<%BANNER%>

Establishment and Management of Native Wildflowers on Florida Roadsides and Former Pastures

Permanent Link: http://ufdc.ufl.edu/UFE0022660/00001

Material Information

Title: Establishment and Management of Native Wildflowers on Florida Roadsides and Former Pastures
Physical Description: 1 online resource (128 p.)
Language: english
Creator: Frances, Anne
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: bahiagrass, coreopsis, cutting, disturbance, establishment, gaillardia, herbicide, lanceolata, leavenworthii, management, microsite, mowing, native, notatum, paspalum, pasture, pulchella, right, roadside, seed, wildflower
Horticultural Science -- Dissertations, Academic -- UF
Genre: Horticultural Science thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Establishing native wildflowers into areas dominated by Paspalum notatum var. saurae (bahiagrass) is a common goal of roadside management and ecological restoration. Yet, there is limited information on establishment and management practices for local ecotype seeds. This research sought to determine the effects of competition and disturbance on native wildflower establishment in areas dominated by P. notatum. Effects of planting season and post-planting disturbance (cutting) were assessed on competitive interactions between P. notatum and two Florida native congeners, Coreopsis lanceolata (lanceleaf tickseed) and C. leavenworthii (Leavenworth's tickseed) in two-species competition experiments. Coreopsis survival was greater in fall- than spring-established plants. In fall-established plants, C. lanceolata had higher survivorship than C. leavenworthii, although C. leavenworthii biomass and flower number were greater than that of C. lanceolata. Paspalum notatum reduced C. lanceolata biomass 58% and C. leavenworthii biomass 41%; however, conspecific neighbors reduced biomass of both Coreopsis species by at least 81%. Cutting decreased above- and belowground Coreopsis biomass by 55% and 30%, respectively. Seed and microsite limitations to establishment were assessed by seeding C. lanceolata at 100, 600, and 1100 live seeds/m2 and altering microsites with disturbance (none, sethoxydim herbicide, glyphosate herbicide, topsoil removal) and irrigation (none, pre-seeding, pre- and post-seeding) treatments. By the end of the study, microsite limitation was greater than that of seed limitation with greater C. lanceolata establishment in the glyphosate treatment than other disturbance treatments. Coreopsis lanceolata establishment was limited when seeded at 100 seeds/m2 but not at 600 seeds/m2. Seeding at 1100 seeds/m2 provided little increase in establishment compared to 600 seeds/m2. Effects of pre-planting herbicide treatments (none, glyphosate, imazapic) and post-planting mowing frequencies (two and six times/year) were assessed for three wildflower species (C. lanceolata, C. leavenworthii, and Gaillardia pulchella (firewheel)) on simulated roadsides at three sites in Florida. The glyphosate treatment resulted in greater wildflower establishment than the imazapic and control treatments. The imazapic treatment improved establishment of C. lanceolata only, which also had moderate cover in the control treatment. Mowing frequency did not affect wildflower percent cover or seed bank density, perhaps because mowing was reduced during wildflower blooming and seed dispersal.
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 Anne Frances.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Norcini, Jeffrey G.
Local: Co-adviser: Reinhardt Adams, Carrie H.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022660:00001

Permanent Link: http://ufdc.ufl.edu/UFE0022660/00001

Material Information

Title: Establishment and Management of Native Wildflowers on Florida Roadsides and Former Pastures
Physical Description: 1 online resource (128 p.)
Language: english
Creator: Frances, Anne
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: bahiagrass, coreopsis, cutting, disturbance, establishment, gaillardia, herbicide, lanceolata, leavenworthii, management, microsite, mowing, native, notatum, paspalum, pasture, pulchella, right, roadside, seed, wildflower
Horticultural Science -- Dissertations, Academic -- UF
Genre: Horticultural Science thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Establishing native wildflowers into areas dominated by Paspalum notatum var. saurae (bahiagrass) is a common goal of roadside management and ecological restoration. Yet, there is limited information on establishment and management practices for local ecotype seeds. This research sought to determine the effects of competition and disturbance on native wildflower establishment in areas dominated by P. notatum. Effects of planting season and post-planting disturbance (cutting) were assessed on competitive interactions between P. notatum and two Florida native congeners, Coreopsis lanceolata (lanceleaf tickseed) and C. leavenworthii (Leavenworth's tickseed) in two-species competition experiments. Coreopsis survival was greater in fall- than spring-established plants. In fall-established plants, C. lanceolata had higher survivorship than C. leavenworthii, although C. leavenworthii biomass and flower number were greater than that of C. lanceolata. Paspalum notatum reduced C. lanceolata biomass 58% and C. leavenworthii biomass 41%; however, conspecific neighbors reduced biomass of both Coreopsis species by at least 81%. Cutting decreased above- and belowground Coreopsis biomass by 55% and 30%, respectively. Seed and microsite limitations to establishment were assessed by seeding C. lanceolata at 100, 600, and 1100 live seeds/m2 and altering microsites with disturbance (none, sethoxydim herbicide, glyphosate herbicide, topsoil removal) and irrigation (none, pre-seeding, pre- and post-seeding) treatments. By the end of the study, microsite limitation was greater than that of seed limitation with greater C. lanceolata establishment in the glyphosate treatment than other disturbance treatments. Coreopsis lanceolata establishment was limited when seeded at 100 seeds/m2 but not at 600 seeds/m2. Seeding at 1100 seeds/m2 provided little increase in establishment compared to 600 seeds/m2. Effects of pre-planting herbicide treatments (none, glyphosate, imazapic) and post-planting mowing frequencies (two and six times/year) were assessed for three wildflower species (C. lanceolata, C. leavenworthii, and Gaillardia pulchella (firewheel)) on simulated roadsides at three sites in Florida. The glyphosate treatment resulted in greater wildflower establishment than the imazapic and control treatments. The imazapic treatment improved establishment of C. lanceolata only, which also had moderate cover in the control treatment. Mowing frequency did not affect wildflower percent cover or seed bank density, perhaps because mowing was reduced during wildflower blooming and seed dispersal.
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 Anne Frances.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Norcini, Jeffrey G.
Local: Co-adviser: Reinhardt Adams, Carrie H.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022660:00001


This item has the following downloads:


Full Text

PAGE 1

1 ESTABLISHMENT AND MANAGEMENT OF NATIVE WILDFLOW ERS ON FLORIDA ROADSIDES AND FORMER PASTURES By ANNE FRANCES A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2008

PAGE 2

2 2008 Anne Frances

PAGE 3

3 To my yaya, Emily Osmou Velelli.

PAGE 4

4 ACKNOWLEDGMENTS I would like to express my grat itude to m y chair, Dr. Jeff No rcini, and my cochair, Dr. Carrie Reinhardt Adams, for their guidance and for the many hours they contributed to this project. I would like to acknow ledge Dr. Jeff Norcini for his commitment to conserving Florida native wildflowers and for conceiving of the gr ant project that funded this research. I am grateful to Dr. Carrie Reinhardt Adams for helping me keep my eyes on the prize and for devoting as much energy to being a cochair as she would a chair. I would like to thank my committee members, Dr. Doria Gordon, Dr. Debbie Miller, and Dr. Sandy Wilson, who have graciously lent their expertise and provided guidance throughout the project. I am also indebted to Dr. Bijan Dehgan for serving as my cochair du ring the first two years of my program and for his continued mentorship. I am fortunate to have had such a supportive and involved committee. This project was made possibl e through the financial support of The Florida Department of Transportation. I am also indebted to Megha n Brennan, Mary Christman, and Ramon Littell for their statistical expertise that contributed to the project design and analysis. However, any errors contained herein are my own. My deepest gratitude goes to the wildflower crew, who cont ributed to this project above and beyond the call of duty. Ethan Stonerook assist ed with the seed bank study and with the setup of field plots in Citra. Kate Turner helped with virtually every aspe ct with this project and demonstrated her willingness to learn techniques in a subject far from her field of journalism. I am grateful to Claudia Peuela for helping to coordinate and impleme nt cutting treatments, necessary weeding, and data entry, also in a fiel d that was not her own. I would like to thank Emily Austen for approaching this research with as much attention to detail and dedication as if it were her own. I am indebted to the wildflower crew as a whole for their commitment to getting the job done well, which often included spending many hours in the heat, cold, or rain,

PAGE 5

5 with a seemingly endless number of pots, plots, and plants. In addition to their invaluable contribution to this project, I am also gr ateful for their friends hip and good company. Experiments conducted in Citra would not have been possible without the help of many capable people at the University of Floridas Pl ant Science Research and Education Unit. I am most indebted to Jim Boyer, who improved the project design with his knowledge of field experiments and willingly coordinated all of the experimental treatments. I am grateful to everyone who helped maintain the research pl ots and graciously shar ed their lunch room, including: Pete Fast Eddie Brown, Walt Davis, Ronald Tom Cutler, Tony Skinner, Roger Vath, Mike Durham, Carl Vining, Dan Beach, and anyone else I may have inadvertently forgotten. Their dedication made my work much eas ier and more enjoyable. I would also like to thank inmates from the Marion County Jail for thei r contribution to field work. I am especially grateful to the inmates who helped with the biomass harvest for their attention to detail and for teaching me the Soulja Boy dance. The experiment conducted in Fort Pierce bene fited from the help of many people at the University of Floridas Indian River Research an d Education Center. I am particularly grateful to Keona Muller, who coordinated and supervised all the experimental treatments, regularly monitored and photographed the research plots, and or ganized and participated in data collection. Pat Frey, Sandy Wilson, Carl Frost, and several others provided valuable assistance in the field. I give special thanks to Randy Burton for carry ing out the experimental treatments and for designing the hedge trimmer on wheels to mow the research plots. I am also grateful to Pat Frey and Judy Gersony for their gracious hospitality. I am much obliged to Jim Aldrich, Ama nda Brock, Barron Riddle, Tom Batey, Jeff Norcini, Tom Bolton, and the farm crew for assi sting with field work at the North Florida

PAGE 6

6 Research and Education Center in Quincy, Florid a. I am grateful not only for the many hours of work they contributed, but also fo r their efforts to make me feel welcome. I give special thanks to the Quincy dorm, for providing a wealth of in teresting stories and a greater appreciation for my own bed. Experiments conducted in Gainesville, Florid a benefited from the expertise of Jason Haugh, Joe Vasquez, Robert Smith, Brian Owens, and Doug Prevatt, who provided invaluable help in the greenhouse complex including buildi ng benches, installing irrigation, and applying pesticide. Marie Nelson, Cindy Olejownik, Richard Phelan and the rest of the Gainesville Environmental Horticulture Depa rtment helped with project logistics including reserving the state vehicle, coordinating payrol l for all the field assistants, and helping with computer and network issues. Additionally, the students, f aculty, and staff in the Plant Restoration and Conservation Horticulture Research Consortium provided support a nd a helping hand as needed throughout the project. They incl ude Dr. Hector Perez, Dr. Mike Kane, Scott Stewart, Danielle Watts, Phil Kauth, and Daniela Dutra. I would especially like to thank F Almira and Nancy Steigerwalt for including me in their labs and helping me with project logistics. Harvesting the biomass from the competition st udy (Chapter 2) required the help of many people. I am grateful to everyone who helped for working their hardest and for making it a fun experience. They include: Terry Byatt, Lisa Hager, Cassi e Elowe, Kei Egan, Yaro Neils, Alison Heather, Kara Monroe, Dan Holden, Alison Lugrin, Sheena Olimpo, Aron Guerrero, Stefanie Calvet, Joelle Szerdi, Jason Hyde, and Ma ndy Thomas. I am especially grateful to those who contributed a little extra by working in fall as well as summer, recruiting others, or otherwise giving the project their al l, including: Brittany Borck, Charlene Volpe, Dallas Scott, Beau Frail, Lynette Salas, Lauren Pell, Mark Turner, Juan Carlos Calderon, Mohamed Kouider,

PAGE 7

7 Felipe Osorio, Nancy Steigerwalt, and Julie Sorenson. I would also like to acknowledge the individuals who volunteered their time to the project, including: Kristen Bartlett Grace, Dr. Carrie Reinhardt Adams, Dr. Pete Adams, Regina Frances, and Simon Frances. Although the results of the Miami roadside plot s are not included in this dissertation, the study served as the basis for some of the experiments presented he re. Therefore, I would like to thank Lauren Linares, formerly with the Flor ida Turnpike, for identif ying the study areas and coordinating efforts with the Turnpike. Bruce Ma ntel, Bryan Nipe, and Andres Aquino, all with the Florida Turnpike, helped make the project possible. Kristie Wendelberger, Hannah Thornton, Steve Woodmansee, and Regina Frances provided invaluable fiel d assistance. Hannah Thornton helped with the initial design and patiently taught me how to delineate plots accurately with a compass and measuring tape. Steve Wo odmansee also helped identify plant specimens from the seed bank study. I am indebted to The North Carolina Bota nical Garden for introducing me to the importance of native plant conservation. I would also like to acknowledge Claudia Alta "Lady Bird" Taylor Johnson (1912-2007) for spearheadin g the Highway Beautific ation Act and for her commitment to natural resource conservation. I would like to express my gratitude to Lisa Hager, Natalie Boodram, and Jensen Montambault for their friendship and support over the years. I am especially grateful to Lisa Hager, who not only helped to format the references in this dissertation, but also served as a mentor throughout this process. I would like to thank Brian Frankel for his unconditional support over the past three years and for helping me to keep things in perspective. I am most grateful for his ability to take my ideas seriously while encouraging me to not take myself too seriously.

PAGE 8

8 I cannot express enough gratitude to my parents, who have supported me emotionally and financially, and provided voluntee r field assistance, home-cooked meals, and enough worrying to ensure that nothing could go wrong. Finally, I w ould like to acknowledge the sacrifices made by my grandparents and parents that made it possible for me to pursue this path.

PAGE 9

9 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4LIST OF TABLES................................................................................................................. ........11LIST OF FIGURES.......................................................................................................................12ABSTRACT...................................................................................................................................14 CHAP TER 1 INTRODUCTION..................................................................................................................16Overview and Rationale.........................................................................................................16Use of Local Ecotypes............................................................................................................18Seed and Microsite Limitations to Establishment.................................................................. 20Role of Competition............................................................................................................ ...22Effects of Disturbance............................................................................................................24Chemical Disturbance (Herbicide)..................................................................................25Mechanical Disturbance (Topsoil Removal and Mowing)............................................. 26Native Wildflower Species..................................................................................................... 28Research Design and Objectives............................................................................................292 EFFECTS OF PLANTING SEASON, INITIAL ADVANTAGE, AND CUTTING FREQUENCY ON THE COMPETITIVE INTERAC TIONS BETWEEN A NONNATIVE PASTURE GRASS AND TWO NATIVE WILDFLOWERS............................... 32Introduction................................................................................................................... ..........32Methods..................................................................................................................................35Study Species and Site.....................................................................................................35Experimental Design....................................................................................................... 36Data Collection and Analysis.......................................................................................... 38Results.....................................................................................................................................39Season and Species.......................................................................................................... 39Initial Advantage and Competition................................................................................. 39Cutting Frequency...........................................................................................................40Coreopsis Fitness............................................................................................................. 40Discussion...............................................................................................................................41InterVersus Intraspecific Competition.......................................................................... 41Seasonal Differences.......................................................................................................42Species-Specific Responses.............................................................................................43Disturbance......................................................................................................................44Initial Advantage............................................................................................................. 45Conclusions.............................................................................................................................47

PAGE 10

10 3 EVALUATING SEED AND MICROSITE LIMITATION TO THE ESTAB LISHMENT OF A NATIVE WILDFL OWER IN A NON-NATIVE PASTURE GRASS...................................................................................................................................55Introduction................................................................................................................... ..........55Methods..................................................................................................................................59Study Site and Species..................................................................................................... 59Experimental Design....................................................................................................... 60Data Collection and Analyses......................................................................................... 62Results.....................................................................................................................................63Coreopsis lanceolata Changes Over Time...................................................................... 63Coreopsis lanceolata Establishment...............................................................................64Coreopsis lanceolata Fitness...........................................................................................65Community Vegetation...................................................................................................66Soil Characteristics..........................................................................................................68Discussion...............................................................................................................................68Interactive Effects of Seed and Microsite Availability................................................... 68Disturbance Effects on Microsite Quality....................................................................... 69Requirements for Emergence Versus Establishment...................................................... 72Effects on Community Vegetation.................................................................................. 73Conclusions.............................................................................................................................744 MOWING FREQUENCY AND HERBICI DE EFFECT S ON ESTABLISHING NATIVE WILDFLOWER POPULATIONS ON SIMULATED ROADSIDES................... 82Introduction................................................................................................................... ..........82Methods..................................................................................................................................84Study Sites and Species................................................................................................... 84Experimental Design....................................................................................................... 86Seed Bank Study..............................................................................................................87Statistical Analysis.......................................................................................................... 88Results.....................................................................................................................................89Gaillardia pulchella ........................................................................................................89Coreopsis lanceolata .......................................................................................................90Coreopsis leavenworthii ..................................................................................................91Discussion...............................................................................................................................93Species-Specific Responses.............................................................................................93Effects of Imazapic.......................................................................................................... 95Mowing Treatments......................................................................................................... 96Paspalum notatum Dominance........................................................................................96Conclusions.............................................................................................................................97CONCLUSIONS.................................................................................................................... ......109LIST OF REFERENCES.............................................................................................................118BIOGRAPHICAL SKETCH.......................................................................................................128

PAGE 11

11 LIST OF TABLES Table page 2-1 Effects of species ( Coreo psis leavenworthii C. lanceolata ), cutting frequency (none, once/month, twice/month), planting treatmen t, and their interactions on aboveand belowground biomass of the target (central) Coreopsis in fall-established plants............ 482-2 Effects of species ( Coreopsis leavenworthii C. lanceolata ), cutting treatment (none, once/month, twice/month), planting treatmen t, and their interactions on flower number of the target (central) Coreopsis in fall-established plants...................................483-1 Disturbance treatment effects (control, sethoxydim, glyphosate, scraped) on soil chemistry before seeding Coreopsis lanceolata in a former pasture in Citra, FL............. 753-2 Effects of irrigation (none, pre, full), seeding rate (low, me dium, high), disturbance (control, sethoxydim, glyphosate, scraped), and time (weeks after seeding) on density and percent cover of Coreopsis lanceolata in a former pasture in Citra, FL........ 763-3 Effects of irrigation (none, pre, fu ll), seeding rate (l ow, medium, high), and disturbance (control, sethoxydim, glyphos ate, scraped) on percent cover of Paspalum notatum forbs, and graminoids 10 and 46 weeks after seeding Coreopsis lanceolata in a former pasture in Citra, FL......................................................................................... 774-1 Soil characteristics and types be fore seeding native wildflowers ( Gaillardia pulchella Coreopsis lanceolata C. leavenworthii ) at the study sites located in Quincy, Citra, and Fort Pierce, FL.....................................................................................984-2 Dates of mowing treatments used at each site (Quincy, Citra, Fort Pierce, FL) for each species ( Gaillardia pulchella, Coreopsis lanceolata C. leavenworthii )...................994-3 Effects of establishment treatment (pre -seeding herbicide: control, glyphosate, and imazapic), mowing frequency (two or six times/year), and their interactions on wildflower percent cover of Gaillardia pulchella in Quincy and Citra, FL.................... 1004-4 Effects of establishment treatment (pre -seeding herbicide: control, glyphosate, and imazapic), mowing frequency (two or six times/year), and their interactions on wildflower percent cover of Coreopsis lanceolata in Quincy and Citra, FL................... 1014-5 Effects of establishment treatment (pre -seeding herbicide: control, glyphosate, and imazapic), mowing frequency (two or six times/year), and their interactions on wildflower percent cover of Coreopsis leavenworthii in Quincy, Citra, and Fort Pierce, FL.........................................................................................................................1024-6 Effects of establishment treatment (pre -seeding herbicide: control, glyphosate, and imazapic), mowing frequency (two or six times/year), and their interactions on emerged wildflower seedlings from the soil seed bank in fall 2006 in Quincy, Citra, and Fort Pierce, FL..........................................................................................................103

PAGE 12

12 LIST OF FIGURES Figure page 2-1 Total monthly rainfall and minimum and ma xi mum temperatures at the study site in Gainesville, FL during the experiment.............................................................................. 492-2 The two monocultures and four mixtures included in the planting treatments.................. 502-3 Effects of planting season (fall or spring) and species ( Coreopsis lanceloata or C. leavenworthii) on Coreopsis survival (target plant) ei ght months after planting.............. 512-4 Effects of neighbor plant species on a boveand belowground biomass of the target (center) Coreopsis in fall-established plants...................................................................... 522-5 Effects of cutting frequency on aboveand belowground biomass of the target (center) Coreopsis neighboring Coreopsis and neighboring Paspalum in fallestablished plants............................................................................................................. ..532-6 Effects of cutting frequency and planting treatment on the mean number of C. leavenworthii and C. lanceolata flowers in fall-established plants...................................543-1 Minimum and maximum temperatures at 60 cm, solar radiation, and precipitation at the study site in north-central Florida................................................................................ 783-2 Effects of irrigation, seeding rate, and disturbance treatments on Coreopsis lanceolata density and percent cover ................................................................................ 793-3 Effects of seeding rate and di sturbance treatment on shoot biomass of Coreopsis lanceolata Paspalum notatum forbs, and graminoids 46 weeks after seeding................ 803-4 Effects of seeding rate and dist urbance treatment on percent cover of Coreopsis lanceolata flowers 20 weeks after seeding........................................................................813-5 Effects of seeding rate and disturbance treatment on number of Coreopsis lanceolata seedlings 46 weeks after seeding....................................................................................... 814-1 Solar radiation, temperature, and rainfa ll in Fort Pierce, Citra, and Quincy, FL during the study period.................................................................................................... 1044-2 Effects of establishment treatment (pre-s eeding herbicide: control, glyphosate, and imazapic) on wildflower percent cover by site, season, and year.................................... 1054-3 Composition of Gaillardia pulchella forbs, and graminoids in the seed bank in fall 2006 and as aboveground vegetation in spring 2007 in Quincy and Citra, FL............... 106

PAGE 13

13 4-4 Composition of Coreopsis lanceola ta, forbs, and graminoids in the seed bank in fall 2006 and as aboveground vegetation in spri ng 2007 in Quincy and Citra, FL............... 1074-5 Composition of Coreopsis leavenworthii forbs, and graminoids in the seed bank in fall 2006 and as aboveground vegetation in sp ring 2007 in Quincy, Citra, and Fort Pierce, FL.........................................................................................................................108

PAGE 14

14 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy ESTABLISHMENT AND MANAGEMENT OF NATIVE WILDFLOW ERS ON FLORIDA ROADSIDES AND FORMER PASTURES By Anne Frances August 2008 Chair: Jeffrey G. Norcini Cochair: Carrie Reinhardt Adams Major: Horticultural Science Establishing native wildflow ers into areas dominated by Paspalum notatum var. saurae (bahiagrass) is a common goal of roadside manage ment and ecological restoration. Yet, there is limited information on establishment and management practices for local ecotype seeds. This research sought to determine the effects of co mpetition and disturbance on native wildflower establishment in areas dominated by P. notatum Effects of planting season and post-planti ng disturbance (cutting) were assessed on competitive interactions between P. notatum and two Florida native congeners, Coreopsis lanceolata (lanceleaf tickseed) and C. leavenworthii (Leavenworths tickseed) in two-species competition experiments. Coreopsis survival was greater in fallth an spring-established plants. In fall-established plants, C. lanceolata had higher survivorship than C. leavenworthii although C. leavenworthii biomass and flower number were greater than that of C. lanceolata Paspalum notatum reduced C. lanceolata biomass 58% and C. leavenworthii biomass 41%; however, conspecific neighbors reduced biomass of both Coreopsis species by at least 81%. Cutting decreased aboveand belowground Coreopsis biomass by 55% and 30%, respectively.

PAGE 15

15 Seed and microsite limitations to es tablishment were assessed by seeding C. lanceolata at 100, 600, and 1100 live seeds/m2 and altering microsites with disturbance (none, sethoxydim herbicide, glyphosate herbicide, topsoil removal) and irrigation (none, pre-seeding, preand post-seeding) treatments. By the end of the st udy, microsite limitation was greater than that of seed limitation with greater C. lanceolata establishment in the glyphosate treatment than other disturbance treatments. Coreopsis lanceolata establishment was limited when seeded at 100 seeds/m2 but not at 600 seeds/m2. Seeding at 1100 seeds/m2 provided littl e increase in establishment compared to 600 seeds/m2. Effects of pre-planting herbicide treatments (none, glyphosate, imazapic) and postplanting mowing frequencies (two and six times/year) were assessed for three wildflower species ( C. lanceolata C. leavenworthii and Gaillardia pulchella (firewheel)) on simulated roadsides at three sites in Florida. The glyphosate treatment re sulted in greater wildflow er establishment than the imazapic and control treatments. The im azapic treatment improved establishment of C. lanceolata only, which also had moderate cover in th e control treatment. Mowing frequency did not affect wildflower percent cover or seed bank density, perhaps because mowing was reduced during wildflower blooming and seed dispersal.

PAGE 16

16 CHAPTER 1 INTRODUCTION Overview and Rationale As more land is developed for human use, biodiversity and habitat for native species continue to decline. Given increased huma n population growth and increased per capita consumption of resources, continued development seems inevitable (Tilman et al. 2001). Conversion of land to agriculture and development negatively affects biodiversity conservation through habitat destruction and fragmentation, climate change, pollution, and the spread of invasive species, among others. While traditiona l conservation efforts have focused on areas less disturbed by human development (i.e., natural areas), the presence of altered landscapes has become ubiquitous in many parts of the world. A ltered landscapes may be managed or restored to lessen negative impacts on conservation areas or undeveloped habitat, and, in some cases, conserve biodiversity. Conser vation efforts in altered landsca pes have therefore become a necessity in achieving the ove rall objectives of conservation (Grimm et al. 2008). A consequence of development and population growth is an increase in the number of roads. Within the past century, roads have b ecome a widespread component of our landscape, ubiquitous in almost every part of the United St ates (Forman and Alexander 1998). The area that roads impact is much greater than the space th ey occupy (Forman and Alexander 1998, Watts et al. 2007). Roads not only alter the land they enco mpass, but also impact the terrestrial and aquatic ecosystems in which they occur and influence adjacent land use (Trombulak and Frissell 2000). Roads may also provide corridors for in vasive, non-native species to expand their range (Gelbard and Belnap 2003). Conservation efforts on roadsides are focused on minimizing negative impacts on wildlife and surrounding natu ral areas. Mortality of animals through vehicle collisions has been

PAGE 17

17 reduced by incorporating wildlife crossings into roadside plans (Bissonettea and Adairb 2008). Modifying management activities may also help to protect rare plant species that occur on roadsides. Planting and conserving native speci es on roadsides may decrease the occurrence of non-native species and serve to buffer surroundin g natural areas from th e negative effects of roads. Native wildflowers on roadsides may also help to meet Departments of Transportation other objectives including roadside beautification and reduced costs th rough decreased mowing (Florida Department of Transportation 1998, Harp er-Lore and Wilson 2000). For these reasons, there has been an increased interest in planting native species on roadside s (Gordon et al. 2000). As undeveloped land becomes increasingly scarce, restoring former agricultural land has become a goal of many restoration projects (Zimmerman et al. 2000, Stevenson and Smale 2005, Sampaio et al. 2007, Orrock et al. in press ). Some of the area availa ble for restoration in the southeastern United States is abandoned pastur eland. In Florida alon e, pasture and rangeland occupy over 2.1 million ha and often overlap sens itive wetland areas (Swain et al. 2007). However, several barriers exist to restoring native communities on former pasture. Abandoned agricultural areas often ha ve altered hydrology and levels of soil nutrients, which can affect plant productivity. The dominant species (usually a grass) may prevent the establishment of new species and may be difficult to eradicate. Moreover, native plant propagules to recolonize the site may be lacking. Ecological restoration and road side beautification projects often require the introduction of desirable plant species. Plants provide th e basic structure and pr imary production upon which many other species and processes depend and are therefore an in tegral part of terrestrial revegetation and restorati on projects. In Florida, Paspalum notatum Flgg var. saurae Parodi (bahiagrass) occurs on roadsides as well as on land slated for restoration (Violi 2000).

PAGE 18

18 Therefore, establishing nativ e wildflowers (forbs) into P. notatum is a common goal of roadside management and ecological restoration projec ts. Yet, little is known about effective establishment and management methods. Eviden ce suggests that wildflow er establishment is limited by competition from P. notatum although this has not been experimentally tested. Establishment methods may be different on roadsi des and pastures than natural areas due to altered hydrology and soil nutrien ts. Disturbance treatments that eradicate or reduce P. notatum may increase the establishment and sustainability of native wildflower popul ations, yet this has not been confirmed. Moreover, lo cal ecotype seeds have recently become available, which may require different establishment and management methods than seeds from nonlocal sources. Use of Local Ecotypes There is an increasing concern about the risks of introducing nonlocal native plants in restoration projects and road side plantings, including gene tic swa mping and outbreeding depression of local populations (Booth and Jones 2001). For this reason, restoration protocol often suggests using plant sources of local prove nance or local ecotypes (Booth and Jones 2001). Similarly, the Florida Department of Trans portation would like to expand their current wildflower program to increase the number of local ecotype, native wi ldflower populations on roadsides (Florida Department of Transporta tion 1998). The ecotype co ncept was introduced by Turesson in the 1920s to describe genetic varieties within a single species (Lubchenco and Real 1991). An ecotype is a population of a species that has adapted to a particular set of environmental conditions via natural selection. Adaptation to the introduction s ite is not only essential for th e survival of the introduced individuals but also for the pe rsistence of the popul ation over time. Local ecotypes may be better adapted to the climate at the introduction site than nonlocal ecotypes. For example, in a study examining effects of population size and dist ance of introduced species, Vergeer et al.

PAGE 19

19 (2004) found evidence of local adaptation. In pl ants, ecotypic differentiation can influence a number of processes including freezing tolerance a nd germination. For example, in response to low temperatures, northern ecotypes of two w oody tree species acclimated more rapidly and developed greater tolerance to freezing than sout hern ecotypes (Li et al. 2004, Li et al. 2005). In several species, populations origin ating from sites with severe winters required long periods of chilling to break dormancy while populations from warmer sites germinated very quickly under the same conditions (Allen and Meyer 1998). Mo reover, Norcini et al. (2001) found that local ecotypes of Coreopsis lanceolata in Florida bloomed earlier and had higher survivorship than nonlocal ecotypes. The need to develop safe transfer zones fo r native plant material has been expressed by state and federal agencies (United States De partment of Agriculture and United States Department of the Interior 2002). Transfer zones provide physical boundaries within which ecotypes of species can safely be transferred, without negatively impacting the genetics of plant metapopulations. Several state programs are alrea dy in place. For example, the Iowa Ecotype Project was developed to increase the availability of Iowa-origin seed for roadside plantings and prairie reconstructions (Houseal and Smith 2000). Similarly, Miss ouri started a local ecotype program with 33 species from two prairie ecozones, which were based on climate and soil conditions (Erickson and Navarrete-Tindall 2004) In addition, the Oregon Department of Forestry (2007) developed seed transfer z ones for many tree species. Although some programs are already in place, a consortium of federal ag encies recommended that funding and research be devoted to developing transfer zones (United Stat es Department of Agriculture and United States Department of the Interior 2002).

PAGE 20

20 In Florida, the demand for native wildflower seed for roadside plantings exceeds the availability of local ecotype seed (Florida Depa rtment of Transportation 1998). Within the last five years, local ecotype seeds of several speci es of native wildflowers have become available through Floridas Wildflower Seed and Plant Growers Association, In c. (2008). Currently, limited information exists on effective methods to establish local ecotype native wildflowers on Florida roadsides (but see No rcini and Aldrich 2004, Florida Department of Transportation 2005). Because Florida ecotypes are adapted to Floridas climate, establishment and management practices may be different than those for nonlocal seed sources and on altered landscapes such as roadsides. Seed and Microsite Limitations to Establishment Direct seeding of propagules in to restoration sites is often the most econom ical approach for revegetation (Pfaff and Gonter 1999). However, in order for a plant to become established at a new site, the seed (or propagule) must not only arrive at the site but also germinate and survive the seedling stage of development. For this reason, plant establishment is said to be limited by the availability of seeds, of areas suitable fo r emergence and growth, or of both (Stampfli and Zeiter 1999, Turnbull et al. 2000, Zobel et al. 2000, Holzel 2005). If a species does not exist at a site due to dispersal limitation, the si te is said to be seed limited. In restoration projects and roadside plantings the number of seeds introduced to the site is determined by the seeding rate. Yet, there is limited information on the effects of seeding rate on establishment. Seeding rates are rarely based on parameters th at ensure establishment of the restored population (e.g., Cole 2007) and are more often chosen based on seed availability or limited empirical data. For example, the Florid a Department of Transportation uses the same bulk seeding rate for several wildflower species because species-specif ic requirements are not yet known (Florida Department of Transportation 1998). Seeding at deliberately high rates is

PAGE 21

21 practiced to increase the seeded species compe titive ability over existi ng vegetation. In a study on native grass restoration in California, doubling the seed density resulted in twice as many established grass plan ts (Orrock et al. in press ). While increasing the seeding rate may increase overall emergence, Harkess and Lyons (1998) found that doubling the seeding rate did not affect the number of plants maturing to flower. Mo reover, increasing the s eeding rate of native legumes negatively affected individual plan t establishment (Fischbach et al. 2006). Plant establishment of introduced populat ions not only depends upon the appropriate seeding rate but also upon creating areas suitable for seedling emergence and survival, i.e., microsites, which are also known as safe sites or regeneration niches. Microsites suitable for germination and seedling establishment are sp ecies-specific, but usually provide water and contact with the soil (Isselstei n et al. 2002). In natural system s, water is present in the soil through rainfall, groundwater, or a nearby wetland. However, water availability on degraded sites may be different than adjacent natural systems due to altered hydrology. Therefore, germination of introduced seeds on degraded site s may be particularly limited in these systems by the lack of water at the soil surface. Moreov er, seeding may occur in a different season than natural seed dispersal or when rainfall is limited. In these cases, supplemental irrigation can increase the seeded species survival and esta blishment (Harper-Lore and Wilson 2000, Cox et al. 2004). Microsites for new species may be limited by the existing vegetation (Pfaff and Gonter 1999), which can prevent seed-to-soil contact and limit resources like water, light, and nutrients through competition (Tilman 1994). For exampl e, increased productivity in grasslands prevented seedling establishment of forbs by d ecreasing light levels (F oster and Gross 1998). Seeds of some species require exposure to high ra tios of red to far-red lig ht to germinate (Rees

PAGE 22

22 1997, Baskin and Baskin 2001). Existing canopie s decrease the red to far-red ratio, which inhibits germination of some sp ecies (Rees 1997). Disturbance tr eatments that reduce or remove the existing vegetation have led to increased establishment of desi rable species in other habitats (Holzel 2005, Martin an d Wilsey 2006). Although there is little published information on the esta blishment of native wildflowers in P. notatum in Florida, evidence suggests that na tive forbs introduced into communities dominated by P. notatum often fail to establish (Uridel 1994, Violi 2000). Selected for its drought tolerance and lack of pest problems, P. notatum was originally introduced as a pasture grass but is also commonly planted on roadsides due to its effectiveness in controlling erosion through the formation of de nse mats (Violi 2000). Paspalum notatum was introduced to the United States from Brazil in 1913 (Scott 1920) and was estimated to occupy over 2 million ha in the southeastern U.S. by 1978 (Beatty and Powell 1978). Studies on former pastures in Florida indicate that disturbing or removing P. notatum can lead to increased establishment of native wildflowers (Uridel 1994, Gordon et al. 2000). Ba sed on these field observations, competition is hypothesized to be the driv ing interaction between P. notatum and native wildflowers, although this has not been experimentally tested. Role of Competition Com petition is hypothesized to play an impor tant role in structuring biological communities and has been extensively studied in both agricultural (Radosevich and Rousch 1990, Booth et al. 2003) and natural (Grace and T ilman 1990, Tilman 1990) systems. Processes governing plant-plant interactions have focuse d on competition because species that coexist often compete for limited resources like light, wa ter, and nutrients (Gol dberg 1990). In plants, competition occurs aboveground through shoots and belowground through roots. Aboveground competition is usually expressed as competition for light. Since light is unidirectional, taller

PAGE 23

23 species usually are better competitors for light as they shade out shorter species (Hely and Roxburgh 2005, Lepik et al. 2005). Belowground competition is usually mediated through limiting resources like nutrients and water (Tilman 1994). Aboveground competition is usually stronger when light is limiting and belowground competition is usually stronger when water and nutrients are limiting. Two-species competition e xperiments have been used to determine the stronger competitor under a particular set of conditions (Connolly and Wayne 2005). In field experiments involving P. notatum and native wildflowers, competition is assumed to be limiting establishment. However, competitive interactions may be influenced by initial advantage, seasonal differences and disturbance treatments. Since P. notatum has been planted on roadsides and pastures in the southeaste rn United States for many years, vegetation on these sites often consis ts of well-established monotypic stands of P. notatum Competition between species can be influenced by their order of arrival at a site (Ejrnaes et al. 2006). Regarding competition for light, the first individual s to arrive at a site (or larger individuals) often have a competitive advantage over other individuals (Hely and Roxburgh 2005). Competitive ability between two species can therefore depend on the order of arrival, species identity, or both (Hely and Roxburgh 2005). When native wildflowers are introduced into wellestablished monotypic stands of P. notatum assessing the competitive ability between P. notatum and native wildflowers may be confounded with P. notatums initial advantage over the wildflower species. The importance of arrival order in determining species composition has been called historical contingency (Ejrnaes et al. 2006), while the eff ects of arrival order on competitive interactions are referre d to as initial advantage or th e founder effect (Perry et al. 2003, Hely and Roxburgh 2005).

PAGE 24

24 Seasonality can influence competitive interactions between P. notatum and wildflower species. P. notatum is a long-day plant that is dorman t during the short days of winter (Marousky and Blondon 1995). Flow ering is initiated when day length exceeds 13.5 hours (Marousky and Blondon 1995). In Florida, P. notatum flowering usually begins in May and seed maturation occurs throughout the summer (Watson and Burton 1985). Paspalum notatum height can exceed 100 cm when in flower, compared with approximately 20 cm when dormant (personal observation). The in crease in height during the su mmer months likely intensifies aboveground competition between P. notatum and wildflower species. Planting wildflowers in the fall, when P. notatum is dormant, may lead to increased establishment. Disturbance can also influence co mpetition between wildflowers and P. notatum By reducing asymmetric competition for light (Lep 1999), disturbance can temporarily disrupt competitive exclusion. For example, the compe titive effects of grasses on herbaceous species can be temporarily alleviated by removing the gr ass via herbicide (del-Val and Crawley 2005). If competition from P. notatum is limiting wildflower establishment, disturbing P. notatum may increase wildflower establishment. Effects of Disturbance Disturbance is an integral co m ponent of ecological systems and can be broadly defined as any event that causes changes to community structure and functi on. Plant disturbance is often defined as a discrete event that reduces biomass in a community. Disturba nce regime refers to both the type and frequency of disturbance. The frequency of disturbance can greatly affect the species composition and diversity of an ecosyst em. Ecological theory predicts that an intermediate frequency of disturbance can increase diversity in some habitats (Connell 1978). Increased species richness and diversity in re sponse to moderate disturbance are welldocumented (Dayton 1971, Lep 1999).

PAGE 25

25 While disturbance occurs in all habitats, some habitats are maintained by frequent disturbance. For example, longleaf pine ecosystems in the southeastern United States are maintained by frequent fire (Taylor 1998). Many of Floridas native wildflowers naturally occur in habitats with frequent dist urbance (Taylor 1998). However, the type and frequency of disturbance in natural systems are usually diffe rent than on roadsides and pastures. While roadsides and pastures are subjec t to both natural and artificial disturbance regimes, natural disturbance regimes like fire a nd flooding tend to be prevented or controlled on roadsides for safety reasons. Moreover, management activi ties on roadsides include disturbances such as herbicide applicati on and mowing (Forman et al. 2003). Chemical Disturbance (Herbicide) Disturbance treatm ents include practices that reduce or re move the existing vegetation. Disturbance can be chemical or mechanical and can occur before or after wildflower emergence. Chemical disturbance includes the use of herbicides to eradicat e or reduce undesirable vegetation without disturbing the soil. Non-selective, or broad-spectrum herb icides, are designed to injure or kill any plant they contact, wh ile selective herbicides affect se nsitive species but do not affect tolerant species (Mon aco et al. 2002). Glyphosate (e.g., Roundup) is an example of a non-selective herbicide (Monaco et al. 2002). Applying glyphosate to P. notatum increased the establishment of native species subsequently planted in P. notatum pastures (Gordon et al. 2 000; Uridel 1994). Because glyphosate is non-selective, it should be applie d before wildflower seeding or planting. Sethoxydim (e.g., Poast, Vantage) is a graminicide, i.e., grasses are sensitive to the herbicide but dicots and non-grass monocots are either not sensitive or have very low sensitivity (Monaco et al. 2002). Since dicots a nd non-grass monocots are tolerant of graminicides, sethoxydim can

PAGE 26

26 be used for the postemergence control of grasses with little damage to other species. Thus, sethoxydim can be applied at any time during wildflower establishment. Imazapic (Plateau, Impose, Panoramic 2S L) is a selective herbicide that has improved the restoration of native species in certain habitats (Mas ters et al. 1996). The tolerance of wildflowers to imazapic varies considerably according to species, ecotype, and rate of application (Norcini et al. 2003). Wildflowers are generally more tolerant to imazapic when applied preemergence compared to postemerge nce (BASF Corporation 2003, Norcini et al. 2003). When imazapic is applied at the time of wildflower seeding, it can decrease the density and growth rate of some existing vegetation while minimally impacting the emergence or establishment of certain wild flower species (Beran et al 1999, Norcini et al. 2003). On roadsides, the complete eradication of P. notatum may not be desirable because of the possibility of increased erosion. However, while P. notatum effectively stabilizes roadsides, it continuously produces tall seedheads throughout the summer, which may require frequent mowing to maintain acceptable driv er visibility (Baker et al. 1999). Herbicides at sublethal rates can be used as plant growth retardan ts to limit seedhead production of P. notatum and therefore reduce the frequency of mowing (Baker et al. 1999, Florida Department of Transportation 2005). The reduced P. notatum growth may have the added be nefit of increasing wildflower establishment, without completely eradicating P. notatum Glyphosate, sethoxydim, and imazapic have suppressed seedhead production of P. notatum when applied at low rates (Baker et al. 1999). Mechanical Disturbance (Topsoil Removal and Mowing) Mechanical disturbance includes treatm ents that physically remove the existing vegetation and typically include tilling, disking, and mo wing. Topsoil removal, or scarification, is also an example of mechanical disturbance. Topsoil removal has led to increased native

PAGE 27

27 species establishment in former agricultural ar eas by removing the dominant vegetation as well as decreasing artificially enha nced soil nutrient levels (Holze l 2005, Buisson et al. 2006, Clark and Tilman 2008). Topsoil removal may also remove the majority of weed seeds in the soil seed bank (Roberts 1981, Dalrymple et al. 2003). A combination of repe ated herbicide application and sod removal through disking or plowi ng has been shown to effectively control P. notatum (Violi 2000); however, it is not cl ear if combining herbicide tr eatments with mechanical sod removal results in greater P. notatum control than either treatment alone. Mowing is another type of m echanical disturbance. Unlik e topsoil removal or tilling, mowing disturbs the existing vegeta tion but does not disturb the soil. Mowing of roadside vegetation is the most common and widespread fo rm of roadside management, and is primarily utilized for driver safety. Th e purpose of mowing is to reduce the height of vegetation, which can interfere with driver visibi lity, and to prevent the growth of woody species, which could harm drivers who pull off of the road. The frequency, timing, and height of mowing can influence plant species composition on roadsides. Grasses tend to thri ve in areas that are frequently mowed, due to their ability to rapidly grow new shoots from a ground-level apical meristem (Forman et al. 2003). Mowing has been shown to increase species diversity (P arr and Way 1988, Collins et al. 1998, Hoffmann et al. 2004) in some habitats. In a long-term study comparing mowing frequencies, the diversity of roadside vegetation was greater in areas mowed tw o times per year than once per year or not at all, and the lowest diversity was found in area s that were not mowed (Parr and Way 1988). Although the frequency of roadside mowing varies, it is generally mu ch more frequent in Florida (two-twelve times per year) than na turally occurring disturbance.

PAGE 28

28 Mowing of roadsides can affect the recruitm ent of wildflower populations positively by providing regeneration niches and negatively by limiting repr oduction. Mowing and removing litter can increase germination and establishmen t of seedlings (Jensen and Meyer 2001, Jutila and Grace 2002). However, the time of year an area is mowed can significantly impact the growth and recruitment of a species, especially if mowing precedes seed set (Brys et al. 2004). In both annual and perennial species, mowing that precedes seed maturation will likely limit recruitment and decrease the sustainability of the population. If the species is an annual, the population would only persist at that site if viable seeds were stored in the soil. Conversely, mowing that occurs after seed maturation coul d facilitate seed disp ersal and subsequent emergence of dispersed seeds. Flower and seed phenology vary according to wildflower species, weather, and site conditions, ma king specific mowing recommendati ons difficult to predict. Native Wildflower Species Many wildflowers native to Florida are annuals to short-lived perennial s that invest in reproductive growth early in life (Taylor 1998, Wunderlin and Hansen 2003). In fact, wildflowers are planted or m aintained on roadside s for their attractive display of flowers. However, different wildflower species ma y respond differently to abiotic conditions, disturbance, and competition from P. notatum Different plant traits or life-history strategies within a plant community may promote specie s coexistence (Coomes and Grubb 2000, Booth et al. 2003). However, plant survival strategies normally involve trade-offs. For example, plants that are good colonizers are usually not good competitors over time (Coomes and Grubb 2000). Therefore, some wildflower species may be stronger competitors while other wildflowers may recruit more successfully. Additionally, establishm ent of each wildflower species may be greater in certain habitats than others.

PAGE 29

29 Since species-specific traits can influence the establishment of native wildflowers, three native wildflower species were included in this study. All three species ar e in the Asteraceae and include: Gaillardia pulchella Foug. (firewheel), Coreopsis lanceolata L. (lanceleaf tickseed), and Coreopsis leavenworthii Torr. & A. Gray (Leavenwort hs tickseed) (United States Department of Agriculture 2008). Gaillardia pulchella is an annual to short-lived perennial commonly occurring throughout Florida in di sturbed uplands (Wunderlin and Hansen 2003) Coreopsis lanceolata is an evergreen, short-lived perennial that occurs thr oughout much of the United States (Flora of North America Editor ial Committee 2006, United States Department of Agriculture 2008). In Florida, C lanceolata occurs in northern and north-central regions in sandhill and disturbed habitats (Wunderlin and Hansen 2003) Coreopsis leavenworthii is an annual to short-lived perennial and a faculta tive wetland species (Wunderlin and Hansen 2008) Coreopsis leavenworthii is common throughout Florida and occurs in depression marshes, disturbed wetland, marl prai rie, pine rockland, wet flatwood, and wet prairie habitats (Gann et al. 2008). Research Design and Objectives Effective methods are needed to establish and manage native wildflowers on Florida roadsides and for mer pastures. Local ecotype se eds have recently become available, which may require different establishment and management methods than seeds from nonlocal sources. Moreover, conditions on roadsi des and pastures often do not mimic the biotic and abiotic characteristics of native habitats Anecdotal evidence suggests that wildflower establishment is limited by competition from P. notatum although this has not been experimentally tested. Moreover, the effects of preand post-plant ing disturbance on planted native wildflower populations are unknown. This research seeks to determine the effects of competition and disturbance on native wildflower establishmen t and recruitment on roadsides and pastures

PAGE 30

30 dominated by P. notatum Results from this research will provide practical information to the Florida Department of Transportation to guide the establishment and management of local ecotype native wildflower species on roadsides. To fulfill these objectives, three experiments were conducted, which are organized into separate chapters for this dissertation: In Chapter 2, I examine the competitive effects of P. notatum on two Florida native congeners, C. lanceolata and C. leavenworthii This experiment t ook place under controlled conditions, with initial establishment and density held constant in order to focus on competition. The competitive ability of each species was assessed at different frequencies of disturbance (post-planting) in a seri es of two-species competition experiments. The disturbance treatment involved cutting aboveground growth to simulate mowing. The effects of initial advantage and season of planting on competition were also assessed. In Chapter 3, I examine the relative importa nce of seed and microsite limitation to wildflower establishment in a P. notatum pasture. By focusing on the establishment of one native wildflower species, C. lanceolata I was able to test severa l levels of both seed and microsite limitation simultaneously. To test for seed limitations to establishment, C. lanceolata was seeded at low, medium, and high rates. To test for microsite availability, two factors were examined: supplemental irrigation and disturban ce of existing vegetation. The irrigation treatment had three levels: no irri gation, a pre-seeding soak, and pr eand post-seeding irrigation. The disturbance treatment had four levels, lis ted in increasing level of disturbance: no disturbance, sethoxydim herbicide, glyphosate herbic ide, and topsoil removal. The objectives of the experiment were to determine the relative and interactive effects of seeding rate, supplemental irrigation, and disturbance on C. lanceolata establishment in a P. notatum pasture.

PAGE 31

31 In Chapter 4, I examine both establishment and management methods required for creating self-sustaining populations of native wildflowers on roadsides or former pastures. Specifically, I studied the effects of pre-planting herbicide treatments (establishment) and postplanting mowing frequencies (management) on three wildflower species: C. lanceolata C. leavenworthii, and G. pulchella To examine treatment effects across a range of climate and soil types, the experiment was conducted on simulated road sides in three different regions of Florida. Sites were located in north (Q uincy), north-central (Citra), and south-central (Fort Pierce), Florida. To facilitate evalua tion of effects on wildflower esta blishment as well as subsequent recruitment, the experiment was conducted for two years.

PAGE 32

32 CHAPTER 2 EFFECTS OF PLANTING SEASON, INITIAL ADVANTAGE, AND CUTTING FREQUENCY ON THE COMPETITIVE IN TERAC TIONS BETWEEN A NON-NATIVE PASTURE GRASS AND TW O NATIVE WILDFLOWERS Introduction Non-native species comprise the do minant vegetation of many areas targeted for revegetation and restoration proj ects. While replacing non-native sp ecies with natives is often a goal of these projects, the establishment of native species may be limited by competition from non-native species. Competition between nati ves and non-natives may be influenced by disturbance, species-specific tr aits, initial advantage, season, and site characteristics, among others. Confirming the role of competition be tween non-native and native species in contextspecific situations can direct management activi ties to improve the establishment of desirable native species. In Florida, Paspalum notatum Flgg var. saurae Parodi (bahiagrass, hereafter Paspalum ) occurs on roadsides and on some of the land slated for restoration (Violi 2000). Paspalum is a rhizomatous, mat-forming grass that can persist for decades and prevent native species establishment (Uridel 1994, Violi 2000). In restoration sett ings or roadside wildflower plantings in Florida, native forbs introduced into communities dominated by Paspalum often fail to establish. Disturbing or removing Paspalum can lead to increased establishment (Uridel 1994, Violi 2000). Based on these field observations, competition is hypothesized to be the driving interaction between Paspalum and native forbs, although this has not been experimentally tested. In many restoration projects and roadside pl antings, seeds of nativ e forbs are introduced into Paspalum communities that are generally well-established monotypic stands. Assessing the competitive ability of Paspalum in these settings is confounded with the fact that Paspalum has the initial advantage. The importance of arri val order in determining species composition has

PAGE 33

33 been called historical contingency (Ejrnaes et al. 2006), while the eff ects of arrival order on competitive interactions are referre d to as initial advantage or th e founder effect (Perry et al. 2003, Hely and Roxburgh 2005). Competition between species can be influenced by their order of arrival at a site (Ejrnaes et al. 2006). For example, in competition for light, the first individuals to arrive at a site (or larger individuals) often have a competitive advantage over other individuals (Hely and Roxburgh 2005). Competitive ability between two species can therefore depend on the order of arrival, spec ies identity, or both (Hely and Roxburgh 2005). Initial size seems to affect comp etitive ability most when the competition is aboveground rather than belowground (Gerry and W ilson 1995, Perry et al. 2003). Plant competition has been extensively studi ed in both agricultural (Radosevich and Rousch 1990, Booth et al. 2003) and natural (G race and Tilman 1990, Tilman 1990) systems. Processes governing plant-plant interactions have focused on competition because species that coexist often compete for limite d resources like light, water, and nutrients (Goldberg 1990). Generally, competition between two species is described as symmetric or asymmetric. If competition between two species is symmetric, res ources are shared equally between the species. If competition is asymmetric, one species will have a competitive advantage over a coexisting species. Over time, the species with the comp etitive advantage will likely exclude the coexisting species from the area (Lepik et al. 2005). Asymmetry is far more common than symmetry in competitive interactions because one species is us ually better able to allocate limiting resources than the other species in a pa rticular environment (Ricklef s 1990). Aboveground, asymmetric competition occurs when larger or taller plants with more access to light competitively exclude shorter species through shading (Lepik et al. 2005, Hely and Roxburgh 2005). Belowground competition is usually mediated through limiting resources like nutrients and water (Tilman

PAGE 34

34 1994). Two-species competition experiments have been used to determine whether asymmetric competition occurs between two species, and if so which species will competitively exclude the other (Connolly and Wayne 2005). Competitive interactions may shift under different disturbance frequencies and environmental conditions. Disturbance can disrupt competitive exclusion by reducing asymmetric competition for light (Lep 1999). Ecol ogical theory predicts that an intermediate frequency of disturbance can increase diversity in some habitats (C onnell 1978). Increased species richness and diversity in response to mo derate disturbance are well-documented (Dayton 1971, Lep 1999). Disturbance treatments reduce or remove the existing vegetation and can lead to increased establishment of desirable specie s in a restoration setting (Holzel and Otte 2003, Martin and Wilsey 2006). Many of Floridas native wildflowers naturally o ccur in habitats with frequent disturbance (Taylor 1998) However, the type and freque ncy of disturbance in natural systems are usually different than on managed systems like roadsides and pastures. While roadsides and pastures are subjec t to both natural and artificial disturbance regimes, natural disturbance regimes like fire a nd flooding tend to be prevented or controlled on roadsides for safety reasons. Moreover, management activi ties on roadsides include disturbances such as herbicide applicati on and mowing (Forman et al. 2003). Identifying the conditions that facilitate na tive forb establishment in areas dominated by Paspalum would provide insight into improving re storation projects and native roadside plantings. The objectives of this study were to 1) assess the competitive effects of Paspalum notatum on the growth and reproduction of two native forbs, Coreopsis lanceolata and C. leavenworthii, and 2) determine the effects of initial advantage, disturbance, and season of planting on competitive interactions betw een the species. I hypothesized that Paspalum would

PAGE 35

35 decrease growth of both Coreopsis species but that the co mpetitive ability of the Coreopsis species would increase with the initial advant age and a moderate amount of disturbance. Methods Study Species and Site I exam ined effects of Paspalum competition on two congeners, Coreopsis leavenworthii and C. lanceolata Paspalum was introduced to the United States from Brazil in 1913 (Scott 1920) and was estimated to occupy over 2 million ha in the southeastern U.S. by 1978 (Beatty and Powell 1978). Selected for its drought tolerance and lack of pest problems, Paspalum was originally introduced as a past ure grass but is also commonly planted on roadsides due to its effectiveness in controlling erosion through th e formation of dense mats (Violi 2000). While several cultivars have been developed since its introduction, Pensacola is the most common cultivar in Florida (Violi 2000) and was the cultivar used in this study. For this study, scarified seeds of Paspalum were purchased from a local agricultural supply store. Coreopsis lanceolata L. (lanceleaf tickseed) is an ever green, short-lived pe rennial that is native to much of eastern North America. Ranging from Florida west to New Mexico and north to Ontario, C. lanceolata typically occurs on sandy soils, ditches, roadsides, and disturbed sites and flowers from May through July (Flora of North America Editorial Committee 2006). In north-central Florida, C. lanceolata occurs in sandhills and disturbed sites and blooms in spring (Wunderlin and Hansen 2003). Coreopsis leavenworthii Torr. & A. Gray (Leavenworths tickseed) is an annual to shortlived perennial that is common throughout Florida (Wunderlin and Hansen 2003). A facultative wetland species, C. leavenworthii occurs in depression marshes, disturbed wetland, marl prai rie, pine rockland, wet flatwood, and wet prairie habitats (Gann et al. 2008). Flowers may be produced year-round in sout h Florida, but most flowers are produced in

PAGE 36

36 spring and early to mid-summer (Osori o 2001). Florida ecotype seeds of both Coreopsis species were purchased from Floridas Wildflower S eed and Plant Growers Association, Inc. (2008). The study site was located at the University of Florida in Gainesville, FL, USA (lat 29 38 N, long 82 21 34 W; elevation 21 m). Pots were located on benches outdoors, exposed to natural rainfall and in full sun. Rainfall av erages 123 cm annually and average minimum and maximum temperatures are 6 and 32 C, respectively (Figure 2-1). Experimental Design The experim ent was a factorial (72 treatment combinations) arranged in a completely randomized design and replicated 10 times. The four treatments were season (fall, spring), species (C. leavenworthii C. lanceolata ), cutting frequency (none, moderate, high), and planting treatments (two monocultures and four mixtur es, Figure 2-2). To examine differences in competition based on seasonality, the experiment was planted in fall 2006 and repeated in spring 2007. Fall and spring planting dates were November 9 and March 22 for C. leavenworthii and November 21 and March 28 for C. lanceolata respectively. Planting wa s staggered to allow for timely treatment application and data collection. Because I was primarily interest ed in testing the effects of Paspalum competition on Coreopsis growth, I used a partial a dditive design with a target-nei ghbor approach (Barbour et al. 1987, Gibson et al. 1999, Freckleton and Watkinson 2000). With this approach, the focus of the response variable is the target individual located in the cen ter of the pot, in this case C. leavenworthii or C. lanceolata (Figure 2-2). Plugs of each Coreopsis species were grown alone (monoculture) or in combination with Paspalum (mixture). The low density monoculture treatment consisted of one Coreopsis plug planted in the center of the pot. To test effects of Coreopsis intraspecific competition, a second monoculture was planted with seven Coreopsis propagules (high density). Mixt ures also contained seven propagules but consisted of one

PAGE 37

37 Coreopsis planted in the center of the pot surrounded by six Paspalum propagules (Figure 2-2). The initial advantage treatment was implemented by planting mixtures with plugs of different ages. Plugs that had germinated 40 7 days before planting were considered older; those germinated 10 2 days before seeding were considered younger (Figure 2-2). The species planted with an older plug was considered to have the initial advantage in a mixture if the species neighbors were younge r plug(s) (Figure 2-2). Plugs were produced in plastic germinati on plug trays (27.9 X 54.6 X 5.1 cm, 128 square cells) filled with soilless medium (Fafard #2 Mix, Conrad Fafard, Inc.). One seed (of each species) was placed in the center of a cell. Tray s were kept in a greenhouse in Gainesville, FL (minimum/maximum temperatures/39 C fall, 9/ 33 C spring) and sub-irrigated as needed until saturated. Paspalum plugs for the spring planting were germinated on thermostatically controlled propagation mats (25 C ). Seedlings were marked with different colored toothpicks within 2 days of emergence to accurately track seedling age. To simulate disturbance, shoot biomass ta ller than 10 cm was removed throughout the study twice per month (frequent), once per month (moderate), or not at all (none). Cutting treatments were initiated when th e majority of plants reached a height of greater than 10 cm. This corresponded to 17 and 10 w eeks after planting (WAP) for th e fall and spring treatments, respectively. Cutting treatments continued until 27-28 WAP and were terminated 6-8 weeks before plants were harvested for biomass. Plugs were planted in 26.6 L pots (34 cm di ameter by 29 cm deep) filled with a soilless medium (Fafard #2 Mix, Conrad Fafard, Inc.). Propagules in mixtures and high density monocultures were planted 10.5 cm on center. To focus on effects of disturbance, fertilizer and supplemental irrigation were applie d so that nutrients and water would not be limiting factors to

PAGE 38

38 growth. Pots were fertilized 8-10 WAP with one-half the low label rate of Osmocote 18-6-12 (Scotts Miracle-Gro Company) for 26.6 L pots. Supplemental irrigation was provided as needed through pressure-compensated dribble rings so th at the amount and dist ribution of water among and within pots was uniform. Pots were watered until saturated. Pesticides were used to control disease and insect problems as they arose. Data Collection and Analysis All plants were harvested 33-35 WAP. Target plants were separated from neighbor plants. Aboveand belowground growth were sepa rated at the soil line, washed, and dried at 60 to 70 C for one week. Belowground growth that could not be separate d to species comprised less than 4% of total belowg round biom ass and was therefore not included in the analysis. Coreopsis fitness was assessed by the number of flowers. In the high monoculture treatment, the number of flowers for the target Coreopsis was estimated by counting all flowers in the pot and dividing by the tota l number of plants present. Fl owers of fall plants were counted 29 WAP for both Coreopsis species. I analyzed data using mixed models with restricted maximum likelihood methodology (PROC MIXED, version 9.1; SAS Institute, Cary No rth Carolina, USA). The fixed effects were season, species, planting treatment, cutting freque ncy, and their interactions. Since this was a completely randomized design, there were no random effects in the model. Biomass results of the target Coreopsis (aboveand belowground) were square root transformed and Coreopsis flower counts were log ( x + 1) transformed to meet assumptions of normality and homogeneity of variance. Means were separa ted using least squares means (with PDIFF option) as part of the mixed models analyses; P values were adjusted using the Bonferroni method.

PAGE 39

39 Results Season and Species Season of planting had a pronounced e ffect on survival of the target Coreopsis, which was m uch greater in fall-established plants than spring-established plants for both Coreopsis species (Figure 2-3). Moreover, survival of C. lanceolata was greater than that of C. leavenworthii for both seasons, even though C. leavenworthii biomass was greater than that of C. lanceolata in fall-established plants (Table 2-1, Figure 2-3). Both aboveand belowground biomass of spring-established C. lanceolata plants were at least three times less than those of fall plants (data not presented). However, because the high mortality in spring-established plants resulted in small sample sizes, statistical an alysis of biomass was performed only on fallestablished plants. Initial Advantage and Competition Initial advan tage treatments did not in fluence mean biomass of the target Coreopsis species (P < 0.05 for all treatment combinations within mixtures), so mixtures were combined for presentation. Target Coreopsis biomass was reduced by both neighboring Coreopsis and neighboring Paspalum compared to Coreopsis grown alone (without neighbors) (Table 2-1, Figure 2-4). Neighboring Coreopsis reduced aboveground C. leavenworthii 87% and C. lanceolata 88%, compared to each (target Coreopsis) species planted alone (Figure 2-4). Paspalum reduced aboveground C. leavenworthii 41% and C. lanceolata 58%, less than the reductions caused by intraspecific competition. Coreopsis lanceolata belowground biomass was also reduced, by 81% and 64 % when planted with Coreopsis and Paspalum neighbors, respectively (Figure 2-4). Neighboring Coreopsis and Paspalum reduced belowground C. leavenworthii 87% and 52%, respectively, although th e difference between the treatments was not significant (P = 0.0754, Figure 2-4).

PAGE 40

40 Cutting Frequency Cutting decreased aboveground biom ass of the target Coreopsis by 55% and belowground biomass by 30% (Table 2-1, averaged across Coreopsis species). In both Coreopsis species, aboveground biomass of the target species was greatest with no cutting, while there was no difference between cutting once or twice per month (Figure 2-5). The same trend was observed in belowground biomass of C. lanceolata but C. leavenworthii belowground biomass did not differ among cutting treatments (F igure 2-5). Cutting fre quency did not affect aboveor belowground biomass of neighboring C. lanceolata or C. leavenworthii (high density monoculture treatment, Figure 2-5). Aboveand belowground biomass of neighboring Paspalum was not affected by cutting when grown with C. leavenworthii However, cutting once per month resulted in incr eased aboveground biomass of Paspalum when grown with C. lanceolata compared to no cutting and cutting twice per month (Figure 2-5). Coreopsis Fitness The number of target Coreopsis flowers varied by Coreopsis species, cutting frequency, and planting treatment in fall-establishe d plants (Table 2-2). The number of C. leavenworthii flowers was greater than that of C. lanceolata ( P < 0.0001, Table 2-2, Figure 2-6). Generally, the presence of neighbors as well as cutt ing treatments decreased the number of Coreopsis flowers. However, the effect of competition from neighbors was influenced by cutting treatment and Coreopsis species, which led to significant sp ecies by planting and species by cutting interactions (Table 2-2). Theref ore, analyses were conducted separa tely for each species. In the overall model, there were no effects of initial advantage on flower number so mixture treatments (with Paspalum as the neighboring plant) were combined for presentation. The presence of Paspalum neighbors reduced C. leavenworthii flowers (of the target plant) in fall-established plants ( P = 0.0167, Table 2-2, Figure 2-6); however Coreopsis

PAGE 41

41 neighbors did not reduce target C. leavenworthii flowers compared to plants grown alone (Figure 2-6). From a population perspective, the tota l number of C. leavenworthii flowers in the high density monoculture was greater than in Paspalum mixtures in fall ( P = 0.0011, data not presented). Cutting once or twice per month reduced the number of C. leavenworthii flowers compared to no cutting ( P < 0.0001, Table 2-2, Figure 2-6). For C. lanceolata the presence of both Coreopsis and Paspalum neighbors reduced the number of flowers compared to plants grown alone in fall-established plants (P < 0.0001, Table 2-2, Figure 2-6). However, Coreopsis neighbors reduced target C. lanceolata flowers more than Paspalum neighbors (Figure 2-6). From a population pe rspective, the total number of flowers in the high density monoculture was equal to the number of flowers produced by C. lanceolata plants grown alone ( P = 1.0000, data not presented). Cuttin g twice per month decreased the number of C. lanceolata flowers compared to no cutt ing or cutting once per month ( P < 0.0001, Table 2-2, Figure 2-6). Discussion InterVersus Intraspecific Competition Inter specific competition from Paspalum reduced biomass of both Coreopsis species established in the fall and contributed to high leve ls of mortality in spring-established plants. However, intraspecific competition decreased biomass of the target Coreopsis more than interspecific competition in most cases. Greater biomass in mixture pots could be a result of resource partitioning among roots of the different species. However, root growth and structure were greatly affected by the c ontainer by the end of the study, complicating efforts to assess spatial partitioning of roots between species. Although intraspecific competition was more damaging to the target Coreopsis than interspecific competition, overall Coreopsis fitness (measured by flower number) in the high density monocultures was equal to or greater than

PAGE 42

42 Coreopsis plants grown alone. In many instances, the total number of flowers per pot was similar, regardless of the numb er of propagules present. Although competition from Paspalum reduced Coreopsis biomass, Coreopsis biomass was substantial in fall-established pl ants. On a plant-for-plant basis, C. leavenworthii was equal to or greater than Paspalum biomass. With no disturbance, C. leavenworthii biomass was double that of Paspalum (Figure 2-5). Coreopsis lanceolata biomass was only comparable to Paspalum biomass when plants were not cut (F igure 2-5). Although the presence of Paspalum neighbors generally decreased the nu mber of flowers of both Coreopsis species, C. leavenworthii produced more flowers than C. lanceolata in both seasons (Figure 2-6). While these results suggest that C. leavenworthii is a stronger competitor than C. lanceolata C. leavenworthii survivorship was much lower than that of C. lanceolata especially in spring-establi shed plants. Comparing the competitive strength of two species in a shortterm, two-species competition study may be more complex than evaluating relative biomass (or growth rate) of the two species, as is traditionally done. In this case, the competitive strength of the Coreopsis species depended upon the season of planting and the manner in which competitive strength is evaluated. Seasonal Differences Increased mortality in sp ring-established pl ants may be explained by the response of Paspalum to seasonal changes in environmental c onditions. The majority of the spring experiment occurred during the rainy season, with increased temperatures and day length (Figure 2-1). Paspalum ( notatum ) is a long-day plant th at is dormant during the short days of winter (Marousky and Blondon 1995). Flow ering is initiated when day length exceeds 13.5 hours (Marousky and Blondon 1995), which occurs in May at the study site. In Florida, Paspalum seed maturation begins in June and conti nues throughout the summer (Watson and Burton 1985). Fall-established Coreopsis were planted in November and grew for approximately 26 weeks

PAGE 43

43 before Paspalum began to flower. In contrast, spring-es tablished plants grew for approximately 7 weeks before Paspalum began to flower. Paspalum height can exceed 100 cm when in flower, compared with approximately 20 cm when dorma nt (personal observation). The increase in height during the summer months likely intensified aboveground competition between Paspalum and Coreopsis This increased growth likely had a grea ter effect on spring-established plants because they were younger (and smaller) and expos ed to more intense competition for a longer period of time than fall-established plants. While competition with Paspalum decreased survivorship of Coreopsis in mixtures, there was also decreased survivorship of Coreopsis grown in monocultures in spring-established plants. Coreopsis leavenworthii mortality in response to co mpetition was confounded with mortality due to completion of the C leavenworthii life cycle. Mortality of C. lanceolata in monocultures occurred only in spring-established plants and may have been a result of overwatering. The increased Paspalum growth in mixture pots over the summer required substantial amounts of supplemental irrigation, so that water would not be limiting. However, since Coreopsis plants were generally smaller than Paspalum plants, Coreopsis monocultures did not require as much water as mixture pots. Th e growing medium in th e monoculture pots was consistently wet between watering, while the grow ing medium in mixture pots was dry. Because C. lanceolata is an upland species, the increased moisture in monocultures likely affected C. lanceolata more than C. leavenworthii which is a facultative wetland species. Species-Specific Responses Differences in growth form and life history strategies may account for differences in competitive strength and survivorship among the two Coreopsis species. At a community level, different plant traits or lifehistory strategies may promote species coexistence (Coomes and Grubb 2000, Booth et al. 2003). However, plant surviv al strategies normally involve trade-offs.

PAGE 44

44 For example, plants that are good colonizers are usually not good competitors over time (Coomes and Grubb 2000). Coreopsis leavenworthii is an annual to short-lived perennial while C. lanceolata is an evergreen short-live d perennial. In this study, C. leavenworthii seemed to be more of an annual than short-liv ed perennialflowering, seeding pr olifically, and then dying. In contrast, C. lanceolata produced fewer flowers and biomass than C. leavenworthii but had higher overall survivorship. Coreopsis lanceolata appeared to be able to to lerate the stress of reduced light for longer periods of time than C. leavenworthii by surviving as a sma ll plant underneath the canopy of Paspalum The trade-off between competitive ability and recruitment opportunities is often related to trade-offs between seed si ze and number (Jakobsson and Eriksson 2000). Plants that produce many, sm all seeds increase recruitment opportunities whereas plants that produce few, large seeds are good competitors (Jakobsson and Eriksson 2000). Although the seeds of C. lanceolata seeds are relatively small, their mass is approximately seven times that of C. leavenworthii (based on a sample of seeds weighed before the study). The results from this study suggest that C. leavenworthii is a better colonizer than C. lanceolata while C. lanceolata is a better long-term competitor than C. leavenworthii. Disturbance Cutting generally resulted in decreased a boveground biom ass and flower production of the target Coreopsis for both species (Figures 2-5 and 2-6). However, cutting had little effect on the biomass of neighboring Paspalum or Coreopsis (Figure 2-5). Neighb oring plants may have been able to compensate for loss of height by increasing in width, si nce aboveground growth of neighboring plants was able to extend beyond th e boundaries of the pot. Moreover, cutting may have resulted in the lo ss of apical dominance and promot ed branching. In the case of Paspalum much of its biomass is stored in rhizomes at th e soil surface; these rhizomes were not affected by the cutting treatments. Additio nally, cutting treatments were st opped 6-8 weeks before plants

PAGE 45

45 were harvested. This time period corresponde d with increased growth and flowering in Paspalum in fall-established plants, which may have been able to compensate for the loss of biomass imposed by cutting before the biomass harvest. Contrary to the hypothesis, cutting did not in crease the competitive ability of the target Coreopsis An increase in disturbance can disrup t competitive exclusion by decreasing the number of occupied patches a nd by increasing light (Cordonnier et al. 2006). However, the cutting treatments in this study di d not decrease occupied patches. Additionally, the ability of disturbance to prevent competitive exclusion is based on life-history and growth strategy tradeoffs of the competing species. Since each species has limited resources to allocate to growth and reproduction, there is a trade-off between competitive ability and colonization success and/or stress tolerance (Cordonnier et al. 2006). In a short-term competition study where colonization space is restricted, the longer-term effects of dist urbance and trade-offs may not be evident. Moreover, although cutting may have increased light it simultaneously decreased the biomass of the target Coreopsis Initial Advantage The absence of an initial advantage effect on interspecific com petition in the present study may indicate that light was not a limiting factor during establishment. Consequently, seedling age may not have been a good surrogate for first arrival or earlier germination under field conditions. Although initia l advantage has been shown to increase competitive ability, this increase was only important for seedling establishment (Weigelt et al. 2002). After seedling establishment, species-specific attributes (particularly allocation to biomass) determined the outcome of competitive interact ions (Weigelt et al. 2002). W ilson (1988) found that initial advantage was only relevant in competition for light. In the present study, plugs (seedlings) were not competing for light at the beginning of the study because they were planted in pots

PAGE 46

46 large enough to accommodate eight months of growth. Therefore, although the older seedlings were larger (taller and had mo re leaves) than the younger seedlings, the older seedlings did not shade the younger seedlings. Additionally, one of th e assumptions of the founder effect model is that a species cannot invade a pa tch that is already occupied (C ordonnier et al. 2006). However, in this study, Paspalum frequently occupied additi onal patches by growing over Coreopsis plants. This suggests that competition in this experiment follows a hierarchical competition model, where species can be re placed (Cordonnier et al. 2006). In a hierarchical model, coexistence between species occurs if there is a trade-off betw een competition and colonization. Despite the lack of difference in initial advantage treatments, Paspalum decreased Coreopsis growth and fecundity within one growing season. This suggests that th e lack of native forb establishment in communities dominated by Paspalum in the field is likely influenced by competition. However, it is not clear if initial advantage in the field increases Paspalum competitive advantage. Although results of this study de monstrate competition between Paspalum and both Coreopsis species, applying these results to field c onditions should be done with caution. The present study was conducted in an environment where nutrients and water were not limiting factors, but growth, especially belowground, wa s constricted by pots. In contrast, many of Floridas upland communities, where Paspalum is most often planted, are dry and nutrient poor. Different environmental conditions and nutrient availability can shift or reverse competitive interactions (Barbour et al. 1987). Dry and nutrient poor conditions would likely increase the intensity of belowground competition, which could increase the competitive ability of P. notatum due to its dense root system.

PAGE 47

47 Conclusions Based on results of this experim ent, season of planting and species of Coreopsis greatly influenced competitive interactions with Paspalum notatum Coreopsis lanceolata survivorship was greater than that of C. leavenworthii, especially in spring-estab lished plants. However, biomass of C. leavenworthii was less affected by interspecific competition than C. lanceolata Additionally, C. leavenworthii produced more flowers than C. lanceolata Cutting treatments did not increase species coexistence, at least within the time limits of this experiment. Competitive interactions between Coreopsis and Paspalum were strong regardless of initial advantage (as tested thro ugh seedling age). Intraspecific comp etition was as strong, or stronger, than interspecific competition in terms of biomass, but fitness of the overall population was not reduced by intraspecific competition as it was by interspecific competition. Based on these results, C. lanceolata may result in a more successful longterm establishment in a restoration context than C. leavenworthii. While intraspecific competition limits Coreopsis establishment more than interspecific competition, planting in the fall and limiting disturbance may increase Coreopsis survival.

PAGE 48

48 Table 2-1. Effects of species (Corepsis leavenworthii C. lanceolata ), cutting frequency (none, once/month, twice/month), planting treatment (see Figure 2-2), and their interactions on aboveand belowground biomass of the target (central) Coreopsis in fallestablished plants. Means were squa re root transformed for analysis. Aboveground Belowground Factor df F P F P Species 1, 20210.400.0015 33.74 <0.0001 Cutting 2, 20227.23<0.0001 6.16 0.0025 Planting 5, 20218.76<0.0001 14.08 <0.0001 Species X cutting 2, 2021.390.2523 0.84 0.4348 Species X planting 5, 2020.710.6182 1.08 0.3744 Cutting X planting 10, 2020.480.9022 0.40 0.9464 Species X cutting X planting 10, 2020.800.6281 1.27 0.2470 Table 2-2. Effects of species (Coreopsis leavenworthii C. lanceolata ), cutting treatment (none, once/month, twice/month), planting treatment (see Figure 2-2), and their interactions on flower number of the target (central) Coreopsis in fall-established plants. Flower numbers were log ( x + 1) transformed for analysis. Coreopsis flowers Factor df F P Species 1, 27217.54<0.0001 Cutting 2, 27296.20<0.0001 Planting 5, 2728.45<0.0001 Species X cutting 2, 27219.33<0.0001 Species X planting 5, 2722.450.0341 Cutting X planting 10, 2721.980.0358 Species X cutting X planting 10, 2720.540.8574 Coreopsis leavenworthii Cutting 2, 10053.74<0.0001 Planting 5, 1002.920.0167 Cutting X planting 10, 1001.060.4037 Coreopsis lanceolata Cutting 2, 17228.56<0.0001 Planting 5, 17210.95<0.0001 Cutting X planting 10, 1721.420.1765

PAGE 49

49 Rainfall (cm) 0 2 4 6 8 10 12 14 16 18 Month (2006-2007) NovDecJanFebMarAprMayJunJulAugSepOctNov Temperature (C) -10 0 10 20 30 40 Figure 2-1. Total monthly rainfa ll and minimum and maximum temp eratures at the study site during the experiment. Monthly rainfall fr om The University of Floridas W4DFU Weather Station in Gainesvi lle, FL, approximately 1.5 km from the study site, accessed through Weather Underground (http ://www.weatherunderground.com, ID KFLGAINE10). Temperatures recorded on si te with a HOBO data logger (Onset Computer Corporation, Bourne, MA).

PAGE 50

50 C C C C C C C C C P P P P P P C P P P P P P C P P P P P P C P P P P P P Mixture No advantage Younger Mixture No advantage Older Monoculture High density Younger Monoculture Low density Younger Mixture Coreopsis advantage Mixture Paspalum advantage Figure 2-2. The two monocultures a nd four mixtures included in the planting treatments. C indicates Coreopsis and P indicates Paspalum Shaded circles represent younger plants (10 2 days old); white circles represent older plants (40 7 days old). The

PAGE 51

51 center plant in each pot is the target species ( C. leavenworthii or C. lanceolata ) while the surrounding plants are neighbor species ( Coreopsis or Paspalum ). Planting season Fall SpringCoreopsis survival (%) 0 20 40 60 80 100 C. lanceolata C. leavenworthii Figure 2-3. Effects of planting season (fall or spring) and species (Coreopsis lanceolata or C. leavenworthii) on mean ( 1 SE) Coreopsis survival (target plan t) eight months after planting. Percent survival was averaged across planting and cutting treatments.

PAGE 52

52 Neighbor plant species NoneCoreopsisPaspalumC. lanceolata biomass (g) 0 20 40 60 80 100 120 None a c b x z y NoneCoreopsisPaspalumC. leavenworthii biomass (g) 0 20 40 60 80 100 120 140 160 Aboveground Belowground a c b None x y y Figure 2-4. Effects of neighbor pl ant species (none, conspecific Coreopsis or Paspalum ) on aboveand belowground bioma ss of the target (center) Coreopsis in fall-established plants. Means ( 1 SE) with different le tters within a graph (target species) and aboveand belowground biom ass differ significantly ( P < 0.05, Bonferroni correction). Aboveand belowground biomass of the target Coreopsis were square root transformed for analysis; untransforme d means are presented here. Means were averaged across cutting treatments.

PAGE 53

53 Biomass (g) 0 20 40 60 80 100 120 140 160 NoneTwice OnceNoneTwice OnceNoneTwice OnceC. leavenworthiia b b x x x a a a xx x a a a x x x None, once, or twice = cutting frequency/month Aboveground Belowground Neighbor Coreopsis Neighbor Paspalum Target Coreopsis Biomass (g) 0 20 40 60 80 100 NoneTwice OnceNoneTwice OnceNoneTwice OnceC. lanceolataa xy b b x y y y x b a b a a a x x x Figure 2-5. Effects of cutting frequency on aboveand belowground biomass of the target (center) Coreopsis neighboring Coreopsis and neighboring Paspalum in fallestablished plants. Coreopsis indicates C. leavenworthii in the top graph and C. lanceolata in the bottom graph. Means ( 1 SE) with different letters within a group (target, neighbor Coreopsis or neighbor Paspalum ) and aboveand belowground biomass differ significantly (P < 0.05, Bonferroni correction). Aboveand belowground biomass of the target Coreopsis were square root transformed for analysis; untransformed means are presente d here. Total neighbor biomass per pot was divided by number of neighbors (generally six) to compare with target biomass on a per plant basis. Means for cutting frequencies were averaged across planting treatments.

PAGE 54

54 Neighbor plant species NoneCoreopsisPaspalumLog number of flowers 0 1 2 3 4 5 NoneCoreopsisPaspalumLog number of flowers 0 1 2 3 4 5 None CoreopsisPaspalum Coreopsis leavenworthii b ab a NoneOnceTwice 0 1 2 3 4 5 Coreopsis leavenworthii a b c Cutting frequency/month NoneOnceTwice 0 1 2 3 4 5 Coreopsis lanceolata b a a Coreopsis lanceolata c a b None CoreopsisPaspalum Figure 2-6. Effects of cutting freq uency and planting treatment on th e mean ( 1 SE) number of Coreopsis lanceolata (top graphs) and C. leavenworthii (bottom graphs) flowers. Means with different letters within a cutting or planting treatment differ significantly ( P < 0.05, Bonferroni correction). Nu mber of flowers was log ( x + 1) transformed for analysis.

PAGE 55

55 CHAPTER 3 EVALUATING SEED AND MICROSITE LIMITA TION T O THE ESTABLISHMENT OF A NATIVE WILDFLOWER IN A NON -NATIVE PASTURE GRASS Introduction Ecological restoration and roadsi de revegetation projects ofte n require the introduction of desirable plant species. In Florida, P aspalum notatum Flgg var. saurae Parodi (bahiagrass) occurs on roadsides as well as on some past ures slated for restoration (Violi 2000). Paspalum notatum was originally introduced as a pasture grass but is also commonly planted on roadsides due to its effectiveness in controlling erosi on (Violi 2000). Therefore, establishing native wildflowers (forbs) into P. notatum is a common goal of ecologica l restoration and roadside management projects. However, several barriers exist to establishing native plants in former pastures (Zimmerman et al. 2000). Abandoned ag ricultural areas often have altered hydrology and soil nutrients, which can aff ect plant productivity. The dominant species (usually a grass) may prevent the establishment of new species an d may be difficult to eradicate (Stevenson and Smale 2005). Moreover, native plant propagules to r ecolonize the site may be lacking (Orrock et al. in press ). If the former pasture lacks native plant propagules to reco lonize the site, introducing native plants may be necessary. Establishing nativ e species at the time of pasture removal may also slow or prevent the invasi on of undesirable ruderal and inva sive species. Yet, introducing native plants does not guarantee their establishment. In order for a plant to become established at a new site, the seed (or propa gule) must not only arrive at the site, but also germinate and survive the seedling stage of development. For this reason, plant estab lishment is limited by the availability of seeds, of areas suitable for emer gence and growth, or of both (Stampfli and Zeiter 1999, Turnbull et al. 2000, Zobel et al. 2000, Holzel 2005).

PAGE 56

56 Seed addition experiments test seed li mitation by introducing seeds to a site and determining if the population increa ses as a result. Turnbull et al. (2000) distinguish between two types of seed limitation experiments. Seed augm entation involves adding seeds of species that are already part of the target community, wher eas seed introduction i nvolves adding seeds of non-resident species to the community. In a meta-analysis of seed limitation experiments, Clark et al. (2007) suggest testing the ma gnitude of seed addition effects, rather than simply detecting the occurrence of seed limitation. In a restoration se tting, understanding th e magnitude of seed addition effects could help determine criteria fo r establishment of a self -sustaining population. Direct seeding of propagules in to restoration sites is often the most economical approach to revegetation (Pfaff and Gonter 1999), yet there is limited inform ation on the effects of seeding rate on establishment. Seeding rates are rarely based on parameters that ensure establishment of the restored population (e.g., Cole 20 07) and are more often based on seed availability (Burton et al. 2006). For example, the Florida Department of Transportation uses the same bulk seeding rate, regardless of the wildflow er species being seeded, because species-specific requirements are not yet known (Florida Department of Transportation 1998). However, given differences in viability of seed lots and differences in seed mass by species, the number of viable seeds used per project may vary widely. Seeding at delibe rately high rates is practiced to increase the seeded species competitive ability over ex isting vegetation. In a study on native grass restoration in California, doubling the seed density resulted in twice as many established grass plants (Orrock et al. in press ). While increasing the seeding rate may increase overall emergence, Harkess and Lyons (1998) found that doubling the seeding rate did not affect the number of plants maturing to flower. Moreover, increasing the seeding rate of native legumes negatively affected individual plant esta blishment (Fischbach et al. 2006).

PAGE 57

57 Plant establishment of introduced populat ions not only depends upon the appropriate seeding rate but also upon creating areas suitable for seedling emergence and survival, i.e., microsites, which are also known as safe sites or regeneration niches (Nathan and Muller-Landau 2000). I define microsite in a broad sense to in clude the environmental conditions necessary for germination and emergence as well as the abse nce of competitors, predators, and pathogens (Harper 1977, referred to as saf e site). Microsites suitable for germination and seedling establishment are species-specific, but usually provi de water and contact with the soil (Isselstein et al. 2002). Water availability on degraded sites may be different than on adjacent natural systems due to altered hydrology. Therefore, ge rmination of introduced seeds on degraded sites may be limited by the lack of water at the soil surface. Moreover, seeding may occur in a different season than natural seed dispersal or when natural rainfall is limited. In these cases, supplemental irrigation can increase the seeded sp ecies survival and esta blishment (Harper-Lore and Wilson 2000, Cox et al. 2004). Microsites for new species may be limited by the existing vegetation (Pfaff and Gonter 1999), which can prevent seed-to-soil contact and limit resources like water, light, and nutrients through competition (Tilman 1994). For example, increased productivity in grasslands can prevent seedling establishment of forbs by decreasing light levels (Foster and Gross 1998). Releasing competition by disturbing the existing vegetation often leads to increased seedling establishment (Holzel 2005, Martin and Wilsey 2006). The competitive effects of grasses on herbaceous species can be at least temporarily alleviated by removing the grass via herbicide (del-Val and Crawley 2005). Removing the topso il has also been shown to increase native species establishment, especially in former agricultural areas, by decreasing artificially enhanced soil nutrient levels as well as removing domi nant vegetation (Holzel 2005, Buisson et al. 2006,

PAGE 58

58 Clark and Tilman 2008). Topsoil removal may also remove the majority of weed seeds in the soil seed bank (Roberts 1981, Dalr ymple et al. 2003). In Florida, P. notatum pastures occupy some of the area slated for restoration (Violi 2000). Paspalum notatum is a rhizomatous, mat-forming grass that can persist for decades and prevent native species establishment (Uridel 1994, Violi 2000). Applying glyphosate, a nonselective herbicide, to P. notatum increased the establishment of native species subsequently planted in P. notatum pastures (Uridel 1994). A combinati on of repeated herb icide application and sod removal through disking or pl owing also can effectively control P. notatum ; however, undesirable ruderal or non-nati ve species then typically co lonize the site (Violi 2000). Additionally, it is not clear if combining herbic ide treatments with mechanical sod removal results in greater P. notatum control than mechanical tr eatments alone (Violi 2000). Many species are often limited by both seed and microsite limitation (Stampfli and Zeiter 1999, Juenger and Bergelson 2000). Eriksson (1992) suggests that seed and microsite limitation represent extremes on a spectrum and that a simp le dichotomy between the two does not describe the limitations of most species to establishment. Similarly, Clark et al. (2007) stress the relative importance of seed and microsite limitation to plan t establishment. It is possible that seed and microsite availability can interactively affect establishment, where the availability of one may compensate for the lack of the other. For example, rolling or raki ng the soil after seeding increased establishment by improving seed-tosoil contact (Harkess and Lyons 1998, Gordon et al. 2000). This increase in establishment may al low for a reduction in seeding rate (Harkess and Lyons 1998, Harper-Lore and Wilson 2000, Grabow ski 2005). A better unde rstanding of the relative and interactive effects of seed and microsite limitation may help guide revegetation efforts that are limited by time and resources.

PAGE 59

59 This study focuses on seed and microsite lim itation to the establishment of a native wildflower species in a former pasture. Specifi cally, the objectives of the experiment were to determine the relative and inte ractive effects of supplementa l irrigation, seeding rate, and disturbance on C. lanceolata establishment in a P. notatum pasture. By focusing on the establishment of one species, I was able to test several levels of both seed and microsite limitation simultaneously. I hypothesized that establishment would increase with increasing seed and microsite availability. Methods Study Site and Species The study site was located at the Univers ity of Floridas Institute of Food and Agricultural Sciences P lant Science Research an d Education Unit in Citr a, Florida (lat 29 24 35 N, long 82 08 26 W; elevation 21 meters). Before its conversion to agriculture, the site was likely part of the sandhill ecosystem (M yers 1990), an open savanna dominated by Pinus palustris Mill. (longleaf pine), Quercus laevis Walter (turkey oak), and Aristida stricta Michx. var. beyrichiana (Trin. & Rupr.) D.B.Ward (wiregrass). A diverse assemblage of grasses and forbs generally comprises the groundcover of sa ndhills (Myers 1990). Rainfall averages 132 cm annually and average minimum and maximum te mperatures are 14 and 27 C, respectively (Figure 3-1). The soil is a Candler sand (0-5% slopes) with low nitrogen levels, both in the form of available nitrates and ammonium (Table 3-1). The study area was a P. notatum pasture that had not been grazed or fertilized since 1995. Paspalum notatum was introduced to the United States fr om Brazil in 1913 (Scott 1920) and was estimated to occupy over 2 million ha in the southeastern U.S. by 1978 (Beatty and Powell 1978). While several cultivars have been de veloped since its introduction, Pensacola is commonly used in Florida pastures and along roadsides and the cultivar planted at the study site

PAGE 60

60 (Violi 2000). At the start of the study, P. notatum was the dominant species, with forbs and other graminoids comprising less than 5% of cover. Coreopsis lanceolata L. (lanceleaf tickseed) is an ever green, short-lived pe rennial that is native to much of eastern North America. Ranging from Florida west to New Mexico and north to Ontario, C. lanceolata typically occurs on sandy soils, ditches, roadsides, and disturbed sites and flowers from May through July (Flora of North America Editorial Committee 2006). In north-central Florida, C. lanceolata occurs in sandhills and disturbed sites and blooms in spring (from March through May, Wunderlin and Hansen 2003). Although C. lanceolata occurs in two counties adjacent to the study si te (Wunderlin and Hansen 2008), C. lanceolata was not found at the study site at the start of the experiment. A seed bank study conducted in a nearby field confirmed that C. lanceolata was not present in the seed bank (data not presented). For this study, seeds of C. lanceolata (north Florida ecotype) were purch ased from Floridas Wildflower Seed and Plant Growers Association, Inc. (2008). Seed viability was independently determined using a tetrazolium test (AOSA 2006). Experimental Design This experiment tested the effects of supplemental irrigation, seeding rate, and disturbance on C. lanceolata establishment in a P. notatum pasture. The experiment was arranged in a completely randomized split plot de sign with irrigation as th e main plot treatment and seeding rate and disturbance re gime as factorial sub plots. E ach replicate sub plot was 5 X 5 m with 1-m and 9-m buffers between sub and ma in plots, respectively. The experiment was conducted in two adjacent fields and there were three replicates per field. To test the sensitivity of seedling establishmen t due to seeding rate, plots were seeded at 100 (low), 600 (medium), and 1100 (high) live seeds/m2 with a no-till, hydraulic seed drill on November 20, 2006. The medium seeding rate is recommended by the Florida Department of

PAGE 61

61 Transportation for right-of-way pl antings (pure live seed equivale nt of recommended bulk rate 11.2 kg/ha, Florida Department of Transportation 1998). Microsites were experimentally altered with two factors: irrigati on and disturbance of existing vegetation. The irrigation treatment had three levels: no ir rigation (none), a pre-seeding soak (pre), and preand post-seeding irrigation (full). For the pre-seeding soak, plots were irrigated to approximately 5 cm immediately before seeding. The preand post-seeding irrigation included the same pre-seeding soak plus 2.5 cm irrigation twice/week for 6 weeks following seeding (Figure 3-1). The disturbance treatment had four levels, lis ted in increasing level of disturbance: no disturbance (control), sethoxydim herbicide (s ethoxydim), glyphosate herbicide (glyphosate), and topsoil removal (scraped). To test competitive effects of P. notatum on the seedling stage of C. lanceolata development, sethoxydim (Poast; BASF) a grass herbicide, was applied to actively growing P. notatum 23 and 30 weeks after seeding at 2.6 L/ha (with a surfactant of 83% paraffin base petroleum oil at 2.3 L/ha). Gl yphosate (Roundup Original Max; Monsanto) was applied 4 and 2 weeks prior to seeding at 2.3 L/ha, after which the vegetation was mowed to 5 cm (2 days prior to seeding). Both sethoxydi m and glyphosate were applied with a Cushman Spraytek (Jacobsen Co.) that delivered 281 L/ha at 310 kPa. The scraped treatment consisted of removing the top 13 cm of vegetation and soil one week prior to s eeding. The soil surface was smoothed down with a turf roller a few days before planting. I collected soil samples after implementing the disturbance treatments and just prior to seeding. I haphazardly collected one core (6 cm diameter X 5 cm depth) from each sub plot. Cores were aggregated by disturbance treatment with in each main plot.

PAGE 62

62 Data Collection and Analyses I recorded percent cover and density data 10, 20, 39 and 46 weeks after seeding (W AS). Aboveground biomass was sampled 46 WAS (the end of the study). I cate gorized biomass into C. lanceolata P. notatum forbs (excluding C. lanceolata ), or graminoids (excluding P. notatum ). Percent cover categories included those of biomass as well as standing dead, bare ground, and an estimate of the cover occupied by C. lanceolata flowers. To visually estimate percent cover, I randomly placed three 1 X 1-m subsamples within each plot. Density was counted for C. lanceolata only in 0.25 X 0.25-m subsamples nested within the meter squared subsamples for percent cover. Biom ass was collected from circular 0.15-m2 (diameter 0.44 m) subsamples nested within the meter squared co ver subsamples but not overlapping with the density subsamples. Aboveground vegetation root ed in each biomass subsample was cut at ground level, categorized, and dr ied at 60 C for one week. I analyzed data using mixed models w ith restricted maximum likelihood methodology (PROC MIXED, version 9.1; SAS Institute, Cary No rth Carolina, USA). The fixed effects were seeding rate (low, medium, high) irrigation (none, pre, full), di sturbance (control, sethoxydim, glyphosate, scraped), and the in teractions among them. The ra ndom effects were block (for field), main plots within blocks and main plot by sub plot with in blocks. Degrees of freedom were approximated with Kenward-Rogers method. Response variables included density, percent cover, and biomass as explained above. To identify changes in percent cover and density over time, a repeated measures analysis was a dded to the model with an exponential spatial covariance structure (SP(EXP)). Means were se parated using least squa res means (with PDIFF option) as part of the mixed models analyses; P values were adjusted using the Bonferroni method. Percent cover and dens ity data were averaged by subpl ot for the repeated measures analysis and biomass data were averaged by subpl ot and converted to a per meter squared basis

PAGE 63

63 before analysis. Density data were log ( x + 1) transformed, percen t cover data were arcsine square root transformed and biomass of P. notatum forbs, and graminoids were square root transformed to meet assumptions of normality and homogeneity of variance. Results Coreopsis lanceolata Changes Over Time Increasing seeds and m icrosites resulted in greater C. lanceolata establishment. However, the effects of irrigati on, seeding rate, and disturbance on C. lanceolata density and percent cover changed over time (Figure 3-2). In general, density decreased while cover increased; however, the extent of these tem poral changes depended on the treatment (Figure 32). Between 10 and 46 WAS, density decrease d 50% at the high seed ing rate and 39% in glyphosate treated plots but did not d ecrease at the low seeding rate or in scraped plots. Percent cover increased eight-fold in the glyphosate treate d plots compared to a five-fold increase in the control and scraped plots. Different temporal responses by treatment re sulted in significant interactions between WAS and each of the main effects so main eff ects were analyzed separately within each time period (Table 3-2). There were a number of significant interactions among seeding rate, irrigation, and disturbance treatments for both de nsity and percent cover 10 WAS (Table 3-2). By 46 WAS, each of the main e ffects significantly affected C. lanceolata density and cover but there were no significant interactions. Theref ore, means of density and cover over time are presented separately for each main effect (Figure 3-2). At 10 WAS, density and cove r differences among irrigation treatments varied by seeding rate (Table 3-2). Density and c over were twice as high in the full irrigation treatment as in the noand pre-irrigation treatments but only at the medium and hi gh seeding rates. At the low

PAGE 64

64 seeding rate, density and cove r were equally low among irriga tion treatments, resulting in significant rate by irrigation interactions ( P = 0.0116 for density, P < 0.0001 for cover, Table 32). Similarly, density in the scraped plots wa s lower than in the ot her three disturbance treatments at the medium and high seeding rates. However, at the low seeding rate, density was equally low among disturbance treatments, resulting in significant rate by disturbance interaction ( P = 0.0023, Table 3-2). The full irrigation treatment increased percent cover compared to the no-and preirrigation treatments (Figure 3-2). However, this increase was more pronounced in the glyphosate treated plots than in the other di sturbance treatments l eading to a significant irrigation by disturbance interaction ( P = 0.0004, Table 3-2). Coreopsis lanceolata Es tablishment Despite significant interac tions among main effects at the beginning of the study, disturbance treatments resulted in the largest differences in C. lanceolata cover and biomass by the end of the study. While percent cover was gr eatest in the glyphos ate treated plots throughout the study, by 46 WAS, cover in glyphosate treated plots wa s more than double that of the control, sethoxydim, and scraped plots (Figur e 3-2). Effects on biomass mirrored those of cover; biomass in glyphosate tr eated plots was 7, 6, and 2.5 time s greater than in control, sethoxydim, and scraped plots, respectively (Fig ure 3-3). At 46 WAS, density in glyphosate treated plots was 17% greater than in control plots but there was no difference in density between glyphosate and control plots at 10 and 20 WAS (Figure 3-2). The scraped treatment affected C. lanceolata density, biomass, and cover differently. Density in the scraped plots remained low thr oughout the study. In contrast, biomass (at 46 WAS) in the scraped plots was more than double that of the control and sethoxydim treated plots (Figure 3-3). Cover in the scraped plots was greater than the control 20 and 39 WAS. However,

PAGE 65

65 by 46 WAS, cover in the scraped plots was equa l to the control and lo wer than the sethoxydimtreated plots (Figure 3-2). In general, the sethoxydim treatment did not increase C. lanceolata establishment. While the sethoxydim treatment increased cover slightly compared to th e control at 39 WAS, there was no difference by 46 WAS (Figure 3-2). Density in the sethoxydim treatment remained equal to that of the control throughout the study (Figure 3-2). At 46 WAS, there was no difference in biomass between the sethoxydim and cont rol treatments (Figure 3-3). Coreopsis lanceolata density and cover increased with each increase in seeding rate (Figure 3-2). The difference between the low an d medium seeding rates was larger than the difference between the medium and high seeding rates. For cover, this trend became more apparent over time. Effects of seeding rate on biomass varied by disturbance treatment, resulting in a significant rate by disturbance interaction (Figure 3-3; P = 0.024). In glyphosate treated plots, biomass was greater at the mid and high seeding rates compared to the low rate. However, seeding rate did not result in biomass differences among the other three disturbance treatments (Figure 3-3). The full irrigation treatment resulted in the highest density, with no difference between the noand preirrigation trea tments (Figure 3-2). This difference was consistent throughout the study, despite the decline in density over time. Pe rcent cover was greatest w ith the full irrigation treatment at 10 and 46 WAS but there was no difference among irrigation treatments at 20 and 39 WAS (Figure 3-2). Irrigati on treatments did not affect C. lanceolata biomass at 46 WAS (data not presented). Coreopsis lanceolata Fitness The percent cover of C. lanceolata flowers was recorded 20 WAS, which was estim ated to be peak bloom in glyphosate treated plots a nd in a nearby planted population. Percent cover

PAGE 66

66 of flowers ranged from 2.8% in the low seedi ng rate to 4.2% in the high seeding rate in glyphosate treated plots (Figure 3-4). In the scra ped plots, flower cover ranged from 0.8% in the low seeding rate to 3.1% in the high seedi ng rate (Figure 3-4). Percent cover of C. lanceolata flowers was less than 0.5% in th e control and sethoxydim treatment s regardless of seeding rate (Figure 3-4). Between 20 and 39 WAS, the C. lanceolata population consisted of adult plants originating from the fall seeding. However, be tween 39 and 46 WAS, I observed recruitment (emerged seedlings). At 46 WAS, the number of seedlings (per 0.25 X 0.25-m quadrat) averaged 2.5 in scraped plots but less than 1 in control, set hoxydim, and glyphosate treatments (Figure 3-5). Seedlings likely represent new re cruitment from the planted population because 1) no seeds were observed on the soil surface 10-20 WAS and 2) numerous seeds were visible on the soil surface following seed maturation and dispersal from the adult plants. However, it is possible that some of the seedlings originated from dormant seeds from the fall seeding. Community Vegetation The percent cover of P. notatum forbs, and gram inoids changed significantly over the experiment. However, these changes over time varied by treatment factor, resulting in significant interactions between main effects and WAS (data not presented) Therefore, percent cover results were analyzed separately within each time period (WAS). Paspalum notatum cover in glyphosate and scraped treatments remained significantly lower than control and sethoxydim treatme nts throughout the study. At 10 WAS, P. notatum cover was less than 1% in glyphos ate and scraped treatments comp ared to 79% in control plots (Table 3-3). By week 46, P. notatum cover was less than 4% in glyphosate and scraped treatments compared with greater than 81% in control and 75% in sethoxydim treatments, respectively (Table 3-3). Similarly, P. notatum biomass was more than 20 times greater in the

PAGE 67

67 control and sethoxydim treatments as compared to the glyphosate and scraped plots (Figure 3-3). While the sethoxydim treatment significantly reduced P. notatum cover by approximately 7% compared to the control 46 WAS, this reduction wa s slight when compared to the effects of the glyphosate and scraped treatments on P. notatum cover (Table 3-3). Increases in C. lanceolata seeding rate decreased P. notatum cover and biomass 46 WAS (Table 3-3). The decrease in biomass was only significant in the control and glyphosate treatmen ts which led to a significant rate by disturbance interaction ( P < 0.05; Figure 3-3). Paspalum notatum cover was lower with the full irrigation treatment compared to no irrigation 46 WAS (Table 3-3). Disturbance treatments resulted in the greate st differences in forb cover. At 10 WAS, forb cover was greater in glyphosat e treated plots than control a nd scraped plots (Table 3-3). However, by 46 WAS, forb cover was 58% in scra ped plots, much greater than in all other disturbance treatments. The low seeding rate resu lted in greater forb cove r than the high rate at 10 WAS (Table 3-3). This trend continued 46 WAS, but only in the glyphosate treatment. There was no difference in forb cover or biomass among seeding rates in the control, sethoxydim, and scraped treatments (Figure 3-3), resulting in significant rate by disturbance interactions for both cover and biomass of forbs 46 WAS ( P < 0.01, for both). In the scra ped plots, forb cover and biomass were lowest with the full irrigation treatment at 46 WAS. However, there was no difference in forb cover among irrigation treatme nts in the control, sethoxydim, and glyphosate treatments, resulting in a significant irri gation by disturbance in teraction 46 WAS ( P < 0.0001 for cover, P < 0.05 for biomass). Graminoids, excluding P. notatum occupied a small percentage of the community vegetation (Table 3-3). The scraped treatmen t resulted in the greatest graminoid cover and biomass compared to the control, sethoxydim, and glyphosate treatments (Table 3-3, Figure 3-3,

PAGE 68

68 P < 0.0001 for cover at 10 WAS and cove r and biomass at 46 WAS). In the scraped treatment, graminoid cover and biomass were greatest with th e full irrigation treatment compared to the noand pre-irrigation treatments (T able 3-3). However, there we re no differences among irrigation treatments in the control, sethoxydim, and scraped treatments, resulting in significant irrigation by disturbance interactions fo r cover at 10 and 46 WAS ( P < 0.001 and P < 0.0001, respectively). Averaged over disturbance and irrigation treatments, graminoid biomass was greater in the low seeding rate than the medium and high rates ( P < 0.01; Figure 3-3). Soil Characteristics Re moving the topsoil resulted in decreased pH and organic matter and increased potassium as compared to the control, sethoxyd im, and glyphosate treatments (Table 3-1). There was no difference in nitrogen (as available nitr ates and ammonium) or phosphorus levels among disturbance treatments (Table 3-1). Discussion Interactive Effects of Seed and Microsite Availability Seed and m icrosite availability interactively affected C. lanceolata emergence and seedling growth as observed 10 WAS. While in creases in seeding rate generally increased C. lanceolata establishment, differences in microsite av ailability (i.e., dist urbance and irrigation treatments) were often apparent only at the medium and high seed ing rates. Hence, microsite limitation may only become biologically relevant when a minimum number of seeds is present. Although seed and microsite limita tions interactively affected C. lanceolata establishment at the beginning of the study, the relative importance of microsite limitation was greater than that of seed limita tion by the end of the study. When the extant vegetation was not disturbed, seeding at the high ra te (essentially eliminating seed limitation) did not increase C. lanceolata biomass as compared to the medium or low rates (Figure 3-3) This study provides

PAGE 69

69 further support that increasing seed availability may not necessarily lead to increased recruitment if microsites are not avai lable (Holl et al. 2000). Disturbance Effects on Microsite Quality C. lanceolata estab lishment was limited by th e extant pasture vegetation ( P. notatum ), and not surprisingly, distur bance treatments increased C. lanceolata establishment compared to the control. However, the three disturbance treatments affected C. lanceolata establishment differently. Treating P. notatum pastures with glyphosate resulted in the greatest C. lanceolata establishment, regardless of seeding rate or irrigation treatment (Figure 3-2). Additionally, percent cover of flowers wa s greater in plots where P. notatum had been removed compared to plots where P. notatum was present (Figure 3-4) In other studies, P. notatum was a strong competitor (Rich et al. 2003) and limited estab lishment of native forbs (Uridel 1994, Violi 2000). Contrary to my hypothesis, the scraped plots did not result in the greatest C. lanceolata establishment. Both the glyphosate and scraped disturbance treatments reduced P. notatum biomass; however, this decrease alone did not result in an equi valent increase in wildflower cover or biomass (Figures 3-2 and 3-3). Most studies on sessile orga nisms equate microsites with available space for growth (Erik sson 2005). However, the differences in C. lanceolata establishment between the glyphosate and scraped tr eatments suggest that microsite requirements are more complex than providing adequate space and likely include abiotic as well as biotic components (Tilman 1994). The difference in C. lanceolata establishment between glyphosate and scraped treatments was largely due to lower emergence in the scra ped plots. Removing the topsoil resulted in greater than 90% cover of bare ground, compared with 5% in the glyphosate treatment. While seed-to-soil contact generally increases establishment, multiple, small gaps result in greater establishment than fewer, larg e gaps (Burke and Grime 1996). Topsoil removal also likely

PAGE 70

70 exposed native sandhill soils, which are deep, we ll-drained deposits of sand. Sandhill soils generally do not have horizon layers; organic matter, if present, is superficial and associated with existing vegetation (Myers 1990). The topsoil remova l resulted in a harsh environment with little substrate stability and reduced organic matter. In fact, the sandhill environment has been likened to a desert for establishing seedlings (Myers 1990) In arid environments, seedling survivorship has been shown to be greater under shrub c over than in open areas (De Jong and Klinkhamer 1988). In a study on pasture restoration in Puerto Rico, removal of pasture vegetation resulted in decreased germination of four species and incr eased soil surface temperat ure (Zimmerman et al. 2000). While topsoil removal may be an effective way to remove P. notatum glyphosate was as equally effective in removing P. notatum and much more effective in establishing C. lanceolata The glyphosate treatment effectively killed P. notatum but left a layer of litter, comprised of dead thatch and P. notatum stems. The litter present in gl yphosate treated plots seemed to provide better microsites for germination and em ergence than the bare sand present in scraped plots. Because of the litter layer, microsite characteristics in glyphosat e treated plots may not have been different than control plots. Soil characteristics of glyphos ate treated plots were equivalent to control plots, and both had higher organic matter than in scraped plots (Table 3-1). Additionally, there was no differe nce in percent cover of litter or of bare ground between the glyphosate and control treatments (data not pr esented). Existing vegetation can actually facilitate germination in some species by bufferi ng relative humidity and temperature on the soil surface (Holmgren et al. 1997, Kennedy and Sousa 2006). The litter left by the glyphosate treatment may also have ameliorated the low moisture and nutrient levels characteristic of sandhills (Foster and Gross 1998).

PAGE 71

71 Reducing P. notatum through a grass herbicide treatment at the rate applied did not increase C. lanceolata establishment. While sethoxydim decreased P. notatum cover by approximately 7%, the reducti on resulted in only a small a nd short-lived increase in C. lanceolata cover. Despite the small reduction in cover, P. notatum was still the dominant species in the sethoxydim treatment. Most grass herb icides, including sethoxydim, work best when applied during active grow th of the target species. In this study, P. notatum was not actively growing until approximately 20 WAS, at which point C. lanceolata was already flowering in glyphosate plots. Existing vegetation may serve as an effective nurse cr op, helping to control undesirable ruderals while desi rable natives become established (Ewel and Putz 2004). Although P. notatum suppressed the growth of unwanted rudera ls in this case, it also suppressed the growth and flowering of the desirable native species. Reducing P. notatum cover after C. lanceolata was established was not as effective as removing P. notatum prior to seeding. By the end of the study, supplemental irriga tion did not result in large increases in C. lanceolata cover or biomass, but this result may be confounded by the particularly wet winter during which this study was conducte d (El Nio year with high rainfa ll). Effects of irrigation on cover were apparent at 10 a nd 46 WAS but not between 10 and 46 WAS (Figure 3-3). Perhaps differences between the full irrigation treatm ent compared to the preand no-irrigation treatments were too small to be detected. At 46 WAS, differences in cover may have been more apparent than at 20 and 39 WAS due to incr eased rain following a drought (Figure 3-2). Coreopsis lanceolata has thick, leathery leaves that curl in response to dr ought, resulting in decreased cover. Plants may have recovered by 46 but not 39 WAS. Although the full irrigation treatment did not result in large increases in C. lanceolata cover compared to the no-and pre-

PAGE 72

72 irrigation treatments, C. lanceolata density was consistently higher in the full irrigation treatment than the other two irrigation treatments. Requirements for Emergence Versus Establishment Different responses of C. lanceola ta density and cover by treatment and over time indicate different requirements for seedli ng emergence versus subsequent growth. Coreopsis lanceolata emergence occurred within 10 weeks of s eeding, from which point density decreased over time (Figure 3-2). Howe ver, percent cover increased throughout the experiment. The decrease in density coupled with the increase in cover indicates that increases in stand establishment through time resulted from emerge d seedlings increasing in size rather than number. While growth of C. lanceolata was limited by P. notatum emergence was not. At 10 WAS, there were no differences in density be tween the control and glyphosate treatments; however, cover was greater in gl yphosate than control treatments (Figure 3-2). These results indicate that C. lanceolata was able to germinate in P. notatum but the resulting seedlings were smaller than those in the glyphos ate treated plots. Plants are often more sensitive to environmental conditions and competition during the seedling stage than during germination or as adults (Isselstein et al. 2002, Holzel and Ot te 2003). A study focusing on the establishment of annuals in an arid environment found that plant-pl ant interactions shifted from facilitation in the seedling stage to competition during the reprodu ctive stage (Schiffers and Tielborger 2006). In this study, P. notatum had a neutral effect on C. lanceolata emergence but a competitive effect on seedling and adult growth. While adding seeds and supplemental irriga tion did not compensate for the lack of microsites created by the glyphosate treatmen t by the end of the study, the significant interactions among treatments 10 WAS suggest a greater interplay between seed and microsite limitation early in C. lanceolata establishment. Changes to micr osite availability have been

PAGE 73

73 shown to affect different grow th stages of the same plants differently. The present study supports these results, but also suggests a mo re complex relationship. The differences in C. lanceolata establishment between the glyphosate and scraped treatments and between the full and noor pre-irrigation treatments suggest that microsites repr esent a complex interaction of factors, not just the availability of space, water, or nutrients. Effects on Community Vegetation Treatm ents that improved wildflower establ ishment affected other vegetation in the community positively and negatively. For example, increasing the seeding rate not only resulted in increased C. lanceolata cover, but also decreased cover of P. notatum forbs, and graminoids. The full irrigation treatment increased C. lanceolata cover and density but also increased graminoids and decreased P. notatum Since P. notatum is dormant during winter, the decrease in P. notatum due to irrigation appears to be a resu lt of increased wildflower density. Removing the topsoil did not prevent colonizatio n of weedy forbs in the scraped plots. Although not tested for this site, several studies ha ve demonstrated that the majority of the seed bank is found in the top 5 cm of soil (Roberts 19 81). In this study, the top 13 cm was removed but forb establishment was still high, especially 46 WAS. Although some forbs were native species representative of sandhill communities like Asimina Adans sp. (paw paw) and Asclepias L. sp. (milkweed), the majority of forbs were agricultural weeds like Rumex L. sp. (dock) and Indigofera hirsuta L. (roughhairy indigo). Most of the weedy forbs that colonized the scraped plots occurred in nearby agricu ltural fields. Removing topsoil creates large patches of bare ground where colonizing plants can quickly beco me established. In agricultural and heavilyfragmented areas, propagule pressure from undesira ble ruderal or invasive species can be high. Therefore, topsoil removal in these settings may not be as beneficial as in less disturbed settings, where recruitment from desirable natives is more likely.

PAGE 74

74 Graminoids (other than P. notatum ) comprised a small percentage of the community vegetation and were mostly represented by common sedges ( Cyperus spp.) This indicates that native grasses, which are an essential component of the native understory, were largely absent from the study site. These resu lts suggest that while introducing a single forb species may be a short term solution in converting pasture grass to native habitat, seed introductions of a suite of native grasses and forbs would be necessary to de velop this site into a species rich native community. Conclusions W hile seed and microsite limitations interactively affected C. lanceolata establishment at the beginning of the study, the relative importance of microsite limitation was greater than that of seed limitation by the end of the study. Alleviation of microsite limita tion depended strongly on the quality of microsite. Speci fically, disturbing the extant ve getation with glyphosate herbicide prior to seeding resulted in greater establis hment than any other disturbance treatment. Removing the topsoil increased C. lanceolata somewhat compared to the control but also increased colonization by weedy forbs. By the end of the study, C. lanceolata establishment was not affected by supplemental irrigation. Coreopsis lanceolata establishment was limited when seeded at 100 live seeds/m2 but not at 600 live seeds/m2. Seeding at 1100 live seeds/m2 provided little additional benefits to overall C. lanceolata establishment compared to seeding at 600 live seeds/m2. Different treatment responses between C. lanceolata density and percen t cover, at the beginning and end of the study, suggest that re quirements for seedling emergence are less stringent than those for plant establishment.

PAGE 75

75 Soil characteristics Disturbance treatment pH Organic matter (%) NO3-N (mgkg-1) NH4-N (mgkg-1) P (mgkg-1) K (mgkg-1) Control 5.89a 2.02a <1.00 0.95 31.62 37.12a Sethoxydim 5.91a 2.01a <1.00 0.96 30.08 39.53a Glyphosate 5.88a 1.88a <1.00 0.95 31.21 43.44a Scraped 5.62b 1.37b <0.10 1.00 30.17 46.65b Means with different letters within soil characteristics differ significantly as determined by ANOVA ( P <0.0001, Bonferroni correction); no means comparison was performed if a soil characteristic was not significant. Analysis was not performed on NO3-N because values were too low to be detected with greater accuracy. Table 3-1. Disturbance treatment effects (con trol, sethoxydim, glyphosate, scraped) on soil chemistry before seeding Coreopsis lanceolata in a former pasture in Citra, FL.

PAGE 76

76 Table 3-2. Effects of irrigation (none, pre, full) seeding rate (low, me dium, high), disturbance (control, sethoxydim, glyphosate, scraped), and time (weeks after seeding) on density and percent cover of Coreopsis lanceolata in a former pasture in Citra, FL. Analysis was performed on log (x +1) transformed density data a nd arcsine square root percent cover data. Model Density Percent cover Factor df F P F P Repeated measures model Irrigation 2, 1426.47<0.00018.89 0.0032 Seeding rate 2, 165244.75<0.0001142.07 <0.0001 Disturbance 3, 16528.69<0.0001153.08 <0.0001 Weeks after seeding 3, 54043.90<0.0001960.09 <0.0001 Irrigation X rate 4, 1652.690.03300.90 0.4650 Irrigation X disturbance 6, 1651.040.39820.79 0.5802 Rate X disturbance 6, 1652.470.02560.48 0.8225 Irrigation X rate X disturbance 12, 1650.680.76930.29 0.9909 Irrigation X weeks 6, 5406.60<0.00013.16 0.0047 Rate X weeks 6, 54020.7<0.000124.36 <0.0001 Disturbance X weeks 9, 5407.38<0.000169.66 <0.0001 Irrigation X rate X weeks 12, 5401.800.04530.73 0.7184 Irrigation X disturbance X weeks 18, 5400.640.86960.91 0.5706 Rate X disturbance X weeks 18, 5401.950.01101.57 0.0624 Irrigation X rate X disturbance X weeks 36, 5400.680.92260.77 0.8310 Ten weeks after seeding Irrigation 2, 1421.47<0.000148.53 <0.0001 Seeding rate 2, 165130.94<0.0001152.16 <0.0001 Disturbance 3, 16529.43<0.000135.55 <0.0001 Irrigation X rate 4, 1653.340.01166.51 <0.0001 Irrigation X disturbance 6, 1650.860.52634.42 0.0004 Rate X disturbance 6, 1653.580.00231.79 0.1050 Irrigation X rate X disturbance 12, 1650.490.92081.22 0.2705 Forty six weeks after seeding Irrigation 2, 1414.15<0.00015.62 0.0161 Seeding rate 2, 165134.03<0.000183.51 <0.0001 Disturbance 3, 16526.81<0.0001118.46 <0.0001 Irrigation X rate 4, 1652.300.06080.59 0.6725 Irrigation X disturbance 6, 1650.690.65430.91 0.4888 Rate X disturbance 6, 1651.150.33490.80 0.5738 Irrigation X rate X disturbance 12, 1650.620.82610.49 0.9167

PAGE 77

77Table 3-3. Effects of irrigation (none, pre, full), seeding rate (low, medium, high), and disturbanc e (control, sethoxydim, gly phosate, scraped) on percent cover of Paspalum notatum forbs, and graminoids 10 and 46 weeks after seeding Coreopsis lanceolata in a former pasture in Citra, FL. Factor Vegetation category level Paspalum notatum Forbs Graminoids Weeks after seeding Weeks after seeding Weeks after seeding 10 46 10 46 10 46 Irrigation ns ** ns ns ns None 38.8 42.8 a 3.2 19.9 0.5 1.5 ab Pre 39.9 40.9 ab 3.7 20.3 0.6 1.2 a Full 40.7 38.9 b 4.3 15.9 0.9 2.2 b Rate (live seeds/m 2 ) ns *** ** *** ns + 100 40.1 45.7 a 4.7 a 24.6 a 0.8 1.8 600 39.3 40.1 b 3.7 ab 15.8 b 0.7 1.7 1100 39.9 36.8 c 2.8 b 15.8 b 0.6 1.3 Disturbance *** *** *** *** *** *** Control 79.0 a 81.6 a 2.3 a 2.5 a 0.6 a 1.6 a Sethoxydim 79.4 a 75.1 b 2.0 a 3.9 a 0.7 a 1.7 a Glyphosate 0.5 b 3.4 c 7.3 b 11.0 b 0.3 a 1.1 a Scraped 0.3 b 3.5 c 3.3 a 57.5 c 1.1 b 2.1 b Significance of F-test Irrigation*rate ns ns ns ns ns ns Irrigation*disturbance ** ns ns *** ** *** Rate*disturbance ns ns ns ** ns ns Irrigation*rate*disturbance ns ns + ns ns ns Significant main effects and interactions are indicated by: *** ( P < 0.0001), ** ( P < 0.01), ( P < 0.05). Marginally significant effects (0.10 > P > 0.05) are indicated by +. Means with different letters within vegetation categories are significantly different ( P < 0.05, Bonferroni correction). Analysis was pe rformed on arcsine square root transformed data; untransformed means are presented here.

PAGE 78

78 Weeks after seeding (month) 01020304050Precipitation (cm) 0 1 2 5 Rainfall Irrigation (Nov)(Jan) (Apr) (Jun)(Aug)(Nov) 'Pre' irrigation 'Full' irrigation 01020304050Solar radiation (w/m 2 ) 40 60 80 100 120 140 160 180 200 220 240 260 2 nd sethoxydim application 1 st sethoxydim application 01020304050Temperature (C) 0 5 10 15 20 25 30 35 40 Figure 3-1. Minimum and maximum temperatures at 60 cm, solar radiation, and precipitation at the study site in north-central Florida from November 2006 to November 2007. Data from FAWN, Florida Automated Weathe r Network (http://fawn.ifas.ufl.edu).

PAGE 79

79 Sethoxydim Sethoxydim Density/m 2 20 40 60 80 100 120 140 160 Percent cover/m 2 0 20 40 60 80 Percent cover/m 2 0 20 40 60 80 Density/m 2 0 20 40 60 80 100 120 140 160 Weeks after seeding 01020304050Density/m 2 20 40 60 80 100 120 140 160 Weeks after seeding 01020304050Percent cover/m 2 0 20 40 60 80 Cover DensityF E D C Disturbance treatment None Sethoxydim Glyphosate Scraped Irrigation treatmentNone Pre Full A B a b b a b b a a a a b ab a b b a b b a b b a b b a c b a c b a c b a c b a c b a c b a c b a c b b a b b c a c b c a b b c a bc b b a a a b a a a c a bc b c a b b 100 600 1100 Seeding rate: live seeds/m 2 Figure 3-2. Effects of irrigation (graphs A and B), seeding rate (graphs C and D), and disturbance (graphs E and F) treatments on mean ( SE) Coreopsis lanceolata density (individuals/m2; graphs A, C, E) and percent cover (graphs B, D, F). Different letters within each column indicate significant differences ( P < 0.05; Bonferroni correction). Analysis was performed on log ( x + 1) transformed density data and arcsine square root cover data; untransfor med means are presented here. Means are presented separately for each main effect because th ere were no significan t interactions by the end of the study (see Table 3-2).

PAGE 80

80 Disturbance treatment Co n t r o l S e t h o x y d i m G l y p h o s a t e S c r a p e d 0 2 4 6 8 10 Disturbance treatment Co n t r o l S e t h o x y d i m G l y p h o s a t e S c r a p e d Shoot biomass (grams/m 2 ) 0 100 200 Shoot biomass (grams/m 2 ) 0 100 200 300 0 100 200 300 a ab b ab b b c de e de de cd Graminoids Paspalum notatum Coreopsis lanceolata Forbs 100 1100 Coreopsis lanceolata seeding rate Number of live seeds/m 2 600 ef def cde cd c e e b a a ef def c c c c c c b c c a ab ab b ab ab ab ab ab ab ab ab ab a a Figure 3-3. Effects of seeding ra te and disturbance treatment on mean (+ SE) shoot biomass of Coreopsis lanceolata Paspalum notatum forbs, and graminoids 46 weeks after seeding. Means with different letters within each specie s/guild indicate significant differences ( P < 0.05; Bonferroni correction). Analysis of all response variables except Coreopsis lanceolata was performed on square root transformed data; un transformed means are presented here. Note differences in y-axis scale for forbs and graminoids.

PAGE 81

81 Disturbance treatment ControlSethoxydimGlyphosateScrapedPercent cover/m2 0 1 2 3 4 5 100 1100 Coreopsis lanceolata seeding rate Number of live seeds/m 2 600 Figure 3-4. Effects of seeding ra te and disturbance treatment on mean (+ SE) percent cover of Coreopsis lanceolata flowers 20 weeks after seedi ng. Means were not compared statistically due to the high proportion of zero values. Disturbance treatment ControlSethoxydimGlyphosateScrapedSeedling density/0.25 x 0.25 m 0 1 2 3 4 5 6 7 100 1100 Coreopsis lanceolata seeding rate Number of live seeds/m 2 600 Figure 3-5. Effects of seeding rate and disturbance treatment on mean (+ SE) number of Coreopsis lanceolata seedlings 46 weeks after seeding. Means were not compared statistically due to the high proportion of zero values.

PAGE 82

82 CHAPTER 4 MOWING FREQUENCY AND HERBICIDE EFFECT S ON ESTABLISHING NATIVE WILDFLOWER POPULATIONS ON SIMULATED ROADSIDES Introduction Roads have becom e a ubiquitous part of our landscape, comprising more than 4 million ha of land in the United Stat es (Forman and Alexander 1998, Harper-Lore and Wilson 2000). Roads cause habitat fragmentation (Heilman et al. 2002) and provide corr idors for invasive, nonnative species to expand their range (Gelbard and Belnap 20 03, Hansen and Clevenger 2005). However, since the 1970s, there has been more of an ecological approach to roadside management, with increased efforts to increase biodiversity and manage for wildlife (HarperLore and Wilson 2000). Increasing native wildflow er populations on roadsides can decrease the occurrence of non-native, invasive species as well as improve aesth etics (Florida Department of Transportation 1998, Harper-Lore and Wilson 2000). Moreover, roadside maintenance costs can be reduced when wildflower populations are ma naged appropriately (Mar kwardt 2005). Hence, roadside vegetation managers are becoming increasingly interested in planting native wildflowers along roadsides (Harper-Lore and Wilson 2000). Roadsides are often planted with grass monocu ltures that effectivel y control erosion and can tolerate frequent mowing (For man et al. 2003). However, the establishment of new species into established grass monocultu res may be limited by competition (Forman et al. 2003). In Florida, roadsides are commonly planted with Paspalum notatum Flgg var. saurae Parodi (bahiagrass), a non-native pasture grass introduced to the United States from Brazil (Violi 2000). Paspalum notatum forms dense mats of vegetation, ofte n preventing native plant establishment (Violi 2000). Applying herbicide to extant ve getation (postemergence) before seeding may increase wildflower establishment (del-Val and Crawley 2005). The application of glyphosate, a non-selective herbicide, to P. notatum increased establishment of native species subsequently

PAGE 83

83 planted in P. notatum pastures (Uridel 1994). When applied at the time of wildflower seeding, imazapic, a selective herbicide, decreased the density and growth rate of existing vegetation while not substantially impacting the establishmen t of some wildflower species (Beran et al. 1999, Norcini et al. 2003). However, the tole rance of wildflowers to imazapic varies considerably according to wildflower species ecotype, and rate of application (BASF Corporation 2003, Norcini et al. 2003). Limited information is available on the esta blishment and manageme nt of local ecotypes of native wildflowers on Florida roadsides. Seeds have been traditionally purchased from nonlocal sources (Florida Department of Transportation 1998), but local ecotype seeds have recently become available via Floridas Wildflow er Seed and Plant Growers Association Inc. (2008). Local ecotypes may be better adapted to the area in which they o ccur than more distant ecotypes (Norcini et al. 2001). For example, Norcini et al. (2 001) found that local ecotypes of Coreopsis lanceolata in Florida bloomed earlier and had hi gher survivorship than nonlocal ecotypes. However, environmental conditions on roadsides are altered and often do not mimic the biotic and abiotic character istics of native habitats (For man and Alexander 1998, Trombulak and Frissell 2000, Jenkins et al. 2004). Although Florida ecotypes are adapted to Floridas climate, establishment and management practices may be different than t hose for nonlocal seed sources and on altered landscapes such as roadsides. Roadsides are managed to meet several ob jectives including driver safety, road stabilization, and maintenance access. Mowing of roadside vegetation is the most common and widespread form of roadside management, and is primarily utilized for driver safety. The frequency, timing, and height of mowing can influence plant species composition on roadsides. Mowing can increase species diversity (Pa rr and Way 1988, Collins et al. 1998, Hofmann and

PAGE 84

84 Isselstein 2004). In a long-term study comparing mowing frequencie s, the diversity of roadside vegetation was greater in areas mowed two times pe r year than once per year or not at all, and the lowest diversity was found in areas that were not mowed (Parr and Way 1988). In addition to increasing diversity, mowing and removing litter can increase germination and establishment of seedlings (Jensen and Meyer 2001, Jutila and Grace 2002) However, the time of year an area is mowed can significantly impact the growth and r ecruitment of a species, especially if mowing precedes seed set (Brys et al. 2004). The objective of this study was to determine the effect s of pre-seeding herbicide application and mowing frequency on native wildfl ower establishment on simulated roadsides in Florida. I hypothesized that a moderate amount of disturbance (mowi ng twice per year) and reduced competition through herbicide treatment s would result in increased wildflower establishment. Methods Study Sites and Species To examine treatment effects across a range of climate and soil types, the experiment was conducted in three different regions of Florida (Table 4-1, Figure 4-1). Sites were located in north (Quincy), north-central (Citra), and south-central (Fort Pierce) Flor ida (Table 4-1). All sites are research centers of the University of Floridas Institute of Food and Agriculture Sciences. Conducting the experiments at rese arch centers allowed for more controlled conditions and easier access to equipment than would be possible on roadsides. The study areas were planted in P notatum approximately 10-20 years before the beginning of the study. Selected for its drough t tolerance and lack of pest problems, P. notatum was originally introduced as a pasture grass (Scott 1920). Paspalum notatum is now commonly planted on roadsides due to its effectiveness in controlling erosion through the formation of a

PAGE 85

85 dense root system (Violi 2000). While seve ral cultivars have been developed since its introduction, Pensacola is the cultivar planted at all three study sites and frequently planted on pastures and roadsides in Florida (Violi 2000). Native wildflowers in this study includ e three species in the Asteraceae: Gaillardia pulchella Foug. (firewheel), Coreopsis lanceolata L. (lanceleaf tickseed), and Coreopsis leavenworthii Torr. & A. Gray (Leavenworths tick seed) (United States Department of Agriculture 2008). Gaillardia pulchella is an annual to short-lived perennial commonly occurring throughout Florida in disturbed uplands (Wunderlin and Hansen 2003) It blooms from spring to summer in north Florida a nd year-round in south Florida (Osorio 2001). Coreopsis lanceolata is an evergreen, short-lived perennial that occurs thr oughout much of the United States and blooms in spring (Flora of North America Editorial Committee 2006, United States Department of Agricu lture 2008). In Florida, C lanceolata occurs in northern and northcentral regions in sandhill and distur bed habitats (Wunderlin and Hansen 2003) Coreopsis leavenworthii is an annual to short-lived perennial and a facultative wetla nd species (Wunderlin and Hansen 2008) Previously thought to be endemic to Florida, C. leavenworthii is now known to occur in Alabama (United States Department of Agriculture 2008). Coreopsis leavenworthii is common throughout Florida and occurs in depression marshes, disturbed wetland, marl prairie, pine rockland, wet flatwood, and wet prairie habitats (Gann et al. 2008). Flowers may be produced year-round in south Florida, but most fl owers are produced in spring and early to midsummer (Osorio 2001). Local ecotype seeds of each wildflower sp ecies were mixed with moistened sand and hand broadcast into P. notatum pastures in early October 2005. The seeding rate (pure live seed) varied by species: C leavenworthii 3.1 kg/ha (~2007 seeds/m2); C. lanceolata 6.2 kg/ha

PAGE 86

86 (~554 seeds/m2); and G. pulchella 1.9 kg/ha (~121 seeds/m2). All three wildflower species were seeded in Citra and Quincy; only C leavenworthii was seeded in Fort Pierce. ( C. lanceolata is not native to south-centr al Florida and local ecotype G. pulchella seeds were not available for this ecoregion.) Within three weeks of seeding, Hurrican e Wilma made landfall near the Fort Pierce site and study plot s were inundated for several days. Experimental Design I evaluated the effects of mowing regime and establishment methods on three native Florida wildflower species for two years at thr ee sites in Florida. Treatment effects were assessed separately for each specie s and there were four replicates per site. Replicate plots were 12 m2 (3 X 4 m) with 1-m perimeter buffers within each plot, 1-m alleys between each plot, and a 10-m buffer strip between each experiment. Two mowing regimes and three establishment methods were evaluated in a split plot design, with mowing regime as the main plot treat ment and establishment method as the sub plot treatment. The two mowing regimes differed in the number of times each plot was mowed, either two or six times per year. The timing of mowi ng treatments was adjusted for flowering and seeding phenology of each wildflow er species (Table 4-2). The three establishment methods evaluated the use of herbicides (glyphosate, imazapic, and an untreated control) on extant vegeta tion. Glyphosate (Roundup, Monsanto) was applied at 2.3 L ai/ha approximately 4 and 2 weeks be fore seeding. Imazapic (Plateau, Impose, Panoramic 2SL, BASF) was applied at 0.07 kg ai/h a within one day of seeding. In Citra, herbicide was applied with a Cushman Sprayte k (Jacobsen Co.) that delivered 281 L/ha at 310 kPa. In Quincy and Fort Pierce, herbic ide was applied with a compressed air backpack sprayer (Teejet flat fan nozzle, Spraying System s Co.) that delivered 3 74 L/ha at 138 kPa.

PAGE 87

87 Sites were prepared by mowing the extant vegetation to 5 cm and removing thatch from the plots within one week before seeding. Because P. notatum was especially dense in Fort Pierce, vegetation was scarified with a conservation planter before seeding. After seeding, plots were rolled with a turf roller to increase seed-to-soil contact and irrigated overhead for approximately two weeks. At the start of the study, P. notatum was the dominant species with greater than 70% cover at each site. Forbs comp rised less than 2% and 9% cover in Citra and Quincy, respectively. In Fort Pierce, forbs comp rised 32% cover mainly due to the occurrence of Arachis glabrata Benth. (perennial peanut, a forage crop) that was previously planted. None of the planted wildflower species we re growing at the study sites pr ior to seeding. A preliminary seed bank study conducted in fall 2005 confirmed that wildflower species were not present in the seed bank prior to the experiment. Research plots were evaluated in fall (November/December), spring (April/May), and summer (August/September) from 2005 through 2007. Density and pe rcent cover were evaluated in the center square meter of each plot. Density counts were for all individuals (adults and seedlings) of the wildflower species. Per cent cover was visually estimated (to the nearest percent) by the same observer at each data collection to minimize observer bias (Korb et al. 2003). Percent cover categories included the wildfl ower species, forbs (exc luding the wildflower species), graminoids, bare ground, and standing dead /litter. To allow for future evaluations, only nondestructive sampling was conducted. Seed Bank Study In fall 2006, eight soil cores (6 cm diameter by 5 cm deep, 141.3 cm3) were randomly sampled from each plot (excluding the center sq uare meter), aggregated, and a subsample (282.6 cm3) was selected for the seed bank study. Soil samples were spread evenly, to no more than 2 cm deep, over a soilless medium (Fafard #2 Mix, Conrad Fafard, Inc.) in pots (18 cm diameter

PAGE 88

88 by 11 cm high). Initially, soil was misted until satu rated and subsequently irrigated by drip lines placed underneath the top 2 cm of soil. Pots we re arranged in a randomized complete block design in a greenhouse at the Univers ity of Florida in Gainesville. A control pot filled only with a soilless medium was included within each bloc k. After 2 months, emerged seedlings were removed from each pot and counted into groups of wildflower species, forbs, and graminoids. The surface soil was stirred in the pots and ir rigated. After another 2 months, all emerged seedlings were counted again. The first and second data collections were combined for analysis. Statistical Analysis I analyzed aboveground vegeta tion data using mixed models with restricted maximum likelihood methodology (PROC MIXED, version 9.1; SAS Institute, Cary North Carolina, USA). Preliminary results indicated signi ficant differences among seasons of data collection; therefore, each season was analyzed separately. Within each season, the fixed factors were mowing regime, establishment method, site, and year of data collection. Rando m factors were block (from split plot design), block by mowing regime, and block by site. To identify changes in wildflower percent cover and density over time, a repeated measures analysis was added to the model. Density data were log ( x + 1) transformed and percent c over data were arcsine square root transformed to meet assumptions of norma lity and homogeneity of variance. Treatment effects on the seed bank (emerged seedlings) were determined as above, but without the repeated measures analysis. Counts of emerged seedlings were converted from cm3 to m2 (McBurney 2005), and log ( x + 1) transformed to meet assumptions of normality and homogeneity of variance. Means were separated using least squares means (with PDIFF option) as part of the mixed models analyses; P values were adjusted using the Bonferroni method.

PAGE 89

89 Results Mowing frequency did not affect the percen t cover or seed bank density of any of the wildflower species (Tables 4-3, 4-4, and 4-5). However, establishment method was significant for every species during every season of data collection (P < 0.0001, Tables 4-3, 4-4, and 4-5). In most cases, the glyphosate treatment increas ed wildflower cover and seed bank density compared to the control, but results varied by wild flower species and site. Therefore, results of each species are presented separately. Gaillardia pulchella Mowing frequency had no effect on G. pulchella emergence or growth; however, establishment treatment significantly affected per cent cover, density, and seed bank density. The glyphosate treatment resulted in greater G. pulchella cover compared to the control and imazapic treatments (Table 4-3, Figure 4-2). Cover wa s less than 3% in the control and imazapic treatments and did not differ between the two tr eatments. During spring, cover was greater in 2007 than 2006 in the glyphosate treatment but s howed no response by year in the control and imazapic treatments (Table 4-3, Figure 4-2). During fall, cover decreased in the glyphosate treatment as a result of G. pulchella mortality (or aboveground di eback) following the growing season. The seasonal reduction in cover was greater in Citra than in Quincy in 2006, resulting in a significant establishment by site by year interaction (Table 4-3, Figure 4-2). I observed no difference in density among es tablishment treatments the first year after seeding. However, in the second year of the e xperiment (starting fall 2006 in Quincy and spring 2007 in Citra), density was greater in the glyphosate treatment than in the control and imazapic treatments ( P < 0.001; data not presented). Similarly, the number of seedlings emerging from the soil seed bank was greater in th e glyphosate treatment than the control and imazapic treatments in Citra (Table 4-6, Figure 4-3, based on result s of seed bank study). However, there was no

PAGE 90

90 difference in seedling emergence in Quincy, resu lting in a significant site by establishment interaction (Table 4-6, Figure 4-3). Regardless, G. pulchella contributed less than 4% to the overall seed bank dens ity (Figure 4-3). The establishment treatment affected both the aboveand belowground composition of forbs and graminoids (Figure 4-3). The glyphosate treatment decreased graminoid cover but increased forb cover in spring 2007 (Figure 4-3). Graminoids generally dominated the aboveground vegetation, yet only comprised a small percentage of the seed bank. The majority of seeds in the seed bank consisted of forb s, which comprised less than 13% of aboveground cover in the control and imazapic treatments and up to 40% cover in the glyphosate treatment (Figure 4-3). Coreopsis lanceolata Similar to G. pulchella C. lanceolata was significantly affected by establishment treatments but not mowing treatments. Coreopsis lanceolata cover in the glyphosate treatments was greater than the control and imazapic treatmen ts in all seasons (Table 4-4, Figure 4-2). The imazapic treatment increased C. lanceolata cover compared to the control treatment in Quincy but not in Citra (Figure 4-2), resulting in significant ( P < 0.05) establishment by site interactions for spring and summer (Table 4-4) Until one year after seeding, C. lanceolata cover generally increased (Figure 4-2). However, the extent of the increase was affected by site, establishment treatment, and season, resulting in several significant inte ractions (Table 4-4). In the glyphosate treatment, C. lanceolata cover did not change in the second year of the experiment (Figure 4-2, fall 2006-fall 2007). In the control and imazapic treatments, however, cover increased from 2006 to 2007 in Citra but there was no change in cover in Quincy (Figure 4-2). During the second year of the experiment (starting fall 2006), density was consistently greater in the glyphosate treatment than the control and imazapic treatments ( P < 0.0001; data not

PAGE 91

91 presented). In the glyphosate treatmen t, density averaged 44 and 491 individuals/m2 in Citra and Quincy, respectively. Density in the cont rol and imazapic treatments averaged 13 and 75 individuals/m2 in Citra and in Quincy, respectively (averages include fall 2006 and spring and fall 2007). While C. lanceolata seed bank density was greater in the glyphosate treatment than the control and imazapic treatments, C. lanceolata only comprised a small proportion of the overall seed bank (Tab le 4-6, Figure 4-4). Establishment treatments affected both aboveand belowground community composition (Figure 4-4). Coreopsis lanceolata dominated the aboveground ve getation in the glyphosate treatment while graminoids dominated the a boveground vegetation in the control and imazapic treatments (Figure 4-4). The gl yphosate treatment decr eased graminoid cover but did not affect seed bank density compared to the control and imazapic treatments ( P < 0.001, Figure 4-4). Moreover, graminoid cover was lower in the im azapic treatment than th e control in fall and spring, regardless of year ( P < 0.05, Figure 4-4). In all establishment treatments, forbs comprised the majority of the seed bank but a small minority of aboveground vegetation (Figure 4-4). Coreopsis leavenworthii Effects of establishm ent on C. leavenworthii cover varied by site, season, and year (Table 4-5). In general, C. leavenworthii cover decreased over time, despite seasonal fluctuations (Figure 4-2). Additionally, cover was typically greater in the glyphosate treatment than in the control and imazapic treatments, but this varied co nsiderably by site (Figure 4-2). In Quincy, cover in the glyphosate treatment was greater than the control and imazapic treatments, except in fall 2006 when there was no difference between the glyphosate and imazapic treatments ( P <0.05 for all seasons and years, Figure 4-2). In Citra, C. leavenworthii cover was greater in glyphosate treated plots than in control and imazapic treated plots in summer and fall 2006 ( P

PAGE 92

92 <0.05 for all treatments) but there was no differe nce in cover among establishment treatments before or after this time period (Figure 4-2) There was no difference in cover between the imazapic and control treatments in Quincy or Citra. In Fort Pierce, cover was greater in the glyphosate treatment than the control and imazapic treatments from fall 2005 through summer 2006 ( P < 0.05 for all seasons and treatments, Figure 4-2). Coreopsis leavenworthii cover remained below 1% in imazapic treated plots in Fort Pierce throughout the study (Figure 4-2), signifi cantly lower than control and glyphosate treated plots from fall 2005 through spring 2007 ( P < 0.05 for all seasons and treatments). In summer and fall 2007, wildflow er cover in control and glyphosate treated plots decreased, resulting in no differences among an y establishment treatment (Figure 4-2). Mowing frequency alone did not affect cover of C. leavenworthii ; however, there was a mowing by site interaction during summer ( P = 0.0064, Table 4-5). This resulted from a difference in mowing treatment in Fort Pierce only. During summer 2006, cover was estimated within two weeks of a mowing treatment, result ing in a short-term difference between mowing treatments. Coreopsis leavenworthii density fluctuated widely, varying by site, season, and establishment treatment. For example, density averaged 3,146 individuals/m2 in the glyphosate treatment in Quincy in fall 2006 compared to 1,127 and 285 individuals/m2 in control and imazapic treatments, respectively. Despite this increase in density, cover decreased in glyphosate treated plots during the same time pe riod due to mortality (or aboveground dieback) of adults following the growing season (Figure 4-2). At all sites, the glyphosate treatment increased C. leavenworthii seed bank density compared to the control and imazapic treatments (Table 4-6, Figure 4-5). In Quin cy, there was no difference in the C. leavenworthii seed bank

PAGE 93

93 density between the control and imazapic treatm ents. However, the imazapic treatment increased the seed bank density in Citra and decreased the s eed bank density in Fort Pierce compared to the control treatment (establishment by site interaction, Table 4-6, Figure 4-5). Effects of establishment treatments on the aboveand belowground community composition varied by site (Figure 4-5). In general, graminoids dominated the aboveground vegetation. However, in Quincy, C. leavenworthii was dominant in the glyphosate treated plots. In Fort Pierce, forbs dominated the imazapic treated plots. Coreopsis leavenworthii comprised the majority of the seed bank in the glyphosate trea tment in Fort Pierce and Quincy (Figure 4-5). However, in Citra, forbs dominated the glyphos ate treatment, despite a high density of C. leavenworthii in the seed bank (Figure 4-5), resulti ng in significant site by establishment interactions (Table 4-6). Forb s also dominated the seed bank in the control and imazapic treatments (Figure 4-5). Discussion Species-Specific Responses Treating P. notatum with glyphosate in the early fall before wildflower seeds were sown increased es tablishment of all three wildflower species compar ed to the control and imazapic treatments. However, overall wildflower establishment varied by species. Coreopsis lanceolata was able to establish without any disturbance to the extant vegetati on, and the cover of the established population genera lly increased over time. Coreopsis lanceolata cover in the glyphosate treated plots remained high throughout the study, and C. lanceolata was often the dominant species. Coreopsis lanceolata cover was also consiste ntly high throughout the study, with few seasonal fluctuations (Figure 4-2). Although the glyphosate treatment increased G. pulchella cover, establishment of this species was lower than that of the ot her two wildflower species. Cover of G. pulchella remained

PAGE 94

94 below 25% throughout the study. This may have been caused by seeding at a lower seeding rate than the other two species. Although the G. pulchella population was small, it did not appear to decrease over time, aside from seasona l fluctuations in cover (Figure 4-2). Coreopsis leavenworthii establishment varied by site and over time. Establishment was lower in Citra than in Quincy and Fort Pierce. Coreopsis leavenworthii is a facultative wetland species and Citra is part of an upland, xeric habitat. Despit e sufficient seedling emergence in the second year of the study, C. leavenworthii cover decreased over time. Coreopsis leavenworthii is an annual to short-lived perennial th at often dies after flowering. The seasonal mortality of this species may have created gaps that were colonized by P. notatum and ruderal species from the seed bank. These species, in turn, likely limited the gr owth (and increase in cover) of C. leavenworthii seedlings. The results from the seed bank study provide further evidence of species-specific responses to establishment treatme nts. Species that maintain vi able seeds in the soil seed bank may be more likely to maintain stable populat ions over time (Yates and Ladd 2005), provided that the seed bank serves as a source rather than a sink of new i ndividuals (Rees 1997). However, in this study, the aboveground vegeta tion differed markedly from the seed bank composition. One year after initial seeding, C. leavenworthii seeds dominated the seed bank in the glyphosate treatment (Figure 45). However, this was not e xpressed in aboveground growth for C. leavenworthii in Fort Pierce or Citra. C onversely, in glyphosate treated plots, C. lanceolata was the dominant species aboveground but did not comprise the majority of seeds in the seed bank (Figure 4-4). Gaillardia pulchella neither dominated the seed bank community nor the aboveground vegetation (Figure 4-3). Studi es on the establishment of perennial native

PAGE 95

95 forbs in grass-dominated communities also f ound differential effects by species (Fenner 1978, Brown and Bugg 2001). Effects of Imazapic In contrast to the hypothesis and other studies (Beran et al. 1999, Washburn and Barnes 2000), imazapic resulted in minimal wildflower establishment over the control treatment. Applied at low rates, imazapic reduces P. notatum vegetative growth and seedhead development (Baker et al. 1999, BASF Corporation 2003). Although P. notatum seed production was not measured in this study, imazapic did not reduce P. notatum cover or seed bank density, except in Fort Pierce (Figures 4-3, 4-4, and 4-5). Howe ver, the effects of imazapic on wildflower establishment varied by site and species. For example, C. lanceolata cover was greater in imazapic treated plots in Quincy but not in Citra (Figure 4-2). Coreopsis lanceolata may be more tolerant to imazapic than the other wildfl ower species. In a study by Norcini et al. (2003), imazapic stunted the growth of G. pulchella more than C. lanceolata In Fort Pierce, C. leavenworthii did not establish in the imazapic treatment Seedlings emerged within a few days of seeding but died shortly thereafter. Imazapic can cause stand-thinning, which may be corrected by increasing the seeding rate (althoug h not tested in this study) (BASF Corporation 2003, Norcini et al. 2003). Moreover, tolerance to imazapic has been shown to vary according to local soil and environmental conditions (BASF Corporation 2003). In a study on imazethapyr, an herbicide in the same family as imazapic, persistence in the soil increased as soil pH decreased (Loux and Reese 1993). While soil pH in Fort Pierce was lower than the other two sites, it is not clear why the effect of imazapic on C. leavenworthii was different in Fort Pierce than in Quincy and Citra.

PAGE 96

96 Mowing Treatments Mowing frequency did not affect cover, de nsity, or number of seedlings emerging from the seed bank of any of the wildflower species in this study (Table 4-6). This result contrasts with other studies in which diversity increase d in response to mowing or grazing (Hofmann and Isselstein 2004, Williams et al. 2007); however, mowing was typically more frequent than six times per year. Additionally, mowing in the present study was adjusted for the flowering and seeding phenology of each of the wildflower speci es, likely reducing the effects of mowing. For example, because C. lanceolata blooms in early spring, plots were not mowed from March 1 through May 15 in the mow six times per year treatment (Table 4-2). The time of year an area is mowed can significantly impact the growth and survival of a species (Brys et al. 2004). In both annual and perennial species, if mowing precedes the seed maturation of a species, recruitment will be limited and the population will decrease in si ze. If the species is an annual, the population would only continue at that site if viable seeds were stored in the soil (Rees 1997, Norcini et al. 2003). Several studies have shown that mowi ng and removing litter can increase germination and establishment of seedlings (Jensen and Meyer 2001, Jutila and Grace 2002). Although litter was not removed in this study, mowing frequency had no effect on wildflower germination or seedling establishment. Paspalum notatum Dominance Establishm ent of wildflowers was likely limited by competition from P. notatum (Chapter 2, Violi 2000). The ability of P. notatum to limit establishment of other species is evident by comparing the aboveand belowground community composition. In undisturbed plots, graminoids (the majority of which were P. notatum ) dominated the aboveground vegetation, yet forbs comprised the ma jority of the seed bank. When P. notatum was disturbed by glyphosate, wildflower and forb es tablishment generally increased. Paspalum notatum is a

PAGE 97

97 rhizomatous, fast-growing, tropical grass that often forms a dense, monotypic stand (Violi 2000). In contrast, most grasses nativ e to this region are slow grow ing bunchgrasses that leave more patches of bare ground for other species to establish (Myers 1990). Moreover, the small proportion of graminoids emerging from the seed bank suggests that regeneration of graminoids, and in particular P. notatum after glyphosate treatment occurs vegetatively rather than through seeds. Forbs occupied the majority of the seed bank, reinforcing the benefits of a no-till system on roadsides or simulated roadsides. While tilling would decrease P. notatum competition, it would also increase competition from weedy forbs emerging from the seed bank. Conclusions Reducing extant vegetation through glyphosate re sulted in the greatest establishm ent of native wildflowers. Seeding wildflowers into P. notatum without disturbing the extant vegetation resulted in little to no establishment for G. pulchella and C. leavenworthii (except at Fort Pierce) and a moderate amount of cover for C. lanceolata An imazapic treatment at the time of seeding did not in crease establishment of G. pulchella and C. leavenworthii but did result in increased cover of C. lanceolata in Quincy. Differences in establishment by species suggest that perennials, particularly evergreen species, may persist longer in a competitive environment than annuals. Mowing frequency did not affect the percent cover or seed bank de nsity of any of the wildflower species; however, the timing of mowing was adjusted for the phenology of each wildflower species at each site. Since roadside mowing in Florida is timed typically to control the height of the dominant grass sp ecies rather than to facilitate wildflower seeding, it is unclear whether the mowing frequencies in this study would have the same effect on roadside wildflower populations.

PAGE 98

98 Table 4-1. Soil characteristics and type s before seeding na tive wildflowers ( Gaillardia pulchella Coreopsis lanceolata C. leavenworthii ) at the study sites located in Quincy, Citra, and Fort Pierce, FL. Means for soil chemistr y are presented with standard errors in parentheses. Site Site characteristic Quincy Citra Fort Pierce Location in Florida (Lat, Long) North (30N, 84W) North-central (29N, 82W) South-central (27N, 80W) Elevation (m) 70 20 6 Soil type Norfolk loamy fine sand (0-5% slopes) Candler sand (0-5% slopes) Ankona sand (0-2% slopes) pH 5.78 (0.02) 6.62 (0.04) 5.16 (0.05) Organic matter (%) 2.56 (0.12) 2.21 (0.07) 1.25 (0.02) NO3-N (mg/kg) 0.16 (0.02) 0.39 (0.02) 6.39 (0.86) NH4-N (mg/kg) 4.12 (0.12) 4.86 (0.46) 2.08 (0.20) P (mg/kg) 3.11 (0.25) 35.67 (1.52) 10.79 (0.56) K (mg/kg) 53.38 (3.57) 31.26 (2.07) 11.13 (0.55)

PAGE 99

99 Table 4-2. Dates of mowing treatments used at each site (Quincy, Citra, Fort Pierce, FL) for each wildflower species ( Gaillardia pulchella Coreopsis lanceolata C. leavenworthii ). The timing of mowing treatments was adjusted to allow for flowering and seeding of each wildflower species. Mowing dates were adjusted slightly for C leavenworthii in Fort Pierce because of earlier flow ering due to the warmer climate. Species Site Mowing Treatment Gaillardia pulchella Coreopsis lanceolata Coreopsis leavenworthii Year Quincy & Citra Quincy & Citra Quincy & Citra Fort Pierce Mow 2 times/ year 2006 Mar 15 Mar 1 Apr 1 Mar 1 Oct 15 Oct 15 Oct 15 Oct 1 2007 Mar 15 Mar 1 Apr 1 Mar 1 Oct 15 Oct 15 Oct 15 Oct 1 Mow 6 times/ year 2005 Dec 1 Nov 15 Nov 15 2006 Jan 1 Jan 1 Jan 1 Jan 1 Feb 15 Mar 1 Mar 1 Mar 1 Mar 15 May 15 Apr 1 Jul 15 Jul 1 Jul 15 Aug 1 Aug 21 Oct 15 Sep 1 Oct 15 Oct 1 Nov 15 Oct 15 Dec 1 Dec 1 Dec 15 2007 Feb 15 Jan 1 Jan 1 Jan 1 Mar 15 Mar 1 Mar 1 Mar 1 Jul 1 May 15 Apr 1 July 15 Oct 15 Jul 1 Aug 1 Aug 21 Sep 1 Oct 15 Oct 1 Oct 15

PAGE 100

100Table 4-3. Effects of establishment treatm ent (pre-seeding herbicide: control, glyphos ate, and imazapic), mowing frequency (two or six times/year), and their interactions on wildflower pe rcent cover (arcsine square root transformed) of Gaillardia pulchella in Quincy and Citra, FL. Fall season included three years of data collection; spring and summer included two years. Season Treatment factors Fall Spring Summer df F P df F P df F P Mowing 1, 30.230.6659 1, 31.35 0.3288 1, 30.020.9018 Establishment 2, 99131.69<0.0001 2, 63116.54 <0.0001 2, 63188.96<0.0001 Site 1, 30.160.7145 1, 33.38 0.1635 1, 31.670.2867 Year 2, 9911.78<0.0001 1, 630.68 0.4124 1, 630.330.5653 Mow X estab 2, 990.700.4971 2, 630.18 0.8321 2, 630.150.8613 Mow X site 1, 991.710.1945 1, 630.30 0.5828 1, 631.910.1716 Mow X year 2, 990.470.6272 1, 631.11 0.2954 1, 630.010.9405 Estab X site 2, 990.130.8821 2, 632.69 0.0758 2, 632.680.0762 Estab X year 4, 9920.57<0.0001 2, 6316.50 <0.0001 2, 630.280.7543 Site X year 2, 9926.05<0.0001 1, 631.99 0.1629 1, 630.070.7966 Mow X estab X site 2, 990.320.7243 2, 630.62 0.5394 2, 631.510.2285 Mow X estab X year 4, 990.440.7802 2, 630.78 0.4611 2, 630.570.5665 Mow X site X year 2, 990.790.4578 1, 630.11 0.7464 1, 630.050.8258 Estab X site X year 4, 9912.25<0.0001 2, 630.25 0.7775 2, 631.120.3314 Mow X estab X site X year 4, 990.340.8492 2, 631.12 0.3315 2, 630.110.8919

PAGE 101

101Table 4-4. Effects of establishment treatm ent (pre-seeding herbicide: control, glyphos ate, and imazapic), mowing frequency (two or six times/year), and their interactions on wildflower pe rcent cover (arcsine square root transformed) of Coreopsis lanceolata in Quincy and Citra, FL. Fall se ason included three years of data collection; spri ng and summer included two years. Season Treatment factors Fall Spring Summer df F P df F P df F P Mowing 1, 30.060.82561, 30.46 0.54601, 30.020.8949 Establishment 2, 9897.81<0.00012, 6389.99 <0.00012, 63123.71<0.0001 Site 1, 32.010.25151, 323.48 0.01681, 35.270.1054 Year 2, 98484.10<0.00011, 6391.99 <0.00011, 6320.17<0.0001 Mow X estab 2, 980.280.75892, 630.36 0.69972, 630.480.6183 Mow X site 1, 980.120.72961, 630.65 0.42331, 630.420.5218 Mow X year 2, 980.230.79521, 631.93 0.16951, 630.660.4183 Estab X site 2, 981.900.15502, 634.29 0.01792, 637.810.0009 Estab X year 4, 9818.15<0.00012, 634.46 0.01552, 635.460.0065 Site X year 2, 9823.97<0.00011, 631.80 0.18501, 636.710.0119 Mow X estab X site 2, 980.020.97772, 630.16 0.85462, 630.690.5046 Mow X estab X year 4, 980.760.55152, 630.39 0.67662, 630.000.9995 Mow X site X year 2, 980.430.64861, 630.11 0.74361, 630.080.7773 Estab X site X year 4, 984.230.00332, 632.11 0.12962, 630.800.4544 Mow X estab X site X year 4, 980.190.94422, 630.16 0.85092, 630.280.7596

PAGE 102

102Table 4-5. Effects of establishment treatm ent (pre-seeding herbicide: control, glyphos ate, and imazapic), mowing frequency (two or six times/year), and their interactions on wildflower pe rcent cover (arcsine square root transformed) of Coreopsis leavenworthii in Quincy, Citra, and Fort Pierce, FL. Fall season incl uded three years of data co llection; spring and summer included two years. Season Treatment factors Fall Spring Summer df F P df F P df F P Mowing 1, 30.090.78091, 30.21 0.67601, 31.330.3323 Establishment 2, 15081.55<0.00012, 96103.04 <0.00012, 96235.23<0.0001 Site 2, 68.160.01942, 628.04 0.00092, 622.210.0017 Year 2, 15023.55<0.00011, 9631.37 <0.00011, 96288.86<0.0001 Mow X estab 2, 1500.500.60592, 960.55 0.57942, 960.280.7587 Mow X site 2, 1500.770.46582, 961.82 0.16832, 965.320.0064 Mow X year 2, 1500.060.94641, 960.85 0.35811, 963.300.0725 Estab X site 4, 1508.66<0.00014, 9638.96 <0.00014, 9630.16<0.0001 Estab X year 4, 1501.320.26692, 960.49 0.61312, 9663.43<0.0001 Site X year 4, 15019.67<0.00012, 960.20 0.82222, 967.410.0010 Mow X estab X site 4, 1500.140.96604, 961.22 0.30714, 962.150.0811 Mow X estab X year 4, 1500.080.98732, 960.08 0.92012, 960.960.3861 Mow X site X year 4, 1500.110.97922, 962.28 0.10742, 961.130.3258 Estab X site X year 8, 1505.70<0.00014, 968.68 <0.00014, 968.03<0.0001 Mow X estab X site X year 8, 1500.790.61624, 961.03 0.39424, 961.380.2471

PAGE 103

103Table 4-6. Effects of establishment treatment (pre-seeding herbic ide: control, glyphosate, and imazapic), mowing frequency (two or six times/year), and their interactions on emerged wildflower seedlings from the so il seed bank (per meter squared) in fall 2006 in Quincy, Citra, and Fort Pierce, FL. Species Site Gaillardia pulchella Coreopsis lanceoata Coreopsis leavenworthii Factor level Quincy Citra Quincy Citra Quincy Citra Ft. Pierce Establishment Control 0 a 0 a 133 a 199 a 686 a 199 a 4,313 a Glyphosate 88 a 1,327 b 2,565 b 3,118 b 30,786 b 26,362 b 31,980 b Imazapic 22 a 22 a 265 a 177 a 929 a 973 c 376 c Significance of F-test Establishment *** *** *** Mowing ns ns ns Site ns ns Establishment x mowing ns ns ns Establishment x site *** ns *** Mowing x site ns ns ns Establishment x mowing x site ns ns ns Significant main effects and interactions are indicated by: *** ( P <0.0001) and ( P <0.05). Means with different letters within a species and site are significantly different ( P <0.05, Bonferroni correction). Analysis was performed on log transformed data; untransformed means are presented here.

PAGE 104

104 Solar Radiation (w/m 2 ) 0 100 200 300 400 500 Quincy Citra Ft. Pierce Month and year Sep Jan May Sep Jan May Sep Rainfall (cm) 0 5 10 15 20 25 30 Temperature (C) -10 0 10 20 30 40 2006 2005 2007 Figure 4-1. Solar radiation (monthly average) temperature (monthly average minimum and maximum), and rainfall (monthl y total) in Fort Pierce, Citra, and Quincy, FL during the study period. Data from FAWN, Florida Automated Weather Network (http://fawn.ifas.ufl.edu).

PAGE 105

105 Percent cover/m 2 0 5 10 15 20 25 30 Season and year F a l l 0 5 S p r i n g 0 6 S u m m e r 0 6 F a l l 0 6 S p r i n g 0 7 S u m m e r 0 7 F a l l 0 7Percent cover/m 2 0 5 10 15 20 25 30 Season and year F a l l 0 5 S p r i n g 0 6 S u m m e r 0 6 F a l l 0 6 S p r i n g 0 7 S u m m e r 0 7 F a l l 0 7 0 20 40 60 80 0 20 40 60 80 0 20 40 60 80 Season and year F a l l 0 5 S p r i n g 0 6 S u m m e r 0 6 F a l l 0 6 S p r i n g 0 7 S u m m e r 0 7 F a l l 0 7Percent cover/m 2 0 20 40 60 80 Quincy Citra Fort Pierce Coreopsis leavenworthii Coreopsis lanceolata Gaillardia pulchella 0 20 40 60 80 Glyphosate Control Imazapic Figure 4-2. Effects of establishm ent treatment (pre-seeding herbicide: control, glyphosate, and imazapic) on mean ( 1 SE) perc ent cover of Gaillardia pulchella Coreopsis lanceolata and C. leavenworthii by site, season, and year.

PAGE 106

106 0 10000 20000 30000 40000 50000 60000 Quincy Establishment treatment C o n tr o l G l y p h o s a te Im a z a p i c 0 20 40 60 80 100 Quincy G. pulchella forbs graminoids Seedlings/m 2 Citra Establishment treatment C o n tr o l G l y p h o s a te Im a z a p i cPercent cover/m 2 Citra G. pulchella forbs graminoids Figure 4-3. Composition of Gaillardia pulchella forbs, and graminoids in the seed bank in fall 2006 (emerged seedlings/m2, top graphs) and as aboveground vegetation in spring 2007 (percent cover/m2, bottom graphs) in Quincy and Citra, FL.

PAGE 107

107 Establishment treatment C o n tr o l G l y p h o s a te I m a z a p i c 0 20 40 60 80 100 120 Quincy 0 10000 20000 30000 40000 50000 60000 Quincy Establishment treatment C o n t r o l G l y p h o s a t e Im a z a p i cPercent cover/m 2 Citra Seedlings/m 2 Citra C. lanceolata forbs graminoids Figure 4-4. Composition of Coreopsis lanceolata, forbs, and graminoids in the seed bank in fall 2006 (emerged seedlings/m2, top graphs) and as aboveground vegetation in spring 2007 (percent cover/m2, bottom graphs) in Quincy and Citra, FL.

PAGE 108

108 Establishment treatment C o n tr o l Gl y p h o s a te Im a z a p i cPercent cover/m 2 0 20 40 60 80 100 120 140 Seedlings/m 2 0 20000 40000 60000 80000 Establishment treatment C o n tr o l G l y p h o s a te I m a z a p i c Establishment treatment C o n tr o l G l y p h o s a te Im a z a p i c Citra Citra Fort Pierce Quincy Fort Pierce Quincy C. leavenworthii forbs graminoids Figure 4-5. Composition of Coreopsis leavenworthii forbs, and graminoids in the seed bank in fall 2006 (emerged seedlings/m2, top graphs) and as aboveground vegeta tion in spring 2007 (percent cover/m2, bottom graphs) in Quincy, Citra, and Fort Pierce, FL.

PAGE 109

109 CHAPTER 5 CONCLUSIONS Results from this study indicate that nati ve wild flower establis hment on roadsides and pastures dominated by P. notatum is limited by competition (C hapter 2). Disturbing P. notatum with glyphosate herbicid e prior to wildflower seeding grea tly improved establishment of all species (Chapters 3 and 4). Ot her herbicide treatments, includi ng applying imazapic at the time of wildflower seeding and a pplying sethoxydim post wildflower emergence, resulted in no to little increase in wildflow er establishment in most cases. A moderate amount of post-planting disturbance (mowing two or six times per year) di d not affect the short-term sustainability of wildflower populations, perhaps because the tim ing of disturbance was adjusted for the phenology of each wildflower species (Chapter 4). However, frequent disturbance (cutting 12 or 24 times per year) that was not adjusted to each wildflower species phenology reduced the growth and fitness of wildflower populations (Chapter 2). Although overall effects of establishment and management were similar among the wildflower species, some speciesspecific differences were noteworthy and may help to guide future research. While competition from P. notatum limited wildflower establishment and growth, aboveground competition did not account for the to tal reduction in growth. In Chapter 3, both the glyphosate and scraped disturbance treatments reduced P. notatum biomass; however, this decrease alone did not result in an equivalent increase in wildflower cover or biomass. Most studies on sessile organisms equate microsites with available space for growth (Eriksson 2005). However, the differences in C. lanceolata establishment between the glyphosate and scraped treatments suggest that microsite requirements are more complex than providing adequate space and likely include abiotic as well as biotic components (Tilman 1994). Moreover, in Chapter 2, intraspecific competition resulted in greater reduction of Coreopsis biomass than interspecific

PAGE 110

110 competition, yet conspecific neighbor biomass was smaller than that of Paspalum neighbors. In the field experiment in Chapter 3, scraped plots had less organic matter than glyphosate treated plots. Additionally, although not measured in the study, there was likely less moisture at the soil surface of scraped plots than glyphosate treated plots. In the study presented in Chapter 2, fertilizer and supplemental irriga tion were applied so that re sources would not be limiting. However, because conspecific individuals have the same resource needs, those resources were likely depleted more quickly with intraspe cific rather than interspecific competition. Wildflower establishment was aff ected by seasonal differences. Coreopsis species planted in the fall had greater survivorship, biomass, and numb er of flowers compared to Coreopsis planted in spring (Chapter 2). Although wildflowers were seeded only in fall in the field experiments, wildflower establishment di d not seem to be limited in glyphosate treated plots following seeding. Seasonal differences in wildflower establishment appeared related to P. notatum phenology. P. notatum is a long-day plant that is dormant during the short days of winter and initiates flowering when day length exceeds 13.5 hours (Marousky and Blondon 1995). Wildflowers planted in fall are able to establish when P. notatum is dormant, which likely limits competition. Seasonal differences may also help to explain species-specific responses of wildflowers. The fitness and perhap s subsequent recruitment of species that bloom in spring, prior to P. notatum flower initiation, may be greater than species that bloom later due to decreased competition from P. notatum The continued recruitment of Phlox drummondii Hook. and C. basalis (A.Dietr.) S.F.Blake, both of wh ich bloom in spring (Wunderlin and Hansen 2003), may also be due to their production of flowers while P. notatum is dormant. Although neither P. drummondii nor C. basalis is native to Florida (Wunderlin and Hansen 2003), they are both commonly planted on Florida roadsides.

PAGE 111

111 Seeding rates can greatly influence wildflower establishment, although this was only explicitly tested for one wild flower species in this study ( C. lanceolata Chapter 3). Seeding at too low of a rate results in low levels of es tablishment while seeding at too high of a rate provides no benefit for the additi onal resources expended (Chapter 3, Burton et al. 2006). Pure live seed rates are often expresse d as seed mass per unit area seed ed (e.g., kg/ha). Although this type of pure live seed rate accounts for differenc es in viability, it does not take into account differences in seed mass by species. Therefore, ev en if the same pure live seed rate is used for different species, the number of viable seeds per unit area may vary widely among species. For example, because the mass of C. leavenworthii seeds is smaller than C. lanceolata seeding at 6 kg pure live seed/ha results in more C. leavenworthii seeds/ha than C. lanceolata seeds/ha. For species like C. leavenworthii that have small seeds and are s hort-lived, it may be more costeffective to seed at lower rates for consecutive years than at a high rate only once. To avoid seeding at too high of a rate, species-speci fic recommendations for seeding rates can be developed (e.g., Texas Department of Transportation, Markwardt 2005). Until species-specific seeding rates are established for Florida ecotype seeds, the number of live seeds per unit area can be approximated using the pure live seed rate a nd the seed mass (Burton et al. 2006). Using the appropriate seeding rate for each species may increase wildflower establishment and also decrease costs. The establishment of C. lanceolata was affected by both seed and microsite limitation. While there were interactive effect s of seeding rate, irrigation, a nd disturbance treatment early in wildflower establishment, by the end of the study, lack of disturbance to the existing vegetation (i.e., microsite limitation) proved to be the factor most limiting to establishment (Chapter 3). Although increasing the seeding rate and provi ding supplemental irrigation may improve

PAGE 112

112 establishment, these practices did not comp ensate for lack of disturbance. While P. notatum did not limit wildflower emergence, it did limit subse quent growth and reproduction of wildflowers (Chapters 3 and 4). Due to the possibility of increased erosion on roadsides, complete eradication of P. notatum with glyphosate may not be desirable. Eradication of P. notatum after wildflower establishment but sti ll early in the wildflower life cycle may result in established wildflower populations with little erosion. However, the selectiv e herbicides in this study were not effective in increasing wildflower establishment. Results from the herbicide treatments in Chap ters 3 and 4 provide further evidence that P. notatum is a competitive, dominant species that rest ricts growth of other species. The ability of P. notatum to limit establishment of other specie s was evident by comparing the aboveand belowground plant community composition (Chapter 4). In undisturbed plot s, graminoids (the majority of which were P. notatum ) dominated the aboveground vegetation, yet forbs comprised the majority of the seed bank. When P. notatum was disturbed by glyphosate, a non-selective herbicide, wildflower and forb establishment in creased. The limited effectiveness of imazapic and sethoxydim, both selective herbicides, in in creasing wildflower esta blishment demonstrate that a slight reduction in P. notatum growth does not result in competitive release (Chapters 3 and 4). Paspalum notatum is a rhizomatous, fast-growing, tr opical grass that often forms a dense, monotypic stand (Violi 2000). In contra st, most grasses native to Florida are slow growing bunchgrasses that leave more patches of bare ground in the landscape (Myers 1990). Native wildflowers may be able to coexist better with native grasses than with P. notatum because patches of bare ground of ten provide suitable microsites in which native wildflowers can become established.

PAGE 113

113 The effectiveness of the glyphosate treatment may have been relate d to seasonal effects on P. notatum phenology, although season of application wa s not tested in this study. Since glyphosate is applied as a spray application to aboveground plant growt h, foliar absorption is required for glyphosate activity. Because of this, rapidly growing plants are the most sensitive (Monaco et al. 2002). However, the timing of gl yphosate application can a ffect the efficacy of the herbicide (Adams and Galatowitsch 2006). Although rapidly growing plants are the most sensitive, better control of perennial plants may be achieved wh en plants are not rapidly growing. Many perennial plants, including P. notatum invest resources in stor ed carbohydrates, often in roots or rhizomes. For this reason, perennial pl ants are often able to resprout the season following glyphosate application. Because plant re source allocation to stored carbohydrates often follows a seasonal pattern, applying glyphosate when plants move carbohydrates to stored reserves may help to increase translocation of the herbicide itself (Adams and Galatowitsch 2006). This may result in greater control over the long-term. Post-planting disturbance (mow ing two or six times per year) did not affect the shortterm sustainability of wildflower populations, perhaps because the timing of disturbance was adjusted for the phenology of each wildflower species (Chapter 4). However, frequent disturbance (cutting 12 or 24 times per year) that was not adjusted to each wildflower species phenology reduced the growth and fitness of wildfl ower populations (Chapter 2). There were no interactive effects of mowing or cutting on wildflower esta blishment (Chapter 4) or growth (Chapter 2), suggesting that the effects of mowi ng and competition are additive. Wildflower establishment would be expected to increase with either reduced competition, less frequent disturbance, or both. Moreover, mowing too frequently can not only negatively impact newlyestablished wildflowers but also can reduce the sustainability of longestablished wildflower

PAGE 114

114 populations. Therefore, mowing of roadside vegetation in areas with wildflower populations should be carefully timed to adjust for the flowering and seeding phenology of each wildflower species at each site. Although si milar guidelines are included in the Florida Department of Transportations Wildflowers in Florida (1998), these guidelines appear to be followed inconsistently in Florida (per sonal observation, Gordon et al. 2000). Management practices on Florida roadsides include mowing frequently and close to the ground (personal observation), although this is not beneficial to grasses or wildflowers. Regardle ss of wildflower establishment, it seems that the Florida Department of Tran sportation could reduce mowing frequency without affecting driver safety (persona l observation, Gordon et al. 2000) which would likely result in decreased maintenance costs. Native wildflowers responded differently to establishment and management practices in this study. For example, C. lanceolata became established without a ny disturbance to the extant vegetation, and the cover of the established population gene rally increased over time. Coreopsis leavenworthii establishment varied by site but tended to decrease over time, while establishment of G. pulchella was lower than that of the other two wildflower species (Chapter 4). Moreover, the imazapic treatment incr eased establishment of C. lanceolata only (Chapter 4). The results from the seed bank study provide further evidence of species-specific resp onses to establishment treatments. One year after initial seeding, C. leavenworthii seeds dominated the seed bank in the glyphosate treatment. However, this was genera lly not expressed in aboveground growth. The seasonal mortality of C. leavenworthii may have created gaps that were colonized by P. notatum and ruderal species, which lik ely limited the growth of C. leavenworthii seedlings. Conversely, in glyphosate treated plots, C. lanceolata was the dominant speci es aboveground but did not comprise the majority of seeds in the seed bank. Due to its ever green, perennial nature, C.

PAGE 115

115 lanceolata may have been better able to compete for space than C. leavenworthii (Chapter 4). Species-specific responses were also ev ident in response to competition from P. notatum (Chapter 2). Although survivorship of C. lanceolata was greater than that of C. leavenworthii C. lanceolata produced less biomass and fewer flowers than C. leavenworthii (Chapter 2). These results suggest that C. lanceolata may persist longer in a community dominated by P. notatum than its congener. Although this study was limited to three wild flower species, results suggest that evergreen, perennial wildflower species may be stronger competitors with P. notatum than annual species or perennials with seasonal growth fluctuations. Moreover, species that bloom in the spring while P. notatum is dormant may persist for longer on roadsides planted with bahiagrass. In areas where it is necessary to maintain aggressive mat-forming grasses on roadsides (i.e., for erosion control), native wildfl owers that are evergreen and/or bloom early in the spring may compete better than other wildflower species. However, many of Floridas native wildflowers are not evergreen or spring-bloomers. Therefore, where possible, limiting the planting and expansion of aggressive grasses will not only benefit native wildflowers, but will likely be necessary to establish new populati ons of many species of native wildflowers. While a fall application of glyphosate effectively increased the establishment of native wildflowers, the resulting plant community was rarely diverse. This study focused on the introduction of a single wildflower species into a P. notatum pasture because 1) the Florida Department of Transportation t ypically plants wildflower monoc ultures and 2) the source of competition between species is more apparent in two-species models than multi-species models. However, native plant communities in Florida ar e generally diverse and contain a suite of interacting species, especially in the understory. From a restorat ion perspective, introducing one

PAGE 116

116 native species to the site would likely not fulf ill the goals of the restoration project. An alternative to introducing wild flower monocultures for beautification is to restore a diverse native plant community to roadsides. Other states have successfully incorp orated restoration into roadside management (Houseal and Smith 2000, Brown and Bugg 2001, Markwardt 2005). There are ecological benefits to this approach (Ries et al. 200 1), but there may be practical benefits as well. For example, diverse plant co mmunities tend be more resilient to disturbance and stable over the long-term than less diverse communities or monocultures (Tilman et al. 2006). The lack of diversity observed in glyphosate treated plots in Chapters 3 and 4 is also likely due to the lack of propagules at the st udy sites, which had been under agricultural production before the study. In areas depaupera te of native propagules, introducing a suite of native species in addition to tr eating with glyphosate will likely be necessary to establish a diverse native plant communit y. In addition, eradicating P. notatum without subsequently establishing desirable species may result in the colonization by weedy, or possibly, invasive species (as seen in the scraped plots in Chap ter 3). Re-establishing the dominant native groundcover species seems essential to the establ ishment of an herbaceous, native community. In Florida, the dominant native groundcover species is usually a grass (Myers 1990). However, native grasses are rarely used in roadside planti ngs in Florida (Jenkins et al. 2004). Harper-Lore (2000) recommends including native grasses with wildflowers to visually compliment and physically give structure to the wildflower pl anting. While many state wildflower programs include grasses in wildflower plantings (Harper-Lore 2000), grasses and wildflowers are usually not seeded or planted at the same time in Flor ida (Florida Department of Transportation 1998).

PAGE 117

117 Based on the results of this study, recommenda tions for native wildfl ower establishment and management include: 1) limiting P. notatum establishment and occurrence to areas where erosion control is absolutely necessary, lik e areas closest to the road, 2) removing P. notatum where it is not necessary or beneficial, with two glyphosate applications in the fall, and 3) carefully timing roadside mowing or other post -planting disturbance for the phenology of each wildflower species. In additi on, recommendations for future res earch include: 1) studies on native wildflowers that can compete with aggr essive grasses including evergreen perennials and/or species that bloom in early spring, 2) sp ecies-specific seeding rates based on the number of live seeds per unit area as we ll as species life history traits, 3) continued research on native grasses appropriate for roadsides and restorat ion (see Jenkins et al. 2004), 4) restoration approaches to roadside management, and 5) studies that include long-term results of establishment and management methods.

PAGE 118

118 LIST OF REFERENCES Ada ms, C. R. and S. M. Galatowitsch. 2006. Incr easing the effectiveness of reed canary grass ( Phalaris arundinacea L.) control in wet meadow rest orations. Restoration Ecology 14: 441-451. Allen, P. S. and S. E. Meyer. 1998. Ecological aspects of seed dormancy loss. Seed Science Research 8:183-191. AOSA. 2006. Rules for testing seeds. Association of Official Seed Analysts, Stillwater, OK. Baker, R. D., L. B. McCarty, D. L. Colvin, J. M. Higgins, J. S. Weinbrecht, and J. E. Moreno. 1999. Bahiagrass ( Paspalum notatum ) seedhead suppression following consecutive yearly applications of plant grow th retardants. Weed Technology 13:378-384. Barbour, M. G., J. H. Buck, and W. D. P itts. 1987. Terrestrial plant ecology 2nd edition. Benjamin/Cummings Publishing Company, Menlo Park, CA. BASF Corporation. 2003. Plateau labe l. Research Triangle Park, NC. Baskin, C. C. and J. M. Baskin. 2001. Seed s: Ecology, biogeography, and evolution of dormancy and germination Academic Press, New York. Beatty, E. R. and J. D. Powell. 1978. Growth a nd management of Pensacola bahiagrass. Journal of Soil and Water Conservation 33:191-192. Beran, D. D., R. E. Gaussoin, and R. A. Mast ers. 1999. Native wildflower establishment with imidazolinone herbicides. Hortscience 34:283-286. Bissonettea, J. A. and W. Adairb. 2008. Restoring habitat permeability to roaded landscapes with isometrically-scaled wildlife crossings. Biological Conservation 141:482-488. Booth, B. D., S. D. Murphy, and C. J. Swant on. 2003. Weed ecology in natural and agricultural systems. CABI, Cambridge. Booth, D. T. and T. A. Jones. 2001. Plants for ec ological restoration. Native Plants Journal 2:1220. Brown, C. S. and R. L. Bugg. 2001. Effects of es tablished perennial gra sses on introduction of native forbs in California. Restoration Ecology 9:38-48. Brys, R., H. Jacquemyn, P. Endels, G. De Blus t, and M. Hermy. 2004. The effects of grassland management on plant performance and demography in the perennial herb Primula veris. Journal of Applied Ecology 41:1080-1091.

PAGE 119

119 Buisson, E., K. D. Holl, S. Anderson, E. Corcket, G. F. Hayes, F. Torre, A. Peteers, and T. Dutoit. 2006. Effect of seed source, topsoil removal, and plant neighbor removal on restoring California coastal prai ries. Restoration Ecology 14:569-577. Burke, M. J. W. and J. P. Grime. 1996. An e xperimental study of plant community invasibility. Ecology 77:776-790. Burton, C. M., P. J. Burton, R. Hebda, and N. J. Turner. 2006. Determining the optimal sowing density for a mixture of native plants us ed to revegetate degraded ecosystems. Restoration Ecology 14:379-390. Clark, C. J., J. R. Poulsen, D. J. Levey, and C. W. Osenberg. 2007. Are plant populations seed limited? A critique and meta-analysis of seed addition experiments. The American Naturalist 170:128-142. Clark, C. M. and D. Tilman. 2008. Loss of plan t species after chronic low-level nitrogen deposition to prairie gra sslands. Nature 451:713-715. Cole, D. N. 2007. Seedling establishment and surviv al on restored campsites in subalpine forest. Restoration Ecology 15:430-439. Collins, S. L., A. K. Knapp, J. M. Briggs, J. M. Blair, and E. M. Steinauer. 1998. Modulation of diversity by grazing and mowing in native tallgrass prairie. Science 280:745-747. Connell, J. H. 1978. Diversity in tropical rain forests and coral reefs. Science 199:1302-1310. Connolly, J. and P. Wayne. 2005. Assessing determinants of comm unity biomass composition in two-species plant competition studies. Oecologia 142:450-457. Coomes, D. A. and P. J. Grubb. 2000. Impacts of root competition in forests and woodlands: A theoretical framework and review of e xperiments. Ecological Monographs 70:171-207. Cordonnier, T., B. Courbaud, and A. Franc. 2006. The effect of colonization and competition processes on the relation betw een disturbance and diversity in plant communities. Journal of Theoretical Biology 243:1-12. Cox, A. C., D. R. Gordon, J. L. Slapcinsky, a nd G. S. Seamon. 2004. Understory restoration in longleaf pine sandhills. Natural Areas Journal 24:4-14. Dalrymple, G. H., R. F. Doren, N. K. OHare M. R. Norland, and T. V. Armentano. 2003. Plant colonization after complete and partial remova l of disturbed soils for wetland restoration of former agricultural fields in Ever glades National Park. Wetlands 23:1015-1029. Dayton, P. K. 1971. Competition, disturbance, and community organization: The provision and subsequent utilization of space in a rocky intertidal community. Ecological Monographs 41:351-389

PAGE 120

120 De Jong, T. J. and P. G. L. Klinkhamer. 1988. Seedling establishment of the biennials Cirsium vulgare and Cynoglossum officinale in a sand-dune area: th e importance of water for differential survival and growt h. The Journal of Ecology 76:393-402. del-Val, E. and M. J. Crawley. 2005. What limits herb biomass in grasslands: Competition or herbivory? Oecologia 142:202-211. Ejrnaes, R., H. H. Bruun, and B. J. Graae. 2006. Community assembly in experimental grasslands: Suitable environment or timely arrival? Ecology 87:1225-1233. Erickson, B. and N. E. Navarrete-Tindall. 2004. Missouri Native Ecotype Program: Increasing local-source native seed. Natural Areas Journal 24:15-22. Eriksson, O. 1992. Seed and microsite limitation of recruitment in plant populations. Oecologia 91:360-364. Eriksson, O. 2005. Game theory provides no explanat ion for seed size variation in grasslands. Oecologia 144:98-105. Ewel, J. J. and F. E. Putz. 2004. A place for alien species in ecosystem restoration. Frontiers in Ecology and the Environment 2:354 Fenner, M. 1978. A comparison of the abilities of colonizers and closed-turf species to establish from seed in artificial swar ds. Journal of Ecology 66:953-963. Fischbach, J. A., N. J. Ehlke, P. R. Peterson, D. L. Wyse, D. R. Swanson, and C. C. Sheaffer. 2006. Seeding rate affects establishment of native perennial legumes. Native Plants Journal 8:61-68. Flora of North America Editorial Committee, editor. 2006. Flora of North America North of Mexico. Oxford University Press, New York. Florida Department of Transportation. 1998. W ildflowers in Florida 4th edition. Florida Department of Transportation, Tallahassee. Florida Department of Transportation. 2005. District seven vegetation management plan. Florida Department of Transportation, Tallahassee. Forman, R. T. T. and L. E. Alexander. 1998. Ro ads and their major ecological effects. Annual Review of Ecology And Systematics 29:207-231. Forman, R. T. T., D. Sperling, J. A. Bissonette, A. P. Clevenger, and C. R. Goldman. 2003. Road ecology: Science and solutions. Island Press, Washington, D.C. Foster, B. L. and K. L. Gross. 1998. Species ri chness in a successional grassland: Effects of nitrogen enrichment and plant litter. Ecology 79:2593-2602.

PAGE 121

121 Freckleton, R. P. and A. R. Watkinson. 2000. De signs for greenhouse studies of interactions between plants: An analytical perspective. Journal of Ecology 88:386-391. Gann, G. D., K. A. Bradley, and S. W. Wood mansee. 2008. The floristi c inventory of south Florida database online. The Institute fo r Regional Conservation, Miami. World Wide Web site: www.regionalconservation .org [Date accessed: May 29, 2007]. Gelbard, J. L. and J. Belnap. 2003. Roads as conduits for exotic plant invasions in a semiarid landscape. Conservation Biology 17:420. Gerry, A. K. and S. D. Wilson. 1995. The influence of initial size on the competitive responses of 6 plant species. Ecology 76:272-279. Gibson, D. J., J. Connolly, D. C. Hartnett, a nd J. D. Weidenhamer. 1999. Designs for greenhouse studies of interactions between plants. Journal of Ecology 87:1-16. Goldberg, D. E. 1990. Components of resource competition in plant communities. Pages 27-49 in J. B. Grace and D. Tilman, editors. Perspectives on plant competition. Academic Press, Inc., San Diego. Gordon, D. R., M. J. Hattenbach, G. S. Seamon, K. Freeman, and D. A. Jones. 2000. Establishment and management of upland native plants on Florida roadsides. Final report to the Florida Department of Transportation, Tallahassee. Grabowski, J. 2005. Native wildflower seed production techniques in Missi ssippi. Native Plants Journal 6:73-75. Grace, J. B. and D. Tilman, editors. 1990. Persp ectives on plant competition. Academic Press, Inc., San Diego. Grimm, N. B., S. H. Faeth, N. E. Golubiewski, C. L. Redman, J. Wu, X. Bai, and J. M. Briggs. 2008. Global change and the ecology of cities. Science 319:756-760. Hansen, M. J. and A. P. Clevenger. 2005. The influence of disturba nce and habitat on the presence of non-native plant species along transport corridor s. Biological Conservation 125:249-259. Harkess, R. L. and R. E. Lyons. 1998. Establishm ent success and relative costs of four annual species for roadside pl anting. HortTechnology 8:583-585. Harper, J. L. 1977. Population biology of plants. Academic Press, London. Harper-Lore, B. L. 2000. Incorporating grasses into clear zones. Pages 21-22 in B. L. HarperLore and M. Wilson, editors. Roadside use of native plants. Isla nd Press, Washington, D.C.

PAGE 122

122 Harper-Lore, B. L. and M. Wilson, editors. 2000. Roadside use of native plants. Island Press, Washington, D.C. Heilman, G. E., J. R. Strittholt, N. C. Slosser, and D. A. Dellasala. 2002. Forest fragmentation of the conterminous United States: Assessing forest intactness through road density and spatial characteristics. Bioscience 52:411-422. Hely, S. E. L. and S. H. Roxburgh. 2005. The interactive effects of elevated CO2, temperature and initial size on growth and competition be tween a native C-3 and an invasive C-3 grass. Plant Ecology 177:85-98. Hoffmann, W. A., B. Orthen, and A. C. Franco. 2004. Constraints to seedling success of savanna and forest trees across the savanna -forest boundary. Oecologia 140:252-260. Hofmann, M. and J. Isselstein. 2004. Seedling recruitment on agricultu rally improved mesic grassland: the influence of disturbance and management schemes. Applied Vegetation Science 7:193-200. Holl, K. D., M. E. Loik, E. H. V. Lin, and I. A. Samuels. 2000. Tropical montane forest restoration in Costa Rica: overcoming barriers to dispersal and establishment. Restoration Ecology 8:339-349. Holmgren, M., M. Scheffer, and M. A. Huston. 1997. The interplay of facilitation and competition in plant communities. Ecology 78:1966-1975. Holzel, N. 2005. Seedling recruitment in flood-mea dow species: The effects of gaps, litter and vegetation matrix. Applied Vegetation Science 8:115-124. Holzel, N. and N. Otte. 2003. Restoration of a species-rich flood meadow by topsoil removal and diaspore transfer with plant material Applied Vegetation Science 6:131-140. Houseal, G. and D. Smith. 2000. Source-identifi ed seed: The Iowa roadside experience. Ecological Restoration 18:173-183. Isselstein, J., J. R. B. Tallowin, and R. E. N. Smith. 2002. Factors affecting seed germination and seedling establishment of fen-meadow species. Restoration Ecology 10:173-184. Jakobsson, A. and O. Eriksson. 2000. A comparative study of seed number, seed size, seedling size and recruitment in grassland plants. Oikos 88:494-502. Jenkins, A. M., D. R. Gordon, and M. T. Renda. 2004. Native alternatives for non-native turfgrasses in central Florida: Germina tion and responses to cultural treatments. Restoration Ecology 12:190-199.

PAGE 123

123 Jensen, K. and C. Meyer. 2001. Effects of light competition and litter on the performance of Viola palustris and on species composition and dive rsity of an abandoned fen meadow. Plant Ecology 155:169-181. Juenger, T. and J. Bergelson. 2000. Factors lim iting rosette recruitmen t in scarlet gilia, Ipomopsis aggregata : seed and disturbance lim itation. Oecologia 123:358-363. Jutila, H. M. and J. B. Grace. 2002. Effects of disturbance on germination and seedling establishment in a coastal prairie grassland: a test of the competi tive release hypothesis. Journal of Ecology 90:291-302. Kennedy, P. G. and W. P. Sousa. 2006. Forest en croachment into a Californian grassland: Examining the simultaneous effects of facilitation and competition on tree seedling recruitment. Oecologia 148:464-474. Korb, J. E., W. W. Covington, and P. Z. Ful 2003. Sampling techniques influence understory plant trajectories after restoration: An example from Ponderosa Pine restoration. Restoration Ecology 11:504-515. Lepik, M., J. Liira, and K. Zobel. 2005. High shoot plasticity favours plant coexistence in herbaceous vegetation. Oecologia 145:465-474. Lep, J. 1999. Nutrient status, disturbance and competition: An experimental test of relationships in a wet meadow. Journal of Vegetation Science 10:219-230. Li, C. Y., O. Junttila, P. Heino, and E. T. Pa lva. 2004. Low temperature sensing in silver birch ( Betula pendula Roth) ecotypes. Plant Science 167:165-171. Li, C. Y., Y. Q. Yang, O. Junttila, and E. T. Palva. 2005. Sexual differences in cold acclimation and freezing tolerance development in sea buckthorn ( Hippophae rhamnoides L.) ecotypes. Plant Science 168:1365-1370. Loux, M. M. and K. D. Reese. 1993. Effect of so il type and pH on persis tence and carryover of imidazolinone herbicides. Weed Technology 7:452-458. Lubchenco, J. and L. A. Real. 1991. Experimental manipulations in lab and field systems. Pages 715-733 in L. A. Real and J. H. Brown, editors Foundations of Ecology. The University of Chicago Press, Chicago. Markwardt, D. 2005. Texas roadside wild flowers. Native Plants Journal 6:69-71. Marousky, F. J. and F. Blondon. 1995. Red and far-red light influence carbon partitioning, growth and flowering of bahiagrass ( Paspalum notatum ). The Journal of Agricultural Science 125:355-359.

PAGE 124

124 Martin, L. M. and B. J. Wilsey. 2006. Assessing grassland restoration success: relative roles of seed additions and native ungulate activ ities. Journal of Applied Ecology 43:1098. Masters, R. A., S. J. Nissen, R. E. Gausso in, D. D. Beran, and R. N. Stougaard. 1996. Imidazolinone herbicides improve restora tion of Great Plains grasslands. Weed Technology 10:392-403. McBurney, M. E. 2005. Seed bank and vegetation re lationships in Ritchies vegetation zones on the Churchill River Estuary. Master's Th esis. The Universi ty of Winnipeg. Monaco, T. J., S. C. Weller, and F. M. As hton. 2002. Weed science prin ciples and practices Wiley, New York. Myers, R. L. 1990. Scrub and high pine. Pages 150-193 in R. L. Myers and J. J. Ewel, editors. Ecosystems of Florida. University of Central Florida Press, Orlando. Nathan, R., and H. C. Muller-Landau. 2000. Spatial patterns of seed dispersal, their determinants and consequences for recruitment. Trends in Ecology and Evolution 15:278. Norcini, J. G. and J. H. Aldrich. 2004. Establishment of native wildflower plantings by seed. Institute of Food and Agricultural Sciences, University of Florida. Norcini, J. G., J. H. Aldrich, and F.G. Marti n. 2001. Seed source effects on growth and flowering of Coreopsis lanceolata and Salvia lyrata Journal of Environmental Horticulture 19:212-215. Norcini, J. G., J. H. Aldrich, and F. G. Ma rtin. 2003. Tolerance of native wildflowers to imazapic. Journal of Environmental Horticulture 21:68-72. Oregon Department of Forestry. 2007. Seed zone maps. World Wide Web site: http://www.oregon.gov/ODF/FIELD/Nursery/Z oneMaps.shtml [Date accessed: May 29, 2007]. Orrock, J. L., M. S. Witter, and O. J. Reichman. in press Native consumers and seed limitation constrain the restoration of a native perenni al grass in exotic habitats. Restoration Ecology. Osorio, R. 2001. A Gardener's Guide to Florida's Native Plants. University Press of Florida Gainesville. Parr, T. W. and J. M. Way. 1988. Management of roadside vegetation: The long-term effects of cutting. The Journal of Applied Ecology 25:1073-1087. Perry, L. G., C. Neuhauser, and S. M. Galatowitsch. 2003. Founder control and coexistence in a simple model of asymmetric competition for light. Journal of Theoretical Biology 222:425-436.

PAGE 125

125 Pfaff, S. and M. A. Gonter. 1999. Direct seed ing native species on reclaimed phosphate-mined lands. in Second Eastern Native Grass Symposiu m. United States Department of Agriculture (USDA)-Agricultural Resear ch Service and USDA-Natural Resources Conservation Service, Baltimore. Radosevich, S. R. and M. L. Rousch. 1990. The ro le of competition in agriculture. Pages 341363 in J. B. Grace and D. Tilman, editors. Perspectives on plant competition. Academic Press, Inc., San Diego. Rees, M. 1997. Seed dormancy. Pages 214-238 in M. J. Crawley, editor. Plant ecology. Blackwell Science Ltd, Malden, MA. Rich, J. R., F. M. Rhoads, S. M. Olson, and D. O. Chellemi. 2003. A low input sustainable fresh market tomato production system. Institute of Food and Agricultural Sciences, University of Florida. Ricklefs, R. E. 1990. Ecology. 3rd edition. W.H. Freeman, New York. Ries, L., D. Debinski, and M. Weiland. 2001. Conser vation value of roadside prairie restoration to butterfly communities. Conservation Biology 15:401-411. Roberts, H. A. 1981. Seed banks in so il. Advances in Applied Biology 6:1-55. Sampaio, A. B., K. D. Holl, and A. Scariot. 2007. Does restoration en hance regeneration of seasonal deciduous forests in pastures in central Brazil? Restoration Ecology 15:462-471. Schiffers, K. and K. Tielborger. 2006. Ontogeneti c shifts in interactio ns among annual plants. Journal of Ecology 94:336-341. Scott, J. M. 1920. Bahiagrass. University of Florida Agricultural Experiment Stations, Gainesville, FL. Stampfli, A. and M. Zeiter. 1999. Plant species decline due to abandonm ent of meadows cannot easily be reversed by mowing. A case study from the southern Alps. Journal of Vegetation Science 10:151-164. Stevenson, B. A. and M. C. Smale. 2005. Seed be d treatment effects on vegetation and seedling establishment in a New Zealand pasture one year after seeding with native woody species. Ecological Restoration 6:124-131. Swain, H. M., P. J. Bohlen, K. L. Campbell, L. O. Lollis, and A. D. Steinman. 2007. Integrated ecological and economic analysis of ranch management systems: An example from south central Florida. Rangeland Ecology Management 60:1-11. Taylor, W. K. 1998. Florida wildflowers in th eir natural communities. University Press of Florida, Gainesville.

PAGE 126

126 Tilman, D. 1990. Constraints and tradeoffs toward a predictive theory of competition and succession. Oikos 58:3-15. Tilman, D. 1994. Competition and biodiversity in spatially struct ured habitats. Ecology 75:2-16. Tilman, D., J. Fargione, B. Wolff, C. D'Ant onio, A. Dobson, R. Howart h, D. Schindler, W. H. Schlesinger, D. Simberloff, and D. Swackhamer. 2001. Forecasting agriculturally driven global environmental change. Science 292:281-284. Tilman, D., P. B. Reich, and J. M. H. Knops 2006. Biodiversity and ecosystem stability in a decadelong grassland experiment. Nature 441:629-632. Trombulak, S. C. and C. A. Fr issell. 2000. Review of ecological effects of roads on terrestrial and aquatic communities. Conservation Biology 14:18-30. Turnbull, L. A., M. J. Crawley, and M. R ees. 2000. Are plant populations seed-limited? A review of seed sowing experiments. Oikos 88:225-238. United States Department of Agriculture. 2008. Plants Database. World Wide Web site: www.usdaplants.gov [Date accessed: June 14, 2008]. United States Department of Agriculture and Un ited States Department of the Interior. 2002. Interagency program to supply and manage na tive plant materials for restoration and rehabilitation on Federal lands. Uridel, K. W. 1994. Restoration of native herbs in abandoned Paspalum notatum (bahiagrass) pastures. Master's Thesis. University of Florida, Gainesville. Vergeer, P., E. Sonderen, and N. Ouborg. 2004. Intr oduction strategies put to the test: Local adaptation versus heterosis. Conservation Biology 18:812-821. Violi, H. 2000. Element Stewardship Abstract for Paspalum notatum Flgg. The Nature Conservancy, Arlington, VA. Washburn, B. E. and T. G. Barnes. 2000. Native warm-season grass and forb establishment using imazapic and 2,4-D. Native Plants Journal 1:61-69. Watson, V. H. and B. L. Burton. 1985. Bahiagrass carpetgrass and dallisgrass. Iowa State University Press, Ames, IA. Watts, R. D., R. W. Compton, J. H. McCammon, C. L. Rich, S. M. Wright, T. Owens, and D. S. Ouren. 2007. Roadless space of the conterminous United States. Science 316:736-738. Weigelt, A., T. Steinlein, and W. Beyschlag. 2002. Does plant competition intensity rather depend on biomass or on species identity ? Basic and Applied Ecology 3:85-94.

PAGE 127

127 Wildflower Seed and Plant Growers Associ ation Inc. 2008. World Wide Web site: www.floridawildflowers.com [Date accessed: June 14, 2008]. Williams, D. W., L. L. Jackson, and D. D. Sm ith. 2007. Effects of frequent mowing on survival and persistence of forbs seeded into a speci es-poor grassland. Restoration Ecology 15:24-33. Wilson, J. B. 1988. The effect of initial advant age on the course of plant competition. Oikos 51:19-24. Wunderlin, R. P. and B. F. Hansen. 2003. Guide to the vascular plants of Florida 2nd edition. University Press of Florida, Gainesville. Wunderlin, R. P. and B. F. Hansen. 2008. Atlas of Florida vascular plants. Institute for Systematic Botany, University of South Florida, Tampa. World Wide Web site: http://www.plantatlas.usf.edu [Date accessed: June 14, 2008]. Yates, C. J. and P. G. Ladd. 2005. Relative importance of reproductive biology and establishment ecology for persistence of a rare shrub in a fragmented landscape. Conservation Biology 19:239-249. Zimmerman, J. K., J. B. Pascarella, and T. M. Aide. 2000. Barriers to forest regeneration in an abandoned pasture in Puerto Ri co. Restoration Ecology 8:350-360. Zobel, M., M. Otsus, J. Liira, M. Moora, and T. Mols. 2000. Is small-scale species richness limited by seed availability or micr osite availability ? Ecology 81:3274-3282.

PAGE 128

128 BIOGRAPHICAL SKETCH Anne Frances com pleted her doctoral degree in Environmental Horticulture at the University of Florida. Her doctoral research was supported by a grant from the Florida Department of Transportation. She complete d a Master of Science degree in 2003 through Florida International Universitys Biology Department, with a focus on ethnobotany. She received a Bachelor of Arts degree in biology from the University of North Carolina Chapel Hill in 1996. Before starting her masters degree, Anne interned at The North Carolina Botanical Garden in Chapel Hill, NC (1997) and worked at the National Fish and Wildlife Foundation in Washington, DC (1998-2000). After completing her masters degree, Anne worked at Fairchild Tropical Botanic Garden and The Institute for Re gional Conservation in Miami, FL (2003-2004).