|USFDC Home | USF Electronic Theses and Dissertations||| RSS|
This item is only available as the following downloads:
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001795707
007 cr mnu|||uuuuu
008 070711s2006 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001580
Ansley, Shannon Elizabeth.
Secondary seed dispersal of longleaf pine, Pinus palustris, and Sand Live Oak, Quercus geminata, in Florida sandhill
h [electronic resource] /
by Shannon Elizabeth Ansley.
[Tampa, Fla] :
b University of South Florida,
ABSTRACT: Studies of secondary seed dispersal by small mammals have largely been focused on the interaction between nut-bearing tree species and sciurid rodents such as squirrels, and on heteromyid rodents in the southwestern United States. However, there is now evidence that wind-dispersed tree species such as pines also undergo a process of secondary seed dispersal, where animals redistribute (cache) seeds that have already fallen to the ground, often in microhabitats more suitable for successful seed germination. In Florida sandhill, where fire suppression has threatened wind-dispersed longleaf pine (Pinus palustris) by encouraging the encroachment of hardwoods such as sand live oak (Quercus geminata), secondary seed dispersal may be an important factor in determining community composition and persistence of longleaf pine systems. Using a combination of seed depots and seed predator exclosures, I looked at both longleaf pine and sand live oak in terms of whether small animals^ such as squirrels (Sciurus carolinensis) and cotton mice (Peromyscus gossypinus) cache the seeds, and where the seeds of these two tree species best germinate. Since sand live oak acorns are prone to infestation by weevils (Curculio spp.), I also examined whether nut condition affects acorn germination potential. I found that longleaf pine seeds are cached by small mammals to a small degree. While these seeds are not moved great distances from where they originate, they are often redistributed into microhabitats that promote successful seed germination. Caging experiments indicated that seeds were most likely to germinate when buried in open areas between adult trees, and to some degree, under shrub cover. On the other hand, sand live oak acorns appear to face heavy predation by large seed predators such as raccoons (Procyon lotor) and wild pigs (Sus scrofa). Those acorns that do escape predation, including weevil-infested acorns, may provide an opportunity for seedling establish ment. However, it appears that sand live oak depends heavily on vegetative sprouting for regeneration. This suggests that even in the absence of fire, longleaf pines in Florida sandhill are able to persist through secondary seed dispersal by small animals coupled with heavy seed predation on competing sand live oak.
Thesis (M.A.)--University of South Florida, 2006.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 48 pages.
Adviser: Gordon A. Fox, Ph.D.
t USF Electronic Theses and Dissertations.
Secondary Seed Dispersal of Longleaf Pine, Pinus palustris and Sand Live Oak, Quercus geminata in Florida Sandhill by Shannon Elizabeth Ansley A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science Department of Biology College of Arts and Sciences University of South Florida Major Professor: Gordon A. Fox, Ph.D. Susan S. Bell, Ph.D. Henry R. Mushinsky, Ph.D. Ronald J. Sarno, Ph.D. Date of Approval: April 6, 2006 Keywords: caching, predation, small mammals wind dispersal, microhabitat, weevil infestation Copyright 2006 Shannon Elizabeth Ansley
Dedicated to my mother, Janice Ansley, fo r encouraging me to pursue my dreams and enabling me to do so, and to my daughter, S ophie Ansley, that I can give her the same opportunities.
ACKNOWLEDGEMENTS I would like to thank my advisor, Gordon Fox, for all of his guidance, support, generosity, and statistical assistance. For th eir untiring patience and valuable advice, I also wish to thank my committee members: Susan Bell, Henry Mushinsky, and Ronald Sarno. An extra thanks goes to Dr. Mushin sky, for the loan of fieldwork supplies. I would like to thank the members of the Fox lab who offered much discussion, support, and friendship, includi ng Sarah Barry, Nilda Felician o, Gabe Herrick, Samantha Hord, Angela Machado, Evette Mazzo, Juliann e Robinson, Pam Tingiris, and Alia Tsang; USF faculty and staff for their assistance over the years, especially Christine Smith, who always had the time to help me, and Marilyn Whetzel, for her advice and lasting friendship; and George Kish of U.S.G.S. and the Florida Native Plant Society, for additional data collection, photos, and for giving my greenhouse seedlings a home. Thank you to all of my family and friends for unconditional love and encouragement, especially my mother, Jani ce Ansley, who took everything in stride, never questioned my decisions, a nd always believed that I co uld succeed. I owe much of this success to Matthew Carbone, who was the be st field help anyone could ever ask for. Besides contributing ideas, critical review and endless love and support, he also undertook much of the demanding physical labo ur involved in my fieldwork. Without him, this would have been a much longer (and harder) process. And my heart goes out to Lucky, who never resented all the hours I wa s away from home, and who always greeted me with a wagging tail wh en I finally returned.
i TABLE OF CONTENTS LIST OF TABLES iii LIST OF FIGURES iv ABSTRACT v INTRODUCTION 1 Diplochory 1 Seed Caching and Secondary Seed Dispersal by Small Mammals 2 The Sandhill Community 4 Objectives 6 Study Site 7 CHAPTER ONE: SECONDARY SEED DISPERSAL OF LONGLEAF PINE 8 Introduction 8 Methods 11 Caching by Small Mammals 11 Germination Success by Microhabitat and Presence of Small Mammals 12 Statistical Analyses 13 Results 14 Caching by Small Mammals 14 Germination Success by Microhabitat and Presence of Small Mammals 20 Protection: Caged vs. Uncaged 21 Habitat: Open Ground vs. Under Shrub Cover 21 Microsite: Buried vs. Soil Surface 22 The Effects of Ha bitat and Protection 22 The Effects of Habitat and Microsite 22 The Effects of Protection and Microsite 22 The Effects of Habitat, Protection, and Microsite 22 Discussion 23 CHAPTER TWO: SECONDARY SEED DISPERSAL OF SAND LIVE OAK 28 Introduction 28 Methods 30 Removal and Caching by Small Mammals 30 Germination Success by Microhabitat and Presence of Small Mammals 31 The Effects of Weevil Infestation on Removal and Germination 32
ii Transect Experiment 32 Greenhouse Experiment 32 Statistical Analyses 32 Results 33 Removal and Caching by Small Mammals 33 Germination Success by Microhabitat and Presence of Small Mammals 34 The Effects of Weevil Infestation on Removal and Germination 34 Transect Experiment 34 Greenhouse Experiment 36 Discussion 37 DISCUSSION 41 REFERENCES 43
iii LIST OF TABLES Table 1. Number and density of seeds at varying distances of depots from source tree to simulate a wind -dispersed seed shadow. 11 Table 2. Spatial pattern of longleaf pine seed caches around source trees as a function of nearestneighbor distances. 17 Table 3. Spatial pattern of longleaf pine seed caches around source trees as a function of second-near est-neighbor distances. 17 Table 4. Number and size of longleaf pine seed caches made by small mammals. 18 Table 5. Longleaf pine seed move ments and dispersal distances. 18 Table 6. The effects of grass cover and shrub cover on the fate of longleaf pine seeds. 20 Table 7. Likelihood ratio statistics for Type III analysis of the effects of fate, habitat, protection, and microsite on germination of longleaf pine seeds. 23
iv LIST OF FIGURES Figure 1. Survival distribution and hazard functions ( 1 S.E.) for longleaf pine seeds. 15 Figure 2. The locations of longleaf pine seed caches made by small mammals around each of five source trees, (a) tree one, (b) tr ee two, (c) tree three, (d) tree four and (e) tree five. 16 Figure 3. Percent cover of grass and sh rubs around source longleaf pines. 19 Figure 4. The separate effects of prot ection, habitat, and microsite on the germination of longleaf pine seeds. 20 Figure 5. The effects of ha bitat (open areas vs. under sh rub cover) and protection (uncaged controls vs. caged trea tments) on germination of buried and ground-surface longleaf pine seeds. 21 Figure 6. Survival distribution and hazard functions ( 1 S.E.) for sand live oak acorns. 33 Figure 7. Survival probabilities ( 1 S. E.) of sound and weevil-infested sand live oak acorns from transects. 35 Figure 8. Hazard function ( 1 S.E.) for sound and weevil-infested sand live oak acorns from transects. 35 Figure 9. Mean ( 1 S.E.) germination times of sound and infested sand live oak acorns. 36 Figure 10. The cumulative probability of survival to germination for sound and weevil-infested acorns. 37
v Secondary Seed Dispersal of Longleaf Pine, Pinus Palustris and Sand Live Oak, Quercus Geminata in Florida Sandhill Shannon Elizabeth Ansley ABSTRACT Studies of secondary seed dispersal by small mammals have largely been focused on the interaction betwee n nut-bearing tree species and sciu rid rodents such as squirrels, and on heteromyid rodents in the southweste rn United States. However, there is now evidence that wind-dispersed tree species su ch as pines also u ndergo a process of secondary seed dispersal, where animals re distribute (cache) seeds that have already fallen to the ground, often in microhabitats more suitable for successful seed germination. In Florida sandhill, where fire suppression ha s threatened wind-dispersed longleaf pine ( Pinus palustris ) by encouraging the encroachment of hardwoods such as sand live oak ( Quercus geminata ), secondary seed dispersal may be an important factor in determining community composition and persistence of longl eaf pine systems. Using a combination of seed depots and seed predator exclosures I looked at both longleaf pine and sand live oak in terms of whether small animals such as squirrels ( Sciurus carolinensis ) and cotton mice ( Peromyscus gossypinus ) cache the seeds, and where the seeds of these two tree species best germinate. Since sand live oa k acorns are prone to infestation by weevils ( Curculio spp.), I also examined whether nu t condition affects acorn germination potential. I found that longl eaf pine seeds are cached by small mammals to a small degree. While these seeds are not moved great distances from where they originate, they
vi are often redistributed into microhabitats that promote successful seed germination. Caging experiments indicated that seeds were most likely to germinate when buried in open areas between adult trees, and to some degree, under shrub cover. On the other hand, sand live oak acorns appear to face heavy predation by large seed predators such as raccoons ( Procyon lotor ) and wild pigs ( Sus scrofa ). Those acorns that do escape predation, including weevil-infested acorns, may provide an opportunity for seedling establishment. However, it appears that sand live oak depends heavily on vegetative sprouting for regeneration. This suggests that even in the absence of fire, longleaf pines in Florida sandhill are able to persist through secondary seed dispersal by small animals coupled with heavy seed predat ion on competing sand live oak.
1 INTRODUCTION Diplochory Diplochory is defined as bei ng two or more seed dispersal phases with a different dispersal mechanism in each phase. This multi-step dispersal process is thought to increase the probability of seedlings establis hing from just one phase alone (Vander Wall and Longland, 2004). Phase one occurs when the seeds are moved away from the parent in an attempt to decrease competition with, and predation under, the parent tree. Phase one is random with respect to where seeds land, often depositing th e seeds in clumps which can increase predation risk, or in sites that decrease germination probabilities, such as into dense litter or on the soil surface (Vander Wall and Longland, 2004). For species such as pines, wind dispersal would be phase one; the gravitational drop of acorns from oaks is also an example of this phase. Phase two concerns the moveme nt of the seeds from thei r initial landing spot to safe sites, where they are somewhat pr otected from predation and environmental effects, and where there is an increased chance of seed germination. Often termed directed dispersal or sec ondary dispersal, this phase is seen as a higher quality dispersal process that redistri butes seeds into sites that ar e usually favorable for seedling establishment (Lanner 1998). A small mamm al scatterhoarding seeds that it has harvested from the ground would re present phase two of diplochory.
2 Seed Caching and Secondary Seed Dispersal by Small Mammals As far back as the 1940s, studies have documented animals storing seeds for future use (Hatt, 1943; Spencer, 1941). Despit e the importance of animal seed dispersal and the interaction between small mammals and plants, however, investigators have largely ignored this st age of seed dispersal except with respect to the general interaction between squirrels and acorns (Plucinski and Hunter, 2001; Steele et al., 2001; Steele et al., 1996; Weckerly et al., 1989; Stapanian a nd Smith, 1984) and heteromyid rodents in the southwestern United States (Longland et al., 2001; Price et al., 2000; Longland and Clements, 1995; Reichman, 1979; but see: Wang et al., 2004; Manson et al., 1999; Vander Wall, 1994; Price and Jenkins, 1986). As Vander Wall (1992) suggested, a tendency exists to focus on the first stage of di spersal, from parent plant to contact with soil. Generally, where small mammals have be en investigated, the studies have usually been directed toward their roles as seed predators (Wenny, 2000; DeSimone and Zedler, 1999). Although most rodents are indeed seed predators, th ey also tend to be seed dispersers (Brewer, 2001; Pr ice and Jenkins, 1986), when a proportion of their seed caches are not consumed. Seed caching can potentially be beneficial to the plant species, the mammals that cache them, or both. Seed caching allows sma ll mammals to guard against food scarcity and compete for limiting resources (Vander Wall, 1990; Jensen and Nielsen, 1986; Price and Jenkins, 1986), while the seeds may be nefit through decreased predation by noncaching animals, decreased intraspecific comp etition with larger plants, and the potential for being buried in favorable germination s ites (Forget et al., 2000; Vander Wall, 1990; Jensen and Nielsen, 1986; Pric e and Jenkins 1986). In fact recolonization of cleared
3 sites has (in some cases) been attributed to mammal seed cachesin these cases, caching in open sites decreases chance of cache robbery, and simultaneously increases germination and survival probabilities (K ollmann and Poschlod, 1997; Dickman and Doncaster, 1989). Borchert and Jain (1978) f ound that in experimental sites in California grassland where rodents were present, recrui tment of both small a nd large-seeded grass species was enhanced, while in sites where rodents were excluded, there was a negative effect on the abundance of small-seeded species because of interspecific competition with the larger-seeded species. Therefore, it a ppears that rodent seed-cachers may have an important role in the establis hment of plant populations, espe cially in cleared sites. However, animals may cache seeds in sites that are not necessarily favorable for germination. Animals that cache seeds in these bad sites thus have a negative effect on seedling establishment (Hollander and Vander Wall, 2004). Directed seed dispersal by animals is not always towards the best germin ation sites, but it may be more likely to occur than with random wind di spersal (Purves and Dushoff, 2005). The proportion of cached seeds that su rvive to germination is unknown, although estimates have ranged from 2.2% (Guariguata et al., 2000) to 15% (West, 1968), and as high as 99% (Vander Wall, 1994). However, Vander Wall (1990) argued that although only a small proportion of seeds may escape predation through caching, this dispersal method is so effective that some plants are primarily dependent on small mammals for seed dispersal. Indeed, nut morphology (i.e. being large and nutritious ) and the fact that nuts and seed fall to the ground appear to be adaptations for dispersal by ground-dwelling small mammals (Vander Wall, 1990).
4 Probability of germination is likely a function of a number of factors such as plant species (Vander Wall 1990), seed size (Bre wer, 2001; Price a nd Jenkins, 1986), and whether it is a masting year, where an increase in the seed crop can l ead to an increase in the number and size of caches, as well as a higher probability of cached seeds going unrecovered (Vander Wall, 1992; Price and Je nkins, 1986), although high year-to-year variability in annual seed production (Herrera et al., 1998) may render this effect highly unpredictable. In addition to environmental stochasticity, however, microsite appears to be the single most important factor in dete rmining whether a seed will germinate into a seedling (Xiao et al., 2005; Hollander and Vander Wall, 2004; Kollman and Pochslod, 1997). Increasingly, researchers are looking at the intera ctions between seeds and animals beyond seed predation, to secondary seed dispersal and the processes involved in diplochory (Vander Wall et al., 2005; Brew er, 2001). The study of secondary seed dispersal allows a better understanding of pl ant population biology and the ultimate fate of seeds. The Sandhill Community Sandhill vegetation is a fire-adapted system of longleaf pine ( Pinus palustris ) over an understory of wiregrass ( Aristida stricta ) and scattered oaks, including turkey oak ( Quercus laevis ), and sand live oak ( Q. geminata ) (Myers and Ewel, 1990). The sandhill ecosystem occurs on the xeric, well-drained soil s of the southeastern coastal plain of the United States, into the Florida panhandle and northern and central pe ninsular Florida. Sandhill vegetation is dependent on frequent fires; in fact, many of the vegetation species
5 have life cycles that are intrinsically linked with fire (Platt, 1998; Myers and Ewel, 1990; Rebertus, 1988). Because of its long association with sandhi ll, fire is an important component of the disturbance regime. Historically, sandhi ll probably burned ever y 2-5 years in the early summer (May-June), when there is the highest probability of afternoon thunderstorms (Mushinsky and Gibson, 1991; My ers and Ewel, 1990; Rebertus, 1988). Fire is necessary to maintain the stability of the system, by decreasing canopy cover and litter, and increasing wiregrass and herbaceo us plant densities (Streng et al., 1993; Mushinsky and Gibson, 1991). As well, if fire s are frequent and lo w-intensity during the growing season, there is a higher probabil ity of oak mortality, slowing hardwood invasion of sandhill. On the other hand, if the natural fire regime is interrupted and time between burns increases, there is increased canopy closure, greater litter depth, and an increase in the number of hardwood speci es (Mushinsky and Gi bson, 1991; Rebertus, 1988). For example, Mushinsky and Gibson (1991) found that an annual burn plot had only 4% canopy cover, while an unburned plot had 45-60% canopy cover. In the absence of fire, hardwoods such as oaks quickly invade the ecosystem (Myers and Ewel, 1990; Humphrey, 1982). The sandhill ecosystem is now considered endangered over much of its range. About 75-95% of the original 24 million hectar es has been exploited for citrus culture, urban, and commercial development, reflecting th e suitability of its we ll-drained soils for anthropogenic purposes (Richardson and Runde l, 1998; Myers and Ewel, 1990; Rebertus, 1988; Humphrey, 1982). Most of the remain ing sandhill vegetation has been highly disturbed. In fact, even with the return of prescribed burns aimed toward restoring
6 longleaf pine and wiregrass, turkey oaks tend to dominate the system, and wiregrass populations are slow to regenerate (Rebertu s, 1988). A number of hardwood species present in sandhill have seeds that are mainly mammal-dispersed, including oaks, hickory, dogwood, black cherry, southern Ma gnolia, and American holly (Myers and Ewel, 1990), which has implications for hard wood encroachment of sandhill in the absence of fire. Objectives 1. To determine the extent to which diplochor y occurs in the major sandhill tree species (longleaf pine, Pinus palustris ) and an associated hardwood species, sand live oak ( Quercus geminata ). I examined whether secondary seed dispersal occurs with respect to both of these tree species, and the implications of diplochory on the composition of sandhill. No data exist on secondary seed dispersal of longleaf pine, but since mammals have been known to cache seeds of wind-dispersed species (Thayer and Vander Wall, 2005; Vander Wall, 1992), I felt that it was likely that longleaf pine, too, is secondarily disper sed. Small mammals readily cache acorns (Smallwood et al., 2001; Steel e et al., 2001) and thus I hypothesized that mammals would cache sand live oak acorns (Abrahamson and Layne, 2003), which could lead to oak establishment in unoccupied patches. 2. To examine what factors affect germinati on probabilities of longl eaf pine seeds and sand live oak acorns. Whether seeds are actively cached by animals or not, which microsites are most favourable to seed ling establishment, and what are the consequences of these factors? With longl eaf pine, it has been suggested that seeds
7 germinate best when in contact with soil (Boyer, 1990) in open areas (Platt et al., 1988). I hypothesized that longleaf pine seed s were more likely to germinate in open sites, and that burial would enhance germination. Estimates of seedling establishment for sand live oak are difficult to obtain, but other oak species germinate well when buried (Price and Jenkins, 1986) in open sites (Abrams, 1996), and I felt that this too would be the case for sand live oak. Study Site This study took place at the University of South Floridas Ecological Research Area (Eco Area) (2805N, 8220W) located in Tampa, Florida. The EcoArea is a one square mile tract of land that consists of sandhill, flatwoods, and swamps. In the southern portion of the tract, the Eco Area is divi ded into experimental burn plots of 6000m2 to 20000m2 in size. The vegetation in the Eco Area burn plots, as well as in patches north of these plots, is typical of the southern sandhill association found in xeric upland areas with sandy, fairly infertile soil. An open canopy is dominated by longleaf pine interspersed with turkey oa k and sand live oak. The ground cover consists of various grasses ( Andropogon spp., Aristida stricta ), herbaceous plants ( Pityopsis spp., Liatris spp., Eupatorium spp., Aster spp.), and saw palmetto ( Serenoa repens ). The Eco Area is home to a number of seed predators and potential seed dispersers, including grey squirrels ( Sciurus carolinensis ), cotton mice ( Peromyscus gossypinus ), cotton rats ( Sigmodon hispidus ), wild pigs ( Sus scrofa ), raccoons ( Procyon lotor ), white-tailed deer ( Odocoileus virginianus ), and wild turkey ( Meleagris gallopavo osceola ).
8 CHAPTER ONE SECONDARY SEED DISPER SAL OF LONGLEAF PINE Introduction Longleaf pine, of the Family Pinaceae, is found in the upland sandy soils of the United States coastal plain, from Virginia to Texas and into south-central Florida. Longleaf pine is extremely drough t-tolerant, because of a large taproot, but is also quite shade-intolerant, preferring ope n canopies maintained by freque nt fires in order for its seedlings to flourish (Fralish and Frankli n, 2002), although seedlings and juveniles can persist in the shade (Gordon Fox, pers. comm.). Its entire lifecycle is intrinsically linked to fire. Where fire is suppressed by hu mans or does not occu r naturally, hardwood species are able to invade into open areas, d ecreasing the number of patches available for colonization by longleaf pine seedlings. Only 5% of the original l ongleaf pine woodlands still remain, reflecting fire suppression and agricultural practic es (Richardson and Rundel, 1998). Longleaf pine masts every 7-10 years, producing winged seeds that are dispersed primarily by wind. Longleaf pine releases it s seeds in the fall, and germination occurs during the relatively dry winter (Whitney et al., 2004). Longleaf pine recruits mostly into open spaces between adult trees; seeds do not germ inate well if shaded by adult trees or if under heavy litter (Platt et al., 1988), or if there is no c ontact between seed and soil
9 (Boyer, 1990). Early growth of the seed lings is heavily weighted toward the development of a deep taproot that reaches far into the soil to maximize uptake of nutrients and water (Whitney et al., 2004). In xeric sandhill, water limitation may strongly affect the distribution of longleaf pine seedlings if they are competing with neighboring seedlings for limited soil nutrients and moisture. It is also likely that seedlings compete strongly for light. Whitney et al. (2004) argued that any aggregation of seedlings that occurs as a result of dispersal eventually shifts to a more diffuse distribution as a result of this competition, bu t it may depend more on the amount of light available, as many open spaces in longleaf pi ne woodlands are occupied by clusters of pine seedlings. Longleaf pine seedling biology is largel y unknown. Seedlings are thought to spend the first 2 to 12-15 years in what is known as the grass stage (Burns and Honkala, 1990a), where a grass-like bunch of need les surrounds the fire -resistant apical bud. After accumulating sufficient carbohydrate stores, the seedling bolts in a rapid growth phase that takes the apical meristem out of the range of most fires within 2-3 years (Keeley and Zedler, 1998). Because long leaf pine self-prunes its lower branches, once a tree is above the critical height of approximately 1.5m, it increases its odds of escaping major damage from the low-intens ity ground fires that naturally occur in longleaf pine stands (Agee, 1998), although trees at least twice this he ight can be killed by fire (Gordon Fox, pers. comm.). Recruitment of longleaf pine is highly epis odic and is most likely to occur during mast years (Platt et al., 1988). Wind causes the seeds to fall randomly with respect to site from the source tree. The openness of longleaf pine stands makes it likely that some
10 seeds fall into open spaces, but many more fall into less suitable germination sites, such as into shrubs or in areas of dense litter. The majority of seeds on the ground likely are subject to predation by animals. Also likel y, some proportion of these seeds is carried away by scatterhoarding animals into ground cach es (Lanner, 1998). Seeds in caches that are unrecovered by the animal that cached th em or by a nave forager, if placed in suitable microsites, can germinate, adding to the existing pine population. Several studies have documented small mammals ac tively caching wind-dispersed pine seeds (Thayer and Vander Wall, 2005; Vander Wall, 2002; Abbott and Quink, 1970). In fact, caching may be extremely important where seed s and seedlings are faced with moisture limitation, as buried seeds are less likely to de hydrate than are seed s on the soil surface (Vander Wall, 1990). In sandhill, caching of lo ngleaf pine seeds may be advantageous in promoting seedling establishment, especially in open sites. The traditional forestry theory is that wind-dispersed seeds fall into cracks and depressions in the soil in orde r to reach safe sites for germ ination (Lanner, 1998). Lanner (1998) points out that because there is no eviden ce for this theory, it is most likely based on the reasoning that seeds would not be able to survive on the so il surface because of high levels of dehydration and because of preda tion. While it is possible that seeds could be dispersed into soil cracks, it would be ha rd to gather evidence confirming this, and it seems more likely that many seeds are activel y carried away by small animals and buried in sites that may be favorable for germina tion, in the second phase of diplochory. As diplochory in wind-dispersed pines has not been well-documented, my first objective was to determine to what extent dipl ochory occurs in longleaf pine.
11 Regardless of how a seed arrives at a pot ential germination site, environmental factors play an important role in determining the likelihood th at the seed will be able to germinate (Xiao et al., 2005). My second obj ective was to determine how much more likely longleaf pine seeds are able to germinat e in ideal open areas as opposed to under shrub cover, and the differences that result from being buried (as in an animal cache) as to lying on the soil surface (as in random wind dispersal). Methods Caching By Small Mammals In October 2005, I chose five mature l ongleaf pines within the study area. At each tree, I set up 20 seed depots (a Petri dish glued to a 30.5cm x 35.6cm tray) in a pattern that simulated wind dispersal (Table 1). This wind di spersal pattern was adapted from Greene and Johnsons (1989) c onifer seed wind dispersal model and corresponding experimental results. The seeds at each depot were marked with fluorescent powder (Radiant Color, Richmond, CA ); each tray was filled with fluorescent powder to mark the feet of a ny animals visiting the depots. Table 1. Number and density of seeds at var ying distances of depots from source tree to simulate a wind-dispersed seed shadow. Distance From Tree (m) Number of Depots Number of Seeds/Depot Density (Seeds/m2) 4 8 55 8.75 8 8 58 3.08 12 4 25 0.40
12 I checked the depots every two days and counted the number of seeds removed: eaten (empty hulls) or gone (completely ea ten or carried away from the depot). I also went out at night on days 5 and 10 with a handheld ultraviole t light to track powder trails and find cache sites. I marked each cache site with a flag, and returned the next day to measure the following cache characteristic s, after checking that the cache was still present: cache size, distance from depot, and distance from tree. I measured percent cover of grass and shrubs to use in my statistical analyses as measures of vegetation structure. I walk ed four 24m transects in an area that encompassed the seed depots around each tree. At 8m intervals along each transect, I determined the percent cover of grass and th e percent cover of shrubs. I did this by estimating the amount of ground covered by grass and shrubs in a 1m2 quadrat and converting my estimates into percent cover. Germination Success by Microhabitat and Presence of Small Mammals This experiment took place in the winter months, from December 2005 to March 2006. I established a 3-way factorial design with the following levels: open ground vs. under shrub canopy, caged vs. uncaged, and buried vs. unburied seeds. Each cage measured 0.5m x 1.0m x 0.3m and was made of fiberglass screeni ng stapled to wooden stakes. The screening was buried > 5cm into the ground at the bottom of the cages to prevent animals from burro wing under the edges. Fifteen cages were placed in open habitats (open canopy with a mixture of sandy/grassy soil surface). The other 15 cages were placed in shrub habitats (in shrubby stands of bushes/oak trees). On e half of each cage was randomly assigned
13 buried and the other half was designated as unburied. Twenty -five longleaf pine seeds were placed within each of these treatm ents (50 total per cage) in a 5 x 5 pattern, with the buried seeds being buried 1 cm in the ground and covered with a light layer of topsoil, and the unburied s eeds being placed directly on the soil surface. Fiberglass screening was stapled over the top of the cages to prevent birds and other seed predators from removing the seeds. This screening caused minimal shading effects. Controls (uncaged treatments) were established with in 1m of each of the cages. Cages were monitored every few days for evidence of germination. Statistical Analyses Statistical analyses were performed in SAS (v9.1) and SigmaStat (v3.1). For the seed depot and caching data, I examined seed removal rates via life table analyses of survival probability and hazard functions. Thompsons chi-square test for spatial patterning using nearest-neighbor methods was used to determine the spatial distribution of cache sites (Thompson, 1956). The observed 2 value is tested against two alternative hypotheses, where a signif icantly small value of 2 (i.e. less than 2 0.975 at = 0.05) signifies an aggregated pattern, a nd a significantly large value of 2 (i.e. more than 2 0.025 at = 0.05) implies uniform patterning. Any 2 values that fall within this range indicate that the data is randomly patterned. I us ed both nearest-neigh bor and second-nearest neighbor methods to analyze spatial pattern, as both can provide information on spatial distribution. For example, if there is an office building with two people in each office, using nearest-neighbor methods would likely indicate that i ndividuals are in a clumped pattern. Using second-neares t-neighbor distances would likely show a more uniform
14 pattern. Increasing levels of nearest-nei ghbor distances would probably not add much more information with respect to spatial pa ttern. In the case of seed caches, nearestneighbors may be clumped, but on the scale of second-nearest-neighbor mean distances, caches may demonstrate a more regular pa ttern as a whole. A one-way ANOVA and Holm-Sidak pairwise comparisons were used to determine whether caches were made disproportionately with respect to habitat. I used logistic regr ession to relate seed fate to vegetation cover. A log-linear model for count data, assuming a Poisson di stribution and utilizing a log link function, was used to examine the eff ects of protection level (caged or uncaged), habitat (open site or under shrub cover), and microsite (buried or on top of soil) on germination of longleaf pine seeds. The m odel tested the individual effects and their interactions on the fate (i.e. germ ination or not) of pine seeds. Results Caching By Small Mammals Within 10 days, 98.9 1.37% (mean 1 SD) (97.21-99.9%) of longleaf pine seeds were removed, with a mean removal rate of 99.94 seeds per day. Empty hulls indicated that 5.57 6.79% (0.6017.13%) of the seeds were eat en at the depots. Life table survival es timates indicate that there was a short delay immediately following the placement of the seeds in the depots, followed by a rapid but steady decrease in survival of the seeds over time (Figure 1). The hazard function, which is the per capita probability of death at any one time given su rvival up to that time, increas ed sharply after an initially steady rate (Figure 1).
15 Time (days) 0246810 Probability 0.0 0.2 0.4 0.6 0.8 1.0 1.2 Survival Hazard Figure 1. Survival distribution and hazard functions ( 1 S.E.) for longleaf pine seeds. Error bars are not always visually apparent because sample size is on the order of 5000. I was able to track 1.91 0.34% (1.392.29%) of the seeds to caches made by small animals. Figure 2 shows the location of these caches relative to each individual tree.
16 (a) (b) (c) (d) (e) Figure 2. The locations of longleaf pine seed caches made by small mammals around each of five source trees, (a) tree one, (b) tree two, (c) tree three, (d) tree four, and (e) tree five. Source tree is repres ented by the cross-hatched circle. Filled circles represent locations of seed depots. Open circles represent locations of caches.
17 Using nearest-neighbor data, I determined that caches around trees one and two were distributed in a clumped pattern ( 2 44 = 22.65, p < 0.05 and 2 34 = 14.01, p < 0.05 respectively; Table 2). Caches around tree three were rando mly distributed ( 2 38 = 39.84, p = 0.388; Table 2). Trees four and five had caches that were uni formly distributed ( 2 32 = 59.37, p < 0.05 and 2 26 = 38.33, p < 0.05 respectively; Table 2). Table 2. Spatial pattern of longleaf pi ne seed caches around source trees as a function of nearest-neighbor distances. Tree df 2 p Spatial Pattern 1 44 22.65 0.0032 Clumped 2 34 14.01 0.001 Clumped 3 38 39.84 0.3882 Random 4 32 59.37 0.0023 Uniform 5 26 38.22 0.0578 Uniform When second-nearest-neighbor data were used, the distribution of caches were uniformly distributed around trees one, three, four and five (Table 3). Tree two had randomly distributed cache sites (Table 3). Table 3. Spatial pattern of longleaf pi ne seed caches around source trees as a function of second-nea rest-neighbor distances. Tree df 2 p Spatial Pattern 1 88 160.53 <0.0001Uniform 2 68 74.89 0.2648 Random 3 76 181.05 <0.0001Uniform 4 64 122.83 <0.0001Uniform 5 52 88.47 0.0012 Uniform I analyzed third-nearest-neighbor data fo r tree two, and found that on this scale, caches were distributed in a random pattern around the tree ( 2 102 = 125.73, p = 0.0555).
18 Although this p value is not quite significant, it indicates that as higher-level nearestneighbor distances are used, caches show a more uniform distribution. The number of seeds cached per tree ranged from 13 to 22 seeds (mean = 17.4 3.36 seeds) (Table 4). The majority of thes e caches had one seed in them (mean = 1.09 0.03 seeds/cache; range 1-2) (Table 4). Table 4. Number and size of longleaf pine seed caches made by small mammals. No significant difference exists among trees. Number of Number of Cache Size Tree Caches Made Seeds Cached Mean SD 1 22 23 1.05 0.21 2 17 20 1.12 0.33 3 19 21 1.11 0.32 4 16 18 1.13 0.34 5 13 14 1.08 0.28 Seeds were not moved very far from th eir originating depots (mean = 0.93 0.29m), but were generally moved away from the trees toward the canopy edges (a mean net distance of 0.33 0.28m away from the tree) (Table 5). Seed caches appeared to be somewhat clustered in distribution. Table 5. Longleaf pine seed m ovements and dispersal distances. Seed Movements Relative To Their Orig in Secondary Dispersal Distance (m) Tree Toward source tree Away from source tree Mean* SD Shortest Longest 1 12 10 -0.084 0.56 0.08 3.59 2 7 10 0.25 0.84 0.07 2.55 3 5 14 0.33 1.02 0.08 2.96 4 3 13 0.49 0.56 0.20 1.45 5 3 10 0.65 0.74 0.26 1.82 A negative dispersal distance indicates that seeds were mo ved toward the source tree.
19 When cache sites were characterized by microhabitat, I found that 48.28% of the caches were made in open, grassy sites, 33.33% were made in bare sandy soil, and 18.39% were made under shrubs and dense vegetation. Numbers of caches made were significantly different as a function of microhabitat, as shown by both a 2 test and a oneway ANOVA (F2,12 = 4.105, p < 0.05). Holm-Sidak pairwise multiple comparison procedures showed a significant difference between the number of caches made in grass as opposed to in shrub microhabitats (p < 0.05). Grass cover among sites ranged from 60.31-86.63% (mean = 69.39 10.70%), while shrub cover was 9.94-39.69% (mean = 29.18 11.61%) (Figure 3) A generalized logistic model indicated that grass cover, shrub cover, and the interaction between the two types of cover, all had significant effects (p < 0.05) on the fate of seeds, i.e. whether seeds were eaten immediately or cached (Table 6). Tree 0123456 % cover 0 20 40 60 80 100 120 Grass Shrub Figure 3. Percent cover of grass and shrubs around source longleaf pines. Grass cover and shrub cover are not mutually exclusive as both can be present in a given location.
20 Table 6. The effects of grass cover and sh rub cover on the fate of longleaf pine seeds. Source df 2 p Grass cover 1 4.11 0.0426 Shrub cover 1 12.93 < 0.001 Grass cover x shrub cover 1 20.45 < 0.001 Germination Success by Microhabitat and Presence of Small Mammals The mean number of germinated seeds for each main effect of protection, habitat, and microsite are shown in Figure 4. Mean number of germinated seeds 0 1 2 3 Caged Uncaged Open Shrub Buried Ground Figure 4. The separate effects of prot ection, habitat, and microsite on the germination of longleaf pine seeds. Error bars represent standard error. The combined effects of protection and habitat on the mean number of longleaf pine seeds that germinated when either buried or on soil surface are shown in Figure 5.
21 Mean number of germinated seeds 0 1 2 3 4 5 6 7 Caged, open Caged, shrub Uncaged, open Uncaged, shrub BuriedSoil surface Figure 5. The effects of habitat (open areas vs. under shrub cover) and protection (uncaged controls vs. caged treatments ) on germination of buried and groundsurface longleaf pine seeds. Error bars represent standard error. A log-linear model (a generalized lin ear model assuming a Poisson distribution and utilizing a log link function) tested the main effects protection, habitat, and microsite, the three two-way interactions, a nd the three-way interaction, in relation to pine seed germination. Protection: Caged vs. Uncaged When caged, 2.20 2.69 seeds germinated, while only 0.38 0.80 uncaged seeds germinated. Protection from predation signifi cantly affected germination probabilities ( 2 = 57.03, p < 0.05; Table 7). Habitat: Open Ground vs. Under Shrub Cover More seeds germinated in open areas (1.50 2.62 seeds) than did under shrub cover (1.08 1.61 seeds), but not si gnificantly so (Table 7).
22 Microsite: Buried vs. Soil Surface Seeds germinated in signif icantly higher proportions ( 2 = 47.70, p < 0.05; Table 7) when buried (2.23 2.57 seeds) than when on the soil surface (0.35 1.07 seeds). The Effects of Habi tat and Protection The effect of caging was greater in op en areas (2.83 3.15 seeds germinated) than in shrubby areas (1.57 1.98 seeds germinated), but no t significantly (Table 7). Interestingly, the opposite was true in the un caged controls, where more seeds germinated in shrubby sites than in open sites (0 .60 0.93 and 0.17 0.59 seeds, respectively). The Effects of Habitat and Microsite The difference between the number of burie d seeds that germinated and those on soil was greater in open areas (2.7 3.15 buried, 0.30 1.02 on the ground) than in shrubby areas (1.77 1.74 buried, 0.40 1.13 on the ground), but not significantly so (Table 7). The Effects of Protection and Microsite The effect of caging was greater for bur ied seeds than for those on the ground, but not significantly so (Table 7). When seeds were buried, 3.7 2.82 caged and 0.77 1.01 uncaged seedlings arose, while 0.70 1.44 cag ed seedlings came from seeds on the soil surface. No uncaged seeds on the soil surface survived to produce seedlings. The Effects of Habitat, Protection, and Microsite The effects of protection and microsite (i .e. buried vs. soil su rface) were more prominent with seeds that were in open area s. With open areas, the difference between caged and uncaged seeds was much larger for buried seeds than fo r those on the ground. Under shrub cover, the difference between caged and uncaged seeds was only slightly
23 higher with buried seeds. There was basica lly no difference in germination proportions between protection levels with seeds placed on the soil surface, whereas there was a large difference between protection levels for buried seeds. Table 7. Likelihood ratio statistics for Ty pe III analysis of the effects of fate, habitat, protection, and microsite on germination of longleaf pine seeds. Source df 2 p Fate 2 24983.10 < 0.001 Protection x fate 2 57.03 < 0.001 Habitat x fate 2 0.49 0.783 Microsite x fate 2 47.70 < 0.001 Protection x habitat x fate 2 2.13 0.345 Protection x microsite x fate 2 3.23 0.199 Habitat x microsite x fate 2 0.46 0.796 Protection x habitat x microsite x fate 2 3.07 0.216 Discussion Evidence exists that diplochory does o ccur in longleaf pine, and that small animals do cache wind-dispersed seeds. While only a small percentage of seeds were found in caches, these seeds have the potenti al to add to the ex isting pine population. Regardless of how seeds arrive in a potential germination site, those that are in open areas have a better chance of surviving to seedlings Burial also enhan ces germination success by offering more protection from both predati on and dehydration than is afforded those that are lying on the soil surface.
24 Seeds in depots were removed rapidly af ter an initial delay, indicating that seeds lie on the soil surface for a s hort period before being found by animals. After seeds were discovered, hazard for an individu al seed increased substantiall y. It appears that animals have the potential to remove an entire s eed crop within a few weeks, eating some and caching others. More than 90% of the seeds in the depots were unaccounted for, so those seeds not immediately eaten appear to have been transported away from their origin. While the fate of these seeds is not known, seed removal does not necessarily imply seed predation (Vander Wall et al., 2005; Brewer, 2001). Perhaps some of these seeds also were cached. Vander Wall (1993) stressed that misinterpretation of seed removal results can lead to an overestimation of seed predation rates, especially in cases such as this where secondary seed dispersal through caching occurs. Less than two percent of pine seeds were located in caches. This number seems small until longleaf pine life hi story is taken into consideration. Longleaf pines produce their first cones in about their thirtieth year, and depending on the size of an adult, pines can produce between 15 and 65 cones annually, with an average of 35 seeds per cone (Boyer, 1990). Over a lifetime of 70 years, tw o percent of an individual trees seed yield could potentially result in anywhere from 700 to severa l thousand pine seeds being cached. The number of seedlings resulting fr om these caches would of course depend on successful germination and seedling survival However, as longleaf pines can live several hundred years (Boyer, 1990), recruitmen t could likely be quite significant. Seeds were not moved great distance s, but secondary dispersal does not necessarily mean that there is any advantage to the seed in terms of increasing dispersal distances. Instead, seeds are redistributed around the source with an emphasis on
25 directed dispersal, often toward favorable germination sites (Vander Wall and Longland, 2004). Caches close to other caches tended to be aggregated or uniformly distributed. However, second-nearest-neighbor caches were regularly spac ed. This suggests that animals rearrange wind-dispersed seed sh adows into a more uniform distribution, although individual caches may be clumped together to a small degree. In this case, the process of caching actually redistributed th e seeds in a more evenly-spaced distribution than that displayed by wind disper sal (Vander Wall and Longland, 2004). The majority of seeds were cached in open sandy and grassy areas, while only a small proportion were cached under shrubs. While directed dispersal does not always favor the species being dispersed, in this case, the seeds were placed in sites where longleaf pine seeds do better: in open areas between adult trees (Platt et al, 1988). The animals largely avoided caching pine seeds unde r the shade of shrubs where pine seeds often fail to germinate. Thus, directed disp ersal of longleaf pine by small animals may be toward advantageous microsites, which may be important in maintaining the current longleaf pine population and in colonizing open spaces. The caging experiment demonstrated that buried seeds are far more likely to germinate than are seeds placed on the soil su rface. Wang et al. (2004) found that seeds of oil tea ( Camellia oleifera) germinated significantly more when buried than when left on the soil surface. Burial of seeds both enhances germination and decreases predation rates (Guariguata et al., 2000; Crawley, 1997). All of the uncaged seeds on the soil surface and many of the caged ground-surface seeds were removed by animals; some may have been cached but it is likely that many were eaten. Most of the seedlings originated from buried treatments, especially those protected from animals. Also, it is
26 likely that seeds on the soil surface are s ubject to dehydration over hot, dry winters, which may decrease the likelihood of germination (Lanner, 1998). Those seeds in open habitats were also mo re likely to survive to germination than those under shrub cover. This is most likely due to the fact that small mammals prefer some degree of cover to decrease preda tion risk (Vander Wall, 1990), and seeds under shrubs are more accessible than those in th e open. Kollman and Poschlod (1997) suggest that seed survival increases with openness b ecause less cover is available as protection for small seed predators. Many animals also cache their seeds in open areas to decrease cache pilferage (Vander Wall, 2000; Vander Wall 1990), so these seeds are less prone to a nave animal locating them. However, a number of seeds did germinate under shrub cover. Because the literature strongly suggests that pine seedlings in the shade do not survive as well as those in open areas (Keel ey and Zedler, 1998), further study on first and second year seedling survival would help to determine the importance of habitat in seed germination. Both Price and Jenkins (1986) and Shaw (1968) found that seeds protected from mammals were better able to germinate. However, even without caging, a small proportion of seeds, both buried and on the soil surface germinated (Shaw, 1968). In my study, all of the seeds that were uncaged a nd on the soil surface were removed. This suggests that seeds on the ground are prone to heavy predation a nd are not likely to survive to germination. This also suggests th at longleaf pine seedli ngs most likely arise from animal seed caches, since some that were buried were able to establish, regardless of protection level.
27 Thus, diplochory does occur in longleaf pi ne, with a proportion of the seed crop being cached by small animals. Those cached seeds that go unrecovered, if buried in suitable sites such as open areas, and to some degree unde r shrub cover, are able to germinate. These seeds that germinate into s eedlings are the ones that will contribute to the existing pine population.
28 CHAPTER TWO SECONDARY SEED DISPERSA L OF SAND LIVE OAK Introduction Sand live oak ( Quercus geminata Small) is a white oak in the Family Fagaceae. It is a semi-evergreen broadleaf shrubby tree that is found in the lower coastal plain of the United States, from Louisiana to North Caroli na and stretching into Florida. Sand live oak is usually found in xeric to mesic sites s ubject to nutrient and water limitations, such as sandhill (Fralish and Franklin, 2002). Wh ile sand live oak does respond to fire by vegetative resprouting, it actually invades more readily as fire intervals increase (Menges and Kohfeldt, 1995). In the absence of fire, sand live oa k, along with other hardwood species such as turkey oak ( Q. laevis) has the tendency to replace longleaf pine ( Pinus palustris) in the southeastern pine-hardwood a ssociations (Fralish and Franklin, 2002). Sand live oak produces acorns every year and most trees in a given population mast every 2-6 years. These acorns mature and drop in the fall and, like other white oak acorns, germinate shortly after landing on th e ground. This lack of dormancy, coupled with a large amount of stored carbohydrates, makes the acor ns highly palatable to seed consumers (Abrahamson and Abrahamson, 1989). In northern temperate forests, squirrel s and other small mammals readily cache vast quantities of acorns (Smallwood et al., 2001; Steele et al., 2001; Barnett, 1977).
29 However, the amount of acorn caching that occurs in southern sandhill is not known. While resprouting of sand liv e oak may aid in local population maintenance, secondary dispersal via small animals may allow colonization of new patches within the open sandhill system, allowing sand live oak to mo re easily invade and replace longleaf pine stands. Therefore, my first objective was to determine if secondary dispersal of sand live oak acorns occurs in Florida sandhill. I hypothesized that since acorns of other oak species are secondarily dispersed by small mamm als, it is likely that sand live oak acorns will undergo a similar process. Although oak species do not tend to be light-limited, oaks readily establish large numbers of seedlings with rapid growth rates in high light environments (Abrams, 1996). In characteristically-open sandhill, fast gr owth rates offer many opportunities for sand live oak to rapidly regenerate. The amount of light available to sand live oak may affect the success of acorn germination. As we ll, although white oak acorns do tend to germinate shortly after seed dr op, some species of oak are un able to germinate on the soil surface (Price and Jenkins, 1986). Thus, my second objective was to determine the microsite that best enhances sand live oak aco rn germination. I hypothesized that burial in open sites would enhance germ ination of sand live oak acorns. Weevils of the Family Curculionidae atta ck acorns of different species to varying degrees (Branco et al., 2002; Steele et al., 1996; Ander sson, 1992). The adult female oviposits eggs into maturing acorns. These e ggs hatch within 1-2 weeks, and the larvae dwell inside the acorn, feeding on the endospe rm for at least two weeks before chewing an exit hole and leaving th e acorn to continue devel opment underground. Weevil larvae can infest up to 70% of an acorn crop in a given year (Steele et al., 1996). Animals are
30 known to actively select seeds and manage th eir caches, often reje cting weevil-infested seeds (Crawley and Long, 1995; Vander Wall, 1990). Roth and Vander Wall (2005) observed a chipmunk picking out infested bush chinquapin seeds while carrying away sound seeds. Rejected weevil-infested acorns have the potential to germinate and provide a means of seedling establishment for an oak population, if the acorn embryo is not damaged and if enough of the endosperm re mains (Branco et al., 2002). My third objective was to determine whether uninfested acorns are preferred over weevil-infested acorns, and to what extent weevil-infested acorn s are able to germinate. I felt that sound acorns would be taken preferentially over infe sted acorns, and that infested acorns would be able to germinate despite damage to the endosperm. Methods Removal and Caching By Small Mammals This experiment took place in October and November 2005, during natural seed fall. I chose five mature sand live oak ( Q. geminata ) trees within the study area. At each tree, I set up 10 seed depots (a Petri dish glued to a 30.5cm x 35.6cm tray) in a circle around the tree. The radius of this circle was 5m; the trays were approximately 3.14m apart. This circular pattern is similar to that shown by gravitationa l dispersal of acorns. At each depot, I placed 80 acorns mark ed with fluorescent powder (Radiant Color, Richmond, CA). I also filled each tray with fluorescent powder in order to mark the feet of any animals visiting the depots. I checked the depots every other day and counted the number of acorns removed from each depot. Removed acorns were further categorized as eaten (empty acorn hulls pr esent at or around de pot) or gone (either
31 completely eaten or carried away from the de pot). On day 6 of this experiment, I went out at night with a handheld ultraviolet light to track any fluorescent powder trails. Germination Success by Microhabitat and Presence of Small Mammals This experiment took place over the wint er, from December 2005 to March 2006. I established a 3-way factorial design with th e following levels: caged vs. uncaged, open ground vs. under shrub canopy, and buried vs. unburied. Each cage measured 0.5m x 1.0m x 0.3m and was made of fiberglass sc reening stapled to wooden stakes. The screening was buried > 5cm into the ground at the bottom of the cages to prevent animals from burrowing under the edges. Fifteen cages were placed in open habitats (open canopy with a mixture of sandy/grassy soil surface). The other 15 cages were placed in shrub habitats (in shrubby stands of bushes/oak trees). On e half of each cage was randomly assigned buried and the other half was designated as unburied. Twenty-five acorns were placed within each of these treatments (50 to tal per cage) in a 5 x 5 pattern, with the buried acorn being buried 3-4 cm in the ground and covered with a light layer of topsoil, and the unburied acorns being plac ed directly on top of the soil surface. Fiberglass screening was stapled over the top of the cages to prevent birds and other seed predators from removing the seeds. This screening caused minimal shading effects. Controls (uncaged treatments) were establis hed within 1m of each of the cages. Cages were monitored every few days for evidence of germination.
32 The Effects of Weevil Infestat ion on Removal and Germination To determine infestation rates of a typical crop of acorns, I co unted the number of acorns with weevil holes from a sample of 1000 acorns. Acorns were collected in October 2005, at the time of natural seed fall. I used the presence or absence of weevil holes as an estimate of the maximum amount of damage that a weevil larva could inflict before exiting an acorn. Transect Experiment In November 2005, I established ten 120m transects throughout the study area. Each transect consisted of 24 stations placed 5m apart, with one acorn per station. On each transect, half of the acorns were sound (i ntact nut with no weev il holes) and half of the acorns were infested (visible weevil hole) and I assigned thei r positions randomly. Transects were monitored daily to determine removal rates. Greenhouse Experiment This part of the study took place at the greenhouse in the University of South Floridas Botanical Gardens from November 2005 to March 2006. I randomly assigned 120 sound and 120 infested acorns to 240 pots in the greenhouse. I planted individual acorns at a depth of 2.5 times the width of th e acorn in potting soil within these pots, and watered them every other day. I monitored these acorns every few days for emergence and germination proportions and rates. Statistical Analyses Statistical analyses were performed in SAS (v9.1) and SigmaStat (v3.1). I used life table analyses to examine removal rate s for both acorns from depots, and for acorns
33 from transects. Logistic regression was used to determine the effect of nut condition on germination in the greenhouse. Results Removal and Caching By Small Mammals Within 16 days, 99.58 0.61% (mean 1 SD) (98.5-100%) of the acorns were removed, with a mean removal rate of 49.79 acorns per day. 15.6 12.22% (3.2529.63%) of the acorns were eaten at the de pots (shown by empty hulls at or around the depot). Life table survival es timates indicate that survival probabilities initially remain high, then begin to steadily decrease (Figure 6). Hazard, the probability of non-survival at a time x given survival until time x, in creases gradually before suddenly peaking, and then declines sharply before increasing quickly again (Figure 6). Time (days) 0246810121416 Probability 0.0 0.2 0.4 0.6 0.8 1.0 1.2 Survival Hazard Figure 6. Survival distribution and hazard functions ( 1 S.E.) for sand live oak acorns.
34 It was impossible to tell, however, if any acorns were carried away by small animals, because one or more raccoons visi ted each depot at all 5 trees, strewing the fluorescent powder around and making it difficult to track anything other than raccoon prints, which are easily identifiable in the field. Germination Success by Microhabitat and Presence of Small Mammals Only 0.23% of the acorns survived to germ ination. Twenty-eight of the thirty cages were knocked over and uprooted by what looked like wild pigs and raccoons. As well, more than 90% of the acorns that were used in this experiment were removed. Of the seven acorns that germinated, 100% of them had been buried. Six acorns (85.7%) were in open sites, while the remaini ng acorn was under shrub cover. Four of the seven seedlings (57%) came from caged tr eatments, and the othe r three (43%) from uncaged controls. The Effects of Weevil Infestat ion on Removal and Germination 3.6% of the acorns that I surveyed fo r weevil infestati on had weevil holes. Transect Experiment Within 8 days, 94.2% of sound acorns and 77.5% of infested acorns were removed. Both sound and infested acorns were removed at a steady rate after the first day (Figure 7). Nut condition significantly affected daily rem oval rates (Wilcoxon chisquare, p < 0.05). The mean survival ti me of sound acorns was 3.54 2.14 days, while the mean survival time of infested aco rns was 4.78 2.38 days. A Wilcoxon test indicated that nut condition had a significant effect on mean survival time between the
35 two groups (p < 0.05). The hazard function, which estimates the probability of death given survival up to that time, indicates that while infested acorns are prone to a slight increase in removal over time, sound acorns are much more at risk, with a large increase in the probability of being eaten the longer that they re main on the ground (Figure 8). Time (days) 0246 Probability of survival 0.0 0.2 0.4 0.6 0.8 1.0 1.2 Sound Infested Figure 7. Survival probabilities ( 1 S. E.) of sound and weevil-infested sand live oak acorns from transects. Time (days) 0246 Probability 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 Sound Infested Figure 8. Hazard function ( 1 S.E.) fo r sound and weevil-infe sted sand live oak acorns from transects.
36 Greenhouse Experiment Within 118 days, 35.8% of sound acorns and 12.5% of infested acorns germinated. Type III analysis through logi stic regression indicated that nut condition significantly affected the number of acorns that germinated (p < 0.05). However, the mean time to germination di d not differ significantly with respect to nut condition (Mann-Whitney rank-sum test p = 0.797) (Figure 9). Sound acorns germinated in 75.2 16.0 days, while infested acorns germinated in 78.3 22.3 days. Mean germination time (days) 0 20 40 60 80 100 120 Sound Infested Figure 9. Mean ( 1 S.E.) germination ti mes of sound and infested sand live oak acorns. The probability of a weevil-infested aco rn surviving to germination is 0.0281, while the probability of a sound acorn surv iving to germination is 0.0208 (Figure 10).
37 Figure 10. The cumulative probability of survival to germination for sound and weevil-infested acorns. Numbers on branches indicate conditional survival probabilities. Numbers in brackets indicate the probability of nut condition and are not included in survival probability calculations. Discussion It is evident that sand live oak acorns are subject to extremely high levels of predation in Florida sandhill. However, thos e acorns that do manage to escape predation, including weevil-infested one s, may provide a means of seedling establishment. Acorns from seed depots were removed fa irly quickly from seed depots. There was a delay between when the depots were put out and when raccoons found the seeds. There was then another delay, which might be explained as a period where the raccoons ate to satiation, rested for a day or so, and then returned to finish eating the seeds. While the depots may have attracted the animals by promising a concentration of food, I feel that this is unlikely, as there were many acorns on the ground around the source trees as Acorn Sound Infested Eaten 0.942 Uneaten Eaten 0.775 Uneaten Germinate 0.0208 Die 0.0372 Germinate 0.0281 Die 0.197 (0.964) 0.942 0.225 0.058 0.775 0.358 0.125 (0.036) 0.642 0.875
38 well. There is the potential that other sma ll animals may have had brief opportunities to move acorns into caches, but unfortunately th ere is no evidence of such. Large mammal exclusion experiments may help determine if small mammals are actually caching acorns, but they would most likely give an overe stimate of seed caching proportions. While raccoons can be seed dispersers of various types of fruit (LoG iudice and Ostfeld, 2002; Cypher and Cypher, 1999), it is unlikely that th ey are dispersers of acorns. The only exception would be if acorns were only par tially consumed, and if the embryo was still intact. In this case, the acorns may still be able to germinate (Branco et al., 2002). The caging experiment again points to hi gh levels of acorn predation. Uncaged acorns were found almost immediately; ev en buried acorns were dug up. Cages were destroyed by large seed predators (most likely wi ld pigs and raccoons). It is important to note that the animals were not cueing into the cages themselves; pine seed cages close by were not destroyed or even t ouched. While further study is re quired, it is likely that these large seed predators ar e attracted to the odor and relative ly high nutritional value of the acorns; it is unlikely, however, that these an imals provide much in the way of secondary seed dispersal. The mechanical disturba nce of the soil caused by these large animals digging around in the soil may cause some seeds to become inadvertently buried and perhaps escape predation (Thayer and Vande r Wall, 2005; Alexander et al., 1986), but it appears that most acorns are eaten. The se ven seedlings that were found to originate from my cages represent a very small (<1%) proportion of the original number of acorns that I put out. While the fact that all of th em emerged from burial sites points to animal caching as providing a means of germination, without further testing it is almost impossible to speculate on the implications of this seedling establishment.
39 However, while only a small proportion of acorns were able to germinate, one must consider the implications of long-ter m acorn production by sand live oak. Sand live oak can produce acorns within the first year (Carey, 1992). While estimates of annual sand live oak acorn production are vague, othe r white oaks can produce from 200 to over 2000 acorns annually (Stransky, 1990). Oaks are long-lived species (Fralish and Franklin, 2002), and an estimate of 70 acorn-p roducing years is conservative, but even over this time period, a sand live oak coul d potentially produce 14,000 to several hundred thousand acorns. While less than 1% of thes e acorns may survive to germination, they represent an addition of 20 to several hundred oak seedlings to the understory, per adult oak. Interestingly, animals were less likely to remove weevil-infested acorns than they were to remove sound acorns. Transect expe riments showed that infested acorns were more likely to spend more time on the ground, offering possibilities for both germination and for small animals to eventually carry o ff and cache. Weevil-infested acorns clearly are able to germinate, although less so than sound acorns (Branco et al., 2002; Fukumoto and Kajimura, 2000; Andersson, 1992). Howe ver, when the probability of an acorn surviving predation is taken into account w ith the probability of germination, there is very little difference between sound and infested acorns; in fact, an infested acorn has a slightly higher probability of surviving to germ inate. If infested acorns are rejected by seed-eating animals, it may provide a way fo r acorns to escape predation and produce viable seedlings (Weckerly, 1989). Since whit e oak acorns germinate so soon after seed drop, those weevil-infested acorn s that are left behind, if ab le to germinate on the soil surface, may enhance that escape mechanism. While less than 5% of the acorns that I
40 surveyed were weevil-infested, that proporti on most likely increases over time (Steele et al., 1996), suggesting that infested acorns may make up a significant proportion of those seedlings that originate from acorns. Further study on seedling viability over time would be useful in determining the extent to whic h infested acorns contribute to a standing oak population. However, despite the large acorn crops that can result from sand live oaks, seed germination and seedling establishment appe ars to be relatively rare in the upland sandhill habitats (Abraham son and Layne, 2003). Sand live oak depends heavily on clonal expansion for population maintenan ce, which most likely counteracts the incredibly high levels of predation to which sand live oak acorns are subject. Sprouting likely allows the persistence of sand live oak in sandhill systems. However, while acorn escape and survival to germination may be rare, it is probably still important in oak recruitment in the long-term (Clark and Hallg ren, 2003), contributing to the presence of sand live oak in the sandhill understory.
41 DISCUSSION Secondary seed dispersal does occur in Fl orida sandhill, at least with respect to longleaf pine. Sand live oak, and most likel y other oak species such as turkey oak ( Q.laevis ), appear much more at risk of seed predation, and secondary seed dispersal is likely rare. During the caging experiments, large seed predators (most likely wild pigs, deer, and raccoons) knocked down cages to get to the acorns. Pine seed cages in close proximity to the destroyed acorn cages were untouched. Large seed predators were probably attracted by the odor of the acorns (Vander Wall, 2000). While a majority of uncaged pine seeds were eaten, the results of the caging experiments i ndicate that when a higher quality food is available, animals will opt for those, affording the other seeds an opportunity to escape predation and germinate. In sandhill, where acorns of several oak species are readily available, pine seeds stand a relatively good chance of surviving to establish seedlings. Acorns are more desirable food sources than are wind-dispersed pine seeds because of their higher moisture and nutri ent content (Fornara and Dalling, 2005). Larger seeds such as acorns are also prone to higher predation risk than are smaller seeds, especially with large seed predators that have large energy needs (Taulman and Williamson, 1994; Price and Jenkins, 1986); acorn s, with their high levels of fat,
42 digestible carbohydrates, and other nutrient s, fulfill these energy needs more than smaller, drier wind-dispersed seeds (Abr ahamson and Abrahamson, 1989). Alternate sources of food may cause differences in fo raging and caching behavior. Thus, when evaluating the fate of seeds, it is important to consider all of the food sources available to seed predators and potential seed di spersers (Roth and Vander Wall, 2005). The sandhill system may be somewhat se lf-sustaining, even in the absence of frequent fire. Longleaf pine has the abili ty to regenerate th rough both random wind dispersal events and through the secondary disp ersal of seeds via an imal caching. It is also likely that large seed predators such as wild pigs, deer, and raccoons, may be indirectly aiding in burying wind-dispersed seeds through the distur bance of soil caused by foraging for larger seeds (Thayer and Va nder Wall, 2005; Alexander et al., 1986). Sand live oak seedling establishment, on the other hand, appears to be suppressed by the presence of these large seed predators. It is likely that the oak species in sandhill counteract the lack of secondary seed disp ersal by vegetatively sprouting (Abrahamson and Layne, 2003). Regardless, a very complex interacti on exists in the sandhill. Although even without frequent fires, longleaf pine appears to be persisting (at least in this study area), sand live oak has a propensity to invade as fi re intervals increase. Frequent fires likely keep the oak population somewhat under control while creating the open spaces necessary for pine seedlings to establish. W ith widespread fire suppression in sandhill, the ability of longleaf pine to continue to colonize open patches may depend to quite some extent on mammalian seed dispersers.
43 REFERENCES Abbott, H. G., and T. F. Quink. 1970. Ecology of eastern white pine seed caches made by small forest mammals. Ecology 51 : 271-278. Abrahamson, W. G., and C. R. Abrahams on. 1989. Nutritional quality of animal dispersed fruits in Florida sand ridge ha bitats. Bulletin of the Torrey Botanical Club 116 : 215-228. Abrahamson, W. G., and J. N. Layne. 2003. Long-term patterns of acorn production for five oak species in xeri c Florida uplands. Ecology 84 : 2476-2492. Abrams, M. D. 1996. Distributi on, historical development a nd ecophysiological attributes of oak species in the eastern United St ates. Annales Des Sciences Forestieres 53 : 487-512. Agee, J. K. 1998. Fire and pine ecosystems. In D. M. Richardson, editor. Ecology and biogeography of Pinus Cambridge University Press, New York, NY. Alexander, L., B. C. Larson, and D. P. Ols on. 1986. The influence of wildfire on eastern white pine regeneration in mixe d hardwood-conifer forests. In D. T. Funk, editor. Eastern white pine: today and tomorrow. General Technical Report WO-51. U.S. Department of Agriculture, Forest Service, Washington, D.C. Andersson, C. 1992. The effect of weevil a nd fungal attacks on the germination of Quercus robur acorns. Forest Ecology and Management 50 : 247-251. Barnett, R. J. 1977. Effect of burial by squirr els on germination and survival of oak and hickory nuts. American Midland Naturalist 98 : 319-330. Borchert, M. I., and S. K. Jain. 1978. Effect of rodent seed predation on four species of California annual grasses. Oecologia 33 : 101-113. Branco, M., C. Branco, H. Merouani, and M. H. Almeida. 2002. Germination success, survival and seedling vigor of Quercus suber acorns in relation to insect damage. Forest Ecology and Management 166 : 159-164.
44 Brewer, S. W. 2001. Predation and dispersal of large and small seeds of a tropical palm. Oikos 92 : 245-255. Boyer, W. D. 1990. Pinus palustris Mill. In R. M. Burns, and B. H. Honkala, technical coordinators. Silvics of North Ameri ca. 1: Conifers. Agriculture Handbook 654. U.S. Department of Agriculture, Forest Service, Washington, D.C. Carey, J. H. 1992. Quercus virginiana In Fire Effects Information System [Online]. U.S. Department of Agriculture, Fore st Service, Rocky Mountain Research Station, Fire Sciences Labor atory (Producer). Available: http://www.fs.fed.us/database/feis/ [April 5, 2006]. Crawley, M. J. 1997. Plant-herbivore dynamics. In M. J. Crawley, editor. Plant ecology. Blackwell Scie nce, Malden, MA. Crawley, M. J., and C. R. Long. 1995. Alternat e bearing, predator sa tiation and seedling recruitment in Quercus robur L. Journal of Ecology 83 : 683-696. Cypher, B. L., and E. A. Cypher. 1999. Ge rmination rates of tr ee seeds ingested by coyotes and raccoons. American Midland Naturalist 142 : 71-76. DeSimone, S. A., and P. H. Zedler. 1 999. Shrub seedling recruitment in unburned Californian coastal sage scrub and adjacent grassland. Ecology 80 : 2018-2032. Dickman, C.R., and C.P. Doncaster. 1989. The ecology of small mammals in urban habitats. II. Demography and disp ersal. Journal of Animal Ecology 58 : 119-127. Forget, P. M., T. Milleron, F. Feer, O. He nry, and G. Dubost. 2000. Effects of dispersal pattern and mammalian herbivores on seedling recruitment for Virola michelii (Myristicaceae) in French Guiana. Biotropica 32 : 452-462. Fornara, D. A., and J. W. Dalling. 2005. Post -dispersal removal of seeds of pioneer species from five Panamanian fo rests. Journal of Tropical Ecology 21 : 79-84. Fralish, J. S., and S. B. Franklin. 2002. Ta xonomy and ecology of woody plants in North American forests. John Wiley and Sons Inc., New York, NY. Fukumoto, H., and H. Kajimura. 2000. Effects of insect predation on hypocotyl survival and germination success of mature Quercus variabilis acorns. Journal of Forest Research 5 : 31-34. Greene, D. F., and E. A. Johnson. 1989. A mode l of wind dispersal of winged or plumed seeds. Ecology 70 : 339-347.
45 Guariguata, M. R., J. J. R. Adame, and B. Finegan. 2000. Seed removal and fate in two selectively logged lowland forests w ith contrasting protection levels. Conservation Biology 14 : 1046-1054. Hatt, R. T. 1943. The pine squirrel in Colorado. Journal of Mammalogy 24 : 311-345 Herrera, C. M., P. Jordano, J. Guitian, and A. Traveset. 1998. Annual variability in seed production by woody plants and the masting concept: reassessment of principles and relationship to pollination and seed dispersal. American Naturalist 152 : 576594. Hollander, J. L., and S. B. Vander Wall. 2004. E ffectiveness of six sp ecies of rodents as dispersers of singl eleaf pinon pine ( Pinus monophylla ). Oecologia 138 : 57-65. Humphrey, S.R. (ed.). 1992. Rare and enda ngered biota of Florida. Volume I: Mammals. University Press, Tampa, Florida. Jensen, T. S., and O. F. Nielsen. 1986. Rodent s as seed dispersers in a heath oak wood succession. Oecologia 70 : 214-221. Keeley, J. E., and P. H. Zedler. 1998. Evolution of life histories in Pinus In D. M. Richardson, editor. Ecology and biogeography of Pinus Cambridge University Press, New York, NY. Kollmann, J., and P. Poschlod. 1997. Popula tion processes at the grassland-scrub interface. Phytocoenologia 27 : 235-256. Lanner, R. M. 1998. Seed dispersal in Pinus In D. M. Richardson, editor. Ecology and biogeography of Pinus Cambridge University Press, New York, NY. LoGiudice, K., and R. S. Ostfeld. 2002. In teractions between mammals and trees: predation on mammal-dispers ed seeds and the effect of ambient food. Oecologia 130 : 420-425. Longland, W. S., and C. Clements. 1995. Use of fluorescent pigments in studies of seed caching by rodents. Journal of Mammalogy 76 : 1260-1266. Longland, W. S., S. H. Jenkins, S. B. Vande rwall, J. A. Veech, and S. Pyare. 2001. Seedling recruitment in Oryzopsis hymenoides : are desert granivores mutualists or predators? Ecology 82 : 3131-3148. Manson, R. H., R. S. Ostfeld, and C. D. Canham. 1999. Responses of a small mammal community to heterogeneity along forest-old-field edges. Landscape Ecology 14 : 355-367.
46 Menges, E. S., and N. Kohfeldt 1995. Life history strategies of Florida scrub plants in relation to fire. Bulletin of the Torrey Botanical Club 122 : 282-297. Mushinsky, H. R., and D. J. Gibson. 1991. The influence of fire periodicity on habitat structure. In S. S. Bell, E. D. McCoy, and H. R. Mushinsky, editors. Habitat structure: the physical arrangement of objects in space. Chapman and Hall, New York, NY. Myers, R. L., and J. J. Ewel. 1990. Ecosystems of Florida. University of Central Florida Press, Orlando, FL. Platt, W. J., G. W. Evans, and S. L. Rathbun. 1988. The population dynamics of a longlived conifer ( Pinus palustris ). American Naturalist 131 : 491-525. Plucinski, K. E., and M. L. Hunter. 2001. Spatia l and temporal patterns of seed predation on three tree species in an oak-pine forest. Ecography 24 : 309-317. Price, M. V., and S. H. Jenkins. 1986. R odents as seed consumers and dispersers. In D. R. Murray, editor. Seed dispersal. Academic Press, New York, NY. Price, M. V., N. M. Waser, and S. McDonald. 2000. Seed caching by heteromyid rodents from two communities: implicat ions for coexistence. Journal of Mammalogy 81 : 97-106. Purves, D. W., and J. Dushoff. 2005. Di rected seed dispersal and metapopulation response to habitat loss and disturbance: application to Eichhornia paniculata Journal of Ecology 93 : 658-669. Rebertus, A.J. 1988. The effects of fire on forest community composition, structure, and pattern in Florida sandhills. Ph.D. Disse rtation, Louisiana State University and Agricultural and Mechanical Co llege, Baton Rouge, Louisiana. Reichman, O. J. 1979. Desert granivore foragi ng and its impact on seed densities and distributions. Ecology 60 : 1085-1092. Richardson, D.M., and P. W. Rundel. 1998. Ecology and biogeography of Pinus : an introduction. In D. M. Richardson, editor. Ecology and biogeography of Pinus Cambridge University Press, New York, NY. Roth, J. K., and S. B. Vander Wall. 2005. Primary and secondary seed dispersal of bush chinquapin (Fagaceae) by scatterhoarding rodents. Ecology 86 : 2428-2439. Shaw, M. W. 1968. Factors affecting na tural regeneration of sessile oak ( Quercus petraea ) in North Wales .2. Acorn losses a nd germination under field conditions. Journal of Ecology 56 : 647-660.
47 Smallwood, P. D., M. A. Steele, and S. H. Faeth. 2001. The ultimate basis of the caching preferences of rodents, and the oak-disper sal syndrome: tannins, insects, and seed germination. American Zoologist 41 : 840-851. Spencer, D. A. 1941. A small mammal community in the Upper Sonoran desert. Ecology 22 : 421-425. Stapanian, M. A., and C. C. Smith. 1986. How fox squirrels influence the invasion of prairies by nut-bearing tr ees. Journal of Mammalogy 67 : 326-332. Steele, M. A., L. Z. HadjChikh, and J. H azeltine. 1996. Caching and feeding decisions by Sciurus carolinensis : Responses to weevil-inf ested acorns. Journal of Mammalogy 77 : 305-314. Steele, M. A., G. Turner, P. D. Smallwood, J. O. Wolff, and J. Radillo. 2001. Cache management by small mammals: experiment al evidence for the significance of acorn-embryo excision. Journal of Mammalogy 82 : 35-42. Stransky, J. J. 1990. Quercus stellata Wangenh. In R. M. Burns, and B. H. Honkala, technical coordinators. S ilvics of North America. 2: Hardwoods. Agriculture Handbook 654. U.S. Department of Agricult ure, Forest Service, Washington, D.C. Streng, D.R., J.S. Glitzenstein, and W.J. Platt. 1993. Evaluating effects of season of burn in longleaf pine forests: a critical lite rature review and some results from an ongoing long-term study. In S. Hermann, editor. Proceedings of the Tall Timbers Fire Ecology Conference, No. 18, The longleaf pine ecosystem: ecology, restoration, and management. Tall Timbers Research Station, Tallahassee, FL. Taulman, J. F., and J. H. Williamson. 1994. Food preferences of captive wild raccoons, Procyon lotor from east Texas. Canadian Field-Naturalist 108 : 170-175. Thayer, T. C., and S. B. Vander Wall. 2005. Interactions between Steller's jays and yellow pine chipmunks over scatter-hoarded sugar pine seeds. Journal of Animal Ecology 74 : 365-374. Thompson, H. R. 1956. Dist ribution of distance to n th neighbour in a population of randomly distributed individuals. Ecology 37 : 391-394. Vander Wall, S. B. 1990. Food hoarding in animals. University of Chicago Press, Chicago, IL. Vanderwall, S. B. 1992. The role of animals in dispersing a wind-di spersed pine. Ecology 73 : 614-621.
48 Vanderwall, S. B. 1994. Seed fate pathways of antelope bitterbrush: dispersal by seedcaching yellow pine chipmunks. Ecology 75 : 1911-1926. Vander Wall, S. B. 2000. The influence of environmental conditions on cache recovery and cache pilferage by yellow pine chipmunks ( Tamias amoenus ) and deer mice ( Peromyscus maniculatus ). Behavioral Ecology 11 : 544-549. Vander Wall, S. B. 2002. Masting in animal-dis persed pines facilitates seed dispersal. Ecology 83 : 3508-3516. Vander Wall, S. B., K. M. Kuhn, and M. J. Beck. 2005. Seed removal, seed predation, and secondary dispersal. Ecology 86 : 801-806. Vander Wall, S. B., and W. S. Longland. 2004. Di plochory: are two seed dispersers better than one? Trends in Ecology & Evolution 19 : 155-161. Wang, Y. S., Z. S. Xiao, and Z. B. Zha ng. 2004. Seed deposition patterns of oil tea Camellia oleifera influenced by seed-caching rodents. Acta Botanica Sinica 46 : 773-779. Weckerly, F. W., D. W. Sugg, and R. D. Se mlitsch. 1989. Germination success of acorns ( Quercus ): insect predation and tannins. Ca nadian Journal of Forest Research 19 : 811-815. Wenny, D. G. 2000. Seed dispersal, seed pr edation, and seedling recruitment of a neotropical montane tree Ecological Monographs 70 : 331-351. West, N. E. 1968. Rodent-influ enced establishment of pondero sa pine and bitterbrush seedlings in central Oregon. Ecology 49 : 1009-1011. Whitney, E., D. B. Means, and A. Rudloe 2004. Priceless Florida: natural ecosystems and native species. Pineapple Press, Sarasota, FL. Xiao, Z. S., Z. B. Zhang, and Y. S. Wang. 2005. The effects of seed abundance on seed predation and dispersal by rodents in Castanopsis fargesii (Fagaceae). Plant Ecology 177 : 249-257.