Ecological Assessments of Impact and Management of Coral Ardisia (Ardisia Crenata), a Shade Tolerant Invasive Shrub in N...

MISSING IMAGE

Material Information

Title:
Ecological Assessments of Impact and Management of Coral Ardisia (Ardisia Crenata), a Shade Tolerant Invasive Shrub in North Central Florida
Physical Description:
1 online resource (117 p.)
Language:
english
Creator:
Celis Azofeifa, Gerardo
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Interdisciplinary Ecology
Committee Chair:
Kitajima, Kaoru
Committee Co-Chair:
Jose, Shibu
Committee Members:
Mack, Michelle C
Macdonald, Greg
Zipperer, Wayne C

Subjects

Subjects / Keywords:
exotic -- herbicide -- invasive -- seedling -- shade -- tolerance -- undestory
Interdisciplinary Ecology -- Dissertations, Academic -- UF
Genre:
Interdisciplinary Ecology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
Undisturbed closed-canopy forests, traditionally thought to be resistant to exotic plant invasion, are shown to be invadable by certain exotic species, primarily shade tolerant trees and shrubs. The potential impacts of understory invaders on community composition, structure, and function of natural forests remain largely unknown. In this dissertation, I investigated several problems relevant for ecology and management of the invasion of closed-canopy hardwood hammock forests of north central Florida by Ardisia crenata, a shade tolerant shrub. First, I investigated the effects of local abundance of A. crenata and abiotic site characteristics on the richness and abundance native understory plants across five mesic forest sites near Gainesville, Florida. In the presence of A. crenata understory species richness declined by 25% and the total understory cover of native species by 34%, affecting all growth forms (trees, shrub, vines, and herbs). Next, I conducted a manipulative field experiment to evaluate the competitive impacts of A.crenata on survival and growth of transplanted seedlings of Quercus virginiana and Q. hemisphaerica in the understory of four forest sites around Gainesville. Seedling survival and growth decreased in the presence of A. crenata over two growing seasons, and the experimental reduction of aboveground competition from A. crenata increased light availability and seedling survival. In the last set of field and greenhouse experiments, I investigated ecological and physiological factors that potentially affected the efficacy of triclopyr, a herbicide widely used for foliar-application to control A. crenata. In the field, I examined root carbohydrate dynamics and efficacy of herbicides as a function of growing season and mowing. I found that herbicide application was effective in growing season regardless of mowing. However,removal of seed sources that occurred with mowing was important for prevention of rapid population recovery. Greenhouse experiments with radio-labeled triclopyr herbicide showed that a the small amount of herbicide was absorbed, but a high proportion was translocated to the roots. In conclusion, my studies support a view that A. crenata has a negative impact on native plants including tree seedlings in the forest understory by competitively reducing light availability. The use of triclopyr herbicide for control is recommended during warm summer months.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Gerardo Celis Azofeifa.
Thesis:
Thesis (Ph.D.)--University of Florida, 2012.
Local:
Adviser: Kitajima, Kaoru.
Local:
Co-adviser: Jose, Shibu.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2013-06-30

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2012
System ID:
UFE0018240:00001


This item is only available as the following downloads:


Full Text

PAGE 1

1 E COLOGICAL ASSESSMENTS OF IMPACT AND MANAGEMENT OF CORAL ARDISIA ( ARDISIA CRENATA ), A SH A DE TOLERANT INVASIVE SHRUB IN NORTH CENTRAL FLORIDA By GERARDO CELIS AZOFEIFA A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2012

PAGE 2

2 2012 Gerardo Celis Azofeifa

PAGE 3

3 To my w ife Gaby Hern ndez my p arents Ana and Rafael and my siblings Ana and Juanra

PAGE 4

4 ACKNOWLEDGMENTS It has been a long process and various institutions and pe rsons have made it possible, enriching and for e most enjoyable to its very end I wish to acknowledge everyone that has be en part of the process. First, Drs. Stephen Hump hrey and Thomas Frazer directors of the School of Natural Resources and the Environment granted me most of institutional support to pursue my graduate studies. Second, I would like to thank my advisor, Dr. Kaoru Kitajima and Co advisor Dr. Shibu Jose for their dedication and support throughout my research. I also wish to express my sincere appreciation of my committee members, Dr s Michelle Mack, Greg MacDonald, and Wayne Zipperer for their valuable insights and of Dr. J. Jack Ewel for support and guidan ce. I would like to thank Michael Meisenburg for help finding field sites and for guidance in the management of exotic invasive plant species For access to sites and permission to conduct research I thank Dr. F. E. (Jack) Putz for the use of his property by the at Biven s Arm and Hogtown Creek, Pam Ganley for Evergreen Cemetery site. Robert Querns for all his help in the lab oratory and greenhouse. The collection of the data a nalyzed in Chapter 2 was supported by the Florida Department of Environmental Protection contract for the period of 2000 2002 to Dr s Alison Fox and Kaoru Kitajima. T he Florida Exotic Pest Plant Council (FLEPPC) funded my herbicide experiment reported in C hapter 4 The University of Florida Natural Area Teaching Laboratory funded research in management of exotic species in natural areas I wish to express m y sincere appreciation to all of these three organization s Finally, I thank my wife, family and frien ds for their unconditional support, without which none of this would have been possible.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ ........ 10 LIST OF TERMS ................................ ................................ ................................ ........... 12 ABSTRACT ................................ ................................ ................................ ................... 13 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 15 In vasion of Forest Understories: T he Role of Exotic Shrubs and Shade Tolerance ................................ ................................ ................................ ............. 15 Forest Understory Invasion ................................ ................................ .............. 15 Shade Tolerant Shrubs of Horticultural Origin ................................ .................. 16 Colonization, Naturalization an d Spread of Shade Tolerant Invaders .............. 17 Overall Objectives of the Study ................................ ................................ ............... 18 2 INVASIVE EXOTIC SHRUB, ARDISIA CRENATA REDUCES NATI VE PLANT DIVERSITY IN FOREST UNDERSTORIES IN FLORIDA ................................ ...... 20 Materials and Methods ................................ ................................ ............................ 23 Design ................................ ................................ ................................ .............. 23 Abiotic Environmental Factors ................................ ................................ .......... 25 Statistical Analyses ................................ ................................ .......................... 26 Results ................................ ................................ ................................ .................... 27 Sites ................................ ................................ ................................ ................. 27 Native Species Richness and Cover ................................ ................................ 28 Multivariate Association of Native Species Cover ................................ ............ 29 Discussion ................................ ................................ ................................ .............. 30 3 INFLUENCE OF SHADE TOLERANT INVASIVE SHRUB, ARDISIA CRENATA ON OAK SEEDLING REGENERATION IN MESIC FOREST IN FLO RIDA ............ 43 Material and Methods ................................ ................................ ............................. 46 Site and Study Design ................................ ................................ ...................... 46 Environ ment Conditions ................................ ................................ ................... 48 Oak Seedlings: Planting and Measurements of Growth and Survival .............. 49 Statistical Analyses ................................ ................................ .......................... 50 Results ................................ ................................ ................................ .................... 51

PAGE 6

6 Site Characteristics ................................ ................................ .......................... 51 Seedling Survival ................................ ................................ .............................. 51 Seedling Biomass ................................ ................................ ............................. 52 Discussion ................................ ................................ ................................ .............. 53 4 DOES HERBICIDE TRANSLOCATION CORRELATE WITH SEASONAL C ARBOHYDRATE BALANCE IN AN EVERGREEN SHRUB ARDISIA CRENATA ? ................................ ................................ ................................ ............. 62 Materials and Methods ................................ ................................ ............................ 66 Field Experiment ................................ ................................ .............................. 66 Herbicide application and efficacy measurements ................................ ..... 68 Biomass allocation and root carbohydrate storage ................................ .... 69 Greenhouse Experiments ................................ ................................ ................. 70 Statistical Analyses ................................ ................................ .......................... 72 Results ................................ ................................ ................................ .................... 74 Field Experiment ................................ ................................ .............................. 74 Effects of season and mowing on root sugar and starch concentrations ... 74 Herbicide efficacy in the field ................................ ................................ ..... 74 Greenhouse Experiments ................................ ................................ ................. 75 Discussion ................................ ................................ ................................ .............. 76 Influence of Herbicide Timing on Efficacy ................................ ......................... 77 Influence of Mowing on Herbicide Efficacy ................................ ....................... 77 Herbicide Translocation ................................ ................................ .................... 78 5 CONCLUSIONS ................................ ................................ ................................ ..... 97 APPENDIX A ADDITIONAL TABLES AND FIGURES for chapter 2 ................................ ............. 99 B ADDITION AL FIGURES FOR CHAPTER 3 ................................ .......................... 104 C ADDITIONAL FIGURES FOR CHAPTER 4 ................................ .......................... 110 LIST OF REFERENCES ................................ ................................ ............................. 112 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 117

PAGE 7

7 LIST OF TABLES Table page 2 1 Study site locations and soil characterisctics ................................ ..................... 34 2 2 Percent ground cover for each growth form in A. crenata presence and absence ................................ ................................ ................................ ............. 35 2 3 Eigenvectors of principal component analysis for A. cren ata cover, understory native species number, understory native species cover, overstory native species cover, soil moisture, percent light, and diversity ......... 35 2 4 Eigenvectors of principal comp onent analysis for A. crenata cover, understory native species number, understory native species cover for growth forms, overstory native species cover, soil moisture, percent light ........ 36 3 1 St udy site location, soil characteristics, and A. crenata biomass ....................... 56 3 2 Logistic mixed model results for seedling survival 240 days after transplanting of the two oak species and two treatment s ( Ardisia crenata presence and absence) ................................ ................................ ..................... 57 3 3 Logistic mixed model results for seedling survival 600 days after transplanting of the two oak species and two treatments ( Ardisia crenata pres ence and absence) ................................ ................................ ..................... 57 3 4 Logistic mixed model results for seedling survival 240 days after transplanting of the two oak species and two treatments ( Ardisia crenata canopy pull down and no pul l down) ................................ ................................ 58 3 5 Logistic mixed model results for seedling survival 600 days after transplanting of the two oak species and two treatments ( Ardisia crenata canopy pull down and no pull down) ................................ ................................ 58 3 6 Linear mixed model results for two oak seedling biomass at 600 days after transplanting comparing three treatments ( Ardisia crenata absent, A. crenata no pull down, and initial harvest) ................................ ................................ ....... 59 3 7 Linear mixed model results of two oak seedling biomass at 600 days after transplanting comparing three treatments ( Ardisia crenata no pull down, A. crenata pull down, and initial harve st) ................................ ............................... 59 4 1 Study site locations ................................ ................................ ............................ 80 4 2 Field experiment biomass and leaf area of harvested Ardisia crenata individuals ................................ ................................ ................................ .......... 80

PAGE 8

8 4 3 Greenhouse experiment biomass and leaf area of harvested Ardisia crenata individuals ................................ ................................ ................................ .......... 81 4 4 Linear mixed model results for roo t starch concentration of Ardisia crenata in mowed and unmowed fields at herbicide application date ................................ 81 4 5 Linear mixed model resutls for root simple sugar concentration of Ardisia crenata in mowed and unmowed fields at herbicide application date ................ 81 4 6 Linear mixed model results for herbicide efficacy index for adult plants after 6 and 12 months following the four herbicide appl ication dates in the mowed and unmowed fields ................................ ................................ ........................... 82 4 7 Linear mixed model results for the herbicide efficacy index for seedlings after 6 and 12 months after the four herbicide application dates in the mowed and unmowed fields ................................ ................................ ................................ .. 82 4 8 Analysis of variance results for root starch concentration of Ardisia crenata plants grown under low and high light treatments in the greenhouse ................ 82 4 9 Analysis of variance results for root simple sugar concentration of Ardisia crenata plants grown under low and high light treatments in the greenhouse ... 83 4 10 Analysis of variance restuls for radioactivity of 14 C triclopyr in Ardisia crenata plants grown under low and high light treatments in April 2011 ......................... 84 4 11 Ana lysis of variance resutls for radioactivity of 14 C triclopyr in Ardisia crenata plants grown under low and high light treatments in October 2011 ................... 85 4 12 Radioactivity for leaf water wash, tot al absorbed, the treated leaf, and translocation in April 2011 ................................ ................................ ................. 86 4 13 Radioactivity for the total recovery and leaf wash in October 2011 ................... 86 4 14 Radioactivity for treated leaf under two light treatments in October 2011 .......... 86 4 15 Analysis of variance resutls for radioactivity of 14 C triclopyr translocated to different plant tissues (leaves, stems, roots, meristems) of Ardisia crenata plants grown under low and high light treatments overtime in April 2011 .......... 87 4 16 Radioactivity under two light treatments in leaves, meristems, stems and roots at April 2011 ................................ ................................ ............................. 87 4 17 Radioactivity across time in leaves, meristems, stems and roots at April 2011 ................................ ................................ ................................ .................. 88 4 18 Analysis of variance resutls for radioactivity of 14 C triclopyr translocated in Ardisia crenata plants grown under low and overtime in October 2011 ............. 88

PAGE 9

9 4 19 Radioactivity fou nd across time in leaves, meristems, stems and roots at October 2011 ................................ ................................ ................................ ..... 89 A 1 Percent cover of native and exotic species for forest understory and overstory ................................ ................................ ................................ ............ 99

PAGE 10

10 LIST OF FIGURES Figure page 2 1 Experimental design and percent cover of Ardisia crenata ................................ 37 2 2 Species accumulatio n curves of forest understory native species ..................... 38 2 3 Forest understory native species richness in relation to A. crenata cover ......... 39 2 4 Forest understory native species cover in A. crenata invaded and uninvaded plots ................................ ................................ ................................ ................... 40 2 5 P rincipal component analysis correlation biplot for all species .......................... 41 2 6 Principal component analysis correlation biplot for growth forms ...................... 42 3 1 Probability of survival for each oak individual seedling at each site for 240 a nd 600 days census based on generalized linear mixed model. ...................... 60 3 2 Oak seedling biomass for initial harvest and treatments ................................ ... 61 4 1 Schematic of proposed mechanism of carbohydrate movement in a forest understory evergreen plant in relation to seasonal light availability ................... 90 4 2 Herbicide field experiment setup ................................ ................................ ....... 91 4 3 Field experiment plots with herbicide barrier ................................ ...................... 92 4 4 Plot photos for October herbicide application (field experiment) ........................ 93 4 5 Seasonal total non structural carbohydrates at each herbicide application date in the field for mowed and unmowed adult A. crenata plants .................... 94 4 6 Herbicide efficacy after 6 and 12 months after herbicide treatment application date in the field for mowed and unmowed adult A. crenata plants .................... 95 4 7 Seasonal total non struct ural carbohydrates at each herbicide application in the greenhouse for shaded and sun A. crenata plants ................................ ...... 96 A 1 Experimental setup for each site ................................ ................................ ..... 103 B 1 Monthly temperatures during study period taken from nearest meteorological station to study sites at Gainesville, Florida, USA ................................ ........... 104 B 2 Monthly precipitation during s tudy period taken from nearest meteorological station at Gainesville, Florida, USA ................................ ................................ 105

PAGE 11

11 B 3 Mean monthly temperatures during 27 years (1984 to 2011) at Gainesville, Florida, USA ................................ ................................ ................................ .... 106 B 4 Monthly precipitation during 27 years (1984 to 2011) at Gainesville, Florida, USA ................................ ................................ ................................ ................. 107 B 5 Oak seedling biomass for initial and treat ments (zeros excluded) ................... 108 B 6 Light availability for plots without Ardisia crenata (Absent), plots with A. crenata canopies pulled down (Pull down), and plots with A. crenata canopy intact ( No Pull down) ................................ ................................ ....................... 109 C 1 Mean monthly temperatures during 27 years (1984 to 2011) at Gainesville, Florida, USA ................................ ................................ ................................ .... 110 C 2 Monthly pr ecipitation during 27 years (1984 to 2011) at Gainesville, Florida, USA ................................ ................................ ................................ ................. 111

PAGE 12

12 LIST OF TERMS S PECIES A populations of organisms c apable of interbreeding and producing fertile offspring. E XOTIC A species found outside its na tive range because of human mediated transportation. I NVASIVE P lant species whose populations expand explosively in new environment, with significant impacts on local species. P ROPAGULE A structure in a plant from which a new individual may rise, such as s eeds. S HRUB Perennial, multi stemmed woody plant that is usually less than 5 meters (16 feet) in height. Shrubs typically have several stems arising from or near the ground, but may be taller than 5 meters or single stemmed under certain environmental cond itions (USDA)

PAGE 13

13 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy ECOLOGICAL ASSESSMENTS OF IMPACT AND MANAGEMENT OF CORAL ARDISIA ( ARDISIA CRENATA ), A SH A DE TOLERANT INVASIVE SHRUB IN NORTH CENTRAL FLORIDA By Gerardo Celis Azofeifa December 2012 Chair: Kaoru Kitajima Cochair: Shibu Jose Major: Interdisciplinary Ecology Undisturbed closed canopy forests, traditionally thought to be resistant to exotic plant invasion, are shown to be invadable by certain exotic species, primarily shade tolerant trees and shrubs. The potential i mpacts of understory invaders on community composition, structure, and function of natural forests remain largely unknown. In this dissertation, I investigated several problems relevant for ecology and management of the invasion of closed canopy hardwood h ammock forests of north central Florida by Ardisia crenata a shade tolerant shrub. First, I investigated the effects of local abundance of A. crenata and abiotic site characteristic s on the richness and abundance native understory plants across five mesic forest sites near Gainesville, Florida. In t he presence of A. crenata understory species richness declined by 25% and the total understory cover of native species by 34% affecting all growth forms (trees, shrub, vines, and herbs). Next, I conducted a ma nipulative field experiment to evaluate the competitive impacts of A. crenata on survival and growth of transplanted seedlings of Quercus virginiana and Q. hemisphaerica in the understory of four forest sites around

PAGE 14

14 Gainesville. Seedling survival and growt h decreased in the presence of A. crenata over two growing seasons and the experimental reduction of aboveground competition from A. crenata increased light availability and seedling survival. In the last set of field and greenhouse experiments, I investi gated ecological and physiological factors that potentially affect ed the efficacy of t riclopyr, a herbicide widely used for foliar application to control A. crenata In the field, I examined root carbohydrate dynamics and efficacy of herbicide s as a functi on of grow ing season and mowing. I found that herbicide application was effective in growing season regardless of mowing However, removal of seed sources that occurred with mowing was important for prevent ion of rapid population recovery. G reenhouse exper iments with radio labeled triclopyr herbicide showed that a the small amount of herbicide was absorbed but a high proportion was translocated to the roots. In conclusion, my studies support a view that A. crenata has a negative impact on native plants in cluding tree seedlings in the forest understory by competitively reducing light availability. The use of t r i clopyr herbicide for control is recommended during warm summer months.

PAGE 15

15 CHAPTER 1 INTRODUCTION Invasion o f Forest Understories: T he Role of Exotic S hrubs a nd Shade Tolerance F orest U nderstory I nvasion Invasion by exotic (non native) plant species has become a growing concern worldwide in recent decades. Invasions occur in a wide range of terrestrial and aquatic ecosystems around the globe. The process of invasion requires an exotic species to disperse to adequate habitat s establish and persist in the new community (Catford, Jansson, & Nilsson 2009) Humans often facilitate dispersal in particular through horticultural and agricultural trades The process of establishment is also facilitated by changes (enrichment or release) of resources in disturbed ecosystem (Davis, Grime, & Thompson 2000) Human induced disturbances, both those ana logous to natural disturbances and novel types, are becoming more prevalent (Vitousek et al. 1997) especially where exotic species propagule pressures are high (Vil & Ibez 2011) leading to increased cases and extents of invasion s by exotic species (Bradley & Mustard 2006) In general, disturbance is required by many exotic species that ar e pre adapted to disturbance and/or high resource conditions for colonization and population growth (Sax & Brown 2000) Because of the disturbance dependent life histor y of many invasive exotics, undisturbed systems are considered to be less vulnerable to invasion. Some consider that undisturbed ecosystems, especially species rich systems such as tropical forests, are resistant to invasion (Elton 1958) because in intact species rich systems, no empty niches are available for alien species to colonize.

PAGE 16

16 However, the growing evide nce suggests the contrary Once abundant, exotic species impacts can affect community structure and ecosystem functions ( Vil et The impact of invasive exotics on species diversity may be dependent on spatial scales; there is a negative association between native and exotic species richness at small spatial scales, whereas at large scales there is a positive association (Fridley et al. 2007) Shade T olerant S hrubs of H orticultural O rigin Mature forested ecosystems with closed canopies are a good example of an undisturbed ecosystem, where resources such as light are a limiting factor fo r plant growth and establishment. Many of the species invading these ecosystems still require natural disturbances such as tree fall gaps to establish and then can continue to survive after canopy closure (Gorchov et al. 2011) These requirements are similar to the life history traits of many resident trees and shrubs (Richardson & Rejmnek 2011) However, t here may be a group of species that do not require disturbances to establish and persist in forest understories. This group is dominated by shade tolerant shrubs (Martin, Canham, & Marks 2008) Human mediated transport is the mechanism of exotic species movement around the world including shrubs The horticulture industry has played an important role in such transport; 31% of all exotic invasive shrubs in the world were introduced by horticul ture (Richardson & Rejmnek 2011) In regions with the pre sence of high numbers of exotic invasive tree and shrub species (more than 100 exotic invasive species) such as North America, 77% of all invaders were introduced by the horticulture industry (Richardson & Rejmnek 2011) In

PAGE 17

17 an effort to limit introductions of potentially invasive plants, the State of Florida has regulations restricting the introduction of exotic species shown to be detrimental (FLEPPC 2011) Ardisia crenata Sims. (Myrsinaceae) represents a clear example of shade tolerant invasive exotic shrub s of horticultural origin. A. crenata was introduced and promoted by the horticul ture industry as an ornamental (Wirth, Davis, & Wilson 2004) Photosynthetic light response curves of A. crenata exhibits a low 2 s 1 which allows it to persist in shade ( Gerardo Celis, unpublished data ) A. crenata is capable of forming dense monodominant patches (Burks & Langelan d 1998) with cover reaching >90% of ground (personal observation) and 300 stems per m 2 (Kitajima et al. 2006) Dispersal is limited, but birds are the main dispersers (Meisenburg 2007) It is native to east Asia (mainly from southern China to southern Japan) and g enetic analyses suggest that A. crenata came to Florida from southeastern China multiple times and t hen spread from there (Niu et al. 2012) but horticultural trade s somewhat obscure this pattern (Dozier 1999; Kitajima et al. 2006) Colonization, N aturalization and S pread of S hade T olerant I nvaders Light availability under closed forest canopies is low (9% of incident radiation in southern hardwood forests; Canham et al. 1990) and it constrains growth and survival of many plants including seedlings of overstory sp ecies. Survival in shade depend s on morphological physiological and genetic characteristics that contribute to maintenance of positive carbon balance. Such characteristics include not just optimization of shade light utilization (Chazdon & Field 1987; Lusk et al. 2011) but also defens ive t raits against herbivores and

PAGE 18

18 pathogens (Kitajima 1994; Alvarez C lare & Kitajima 2007) and storage that allow survival during periods of negative carbon balance (Poorter & Kitajima 2007; Myers & Kitajima 2007) In case of exotic invaders, also important are the traits that allow them to compete with na tive species (Keane & Crawley 2002) For instance, garlic mustard Alliaria petiolata an understory exotic invasive forb exhibits low degrees of herbivory (Ricklefs 2010) Success of exotic invasive species is often attributed to escape from natural enemies, but success may also be the result of successful acquisition of resources including light. W oody exotic species displayed traits significantly more conductive of resource acquisition than native species (higher specific leaf a rea, larger and thinner leaves, lower wood density) (Tecco et al. 2010) An alternative hypothesis by which understory exotic species successfully invade is that they reduce resources available to competitors. The exotic shrub, Lonicera maackii invading forests in eastern United States has shown to reduce the amount of understory light available to other species and therefore fac ilitate its own invasion by competitive suppression (Miller & Gorchov 2004) Overall O bjective s of the S tudy Given the importance and the potential impact s of shade tolerant exotic invasion on forest understories an integral approach that consider s ecological and life history characteristics of th ese type s of invaders is needed for effective management. The process in search of such an approach should include 1) i dentif ication of the impacts of an exotic species on ecosystems, 2) understanding of the mechanisms by which exotic species produce impacts, 3) identification of the be st control methods to reduce impacts and 4) evaluat ion of

PAGE 19

19 the economic costs and benefit (Epanchin Niell & Hastings 2010) and public willingness to address control methods (Garca Llorente et al. 2008) This di ssertation is an effort toward development of such an integral approach. Chapter 2 assesses the impacts of a shade tolerant exotic invasive shrub, A. crenata in the understory community of a closed canopy forest; how the native understory species richness and cover are associated with the local abundance of A. crenata when abiotic environmental variables such as soil moisture and light availability are simultaneously considered. I also ask how these associations may differ among plant growth forms. In chap ter 3, I evaluate one of the potential mechanisms by which A. crenata displaces native species the impact of light competition from A. crenata on survival and growth of seedlings of two common canopy tree species when the influence of micro environmental varia tions are considered simultaneously. Chapter 4 evaluates factors that influence efficacy of herbicide control of A. crenata ; explor ing the effects of mowing on herbicide efficacy and the impact of seasonal variation o n herbicide translocation

PAGE 20

20 CHAP TER 2 INVASIVE EXOTIC SHRU B, ARDISIA CRENATA REDUCES NATIVE PLA NT DIVERSITY IN FOREST UNDERSTORIES IN FLOR IDA Exotic plant invasions occur in a wide range of terrestrial and aquatic ecosystems around the globe. The process of invasion requires exotics spe cies to disperse to adequate habitat s establish and persist in the new community. Invasion by exotic species is generally linked to disturbance associated resources pulses in the ecosystem (Davis et al. 2000) Disturbances can be nat ural or anthropogenic, and can vary in magnitude, ranging from branch and tree falls to large blow downs by hurricanes in forested ecosystems. However, anthropogenic disturbances are becoming more prevalent (Vitousek et al. 1997) especially where exotic species propagule pressures are high (Vil & Ibez 2011) and consequent ly are particularly conductive to exotic invasion (Bradley & Mustard 2006) While disturbance associated resource fluctuations are important in facilitating colonization and initial population growth by exotic species adapted to these changes (Sax & Brown 2000) once established, such species are shown to alter community structure and ecosystem functions to further promote their population growth On the other hand, undisturbed ecosystems, in particular species rich ecosystems, are considered to be more resistant to inv asion by exotic species (Levine, Adler, & Yelenik 2004) The basis of this view is that all potential ecological niches are occupied by species already present in the community, and resource competition among them results in resistance against invasion by exotic species (Elton 1958) The understories of closed canopy forests, where resources such as light are a limiting factor for plant growth and establishment,

PAGE 21

21 are often viewed as relatively undisturbed and invasion resistant. Many o f the species invading the forest understory still require natural disturbances such as tree fall gaps to establish, even though they may continue to survive after canopy closure (Gorchov et al. 2011) These requirements are similar to the life history traits of many resident trees and shrubs (Richardson & Rejmnek 2011) However, there may be a group of in vaders that do not require disturbances to establish and persist in forest understories. This group is dominated by shade tolerant shrubs (Martin et al. 2008) Their impacts to understory comm unity as well as the recruitment of overstory species need to be evaluated. invasion of shade tolerant invasive species. Hardwood hammocks are characterized by multiple layers of tree s, shrubs and herbs, dominated by a dense canopy consisting of a mix of evergreen and deciduous trees (Veno 1976) They can be further classified by the degree of water availability (xeric, mesic, and hydric) (Vince, Humphrey, & Simons 1989) The north central Florida hardwood hammocks are in a transitional zone from the southern mixed hardwood forest s to the tropical forest s of southern Florida (Platt & Schwartz 1990) Do minant species are broad leaved evergreen ( e.g., Quercus virginiana and Magnolia grandiflora ) needle leaved evergreens ( e.g., Pinus glabra and P. taeda ), and deciduous hardwoods ( e.g., Liquidambar styraciflua and Carya glabra ). Florida has a long history of exotic species introductions (Gordon & Thomas 1997) and about 1,400 species currently form part of the resident flora.

PAGE 22

22 Approximately 11% have become invasive (FLEPPC 2011) and threaten many of Despite having high species diversity (Monk 1965) hardwood hammocks are being invaded by exotic trees ( e.g., Cinnamomum camphora ) vines ( e.g., Discorea bulbifera ), herbs ( e.g., Tradesca ntia fluminensis ) and shrubs including Ardisia crenata (Burks & Langeland 1998) A. crenata is a shade tolerant shrub, which can grow and reproduce under very low light conditions (Kitajima et al. 2006) In spite of growing recognition of the potential impact s of forest understor y invasion by shade tolerant shrubs, quantitative assessments of their impacts are rare co mpared to invaders of other types of habitats In this study, we quantified how A. crenata may affect diversity, richness and structure of native plant communities in the understory of hardwood hammocks More specifically we addressed the following three questions : 1. How are understory native species richness and cover associated with presence and abundance of A. crenata ? 2. How are forest understory species richness and cover, as well as A. crenata cover, associated with abiotic environmental variables such a s soil moisture and light availability? 3. Are these associations similar regardless of plant growth form ? We predicted a negative association between A. crenata abundance and native understory species richness and cover. This negative association is expecte d to be stronger with native trees and shrubs growth forms than herbs and vines because similar life forms with similar resource requirements may compete more with each other

PAGE 23

23 Materials and Methods Exotic invasive plant species impacts on ecosystems are s ometimes difficult to quantify due to the lack of prior knowledge of the state of the ecosystem before plant invasion Plant invasion usually occurs from a focal point and then begins to spread to peripheral areas. The spread will be determined by dispersa l capacity of the species into new areas. A. crenata fruits are rarely dispersed and they can persist up to a year on the plant (Meisenburg 2007) The limited dispersal results in high conc entration of seedlings (~ 600 individuals per square meter ) can be found under adult plants (Kitajima et al. 2006) and slow spatial spread. Personal observations of heavily inv aded sites around Gainesville over multiple years has witnessed a reas adjacent to the focal points of A. crenata invasion under the similar environmental conditions became invaded overtime Design W e selected five mesic hardwood forest sites near Gainesvi lle, Florida where dense patches of A. crenata appeared to be actively expanding (i.e., many large reproductive adults surrounded by smaller individuals at the periphery) All sites were relatively undisturbed forests dominated by broadleaf evergreen and deciduous canopy trees, such as Quercus spp. Two sites were within protected natural areas (San Felasco State Forest (SF) Coclough Pond Nature Park (NL) ) ; while others were private lands adjacent to public natural areas (Micanopy (MC), ake (NL), ; see Table 2 1). In communication with the landowners and land managers of these sites, we ensured that there were no active removal efforts before the end of 2001 when this study was completed. At each site, we located a dense pa tch of A. crenata

PAGE 24

24 and approximated the position of the central invasion point according to the presence of large sized reproductive adults of A. crenata (e.g., multi stemmed individuals > 1 m in height with fruits). This position was marked with a rebar f or the duration of the study. A. crenata stem density and native plant cover within and around each patch were estimated in four 1 m wide transects radiating from the central invasion point in a stratified randomized manner; one transect radiated from the center point in a randomly chosen compass direction within each 90 o quadrant (0 90, 90 180, 180 270, and 270 360 o ). Within each transect, we recorded presence of A. crenata individuals greater than 20 cm in height at every 1 m segment A. crenata may be potentially reproductive above 20 cm in height (Kitajima et al. 2006) and this minimum size threshold also ensured consistent detection threshold because smaller individuals c an be easily overlooked. Each transect was extended 10 m beyond the distance at which the last A. crenata individual was observed (e.g., if the last A. crenata was observed at 16 m from the central point along a particular transect, the length of this tran sect was 26 m including 10 m in which no A. crenata occurred). The size of the invaded area was a polygon defined by this location of the last A. crenata along the four radiating transects (Figure 2 1), while area beyond this zone was considered to be in t was also terminated. We chose five random distances within the invaded area and three in the uninvaded area to set 1 m x 1 m plots along each transect. In each plot, we

PAGE 25

25 quantified th e density and percent cover of A. crenata and recorded the identity, percent cover of all native plant species in the understory (below 0.8 meters above the ground within 1 m x 1 m PVC pipe frame ) and overstory (by visual approximation). Other exotic und erstory species were accounted for in each plot ( a total of 8 species across sites, with the average cover of 1.7% ) but were excluded from statistical analysis. Also recorded was the percent cover of the bare ground if present. The survey was repeated in spring (April) and fall (August) of 2001, so that both early emerging and late emerging species could be accounted for. A total of 99 species were recorded across the four sites, including six species that were encountered only in the fall survey ( Appendix A Table A 1). The percent cover in the overstory, including all vegetation above 0.8 meter s height, was approximated in the same manner, often resulting in greater than 100 percent cover due to overlapping foliage. Abiotic E nvironmental F actors Soil was sampled with a 2.5 cm diameter x 20 cm deep corer after removing litter layer. Two cores were collected from each transect, one randomly chosen plot inside and another from the un invaded area. Thus, the total number of soil cores per site was eight. The s oil was brought back to the lab to determine gravimetric soil moisture after drying to a constant mass at 105C. Soil moisture (volumetric) was also measured in all plots with a soil moisture probe ( Theta probe type ML1 Delta T devices, Cambridge England ) calibrate d against gravimetric soil moistures of sample cores from the same locations Soil was sampled once within two days in May. No rain events occurred in the area during four days prior to the measurements. Four replicate measurement s were

PAGE 26

26 taken w ithin each plot to account for spatial heterogeneity and the plot mean moisture was used for statistical analysis. Dried core samples were homogenized and analyzed for nutrient contents in the Analytical Research Laboratory of the University of Florida. M ehlich 1 extraction was made from each subsample of 5 g, for which phosphorus (P), potassium (K), calcium (Ca) and Magnesium (Mg) concentrations were determined with the ICP method following EPA Method 200.7. Total organic matter content was estimated with the Walkley Black (WB) method. Light (photosynthetic active radiation, PAR) was measured at 0.8 meters above the ground (above A. crenata canopy) with a line quantum sensor (LI COR Inc., Nebraska USA) once during Spring season and expressed relative to the r eference PAR taken simultaneously with a quantum sensor and data logger in the nearest site under 100% open sky. Statistical A nalyses All statistical analyses were conducted using R (R Development Core Team 2012) and used a significance level of P = 0.05. A linear mixed model was used to test differences of soil characteristics between invaded and uninvaded area s The model included invaded status (in and out) as fixed effect and site as a random effect Variables evaluated were s oil moisture and nutrients and were transformed (natural log for nutrients ) to satisfy the normality assumption A T ukey post hoc test was used to identify differences among levels. A. crenata cover was compared among sites with a non parametric tests of Kruskal Wallis and a Nemenyi Damico Wolfe Dunn po st hoc test (Hollander & Wolfe 1999) w as implemented to test differences between sites. Species

PAGE 27

27 accumulation curves were estimated to compare species richness between plo ts with A. crenata presence and absence using R package vegan (Oksanen et al. 2012) which calculates mean and standard deviation from random permutations of the data. A Sha n non Wein er index (Diversity) was calculated for each individual plot. A generalized linear mixed model assuming Poisson distribution and treating site as a random factor was used to evaluate the relationship of native species richness and A. crenata cover. In orde r to account for site variability, site was set as a random effect. In order to summarize how native species cover, richness and diversity were associated with soil moisture, percent light, overstory cover and A. crenata cover we used a principal componen t analysis (PCA). The Kraiser Guttman criterion was used to determine significan ce of eigenvalues. Further, PCA was run after separating native species richness by growth forms ( Tree Herb Vine and Shrub ) to examine which of these life forms may show stron ger negative association with A. crenata abundance Light and soil moisture were log transformed to approximate normality. Results Sites A. crenata the five sites; BM = 550.3 m 2 CP = 510.7 m 2 NL = 3,084.8 m 2 PR = 133.2 m 2 SF = 989.7 m 2 ( Appendix A Figure A 1). The five sites also differed in soil nutrients; phosphorus (F 4, 34 = 17.1 P < 0.001), potassium (F 4, 34 = 4.9 P = 0.001), magnesium (F 4, 33 = 5.8 p = 0.008) and calcium (F 4, 34 = 4.2 P = 0.006). Organic matter did not differ (F 4, 34 = 0.5 P = 0.7 7 ) with ake (NL)

PAGE 28

28 being the least fertile among the four (Table 2 1). Soil moisture was the lowest at whereas other three sites had similar levels (F 4,1 52 = 21. 2 P < 0.001) (Table 2 1). A. crenata cover significantly differed among the five sites ( = 20. 8 P < 0.0 01 ) L ake (NL) compared to other three sites (Figure 2 1 ). Among abiotic environmental factors measured, only soil moisture was different between A. crenata invaded plots with uninvaded plots (F 1,1 51 P = 0.0 3 ) uninvaded plots having higher soil moisture Native S pecies R ichness and C over Native understory species richness in presence of A. crena ta was 25% lower (61 species, 84 plots) than where A. crenata was absent (81 species, 73 plots) (Figure 2 2 ). Based on the species accumulation curves, sampling effort was sufficient to detect differences between plots with A. crenata presence and absence. A minimum of 33 m 2 of sampled area or 33 plots was required to detect differences. The mean species richness for forest understory species was compared among sites at a comparable sample area, and it was highest at San Felasco State Preserve (SF) 49 speci es (32 m 2 sampled area) followed by Colclough Pond (CP) 39 species (32 m 2 ), Micanopy ( MC ) 37species (32 m 2 ), 34 species (32 m 2 ), and ake (NL) 3 1 species (29 m 2 ). Native understory species richness was negatively associated to A. crenata cover ( = 23.5, P < 0.00 1), indicating that increases in A. crenata cover reduce d native species richness. This trend was similar among sites (Figure 2 3 ). Native understory species cover was reduced by the presence of A. crenata Invaded plots had on average 34% less cover (Figure 2 4 ) However, A.

PAGE 29

29 crenata was not merely replacing native species cover but it increase d total ground cover by reducing bare ground Overall m ean total cover per plot including A. crenata was 9.3% high er ( = 11.2, P < 0.001) than plots without A. crenata A ll growth forms were negatively associated with the presence of A. crena ta ; herbs ( = 5.5 P = 0.0 2), vines ( = 4.0 P = 0.0 46), and most notably for trees ( = 8.6 p = 0.00 3), and shrubs ( = 10.4 P = 0.00 1) (Table 2 2). Multivariate A ssociation of N ative S pecies C over The PCA plot shows that native species richness and cover were negatively associated with A. crenata cover, but were largely independent of soil moisture, light, and overstory cover (Figure 2 5 Table 2 3). A. crenata cover (Accov) had the highest loading on principal component 1 (PC1), which accounted for 33% of the total variation. This strong contribution of A. crenata cover to PCA was not unexpected given the sampling design that attempted to span a wide variation in local density of A. crenata Both richness (UnspNo) and cover (Uncov) of understory native species showed strong negative loading to PC1, and Shan n on Weiner dive rsity index showed weaker negative loading with PC1 By contrast, overstory cover showed little relationship with PC1. The second principal component (PC2) accounted for 22% of the total variation mostly in relation to abiotic environmental variables of s oil and light Overstory cover (Ovcov) and moisture (Moist) had negative loadings while light availability (Light) had positive loading, indicating greater overstory cover meant lower light availability and higher soil moisture (Figure 2 5 )

PAGE 30

30 When understo ry cover was considered separately by growth forms (tree, shrub, vine and herb) in PCA (Figure 2 6 ), PC1 accounted for 22 % of the total variation, again largely explained by A. crenata cover (Accov) with high positive loading and understory native species richness (UnspNo) and herbaceous cover (Herb) with large negative loadings (Table 2 4 ). Shrub cover (Shrub) and Tree cover (Tree) were also negatively associated with PC1, but less strongly so. PC2 accounted 18 % of the total variation, reflecting variation s in understory tree species cover (Tree, positive), light (Light, positive) and Overstory and shrub cover (Ovcov and Shrub, negative) in the order of factor loading (Table 2 4 ). These relationships indicate a difference in light requirements between tree s and shrubs. Moisture was not significantly associated with the first two principal components. Discussion Our findings demonstrate that mesic hardwood hammock forest understories are not immune to the invasion of shade tolerant exotic shrubs and their imp acts. The presence of Ardisia crenata shrub was negatively associated with understory community species richness and cover. Overall, native species richness was reduced by 25% and cover by 34% indicating that A. crenata modified understory community struc tu re In a temperate forest, a shrub, Lonicera maack ii is shown to reduce species richness by 53% and cover by 63% on average with greater reduction with increasing residence time of L. maackii (Collier, Vankat, & Hughes 2002) We did not know the residence t ime of A. crenata at each site, but within and across the five sites, native species richness was negatively associated with local abundance of A. crenata Not only was A.

PAGE 31

31 crenata decreasing native cover, it also increas ed total cover by 9.3% making unders tor ies more dense ly vegetated Invasion by exotics may cause a species composition shift in favor of specific growth forms. A. crenata had a negative effect on all growth forms (trees, shrubs, vines and herbs). We found that taking into account the influen ce of A. crenata understory cover of t ree species was more strongly associated with light than shrubs. This means that tree species require more light than shrubs for persistence in the understory (Herault et al. 2011) The shrub and t ree understory dynamics of hard wood forests in the northeast U.S. are negatively associated to each other (Ehrenfeld 1980) where understory species suppress overstory species and vice versa. Th is interaction between shrubs and trees is disrupted by A cre n a ta impact ing all growth form s in southern hardwood hammocks O ver time this could lead to a homogenization of species in the understory. Exotic species impacts are sometimes confounded with c hanges in environmental variables, thus exaggerating the impact attributable to exotic species (Surrette & Brewer 2009) The unaccounted factors can sometimes result in positive correlation s between native and exotic species (Gilbert & Lechowicz 2005) Alterations in community species composition where exotic species have invaded can be due to interaction of exotic species and ecosystem change This is likely the case of disturbed ecosystems in which exotic ruder al species invade. However, exotic species can be the direct cause of change in the species composition by introducing novel traits or functions to an ecosystem (Bauer 2012) The impact s of A. crenata on native species richness and cover

PAGE 32

32 were independent of abiotic micro environmental variable s affecting key resources in the understory, such as light. A. crenata was not on ly replacing n ative species cover, but also making understories more densely vegetated. Hardwood hammock forests of Florida have a history of complex disturbance regimes (fires and hurricanes) of varying scales over space and time These ha ve creat ed multilayered struct ured communities rich in species and depend ent o n cyclic disturbance regime s (Platt & Schwartz 1990) E xploitation of light after disturbances in the understories plays an important role in growth and survival of individuals often facilitating population establishment of invasive exotic plants Subsequently f orest understory communities can be affected by changes in the environmental condition created by invading species. For example, shade cast by Acer negundo reduce s native species richness by 45% and aboveground biomass by 81% in riparian forests in SE Eur ope where A. negundo is an exotic invasive tree (Bottollier C urtet et al. 2012) The authors of this paper consider that the observed changes to light levels were novel to the system and hence reduction of native species Similarly, A. crenata has crown leaf display characteristics that reduc e light availability un derneath (Kitajima et al. 2006) The results of this study suggest that A. crenata impacts on hardwood hammocks community can threaten key processes and functions of this ecos ystem. What are some of the mechanisms that allow A. crenata compete with understory native species? Further research is needed to understand these

PAGE 33

33 mechanisms, which will enable us to quantify impacts and to design effective mitigati on effects One of such mechanisms will be examined in the next chapter.

PAGE 34

34 Table 2 1. Geographical coordinates soil characteristics and %light relative to a nearby open site ( measured at height of 80 cm above ground) of the five study sites in Alachua County, Florida, USA. Means and standard deviations, significant statistical differences between sites or A. crenata invaded zone are identified by different superscript letters Means ( standard deviation) are shown for each site, as well as for the data pooled across sites for inv aded vs. uninvaded plots. Variables were transformed using natural log to satisfy normality. Site (abbreviation) Latitude & Longitude Soil Moisture (m 3 /m 3 ) P (mg/kg) K (mg/kg) Ca (mg/kg) Mg (mg/kg) Org. M. (%) Light (%) Micanopy (MC) 2935'N, 8222'W 0.07 8 (0.018) 370.0 (321.1) a 35.7 (28.7) ab 1395.5 (1341.8) ab 169.6 (191.5) a 1.79 (0.67) a 5.7 (7.9 ) a Colclough Pond (CP) 2937'N, 8219'W 0.100 (0.020) 283.1 (64.9) a 61.7 (83.5) a 1979.4 (977.3) a 72.8 (58.8) a 1.84 (1.07) a 8.1 (20. 2 ) b Lake (NL) 2937'N, 8212'W 0.062 (0.124) 41.3 (75.4) b 16.8 (18.0) b 1028.9 (2308.3) b 30.8 (47.6) b 1.77 (0.99) a 14.9 (22.5 ) a Prairie (PR) 2935'N, 8221'W 0.053 (0.009) 40.3 (11.6) bc 60.8 (39.4) a 1103.4 (733.7) a b 76.4 (49.1) a 1.55 (0.61) a 7.3 (14.4) a b San Felasco State Preserve (SF) 2943'N, 8227'W 0.076 (0.014) 136.9 (104.0) ac 30.1 (8.6) ab 1225.7 (997.0) ab 64.7 (17.9) a 1.95 (0.36) a 7.6 (9.8) a Invaded 0.074 (0.023 ) b 174.3 ( 166.2) a 43.9 (60.9) a 1370.5 (1116.1) a 72.5 (53.0) a 1.91 (0.78) a 4.8 (6.1) b U ninvaded 0.103 (0.091 ) a 180 (226.3) a 40.1 (33.1) a 1343.8 (1474.1) a 95.2 (128.3) a 1.69 (0.73) a 15.3 (23.9) a

PAGE 35

35 Table 2 2. P ercent ground cover (Mean and ran ge) for each growth form in A. crenata presence and absence plots across the five sites in Alachua County, Florida, USA Growth form Absent Present Tree 10.7 (0.0 70.0) 4.2 (0.0 50.0) Vine 5.8 (0.0 54.0) 4.0 (0.0 30.5) Shrub 5.4 (0.0 60.4) 5.9 (0.0 36.0) Herb 5.9 (0.0 27.1) 5.6 (0.0 50.8) Table 2 3 Eigenvectors (factor loading; eigenvector is scaled to the square root of its eigenvalue) of Principal Component Analysis of 157 plots for A. crenata cover (Accov), Understory native speci es number (UnspNo), Understory native species cover (Uncov), Overstory native species cover (Ovcov), soil moisture (Moist), percent light (Light) and Shan n on Weiner index (Diversity) across five sites in Alachua County, Florida, USA PC1 PC2 PC3 A. cren ata cover 1.503 0. 612 0 .377 Un derstory native species richness 1.949 0. 154 0. 171 Un derstory native cover 0 .770 1.181 1.347 Ov erstory native cover 0. .560 1.231 1.077 Moist ure 0. 840 1.098 0. 697 Light 0.021 1.626 0. 887 Diversity index 1.828 0. 301 0. 697 Eigenvalue 2.333 1.536 1.039 Proportion of Variance 0.333 0.219 0.148 Cumulative proportion 0.333 0.552 0.700

PAGE 36

36 Table 2 4 Eigenvectors (factor loading) of Component Analysis of 157 plots for A. crenata cover (Accov), Understory native species number (UnspNo), Understory native species cover growth form s; Trees (Tree), Herbaceous plants (Herb), Shrubs (Shrub), Vines (Vine), Overstory native species cover (Ovcov), soil moisture (Moist), percent light (Light) across five sites in Alachua County, Florida, USA PC1 PC2 PC3 A. crenata c ov er 1.538 0.1 17 0.288 Un derstory native species richness 1.660 0. 178 0.2 32 Ov erstory native cover 0. 352 1.277 0. 735 Moist ure 0. 595 0. 300 1.550 Light 0. 180 1.10 1.181 Shrub cover 0. 104 1.264 1.028 Tree cover 0. 800 0.975 0. 424 Vine cover 0.755 0.502 0.376 Herb cover 1.129 0.948 0. 661 Eigenvalue 1.952 1.604 1.510 Proportion of variance 0.217 0.178 0.169 Cumulative proportion 0.217 0.395 0.564

PAGE 37

37 Figure 2 1 Plot location at each o f the five site s showing cardinal orientations and length of transects in Alachua County, Florida, USA Each dot indicates the location of a plot and its size (and shade of color) indicates percent cover of Ardisia crenata The first five plots were rand omly selected along each transect within the invaded area, but a given plot within the invaded area may not contain any A. crenata individuals due to heterogeneous distribution of individuals.

PAGE 38

38 Figure 2 2 S pecies accumulation curve s of forest understor y native species for plots in the area invaded and uninvaded by A. crenata across five study sites in Alachua County, Florida, USA Solid line indicates the mean species richness and s hade d bands indicate 95% confidence interval s

PAGE 39

39 Figure 2 3 Forest unde rstory native species richness in relation to A. crenata cover at each of the five site s ( indicated by different colors) in Alachua County, Florida, USA Lines are fitted mixed model predictions and light colored shaded area is 95% confiden ce interval for each site (See T able 2 1 for site codes). Points are native understory species richness for each individual plot. Native species richness was significantly and negatively related to A. crenata cover (X 2 = 23.5, P < 0.001).

PAGE 40

40 Figure 2 4 Forest understory native species cover in A. crenata invaded a nd uninvaded plots at each of the five site s in Alachua County, Florida, USA ( see T able 2 1 for abbreviation definitions of sites ) A. crenata presence had a significant effect on native species cover ( X 2 = 18.2, p < 0.001). Red star s are mean s. The top and the bottom of each box correspond to the first and third quartiles (the 25th and 75th percentiles) The median is indicated by the thick horizontal line. Whiskers indicate the highest/lowest values that is with in 1.5 IQR of the box boarder where IQR is the inter quartile range, or distance between the first and third quartiles. Black dots are outliers.

PAGE 41

41 Figure 2 5 Principal component analysis correlation biplot of the first two principal components (perce nt of total variation) for vegetation and environmental characteristics measured during the spring sampling across the five study sites in Alachua County, Florida, USA Loadings for the explanatory variables: understory species number (UnspNo), understory species diversity (Diversity), understory native species cover (Uncov), o verstory natives species cover (Ovcov), s oil moisture (Moist), percent light (Light), and Ardisia crenata cover (Accov) are shown as vectors and the scale is on the left and bottom ax es. Each point represents a n invaded or uninvaded plot.

PAGE 42

42 Figure 2 6 Principal component analysis correlation biplot of the first two principal components (percent of total variation) for vegetation and environmental characteristics during the spring s ampling across the five study sites in Alachua County, Florida, USA Loadings for the explanatory variables: understory species number (UnspNo), native tree species cover (Tree), native herb, forb and graminoid species cover (Herb) native vine cover (Vine) o verstory natives species cover (Ovcov), s oil moisture (Moist), percent light (Light), and Ardisia crenata cover (Accov) are shown as vectors and the scale is on the left and bottom axes. Each point represents an invaded or uninvaded plot.

PAGE 43

43 CHAPTER 3 I N FLU E N CE OF SHADE TOLERANT INVASIVE SHRUB, ARDISIA CRENATA ON OAK SEEDLING REGENER ATION IN MESIC FORES T IN FLORIDA Low l ight availability under closed forest canopies (9% of incident radiation in southern hardwood forests; Canham et al. 1990) constrains growth and survival of many plants including seedlings of overstory species. In fact, native tree species capable of persisting in u nderstories show some level of shade tolerance i.e., the ability to withstand low light levels during some part of their life cycle (Valladares & Niin emets 2008) Degrees of shade tolerance varies among species and this variation is linked to traits such as high tissue density that enhances leaf lifespan and stem survival (Alvarez Clare & Kitajima 2007) and carbohydrate storage that enhances survival and recovery from episodes of negative carbon balance (Myers & Kitajima 2007) These traits allow shade tolerant seedlings to persist under limited light availability near their light compensation points (Givnish 1988; Baltzer & Thomas 2007) where even small change in light availability can significantly influence carbon balance of seedling performance. Hence, within species variation in seedling growth and survival of shade tolerant species can be linked to temporal and spatial variations in light within the forest understory in tropical (Montgom ery & Chazdon 2002) and temperate forest (Canham 1989) The shade stress in understories of closed canopy forests is long believe d as a barrier to in vasion by exotic invasive plants, many of which are disturbance dependent (Davis et al. 2000) Yet, increasing number of studies report that undisturbed closed canopy understories are invaded by shade toleran t exotic species (Martin et al. 2008) The negative impact of these shade tolerant invaders on the understory communities occurs in terms of reduction of species richness, cover, and biomass of native trees

PAGE 44

44 (Bottollier Curtet et al. 2012) and of native shrubs (Collier et al. 2002) ( Chapter 2). However, most of these studies did not test the mechanisms or processes by which the invasive species exclude native species or their broader impacts to ecosystem properties Competition for limiting resources has been suggested, but rarely tes ted (Levine et al. 2003) These species that invade undisturbed forests are predominantly shrubs (28% of global species, followed by t re e s 23%, herbs 36%, vines 17%, and grasses 11%) (Martin et al. 2008) Further, these species exhibit greater shade tolerance where they are non native compared to where they are native, most li kely due to enhanced carbon balance associated with the from host specialized herbivores and pathogens (DeWalt, Denslow, & Ickes 2004) Thus, the lack of natural enemies enabl e s invasive species to allocate resources to growth or further enhance capacity to capture limiting resource giving it an advantage over native species. In addition, the competitive the soil by the invaders, to which indigenous species are not adapted and thus experience reduced growth or seed germination (Callaway & Ridenour 2004; Cipol lini, McClain, & Cipollini 2008) In this paper, we investigate d native species displacement mechanisms and processes associated with invasion by shade tolerant shrubs. As an example of such invader, we chose A. crenata which invade s the shade understory under the closed canopy of hardwood hammock forests in north central Florida (C hapter 2 ) It is speculated that the leaf display patterns of A. crenata locally cast s a deep shade enhancing its competitive ability for light over native species (Kitajima et al. 2006) A

PAGE 45

45 reduction of light availability in closed canopy forests understories can have an important implication in regeneration of overstory species sensitive to changes i n light availability (Poorter 2007) A. crenata Sims. (Myrsinaceae) is a shade tolerant evergreen shrub that was introduced and promoted by the horticulture industry as an ornamental (Wirth et al. 2004) It is classified as a Category 1 Pest Plant by the Florida Exotic Pest Plant Council (FLEPPC 2011) A. crenata is capable of forming dense monodominant patches (Burks & Langeland 1998) with cover reaching >90% of ground (Chapter 2) and 300 stems per m 2 (Kitajima et al. 2006) Dispersal is limited, but birds are the main dispersers (Meisenburg 2007) Genetic analyses suggest that A. crenata ori ginated from southeastern China multiple times and then spread from there (Niu et al. 2012) In order to test whether light competition is one of the mechanism by which A. crenata suppress natives, a manipulative field experiment was conducted. The overall objective of this study was to explore the influence that A. crenata has on seedling regeneration of oaks that currently dominate the overstory of hardwood mesic forests in north central Florida. T wo oak species evaluated were Quercus hemispha e rica and Q. virginiana These are common species found in the overstory of these forests. More specific objectives were the following : 1. To d etermine the effects of A. crenata invasion on survival and growth of seedlings of Quercus hemisphaerica and Q. virginiana two common canopy oak species. 2. To assess the effects of aboveground competition on growth and survival of oak seedlings in the dense stand of A. crenata by comparing the intact plots vs. plots in whic h A. crenata stems are pulled down 3. To e stablish the influence of micro environmental varia tions and their potential interaction with A. crenata invasion on oak seedling survival

PAGE 46

46 We hypothesize d that A. crenata would have a negative effect on survival an d growth of both oak species. Further, we hypothesized that the negative effect would be at least partly attribut able to aboveground competition for light availability and predicted that pulling down A. crenata stems would enhance light and seedling perfo rmance. We also hypothesize d that Q. virginiana the more light demanding of the two species (Spector & Putz 2006) would exhibit greater negative effects of A. crenata presence. Seedling growth and survival were expected to be improved by light availability and soil moisture. Material and Methods Site and S tudy D esign communities are characterized by dominance of evergreen broadleaf trees, often mixed with some evergreen conifers and deciduous trees, under which layers of subcanopy and understory vegetation are present. They are further classified by soil water available, such as xeric, mesic, and hydric hammocks (Vince et al. 1989) The hardwood hammocks of north central Florida occur in a transitional zone between the mixed hardwood forests in the south eastern US and the tropical evergreen forests of south Florida (Platt & Schwartz 1990) Examples of common dominant s pecies include broadleaf evergreen species such as Quercus spp. and Magnolia grandiflora evergreen conifers such as Pinus glabra and P. taeda and deciduous hardwood species, such as Liquidambar styraciflua and Carya glabra We selected four sites from m esic hardwood hammocks near Gainesville, Florida, where we could locate dense patches (> 80% of understory cover) of actively invading populations of Ardisia crenata (See Table 3 1 for site names and locations). These sites

PAGE 47

47 were at least 1.5 km from each o ther, with a maximum distance of 15 km. All sites were closed canopy forest fragments that exhibit typical species composition for the hardwood hammocks in the area, without any sign of major disturbance. In communication with the landowners and land manag ers of these sites, we ensured that there would be no active removal efforts before and through the end of 2011 when this study was completed. At each site we established 36 plots each measuring 1.5 m by 1.5 m Of these, 24 plots were established within densely invaded patches ( A. crenata 12 additional control plots in adjacent areas where A crenata was yet to invade ( A crenata A. crenata individual s Within the invaded area the 24 invaded plots were paired by proximity, and down reduce above ground competition, in which A cre n a ta stems taller than 25 cm were tied to plastic coated wires and pulled down toward outside of the plot. The intended effect down light availability while maintaining competition in the rooting zone of the oak seedlings to be transplanted (see oak seedling s section below). Although, t ll influence d belowground competition, we expected that it was minimal ly influenced by the treatment because the pulled down stems remained alive and sometimes resprouted from the base A crenata plants in the second of eac h plot pair were left down The rational for pairing plots was the importance of stratified randomization to ensure that microenvironmental factors were comparable between the pull down and no pull down treatments.

PAGE 48

48 Environment C ondi tions After seedlings were planted, light availability was measured as photosynthetically active radiation (PAR) in all plot s at 35 cm (average height of seedlings) above the soil surface with a line quantum sensor (Li COR, Lincoln, Nebraska) In the p lot s with A. crenata PAR was also measured at 10 cm above A. crenata of the tallest indiv id ual within the plot to evaluate effect on light availability Simultaneously we continuously monitored reference light in a completely open area nearest t o each site to express the measured PAR as %PAR transmission relative to the light in a nearby open area Measurements were taken under clear sky conditions from 11 am to 3pm once after seedlings were planted After seedlings were planted, volumetric s oil moisture was estimated once in each plot with a soil moisture probe ( Theta probe type ML1 Delta T devices, Cambridge England). Four measurements were taken from different positions within each plot to account for soil heterogeneity. At each site 6 soil c ore s (diameter 5 cm and depth 10 cm, volume 196 cm 3 ) were sampled one chosen randomly from the 3 invaded plots and another from the 3 non invaded plots. Each core sample was measured for bulk density, gravimetric water content and nutrients. After homogen izing, a subsample of 5 g from each soil core was used to quantify availability of phosphorus (P), potassium (K), calcium (Ca) following extraction with 20 mL of the Mehlich 1 extraction solution (0.0125M H 2 SO 4 and 0.05M HCl) and nitrate and ammonium n itr ogen ( NO 3 + and NH 4 + ) following extraction with 1M KCl The extracts were filter ed through Whatman 42 filter paper and sent to the University of Florida IFAS Analytical Research Laboratory for determination of P, K and Ca concentration with ICP (Inductive ly Coupled Plasma

PAGE 49

49 Spectromet ry EPA Method 200.7) and NO 3 N and NH 4 N with an Alpkem Auto Analyzer (EPA Method 353.2). Oak S eedlings : P lanting and M easurements of G rowth and S urvival In April 2009, b are root seedlings of similar size 350 each of Q. hemis ph a erica (10 month old) and Q. virginiana (12 month old) were purchased from a local nursery, Central Florida Lands and Timber N ursery, L.L.C. Seedlings were planted in the field at the end of April 2009 at all sites with 3 seedlings of each species within each plot separated by 50 cm. A dibble bar ( 7 cm wide, 20 cm long, and 1.9 cm thick) was used to create a 24 cm deep hole in the ground to plant seedlings S oil and surrounding vegetation disturbance was kept at a minimum. Seedlings were tagged with flag ging tape to prevent confusion with other oak seedlings that were present in each plot prior to planting. A month after planting height was measured for all seedlings to the nearest mm. Two additional height measurements were taken, one in December 2009 ( 240 days after planting) and another in December 2010 (600 days after planting) when survival was also recorded Randomly selected 2 0 seedlings per species were destructively harvested at the time of transplanting, and separated to root s stem s and lea ves Leaves were scanned with a flatbed scanner, and leaf area was calculated using Scion Image ( Scion Corporation, Frederick, Maryland, USA) to the nearest mm 2 Dry mass was measured after drying at 60 o C for 72 hours. We harvested all surviving seedlings afte r the December 2010 census, following the same method as the initial harvest. After this final harvest, we also harvested and determined all aboveground biomass of adults (> 20 cm height) and seedlings (< 20 cm) of A. crenata within the plots.

PAGE 50

50 Statistical A nalyses All statistical analyses were conducted with R (R Development Core Team 2012) and used a significance level of P = 0.05. We analyzed environmental variables including soil nutrients, PAR, and A crenata biomass using a one way ANOVA to test for differences among sites, after transforming to achieve normality with natural log transformatio n (K) or Box Cox power (P, Ca ) if necessary Tukey post hoc test was used to identify differences between sites For NH 4 + N, which could not be normalized after any transformation w e used Kruskal Wallis test followed by Nemenyi Damico Wolfe Dunn post hoc test (Hollander & Wolfe 1999) Seedling survival was tested with logistic regression, in which the respo nse variable was the fate of each seedling (alive vs. dead). In the first analysis, survival recorded at each census (240 and 600 days after planting) was compared Pull down and No pull down plots were within the same A. crenata patch and Absent plo ts were outside of this patch. Due to the lack of independence two statistical analyses were conducted down ull For the first analysis a generalized linear mixed mod el with binomial distribution was used to test difference s in seedling survival, in which A. crenata presence and oak species identity were the main treatment factors, seedling initial height, soil moisture, and light availability were covariates, and site and plot nested within site were considered as random variables. The second analysis had the down down instead of A crenata presence vs. absence Change in biomass was de termined by comparing initial harvest of seedlings with harvest of surviving individuals at the end of the experiment. Due to low survival and

PAGE 51

51 insufficient sample size, biomass of live seedlings could not be compared statistically. Thus, to analyze the imp acts of A crenata on biomass accumulation by oak seedlings, we assumed the final biomass of dead seedling to be zero, then used a g eneralized linear mixed model with a Tweedie distribution which allows analysis of zero inflated data. The results for seed ling biomass per live seedling at the end of the experimen t are reported in the Appendix C ( Figure C 5). Results Site C haracteristics The four site s differed in soil characteristics (Table 3 1) ; Potassium ( F 3,20 = 16.9, P < 0.001 ), Calcium ( F 3,20 = 15.2, P < 0.001), NO 3 N ( F 3, 20 = 39.0, P < 0.001), moisture ( F 3, 144 = 54.8, P < 0.001) and bulk density ( F 3, 31 = 4.4, P = 0.01) The Lake site had the lowest fertility and the most fertile site was the Cemete ry. However, soil fertility had no appare nt relationship with A. crenata biomass ite had the lowest A. crenata biomass per area and Hogtown had about 2.5 times more b iomass per area. Overall invaded plot s had higher phosphorus ( X 2 = 13.2, P < 0.01) and calcium ( X 2 = 13.7, P < 0 .01). Seedling S urvival Oak seedling survival was significantly lower in the presence of A. crenata at the first census (240 days, P = 0.01) and did not differ significantly between Q. hemisphaerica and Q virginiana ( Tables 3 2 & 3 3 ; Figure 3 1 ). S oil mo isture measured at the time of seedling planting significantly influenced seedling survival, which was lower at lower soil moisture in both censuses ( P < 0.001) Light availability at 35 cm and seedling initial height were not significant factor s in the mo del (P = 0.43 and P = 0.17, respectively).

PAGE 52

52 Similarly, oak seeding survival at the second census (600 days, P < 0.001 ) also did not differ significantly between Q. hemisphaerica and Q virginiana (Table 3 2 & 3 3 ; Figure 3 1 ). S oil moisture measured at the time of seedling planting significantly influenced seedling survival, which was lower at lower soil moisture in both censuses ( P = 0.003 respectively). Light availability at 35 cm and seedling initial height were not significant factor s in the model (P = 0.11 and P = 0.16 respectively). Within invaded plots, pull down treatment significantly increased oak seedling survival co m pared to no pull down treatment at the first census ( P = 0.02) and Q. hemisphaerica had lower survival compared to Q. virginian a (Table 3 4 & 3 5 ; Figure 3 1 ). Greater initial l ight availability at 35 cm ( mean PAR = 13.1%) enhanced seedling survival in the first census ( P = 0.001 ; Appendix B Figure B 6 ) I nitial soil moisture and A. cre n a ta biomass per plot were not significant. In the second census as well, the pull down treatment increased seedling survival ( P = 0.02) and Q. hemisphaerica had lower survival compared to Q. virginiana (Table 3 4 & 3 5 ; Fi gure 3 1 ). I nitial l ight availability at 35 cm did not enhance seedling survival assessed at the second census. Initial soil mo i s ture became significant at the second census ( P < 0.001) whe re low er soil moisture was associated with lower seedling survival (Table 3 5) Greater A. crenata adult biomass in the plot also had a negative influence in survival Seedling B iomass Total mass per seedling decreased significantly from the initial values to 600 days after planting in the analysis treating dead s eedling mass as zero (Figure 3 2 ). Overall, the presence of A. crenata had a negative effect on biomass (P < 0.001 Table 3 6). Q.

PAGE 53

53 hemisphaerica had a lower biomass than Q. virginiana and this difference was maintained regardless of the presence of A. crenata (Figu re 3 2 ). Within invaded plots, the pull down treatment A. crenata had a significant positive effect on biomass of both oaks ( Table 3 7, P < 0.001). However, treatment; Q. virginiana had a positive response ll where as Q. hemisphaerica did not (P = 0.004). Surviving Q. virginiana seedlings were larger than Q. hemisphaerica ( Appendix B Figure B 5 ) Also a small number of surviving seedling s in A. crenata presence plot with no pull down were larger an d their final biomass was similar to the initial biomass. Surviving seedlings in pull down and A. crenata absent plots were smaller on average Discussion The results of this study suggest that the invasive shrub Ardisia crenata is likely to reduce the r ecruitment of canopy tree species in the understory of mesic hardwood forests of north central Florida The presence of A. crenata reduced the survival and growth of seedlings of two oak species : Q. virginiana and Q. hemisphaerica The responses to the pu ll down treatment were consistent with the hypothesis that A. cren ata imposes significant a boveground competition to other understory plants In the long run the invasion of A. crenata might alter the structure and species composition of the mesic hardwoo d hammocks. Light in the understories of closed canopy forests is a limiting resource, and even small reductions in the availability of light can have direct negative effects on growth and survival of understory species (Montgomery & Chazdon 2002) Despite this limitation, A. crenata is capable of invading, reproducing and further reducing the understory light availability. This study is the first to demonstrate experimentally that

PAGE 54

54 res ource competition for light imposed by an invasive understory shrub negatively affects performance of native tree seedlings although similar effects were observed for invasive canopy tree, Acer platanoides which alters both quantity and quality of light under their canopies, influencing survival of natives in riparian communities of western Montana (Reinhart et al. 2006) However, native species that dif fer in light demands are likely to respond differently to shading by invasive shrubs. Native shrubs in pine dominated forest s of northern Minnesota have differential effects on tree seedling survival; survival of light demanding species are most strongly a ffected by shrubs that reduce light availability by 30% than by other growth forms (Montgomery, Reich, & Palik 2010) Many tree species in closed canopy hardwood hammock forests depend on disturbances to reach canopy stature, but light requirements differ among species (Platt & Schwartz 1990) Of the two oak species we examined, Q virginiana is considered to be more light demanding (Spector & Putz 2006) but its survival under A. crenata was better than Q. hemisphaerica This difference was the opposite of our initial expectation. But, it could be attributed to the greater initial size of Q. virginiana compared to Q. hemisphaerica Larger size could be associated with larger carbohydrate stor age which might enhance the tolerance to shade and environmental stresses (Myers & Kitajima 2007) Also, Q. virgin iana seedling heights were taller on average (343.5 m m) at initial planting and could possibl y have greater access to light than Q. hemispha e rica (242.9 m m) Alternati vely, Q. hemisphaerica could be more sensitive to belowground conditions in the area invaded by A. crenata Species interactions are the net effect of both above and belowground competition (Gorchov & Trisel 2003) A. crenata r oots were left intact

PAGE 55

55 in the pull down treatment and hence they impose d competition for belowground resources and modify soil conditions although root competition may be less compared to the A. crenata present plot without this treatment There were only detectable differences in phosphorus and calcium for soil chemical characteristics between inside and outside of A. crenata invaded areas. Access to soil resources can be enhanced by belowgrou nd mutualisms and A. crenata establishes effective symbiotic relationships with native mycorrhizal fungi, enhancing its competitive advantage over native species (Bray, Kitajima, & Sylvia 2003) The ecological impacts of forest invasion by understory shrubs may be less obvious than invasion by canopy dominant species yet reduction of tree seedling regeneration can have a long term impact on forest community structures. A. crenata reduces the diversity of native plants including seedlings of canopy trees in the forest understory (Chapter 2). The results of this study demonstrate that one of the potential mechanism s with which A. crenata reduces recruitment of canopy tree seedlings is aboveground competition. Over time, suppression of seedling recruitment can significantly change the forest community structure as overs tory trees die and become replaced Land managers may want to take into consideration the lack of recruitment of oaks and other native species when managing the natural forests invaded by A. crenata In addition to reduction of A. crenata density by manual removal or herbicide, enrichment planting of seedlings of native species may increase the probability of success for biodiversity conservation.

PAGE 56

56 Table 3 1 L ocation s their soil characteristics and A rdisia crenata biomass (m eans standard deviation ) for the four study sites in Alachua County, Florida, USA D ifferent superscript letter s indicate significant difference between mean values by Tukey post hoc test. Site Latitude and longitude Bulk density (g/cm 3 ) Soil moisture (m 3 /m 3 ) 1 P (mg/kg) 2 K (mg/kg) 3 Ca (mg/kg) 2 NH 4 + (mg/kg) 4 NO 3 (mg/kg) A. crenata Mass per plot (g) 5 Biven s Arm (BA) 1.07 a (0.10) 0.026 b ( 0.010 ) 20.8 a (11.2) 78.7 a (30.9) 897.5 a (254.8) 12.5 a (3.9) 6.10 a (1.22) 1420 d (530) Evergreen Cemetery (E C) 2937'44.18"N, 8219'05.75"W 1.02 ab (0.14) 0.059 a ( 0.024 ) 9.0 b (1.2) 69.5 a (16.6) 936.7 a (354.8) 14.6 a (3.6) 4.62 b (0.88) 2735 b (437) Hogtown Creek (HC) 2941'53.15"N, 8220'36.23"W 1.08 a (0.09) 0.029 b (0 .011 ) 17.3 a (5.72) 52.4 ab (18.1) 752.7 a (86.0) 9.2 a (0.5) 2.58 c (0.65) 3629 a (604) Lake (NL) 2937'54.62"N, 8212'14.47"W 0.88 b (0.20) 0.057 a ( 0.014 ) 7.4 b (0.51) 34.6 b (6.2) 362.2 b (69.3) 10.9 a (1.5) 1.33 c (0.28) 1700 c (369) Ardisia present 6 0.99 a (0.15 ) 0.039 a (0.018 ) 17.3 a (2.0 ) 61.0 a (18.1 ) 850.9 a (330.7 ) 11.4 a (3.9 ) 3.61 a (2.32 ) Ardisia absent 6 1.03 a (0.16 ) 0.052 a (0.025 ) 9.9 b (10.3 ) 56.7 a (31.7 ) 623.7 b (205.6 ) 12.1 a (2.6 ) 3.70 a (1.80 ) 1 Volumetric s oil m oisture m easured in all plots with a Theta probe 2 Box Cox transformation. 3 Log transformation. 4 Non parametric test. 5 Aboveground biomass of A. crenata was determined for all invaded plots (2.25 m 2 ), including both pull down and no pull down treatments. 6 Di fference between plots with Ardisia present (n = 12) or absent (n = 12); Soil moisture Ardisia present (n = 96) or absent (n = 52)

PAGE 57

57 Table 3 2 Logistic mixed model results for seedling survival 240 days after transplanting of the two oak species ( Q v i rgin iana and Q. hemisph a erica ) and two treatments ( Ardisia crenata presence and absence, excluding pull down treatment) across the four sites in Alachua County, Florida, USA Seedling height, light measured 35 cm above the ground, and volumetric soil moistu re per plot at transplanting time are used as covariates. df X 2 value P value A. crenata presence 1 6.1 P = 0. 01 Initial h eight 1 1.8 P= 0.18 Initial l ight 1 0 .6 P= 0.44 Species 1 3.8 P= 0. 05 Initial soil m oisture 1 7.1 P < 0.00 1 Table 3 3 Logistic mixe d model results for seedling survival 600 days after transplanting of the two oak species ( Q v i rginiana and Q. hemisph a erica ) and two treatments ( Ardisia crenata presence and absence, excluding pull down treatment) across the four sites in Alachua Count y, Florida, USA Seedling height, light measured 35 cm above the ground, and volumetric soil moisture per plot at transplanting time are used as covariates. df X 2 value P value A. crenata presence 1 14.3 P < 0. 001 Initial height 1 1.9 P= 0.16 Initial l igh t 1 2.6 P= 0.11 Species 1 0.001 P= 0. 97 Initial soil m oisture 1 8.5 P = 0.00 3

PAGE 58

58 Table 3 4 Logistic mixed model results for seedling survival 240 days after transplanting of the two oak species ( Q v i rginiana and Q. hemisph a erica ) and two treatments ( Ardisi a crenata canopy pull down and no pull down ) across the four sites in Alachua County, Florida, USA Seedling height, light measured 35 cm above the ground, and volumetric soil moisture per plot at transplanting time are used as covariates. As well as, A. crenata seedling and adult aboveground biomass at the 600 days. df X 2 value P value Pull down 1 5.9 P= 0.02 Initial h eight 1 0.09 P= 0.77 Initial l ight 1 10.5 P= 0.001 Species 1 4.5 P= 0.0 3 Ardisia a dult mass 1 0.2 P= 0.66 Ardisia seedling mass 1 0 .4 P= 0. 52 Initial soil m oisture 1 0.3 P = 0. 57 Table 3 5 Logistic mixed model results for seedling survival 600 days after transplanting of the two oak species ( Q v i rginiana and Q. hemisph a erica ) and two treatments ( Ardisia crenata canopy pull down a nd no pull down ) across the four sites in Alachua County, Florida, USA Seedling height, light measured 35 cm above the ground, and volumetric soil moisture per plot at transplanting time are used as covariates. As well as, A. crenata seedling and adult aboveground biomass at the 600 days. df X 2 value P value Pull down 1 8.4 P =0.004 Initial h eight 1 0.008 P= 0.93 Light 1 1.2 P= 0.2 7 Species 1 4.8 P= 0.0 3 Ardisia a dult mass 1 5.5 P= 0.02 Ardisia seedling mass 1 0.8 P= 0. 37 Initial soil m oisture 1 12. 6 P < 0.00 1

PAGE 59

59 Table 3 6 Linear mixed model with a Tweedie distribution results of two oak seedling biomass ( Quercus v i rginiana and Q. hemisph a erica ) at 600 day s after transplanting comparing three treatments ( Ardisia crenata absent, A. crenata no pull do wn and initial harvest) across the four sites in Alachua County, Florida, USA df X 2 value P value A. crenata presence 2 33.5 P<0.001 Species 1 54.4 P< 0.001 A. crenata presence*Species 2 0. 6 P= 0. 72 Table 3 7 Linear mixed model with a Tweedie dist ribution results of two oak seedling biomass ( Quercus v i rginiana and Q. hemisph a erica ) at 600 day s after transplanting comparing three treatments ( Ardisia crenata no pull down A. crenata pull down and initial harvest) across the four sites in Alachua County, Florida, USA df X 2 value P value Pull down 2 14.1 P<0.001 Species 1 65.2 P< 0.001 Pull down *Species 2 10.9 P= 0. 004

PAGE 60

60 Figure 3 1 Box plot of the probability of survival for each individual seedling in a plot based on generalized linear mixed model of Oaks ( Quercus v i rginiana and Q. hemisph a erica ) in presence and absence of Ardisia crenata and A. crenata stems pull down and no pull down at each site for 240 and 600 days census across the four sites in Alachua County, Florida, USA The top and the bottom of each box correspond to the first and third quartiles (the 25th and 75th percentiles) T he median is indicated by the thick horizontal line. Whiskers indicate the highest/lowest values that is within 1.5 IQR of the box boarder where IQR is the inter quartile range, or distance between the first and third quartiles. B lack dots are outliers.

PAGE 61

61 Figure 3 2 Box plot of biomass per seedling for the two oak species ( Quercus hemisphaerica and Q. virginiana ) includes total biomass f rom the ini tial harvest, plots without Ardisia crenata (Absent), plots with A. crenata canopies pulled down (Pull down ), and plots with A. crenata canopy intact (No Pull down ) across the four sites in Alachua County, Florida, USA The analysis includes biomass of dea d individuals as zeros The top and the bottom of each box correspond to the first and third quartiles (the 25th and 75th percentiles) The median is indicated by the thick horizontal line. Whiskers indicate the highest/lowest values that is within 1.5 I QR of the box boarder where IQR is the inter quartile range, or distance between the first and third quartiles. Black dots are outliers.

PAGE 62

62 CHAPTER 4 DOES HERBICIDE TRANSLOCATION CORRELATE WITH SEASONAL CARBOHYDRATE BALANCE IN AN EVERGREEN SHRUB ARDISIA CR ENATA ? I nvasions by exotic plant species are a growing concern around the globe Exotic invasive plants may incur negative ecological impacts by alter ing disturbance regimens, nutrient cycling, and productivity of ecosystems, and displac ing native species (Vil et al. 2011) Furthermore, they incur economic costs of approximately $120 billion per year in the U.S. alone (Pimentel, Zuniga, & Morrison 2005) Many invasive exotic organisms have been recognized as serious threats to natural ecosystems in Florida (Gordon & Thomas 1997) Of the nearly 1,400 naturalized plant species in Florida about 11% have become (FLEPPC 2011) In an effort to counter this trend, Florida spent approximately $230 million from 1980 to 2006 managing aquatic, wetland and upland exotic invasive species (Schmitz 2007) The methods implemented for control of invasive plants in natural areas include biological control agents (insects and pathogens ), mechanical removal (use of machinery to cut, shear, shred, and crush plants and manual removal ), fire, and herbicides which may be combined to increase the efficacy T o achieve a cost effective measure to control exotic species, it is necessary to con sider species specific attributes including growth forms, life histories, and physiological characteristics. Perennial plants in seasonal environments change patterns of vegetative growth, photosynthate translocation and reproduction during the year. If th e desired control method, such as herbicides requires the plant to be actively growing, one should tune time of

PAGE 63

63 control with the time when the plant is most actively growing. However, the most effective time of the year may differ among control methods. Th erefore, to minimize the costs of control and maximize efficacy, a critical factor is to determine the best time of the year to conduct treatments in relation to treatment methods The current study focuses on Ardisia crenata (Myrsinaceae), which is a good example of a shade tolerant shrub that persist s in the understories of natural closed canopy forest s Such shade tolerant invasive plants, many of which are shrubs, are becoming a growing concern in many forest ecosystems (Martin et al. 2008) A crenata was introduced and promoted by the horticulture industry as an ornamental for more than 100 years (Wirth et al. 2004) but it has been classified as a Category 1 Pest Plant (i.e., those that have the most serious impacts on community structures and ecosystem functions) by the Florida Exot ic Pest Plant Council (FLE PPC 2011) A. crenata forms dense mono dominant patches in the understory, with high density of adult and seedling stems up to 600 per m 2 The impacts of A. crenata in natural areas such as hardwood hammocks include reductions of richness and cover of nat ive species (See chapter 2) and reduction of seedling recruitments of overstory canopy species. Both, in turn, have the potential to modify forest structure in the long term (See chapter 3). To reduce the impact of A. crenata on natural areas, typical met hods of control implemented by land managers include manual removal or mechanical mowing in small populations, whereas mowing and spraying are often combined

PAGE 64

64 with applications of herbicides such as glyphosate ( N (phosdphonomethyl) glycine ) 2,4 D ( 2,4 D Di chlorophenoxyacetic acid ) or triclopyr in large populations (Langeland et al. 2011) Mechanical removal such as mowing, in combination with herbicide application have shown to be useful for control of perennial plants (MacDonald et al. 1994; Mislevy, Mullahey, & Martin 1999) Resprouting following m owi ng reduces storage reserves such as non structural carbohydrates in the roots and a subsequent herbicide application is expected to be more effective because plants would have reduc ed capacity to regrow shoots after the herbicide application (Kalmbacher, Eger, & Rowland Bamford 1993) However, if a species has large reserves of non structural carbohydrates in roots it may be capable of resprouting many times. A crenata has some of the largest concentrations of non structural carbohydrates in roots compared to other woody species, possibly enhancing its capacity to resprout (Kitajima et al., 2006) Of the above mentioned herbicides, triclopyr is preferred by land managers for its greater efficacy on woody plants Triclopyr (3,5,6 trichloro 2 pyridinyloxyacetic acid) is a selective systemic herbicide that mimics the effects of plant hormones (auxins, up to 1000 times natural levels) which disrupt s hormonal balance and alter s growth (Ganapathy 1997; Tu et al. 2001) Triclopyr comes in two formulations: triethylamine salt (TEA) and butoxyethyl e ster (TBEE). Both can be sprayed on leaves for the desired effect of translocat ion to roots and meristems to kill whole plants I n perennial species the movement of foliar herbicides has been positively correlated to movements of non structural carbohydrat es from source regions

PAGE 65

65 (leaves) to sink regions (active growth) of the plant via the phloem (Devine & Hall 1990) The direction of movement can be dictated by seasonal changes in flows of non structural carbohydrates within the plant (Engle & Bonham 1980) For example, during the spring season when production of photosynthates in the leaf is low, there is a net mov ement of non structural carbohydrate from storage sources such as roots to active growth meristems and leaves Hence, there is a lower movement of foliar applied herbicides to other parts in the plant. L and managers have reported the need to spray herbicid es several times to obtain the desired control of A. crenata ( Michael Meisenburg, personal communication). Therefore, a better understanding of the mechanisms of herbicide movement and dynamics of carbohydrates is needed to increase the efficacy of control methods such as timing of spraying, dosage, and use of surfactants and adjuvants In North Central Florida, A. crenata conduct s active vegetative and reproductive growth (new leaves, shoots, flowers and developing fruits) during the summer (June through e arly September) M ature fruits are born on the plant from December for up to a year until a new crop of fruits mature. T he roots accumulate non structural carbohydrates (primarily starch) during the non growing season of winter and spring months (December through April), coinciding with increased light availability in semi deciduous canopy ( K aoru Kitajima unpublished data). From this season flux w e predict that herbicide translocation coincide s with non structural carbohydrate translocation to roots in win ter (Figure 4 1)

PAGE 66

66 The overall objective of this study is to assess the efficacy of mowing and triclopyr herbicide when employed to the exotic invasive shrub A. crenata during different seasons and to determine the absorption and translocation of triclopyr More specific objectives are the following: 1. To examine the effects of mowing on root carbohydrate concentration and herbicide efficacy 2. To test if herbicide effects differ between summer and winter, as predicted by the seasonal carbohydrate dynamics. 3. To quantify the percentage of the applied herbicide that is translocated to the roots. We hypothesize d that control of A. crenata by mowing once and the application of triclopyr herbicide in winter or spring would have the greatest control. Furthermore, we hy pothesize d that this increase in efficacy would be associated with non structural carbohydrate translocation to roots. Materials and Methods Two experiments were conducted to evaluate herbicide efficacy over time, one in the field and the other in the gree nhouse. The field experiment examine d the seasonal trend of herbicide efficacy in relation to root carbohydrate dynamics; the greenhouse experiment quantif ied herbicide movement in the plant in relation to light level and root carbohydrate status Field Ex periment We selected three mesic hardwood forest sites near Gainesville, Florida (Table 4 1) infested with dense patches of A crenata (at least 80% ground cover size of patches > 1 ha ). These three sites were chosen within the vicinity of the field expe riment in C hapter 3 based on the logistic advantages such as

PAGE 67

67 existing municipal permits and implementing experiments activities. All sites were relatively undisturbed forests dominated by broadleaf evergreen and deciduous canopy trees, such as Quercus spp In communication with the landowners and land managers, we were ensured that active removal efforts occurred in these sites before and through the end of 2011 when this study was completed. At each site we located a large dense patch of A. crenata and e stablished two rectangular adjacent blocks, each measuring 11.75 m x 4.25 m for a total of six blocks In one of the blocks at each site A. crenata was treated by mowing individuals to a height of 5 cm June 2009 Any seedlings < 5 cm could be potentially untouched by the mowing blades, but most likely they were damaged and killed by mowing. Cut shoots (stems, leaves, flowers and fruits) of adults (> 20 cm stem height) were removed from the mowed area The intended effect of mowing was to reduce carbohydrat e reserves in the roots by obligating plants to resprout and grow over several months, after which herbicide may be more effective in preventing plants to recover from the roots Each block was subdivided into 25 plots, each measuring 0.75 m by 0.75 m, and separated by 0.5 m (Figure 4 2). At the beginning of the experiment p lots were randomly assigned to one of five treatments (four herbicide application times and one control) and of 5 replicates per treatment. We selected four different times of the year ( October 2009, January 201 0 April 2010, and July 2010) for the herbicide treatment in order to evaluate the effect of different seasons on herbicide efficacy These treatment times were 4, 7, 10 and 13 months after the mowing, respectively. The

PAGE 68

68 first four months (June October) corresponded to the season of shoot extension and leaf development, and individuals mowed in June had vigorously resprouted and regrew shoots mostly prior to the first herbicide application time. The control plots, with no herbicide a pplication were maintained over the entire period Plots were monitored by taking photos to estimate A. crenata cover every 15 days from October 2009 until July 2011, one year after the last herbicide application. Herbicide a pplication and e fficacy m easur ements Triclopyr ester at 2 % v/v (Remedy Ultra, 10.8 g a cid e quivalent L 1 TBEE) with 0.5% nonionic surfactant (DyneAmic) was used for the herbicide treatment Herbicide was applied using a 2 gallon hand pressurized handheld sprayer with a fan nozzle (Rou ndup Ortho Heavy Duty, The Fountainhead Group Inc. New York Mills, New York ) on a spray to wet basis (625 L ha 1 ). Off target application to adjacent plots was prevented using a Styrofoam lamina barrier (1.5 m height) covering 3 sides of the plot (Figure 4 3). Prior to each herbicide application, digital photographs of each plot were taken from a distance of 1.6 m above ground and repeated every 15 days until the end of the experiment. Digital images were used to estimate percent cover of A. crenata using a point intercept method with a grid superimposed on the image (every 7.5 cm to create 12 1 grid points per plot ). At each grid point intersect A. crenata presence was evaluated (present = 1 or absent = 0). Percent cover was number of intersections with A. crenata divided by total intersections (range from 0 to 100%). size corresponding to ca. > 20 cm in height) in the photos, and their cover was estimated separately. The initial cover was used as a baseline for subsequent

PAGE 69

69 photographs of the plot. Herbicide efficacy index was calculated as the difference of A. crenata initial percent cover minus percent cover at the later observation time divided by the initial percent cover (Figure 4 4) B iomass a llocation and r oot c arbohydrate s torage Five A. crenata individuals greater than 20 cm in height were haphazardously s elected from areas between treatment plots in each block at each herbicide application date These individuals were harvested inta ct and separated into leaf, stem, and root. Leaves were scanned and leaf area was determined to the nearest mm 2 from scanned images of each leaf using Scion Image ( Scion Corporation, Frederick, Maryland, USA) Plant biomass allocation was determined after drying at 60 o C for 72 hours and weighed to the nearest 0.01 gram. The concentration of t otal non structural carbohydrate (TNC, the total of simple sugars and starch) per unit root mass ( mg g 1 ) in the primary roots was determined from harvested individuals Dried roots were chopped, homogenized and subsampled prior to grinding with a Wiley mill. From a pproximately 15 mg of ground root s from each individual concentrations of soluble sugars and starch were quantified. Soluble sugars were extracted with 80% e thanol in a shaking water bath at 27 o C for 12 hours followed by two additional repeated extractions with ethanol in a shaking water bath at 27 o C for two hours each. The remaining sample was digested to glucose with a 1.1% hydrochloric acid solution at 1 00 o C for 45 minutes to collect the glucose containing solution This was followed by repeated rinsing of the residual solids with deionized water at room temperature.

PAGE 70

70 Concentration of simple sugars and starch was measured as glucose equivalent using pheno l sulphuric acid colorimetric assay modified from Dubois (1956) Greenhouse Experiments Two repeated herbicide application experiments were conducted to follow herbicide absorption and translocation with 14 C labeled triclopyr herbicide using two s ets of plants grown from two separate seed collections. Three hundred seeds of A crenata were collected from 3 spatially separated populations within Gainesville, Florida on February 2010 and on February 2011 seeds were combined and germinated in rectan gular plastic trays filled with a soil mixture (Fafard Superfine Germinating Mix) in a greenhouse located in Gainesville, Florida. T emperature in greenhouse was controlled so that it did not exceed 29.4 o C and light photoperiod was maintained to not exceed 8 hours of dark A total of 40 seedlings were randomly selected with similar amounts from each site and transplanted into 1 gallon pots filled with the same soil mixture. Plants were grown in the greenhouse for 6 months and then assigned to two light trea tments : shade (90% neutral shade cloth covering a frame approximately 150 cm x 300 cm x 150 cm ) and no shade (sun). Controlled release f ertilizer was applied at an equivalent rate of 112 kg N ha 1 ( Osmocote 1 4 N, 14 P, 14 K ). Plants were kept under these t wo treatments throughout the experiment The first experiment used seedlings germinated in February 2010, with herbicide applications done in April 2011 ( on 14 months old plants) T he second experiment used a cohort of seedlings grown from seeds collected on F ebruary 201 1, and herbicide application was done in October 2011, when seedlings were 8 months old). Although the original intention was to use the same cohort of seedlings to

PAGE 71

71 examine the effects of the seasonal timing of herbicide application, seedlin gs from the seeds collected in February 2010 became too big and root bo u nd by October 2011 Plants were sprayed with a 0.65% solution of herbicide triclopyr amine T B E E (Remedy Ultra, 3.5 g a cid e quivalent L 1 ) to simulate a standard field application I mme diately following this application the third most developed leaf from the apical meristem of the main stem receive d a 6 l drop containing 14 C labeled TE A at 0.25% by volume with a 0.1% nonionic surfactant (DyneAmic) (400,000 dpms) This was placed on the center of the leaf between the edge and the main v e in. Prior to herbicide application 5 individuals were randomly harvested to quantify the initial size, total leaf area and dry mass of leaf, stem, and root. Le af area was determined to the nearest mm 2 fr om scanned images of each leaf using Scion Image ( Scion Corporation, Frederick, Maryland, USA). Plant biomass allocation was determined after drying at 60 o C for 72 hours. The concentration of total nonstructural carbohydrate (TNC, starch and simple sugars) in root s was determined following the same method as in the herbicide field experiment described above After herbicide application, 5 plants from each treatment were randomly assigned to three planned harvests For the April 2011 application, harvests we re scheduled at 1, 4 and 7 days after herbicide application After these harvests it was noticed that translocation of herbicide was very limited. Therefore the October 2011 harvests were scheduled at 7, 14 and 21 days after herbicide

PAGE 72

72 application At th e time of harvest the target leaf treated with 14 C labeled triclopyr ester (TBEE) was washed with 5 ml of water and leaf rinse was collected to determine the amount of herbicide that did not penetrate the leaf (leaf water wash) The leaf then was placed in a vial with 10 ml acetone and agitated 1 min to remove any triclopyr that was trapped in the cuticle of the leaf and was not absorbed (leaf acetone wash) Harvested plants were pressed and oven dried at 60 o C. Dried plants that had received 14 C labeled TB EE were then placed on X ray film for auto radiograph development for 40 days. After this time, tissue was separated to treated leaf, other leaves, stem, meristem, and roots and weighed to the nearest 0.01 gram. A subsample of 0.2 g (or amount available) o f each tissue was oxid ized following the Schniger combustion technique to liberate 14 C as 14 CO 2 This was trapped in sci ntillation fluid and labeled quantification was performed using a Packard scintillation counter against known standard. Statistical A n alyses All statistical analyses were conducted with R (R Development Core Team 2012) and used a significance level of P = 0.05. Treatment effects on root TNC concentration in the field experiment was tested with a linear mixed model in which time of herbicide application and mowing were main treatment factors and site and plot nested within site w ere considered as random variables. Herbicide efficacy in the field experiment was tested with a linear model in which the response variable was the herbicide efficacy index, and ti me of herbicide application, mowing and month after herbicide application were the main treatment factors and site and plot nested within site were considered as

PAGE 73

73 random variables. Months after herbicide application were considered as repeated measures Analysis of variance (ANOVA) was used to test the response of root TNC conce ight treatment interactions A Tukey post hoc test was used to identify differences among levels. Herbicide translocation responses were analyzed separately for the April and October experiments due to the differences in individual size s and harvest dates The herbicide movement within the plant was determined by the amount of labeled 14 C triclopyr herbicide recovered from different organs (treated leaf, other leaves, meristem, stems and roots) a nd treated leaf washes (water and acetone), expressed as the percent of the total initial concentration. From these, the following were calculated: t otal recovery (treated leaf + leaf washes + leaves + meristems + stems + roots) leaf water wash leaf acet one wash absorbed (treated leaf + leaves + meristems + stems + roots) treated leaf absorption and translocated (leaves + meristems + stems + roots). The amount of herbicide translocated from the treated leaf to other parts of the plant was determined for leaves, meristems, stems, and roots. These variables were analyzed as response variables in ANOVA s that examined light treatment and days after treatment (DAT) as main effects after ensuring the assumption of normality and equal variance. A Tukey post ho c test was used to identify differences among levels.

PAGE 74

74 Results Field Experiment Effects of season and mowing on root sugar and starch concentrations In the field, mowing shoots of A. crenata significantly reduced root starch concentration compared to unmowe d plants (P < 0.001 Table 4 4 ), and this difference was significant at each har vest time (Figure 4 5, Table 4 4 ). Starch concentration significantly changed with season in both control and mowed plants (P < 0.001 Table 4 4 ); the lowest in October, follow ed by July, and the highest in January and April (Figure 4 5, Table 4 4 ). Simple sugar concentration was also affected by mowing A. crenata (P < 0.01 Table 4 5 ), but in the opposite direction from starch (Figure 4 5); simple sugar concentration significan tly increased by mowing and this difference was maintained over time (Table 4 4). Simple sugar concentration was higher in July and October (during growing season) than in January and April. Herbicide efficacy in the field Herbicide efficacy index for adu lt plants did not differ between 6 and 12 months after herbicide application (P = 0.63, Table 4 6 ). January application date showed the lowest herbicide efficacy and significantly lower efficacy in mowed plots than in unmowed plots ( P < 0.001, Table 4 6) This month showed few resprouts and plot cover of adult plants was explained by surviving plants that did not get killed with herbicide The other application dates (October, April, and July) did not differ from each other and between mowed and unmowed tre atments (Figure 4 6).

PAGE 75

75 At 6 and 12 months after herbicide application, many new seedlings appeared, all of which were apparently from newly germinated seeds and none from resprouts. Herbicide efficacy index was lower for seedlings in the 12 months after ap plication compared t o 6 months (P < 0.001, Table 4 7 ). Overall, January and April herbicide applications had lower efficacy than October and July applications (Figure 4 6). Mowing treatment increased efficacy compared to unmowed plots except for July appli cation date. Greenhouse Experiments There was a significant light treatment effect on starch concentration (P < 0.001 Table 4 8 ); shade treatment decreased starch concentrations in both April and October experiments (Figure 4 7 ). Root sugar concentrations on the other hand showed no difference between light treatments and between April and October experiments (Table 4 9) Amount of herbicide absorbed into the plants was estimated as the amount not accounted in the leaf wash by water and acetone. In the Ap ril herbicide application, there was no significant difference in the total recovery between light treatments (P = 0.09) and harvest times (P = 0.45; Table 4 10 ). Total recovery was 73.8 %. However, there was a significant decrease in leaf water wash over time (P = 0.01), from 60.1% at 1 day after treatment (DAT) to 47.4% at 7 DAT. Accordingly there was a significant increase in amount absorbed (P = 0.04) and translocated (P = 0.04) over time (Table 4 9 and Table 4 1 2 ). There was no significant treatment or time of harvest effect on percent recovery in leaf aceton e wash and treated leaf. Overall leaf acetone wash was 12.7% and treated leaf 3.5%.

PAGE 76

76 In the October herbicide application, there was a significant decrease in the total recovery over time (P = 0.04; Table 4 10) from 99.6% at 7 DAT to 84.5% at 21 DAT. There was a significant decrease in amount in leaf wash over time (Table 4 11 and Table 4 1 3 ), decreasing from 78.2% 7 DAT to 45.8% 21 DAT. There was no significant effect of light treatment or change ove r time for leaf washes (acetone), absorbed and translocated. Overall leaf acetone wash was 6.4%, absorbed 20%, and translocated 12.1%. There was a significant light treatment effect on the amount absorbed in treated leaf (Table 4 14) lower in sun (9.7%) t han in shade (14.6%). Radioactivity was detected in all portions of treated plants in both April and October herbicide application dates and for all treatments (Table 4 15 & 4 18 ) For the sun treatment, proportion of radioactivity in the roots ( 1.4 %) was significantly higher than in other organs (P <0.001) In the shade only the pair wise difference between vs. was significant ( Table 4 16 ). There were no significant differences in other plant tissues. Overall there was a significant inc rease in translocation in all plant parts over the 7 days period (P = 0.004; Tables 4 17 ); overall average from 0.31% to 0.68%. In the October application date there was only a significant increase of translocation over 21 days for all plant parts (P = 0. 002, Table 4 18) ; overall average from 2.1% to 5.3%. Discussion The results of this study confirm that triclopyr, a popular herbicide for control of woody plants, is effective for the control of A. crenata Although relatively small amount of the herbicide enters the plant, there was a

PAGE 77

77 considerable amount of translocation to the roots. Contrary to what we expected, there was no difference in efficacy of mowing adult individuals in the field, even though mowing and subsequent re growth reduced the root carbo hydrate storage prior to all herbicide application dates 4 13 months later Foliar application killed the treated plants regardless of mowing except for a weak effect in January when some plants were not killed likely due to cold weather However, mowing and shoot removal was effective for reducing the density of seedlings originated from seeds that germinated after herbicide applications. The recovery of A. crenata cover from germinated seeds in unmowed plots were the strongest following the January herbi cide application. Influence of H erbicide T iming on E fficacy Seasonal variation of triclopyr herbicide efficacy have been linked to environmental stress such as drought (Seiler et al. 1993) Other stress such as temperature could also lead to reduced efficacy. In our study, the January application date was during the months with below average mean temperatures (Appendix C Fi gure C 1). A. crenata is susceptible to freezing events, which can lead to stem die back (K. Kitajima unpublished data). Even though efficacy index w as reduced in January application date, they were higher than triclopyr applied to other species such as b l ackberry ( Rubus spp.) 63% control at 12 MAT (Ferrell et al. 2009) and Chinese pri vet ( Ligustrum sinense ) 70.3% applied in December at 24 MAT (Harrington & Miller 2005) Influence of M owing on H erbicide E fficacy Mowing increased herbicide efficacy on seedlings however all seedlings present in the plots were from germinated seeds rather than resp routs. All

PAGE 78

78 mowed adult shoots, including fruits, were removed. Therefore, unmowed plots had more seeds ready to be dropped and germinate in the following year. A new cohort of fruits mature on plants in December January, and fruits may remain on plants for up to one year when the next cohort of fruits mature. High seedling cover after 12 months following the herbicide application in January can be attributed to maximum local fruit density in January; when the plants in the unmowed plots were killed by herbi cide in January, they had more fruits than other times of the year (Meisenburg 2007) Seeds that are dropped to the ground germinate when temperature and moisture conditions are appropriat e, mostly from April through October (Alison Fox, personal communication). In the mowed plots, adult size plants that recovered shoots did not flower until the following year (June of the following year ) and produced no ripe fruit prior to the last herbici de application date was conducted (July) Hence, a smaller number of seedlings found in unmowed plots were likely to have originated from seeds that had been dropped prior to the mowing, or dispersed from adults in the surrounding area. Herbicide T ransloc ation A l arge proportion of applied triclopyr herbicide (> 60%) did not enter the leaf of A. crenata and slow absorption into plants continued over 7 14 days. Hence, a rain event shortly after herbicide application may wash down and compromise its efficac y. In other species, such as the honey mesquite tree ( Prosopis juliflora ), leaves absorb 66% of applied triclopyr ester herbicide within 24 hours. These differences in absorption by leaves are related to leaf developmental stage and relative amount of waxy cuticle (Hess 1987) Leaves

PAGE 79

79 that have incomplete development or reduced waxy cuticle on a leaf surface tend to show gre ater absorption of water soluble herbicides Leaf wash with acetone recovered between 6 and 12% of applied herbicide, which was double of what entered the leaf indicating that the cuticle represents a considerable barrier to the herbicide entry into leaves of A. crenata Nevertheless, the small amount of herbicide that does enter the leaf is likely to be translocated to other plant parts, most importantly to the roots. Triclopyr is highly mobile in plants, in particular under warm conditions (Radosevich & Bayer 1979) The greenhouse environment al conditions were probably optimal for herbicide translocation, while the colder condition of the field in January may have compromised the herbicide efficacy by constraining translocation. In summary, triclopyr is an effective herbicide to control A. cre nata despite the small amount of the herbicide that enters the plant. A method that will increase herbicide penetration could yield better results and could lower rates of herbicide application needed. Mowing was effective for controlling seedlings by rem oving seed sources and possibly multiple mowing treatments could further reduce seed source. Weather can play an important role in efficacy of adult plants and therefore it is not recommended to apply herbicide during cold periods. Regardless of method or timing it is recommended that multiple herbicide treatments be conducted to obtain the desired control to kill both adults and new seedlings

PAGE 80

80 Table 4 1 Study s ite locations Alachua County, Florida, USA Site L atitude and longitude Evergreen Cemetery (E C) 2937'44.18"N, 8219'05.75"W Hogtown Creek (HC) 2941'53.15"N, 8220'36.23"W Lake (NL) 2937'54.62"N, 8212'14.47"W Table 4 2 Field experiment biomass and leaf area (means) of harvested Ardisia crenata individuals in the mowed and unmowed fields at subsequent dates when herbicide applications were administered across the three sites in Alachua County, Florida, USA Application date Treatment Leaf Area (cm 2 ) Leaf (g) Stem (g) Root (g) Flower & Fruit (g) October 2009 Cut 981.2 5.6 4.2 22.1 0.0 Not Cut 1836.1 11.7 18.2 37.5 9.3 January 2010 Cut 949.8 5.8 4.6 26.3 0.0 Not Cut 1267.1 8.3 14.5 28.0 6.2 April 2010 Cut 857.4 5.7 5.4 26.3 0.2 Not Cut 1187.9 8.7 18.2 41.9 5.2 July 2010 Cut 887.4 9.7 9.1 38.2 2.1 Not Cut 1108.3 11.6 26.2 46.1 7.7

PAGE 81

81 Table 4 3 Greenhouse experiment biomass and leaf area (means) of harvested Ardisia crenata individuals grown under sun and shade light treatments in the greenhouse for the April and October 2011 herbicide application experiments across the thr ee sites in Alachua County, Florida, USA Application date Light Leaf Area (cm 2 ) Leaf (g) Stem (g) Root (g) Flower & Fruit (g) April Sun 432.3 5.6 1.8 23.6 0.0 Shade 493.8 3.8 1.1 10.6 0.0 October Sun 139.0 1.8 0.5 2.8 0.0 Shade 194.9 1.5 0.3 1.4 0.0 Table 4 4 Linear mixed model results for root starch concentration of Ardisia crenata as a function of mowing and herbicide application date across the three sites in Alachua County, Florida, USA df X 2 value P value Mowing 1 26.6 P<0.001 Applicat ion date 3 135.9 P< 0.001 Mowing*Application date 3 2.8 P= 0. 42 Table 4 5 Linear mixed model results for root simple sugar concentration of Ardisia crenata as a function of mowing and herbicide application date across the three sites in Alachua County Florida, USA df X 2 value P value Mowing 1 13.3 P<0.001 Application date 3 105.0 P< 0.001 Mowing*Application date 3 0.6 P= 0. 90

PAGE 82

82 Table 4 6 Linear mixed model results for herbicide efficacy index for adult plants after 6 and 12 months following th e four herbicide application dates in the mowed and unmowed fields across the three sites in Alachua County, Florida, USA df X 2 value P value Application date 3 115.4 P < 0.001 Mowed 1 16.5 P < 0.001 Month After treatment (MAT) 1 0.2 P = 0.63 Applic ation date Mowed 3 20.7 P < 0.0 01 Table 4 7 Linear mixed model results for the herbicide efficacy index for seedlings after 6 and 12 months after the four herbicide application dates in the mowed and unmowed fields across the three sites in Alachua C ounty, Florida, USA df X 2 value P value Application date 3 65.4 P < 0.001 Mowed 1 99.5 P < 0.001 Month After treatment (MAT) 1 64.5 P < 0.001 Application date Mowed 3 10.3 P = 0.0 2 Table 4 8 A nalysis of variance results for root starch concent ration of Ardisia crenata plants grown under low and high light treatments in the greenhouse for the April and October 2011 experiments in Alachua County, Florida, USA df F value P value Light 1 94.4 P<0.001 Experiments 1 2.1 P = 0.05 Light*Experiments 1 3.4 P= 0. 09

PAGE 83

83 Table 4 9 A nalysis of variance results for root simple sugar concentration of Ardisia crenata plants grown under low and high light treatments in the greenhouse for the April and October 2011 experiments in Alachua County, Florida, USA df F value P value Light 1 4.3 P=0.06 Experiments 1 4.8 P = 0.05 Light*Experiments 1 0.01 P= 0. 91

PAGE 84

84 Table 4 10 A nalysis of variance results for radioactivity of 14 C triclopyr in Ardisia crenata plants grown under low and high light treatments at 1, 4 and 7 days after herbicide treatment (DAT ) for the April 2011 greenhouse experiment in Alachua County, Florida, USA Response variable Factors df F value P value Total Light 1 3.1 P=0.09 DAT 2 0.8 P = 0.45 Light*DAT 2 1.6 P= 0. 21 Leaf water wa sh Light 1 0.1 P=0.82 DAT 2 5.8 P = 0.01 Light*DAT 2 0.8 P= 0. 44 Leaf acetone wash Light 1 3.9 P=0.06 DAT 2 3.1 P = 0.06 Light*DAT 2 1.5 P= 0. 25 Absorbed Light 1 0.3 P=0.59 DAT 2 3.5 P = 0.04 Light*DAT 2 1.5 P= 0. 24 Treated l eaf Light 1 0.04 P=0.84 DAT 2 2.3 P = 0.13 Light*DAT 2 1.3 P= 0. 29 Translocated Light 1 1.0 P=0.34 DAT 2 3.7 P = 0.04 Light*DAT 2 1.1 P= 0. 36

PAGE 85

85 Table 4 1 1 A nalysis of variance results for radioactivity of 14 C triclopyr in Ardisia crenata pla nts grown under low and high light treatments at 7 1 4, and 21 days after herbicide treatment (DAT) for the October 2011 greenhouse experiment in Alachua County, Florida, USA Response variable Factors df F value P value Total Light 1 0.1 P=0.79 DAT 2 3 .9 P = 0.04 Light*DAT 2 0.4 P= 0. 68 Leaf water wash Light 1 0.01 P=0.93 DAT 2 3.4 P = 0.049 Light*DAT 2 0.02 P= 0. 98 Leaf acetone wash Light 1 1.6 P=0.21 DAT 2 0.8 P = 0.45 Light*DAT 2 0.7 P= 0. 49 Absorbed by the plant Light 1 0 .2 P=0.68 DAT 1 2.8 P = 0.08 Light*DAT 1 0.65 P= 0. 53 Absorbed by the treated leaf Light 1 5.7 P=0.03 DAT 1 0.3 P = 0.77 Light*DAT 1 0.9 P= 0. 41 Translocated Light 1 0.7 P=0.40 DAT 2 2.9 P = 0.07 Light*DAT 2 0.4 P= 0. 64

PAGE 86

86 Table 4 1 2 Radioactivity ( mean and standard errors of the percent of 14 C labeled triclopyr applied) for leaf water wash, total absorbe d, the treated leaf, and translocation which showed significant effects on days after treatments (DAT) with herbicide in the April 2011 experiment (Table 4 9 ) in Alachua County, Florida, USA Different superscript letters indicate significant difference within a column by post hoc Tukey multiple comparisons Days after herbicide application Leaf water wash Absorbed Translocated 1 d ay 60.1 (1.9) a 3.4 (0.4) a 1.2 (0.4) a 4 days 56.7 (2.3) a b 6.1 (2.5) a b 2.0 (0.4) a b 7 days 47.4 (3.5) b 7.0 (2.8) b 2.7 (1.0) b Table 4 1 3 Radioactivity ( mean and standard errors of the percent of 14 C labeled triclopyr applied) for the total recov ery and leaf wash, which showed significant effects on days after treatment (DAT) with herbicide in the October 2011 experiment (Table 4 10) in Alachua County, Florida, USA Different superscript letters indicate significant difference within a column by p ost hoc Tukey multiple comparisons Days after herbicide application Total recovery Leaf water wash 7 day 99.6 (4.8) a 78.2 (10.4) 14 days 88.8 (2.5) a b 69.1 (2.4) 21 days 84.5 (3.7) b 45.8 (10.3) Table 4 1 4 Radioactivity ( mean and standard error of the percent of 14 C labeled triclopyr applied ) found in the treated leaf following in the October 2011 experiment in Alachua County, Florida, USA Overall means for the two light treatments which significantly differed (Table 4 10), across the three days a fter herbicide treatments (DAT) Light treatment Treated leaf Sun 9.7 (3.3) a Shade 14.6 (5.0) b

PAGE 87

87 Table 4 1 5 The results of analysis of variance for radioactivity of 14 C triclopyr translocated to different plant organs (leaves, stems, roots, meristems) of Ardisia crenata plants grown under low and high light treatments at 1, 4, and 7 days after herbicide treatment (DAT) in the April 2011 experiment in Alachua County, Florida, USA Factors df F value P value Light 1 1.6 P = 0.22 DAT 2 5.9 P = 0.004 Pl ant part 3 21.8 P < 0. 001 Light DAT 2 1.7 P = 0. 18 Light Plant part 3 3.6 P =0.02 DAT Plant part 6 1.7 P = 0.13 Light DAT Pant part 6 0.7 P = 0. 67 Table 4 1 6 Radioactivity ( mean and standard errors of the percent of 14 C labeled triclopyr applied) found in leaves, meristems, stems and roots in the April 2011 experiment ( Table 4 14) in Alachua County, Florida, USA Overall means across the three days after herbicide treatments (DAT) for the two light treatments which significantly differed (Table 4 10). Different superscript letters indicate significant difference by post hoc Tukey multiple comparisons within each light treatment. Light treatment Leaves Meristems Stems Roots Sun 0.35 (0.03) a b 0.14 (0.05) a 1.2 (0.06) a b 1.4 (0.3) c Shade 0.46 (0.11) ab 0.14 (0.05) a 0.36 (0.1) ab 0.79 (0.13) b

PAGE 88

88 Table 4 1 7 Radioactivity ( mean and standard errors of the percent of 14 C labeled triclopyr applied) found across the four plant parts, which differed significantly among days after treatment (DA T) in the April 2011 greenhouse experiment (Table 4 14) in Alachua County, Florida, USA Different superscript letters indicate significant difference by post hoc Tukey multiple comparisons Days after herbicide application Overall recovered 1 day 0.31 ( 0.11) a 4 days 0.50 (0.07) a b 7 days 0.68 (0.12) b Table 4 1 8 The results of analysis of variance for radioactivity of 14 C triclopyr translocated in Ardisia crenata plants grown under lo w and high light treatments at 7, 14, and 21 days after herbicide treatment (DAT) in the October 2011 greenhouse experiment in Alachua County, Florida, USA Factors df F value P value Light 1 1.7 P = 0.20 DAT 2 6.6 P = 0.002 Plant part 3 2.2 P = 0. 09 Light DAT 2 1.0 P = 0. 37 Light Plant part 3 0.15 P =0.93 DAT Plant part 6 0.7 P = 0.62 Light DAT Pant part 6 0.6 P = 0. 74

PAGE 89

89 Table 4 1 9 Radioactivity ( mean and standard errors of the percent of 14 C labeled triclopyr applied) found across the four plant parts, which differed significantly among days after tr eatment (DAT) in the October 2011 greenhouse experiment (Table 4 17) in Alachua County, Florida, USA Different superscript letters indicate significant difference by post hoc Tukey multiple comparisons Days after herbicide application All plant parts 7 day 2.1 (0.9) a 14 days 1.5 (0.2) a 21 days 5.4 (1.1) b

PAGE 90

90 Figure 4 1. Schematic of proposed mechanism of carbohydrate movement in a forest understory evergreen plant in relation to seasonal light availability. In the Summ er Fall (June August) period light levels in the understory are reduced (depicted by the size of the open arrow) and movement (black arrows) of stored carbohydrates from the roots to aboveground plant parts. In the Winter Spring (December April) fa ll period light increases with leaf and the plant exploit greater light availability and excess carbohydrate are translocated to the roots. SUMMER WINTER Carbohydrate Light Light Shoot & leaf production +shoot + Root storage +

PAGE 91

91 Figure 4 2. Herbicide field experiment setup: A) Blocks consisted of 11.75 by 4.25 m area, subdivided into 25 pl ots. B) Each plot was 0.75 by 0.75 m with 0.5 m of separation between plots; Five treatments (four herbicide application times and a control) were randomly applied to each plot with a total of 5 replicates per treatment. Experiment was conducted from April 2009 to July 2011 in Alachua County, Florida, USA A ) Block B) Plot Plot

PAGE 92

92 Figure 4 3. Examples of field experiment plots with herbicide barrier in Alachua County, Florida, USA A) Unmowed plot with herbicide barrier. B) Mowed plot with herbicide barrier and C) close up view of mowed plot with barrier.

PAGE 93

93 Figure 4 4. Examples of responses to October herbicide application (field experiment) measured as Ardisia crenata c over in plots at Hogtown Creek (A F, mowed plot). Plot prior to herbi cide application (A and D), 6 months after treatment (B and E), and 12 months after treatment (C and F) Alachua County, Florida, USA

PAGE 94

94 Figure 4 5. Boxplots of seasonal total non structural carbohydrates (TNC, starch and simple sugars) at each herbicide a pplication date in the field for mowed and unmowed adult A rdisia cre n ata plants across the three sites in Alachua County, Florida, USA Stars are means. The top and the bottom of each box correspond to the first and third quartiles (the 25th and 75th perce ntiles) The median is indicated by the thick horizontal line. Whiskers indicate the highest/lowest values that is within 1.5 IQR of the box boarder where IQR is the inter quartile range, or distance between the first and third quartiles. Black dots are outliers.

PAGE 95

95 Figure 4 6. Boxplots of herbicide efficacy after 6 and 12 months after herbicide treatment application date in the field for mowed and unmowed adult A. cre n ata plants. Herbicide efficacy measured as relative amount of A. cre n ata seedling or adults removed from each plot and summarized across the three sites in Alachua County, Florida, USA Stars are means. The top and the bottom of each box correspond to the first and third quartiles (the 25th and 75th percentiles) The median is indicated by the thick horizontal line. Whiskers indicate the highest/lowest values that is within 1.5 IQR of the box boarder where IQR is the inter quartile range, or distance between the first and third quartiles. Black dots are outliers.

PAGE 96

96 Figure 4 7. Boxplots o f seasonal total non structural carbohydrates (TNC, starch and simple sugars) at each herbicide application for shaded and sun A. cre n ata plants in the greenhouse experiment in Alachua County, Florida, USA Star are means and colored box height includes ra nge between first and third quartiles and thick horizontal line is the median. The top and the bottom of each box correspond to the first and third quartiles (the 25th and 75th percentiles) The median is indicated by the thick horizontal line. Whiskers in dicate the highest/lowest values that is within 1.5 IQR of the box boarder where IQR is the inter quartile range, or distance between the first and third quartiles. Black dots are outliers.

PAGE 97

97 CHAPTER 5 CONCLUSIONS The main objective of this research is to explore the phenomenon of forest exotic invasive plants in a more comprehensive and integral fashion through assessment of their impact on cover and richness of understory species, description of mechanisms by which exotic plants competitively suppress native species, and evaluation of efficacy of herbicides as a common method of control. The results suggest that invasive shrub A. c renata in closed canopy hardwood hammock forests of Florida resulted in the reduc tion of understory species richness by 25%, while the total understory cover of native species was lowered by 34% with significant difference found in all growth forms (trees, shrub, vines and herbs) compared to areas uninvaded by A. crenata Shading by A. crenata is an important mechanism by whic h it can suppress seedli n g s overstory species Such effect can potentially have a significant effect on the regeneration of trees. The s urvival and growth of seedlings of t wo oak species Quercus virginiana and Q. hemisphaerica in the understory decreased in the presence of A. crenata after two growth seasons. The reduction in seedling recruitments of overstory canopy species due to A. crenata invasions can potentially impact forest structure in a long term. Hence, for rapid recovery of native species dive rsity, removal of A. crenata may be complemented with enrichment planting of seedlings of native species. H erbicide s such as t riclopyr, are a good method for of control of A. crenata but efficacy may be compromised by weather conditions such as cold tem perature and rain events. Recovery of A. crenata population from seed

PAGE 98

98 germination is a significant concern in herbicide treated area; seed germination post treatment had a significant contribution to regeneration of A. crenata at 6 12 months after the sing le herbicide treatment Thus, a retreatment of sites would be essential to obtain desired control in highly infested sites. Further research will be needed to evaluate how seedlings of other overstory tree species and rare native understory species in hard wood hammocks are impacted by A. crenata Development of methods to improve efficacy of herbicides will be particularly useful for control ling A. crenata by enhancing herbicide entry through leaves. Finally, the management decisions should consider the ev aluat ion of the economic impacts of A. crenata and public willingness to accept employment of particular methods for control ling A. crenata and restore impacted ecosystems.

PAGE 99

99 APPENDIX A AD DIT IONAL TABLES AND FIG URES FOR CHAPTER 2 Table A 1. P ercent cover ( mean) of native and exotic species for forest understory in the presence and absence of Ardisia crenata and overstory of all plots (n= 157 ) Alachua County, Florida, USA Species name Origin Absent Present O verstory Petiveria alliacea L. Native 3.671 3.560 0.000 Smilax sp. Native 3.119 1.489 0.185 Quercus hemisphaerica W. Bartram ex Willd. Native 1.934 0.258 3.408 Toxicodendron radicans (L.) Kuntze Native 1.36 4 1.398 0.127 Sabal palmetto (Walter) Lodd. Ex Schult. & Schult. f. Native 1.247 0.833 2.261 C arex willdenowii Schkuhr ex Willd. Native 1.214 0.433 0.000 Prunus caroliniana (Mill.) Aiton Native 1.138 0.033 0.000 Quercus pumila Walter Native 1.137 0.000 0.000 Parthenocissus quinquefolia (L.) Planch. Native 1.073 0.775 0.064 Carpinus caroliniana Walter Native 1.060 0.207 17.325 Hedera helix L. Exotic 0.973 0.036 0.000 Chasmanthium laxum (L.) Yates Native 0.868 0.095 0.000 Cornus foemina Mill. Native 0.863 0.024 0.127 Vitis rotundifolia Michx. Native 0.832 1.021 3.357 Verbesina virginica L. Na tive 0.808 1.902 0.000 Oplismenus hirtellus (L.) P. Beauv. Native 0.748 0.232 0.000 Quercus nigra L. Native 0.678 0.274 14.057 Pinus glabra Walter Native 0.638 0.000 3.217 Celtis laevigata Willd. Native 0.595 0.088 14.344 Rumex hastatulus Baldwin Nati ve 0.458 0.173 0.000 Ostrya virginiana (Mill.) K. Koch Native 0.384 0.652 20.911 Dichanthelium spp. Native 0.367 0.119 0.000 Elephantopus elatus Bertol. Native 0.351 0.148 0.000 Bignonia capreolata L. Native 0.342 0.202 0.064 Carya glabra (Mill.) Swee t Native 0.300 0.231 10.401 Arisaema dracontium (L.) Schott Native 0.274 0.140 0.000 Lamium amplexicaule L. Exotic 0.271 0.871 0.000 Ruellia caroliniensis (J.F. Gmel.) Steud. Native 0.182 0.006 0.000 Salvia coccinea Buc'hoz ex Etl. Native 0.173 0.056 0 .000 Campsis radicans (L.) Seemann ex Bureau Native 0.164 0.143 0.115 Cinnamomum camphora (L.) J. Presl Exotic 0.151 0.000 3.121 Serenoa repens (W. Bartram) Small Native 0.149 0.083 0.000 Quercus minima (Sarg.) Small Native 0.137 0.000 0.000

PAGE 100

100 Table A 1. C ontinued. Species name Origin Absent Present O verstory Mitchella repens L. Native 0.134 0.095 0.000 Tradescantia fluminensis Vell. Exotic 0.123 0.571 0.000 Sanicula canadensis L. Native 0.122 0.110 0.000 Acer rubrum L. Native 0.114 0.019 0.000 Dio scorea floridana Bartlett Native 0.103 0.020 0.000 Gelsemium sempervirens (L.) W.T. Aiton Native 0.101 0.069 0.127 Galium pilosum Aiton Native 0.084 0.010 0.000 Euonymus americanus L. Native 0.080 0.010 0.000 Galactia volubilis (L.) Britton Native 0.07 5 0.000 0.000 Prunus serotina Ehrh. Native 0.074 0.500 0.159 Viola sororia Willd. Native 0.074 0.085 0.000 Matelea sp. Native 0.068 0.274 0.000 Callicarpa americana L. Native 0.047 0.369 0.465 Asplenium platyneuron (L.) Britton et al. Native 0.045 0.0 06 0.000 Asclepias sp. 0.041 0.005 0.000 Diospyros virginiana L. Native 0.041 0.024 0.000 Erythrina herbacea L. Native 0.041 0.000 0.000 Oxalis spp. 0.041 0.054 0.000 Rubus argutus Link Native 0.040 0.124 0.000 Entodon sp. {moss} Native 0.034 0.000 0.000 Saururus cernuus L. Native 0.032 0.000 0.000 Quercus michauxii Nutt. Native 0.029 0.000 0.975 Viola walteri House Native 0.029 0.005 0.000 Ampelopsis arborea (L.) Koehne Native 0.027 0.000 0.000 Liquidambar styraciflua L. Native 0.027 0.462 12. 567 Ulmus americana L. Native 0.027 0.000 2.611 Pinus palustris Mill. Native 0.026 0.005 0.478 Morus rubra L. Native 0.021 0.018 1.038 Ilex opaca Aiton Native 0.019 0.000 0.318 Sonchus asper (L.) Hill Exotic 0.018 0.030 0.000 Dryopteris ludoviciana ( Kunze) Small Native 0.016 0.000 0.000 Viburnum nudum L. Native 0.016 0.018 0.573 Distichum sp. {moss} Native 0.014 0.000 0.000 Gordonia lasianthus (L.) J.Ellis Native 0.014 0.071 0.000 Stellaria media L. Vill. Exotic 0.014 0.012 0.000 Acer saccharum M arshall Native 0.012 0.036 2.898

PAGE 101

101 Table A 1. Continued Species name Origin Absent Present O verstory Fern sp. 0.012 0.179 0.000 Arisaema triphyllum (L.) Schott Native 0.007 0.000 0.000 Ilex vomitoria Aiton Native 0.007 0.000 0.000 Mnium sp. {moss} Native 0.007 0.000 0.000 Polygonum spp. 0.007 0.071 0.000 Ambrosia artemisiifolia L. Native 0.005 0.000 0.000 Baccharis glomeruliflora Pers. Native 0.005 0.000 0.000 Botrychium biternatum (Savigny) Underw. Native 0.005 0.000 0.000 Digitaria sp. 0.0 05 0.000 0.000 Hypericum sp. Native 0.005 0.000 0.000 Magnolia grandiflora L. Native 0.005 0.000 10.255 Pinus elliottii Engelm. Native 0.005 0.000 0.159 Acer negundo L. Native 0.000 0.026 0.000 Bambusa sp. Exotic 0.000 0.238 1.051 Cercis canadensis L Native 0.000 0.012 0.000 Cornus florida L. Native 0.000 0.238 0.318 Eupatorium sp. 0.000 0.000 0.000 Juglans nigra L. Native 0.000 0.060 0.478 Krigia virginica (L.) Willd. 0.000 0.000 0.000 Lonicera sempervirens L. Native 0.000 0.018 0.000 Lyonia lucida (Lam.) K. Koch Native 0.000 0.143 0.032 Osmunda cinnamomea L. Native 0.000 0.071 0.000 Persea borbonia (L.) Spreng. 0.000 0.000 0.000 Persea palustris (Raf.) Sarg. Native 0.000 0.030 0.318 Physalis sp. 0.000 0.000 0.000 Quercus virginiana Mi ll. Native 0.000 0.048 14.567 Rhapidophyllum hystrix (Pursh) H. Wendl. & Drude ex Drude 0.000 0.000 0.000 Rhododendron spp. Native 0.000 0.143 0.764 Stachys floridana Shuttlew. ex Benth. Native 0.000 0.043 0.000 Fraxinus caroliniana Mill. Native 0.000 0.000 0.510 Nyssa sylvatica var. biflora (Walter) Sarg. Native 0.000 0.000 0.478 Pinus taeda L. Native 0.000 0.000 1.688 Tilia americana L. Native 0.000 0.000 0.127 Ulmus alata Michx. Native 0.000 0.000 2.197 Osmanthus americanus (L.) Benth. & Hook. f.ex A. Gray Native 0.000 0.00 0 0.987

PAGE 102

102 Table A 1. Continued Species name Origin Absent Present overstory Vaccinium sp. L.* Native Native 0.000 0.000 0.064 Ageratina aromatica (L.) Spach* Native 0.000 0.004 0.000 Carex digitalis Willd.* Native 0.2 22 0.139 0.000 Melothria pendula L.* Native 0.090 0.000 0.000 Poinsettia heterophylla (L.) Klotzsch & Garcke ex Klotzch* Native 0.017 0.011 0.000 Trichostema dichotomum L.* Native 0.023 0.000 0.000 total number of species 85 spp 41 spp number of species that occurred only in the fall (+5 spp) (+2 spp) number of Exotic species fall or spring (+8 spp) (+2 spp) Native species absent in the overstory

PAGE 103

103 Figure A 1 Experimental setup for each site showing shape and size of invaded zone calculated polygon created by distances from the origin (black dot) of the first five plots ( A. crena ta invaded) of each transect at the five study sites in Alachua County, Florida, USA. Invaded zones areas are: CP = 510.7 m 2 MC = 550.3 m 2 NL = 3,084.8 m 2 PR = 133.2 m 2 SF = 989.7 m 2

PAGE 104

104 APPENDIX B ADDITIONAL FIGURES FOR CHAPTER 3 Figure B 1. Monthly temperatures during study period taken from nearest 82 16' 0.012" W) Gainesville, Florida. Bl ack line is monthly mean based on daily means. Red line is mean monthly maximum based on daily maximum. Blue line is the mean monthly minimum temperature based on daily minimum temperature. Open circles are extreme temperatures experienced during the month Grey vertical lines indicate initial planting or census dates (240 and 600 days).

PAGE 105

105 Figure B 2. Monthly precipitation during study period taken from nearest meteorological 82 16' 0.012" W) Gainesville, Florida. Grey vertical lines indicate initial planting or census dates (240 and 600 days).

PAGE 106

106 Figure B 3 Mean monthly temperature s during 2 7 years (1984 to 201 1 ) at Gainesville, Red circles are mean monthly maximum temperatures, tr iangles are mean monthly temperatures, and blue squares are mean monthly minimum temperatures. Lines are linear regressions and shaded areas are 95% confidence intervals for each month Black filled points are temperatures during oak seedling experiment (A pril 2009 to December 201 1 ).

PAGE 107

107 Figure B 4 Monthly precipitation during 2 7 years (1984 to 201 1 ) at Gainesville, Florida and shaded areas are 95% confidence intervals for each month Blac k filled points are precipitation during oak seedling experiment (April 2009 to December 20 10 ).

PAGE 108

108 Figure B 5. Box plot of seedling ( Quercus hemisphaerica and Q. virginiana ) harvest total biomass for initial harvest, plots without Ardisia crenata (Absent), plots with A. crenata canopies pulled down (Pull down ), and plots with A. crenata canopy intact (No Pull down ). Star s are means. Data excludes values of dead individuals. The top and the bottom of each box correspond to the first and third quartiles (the 25th and 75th percentiles) The median is indicated by the thick horizontal line. Whiskers indicate the highest/lowest values that is within 1.5 IQR of the box boarder where IQR is the inter quartile range, or distance between the first and third quarti les. Black dots are outliers.

PAGE 109

109 Figure B 6. Light availability for plots without Ardisia crenata (Absent), plots with A. crenata canopies pulled down (Pull down), and plots with A. crenata canopy intact (No Pull down). Light measured as percent photosynth etically active radiation (PAR) relative to an open area at 35 cm (average height of seedlings) above the soil surface Stars are means. Measurements were taken under clear s ky conditions from 11 am to 3pm, across five sites in Alachua County Florida, USA

PAGE 110

110 APPENDIX C ADDITIONAL FIGURES FOR CHAPTER 4 Figure C 1 Mean monthly temperature s during 2 7 years (1984 to 201 1 ) at Gainesville, Red circles are mean monthly maximum temperatures, triangles are mean mon thly temperatures, and blue squares are mean monthly minimum temperatures. Lines are linear regressions and shaded areas are 95% confidence intervals for each month Black points are temperatures during herbicide field experiment period (April 2009 to Dece mber 201 1 ).

PAGE 111

111 Figure C 2 Monthly precipitation during 2 7 years (1984 to 201 1 ) at Gainesville, Florida and shaded areas are 95% confidence intervals for each month Black points are precip itation during herbicide field experiment (April 2009 to December 20 10 ).

PAGE 112

112 LIST OF REFERENCES Alvarez Clare, S. & Kitajima, K. (2007) Physical defence traits enhance seedling survival of neotropical tree spe cies. Functional Ecology 21 1044 1054. Baltzer, J.L. & Thomas, S.C. (2007) Determinants of whole plant light requirements in Bornean rain forest tree saplings. Journal of Ecology 95 1208 1221. ecosystem change? Biological Invasions 14 1 10. Bottollier Curtet, M., Charcosset, J. Y., Poly, F., Planty Tabacchi, A. M. & Tabacchi, E. (2012) Light interception principally drives the understory response to boxelder invasion in riparian forests. Biolo gical Invasions 14 1445 1458. Bradley, B.A. & Mustard, J.F. (2006) Characterizing the landscape dynamics of an invasive plant and risk of invasion using remote sensing. Ecological Applications 16 1132 1147. Bray, S.R., Kitajima, K. & Sylvia, D.M. (2003 ) Mycorrhizae differentially alter growth, physiology, and competitive ability of an invasive shrub. Ecological Applications 13 565 574. Burks, K.C.E. & Langeland, K.A.E. (1998) Identification & biology of non native plants in Florida's natural areas / K .A. Langeland and K. Craddock Burks, eds. University of Florida, Gainesville, FL. Callaway, R.M. & Ridenour, W.M. (2004) Novel weapons: invasive success and the evolution of increased competitive ability. Frontiers in Ecology and the Environment 2 436 44 3. Canham, C.D. (1989) Different Responses to Gaps Among Shade Tolerant Tree Species. Ecology 70 548 550. Canham, C.D., Denslow, J.S., Platt, W.J., Runkle, J.R., Spies, T.A. & White, P.S. (1990) Light regimes beneath closed canopies and tree fall gaps in temperate and tropical forests. Canadian Journal of Forest Research 20 620 631. Catford, J.A., Jansson, R. & Nilsson, C. (2009) Reducing redundancy in invasion ecology by integrating hypotheses into a single theoretical framework. Diversity and Distribu tions 15 22 40. Chazdon, R.L. & Field, C.B. (1987) Determinants of photosynthetic capacity in six rainforest Piper species. Oecologia 73 222 230. Cipollini, K.A., McClain, G.Y. & Cipollini, D. (2008) Separating above and belowground effects of Alliaria petiolata and Lonicera maackii on the performance of Impatiens capensis. The American midland naturalist 160 117 128. Collier, M.H., Vankat, J.L. & Hughes, M.R. (2002) Diminished plant richness and abundance below Lonicera maackii, an invasive shrub. Am erican Midland Naturalist 147 60 71. Davis, M.A., Grime, J.P. & Thompson, K. (2000) Fluctuating resources in plant communities: a general theory of invasibility. Journal of Ecology 88 528 534. Devine, M.D. & Hall, L.M. (1990) Implications of Sucrose Tr ansport Mechanisms for the Translocation of Herbicides. pp. 299 304. DeWalt, S., Denslow, J. & Ickes, K. (2004) Natural enemy release facilitates habitat expansion of the invasive tropical shrub Clidemia hirta. Ecology 85 471 483. Dozier, H. (1999) Plant introductions to invasion: History, public awareness, and the

PAGE 113

113 case of Ardisia crenata. ProQuest Dissertations and Theses 1 113 p. Dubois, M., Gilles, K.A., Hamilton, J.K., Rebers, P.A. & Smith, F. (1956) Colorimetric method for determination of sugars an d related substances. Analytical chemistry 28 350 356. Ehrenfeld, J.G. (1980) Understory response to canopy gaps of varying size in a mature oak forest. Bulletin of the Torrey Botanical Club 29 41. Elton, C.S. (1958) The ecology of invasions by animals and plants. Methuen & Co. Ltd, London. Engle, D.M. & Bonham, C.D. (1980) Nonstructural carbohydrates in roots of Gambel oak sprouts following herbicide treatment. Journal Range Manage 33 390 394. Epanchin Niell, R.S. & Hastings, A. (2010) Controlling est ablished invaders: integrating economics and spread dynamics to determine optimal management. Ecology Letters 13 528 541. Ferrell, J.A., Sellers, B.A., MacDonald, G.E. & Kline, W.N. (2009) Influence of Herbicide and Application Timing on Blackberry Contr ol. Weed Technology 23 531 534. FLEPPC. (2011) List of Invasive Plant Species. Wildland Weeds 14 11 14. Fridley, J.D., Stachowicz, J.J., Naeem, S., Sax, D.F., Seabloom, E.W., Smith, M.D., Stohlgren, T.J., Tilman, D. & Holle, Von, B. (2007) The invasion paradox: Reconciling pattern and process in species invasions. Ecology 88 3 17. Ganapathy, C. (1997) Environmental fate of triclopyr Environmental Monitoring & Pest Management Branch Department of Pesticide Regulation, Sacramento, CA. Garca Llorente, M., Martn Lpez, B., Gonzlez, J.A., Alcorlo, P. & Montes, C. (2008) Social perceptions of the impacts and benefits of invasive alien species: implications for management. Biological Conservation 141 2969 2983. Gilbert, B. & Lechowicz, M. (2005) Invasib ility and abiotic gradients: The positive correlation between native and exotic plant diversity. Ecology 86 1848 1855. Givnish, T. (1988) Adaptation to sun and shade: a whole plant perspective. Functional Plant Biology 15 63 92. Gorchov, D.L. & Trisel, D.E. (2003) Competitive effects of the invasive shrub, Lonicera maackii (Rupr.) Herder (Caprifoliaceae), on the growth and survival of native tree seedlings. Plant Ecology 166 13 24. Gorchov, D.L., Thompson, E., O'Neill, J., Whigham, D. & Noe, D.A. (201 1) Treefall gaps required for establishment, but not survival, of invasive Rubus phoenicolasius in deciduous forest, Maryland, USA. Plant Species Biology 26 221 234. Gordon, D.R. & Thomas, K.P. (1997) Florida's invasion by nonindigenous plants: history, screening, and regulations. Strangers in paradise. Impact and management of nonindigenous species in Florida (eds D. Simberloff, D.C. Schmitz & T.C. Brown Island Press, Covelo, California, USA. Harrington, T.B. & Miller, J.H. (2005) Effects of Application Rate, Timing, and Formulation of Glyphosate and Triclopyr on Control of Chinese Privet (Ligustrum sinense). Weed Technology 19 47 54. Herault, B., Bachelot, B., Poorter, L., Rossi, V., Bongers, F., Chave, J., Paine, C.E., Wagner, F. & Baraloto, C. (2011) Functional traits shape ontogenetic growth trajectories of rain forest tree species. Journal of Ecology 99 1431 1440. Hess, F.D. (1987) Relationship of plant morphology to herbicide application and absorption. Monograph series of the Weed Science Societ y of America 4 19 35.

PAGE 114

114 Hollander, M. & Wolfe, D.A. (1999) Nonparametric Statistical Methods, 2nd Edition Wiley Interscience. Kalmbacher, R.S., Eger, J.E., Jr & Rowland Bamford, A.J. (1993) Response of southern wax myrtle (Myrica cerifera) to herbicides in Florida. Weed Technology 7 84 91. Keane, R. & Crawley, M. (2002) Exotic plant invasions and the enemy release hypothesis. Trends in Ecology & Evolution 17 164 170. Kitajima, K. (1994) Relative importance of photosynthetic traits and allocation patte rns as correlates of seedling shade tolerance of 13 tropical trees. Oecologia 98 419 428. Kitajima, K., Fox, A.M., Sato, T. & Nagamatsu, D. (2006) Cultivar selection prior to introduction may increase invasiveness: evidence from Ardisia crenata. Biologic al Invasions 8 1471 1482. Langeland, K.A., Ferrell, J.A., Sellers, B.A., MacDonald, G.E. & Stocker, R.K. (2011) Integrated Management of Nonnative Plants in Natural Areas of Florida University of Florida IFAS Publication. Levine, J.M., Adler, P.B. & Yel enik, S.G. (2004) A meta analysis of biotic resistance to exotic plant invasions. Ecology Letters 7 975 989. Levine, J.M., Vil, M., D'Antonio, C.M., Dukes, J.S., Grigulis, K. & Lavorel, S. (2003) Mechanisms underlying the impacts of exotic plant invasio ns. Proceedings. Biological sciences / The Royal Society 270 775 781. Lusk, C.H., Perez Millaqueo, M.M., Piper, F.I. & Saldana, A. (2011) Ontogeny, understorey light interception and simulated carbon gain of juvenile rainforest evergreens differing in sh ade tolerance. Annals of Botany 108 419 428. MacDonald, G.E., Brecke, B.J., Colvin, D.L. & Shilling, D.G. (1994) Chemical and mechanical control of dogfennel (Eupatorium capillifolium). Weed Technology 8 483 487. Martin, P.H., Canham, C.D. & Marks, P.L (2008) Why forests appear resistant to exotic plant invasions: intentional introductions, stand dynamics, and the role of shade tolerance. Frontiers in Ecology and the Environment 7 142 149. Meisenburg, M.J. (2007) Reproductive and dispersal ecology of the invasive coral ardisia ( Ardisia crenata ) in northern Florida. Miller, K.E. & Gorchov, D.L. (2004) The invasive shrub, Lonicera maackii, reduces growth and fecundity of perennial forest herbs. Oecologia 139 359 375. Mislevy, P., Mullahey, J.J. & Mart in, F.G. (1999) Preherbicide mowing and herbicide rate on tropical soda apple (Solanum viarum) control. Weed Technology 13 172 175. Monk, C.D. (1965) Southern mixed hardwood forest of northcentral Florida. Ecological Monographs 35 336 354. Montgomery, R. & Chazdon, R. (2002) Light gradient partitioning by tropical tree seedlings in the absence of canopy gaps. Oecologia 131 165 174. Montgomery, R.A., Reich, P.B. & Palik, B.J. (2010) Untangling positive and negative biotic interactions: views from above and below ground in a forest ecosystem. Ecology 91 3641 3655. Myers, J.A. & Kitajima, K. (2007) Carbohydrate storage enhances seedling shade and stress tolerance in a neotropical forest. Journal of Ecology 95 383 395.

PAGE 115

115 Niu, H.Y., Hong, L., Wang, Z.F., Shen, H., Ye, W.H. & Mu, H.P. (2012) Inferring the invasion history of coral berry Ardisia crenata from China to the USA using molecular markers. Ecological Research 27 809 818. Oksanen, J., Blanchet, F.G., Kindt, R., Legendre, P., Minchin, P.R., O'Hara R.B., Simpson, G.L., Solymos, P., Stevens, M.H.H. & Wagner, H. (2012) vegan: Community Ecology Package Pimentel, D., Zuniga, R. & Morrison, D. (2005) Update on the environmental and economic costs associated with alien invasive species in the United Sta tes. Ecological Economics 52 273 288. Platt, W.J. & Schwartz, M.W. (1990) Temperate Hardwood Forests. Ecosystems of Florida University of Central Florida Press, Orlando. Poorter, L. (2007) Are species adapted to their regeneration niche, adult niche, or both? The American Naturalist 169 433 442. Poorter, L. & Kitajima, K. (2007) Carbohydrate storage and light requirements of tropical moist and dry forest tree species. Ecology 88 1000 1011. affner, U. & Vil, M. (2012) A global assessment of invasive plant impacts on resident species, communities and ecosystems: the interaction of impact measures, invading species' traits and environment. Global Change Biology 18 1725 1737. R Development Co re Team. (2012) R: A Language and Environment for Statistical Computing Vienna, Austria. Radosevich, S.R. & Bayer, D.E. (1979) Effect of temperature and photoperiod on triclopyr, picloram, and 2, 4, 5 T translocation. Weed Science 27 22 27. Reinhart, K. O., Gurnee, J., Tirado, R. & Callaway, R.M. (2006) Invasion through quantitative effects: intense shade drives native decline and invasive success. Ecological Applications 16 1821 1831. Richardson, D.M. & Rejmnek, M. (2011) Trees and shrubs as invasive alien species a global review. Diversity and Distributions 17 788 809. Ricklefs, R.E. (2010) Evolutionary diversification, coevolution between populations and their antagonists, and the filling of niche space. Proceedings of the National Academy of Scien ces of the United States of America 107 1265 1272. Sax, D. & Brown, J. (2000) The paradox of invasion. Global Ecology and Biogeography 9 363 371. present Research Pr ogram. Natural Areas Journal 27 251 253. Seiler, J.R., Cazell, B.H., Schneider, W.G., Zedaker, S.M. & Kreh, R.E. (1993) Effect of plant moisture stress on absorption and translocation of triclopyr in oak seedlings. Canadian Journal of Forest Research 23 2213 2215. Spector, T. & Putz, F.E. (2006) Crown retreat of open grown Southern live oaks (Quercus virginiana) due to canopy encroachment in Florida, USA. Forest Ecology and Management 228 168 176. Surrette, S.B. & Brewer, J.S. (2009) Inferring relatio nships between native plant diversity and Lonicera japonica in upland forests in north Mississippi, USA. Applied Vegetation Science 11 205 214.

PAGE 116

116 Tecco, P.A., Daz, S., Cabido, M. & Urcelay, C. (2010) Functional traits of alien plants across contrasting climatic and land use regimes: do aliens join the locals or try harder than them? Journal of Ecology 98 17 27. Tu, M., Hurd, C., Randall, J.M. & Robison, R. (2001) 7k. Triclopyr. Weed control methods handbook: tools & techniques for use in natural areas (eds M. Tu, C. Hu rd & J.M. Randall. USDA. PLANTS Database, http://plants.usda.gov/java/ Valladares, F. & Niinemets, (2008) Shade Tolerance, a Key Plant Feature of Complex Nature and Consequences. Annual Review of Ecology Evolution and Systematics 39 237 257. Veno, P.A (1976) Successional relationships of five Florida plant communities. Ecology 57 498 508. Vil, M. & Ibez, I. (2011) Plant invasions in the landscape. Landscape Ecology 26 461 472. J.L., Pergl, J., plants: a meta analysis of their effects on species, communities and ecosystems. Ecology Letters 14 702 708. Vince, S.W., Humphrey, S.R. & Simons, R.W. (1989 ) The ecology of hydric hammocks: a community profile. Vitousek, P.M., Mooney, H.A., Lubchenco, J. & Melillo, J.M. (1997) Human domination of Earth's ecosystems. Science 277 494 499. Wirth, F.F., Davis, K.J. & Wilson, S.B. (2004) Florida nursery sales an d economic impacts of 14 potentially invasive landscape plant species. Journal of Environmental Horticulture 22 12 16.

PAGE 117

117 BIOGRAPHICAL SKETCH Gerardo Celis was born in Costa Rica. Upon conclusion of high s ch ool, he initiated a program in e nvironmental s tudies at the University of British Columbia in Vancouver. After one year there, he returned to Costa Rica, where he completed his undergraduate studies in b germination of two sympat ric palm species: Chamaedorea tepejilote Liebm. and Chamaedorea costaricana was the result of a pro bono collaboration with the National Museum of Costa Rica. After concluding his undergraduate stu dies he taught b iostatistics at the same university and was selected by the Organization for Tropical Studies (OTS) to participate in the Research Experiences for Un dergraduates (REU) program at La Selva biological s tation. The research conducted was enti Do patterns of seed germination and seedling biomass allocation reflect a shade tolerance syndrome in Gnetum leybodii Tul. (Gnetaceae a teaching assistant under Plantains, iguanas and shamans: an introduction to field e At this point in his career, he felt the need to develop a broader understanding of environmental processes by incorporating the interdisciplinary dimension. Thus, he decided to pursue a m i nte rdisciplinary e cology with emphasis on t ropical c onservation and d evelopment at the University of land with nati ve tree species in Costa Rica: a n ecophysiological approa ch to species to enroll at UF to pursue a Ph.D. in i nterdisciplinary e cology with emphasis on f orest r esources and c onservation and concluded in the fall of 2012