magnetic location to effectively track nuts (see Alverson and Diaz 1989), we inserted a cylindrical neodymium magnet (0.32 cm tall x 0.32 cm diameter; K & J Magnetics, Jamison, PA) into each nut and tracked it using a magnetic locator (DML2000, Dunham and Morrow, Inc., Dulles, VA). In mid-April 2007, we placed 16 acorns and 16 hickory nuts in each plot associated with the Chinquapin Oak and Shellbark Hickory trees used in our previous experiments, and checked caches approximately every other day for 2 weeks, after which we checked plots and caches once per week for 3 weeks, and then once per month for 4 months because over time few nuts remained in the plots and few caches remained unrecovered. We checked plots and caches for the last time in late September 2007. We followed the threads attached to nuts removed from plots until we found the end of the thread. We categorized the fates of the nuts (eaten, cached, removed) as in the previous experiment. If no nut or fragment was attached to the end of the thread, we used the magnetic locator to systematically search the area around the plot for ~25–30 m in all directions. We classified a nut as eaten if the bare magnet was found and as removed if neither magnet nor nut was found. We employed a paired t-test to determine if more hickory nuts than acorns were cached per plot. There was no difference between plots at Chinquapin Oak trees and plots at Shellbark Hickory trees, so we pooled the data. Too few acorns were cached per plot to compare average caching distance per plot for acorns versus hickory nuts. Sapling distribution and herbivory survey To determine the effect of seed dispersers on survival, we surveyed naturally occurring saplings and their location with respect to adult trees. To determine the effect of browsers on Chinquapin Oak and Shellbark Hickory regeneration, we quantified browse damage on these saplings. We used GPS to map the location of every tree (dbh > 5 cm) and sapling at least 30 cm tall within a 5.2-ha area in the center of the Griffith Woods savanna. We recorded the height and the location (canopy or open) of each sapling, and counted the total number of stems that had been browsed. To determine whether Chinquapin Oak saplings were more likely Northeastern Naturalist Vol. 23, No. 4 S.E. Cilles, G. Coy, C.R. Stieha, J.J. Cox, P.H. Crowley, and D.S. Maehr 2016 471 than Shellbark Hickory saplings to be found under the canopy of a possible parent (a conspecific tree), we performed both a Fisher’s exact test (2 x 2 contingency table because 25% of the cells had expected values less than 5) and a Barnard’s exact test on unpooled data (Barnard 1945) using the Barnard package (Erguler 2015) in the R statistical program (R Core Team 2015). To determine whether there was a difference in the average distance between Chinquapin Oak and Shellbark Hickory saplings to their nearest possible parent trees, we performed a t-test, where the sampling unit was parent tree, and the average distances from saplings to trees were weighted by the number of offspring assigned to an individual tree. The distances from saplings to possible parent trees could be affected by differences between the spatial distributions of mature oaks and hickories; thus, we also tested whether there was a difference in the average distance between closest neighboring mature oaks and between mature hickories using a Wilcoxon 2-sample test because our data were non-normal. We used a Barnard’s exact test on unpooled data to compare the occurrence of browsing (browsing versus no browsing) and a Wilcoxon 2-sample test to compare the average number of browsed stems of Shellbark Hickory and Chinquapin Oak seedlings and saplings. Results Foraging-preference experiment Most seeds were taken from our plots during the first 24 h. The removal of 58 out of 100 hickory nuts and 43 out of 100 acorns was captured on video. More predator species took acorns than hickory nuts (Fig. 1), but most nuts, regardless of type, were taken by Sciurus niger L. (Fox Squirrel). Peromyscus leucopus Rafinesque (White-footed Mouse) was the only other predator of hickory nuts. Nut-fate experiment with visitor identification From the digital photographs, we identified 8 mammal species (Fox Squirrel, White-footed Mouse, Procyon lotor L. [Raccoon], Tamias striatus V. Bailey Figure 1. (A) Percentage of acorns (n = 43) and (B) hickory nuts (n = 58) taken by various species of seed predator in the foraging-preference experiment. Northeastern Naturalist 472 S.E. Cilles, G. Coy, C.R. Stieha, J.J. Cox, P.H. Crowley, and D.S. Maehr 2016 Vol. 23, No. 4 [Eastern Chipmunk], Urocyon cinereoargenteus Schreber [Gray Fox], White-tailed Deer, Didelphis virginiana Kerr [Virginia Opossum], and Sylvilagus floridanus J.A. Allen [Eastern Cottontail]) and 4 bird species (Pipilo erythrophthalmus L. [Eastern Towhee], Cardinalis cardinalis L. [Northern Cardinal], Zonotrichia albicollis J.F. Gmelin [White-throated Sparrow] and Junco hyemalis L. [Dark-eyed Junco]) that visited the plots (Fig. 2). In many cases, it was not possible to discern the species of nut taken or if a nut was removed. Therefore, data are displayed simply as the number of times we recorded a seed predator species at each plot. Seed predators often cut the threads attached to the nuts; thus we could determine the fates of only 32% of the nuts (48% of the acorns and 16% of the hickory nuts). An average of 42% of acorns per plot were eaten at the source and 6% were cached. In contrast, no hickory nuts were eaten at the source and 16% were cached. About 2 more hickory nuts than acorns were cached per plot (1-tailed t-test: t = 7.02, P < 0.0001; Fig. 3A), and these hickory nuts were cached an average of 2.2 m farther than acorns (1-tailed paired t-test: t = 2.52, P = 0.043) (Fig. 3B). Our sample size was small (n = 4) because only 4 plots had cached acorns. Half (7 out of 14) of cached acorns had their embryos excised. Because we rechecked caches, we found that 1 acorn and 1 hickory nut were re-cached; all cached nuts had been recovered by the predators by the end of the study. Nut-fate experiment with caching We recovered 146 acorns and 115 hickory nuts of the 320 nuts used in the experiment; 2 hickory nuts had not been removed from the plots when the experiment Figure 2. Seed-predator frequency across all nutfate experiments with visitor- identification plots. Northeastern Naturalist Vol. 23, No. 4 S.E. Cilles, G. Coy, C.R. Stieha, J.J. Cox, P.H. Crowley, and D.S. Maehr 2016 473 ended in early September 2007. An average of 8.3 more hickory nuts than acorns were cached per plot (1-tailed paired t-test: t = 6.85, P < 0.0001; Fig. 4). The average caching distance per plot for hickory nuts was 6.94 m (SE = 0.72), but only 3 out of 160 acorns were cached (at distances of 2.4 m, 0.6 m, and 0.6 m) (data not shown). One out of the 3 cached acorns had its embryo excised. All caches were either recovered or they rotted (some after germinating). Sapling distribution and herbivory survey Although we found 25% (7 out of 28) of oak saplings and ~14% (75 out of 550) of hickory saplings under the canopy of a possible parent, Chinquapin Oak Figure 3. Acorns and hickory nuts cached in nut-fate experiment with visitor identification. (A) Number of acorns and hickory nuts cached per plot (n = 4 pairs; one-tailed paired t-test: t = 7.02, P < 0.0001). (B) Average caching distance for acorns and hickory nuts (n = 4 pairs; 1-tailed paired t-test: t = 2.52, P = 0.043). Mean ± se are shown in both panels. Figure 4. Average number of acorns and hickory nuts cached per plot in the nut-fate experiment with caching (n = 5 pairs; 1-tailed paired t-test: t = 6.85, P < 0.0001). Graph shows mean ± SE. Northeastern Naturalist 474 S.E. Cilles, G. Coy, C.R. Stieha, J.J. Cox, P.H. Crowley, and D.S. Maehr 2016 Vol. 23, No. 4 saplings were no more likely than Shellbark Hickory saplings to be found under the canopy of a possible parent (Fisher’s exact test: P = 0.099; Barnard’s exact test: Wald statistic = -1.34, nuisance parameter = 0.01, P = 0.448). There was also no difference between the distance from Chinquapin Oak saplings to their nearest possible parent and Shellbark Hickory saplings to their nearest possible parent (mean ± SE ; oak distance: 25.25 ± 4.11 m; hickory distance: 20.78 ± 1.60 m; 2-tailed t-test: t = 0.68, P = 0.502). There was also no difference between the average distance of oak trees and their nearest conspecific neighbor and that of hickory trees with their nearest conspecific neighbor (mean ± SE ; oak distance: 26.08 ± 4.16 m; hickory distance: 24.96 ± 5.00 m; 2-tailed Wilcoxon 2-sample test: statistic = 566, P = 0.291). Browsing on Chinquapin Oak saplings was 2 times greater than on hickory saplings, with 3 of 28 (~11%) Chinquapin Oak and 32 of 550 (~6%) of Shellbark Hickory saplings browsed; these proportions did not differ (2-sided Barnard’s exact test on unpooled data: Wald statistic = -0.714, nuisance parameter = 0.01, P = 0.84). There was no difference in the average number of browsed stems per sapling between Chinquapin Oak and Shellbark Hickory (mean number of stems ± SE ; oak: 0.39 ± 1.71 stems per sapling; hickory: 0.13 ± 0.81 stems per sapling; Kruskal-Wallis: χ2 = 1.11, P = 0.291). Discussion The failure of oak regeneration has the potential to change the community composition of many ecosystems, such as the endangered Bluegrass savanna–woodland of central Kentucky. Mammalian herbivores and seed predators and dispersers influence the regeneration of most tree species, and this is especially true for nutbearing species that provide a nutritious reward for dispersers (Gomez et al. 2003, Lorimer 1992, McEwan et al. 2011), but the interaction between the trees and the seed-eating animals is complicated because animals can both negatively affect tree regeneration by consuming seeds and young plants and positively affect tree regeneration by dispersing and caching seeds away from the parent tree. Our study provides evidence that seed dispersal and predation, but not browsing, are the most plausible factors to explain reduction in Chinquapin Oak recruitment in this savanna–woodland remnant. When we compared Chinquapin Oak, which is not regenerating, to Shellbark Hickory, which is regenerating, we found that acorns were consumed more, cached less, and dispersed less than hickory nuts, but that there was no difference between the level of browsing on Chinquapin Oaks and Shellbark Hickory. The preferential caching of hickory nuts over acorns and the frequent immediate consumption of acorns greatly reduces the potential of oaks to regenerate, while probably increasing hickory regeneration. Therefore, management efforts to improve oak regeneration and preserve current Bluegrass savanna–woodland composition and structure should focus on increasing acorn persistence or burial. We found that acorns were more likely than hickory nuts to be eaten immediately instead of being cached. Of the acorns that were cached, 33–50% had excised embryos, which would serve to further reduce oak recruitment because most acorns with excised embryos do not germinate (Wood 1938, Zhang et al. 2014, but see Yi Northeastern Naturalist Vol. 23, No. 4 S.E. Cilles, G. Coy, C.R. Stieha, J.J. Cox, P.H. Crowley, and D.S. Maehr 2016 475 et al. 2014). The lack of acorn caching is probably due to a combination of perishability and behaviors of the suite of animals that removed them. In our study, acorns were the more perishable resource and were consumed more than hickory nuts. In temperate hardwood forests, these caching decisions are based on perishability and not palatability, where animals consume the more perishable type of nut and cache the less perishable nut when given a choice between 2 species (Lichti et al. 2014). Besides the difference in perishability, a different suite of animals removed acorns and hickory nuts. At our study site, a greater number of species removed acorns than hickory nuts, and a much larger proportion of acorns were taken by species that are not scatterhoarders (Fig. 2). Fox Squirrels, estimated at a density of about 10 individuals per km2 at Griffith Woods (J. Cox, pers. observ.), were the only scatterhoarders that visited our seed plots. They removed ~86% of hickory nuts from seed trays but only ~49% of acorns. Conversely, Raccoons removed ~21% of acorns and no hickory nuts, and were also responsible for heavy predation of acorns in the nut-fate experiment with visitor identification. Raccoons typically ate all of the acorns at a particular plot, leaving the shells on the ground within the plot. In these situations, acorns did not have the opportunity to be dispersed by accidental dropping during transport to another location. Hickory nuts were cached farther from the source than were acorns, which could increase the potential for hickory regeneration relative to oak regeneration. Cached food items are protected from desiccation (Barnett 1977, Haas and Heske 2005, Korstian 1927, Smit et al. 2009) and are at a lower density farther from the seed source, which makes them less likely to be discovered and eaten by naïve seed predators (Jansen et al. 2004, Stapanian and Smith 1984). Therefore, if a hickory nut disperser does not recover some of its caches, the possibility that the nut will survive is greater than for acorns. Consequently, we expected the greater caching distance for hickory nuts to result in a greater number of Shellbark Hickory saplings than Chinquapin Oak saplings, a situation that we observed in our sapling survey (550 hickory saplings versus 28 oak saplings). Although we also expected Chinquapin Oak saplings to be closer to potential parent trees than Shellbark Hickory saplings because of the difference in caching distance, we found no difference between the distance from Chinquapin Oak saplings and Shellbark Hickory saplings to their nearest mature conspecific individual. This finding is inconsistent with the results from our 2 dispersal experiments showing that dispersers move hickory nuts farther than acorns. It is possible that our assumption that the parent tree is the nearest mature conspecific individual is unjustifed, which would have caused us to underestimate dispersal distances, especially for a species that gets dispersed over long distances, such as hickory nuts. Also, contrary to what we expected, there was no difference between the number of browsed stems on Shellbark Hickory saplings and Chinquapin Oak saplings. Hickory foliage has been characterized as unpalatable (Graney 1990), whereas Chinquapin Oak is highly palatable to wildlife species (Hannah 1987, Van Dersal 1940). We found an overall low level of browse on both species (only 11% of Chinquapin Oak saplings and 6% of Shellbark Hickory saplings had at least 1 Northeastern Naturalist 476 S.E. Cilles, G. Coy, C.R. Stieha, J.J. Cox, P.H. Crowley, and D.S. Maehr 2016 Vol. 23, No. 4 browsed stem), probably as a result of the low density of White-tailed Deer (8/km2) at Griffith Woods (J. Cox, pers. observ.) or the higher relative abundance of other more-preferred foods. In an earlier study, a greater percentage of individuals of other tree species (e.g., Gleditsia triacanthos L. [Honey Locust], Crataegus mollis Torr. & A. Gray [Red Hawthorn], Celtis occidentalis L. [Common Hackberry], White Ash) were browsed and had many more browsed stems per sapling than did Chinquapin Oak and Shellbark Hickory saplings (Cilles 2008). Unless Whitetailed Deer or smaller herbivores (such as Microtus pennsylvanicus Ord [Meadow Vole]; Ostfield and Canham 1993) are eating entire young seedlings, it seems that herbivory does not influence Chinquapin Oak regeneration in this Bluegrass savanna– woodland remnant, but more work is needed specifically on predation of young seedlings. Based on our data, seed predators and dispersers are likely the main factors influencing tree regeneration in our study area. There may be a tendency for studies to overestimate seed predation. Human scent left on seeds may make them easier for animals to find. Duncan et al. (2002) found that nuts and fleshy fruits with human scent on them sometimes had higher removal rates than those not touched by humans. Acorns were not included in that study, but there was no difference in the removal rate of human-scented versus unscented hickory nuts; therefore, we maintain that the effects of human scent on our results were minimal. Also, many studies equate seed removal with seed predation, which greatly overestimates seed predation in species with seeds that are frequently cached. In our study, we tracked nuts when they were removed from the seed plots and classified seeds removed from primary caches as eaten unless threads were still attached and could be followed to the secondary cache. Secondary caches are often located farther from the source than primary caches (Pizo and Vieira 2004, Theimer 2001), and more secondary than primary caches produce seedlings (Pizo and Vieira 2004). In addition, seedling and sapling survival can increase with increasing distance from the parent tree (Tamura et al. 2005), which may explain why we found no difference between Chinquapin Oak and Shellbark Hickory in the average distance from sapling to possible parent tree. Collectively, our findings suggest seed predation on Chinquapin Oak is intense in this particular remnant of the Bluegrass savanna–woodland community and is at least partially responsible for the lack of regeneration of this important canopy species. Our results are consistent with those of other studies (Haas and Heske 2005, Johnson et al. 1989, Lorimer 1992) that report nearly complete consumption of acorns by seed predators, especially in communities with increased herbaceous groundcover (Kellner et al. 2016), which would be similar to the dense grasses of the Bluegrass savanna–woodland system. Others (Li and Zhang 2003, Sun et al. 2004) have concluded that seed predation and lack of successful dispersal are the primary causes of oak recruitment failure elsewhere. Because oaks are so longlived, a few successful dispersal events may have been enough to maintain the local regeneration of Chinquapin Oak, but current factors may interfere with these rare successful dispersal events to the point where local regeneration is prevented. For example, seed-predator density may be increased due to the dominance of exotic Northeastern Naturalist Vol. 23, No. 4 S.E. Cilles, G. Coy, C.R. Stieha, J.J. Cox, P.H. Crowley, and D.S. Maehr 2016 477 herbaceous groundcover, such as grasses (e.g., Festuca [Fescue] 31) and shrubs (e.g., Lonicera maackii (Rupr.) Herder [Bush Honeysuckle]), but more research is needed to evaluate this interaction. For now, managers may have to resort to the direct planting of acorns, which would reduce discovery by seed predators, and to planting oak seedlings to maintain the relative importance of Chinquapin Oak within this rare community. To match observed densities of the regenerating Shellbark Hickory and to account for an increased browsing rate on Chinquapin Oak saplings, we suggest managed planting of at least 200 acorns per ha. Acknowledgments We thank Nathan Klar, Megan Poulette, Yoriko Saeki, Kay Shenoy, and Tim Sesterhenn for discussions and suggestions; and Carol Baskin, Christopher Moore, and 2 anonymous reviewers for comments on an earlier version of this manuscript. Funding for this research was provided by the University of Kentucky Graduate School (S.E. Cilles and C.R. Stieha), the University of Kentucky Department of Biology (S.E. Cilles and C.R. Stieha), the Gertrude Flora Ribble Fund (S.E. Cilles), and a research grant from Kentucky Society of Natural History (S.E. Cilles). This work was part of a M.Sc. thesis project by S.E. Cilles in Biology at the University of Kentucky, advised by J.J. Cox, P.H. Crowley, and D.S. Maehr. Literature Cited Abrahamson, W.G., and C.R. Abrahamson. 1989. Nutritional quality of animal-dispersed fruits in Florida sandridge habitats. Bulletin of the Torrey Botanical Club 116:215–228. Abrams, M.D. 2003. Where has all the White Oak gone? Bioscience 53:927–939. Alverson, W.S., and A.G. Diaz. 1989. Measurement of the dispersal of large seeds and fruits with a magnetic locator. Biotropica 21:61–63. Alverson, W.S., D. Waller, and S. Solheim. 1988. Forests too deer: Edge effects in northern Wisconsin. Conservation Biology 2:348–358. Asbjornsen, H., L.A. Brudvig, C.M. Mabry, C.W. Evans, and H.M. Karnitz. 2005. Defining reference information for restoring ecologically rare tallgrass–oak savannas in the midwestern United States. Journal of Forestry 103:345–350. Barnard, G.A. 1945. A new test for 2 x 2 tables. Nature 156:177. Barnett, R.J. 1977. The effect of burial by squirrels on germination and survival of oak and hickory nuts. American Midland Naturalist 98:319–330. Beck, D.E. 1992. Acorns and oak regeneration. Pp. 96–104, In D.L. Loftis and C.E. Mc- Gee (Eds.). Oak Regeneration: Serious Problems, Practical Recommendations. General Technical Report SE-84. US Department of Agriculture Forest Service, Southeastern Forest Experiment Station, Asheville, NC. 319 pp. Borchert, M.J., F.W. Davis, J. Michaelsen, and L.D. Oyler. 1989. Interactions of factors affecting seedling recruitment of Blue Oak (Quercus douglasii) in California. Ecology 70:389–404. Bryant, W., M. Wharton, and J. Varner. 1980. The Blue Ash–oak savanna–woodland: A remnant of presettlement vegetation in the Inner Bluegrass of Kentucky. Castanea 45:149–165. Cho, D.S., and R.E.J. Boerner. 1991. Canopy disturbance patterns and regeneration of Quercus species in two Ohio old-growth forests. Vegetatio 93:9–18. Cilles, S.E. 2008. Impacts of vertebrate seed predators, seed dispersers, and herbivores on tree regeneration in a Bluegrass savanna. M.Sc. Thesis, University of Kentucky, Lexington, KY. 75 pp. Northeastern Naturalist 478 S.E. Cilles, G. Coy, C.R. Stieha, J.J. Cox, P.H. Crowley, and D.S. Maehr 2016 Vol. 23, No. 4 Dennis, A.J. 2003. Scatter-hoarding by Musky Rat-kangaroos, Hypsipyrmnodon moschatus, a tropical rainforest marsupial from Australia: Implications for seed dispersal. Journal of Tropical Ecology 19:619–627. Duncan, R.S., D.G. Wenny, M.D. Spritzer, and C.J. Whelan. 2002. Does human scent bias seed-removal studies? Ecology 83:2630–2636. Erguler, K. 2015. Barnard: Barnard’s Unconditional Test. R package version 1.6. Available online at http://CRAN.R-project.org/package=Barnard. Accessed 1 August 2016. Forget, P.M., and D. Wenny. 2005. How to elucidate seed fate? A review of methods used to study seed removal and secondary seed-dispersal. Pp. 379–394, In P.M. Forget, J.E. Lambert, and S.B. Vander Wall (Eds.). Seed Fate. CAB International, Cambridge, MA. 432 pp. Fox, J.F. 1982. Adaptation of Gray Squirrel behavior to autumn germination by White Oak acorns. Evolution 36:800–809. Gomez, J.M., D. Garcia, and R. Zamora. 2003. Impact of vertebrate acorn- and seedlingpredators on a Mediterranean Quercus pyrenaica forest. Forest Ecology and Management 180:125–134. Graney, D. 1990. Carya ovata (Mill.) Shagbark Hickory. Pp. 219–225, In R. Burns and B. Honkala (Technical coordinators). Silvics of North America. Volume 2. Hardwoods. Agriculture Handbook 654. US Department of Agriculture Forest Service, Washington, DC. 876 pp. 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 86:1015–1021. Hadj-Chikh, L.Z., M.A. Steele, and P.D. Smallwood. 1996. Caching decisions by Grey Squirrels: A test of the handling time and perishability hypotheses. Animal Behaviour 52:941–948. Hanberry, B.B., D.C. Dey, and H.S. He. 2014. The history of widespread decrease in oak dominance exemplified in a grassland–forest landscape. Science of the Total Environment 476:591–600. Hannah, P. 1987. Regeneration methods for oaks. Northern Journal of Applied Forestry 4:97–101. Hulme, P.H. 1994. Post-dispersal seed predation in grasslands: Its magnitude and sources of variation. Journal of Ecology 82:645–652. Jansen, P.A., L. Hemerik, and F. Borgers. 2004. Seed mass and mast-seeding enhance dispersal by a Neotropical scatterhoarding rodent. Ecological Monographs 74:569–589. Johnson, P.S., R.D. Jacobs, A.J. Martin, and E.D. Godel. 1989. Regenerating Northern Red Oak: Three successful case histories. Northern Journal of Applied Forestry 6:174–178. Johnson, W.C., C.S. Adkisson, T.R. Crow, and M.D. Dixon. 1997. Nut caching by Blue Jays (Cyanocitta cristats L.): Implications for tree demography. American Midland Naturalist 138:357–370. Jones, R. 2005. Plant Life of Kentucky. The University Press of Kentucky, Lexington, KY. 856 pp. Kellner, K.F., N.I. Lichti, and R.K. Swihart. 2016. Midstory removal reduces effectiveness of oak (Quercus) acorn dispersal by small mammals in the Central Hardwood Forest region. Forest Ecology and Management 375:182–190. Korstian, C.F. 1927. Factors controlling germination and early survival in oaks. Yale University School of Forestry Bulletin 19. 115 pp. Lewis, A.R. 1982. Selection of nuts by Gray Squirrels and optimal-foraging theory. American Midland Naturalist 107:250–257. Northeastern Naturalist Vol. 23, No. 4 S.E. Cilles, G. Coy, C.R. Stieha, J.J. Cox, P.H. Crowley, and D.S. Maehr 2016 479 Li, H.J., and Z.B. Zhang. 2003. Effect of rodents on acorn dispersal and survival of the Liadong Oak (Quercus liaotungensis Koidz.). Forest Ecology and Management 176:387–396. Lichti, N.I., M.A. Steele, H. Zhang, and R.K. Swihart. 2014. Mast-species composition alters seed fate in North American rodent-dispersed hardwoods. Ecology 95:1746–1758. Lorimer, C.G. 1992. Causes of the oak regeneration problem. Pp. 14–39, In D.L. Loftis and C.E. McGee (Eds.). Oak regeneration: Serious problems, practical recommendations. General Technical Report SE-84. US Department of Agriculture Forest Service, Southeastern Forest Experiment Station, Asheville, NC. 319 pp. Lorimer, C.G., J.W. Chapman, and W.D. Lambert. 1994. Tall understory-vegetation as a factor in the poor development of oak seedlings beneath mature stands. Journal of Ecology 82:227–237. Marquis, D.A., P.L. Eckert, and B.A. Roach. 1976. Acorn Weevils, rodents, and deer all contribute to oak-regeneration difficulties in Pennsylvania. USDA Forest Service Research Paper NE-356. Northeastern Forest Experiment Station, Newtown Square, PA. . McCarthy, B.C. 1994. Experimental studies of hickory recruitment in a wooded hedgerow and forest. Bulletin of the Torrey Botanical Club 121:240–250. McEwan, R.W., J.M. Dyer, and N. Pederson. 2011. Multiple interacting ecosystem drivers: Toward an encompassing hypothesis of oak forest dynamics across eastern North America. Ecography 34:244–256. Ostfield, R.S., and C.D. Canham. 1993. Effects of Meadow Vole population density on tree seedling survival in old fields. Ecology 74:1792–1801. Perea, R., A. San Miguel, and L. Gil. 2011. Leftovers in seed dispersal: Ecological implications of partial seed consumption for oak regeneration. Journal of Ecology 99:194–201. Pizo, M.A., and E.M. Vieira. 2004. Palm harvesting affects seed predation of Euterpe edulis, a threatened palm of the Brazilian Atlantic forest. Brazilian Journal of Biology 64:669–676. Plieninger, T., V. Rolo, and G. Moreno. 2010. Large-scale patterns of Quercus ilex, Quercus suber, and Quercus pyrenaica regeneration in central-western Spain. Ecosystems 13:644–660. Pulido, F., D. McCreary, I. Cañellas, M. McClaran, and T. Plieninger. 2013. Oak regeneration: Ecological dynamics and restoration techniques. Pp. 123–144, In P. Campos L. Huntsinger, J.L. Oviedo, P.F. Starrs, M. Diaz, R. Standiford, and G. Montero. (Eds.). Mediterranean Oak Woodland Working Landscapes. Springer, Dordrecht, The Netherlands. 508 pp. R Core Team. 2015. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Available online at http://www.R-project. org/. Accessed 1 August 2016. Ripple, W.J., and R.L. Bescheta. 2007. Hardwood tree decline following large-carnivore loss, Great Plains, USA. Frontiers in Ecology and the Environment 5:241–246. Ross, B., J. Bray, and W. Marshall. 1970. Effects of long-term deer exclusion on a Pinus resinosa forest in north-central Minnesota. Ecology 51:1088–1093. Russell, F.L., and N.L. Fowler. 2004. Effects of White-tailed Deer on the population dynamics of acorns, seedlings, and small saplings of Quercus buckleyi. Plant Ecology 173:59–72. SAS Institute, Inc. 2003. SAS Version 9.1. Cary, NC. Schlesinger, R.C. 1990. Carya laciniosa (Michx. f. Lould.), Shellbark Hickory. Pp. 211– 214, In R. Burns and B. Honkala (Technical coordinators). Silvics of North America. Volume 2. Hardwoods. Agriculture Handbook 654. US Department of Agriculture Forest Service, Washington, DC. 876 pp. Northeastern Naturalist 480 S.E. Cilles, G. Coy, C.R. Stieha, J.J. Cox, P.H. Crowley, and D.S. Maehr 2016 Vol. 23, No. 4 Schopmeyer, C.S. (Technical coordinator). 1974. Seeds of Woody Plants in the United States. US Department of Agriculture, Agriculture Handbook 450. Washington, DC. 883 pp. Smallwood, P.D., M.A. Steele, and S.H. Faeth. 2001. The ultimate basis of the caching preferences of rodents, and the oak-dispersal syndrome: Tannins, insects, and seed germination. American Zoologist 41:840–851. Smit, C., M. Díaz, and P. Jansen. 2009. Establishment limitation of Holm Oak (Quercus ilex subsp. ballota (Desf.) Samp.) in a Mediterranean savanna–forest ecosystem. Annals of Forest Science 66:1–7. Smith, D.W., R.O. Peterson, and D.B. Houston. 2003. Yellowstone after wolves. Bioscience 53:330–340. Sork, V.L. 1983a. Mammalian seed dispersal of Pignut Hickory during three fruiting seasons. Ecology 64:1049–1056. Sork, V.L. 1983b. Distribution of Pignut Hickory (Carya glabra) along a forest-to-edge transect, and factors affecting seedling recruitment. Bulletin of the Torrey Botanical Club 110:494–506. Stapanian, M.A., and C.C. Smith. 1984. Density-dependent survival of scatterhoarded nuts: An experimental approach. Ecology 65:1387–1396. Steele, M.A., and J.L. Koprowski. 2001. North American Tree Squirrels. Smithsonian Institution Press, Washington, DC. 224 pp. Steele, M.A., P. Smallwood, W.B. Terzaghi, J.E. Carlson, T. Contreras, and A. McEuen. 2004. Oak dispersal syndromes: Do red and white oaks exhibit different dispersal syndromes? Pp. 72–77, In M.A. Spetich (Ed.). Upland Oak Ecology Symposium: History, Current Conditions, and Sustainability. General Technical Report SRS-73. US Department of Agriculture Forest Service, Southern Research Station, Asheville, NC. 311 pp. Sun, S., X. Gao, and L. Chen. 2004. High acorn-predation prevents the regeneration of Quercus liaotungensis in the Dongling Mountain region of north China. Restoration Ecology 12:335–342. Takahashi, K., K. Sato, and I. Washitani. 2006. The role of the Wood Mouse in Quercus serrata acorn dispersal in abandoned cutover land. Forest Ecology and Management 29:120–127. Tamura, N., T. Katsuki, and F. Hayashi. 2005. Walnut seed dispersal: Mixed effects of tree squirrels and field mice with different hoarding ability. Pp. 241–252, In P.M. Forget, J.E. Lambert, and S.B. Vander Wall (Eds.). Seed Fate. CAB International, Cambridge, MA. 432 pp. Theimer, T.C. 2001. Seed scatterhoarding by White-tailed Rats: Consequences for seedling recruitment by an Australian rainforest tree. Journal of Tropical Ecology 17:177–189. Van Dersal, W. 1940. Utilization of oaks by birds and mammals. Journal of Wildlife Management 4:404–428. Vander Wall, S.B. 1990. Food Hoarding in Animals. University of Chicago Press, Chicago, IL. 453 pp. Wharton, M.E., and R.W. Barbour. 1991. Bluegrass Land and Life: Land Character, Plants, and Animals of the Inner Bluegrass Region of Kentucky—Past, Present, and Future. The University Press of Kentucky, Lexington, KY. 257 pp. Williams, R.D. 1990. Juglans nigra L., Black Walnut. Pp. 391–399, In R. Burns and B. Honkala (Technical coordinators). Silvics of North America. Volume 2. Hardwoods. Agriculture Handbook 654. US Department of Agriculture Forest Service, Washington, DC. 876 pp. Northeastern Naturalist Vol. 23, No. 4 S.E. Cilles, G. Coy, C.R. Stieha, J.J. Cox, P.H. Crowley, and D.S. Maehr 2016 481 Wood, O.M. 1938. Seedling reproduction of oak in southern New Jersey. Ecology 19:276–293. Yi, X., M. Zhang, A.W. Bartlow, and Z. Dong. 2014. Incorporating cache-management behavior into seed dispersal: The effect of pericarp removal on acorn germination. PLOS ONE 9(3):e92544. Zhang, M., Z. Dong, X. Yi, and A.W. Bartlow. 2014. Acorns containing deeper plumule survive better: How White Oaks counter embryo excision by rodents. Ecology and Evolution 4(1):59–66.