<%BANNER%>

Evaluation of Two Potential Biological Control Agents of Brazilian Peppertree (Schinus terebinthifolius) in Florida

Permanent Link: http://ufdc.ufl.edu/UFE0022038/00001

Material Information

Title: Evaluation of Two Potential Biological Control Agents of Brazilian Peppertree (Schinus terebinthifolius) in Florida
Physical Description: 1 online resource (123 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: biocontrol, episimus, invasive, pseudophilothrips, schinus, weed
Entomology and Nematology -- Dissertations, Academic -- UF
Genre: Entomology and Nematology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Brazilian peppertree, Schinus terebinthifolius Raddi (Anacardiaceae), introduced from South America, invades a variety of habitats in south and central Florida. The leaflet-roller moth Episimus utilis Zimmerman (Tortricidae) and the thrips Pseudophilothrips ichini Hood (Phlaeothripidae) were selected as potential biocontrol agents of Brazilian peppertree. The objective of this study was to examine the factors influencing the effectiveness of these two biocontrol agents against Brazilian peppertree in Florida. Results revealed that Brazilian peppertree genotypes did not affect the performance of E. utilis, but had a strong effect on the performance of two populations of P. ichini. The thrips haplotypes 2 and 3 were well adapted to all Florida genotypes and should be considered as biocontrol agent of Brazilian peppertree. Episimus utilis performed well when reared on plants grown in fresh-water environments, and also when fed on high-nutrient treatments. Therefore, Brazilian peppertrees growing in upland habitats and high-fertility soils (e.g., abandoned farms) will provide high-quality hosts for E. utilis establishment. In addition, temperature-dependent models and GIS mapping predicted a range of 5.8 to 9.7 generations per year for E. utilis throughout Florida. Based on the isothermal lines of the pupal stage (lethal exposure times), establishment of E. utilis may occur throughout Florida, Hawaii, and southern parts of California, Texas and Arizona. Brazilian peppertree seedlings exposed to high levels of defoliation (30 larvae/plant) suffered a significant reduction in number of leaflets, plant height, foliar biomass, shoot: root ratio, and relative growth rate compared to control plants (no herbivory). In addition, these plants were not able to compensate for insect damage two months after herbivory. However, low levels of defoliation (15 larvae/plant) did not affect Brazilian peppertree growth and biomass allocation in the laboratory. Therefore, the effectiveness of E. utilis to suppress this weed will vary in relation to insect densities present in the field. The results obtained in this Dissertation provide a better understanding of the interactions between two biocontrol agents and Brazilian peppertree. This information will be used to select suitable sites for field releases, and assist in developing an effective long-term control of Brazilian peppertree in different habitats in Florida.
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.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Cuda, James P.
Local: Co-adviser: Overholt, William A.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022038:00001

Permanent Link: http://ufdc.ufl.edu/UFE0022038/00001

Material Information

Title: Evaluation of Two Potential Biological Control Agents of Brazilian Peppertree (Schinus terebinthifolius) in Florida
Physical Description: 1 online resource (123 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: biocontrol, episimus, invasive, pseudophilothrips, schinus, weed
Entomology and Nematology -- Dissertations, Academic -- UF
Genre: Entomology and Nematology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Brazilian peppertree, Schinus terebinthifolius Raddi (Anacardiaceae), introduced from South America, invades a variety of habitats in south and central Florida. The leaflet-roller moth Episimus utilis Zimmerman (Tortricidae) and the thrips Pseudophilothrips ichini Hood (Phlaeothripidae) were selected as potential biocontrol agents of Brazilian peppertree. The objective of this study was to examine the factors influencing the effectiveness of these two biocontrol agents against Brazilian peppertree in Florida. Results revealed that Brazilian peppertree genotypes did not affect the performance of E. utilis, but had a strong effect on the performance of two populations of P. ichini. The thrips haplotypes 2 and 3 were well adapted to all Florida genotypes and should be considered as biocontrol agent of Brazilian peppertree. Episimus utilis performed well when reared on plants grown in fresh-water environments, and also when fed on high-nutrient treatments. Therefore, Brazilian peppertrees growing in upland habitats and high-fertility soils (e.g., abandoned farms) will provide high-quality hosts for E. utilis establishment. In addition, temperature-dependent models and GIS mapping predicted a range of 5.8 to 9.7 generations per year for E. utilis throughout Florida. Based on the isothermal lines of the pupal stage (lethal exposure times), establishment of E. utilis may occur throughout Florida, Hawaii, and southern parts of California, Texas and Arizona. Brazilian peppertree seedlings exposed to high levels of defoliation (30 larvae/plant) suffered a significant reduction in number of leaflets, plant height, foliar biomass, shoot: root ratio, and relative growth rate compared to control plants (no herbivory). In addition, these plants were not able to compensate for insect damage two months after herbivory. However, low levels of defoliation (15 larvae/plant) did not affect Brazilian peppertree growth and biomass allocation in the laboratory. Therefore, the effectiveness of E. utilis to suppress this weed will vary in relation to insect densities present in the field. The results obtained in this Dissertation provide a better understanding of the interactions between two biocontrol agents and Brazilian peppertree. This information will be used to select suitable sites for field releases, and assist in developing an effective long-term control of Brazilian peppertree in different habitats in Florida.
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.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Cuda, James P.
Local: Co-adviser: Overholt, William A.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022038:00001


This item has the following downloads:


Full Text

PAGE 1

1 EVALUATION OF TWO POTENTIAL BIOLOGICAL CONTROL AGENTS OF BRAZILIAN PEPPERTREE (SCHINUS TEREBINTHIFOLIUS) IN FLORIDA By VERONICA MANRIQUE 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 Veronica Manrique

PAGE 3

3 To my life companion, Rodrigo Diaz for always being by my side

PAGE 4

4 ACKNOWLEDGMENTS I would like to acknowledge m y major advisor Dr. J. P. Cuda, and my co-advisor Dr. W. A. Overholt for their constant support and help during my research. I am also grateful to my committee members Drs. H. McAuslane and A. Fox for their contributions and advice during my doctoral studies. Special thanks to all the people from the weed biological control laboratory (University of Florida) who help ed me in different ways during my research: J. Gillmore, J. Medal, O. Moeri, F. Wessels, R. Diaz, J. Ma rkle, L. Markle, Y. Valenzuela, F. Soza, D. Gonzalez, A. Samayoa, and B. Anuforom. I w ould like to acknowledge D. Williams (Texas Christian University) for conducting the genetic analysis, and the collab orators in Brazil for providing the insects: H.J. Pedrosa-Macedo (Feder al University of Parana, Brazil), and M.D. Vitorino (University of Blumenau, Brazil). This project was supporte d by grants from the Florida Department of Environmental Protection, South Florida Water Management District, Florida Exotic Pest Plant Council, and the Smithsonian Ma rine Station at Ft. Pierce. My parents, Ana and Carlos, have been always by my side duri ng this journey, and their personal achievements are a source of inspiration for my everyday life. I am also grateful to my brother and sister, Gonzalo and Paula, for their support at all times. I am very thankfull to Rodrigo for sharing with me his excitment and constant motivation for lear ning and science. Finally, special thanks to my loving niece, Josefina, for her enthusiasm and company during my vacations in Argentina.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........8 LIST OF FIGURES.........................................................................................................................9 ABSTRACT...................................................................................................................................10 CHAP TER 1 LITERATURE REVIEW.......................................................................................................12 Introduction................................................................................................................... ..........12 Study Overview................................................................................................................. .....14 Taxonomy and Molecular Phylogenetics of Schin us terebinthifolius Raddi..........................15 Biology and Ecology of Schinus terebinthifolius Raddi ......................................................... 17 Biology and Ecology of Biological Control Candidates........................................................19 Selecting Effective Biological Control Agents....................................................................... 21 Goals and Hypotheses........................................................................................................... ..24 2 EFFECT OF HOST-PLANT GENOTYPES ON THE PE RFORMANCE OF TWO BIOLOGICAL CONTROL AGENTS OF BRAZILIAN PEPPERTREE............................. 31 Introduction................................................................................................................... ..........31 Materials and Methods...........................................................................................................32 Insects and Plants............................................................................................................ 32 Life History Parameters of E. utilis on Different Brazilian Peppertree Genotypes ........34 Life History Parameters of P. ichini on Different Brazilian P eppertree Genotypes and Schinus molle ........................................................................................................35 Feeding Preference of P. ichini Adults on Different Br azilian Peppertree Genotypes and Schinus molle ........................................................................................................36 Data Analysis...................................................................................................................36 Results.....................................................................................................................................37 Life History Parameters of E. utilis on Different Brazilian Peppertree Genotypes ........37 Life History Parameters of P. ichini on Different Brazilian P eppertree Genotypes and Schinus molle ........................................................................................................38 Feeding Preference of P. ichini Adults on Different Br azilian Peppertree Genotypes and Schinus molle ........................................................................................................39 Discussion...............................................................................................................................39

PAGE 6

6 3 INFLUENCE OF HOST-PLANT QUALITY ON THE PERFORMANCE OF THE BIOCONTROL AGENT EPISIMUS UTILIS .......................................................................49 Introduction................................................................................................................... ..........49 Materials and Methods...........................................................................................................51 Insect Rearing..................................................................................................................51 Life History Parameters of E. utilis on Brazilian Peppertr ee Exposed to Different Salinity Levels .............................................................................................................51 Life History Parameters of E. utilis on Brazilian Peppertr ee Exposed to Different Nutrient Levels.............................................................................................................53 Data Analyses..................................................................................................................54 Results.....................................................................................................................................54 Life History Parameters of E. utilis on Brazilian Peppertr ee Exposed to Different Salinity Levels .............................................................................................................54 Life History Parameters of E. utilis on Brazilian Peppertr ee Exposed to Different Nutrient Levels.............................................................................................................55 Discussion...............................................................................................................................56 4 TEMPERATURE-DEPENDENT DE VELOPMENT AND POTENTIAL DISTRIBUTION OF THE BIOCONTROL AGENT EPISI MUS UTILIS........................... 66 Introduction................................................................................................................... ..........66 Materials and methods.......................................................................................................... ..68 Plants and Insects............................................................................................................ 68 Survival and Developmental Times Calculations........................................................... 68 Temperature Thresholds and Degree-days Calculations ................................................. 69 Cold Tolerance of E. utilis ...............................................................................................70 GIS Maps to Predict Number of Generations of E. utilis in the USA ............................. 71 Results.....................................................................................................................................72 Survival and Developmental Time Calculations............................................................. 72 Temperature Thresholds and Degree-day Calculations ..................................................73 Cold Tolerance of E. utilis ...............................................................................................73 GIS Maps to Predict Number of Generations of E. utilis in the USA ............................. 74 Discussion...............................................................................................................................74 5 EFFECT OF HERBIVORY ON GROWTH AND BIOM ASS OF BRAZILIAN PEPPERTREE SEEDLINGS.................................................................................................88 Introduction................................................................................................................... ..........89 Materials and Methods...........................................................................................................91 Plants and Insects............................................................................................................ 91 Experimental Procedure.................................................................................................. 91 Data Analysis...................................................................................................................92 Results.....................................................................................................................................93 Discussion...............................................................................................................................94

PAGE 7

7 6 CONCLUDING REMARKS................................................................................................ 100 LIST OF REFERENCES.............................................................................................................105 BIOGRAPHICAL SKETCH.......................................................................................................123

PAGE 8

8 LIST OF TABLES Table page 2-1 Life history parameters (mean SE) of E. utilis on different Brazilian peppertree Florida genotypes.. .............................................................................................................44 2-2 Life history parameters (m eans SE) of two populations of P. ichini reared on different host plants............................................................................................................45 3-1 Life history parameters of E. utilis (m eans SE) reared on Brazilian peppertree exposed to different salinity levels.................................................................................... 60 3-2 Leaflet nutrient contents (m eans SE) of Brazilian peppe rtree exposed to different salinity levels................................................................................................................ .....61 3-3 Plant parameters (means SE) of Brazilian peppertree exposed to different salinity levels ..................................................................................................................................62 3-4 Life history parameters of E. utilis (m eans SE) reared on Brazilian peppertree exposed to different nutrient levels.................................................................................... 63 3-5 Leaflet nutrient contents (m eans SE) of Brazilian peppe rtree exposed to different nutrient levels.....................................................................................................................64 3-6 Plant parameters (means SE) of Brazilian pepperrtee exposed to different nutrient levels.. ................................................................................................................................65 4-1 Mean ( SE) developmental time (days) of imm ature stages of E. utilis at five constant temperatures......................................................................................................... 80 4-2 Linear regression parameters estimate s describing the relationship between tem perature and developmental rate (1/D) of E. utilis stages............................................ 81 4-3 Logan non-linear model parameters for developmental rate of E. utilis ..........................82 4-4 Mean ( SE) percentage survival of eggs, larvae, and adults of E. utilis at different tem peratures and exposure times (days)............................................................................ 83 5-1 Morphometric plant parameters (mean SE) of Brazilian peppertree exposed to different levels of herbivory...............................................................................................97 5-2 Biomass allocation (g) (means SE) of Brazilian peppertree exposed to different levels of herbivory.............................................................................................................98

PAGE 9

9 LIST OF FIGURES Figure page 1-1 Brazilian peppertree in Florida.......................................................................................... 28 1-2 Episimus utilis .. ..................................................................................................................29 1-3 Pseudophilothrips ichini ...................................................................................................30 2-1 Brazilian peppertree geno types present in Florida............................................................. 46 2-2 Diagram showing the set-up used for t he feeding pr eference trial of P. ichini .................47 2-3 Number of adults (mean SE) of P. ichini feeding on each host plant after a 4-hour trial.. ...................................................................................................................................48 4-1 Percentage survival from larvae to adult of E. utilis at seven constant tem peratures........ 84 4-2 Developmental rates (Days -1) of each immature stage of E. utilis ...................................85 4-3 Pupal survival of E. utilis at different exposure tim es and temperatures (10, 5 and 0 C)....................................................................................................................................86 4-4 Predicted number of generations per year for E. u tilis in the USA................................... 87 4-5 Map showing the isothermal lines (Ltime50 and Ltime90) at 0 and 5C for E. utilis in the USA..............................................................................................................................88 5-1 Relative growth rate (RGR) (means SE) of leaflets and stems of Brazilian peppertree exposed to different levels of herbivory. ......................................................... 99

PAGE 10

10 Abstract of Dissertation Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy EVALUATION OF TWO POTENTIAL BIOLOGICAL CONTROL AGENTS OF BRAZILIAN PEPPERTREE (SCHINUS TEREBINTHIFOLIUS) IN FLORIDA By Veronica Manrique May 2008 Chair: James P. Cuda Cochair: William A. Overholt Major: Entomology and Nematology Brazilian peppertree, Schinus terebinthifolius Raddi (Anacardiaceae), introduced from South America, invades a variety of habitats in south and central Florida. The leaflet-roller moth Episimus utilis Zimmerman (Tortricidae) and the thrips Pseudophilothrips ichini Hood (Phlaeothripidae) were selected as potential biocontrol agents of Brazilian peppertree. The objective of this study was to examine the factors influencing the effectiveness of these two biocontrol agents against Brazilian peppertree in Florida. Results reve aled that Brazilian peppertree genotypes did not affect the performance of E. utilis, but had a strong effect on the performance of two populations of P. ichini The thrips haplotypes 2 and 3 were well adapted to all Florida genotypes and should be considered as biocontrol agent of Br azilian peppertree. Episimus utilis performed well when reared on plants grown in fresh-water environments, and also when fed on high-nutrien t treatments. Therefore, Brazilian peppertrees growing in upland habitats and high-fertility soils (e.g., ab andoned farms) will provide high-quality hosts for E. utilis establishment. In addition, temperature-de pendent models and GIS mapping predicted a range of 5.8 to 9.7 generations per year for E. utilis throughout Florida. Ba sed on the isothermal

PAGE 11

11 lines of the pupal stage (lethal e xposure times), establishment of E. utilis may occur throughout Florida, Hawaii, and southern parts of California, Texas and Arizona. Brazilian peppertree seedlings exposed to high levels of defoliation (30 larvae/plant) suffered a significant reduction in number of leaflets, plant height, foliar biomass, shoot: root ratio, and relative growth rate compared to contro l plants (no herbivory). In addition, these plants were not able to compensate for insect damage two months after herbivory. However, low levels of defoliation (15 larvae/plant) di d not affect Brazilian peppertr ee growth and biomass allocation in the laboratory. Therefor e, the effectiveness of E. utilis to suppress this weed will vary in relation to insect densities present in the field. The results obtained in this Dissertation provide a better und erstanding of the interactions between two biocontrol agents and Brazilian pepper tree. This information will be used to select suitable sites for field releases and assist in developing an effective long-term control of Brazilian peppertree in differe nt habitats in Florida.

PAGE 12

12 CHAPTER 1 LITERATURE REVIEW Introduction Biological invasions are widely recognized as a serious threat to m any natural and agricultural ecosystems worldw ide (e.g., Simberloff 1996, Parker et al. 1999, Mack et al. 2000, Myers and Bazely 2003, Inderjit 2005). People have been moving plants and animals to different places for many centuries, but it was not until th e mid 1900s that scientists began to study the environmental risk associated with nonindigenou s species (Inderjit 2005). Since then, there have been a number of studies on the biology a nd ecology of nonindigenous species aimed at improving the understanding of the processes and dynamics of invasions (e.g., Elton 1958, Williamson 1996, Mooney and Hobbs 2000, Inderjit et al. 2005, Richardson and Pysek 2008). Many hypotheses were developed in or der to explain the establishmen t and spread of species into new areas (e.g., Elton 1958, Tilman 1982, Davi s and Thompson 2000, Blossey and Notzold 1995, Ellstrand and Schierenbeck 2000, Kean e and Crawley 2002). The Enemy Release hypothesis (ERH) predicts the rapi d increase in abundance and distribution of plant species when released from co-adapted natural enemies in the introduced range (Williams 1954, Williamson 1996, Keane and Crawley 2002). A variant of this hypothesis is the Evolution of Increased Competitive Ability (EICA) hypothesis, which predicts that an invader can reallocate resources from defense to growth and reproduction when rel eased from their specialized natural enemies in the introduced range (Blossey and Notzold 1995). T hus, there is an increase in the competitive ability of introduced individuals compared to na tive species. Evidence both for and against ERH and EICA can be found in the literature (e .g. Willis et al. 2000, Siemann and Rogers 2001, Agraval and Kotanen 2003, Blair and Wolfe 2004, Cappuccino and Carpenter 2005, Genton et al. 2005, Joshi and Vrieling 2005).

PAGE 13

13 Invasive species are an incr easing problem in many ecosystem s, and introduced species are estimated to cause annual economic losses of $120 billion in the USA (Pimentel et al. 2005). Different management strategies have been em ployed against invasive plant species (e.g., Hobbs and Humphries 1995, Myers and Bazely 2003, Baker and Wilson 2004, Emery and Gross 2005, Hulme 2006). Classical biological control of weeds involves the intentional intr oduction of hostspecific natural enemies (e.g., insect herbivores, pathogens) from th e native range to suppress the target weed (Harris 1991, va n Driesche and Bellows 1996, Nordlund 1996, DeLoach 1997, Cuda 2004). The reunion of co-adapted natural enemies with their host plants may result in a reduction of weed population density in the introduced range in accordance with th e ERH. The steps in a typical weed biological control programs according to Harley and Forno (1992) are: 1) selection of the target weed, 2) literature review (t axonomy, biology, ecology of the target weed), 3) foreign explorations (search for potential agents in the native range), 4) ecological studies of the weed and natural enemies, 5) preparation of import permits to transport candidates to quarantine facilities, 6) host-specificity testing (no-choice and choice testing), 7) ma ss rearing and release, 8) evaluation and monitoring of agent establishmen t and subsequent control of the target weed, and 9) distribution of agents to new sites. Even though it often takes 10 years or longer to achieve, biological control is an environmentally friendly and sustainabl e approach which may be the best alternative for controlling the spre ad of many invasive pl ant species (Strong and Pemberton 2000, Myers and Bazely 2003). Classical biological control has a long history of success in both terrestrial and aquatic ecosy stems (e.g., Hartley 1990, Julien and Griffith 1998, Grevstad 2006, Barton et al. 2007). Wellknown examples include suppression of Opuntia cacti by Cactoblastis cactorum Berg. in Australia in 1926 (Dodd 1940, Hosking et al. 1988, Julien and

PAGE 14

14 Griffiths 1998), and the control of Hypericum perforatum L. by Chrysolina beetles in the western USA (Huffaker and Kennett 1959, Julien and Griffiths 1998). Study Overview Hawaii and Florida of ten are cited as the two states most vulnerable to invasion in the USA, with over 4000 and 925 intr oduced species, respectively (Cox 1999). The establishment and spread of invasive plant species poses a serious threat to many ecosystems in Florida. Prominent examples include paperbark tree Melaleuca quinquenervia Cavanilles (Myrtaceae) invading Florida wetlands (Bodle et al. 1994, Turner et al. 1998), Australian pine Casuarina equisetifolia L. (Casuarinaceae) growing along coas tal dunes (Morton 1980, Gordon 1998), Old World climbing fern Lygodium microphyllum (Cav.) (Lygodiaceae) that forms large mats covering native species (Nauman and Austin 1978, Pemberton and Ferri ter 1998), and Brazilian peppertree Schinus terebinthifolius Raddi (Anacardiaceae) that is one of the most widespread invaders, especially in the Everglades (Cuda et al. 2006). Brazilian peppertree is an intr oduced perennial plant that ha s become widely established throughout central and south Florid a (Fig. 1-1) (Cuda et al. 2006) This species is native to Argentina, Brazil and Paraguay (Barkley 1944, 1957), and was brought to Florida as an ornamental in the 1840s (Mack 1991, Cuda et al. 2006). In the USA, Brazilian peppertree occurs in Hawaii, California, Arizona, Texas and Fl orida (Habeck et al. 1994, Cuda et al. 2006). Although Brazilian peppertree is still grown as an ornamental in some states (e.g., California, Texas, and Arizona), this plant has been recogn ized as an invasive species in Hawaii (Morton 1978, Randall 1993), California (Randall 2000), a nd Texas (Gonzalez and Christoffersen 2006). In Florida, Brazilian peppertree is considered to be one of the worst invasive species by the Florida Exotic Pest Council, and is recognized as one of the most widespread exotic plants in the state (Cuda et al. 2006).

PAGE 15

15 Although herbicides and mechanical methods (e.g., cutting, burning and flooding) are routinely used for controlling existing Brazilian peppertree st ands (Gioeli and Langeland 1997, Cuda et al. 2006), these methods are extremely labor intensive and can be very expensive, especially for large infestations In addition, chemical and mechan ical controls are unsuitable for some natural areas (e.g., mangrove forests) because they may have negative effects on non-target species (Doren and Jones 1997) and may increas e water pollution. Ther efore, alternative methods for long-term control of Brazilian pe ppertree are urgently needed in Florida. Brazilian peppertree has been recognized as a target for classical biological control since the 1980s. Several phytophagous ins ects and a fungal pathogen were identified from exploratory surveys conducted in northern Ar gentina and southeastern Brazil as potential biological control agents because they clearly damaged the plant a nd appeared to be host specific in their native range (Bennett et al. 1990, Bennett and Habeck 1991, Habeck et al. 1994, Cuda et al. 2004). The following are two promising natura l enemies of Brazilian peppertree that are currently being studied in Florida quarantine laborat ories: 1) the leaflet roller moth Episimus utilis Zimmerman (Lepidoptera: Tortricidae), and 2) th e shoot and flower attacking thrips Pseudophilothrips ichini (= Liothrips ichini ) Hood (Thysanoptera: Phlaeothripidae). Taxonomy and Molecular Phylogenetics of Schinus terebinthifolius Raddi The taxonomy of Brazilian peppertree follows the higher classification scheme published by Mabberley (1997): Kingdom Plantae Division Magnoliophyta Class Dicotyledonae (Magnoliopsida) Subclass Rosidae Order Sapindales Family Anacardiaceae Tribe Rhoeae Genus Schinus L. Subgenus Euschinus Species Schinus terebinthifolius Raddi 1820

PAGE 16

16 The genus Schinus is native to South America an d includes approximately 28 species (Barkley 1944, 1957). Four Schinus species have been introduced into the continental United States: S. longifolius (Lindl.) Spreg. in Texas, S. molle L. (type species) and S. polygamus (Cav.) Cabrera in California, and Brazilian peppertree in California, Florida, Arizona and Texas. Although S. molle is still a popular ornamental, the Ca lifornia Exotic Pest Plant Council has listed this species as a category B invasive species (Cal-IPC 2006). Barkley (1944) described five varieties of Brazilian peppertree in South America based solely on morphological characters. Of the five recognized varieties, three have been in troduced into the United States: S. terebinthifolius var. acutifolius Engl. in California; S. terebinthifolius var. terebinthifolius Raddi in California, Florida, Ha waii and Puerto Rico; and S. terebinthifolius var. raddianus Engl. in Florida and Puerto Rico (Barkl ey 1944). However, many of th e morphological characteristics used to distinguish these differe nt varieties in Florida are broadly overlapping and it is unclear how well differentiated these varieties are in the native range (Cuda et al. 2006). In order to determine the origin and pattern s of introduction of Brazilian peppertree in Florida, Williams et al. (2005, 2007) collected Brazilian peppertree samples from the introduced range in the USA (Florida, Hawaii, Texas, US Virgin Islands, California) and from the native range (Brazil, Argentina, Para guay). These samples were used to characterize six polymorphic nuclear microsatellite loci in Brazilian peppertree, and a region of the chloroplast DNA (cpDNA) was sequenced (Williams et al. 2002, 2005). Tw o cpDNA haplotypes were found in Florida; haplotype A is more common on the west coast while haplotype B is more common on the east coast (Williams et al. 2005, 2007). Nuclear microsate llite DNA analyses revealed that extensive hybridization has occurred between these two types of plants since arriving in Florida (Williams et al. 2005, 2007). In addition, gene tic studies in the native ra nge identified eleven cpDNA

PAGE 17

17 haplotypes (A-K) of Brazilian peppertree in Br azil, and haplotype D is the most common and widespread (Williams et al. 2005, unpub. data). Th e source location of haplotype A has been found along the coast of southeastern Brazil, and the haplotype B has recently been located in the coast of Bahia in northeastern Brazil (Williams et al. 2005, unpub. data). Biology and Ecology of Schinus terebinthifolius Raddi The invasio n and displacement of native speci es by Brazilian peppertree poses a serious threat to biodiversity in many ecosystems of Florida (Morton 1978). Brazilian peppertree is found mainly in disturbed sites such as highway rights-of-way, canals, fallow farmlands, but it also invades natural communities including pinelands, hardwood hammocks, and mangrove forests (Cuda et al. 2006). Recent estimates based on aerial surveys indicate that approximately 2833 km2 of all terrestrial ecosystems in central a nd south Florida have been invaded by this weed (Cuda et al. 2006). Several at tributes of this plant contribut e to its invasiveness, including a large number of fruits produced per female plant, an effective mechanism of dispersal by birds (Panetta and McKee 1997), tolerance to shade (E wel 1978), fire (Doren et al. 1991), and drought (Nilson and Muller 1980b), allelopa thic effects on neighboring pl ants (Gogue et al. 1974, Nilson and Muller 1980a, Morgan and Overholt 2005), and tolerance to salin e conditions (Ewe 2001, Ewe and Sternberg 2002). Brazilian peppertree is dioecious, and the ma in flowering period in Florida occurs from September to October with a much-reduced bl oom occurring from March to May (Ewel et al. 1982). Numerous small, white flowers occur in dense axillary panicles near the end of branches. Flowers produce abundant amounts of pollen and ne ctar, and are primarily insect-pollinated. A massive number of bright red fruits are typica lly produced on the plants from November to February. The fruits are eaten and dispersed primarily by birds and mammals although some dispersal occurs by gravity or water (Ewel et al. 1982). Brazilian peppertree also is capable of

PAGE 18

18 resprouting from above-ground stems and crowns after damage from cutting, fire, or herbicide treatment. Resprouting and sucker ing are often profuse and the gr owth rates of the sprouts are high, which contribute to the formation of de nse clumps (Woodall 1979, Cuda et al. 2006). The reproductive potential of Brazilian peppe rtree is enormous with female trees producing thousands of seeds every year. Seed germination occurs from November to April, but mainly from January to February, and seed viability ranges from 3060% (Ewel et al. 1982). Seedlings are able to tolerate a broad range of extreme soil moisture conditions (Ewel 1978), and survival of established seedlings ranges from 66-100% (Ewel et al. 1982 ). Although the leaflets are present on Brazilian peppertree plants year round, vegetative growth becomes dormant in winter (October to December), corresponding to the flowering period. Under optimal growing conditions, Brazilian peppertree is capable of producing seeds three years after germination (Ewel et al. 1982). Several studies have shown that the success of Brazilian peppertree as an invader in a variety of habitats in south Flor ida is a consequence of its capacity to tolerate and persist during adverse conditions. Greenhouse studies have show n that the growth rates and morphology of Brazilian peppertree are less affected by salinity than white mangrove ( Laguncularia racemosa L.) or some other native freshwater species (Ewe 2001). Brazilian peppertree also was less affected by seasonality and more tolerant to root flooding than so me native plant species growing in the Everglades National Park (Ewe and Sternberg 2002). In addition, total concentration of nutrients (N, P, Zn, and Cu) was higher in distur bed soils (e.g., previous farmed land) compared to undisturbed soils in the Everglades (Li and Nordland 2001). The high correlations between Brazilian peppertree leaflet P and plant-available P in soils suggest that P enrichment in farmed soils facilitated the invasion of this area by Brazilian peppe rtree (Li and Nordland 2001).

PAGE 19

19 Biology and Ecology of Biol ogical Control Candidates The leaf let roller moth E. utilis is native to southeastern Brazil (Zimmerman 1978). The larval stage usually has five instars but occasi onally six; larvae feed on Brazilian peppertree leaflets and can completely defo liate small plants (Martin et al 2004). First to th ird instars tie together young leaflets with silk to feed, while ol der instars feed inside a cylindrical rolled leaflet (Fig. 1-2a). The dimorphic adults are small, gr ayish brown moths with distinctive wing patterns that readily separate males from females (F ig. 1-2b) (Zimmerman 1978). Females oviposit up to 172 eggs during their lifetime (Martin et al. 2004 ). Adults lived on average 7 days, and one generation was completed in 43 days at 22C (Martin et al. 2004). Because E. utilis completes its entire life cycle in the canopy of the host plant, this agent may be an appropriate biological control agent against Brazilia n peppertree growing in sites exposed to seasonal flooding in Florida. In the 1950s, E. utilis was released and established in Hawaii but successful control of Brazilian peppertree populations was not ach ieved (Goeden 1977, Yoshioka and Markin 1991, Julien and Griffiths 1998). Factors such as unfavor able abiotic and biotic conditions present in Hawaii may explain this outcome. For example, high larval mortality of E. utilis caused by introduced and native parasitoids and predator s has been recorded (Davis 1959, Krauss 1963), which probably reduced the effectiveness of this biocontrol agent. Host specificity studies for E. utilis were conducted in Brazil and in the Entomology and Nematology Departments containment facilities Gainesville, FL. In total 61 plants species belonging to 23 plant families were tested (J. P. Cuda, unpubl. data). No-c hoice tests showed that larval feeding damage was restricted to plan ts in the family Anacardiaceae, and Brazilian peppertree suffered the greatest damage. Some development to the adult stage also occurred on European smoketree ( Cotinus coggygria Scop.), Chinese pistache ( Pistacia chinensis Bunge), cultivated pistachio ( Pistacia vera L.), fragrant sumac ( Rhus aromatica Ait.), winged sumac

PAGE 20

20 ( Rhus copallinum L.), poison sumac ( Toxicodendron vernix L.), poisonwood ( Metopium toxiferum L.), and cashew ( Anacardium occidentale L.), which demonstrated that these plant species are in the physiological host range of the insect. Howeve r, choice tests showed that E. utilis clearly preferred Brazilian pe ppertree over the aforementione d non-target species. Even though E. utilis attacks non-target plant species in the laboratory, this has not been observed in the native range or in Hawaii (Cuda et al. 2006). A petition for the release E. utilis in Florida for biological control of Brazilian peppertree is in preparation. The thrips P. ichini is native to southeastern Braz il (Hood 1949), and has been found feeding only on Brazilian peppertree in its nati ve range (Garcia 1977). Th rips feeds by punching their food substrate with the so litary mandibule (left mandibule onl y developed) and sucking the cell contents through the maxillary stylets (M ound 2005). Both larvae and adult stages of P. ichini severely damage the host plant; larval f eeding usually kills the growing shoot tips and adult feeding may cause flower a bortion (Fig. 1-3) (Garcia 1977). Im mature thrips feed on plants during two larval stages, and the remainder of the immature life cycle occurs in the soil. Thrips belonging to the family Phlaeo thripidae are characterized by having three non-feeding pupal instars (the propupa, pupa I and pupa II) instead of two seen in other families of thrips (Mound and Marullo 1996). The adults are black and winge d, and the larvae are mostly red and wingless. Females require a pre-oviposition period (5-15 days), and can lay on average 220 eggs during their lifetime (Garcia 1977). Th ey exhibit arrhenot okous reproduction whereby unmated females produce only male progeny (Mound and Marullo 1996). The entire life cycle from egg to egg can be completed in 38 days at 24C in controlled laboratory conditions (Garcia 1977). In Brazil, P. ichini is polyvoltine and can produce up to f our generation per year (Garcia 1977).

PAGE 21

21 Studies of the biology an d host specificity of P. ichini were conducted in Brazil and in Florida quarantine; 37 plant species belonging to 10 families were tested (Garcia 1977, Cuda el al. 2002). No-choice tests showed that thrips larvae fed and comp leted their development only on Brazilian peppertree and California peppertree ( Schinus molle L.). California peppertree is a popular ornamental in California, but the California Invasive Plan t Council recently listed this tree as invasive (Cal-IPC 2006). Choice tests sh owed that females oviposited on Brazilian and California peppertrees, and a very few eggs we re laid on poisonwood and American smoketree. Larvae that hatched on poisonwood survived only for few days, and all eggs laid on America smoketree were non viable. The thrips P. ichini was recommended for field release by the federal interagency Technical Advisory Group for Biological Control Agents of Weeds (TAG) in 2007. The use of tortricid moths and thrips as biocon trol agents has shown to be effective against many noxious weeds. For example, galling by immature stages of the moth Epiblema strenuana Walk. at early stages of plant growth significantly reduced plant height, stem height, flower production, and biomass of Parthenium hysterophorus L. (Asteraceae) in Australia (McFayden 1992, Dhileepan and McFadyen 2001). In addition, biological control of Rubus argutus Link by the moth Croesia zimmermani Clark has been effective in pa stures and open areas in Hawaii (Julien 1992). Examples of thrips as biocontrol agents includes Amynothrips andersoni ONeill on alligatorweed Alternanthera philoxeroides (Martius) Grisebach (Maddox 1973), Liothrips mikaniae Priesner on mile-a-minute weed Mikania micrantha Kunth (Cock 1982), L. urichi Karny on Kosters curse Clidemia hirta L. (Julien and Griffiths 1998), and Sericothrips staphylinus Haliday on gorse Ulex europaeus L. (Hill et al. 2001, Davies et al. 2005). Selecting Effective Biological Control Agents Biological control program s often have been cri ticized for their lack of predictability in terms of agent establishment and success (Ehl er 1990, Harris 1998). Severa l factors are usually

PAGE 22

22 considered during the process of agent selection, such as host sp ecificity (Follett and Duan 1999, van Driesche et al. 2000, Sheppard et al. 2003), climatic adaptability (Wapshere 1983, Senaratne et al. 2006), and more recently, impact on the ta rget weed (McEvoy and Coombs 1999, Briese et al. 2002, van Klinken and Raghu 2006). Host specifici ty testing has received the most attention by biological control practiti oners (Briese and Walker 2002, Sheppard et al. 2003), but effectiveness of biocontrol agents has received much less consideration (van Klinken and Raghu 2006). There often is a wide range of potential agents available in the native range, but usually only a few of those agents effectively suppre ss the target weed (M cFadyen 2003, van Klinken and Raghu 2006). For example, of the 21 insects rele ased against prickly pear in Australia, only four contributed to the contro l of this weed (Hosking et al 1988, Julien and Griffiths 1998). Therefore, more effort should be given to st udying the effect of candidate agents on plant performance early in the process of agent selection in order to improve the outcome of biological control programs (McClay and Balciunas 2005). There is general agreement among biological co ntrol practitioners that highly specialized natural enemies that share an evolutionary history with their hosts are likely to be the most effective agents for controlling invasive sp ecies (Strong and Pemberton 2000, Myers and Bazely 2003, Williams et al. 2005). Although host specificity has usually been considered at the species level, there is increasing evidence that natura l enemy populations may be locally adapted to specific genotypes of their host (Nissen et al. 1995, Hufbauer and Roderick 2005). Local host or fine-tuned adaptation refers to a process wher eby locally occurring herbivores with shorter generation times than their hosts and poor dispersa l capabilities adapt rapi dly to a specific host genotype compared to other hosts to which they have no regular exposure (Edmunds and Alstad 1978, Ebert 1994, Gandon and Van Zandt 1998). Several studies have shown that the dispersal

PAGE 23

23 capacity of thrips adults is typically limited, and movements occur mostly between neighboring host-plants, which result in low gene flow a nd strong local adaptation (Karban and Strauss 1994, Rhainds and Shipp 2003, Rhainds et al. 2005) For example, thrips populations of Apterothrips secticornis Trybom were specifically adapted to individual clones of their host Erigeron glaucus Ker. (Karban 1989). In the case of the Brazilia n peppertree, two cpDNA haplotypes are found in Florida (A and B), and the majority of indivi duals are intraspecific hybrids between the two original introductions (Williams et al. 2005, 2007) Therefore, the performance of biological control agents should be tested against all plant genotypes presen t in the area of introduction. Host-plant quality (e.g., nitrog en content) often influences the performance of insect herbivores (Wheeler and Center 1996, Pri ce 2000, Hunter 2001, Wh eeler 2001, Schwab and Raghu 2006). As mentioned before, Brazilian peppert ree invades different habitats in Florida (e.g. mangrove forests, pinelands, abandoned farm s) and environmental conditions encountered in different habitats may affect survival and effectiveness of potential biocontrol agents. In addition to host-plant quality, climate (e.g., temp erature) present in the introduced range also may affect the establishment of biocontrol agents (Sutherst 2000, Byrne et al. 2002). Developmental biology studies coupled with degree-day calculat ions have been used to predict many aspects of insect population dynamics, such as temperature requirements for a species and number of generations expected at specific locations (Logan et al 1976, Briere et al. 1999, Herrera et al. 2005). In addition, the potential distribution of a particul ar insect also is associated with its ability to tolerate cold temperatures, which is essential for successful over-wintering and permanent population establishment (Chen a nd Kang 2005, Coetzee et al 2007, Lapointe et al. 2007). Therefore, understanding the relationship betw een development and temperature is useful for predicting the outcome of classical biological control programs.

PAGE 24

24 Many authors have emphasized the importance of measuring herbivore ability to suppress the target weed in order to select the most effective bioc ontrol agent (McEvoy and Coombs 1999, Balciunas 2000, Pratt et al. 2005, van Klinke n and Raghu 2006). However, few pre-release studies have been conducted to measure the effect of herbivore damage on growth, survival, and reproduction of the target weed (McClay and Balciunas 2005). Of the 38 weed biocontrol projects examined by McFadyen (2003), only 54 out of 132 agents released (41%) contributed to the successful control of those weeds. Therefore, greater effo rt should be directed towards selecting agents on the basis of their potential efficacy (Sheppard 2003, McClay and Balciunas 2005). For instance, Treadwell and Cuda (2007) observed that multiple artificial defoliation events significantly affected growth and fru it production of Brazilian peppertree. Although artificial defoliation has been fre quently used to measure plant re sponses to herbi vory (Dhileepan et al. 2000, Broughton 2003, Wirf 2006), plants may respond differently to actual insect damage (Lehtila and Boalt 2004, Schat and Blossey 2005). Th erefore, the effectiveness of biological control candidates of Brazilian peppertree should be evaluate d before field releases are undertaken (Balciunas 2000). Goals and Hypotheses The overall goal of this research was to determine the potential effectiveness of the leaflet roller moth E. utilis and the thrips P. ichini as biological control agents of Brazilian peppertree in Florida. More specifically, insect performance was evaluated on different Brazilian peppertree genotypes found in Florida. In addition, Br azilian peppertree growing under different environmental conditions (e.g., saline environments, rich nutrient soil, etc. ) may affect herbivore survival. Thus, laboratory experiments examin ed the effect of host-plant quality on the development and survival of E. utilis Climate (e.g., temperature) may also affect permanent

PAGE 25

25 establishment of biocontrol agents (Pilkington and Hoddl e 2006, Lapointe et al. 2007). Therefore, the temperature requirements of E. utilis were evaluated in th e laboratory. Finally, the effect of E. utilis herbivory on growth and biomass allocation of Brazilian peppertree seedlings was evaluated. The results obtained in this study w ill be used to select suitable sites for field releases, and assist in developing an effective long-term control strategy for Brazilian peppertree in different habitats in Florida. The specific hypotheses that were tested in this study are: Component 1: Insect performance on different Brazilian peppertree genotypes Hypothesis 1a: The survival and development of E. utilis are similar when feeding on different Brazilian peppertree genotypes. Hypothesis 1b: The survival and development of P. ichini are similar when reared on different Brazilian peppertree genotypes. Host specificity studies have shown that E. utilis is oligophagous a nd feeds on Brazilian peppertree and several plant spec ies in the Anacardiaceae family in the laboratory (Cuda, unpubl. data). In contrast, P. ichini has a narrow host range, and can only develop on Brazilian peppertree and the congener S. molle (Cuda et al. 2002). Theref ore, I hypothesized that E. utilis and P. ichini would develop on all Florida genotypes. Component 2: The effect of host-p lant quality on the performance of Episimus utilis Hypothesis 2a: The survival and development of E. utilis are greater on Brazilian peppertree plants grown in high nutrient environments. Hypothesis 2b: The survival and development of E. utilis are similar on Brazilian peppertree exposed to different salinity levels. Brazilian peppertree invades different habi tats in Florida (e.g., mangrove forests, pinelands, abandoned farms) and environmental conditions encountered in these habitats may affect the development of potential biocontrol agents. The performa nce of insect herbivores is usually affected by the nutrient co ntents of their host-plants, and greater survival occurs on high-

PAGE 26

26 nutritional hosts (e.g., high N contents) (Mat tson 1980, Wheeler and Ce nter 1996, Stiling and Moon 2005). Therefore, I hypothesized that E. utilis would perform better when reared on Brazilian peppertree exposed to high fertility treatments in the laboratory. In addition, E. utilis is found in coastal locations in its native range (Brazil) (Martin et al. 2004 ), which suggests that this insect could tolerate saline environments. Component 3: Temperature requi rements for development of Episimus utilis Hypothesis 3a: The developmental rate of E. utilis will allow multiple generations per year to be completed in Florida, and the upper developmental threshold will be sufficiently high to not preclude its establishment. Hypothesis 3b: The ability of E. utilis to tolerate cold temperatures varies among insect stages, and lethal times at cold temperatures can be used to predict favorable regions for insect establishment. Climate (e.g., temperature) is known to influen ce insect distribution and population growth, and therefore, the ability of a species to establish in new areas (Sutherst 2000, Byrne et al. 2002). Temperature is an important factor that influences insect surviv al and developmental rate (Logan et al. 1976, Briere et al. 1999, Herrera et al. 2005). Therefore, I hypothesized that E. utilis developmental rates will increase with temper ature within the bounds of certain lower or upper developmental thresholds. Geographic Information System (GIS) models will be used to predict the number of generations per year that E. utilis may have at different locations post-release (Pilkington and Hoddle 2006). In addition, the potent ial distribution of a particular insect is associated with its ability to tolerate cold temperatures (Chen and Kang 2003, Lapointe et al. 2007). Thus, lethal times (LT50 and LT90) at low temperatures could be used to predict the northern limit of th e distribution of E. utilis in the continental USA. Component 4: Impact of E. utilis on growth and biomass allocation of Brazilian peppertree seedlings

PAGE 27

27 Hypothesis 4a: Growth and biomass allocation of Brazilian peppertree are negatively affected by insect defoliation. Hypothesis 4b: Brazilian peppertree seedlings cannot compensate for high levels of herbivory. The plants aboveground architecture (e.g., number of leaflets) is important for determining the photosynthetic capacity and grow th rates of plants (Pearcy et al. 1987). Therefore, I hypothesized that de foliation by the biocontrol agent E. utilis will affect the growth and biomass allocation of Brazilian peppertree seed lings. In addition, plant compensatory growth varies with the amount and timing of insect damage (Thomson et al. 2003, Schooler and McEvoy 2006). Young plants are usually mo re vulnerable to insect damage, and therefore Brazilian peppertree seedlings exposed to high levels of herbivory will suffer a reduction in growth and biomass, and seedlings will not be able to compensate following insect damage.

PAGE 28

28 A B Figure 1-1. Brazilian peppertree in Fl orida. A) large stands, B) detail of plant leaflet and fruits.

PAGE 29

29 A B Figure 1-2. Episimus utilis A) larval feeding damage, B) female (right) and male (left) adults.

PAGE 30

30 A B C Figure 1-3. Pseudophilothrips ichini A) larvae, B) pupae, C) adult.

PAGE 31

31 CHAPTER 2 EFFECT OF HOST-PLANT GENOTYPES ON THE PE RFORMANCE OF TWO BIOLOGICAL CONTROL AGENTS OF BRAZILIAN PEPPERTREE Introduction Brazilian peppertree, Schinus terebinthifolius R addi (Anacardiac eae), a woody perennial plant native to South America (Barkley 1944, 1957), was introduced into Florida, USA, as an ornamental between ca. 1898 1900 (Morton 1978). Brazilian peppertree is recognized as one of the most widespread exotic plants in Florida (C uda et al. 2006), and aeri al surveys estimate that approximately 2833 km2 of all terrestrial ecosystems in central and south Florida have been invaded by this noxious weed (Cuda et al. 2006). Chloroplast DNA (cpDNA) and nuclear microsatellite analyses indicated that two di fferent populations of Brazilian peppertree were introduced separately on the east and west coasts of Florida (Williams et al. 2005, 2007). Haplotype A is more common on the west coast whereas haplotype B is more common on the east coast (Fig. 2-1). Microsatellite DNA analys es revealed that exte nsive hybridization has occurred between these two types of plants si nce arriving in Florida (Williams et al. 2005, 2007). In addition, genetic studies in the native range have identified eleven cpDNA haplotypes (A-K) in Brazil, and haplotype D is the most co mmon and widespread (Williams et al. 2005, unpubl. data). A classical biological control program was initiated in the 1980s against Brazilian peppertree in Florida. Several phytophagous ins ects were identified as potential biological control agents from exploratory surveys conducted in southeastern Brazil, including the leaflet rolling moth Episimus utilis Zimmerman (Lepidoptera: Tortrici dae), and the shoot and flower attacking thrips Pseudophilothrips ichini Hood (Thysanoptera: Phlaeoth ripidae) (Bennett et al. 1990, Bennett and Habeck 1991, Habeck et al. 1994, Cuda et al. 1999, Cuda et al. 2006). The larval stages of E. utilis feed on Brazilian peppertree leaflets and can completely defoliate small

PAGE 32

32 plants (Martin et al. 2004). Because E. utilis completes its entire life cycle in the canopy of its host plant, this herbivore may be an appropriate agent for Brazi lian peppertree growing on sites exposed to seasonal fl ooding. In the case of P. ichini both larval and adu lt stages damage the host plant by feeding on the growing shoot tip s and flowers causing fl ower abortion (Garcia 1977). Immature thrips undergo tw o larval stages on the plan t, and three non-feeding pupal stages occur in the soil (Garci a 1977, Cuda et al. in press). Studies of the biology and host specificity of E. utilis and P. ichini were conducted in Brazil and Florida quarantine (Garcia 1977, Harmuch et al. 2001, Martin et al. 2004, Cuda el al. 2006). The thrips P. ichini was recommended for field release by the federa l interagency Technical Advisory Group for Biological Control Agents of Weed s (TAG) and a release petition for E. utilis is in preparation. There is general agreement among ecologists and weed biological control practitioners that highly specialized natural enemies that share an evol utionary history with th eir hosts are likely to be the most effective agents for controlling i nvasive plants (Zwolfer and Preiss 1983, Strong and Pemberton 2000, Myers and Bazely 2003, Williams et al. 2005). However, hybridization between plants from different source populations in the introdu ced range may compromise old associations. Therefore, the objective of this st udy was to examine the performance of the leaflet roller E. utilis and the thrips P. ichini on all Brazilian peppertree ge notypes found in Florida. In addition, the performace of two populations of P. ichini were evaluated on different Brazilian peppertree genotypes found in Florida and in the native range. Materials and Methods Insects and Plants This study was conducted at the Biological C ontrol Research and Containm ent Laboratory (BCRCL) located at the Indian River Research an d Education Center (IRREC) of the University of Florida, Fort Pierce, FL, USA. A colony of E. utilis was initiated in August 2006 at the

PAGE 33

33 BCRCL, and insects were reared on potted Brazilian peppertree plants (Flori da genotypes) (pots: 18 cm height, 17 cm diameter) inside environm ental growth chambers (25 2 C, 60-70% RH, 14L: 10D photoperiod) (for specific details on rearing methods, see Martin et al., 2004). Episimus utilis individuals were originally collect ed in 2003 from Brazilian peppertree haplotypes C and D found in the vicinity of Curitib a, located in Parana Pr ovince of southeastern Brazil. Two different populations of P. ichini were imported to the BCRCL in Florida. Both populations were collected in Brazil; the first population was f ound feeding on Brazilian peppertree Brazil haplotypes C and D (Curitib a), and the second was found on Brazilian peppertree Brazil haplotype A (Vicosa). Prelim inary genetic studies using mitochondrial DNA revealed that these two thrips populations are ge netically distinct; the first population (Curitiba) was categorized as haplotype 5, and the second (Vicosa) as haplotypes 2 and 3 (D. Williams, unpub. data). A colony of the thrips haplotype 5 was initiated in Ja nuary 2007 at the BCRCL. Since poor survival was obtained when thrips 5 were reared on Brazilian peppertree (Florida genotypes), the colony was maintained on Schinus molle L. (Anacardiaceae) potted plants. After all experiments were completed using this th rips population, the colony was terminated. A colony of the thrips haplotypes 2 and 3 was initiated in November 2007 at the BCRCL, and insects were reared on Brazilian peppertree potte d plants (Brazilian haplotype A). Voucher specimens of E. utilis and P. ichini were deposited in the Florida State Collection of Arthropods, Florida Department of Agriculture and Cons umer Services, Gainesville, Florida, USA. Brazilian peppertree plants representing four Florida genotypes were grown from cuttings that were collected in the fi eld from west (haplotype A), east (haplotype B), and hybrids (haplotypes A or B) plants previously conf irmed by genetic analysis (Williams et al. 2002, Williams et al. 2005). Individuals were classified as pure bred eastern (haplotype B) or western

PAGE 34

34 (haplotype A) if they had the ancestry coeffici ent value (q) > 0.90 in a re spective cluster, and as hybrids (haplotypes A or B) if they contained a maximum of q = 0.40 for either western or eastern ancestry. The cut ends from different Fl orida Brazilian peppertree plants (~10 plants per genotype) were coated in rooting hormone powder (Schultz TakeRoot Rooting Hormone, Bridgeton, MO) and planted in po ts (18 cm height, 17 cm diameter) using sterilized soil mix (Fafard germination mix, Agawam, MA). Pots we re placed in the shade and misted every 10 minutes. After 3 months, cuttings with attached roots were transplanted to new pots (18 cm height, 17 cm diameter) containing soil mixture (Faf ard #3B mix), and all plants were placed in the greenhouse under ambient conditions and watere d as needed. Brazilian peppertree seeds from Brazil (~5 plants per genotype) were collected in Blumenau (haplotype A) and in Cutitiba (haplotype D), and grown inside a quarantine greenhouse at the BCRCL. Genetic analyses were conducted on all plants to confirm plant genotypes (Williams et al. 2002, 2005). Schinus molle seedlings were purchased from a nursery in Calif ornia (Sherman Growers, San Marcos, CA), and grown inside a greenhouse at the BCRCL. All plants were fertilized once with 15 g of Osmocote (a slow release fertilizer 15-9-12, N-P-K), and 400 ml per po t of liquid fertilizer (Miracle Grow 24-8-16) monthly. Life History Parameters of E. utilis on Differe nt Brazilian Peppertree Genotypes Experiments were conducted inside an envi ronmental growth chamber (25 2 C, 60-70 % RH, 14L: 10D photoperiod) at th e BCRCL. Four treatments were established (six to eight replicates per treatment): 1) Brazilian peppertr ee Florida haplotype B (BP FL-B), 2) Brazilian peppertree Florida haplotype A (BP FL-A), 3) Brazilian peppertree Florida hybrid A (BP FLHA), and 4) Brazilian peppertree Florida hybr id B (BP FL-HB). Ten neonate larvae of E. utilis were caged on each potted plant inside a clear ac rylic cylinder (45 cm height, 15 cm diameter) with six holes (6 cm diameter) and tops covered by a fine mesh to allow air circulation. After 20

PAGE 35

35 days, the cylinders were removed and all plant fo liage was cut. Last inst ars or pupae were placed individually inside plastic vials (29.5 ml, Bio-Serv, Frenchtown, NJ) containing moist filter paper and one plant leaflet. Upon adult emergen ce, individual pairs from each treatment (4-6 pairs per treatment) were placed inside ovipos ition wax paper cages containing a cotton wick soaked in Gatorade (limon-lime flavor) and one Brazilian peppertree leafle t. Rectangles of wax paper (19 30 cm) were stapled together to form a cage for oviposition (Moeri 2007). These cages were placed inside Ziploc freezer bags (22.5 18 cm) and kept in the environmental growth chamber. After all adults died, the numbers of hatched and unhatched eggs were counted under a dissecting microscope. Several insect para meters were recorded: 1) pupal weight (mg), 2) developmental time to adult (d ays), 3) survival to adult (%), 4) adult longevity (days), 5) fecundity (total eggs laid), and 6) fertility (% eg gs hatched). In a separate experiment, E. utilis development and survival was evaluated on two Brazilian peppertree genotype s: BP FL-HB, and Brazil ha plotype D (BP BZ-D). The E. utilis colony was originally collected from haplotype s D and C present in Curitiba, Brazil, and therefore, BP BZ-D was included in the test. Individual neonate larvae were placed inside vials (29.57 ml, Bio-Serv, Frenchtown, NJ) containing moist filter pape r and a plant leaflet, and 10 vials were used for each replicate (five replicates total). Several insect pa rameters were recorded: 1) developmental time to adult (day s), 2) survival to a dult (%), and 3) adult longevity (days). Life History Parameters of P. ich ini on Different Brazilian Peppertree Genotypes and Schinus molle The experiments described below were conduc ted separately for each thrips population (haplotype 5, haplotypes 2-3) inside an environmental growth chamber (28 2 C, 60-70 % RH, 14L: 10D photoperiod) at the BCRCL in Fort Pier ce, FL. Seven treatments were established (eight replicates per treatment): 1) BP FL-B, 2) BP FL-A, 3) BP FL-HA, 4) BP FL-HB, 5) BP

PAGE 36

36 BZ-A, 6) BP BZ-D, and 7) S. molle Ten neonates of P. ichini were placed inside a plastic vial (11 cm height, 5 cm diameter) containing a pl ant shoot and moist filter paper. Vials were checked every other day, and moisture and food was added as needed. The insect parameters recorded were: 1) survival to adult (%), and 2) developmental time to adult (days). In a separate experiment, adu lt longevity of each population of P ichini was measured using the same host plants mentioned above, in cluding vials with no food (control). Ten newly emerged adults of P. ichini (five females: five males) were placed in each vial containing a plant shoot and moist filter paper, and seven to eight replicates per treatment were used. Vials were checked every other day, and survival and pr e-oviposition period were recorded. Experiments were terminated when all adults died. Feeding Preference of P. ichini Adults on Different Braz ilian Peppertree Genotypes and Schinus molle Newly emerged females (1-2 days old) of each P. ichini population, with no previous feeding experience, were used in the experiments (25 2 C). Six females of P. ichini were released in the center of each Pe tri dish (15 cm diameter) containing six leaflet disks (2 cm diameter) of the different Brazilian peppertree genotypes (BP FL-A, BP FL-B, BP FL-HA, BP BZ-A, BP BZ-D) and S. molle arranged in a circle (Fig. 22). A total of 40 Petri dishes (replicates) were used for each thrips population, and observati ons were conducted every half hour for a total of 4 hours. The obs ervations consisted of recordi ng the number of thrips on each leaflet disk. Thrips were considered feeding on the plant when found standing on the disk. Data Analysis Life-histo ry parameters (avera ged values per replicate) for E. utilis and P. ichini were compared between host plants using one-way analysis of variance (ANOVA) (SAS Institute 1999). Two-way ANOVA was used to compare pupal weight of E. utilis between genders and

PAGE 37

37 plant treatments (SAS Institute 1999). Data expr essed as percentages (e.g., survival and eggs hatched) were transformed using the arcsine square root-transformation (Zar 1999). For the feeding preference experiment, the total number of thrips in each plant treatment was analyzed over time using non-parametric Friedmans an alysis of variance by ranks for each thrips population (Zar 1999). Statistically different me ans were separated using the Student-NeumanKeuls (SNK) test (SAS Institute 1999). A significance level of = 0.05 was used for all statistical analyses. Results Life History Parameters of E. utilis on Differe nt Brazilian Peppertree Genotypes No differences were detected for survival (~54%), developmental time (~32 days) and adult longevity (~9 days) of E. utilis among the Florida Brazilian pe ppertree genotypes (survival: F3,26 = 0.11, P = 0.93; developmental time: F3,29 = 0.12, P = 0.94; longevity: F3,23 = 0.9, P = 0.46; Table 2-1). Pupal weight did not differ among plant genotypes (17.6 0.5 mg), but female pupae (18.27 0.8 mg) were larger than male pupae (17.01 0.5 mg) (genotypes: F3,59 = 2.11, P = 0.11; sex: F1,59 = 5.48, P = 0.02; genotype sex: F3,59 = 0.33, P = 0.8). In addition, no differences were detected in fecundity (84.6 10.5 eggs laid) or fertility (67.7 9.5% eggs hatched) of E. utilis females reared on the Florida genotypes (fecundity: F3,17 = 0.87, P = 0.47; fertility: F3,17 = 0.11, P = 0.95). When E. utilis was reared either on Brazilian peppe rtree Florida hybrid B (BP FL-HB) or Brazilian haplotype D (BP BZ-D), similar results were obtained. No differences between plant genotypes were detected for survival to adult ( 33 7%), developmental time to adult (34.5 0.8 days), or adult longevity (5 0.5 days) (survival: F1,8 = 1.0, P = 0.34; developmental time: F1,31 = 0.07, P = 0.78; longevity: F1,21 = 0.94, P = 0.34).

PAGE 38

38 Life History Parameters of P. ich ini on Different Brazilian Peppertree Genotypes and Schinus molle Survival to the adult stage of P. ichini haplotype 5 differed am ong host-plants tested ( F6,51 = 29.95, P < 0.0001; Table 2-2). The highe st survival was observed on S. molle (75%), followed by Brazil haplotype D (44%) and Brazil haplotype A (24%) (Table 2-2). The lowest survival was obtained on all Brazilian peppert ree Florida genotypes tested (0 to 4%) (Table 2-2). No differences were detected among host plan ts for developmental time to adult ( F3,22 = 1.39, P = 0.27; Table 2-2). In contrast, the thrips haplotypes 2-3 had sim ilar survival on all host plants tested except for Brazi l haplotype D (6%) (F6,53 = 6.6, P < 0.0001; Table 2-2). Similarly, developmental time to the adult stage did not diffe r among host plants tested except for BP BZ-D (20 days) ( F6,51 = 3.03, P < 0.01; Table 2-2). Adult longevity of P. ichini haplotype 5 differed among the host plants tested ( F4,39 = 60.75, P < 0.0001; Table 2-2). Adults survived longer on S. molle followed by Brazil haplotype D and Brazil haplotype A (Table 2-2). Longevity of <10 days was recorded for those adults exposed to either Brazilian peppertree Florida A or no food (control) (Table 2-2). Adults laid eggs on S. molle and all Brazilian peppertr ee Brazil genotypes, but no eggs were laid on BP-FL A. The pre-oviposition period was similar for all the three host plants that received eggs (12 1 days) (F2,23 = 2.45, P = 0.11). In contrast, P. ichini haplotypes 2-3 had similar adult longevity in all host plants tested except for BP BZ-D (6 days) and control trea tments (5 days) (Table 2-2). In addition, thrips haplotypes 2-3 laid eggs on all host plant tested except for BP BZ-D on which no eggs were found. The pre-oviposition period (days) of thrips haplotypes 2-3 varied among host plants ( F4,33 = 3.2, P = 0.027), being shorter on BP BZ-A ( 6.5 0.5) and BP FL-A (7.7 1.2) compared to S. molle (8.5 1.3), BP FL-B (10.5 1.7), a nd BP FL-HA (13 1.6).

PAGE 39

39 Feeding Preference of P. ichini Adults on Different Braz ilian Peppertree Genotypes and Schinus molle Statistical differences were de tected in the number of adults of P. ichini haplotype 5 feeding on different host plants over time, and the interaction of host plant and time was also significant (treatment: F5,1919 = 42.8, P < 0.0001; time: F7,1919 = 4.3, P = 0.0001; treatment time: F35,1919 = 1.9, P = 0.0007). After a 4-hour trial, female s preferred to feed on all Brazil genotypes and S. molle compared to Florida genotypes ( F5,239 = 4.6, P = 0.0005; Fig. 2-3). Different results were obtained when thrips ha plotypes 2-3 were used. Statistical differences were observed of the number of adults feeding on different host plants over time, and its interaction was also significant (treatment: F5,1919 = 141, P < 0.0001; observation: F234,1919 = 9.9, P < 0.0001; time: F7,1919 = 30.2, P < 0.0001; treatment time: F35,1919 = 3.6, P < 0.0001). After a 4-hour trial, females prefe rred to feed on BP-BZ A and S. molle compared to Florida genotypes and BP-BZ D ( F5,239 = 10.4, P < 0.0001; Fig. 2-3). Discussion Classical biological control has provided sustainable and long-term control of many invasive weeds in both terrestrial and aquatic ecosystems (e.g., Hartley 1990, Julien and Griffiths 1998, Grevstad 2006, Barton et al. 2007). Therefore, the introduction of ho st specific natural enemies against Brazilian peppertree may contribute to the control of this invasive species in Florida and elsewhere. Although host specificity has usually been cons idered at the species level, there is increasing evidence that natural enemy populations may be locally adapted to specific genotypes of their host (Nissen et al. 1995, Hufb auer and Roderick 2005). For instance, the performance of multiple haplotypes of the mite Floracarus perrepae Knihinicki differed between haplotypes of their plant host Lygodium microphyllum Cav. (Goolsby et al. 2006). Thus,

PAGE 40

40 the mite haplotype that is best adapted to the particular fern haplotype found in the introduced range (Florida) offers the greatest prospect for cont rol (Goolsby et al. 2006). In the case of the Brazilian peppertree, th e majority of individuals in Florida are intraspecific hybrids between the two original introductions (Williams et al. 2005, 2007). In the native range, cpDNA haplotypes are geographically structured with only one or two closely related haplotypes occurring at a given loca tion (Williams et al. 2005, unpub. data). The source location of haplotype A has been found along the coas t of southeastern Brazil, whereas the origin of haplotype B is not yet known, but is believ ed to be along the northeast coast of Brazil (Williams et al. 2005, unpub. data). Therefore, haplotypes A and B are distantly separated and it is unlikely that hybridization occurs naturally in the nativ e range (Williams et al. 2005, 2007). These hybrid genotypes in the introduced range w ill not have locally adapted natural enemies in the native range, and so insect performance will need to be tested against all plant genotypes present in the area of introduction. Results from this study revealed that Brazilian peppertree genotypes did not affect the perf ormance of the leaflet roller E. utilis but had a strong effect on the performance of two popul ations of the thrips P. ichini Although E. utilis seems to be a biological control agent well ad apted to Brazilian peppertree sensu lato this insect may not be sufficiently host specific to releas e in Florida. Host specificity tests in the laboratory found that E. utilis also feeds on other related plant species in the Anacardiaceae family (J.P. Cuda, unpublished data), but these non-target effects were not observed in the native range or in Hawaii (Cuda et al., 2006). Further studies are being conducted in Florida qua rantine to determine whether E. utilis is safe to release in Florida. This study showed that two populations of P. ichini differed significantly in their ability to utilize different genotypes of their host plant. Poor survival to adult hood (0-4%) and short adult

PAGE 41

41 longevity (<10 days) were obtained for the hapl otype 5 thrips on all Florida genotypes, while higher survival (~50%) and longevity (~30 days) were observed for the haplotypes 2-3 thrips on these Florida genotypes. In additi on, thrips haplotype 5 did not la y eggs on Florida genotypes, whereas thrips haplotypes 2-3 laid eggs on all plants tested except for Brazil haplotype D. Therefore, the thrips haplotypes 2-3 that were originally collected from Brazil haplotype A are well adapted to Florida genotypes and should be considered as potential biocontrol agent of Brazilian peppertree in Florid a. The high performance of P. ichini observed on S. molle a close relative of Brazilian peppertree not presen t in Florida, was not unexpected. Although P. ichini is not found feeding on this plant in its native ra nge, previous laborator y tests showed that S. molle is in the physiological host range of the thrips (C uda et al. 2006). Schinus molle is a popular ornamental in California, but the California Exotic Pest Plant Counc il recently listed this species as a Category B invasive species (Cal-IPC, 2006). Local host or fine-tuned adaptation refers to a process whereby locally occurring herbivores with shorter generation times than their hosts and poor dispersal capabilities adapt rapidly to a specific host genotype compared to other hosts with which they have no regular exposure (Edmunds and Alstad 1978, Hartle y and Forno 1992, Ebert 1994, Gandon and Van Zandt 1998, Goolsby et al. 2006). In th is study, each thrips population of P. ichini performed better on the plant genotype from where it was or iginally collected. Preliminary genetic studies showed that thrips haplotype 5 is only associated with BP BZ-D and C, while haplotypes 2-3 were found on haplotypes A, K, N, and M in the native range (D. Williams unpubl. data). Brazilian peppertree haplotype K is close re lated to haplotype B (Williams et al. 2005, unpubl. data), which may explain the high performance of thrips 2-3 on BP FL-B. Overall, these results suggest that different po pulations of the thrips P. ichini are adapted to particular genotypes of

PAGE 42

42 Brazilian peppertree in th e native range. Similarly, th e thrips populations of Apterothrips secticornis Trybom were specifically adapted to individual clones of their host Erigeron glaucus Ker. (Karban 1989). Several studies have shown that the dispersal capacity of thrips adults is typically limited, and movements occur mostly betw een neighboring host-plants, which result in low gene flow and strong local adaptation (Karban and Strauss 1994, Rhainds and Shipp 2003, Rhainds et al. 2005). In contrast lepidopteran species such as E. utilis are capable of flying over larger distances (Suckling et al. 1994, Showers et al. 2001), and this may explain in part the differences in diet breadth between E. utilis and P. ichini Another important trait usually associated with locally-adapted herbivores is parthenogenic reproduction or haplodiploidy (Rice 1983, Boecklen and Mopper 1998), which occurs in P. ichini (Garcia 1977). Biological control programs have b een criticized for their lack of predictability in terms of agent establishment and success (Ehler 1990, Harris 1998). In order to improve the predictability of biocontrol, the effectiveness of the natural enemies should be examined on all host genotypes found in the introduced range. In addition, genetic studies may help to identify co-adapted natural enemies in the native range that may be the most effective against a particular genotype present in the introduced range (Williams et al. 2005, Goolsby et al. 2006). This study showed that two candidate biological control agents diffe red in their ability to utilize different genotypes of their host-plant. The leaflet roller E. utilis performed well on all Brazilian peppertree genotypes tested (Florida and Br azil), whereas two populations of P. ichini differed in their ability to utilize different host plant genotypes. Poor survival was recorded for the thrips haplotype 5 when reared on the Florida genotypes, while higher performance on all host plants tested, except for Brazil D, was obtained for the thrips haplotypes 2-3. Thes e results suggest that different populations of P. ichini are adapted to different Brazili an peppertree genotypes in the

PAGE 43

43 native range, and highlight the importance of testing biological control agents on different genotypes of the weed from both the native and in troduced range. Future studies should examine the chemical composition (e.g., secondary comp ounds) of the different Brazilian peppertree genotypes, which may provide grea ter insight into the interaction between these herbivores and their host-plants.

PAGE 44

44 Table 2-1. Life history parameters (mean SE) of E. utilis on different Brazilian peppertree Florida genotypes. BP FL-A=Brazilian pe ppertree Florida haplotype A, BP FLB=Brazilian peppertree Flor ida haplotype B, BP FL-HB= Brazilian peppertree Florida hybrid B, BP FL-HA=Brazilia n peppertree Florida hybrid A. Survival to adult (%) Developmental time to adult (days) Adult longevity (days) BP FL-A BP FL-B BP FL-HA BP FL-HB 55.0 8.86 55.0 9.44 56.25 8.22 48.33 11.9 31.9 0.64 32.0 0.52 31.52 0.32 31.89 0.63 8.42 0.78 8.16 0.75 9.58 1.14 9.56 0.2

PAGE 45

45 Table 2-2. Life history parameters (means SE) of two populations of P. ichini reared on different host plants. BP FL-A = Brazilian peppertree Florida haplotype A, BP FL B = Brazilian peppertree Florida haplot ype B, BP FL-HA = Brazilian peppertree Florida hybrid A, BP FL-HB = Brazilian pe ppertree Florida hybrid B, BP BZ-A = Brazilian peppertree Brazil haplotype A, BP BZ-D = Brazilian peppertree Brazil ha plotype D, control = no food. Different le tters in the same column indicate statistical differences be tween plant treatments ( P < 0.05). Host plants Survival to adult (%) Development to adult (days) Adult longevity (days) Thrips 5 Thrips 2-3 Thrips 5 Thrips 2-3 Thrips 5 Thrips 2-3 BP FL-A 3.8 2.6 d 51.2 8.9 a 18.7 1.33 a 15.8 0.5 b 9.09 0.3 c 34.8 1.6 a BP FL-B 0 d 57.6 11.2 a 16.2 0.5 b 27.8 1.3 b BP FL-HA 0 d 68.7 10.4 a 16.4 0.6 b 25.8 1.9 b BP FL-HB 0 d 52.5 4.5 a 17.4 1 b BP BZ-A 24.4 11.5 c 41.2 5.8 a 18.8 0.9 a 16.4 0.5 b 18.4 2.2 b 34.8 1.4 a BP BZ-D 43.7 6.6 b 6.6 2.1 b 17.7 0.3 a 20 1 a 21.4 1.1 b 5.7 0.4 c S. molle 75.5 7 a 42.5 3.6 a 17.3 0.5 a 17.1 0.5 b 33.1 1.4 a 25.8 2.2 b Control 8.4 0.4 c 5.4 0.5 c

PAGE 46

46 Fig. 2-1. Brazilian peppertree genotypes present in Florida West coast (haplotype A) East coast (haplotype B) Hybrids A, B

PAGE 47

47 Fig. 2-2. Diagram showing the set-up used for the feeding preference trial of P. ichini Release point of thrips adults. 15 cm

PAGE 48

48 0 0.5 1 1.5 2 2.5B P B Z D B P B Z A S m o l l e B P F L B B P F L A B P F L H AMean number adults Thrips 5 Thrips 2-3 Figure 2-3. Number of adults (mean SE) of P. ichini feeding on each host plant after a 4-hour trial. Different letters indi cate statistical differences be tween host plants (P < 0.05); lower case letter shows differences for thri ps 5, upper case letter shows differences for thrips 2-3. a a b b b a B A A B B B

PAGE 49

49 CHAPTER 3 INFLUENCE OF HOST-PLANT QUALITY ON THE PERFORMANCE OF THE BIOCONTR OL AGENT EPISIMUS UTILIS Introduction Brazilian peppertree, Schinus terebinthifolius R addi (Anacardiaceae), is an introduced perennial plant that has become widely established throughout centr al and south Florida (Cuda et al. 2006). This species is native to Argentin a, Brazil and Paraguay (Barkley 1944, 1957), and was brought to Florida as an ornamental in the 1840s (Mack 1991). Brazilian peppertree not only colonizes disturbed sites such as highway rights-of-way, canals, fallow farmlands, and drained wetlands, but also invades natural communitie s, including pinelands, hardwood hammocks and mangrove forests in Florida (Myrtinger and W illiamson 1986, Ewe 2001, Cuda et al. 2006). The invasion and displacement of native species by Br azilian peppertree poses a serious threat to biodiversity in many ecosystems of Florida (Mor ton 1978). Several attributes of this plant contribute to its invasiveness, in cluding a large number of fruits produced per female plant, an effective mechanism of dispersal by birds (Pan etta and McKee 1997), tolerance to shade (Ewel 1978), fire (Doren et al. 1991), and drought (Nilson and Muller 1980b), allelopathic effect on neighboring plants (Gogue et al. 1974, Nilson and Muller 1980a Morgan and Overholt 2006), and tolerance to saline conditions (Ewe 2001, Ewe and Sternberg 2002). Although herbicides and mechanical methods (e.g., cutting, burning and flooding) are routinely used for controlling existing Brazilian peppertree st ands (Gioeli and Langeland 1997, Cuda et al. 2006), these methods are extremely labor intensive and can be very expensive, especially for large infestations In addition, chemical and mechan ical controls are unsuitable for some natural areas (e.g., mangrove forests) because they may have negative effects on non-target species (Doren and Jones 1997) and may increas e water pollution. Ther efore, alternative methods for long-term control of Brazilian pepper tree are urgently needed. A classical biological

PAGE 50

50 control program was initiated in 1980s, and the leaflet-roller moth Episimus utilis Zimmerman (Lepidoptera: Tortricidae) was se lected as a potential biocontrol agent of Brazilian peppertree because it clearly damaged the plan t and appeared to be host speci fic in its native range (Martin et al. 2004, Cuda et al 2006). The larvae of E. utilis feed on Brazilian peppertree leaflets and can completely defoliate small plants (Martin et al. 2004). Early instar s tie young leaflets together to feed, whereas older instars feed and pupate inside a cylindrical rolled leaflet. The dimorphic adults are small, grayish brown in color with distinct wing patterns that separate the sexes (Zimmerman 1978). Because the entire life cycle of E. utilis occurs in the canopy of the host plant, this insect may be well adapted to areas subjected to seasonal flooding in south Florida. In the 1950s, E. utilis was released and established in Ha waii but successful control of Brazilian peppertree populations was not achieved (Goe den 1977, Yoshioka and Markin 1991, Julien and Griffiths 1998). However, its inability to contro l the plant in Hawaii should not preclude its release into other areas infested with Brazilian peppertree. Unfavorable ab iotic or biotic factors may have contributed to its failure in Hawaii. For example, high larval mortality of E. utilis by introduced and native parasitoids and predators of agricultural pests wa s documented following its release (Davis 1959, Krauss 1963). Because Brazilian peppertree invades different habitats in Florida (e.g., mangrove forests, pinelands, abandoned farms), the environmental c onditions encountered in different habitats may influence the survival and effectiveness of potenti al biocontrol agents. Ther efore, the objective of this study was to evaluate the performance of E. utilis reared on Brazilian peppertree exposed to different salinity and nutrient levels in the laboratory. Results from this study will provide a better understanding of the interactio n between Brazilian peppertree and E. utilis one of its

PAGE 51

51 principal herbivores. This information will improve the selection of release sites favorable for establishing the insect in Florida foll owing its release from quarantine. Materials and Methods Insect Rearing Laboratory experim ents were conducted at the Biological Control Research and Containment Laboratory (BCRCL) located at the Indian River Research and Education Center (IRREC) of the University of Flor ida, Fort Pierce, FL. A colony of E. utilis was initiated in August of 2006 at the BCRCL, and insects were r eared on individual Brazil ian peppertree plants (Florida genotypes) grown in nur sery pots (18 cm height, 17 cm diameter) inside environmental growth chambers (25 2 C, 60-70 % RH, 14L: 10D photoperiod) (see Mart in et al. (2004) for rearing procedures). Insects were originally coll ected in 2003 in the vicinity of Curitiba located in the Parana Province of southern Brazil, and im ported to the quarantine f acility in Gainesville, FL. Voucher specimens of E. utilis were deposited in the Florida State Collection of Arthropods, Florida Department of Agriculture and Cons umer Services, Gainesville, Florida, USA. Life History Parameters of E. utilis on Braz ilian Peppertree Exposed to Different Salinity Levels Brazilian peppertree plants (24) were grown fr om seeds collected from Floridas east coast saline environment (Fort Pierce, FL) to assure plant adaptability to those conditions and maintained in the same size nursery pots (18 cm height, 17 cm diameter pots). Plants were grown using a mixture of soil (Fafard #3B mix) a nd sand (1:1); all plants were placed in the greenhouse at BCRCL and watered as needed. Pl ants were fertilized once with 15 g of Osmocote slow release fertilizer (15-9-12 N-K-P), and monthly with 400 ml per pot of liquid fertilizer (Miracle Grow 24-8-16). Forty-five days prior to initiating the experiments, three levels of soil salinity were established as follows (eight replicates each ): 1) fresh water environment (0

PAGE 52

52 parts per thousand salinity), 2) low saline envi ronment (6 ppt salinity), and 3) high saline environment (12 ppt salinity). Plants were either irrigated with tap water (fresh water environment) or with water supplemented with s eawater (36 ppt) to obtain 6 or 12 ppt salinity. In order to prevent plant stress, a stepwise increase in salt concentration (1 ppt every 3 days) was used. Plants were irrigated w ith each salinity level for the dur ation of the experiment (~2 months). Several plant parameters were recorded at the beginning of the experiment (four to six replicates per treatment): 1) leaflet nutrient (N P, K) and sodium cont ents (Na), 2) leaflet toughness, 3) specific leaflet area (SLA=leaf area / leaf dry weight), a nd 4) percent leaflet moisture content (LMC=(fresh dried leaf weig ht) x 100 / fresh weight). Leaflet toughness was measured using a modified 300 g Pesola (Forestry Suppliers Inc., Jackson, MS) with a probe that estimates the pressure required to puncture leaflet tissues (average of four punctures per leaflet). Leaflet toughness, SLA and LMC were determ ined for one newly expanded leaflet (2nd from top of the stem), and one mature leaflet (3rd from bottom of the stem) from each treatment plant. For the nutrient and sodium analyses, leaflet samp les from each plant treatment were harvested, oven-dried at 70C for 1 week a nd ground; all samples were sent to the Agricultural Analytical Services Laboratory, Pennsylva nia State University, PA. Experiments were conducted inside an envi ronmental growth chamber (25 2 C, 60-70 % RH, 14L: 10D photoperiod) starting on December 15, 2006. Ten neonate larvae of E. utilis were caged on each potted plant inside a clear ac rylic cylinder (45 cm height, 15 cm diameter) with six holes (6 cm diameter) and tops covered by a fine mesh to allow air circulation. After 15 days, plant foliage was carefully examined and rem oved; mature larvae were placed individually inside plastic vials (29.57 ml, Bio-Serv, French town, NJ) containing moist filter paper and one

PAGE 53

53 plant leaflet. To measure oviposition, newly emer ged adult pairs from each treatment (4-5 pairs per treatment) were placed inside wax paper cag es (rectangles of 19 x 30 cm) containing one Brazilian peppertree leaflet and a cotton wick with Gatorade for food (Moeri 2007). The oviposition cages were placed inside Ziploc freezer plastic bags and kept in the environmental growth chamber (same as before). After all ad ults died, the numbers of hatched and unhatched eggs were counted using a microscope. Several ins ect parameters were recorded: 1) pupal weight (mg), 2) developmental time to adult (days), 3) surv ival to adult (%), 4) adult longevity (days), 5) fecundity (total eggs laid), and 6) fertility (% eg gs hatched). Life History Parameters of E. utilis on Braz ilian Peppertree Exposed to Different Nutrient Levels Brazilian peppertrees (24 plants ) were grown from seeds collect ed in Fort Pierce, FL using a mixture of potting soil (Fafard #3B mix) and sand (1:1), and all plan ts were placed in a greenhouse at BCRCL and watered as needed. Plants were fertilized monthly using 400 ml per pot of liquid fertilizer (Miracle Grow 24-8-16, N-P-K) during th e first 4 months to assure adequate plant growth. Two months before star ting the experiment, thr ee nutrient levels were established as follows (eight replicates each): 1) low nutrient-level (no fe rtilizer), 2) medium nutrient-level (fertilized once per month using 4 mg of fertilizer (Mir acle Grow 24-8-16) per litter of water), and 3) high nut rient-level (fertilized twice pe r month using Miracle Grow 24-816). Experiments were conducted inside an envi ronmental growth chamber (25 2 C, 60-70 % RH, 14L: 10D photoperiod) starting on August 15, 2007. Several plant parameters described previously were recorded at the beginning of the experiment (six to eight replicates per treatment): 1) leaflet nutrient content (N, P, K), 2) leaflet toughness, 3) SLA, and 4) LMC. Ten neonate larvae of E. utilis were caged on each potted plant inside a clear acrylic cylinder. After

PAGE 54

54 15 days, plant foliage was removed and examined for insects; mature larvae were placed individually inside plastic vials (29.57 ml, Bi o-Serv, Frenchtown, NJ) containing moist filter paper and one plant leaflet. The same insect parame ters were recorded as in the salinity study: 1) pupal weight (g), 2) developmental time to adu lt (days), 3) adult surviv al, 4) adult longevity (days), 5) fecundity (total eggs laid ), and 6) fertility (% eggs hatched). Data Analyses Insect p arameters (% survival to adult, adu lt longevity, fecundity, and fertility) and plant parameters (leaflet toughness, SLA, LMC, a nd nutrient content) were compared among treatments (different salinity levels or nutrien t levels) using one-way analysis of variances (ANOVA) (SAS Institute 1999). Th e proportion of individuals surv iving and the proportion of eggs hatched were arcsine square root transformed prior to an alysis (Zar 1999). Pupal weight was compared between plant treatment and gender using two-way ANOVA (SAS Institute 1999). Means were separated using the post-hoc Student-Neuman-Keuls (SNK) test (SAS Institute 1999). A significance level of = 0.05 was used for all statistical analyses. Results Life History Parameters of E. utilis on Braz ilian Peppertree Exposed to Different Salinity Levels Higher survival to the adult stage was obtained when insects were exposed to Brazilian peppertree plant grown in fresh (55%) and low sa linity-environments (36%), while survival was significantly lower under high salinity conditi ons (6%) (arcsine tran sformation survival: F2,23 = 18.6, P < 0.0001; Table 3-1). Time of development to the adult stage also differed between treatments (F2,19 = 24.23, P < 0.0001) with the longest time (44 days) observed for those larvae reared at the highest salinity level (12 ppt) (Table 3-1); howev er, adult longevity was similar between treatments ( F2,15 = 0.17, P = 0.84; Table 3-1). In addi tion, pupal weight differed

PAGE 55

55 between treatments and adult gender, but th e interaction was not significant (treatment: F1,28, = 45.11, P < 0.0001; gender: F1,28 = 4.92, P = 0.03; treatment gender: F1,28 = 0.03, P = 0.85). Higher pupal weight was obtained for E. utilis reared on Brazilian pepper tree plants grown in the fresh-water environment (18.93 0.9 mg) compared to low salinity-environments (13.73 1 mg), and female pupae (17.58 0.9 mg) were heavier than males (15.68 0.8 mg). However, no differences were detected in fecundity (88.9 18 eggs laid) or fertility (81.3 6.4% eggs hatched) between fresh and low saline environments (fecundity: F1,8 = 0.03, P = 0.87; fertility: F1,8 = 0.03, P = 0.86). Pupal weight and reproductive parameters were not examined for E. utilis reared under high-salinity conditions (12 ppt) due to low numbers of pupae and adults obtained. Leaflet nutrient and sodium contents also diffe red between salinity treatments (Table 3-2). Higher N and P were obtained on leaflets from fresh-water environments compared to saline environments, whereas, higher K and Na levels were recorded from low and high-saline environments (Table 3-2). In addition, toughness of mature leaflets (ML) was greater in the highsalinity treatments (12 ppt), whereas specific l eaflet area (SLA) of new expanded leaflets (NEL) was greater in the fresh-water treatment (Table 3-3). Percent leaflet moisture content (LMC) was higher in the fresh-water compared to saline treatments (Table 3-3). Life History Parameters of E. utilis on Braz ilian Peppertree Exposed to Different Nutrient Levels Higher percent survival of E. utilis to the adult stage occurred on Brazilian peppertree plants exposed to medium (25%) and high (40%) nutrient levels compared to the low level (1%) ( F2,23 = 12.8, P = 0.0002; Table 3-4). Time of development to the adult stage and adult longevity was similar on medium and high nutrien t treatments (time of development: F 1,12 = 1.57, P = 0.24; longevity: F 1,11 = 0.97, P = 0.35; Table 3-4). Although pupal we ight did not differ between medium and high nutrient treatments, female pupae (17.9 0.5 mg) weighed more than male

PAGE 56

56 pupae (16.36 1.0 mg) (treatment: F1,19 = 0.42, P = 0.52; gender: F1,19 = 5.93, P = 0.02; treatment gender: F1,19 = 0.00, P = 0.96). No differences were detected in fecundity (~133 total eggs laid) of E. utilis females reared on Brazilian peppertree plants grown in the medium or high nutrient treatments, whereas higher fertility (88% eggs hatched) was obtained in the high nutrient treatment (fecundity: F1,9 = 1.13, P = 0.31; fertility: F1,9 = 6.25, P = 0.03; Table 3-4). Development and reproductive parameters were not examined for E. utilis reared on low nutrient levels due to low numbers of adults obtained. The leaflet nutrient contents likewise differed between fertilization treatments (Table 3-5). Higher N, P, and K levels were detected in Braz ilian peppertree leaflets exposed to high nutrient treatment compared to the medium or low treatments (Table 3-5), confirming that the application of fertilizer resulted in nutritional differences in leaves. In addition, lower toughness of NEL and higher leaflet moisture content of ML were recorded on high nutrient plants compared to medium or low treatments (Table 3-6). Specific leaflet area (SLA) was not influenced by nutrient treatments (Table 3-6). Discussion Several factors influence the establishm ent of biological control agents, such as climate, natural enemies present in the introduced range and host-plant quality (Sutherst and Maywald 1985, Wheeler and Center 1996, Hunter 2001, Wheeler 2001, Byrne et al. 2003, Senaratne et al. 2006). Furthermore, the nutritional status of ma ny weeds has played an important role in successful biological control programs (Mye rs 1987, Julien 1992). For example, initial applications of nitrogen fertilizer helped in the establishment of Cactoblastis moth on prickly pear cactus (Dodd 1940) and the Cyrtobagous weevil on the Salvinia aquatic fern (Thomas and Room 1986). In addition, plant pa latability varies across salinit y gradients (Levine et al. 1998, Moon and Stiling 2000, Goranson and Ho 2004), and high salinity levels may be detrimental to

PAGE 57

57 some insect herbivores (Hemminga and va n Soelen 1988, Moon and Stiling 2002, Schile and Mopper 2006). Results from this study showed that the performance of the candidate biocontrol agent E. utilis was influenced by its host-plant quality. Su rvival to adult was similar when insects were reared either on Brazilian peppertree irrigated with fres hwater (55%) or low salinityenvironments (36%), whereas lower survival was obtained for in the high salinity-environment (6%). In addition to high mortalit y, time of development to adult for E. utilis was longest in the high-salinity environment, which indicates poor host-quality for these plants. Differences in insect performance also were detected on Brazilian peppertree plants exposed to different nutrient levels. Increased survival of E. utilis was observed under medium (25%) and high (40%) nutrient levels compared to low (1.25 %). Higher fertility was obtained when E. utilis was reared on plants in the high nutrient treatments. Therefore, the increase of leafle t nutrient content (N, P, K) improved insect performance overall. Brazilian peppertree invades different habitats in Florida, including transitional mangrove forests where salinity levels range between 0 to 25 ppt depending on the ti me of the year (rainy vs. dry season) (Ewe 2001, Cuda et al. 2006). Thus, th e salinity levels used in this study (6 and 12 ppt) are within the range of values encountered in the fiel d. Brazilian peppertrees exposed to fresh-water environments were higher-quality host plants, as evidenced by the higher survivorship of E. utilis Differences in plant parameters (e.g., leaflet nutrient contents, toughness, etc.) may explain the changes in ins ect survival observed among salinity-treatments. For example, N and P leaflet content were higher in fresh-water environments compared to saline environments, whereas K and Na were higher in low and high-saline environments. Several studies have shown that nitrogen is a key factor limiting the performance of insect herbivores, and an increase of N availability us ually improves larval survival, growth rate and

PAGE 58

58 reproduction of many insect species (Slans ky and Feeny 1977, Wheeler and Center 1996, Bowdish and Stiling 1998, Hinz and Muller-Sch arer 2000, Stiling and Moon 2005). In contrast, large amounts of sodium (Na) may disrupt metabo lic function and could be detrimental to some insect herbivores (Wang et al. 2001, Wang 200 2, Schile and Mopper 2 006). Increased leaflet toughness also may reduce herbivore feeding and th erefore influence inse ct performance (Feeny 1970, Raupp 1985, Wheeler and Center 1996). In this study, Brazilian peppertree plants exposed to high salinity conditions had tougher mature l eaflets and lower moisture content, which negatively affected E. utilis survival. In contrast, Brazilian peppertrees irrigated only with freshwater had a higher specific l eaflet area (SLA), which translat es into thinner leaflets and lower toughness. Similarly, high salinity levels have been shown to adversely affect the performance of coleopteran stem-borers (Hemminga and van Soelen 1988), and were detrimental to leaf-mining insects (Schile and Mopper 2006). In addition, increased salinity and decreased N availability negatively affected the gall density of Asphondylia borrichiae Rossi and Strong (Diptera: Cecidomyiidae) (Moon and Stiling 2002). Brazilian peppertree also is found in high-fertil ity soils such as dist urbed sites or former farmlands in south and central Florida (Li and Norland 2001, Cuda et al. 2006). For example, the invasion of the Everglades National Park by Br azilian peppertree was facilitated by nutrient enrichment in abandoned farm soils in the Ho le-in-the-Donut (Li and Norland 2001). Results from this study showed that Brazilian peppertree nutrient levels influenced the development and survival of E. utilis in the laboratory. Higher survival (40%) and fertility (88% eggs-hatched) occurred in the high nutrient treatments, which contained higher leaflet nutrient content (N, P, K), lower toughness and higher moisture content of leaflets. There often is a positive relationship between nitrogen and insect performance (Slansky and Feeny 1977, Mattson 1980, Wheeler and

PAGE 59

59 Center 1996, Stiling and Moon 2005). In addition, dietary water has a nutritional value for most insect herbivores (Scriber 1979, Waring a nd Cobb 1992, Huberty and Denno 2004), and increase in water content has a positive effect on the gr owth performance of many lepidopteran species (Scriber 1979, Huberty and Denno 2004, Schoonhoven et al. 2005). Moreover, tougher foliage due to water stress can reduce nitrogen availabili ty and adversely affect performance of chewing insects (Huberty and Denno 2004). Another important factor to consider is plant allelochemicals or secondary compounds that may affect ins ect development (Karban and Baldwin 1997). For example, an increase of nutrient uptake by plants may result in higher a llocation to growth and reduce production of carbon-based allelochem icals (Bryant et al. 1983, Herms and Mattson 1992). This study indicates that Brazilian peppertrees growing in high-fertility soils will provide high-quality hosts for E. utilis development and survival, and insect establishment may occur more rapidly at those sites compared to low-fertility soils. Overall, this study provides a better unders tanding of the interaction between Brazilian peppertree and E. utilis a candidate biological control agent of this noxious weed in Florida. Brazilian peppertree plants growing in fres h water environments (e.g., upland pinelands, hammocks forests, along canals) and in hi gh or medium-fertility soils (e.g., abandoned farmlands, disturbed sites) will provide high-quality hosts for E. utilis development and survival. In addition, coastal environments wi th saline levels of ~6 ppt. also will provide suitable sites for insect establishment. This information will help to select the most suitable sites for field releases and successful establishment of E. utilis in Florida. Further studies should evaluate population densities of E. utilis in different habitats in the native ra nge in order to understand better how environmental conditions affect population dynami cs of this species in the field.

PAGE 60

60 Table 3-1. Life hist ory parameters of E. utilis (means SE) reared on Brazilian peppertree exposed to different salinity levels. Diffe rent letters in the same column indicate statistical differences between treatments ( P < 0.05). Salinity levels Survival to adult (%) Time of development (days) Adult longevity (days) 0 ppt. 55 8.86 a 32.3 0.8 c 8.5 0.6 6 ppt. 36.25 6.25 a 39.37 1 b 8.94 0.6 12 ppt. 6.25 3.24 b 44.37 2.32 a 8.75 0.9

PAGE 61

61 Table 3-2. Leaflet nutrient contents (means SE ) of Brazilian peppertree exposed to different salinity levels. Different letter in the same column indicate statistical differences between treatments ( P < 0.05), df = 2, 13. Salinity levels N (%) P (%) K (%) Na (ug/g) 0 ppt. 2.2 0.07 a 0.42 0.07 1.69 0.19 b 4700 465 b 6 ppt. 1.8 0.09 b 0.27 0.01 2.71 0.02 a 10906 1185 a 12 ppt. 1.8 0.09 b 0.21 0.006 2.89 0.11 a 12452 923 a F 6.07 3.97 18.77 28.11 P 0.01 0.05 0.0003 <0.0001

PAGE 62

62 Table 3-3. Plant parameters (means SE) of Brazilian peppertree exposed to different salinity le vels. NEL=new expanded leaflet ML=mature leaflet. Different letters in the same column i ndicate statistical differences between salinity treatments ( P < 0.05), df = 2, 11. Salinity levels Leaflet toughness (g/mm2) Specific leaflet area (cm2 / g) Leaflet moisture content (%) NEL ML NEL ML NEL ML 0 ppt. 22.3 2.7 49.3 4.6 b 135.4 6.5 a 100.3 8.3 75 4.6 a 67.2 0.5 a 6 ppt. 36.9 6.4 61.6 6.4 b 97.6 11.2 b 84.1 7.7 53.5 3.2 b 56.2 1 c 12 ppt. 50 9.9 77.8 3.2 a 86 7.1 b 88.5 8.6 52.8 4.2 b 59.1 0.9 b F 3.9 8.45 9.01 1.03 9.7 45.6 P 0.06 0.008 0.007 0.39 0.005 <0.0001

PAGE 63

63 Table 3-4. Life history parameters of E. utilis (means SE) reared on Brazilian peppertre e exposed to different nutrient levels. Different letters in the same column indicate statistical differences between treatments ( P < 0.05). Nutrient levels Survival to adult (%) Time of development (days) Adult longevity (days) Fecundity (total number eggs) Fertility (% egg hatched) Low 1.25 1.25 b Medium 25 8.23 a 35.76 1.1 11.93 0.5 119.5 16.5 78.9 3.03 b High 40 6.54 a 34.1 0.8 11.28 0.5 148 21.7 88.7 1.1 a

PAGE 64

64 Table 3-5. Leaflet nutrient contents (means SE ) of Brazilian peppertree exposed to different nutrient levels. Different lett ers in the same column indicate statistical differences between treatments ( P < 0.05), df = 2,17. Nutrient levels N (%) P (%) K (%) Low 0.95 0.06 c 0.14 0.01 c 1.89 0.05 b Medium 1.54 0.06 b 0.31 0.02 b 1.72 0.07 c High 2.01 0.05 a 0.45 0.02 a 2.22 0.05 a F 77.11 50.36 20.33 P <0.0001 <0.0001 <0.0001

PAGE 65

65 Table 3-6. Plant parameters (means SE) of Brazilian pepperrtee exposed to different nutrient le vels. NEL=new expanded leaflet ML=mature leaflet. Different letters in the same column indicate statistical differences between treatments ( P < 0.05), df = 2, 23. Nutrient levels Leaflet toughness (g) Specific leaflet area (cm2 / g) Leaflet moisture content (%) NEL ML NEL ML NEL ML Low 26.39 2.77 a 56.29 3.97 145 14.6 124.5 6.8 66.21 1.55 67.24 1.1 b Medium 19.05 2.67 ab 57.94 2.26 145.6 9.1 132.6 14.2 69.1 1.25 70.13 0.9 a High 14.98 2.04 b 56.85 3.36 169.8 7.5 130.7 10.4 70.04 0.53 70.9 0.7 a F 5.25 0.07 1.7 0.15 2.77 4.31 P 0.01 0.93 0.21 0.86 0.08 0.027

PAGE 66

66 CHAPTER 4 TEMPERATURE-DEPENDENT DEVELOPMENT AND POTENTIAL DISTRIBUTION OF THE BIOCONTROL AGENT EPISIMUS UTILIS Introduction Brazilian peppertree, Schinus terebinthifolius R addi (Anacardiaceae), is an introduced perennial plant that has become widely established throughout centr al and south Florida (Cuda et al. 2006). This species is native to Argentin a, Brazil and Paraguay (Barkley 1944, 1957), and was brought to Florida as an ornamental in the 1840s (Mack 1991, Cuda et al. 2006). In the USA, Brazilian peppertree occurs in Hawaii, Cali fornia, Arizona, Texas and Florida (Habeck et al. 1994, Cuda et al. 2006). In Florida, Brazilian peppertree is recognized as one of the worst invasive species by the Florida Exotic Pest Council, and is one of the mo st widespread exotic plants in the state (Cuda et al. 2006). The invasion and displacemen t of native species by Brazilian peppertree poses a serious threat to biodiversity in many ecosystems of Florida (Morton 1978). Brazilian peppertree has been a target for classical biological control in Florida since the 1980s (Habeck et al. 1994, Cuda et al. 2006) During exploratory surveys conducted in southeastern Brazil, the leaflet roller Episimus utilis Zimmerman (Lepidoptera: Tortricidae) was found to be commonly associated with Brazi lian peppertree (Bennett et al. 1990, Bennett and Habeck 1991, Martin et al. 2004). The larvae fe ed on Brazilian peppertree leaflets and can completely defoliate small plants (Martin et al. 2004). First to third instars tie together young leaflets with silk to provide a cryptic feeding si te, while older instars f eed inside a cylindrical rolled leaflet. Adults are small, grayish brown moths with distinct wing patterns that separate males from females (Zimmerman 1978). The entire life cycle of E. utilis occurs in the canopy of Brazilian peppertree, which may be a favorable char acteristic to allow establishment of this agent in areas subjected to seasonal flooding in south Florida. Episimus utilis was released and

PAGE 67

67 established in Hawaii in the 1950s, but successful control of Brazilian peppertree populations has not been achieved (Goeden 1977, Yoshioka and Ma rkin 1991, Julien and Griffiths 1998). Several factors may explain this outcome, such as unf avorable abiotic and biotic conditions for E. utilis in Hawaii. For example, high larval mortality by introduced and native pa rasitoids and predators has been documented (Davis 1959, Krauss 1963). An understanding of temperature-dependent development of biocont rol candidates such as E. utilis is important when predicting potenti al geographic distri bution and population dynamics in introduced areas. Developmental biology studies coupled with degree-day calculations have been used to predict many as pects of insect population dynamics, such as temperature requirements for a species and number of generations expected at specific locations (Logan et al. 1976, Briere et al. 1999, Lewis et al. 2003, Herrera et al. 2005). For example, Geographic Information System (GIS) models were used to predict favorable areas for permanent establishment of the egg parasitoid Gonatocerus ashmeadi Girault (Hymenoptera: Mymaridae), which is the major biocontrol agent of Homalodisca coagulata (Say) (Hemiptera: Cicadellidae) in California and Hawaii (Pilking ton and Hoddle 2006). In addition, the potential distribution of a particular insect also is associated with its ability to tolerate cold temperatures, which is essential for succe ssful over-wintering and permanent population establishment (MacDonald et al. 1999, Bale et al. 2002, Chen and Kang 2003, Coetzee et al. 2007, Lapointe et al. 2007). Therefore, understanding the re lationship between developmen t and temperature is useful for predicting the outcomes of cla ssical biological control programs. The objectives of this study were to determine temperature-dependent development and survival of E. utilis and use this

PAGE 68

68 information to generate a map to predict the number of generations per year of this biocontrol candidate across the full range of Br azilian peppertree in the USA. Materials and methods Plants and Insects Studies were conducted at the Biological Control Research and Containm ent Laboratory (BCRCL) located at the Indian River Research an d Education Center (IRREC) of the University of Florida, Fort Pierce, FL. A colony of E. utilis was initiated in August of 2006 from a culture maintained in Gainesville, Florida originally co llected in 2003. Insects were reared on Brazilian peppertree potted plants inside environmental growth chambers (25 2C, 60-70% RH, 14L: 10D) (for specific details on rearing methods, see Martin et al. 2004). Insects were originally collected in 2003 from the vicinity of Curitiba locat ed in the Parana province of southern Brazil. Brazilian peppertree leaflets were collected from the field in Fort Pierce, washed and placed inside plastic bags with a moist paper towel inside a refrigerator set at 5C. All leaflets used in the experiments were no more than 2 days old. Voucher specimens of E. utilis were deposited in the Florida State Collection of Arthropods, Flor ida Department of Agriculture and Consumer Services, Gainesville, Florida, USA. Survival and Developmental Times Calculations Tem perature-dependent development studies at seven constant te mperatures (10, 15, 20, 25, 30, 33, 35 0.5C) were conducted inside environmental growth chambers (60-70% RH, 14L: 10D). HOBO data loggers were placed in side each environmental growth chamber to confirm the experimental conditions. Neonate larvae were placed individually inside plastic vials (29.57 ml, Bio-Serv, Frenchtown, NJ) containing moist filter paper and a Brazilian peppertree leaflet. A total of 50 replicates (one vial = on e replicate) per temperature was used. Vials were checked every other day, and water was added to maintain humidity and food was added as

PAGE 69

69 needed. Survival and insect stages were record ed until adult emergence. In addition, single pairs of E. utilis adults (> 2 days old) were placed insi de oviposition wax paper cages containing one Brazilian peppertree leaflet and a cotton wick with Gatorade (lemon-lime) as diet (Moeri 2007). Oviposition cages were made using rectan gles of wax paper (19 x 30 cm) that were stapled together to form a cage (Moeri 2007). Ea ch oviposition cage was pl aced inside a Ziploc freezer bag to maintain high humidity (25 2 C, RH=60-70 %, photoperiod 14L: 10D). The wax paper cages with fresh eggs (<1 d old) were pl aced inside environmenta l growth chambers at each of the aforementioned seven temperatures ( 170-244 eggs from at least four females were used per temperature). Eggs were checked daily under the microscope to determine the percentage of eggs hatched and duration of egg development. The general linear model procedure (PROC GLM, SAS Institute 1999) was used to analyze the developmental time of each insect stage and all immature stages combin ed (egg-adult) for each temperature, and means were separated using the Stude nt-Neuman-Keuls (SNK) test ( P < 0.05). Temperature Thresholds and Degree-days Calculations The linear p ortion of the developmental rate curve [R (T) = a + bT] for each insect stage and all immature stages combined was modeled using the least squares linear regression (PROC GLM, SAS Institute 1999); T refers to temperat ure, and a and b are th e intercept and slope estimates, respectively. The lower temperature threshold was estimat ed by the intersection of the regression line at R (T) = 0, T0 = -a/b. Degree-day requirements for each stage and all immature stages combined were calculated using the invers e slope (1/b) of the appropiate fitted linear regression line (Campbell et al. 1974). The non-linear relationship between developmental rate R(T) and temperature T for each insect stage and all immature stages combined was fitted using the Logan 6 model which allows the estimation of the upper developmental th reshold (Logan 1976, Roy et al. 2002). For the

PAGE 70

70 Logan model equation R(T) = [e ( T) e ( TL (TL T)/ T)]; R(T) refers to the developmental rate at temperature T, is the maximum developmental rate, is a constant de fining the rate at optimal temperature, TL is the lethal maximum temperature, and T is the temperature range at which physiological breakdown becomes the ove rriding influence (Logan 1976, Roy et al. 2002). The developmental rate of E. utilis was modeled using the Marqua rdt algorithm of PROC NLIN that determines parameter estimates through part ial derivations (SAS In stitute 1999). The range of temperatures used for the linear and non-lin ear models (15 to 30C) was selected based on those temperatures where E. utilis completed its development. Cold Tolerance of E. utilis Cold tolerance studies we re conducted using egg, fift h instar, pupa, and adult E. utilis stages, and the sam e methodology was employed for each insect stage. Four insects (larvae, pupae or adults) were placed inside a vial (29.57 ml, Bio-Serv, Frenchtown, NJ) containing moist filter paper and a Brazilian peppertree leaflet or Gato rade, and a total of twenty vials per insect stage were used for each temperature. In addition, adult pairs (>2 days old) were placed inside oviposition wax paper cages contai ning one Brazilian peppertree l eaflet and a cotton wick with Gatorade. Wax paper cages (20 for each temperatur e) with fresh eggs (< 1 day) were placed inside Ziplocs freezer bags and used for the experiments (1,200-1,800 eggs per temperature). All insect stages of E. utilis were acclimated to low temperatures by exposing them gradually from 15oC to the final temperature in intervals of 5oC per day. Each insect stage (eggs, fifth instars, pupae, and adults) was exposed to three constant temperatures (10, 5, and 0oC) for 0.5, 1, 2, 4, and 8 days (four replicates per exposur e time). Since pupae were still alive after 8 days, a longer exposure time (ranging from 10 to 24 days) was used until all pupae had died at 0 and 5oC. After each exposure time, insects were placed at room temperature (25 1oC) and survival was assessed 24 h later. Egg mortalit y was measured by examining egg-hatching daily

PAGE 71

71 during 14 days after exposure to cold temperatur es. Percentage survival (arcsine square root transformation) of each insect stage was analyzed by two-way analysis of variance (ANOVA) with exposure time and temperature as factor s (SAS Institute 1999). The pupal stage was the most tolerant to cold temperatures; therefore, this stage was considered the most plausible overwintering stage and was used for further analys is. Effect of temperat ure and exposure time on percent pupal survival was analyzed using the l ogistic regression (SAS Institute 1999). The time required to 50 and 90% mortality (Ltime50 and Ltime90) of the population at a specific temperature was estimated using PROC PR OBIT (SAS Institute 1999). The Ltime50 and Ltime90 for pupae at 5 and 0C were used to develop models to predict isothermal lines with regions unfavorable for E. utilis establishment. The NAPPFAST database stores daily climate information from 1,879 weather stations acro ss North America (Borchert and Magarey 2005), and probability maps were generated using the last 10 years of weathe r data with the above conditions. The maps were imported to ArcGis 9.0 and a line delimiting the frequency of occurrence for 5 out of 10 years was created (Lapointe et al. 2007). GIS Maps to Predict Number of Generations of E. utilis in the USA The m ethodology developed by Pilkington and H oddle (2006) to generate prediction grids using ArcGis was used is this study. GIS maps were generated from all US states where Brazilian peppertree is present. Daily temperat ures from Florida (98 stations), Texas (396 stations), Arizona (61 stations), California (188 stations) and Hawaii (55 stations) were obtained from weather stations recorded by the Applied Climate Information System (CLIMOD, Southeast Regional Climate Center, http://acis.dnr.sc.gov/Climod/ ). Values for minimum and maximum temperatures were averaged separate ly for the last 5 to 11 years depending on the availability of data. A Microsoft Excel application developed by Univers ity of California-Davis (DegDay v.1.01, http://biomet.ucdavis.edu/ ) was used to obtain the accu mulated degree-days for

PAGE 72

72 E. utilis This application calculates the accumulate d degree-days using the single sine method, which employs the upper and lower temperature de velopmental thresholds of an organism, and daily average of minimum and maximum temperatures (Baskerville and Emin 1969) The number of generations per year was predicted by dividing the cumulative degree-days per station by K, the degree-day requirement for egg to adult development. A spreadsheet was generated with the weather station name, latitude, longitude and number of E. utilis generations per year for each state (Flori da, Texas, Arizona, California and Hawaii), saved as IV dBase file and imported into Ar cGis 9.0 (ESRI Inc., Redlands, CA, USA). The imported file was converted into a shapefile using the ADD X-Y DATA function. A shapefile of each state was obtained from the Continental USA database of AWhere (AWHERE, Inc., Denver, CO, USA) to project the geographic ra nge of predictions. The ArcGis Geostatistical Analyst function (ESRI Inc., Redlands, CA, USA) was used to generate prediction grids of E. utilis generations per year. The Inverse Distance Weighted (IDW) deterministic method was used to generate weighted averages of nearby known values, and predict values at unsampled locations (Pilkington and Hoddle 2006). One of the advantages of the IDW methods is that it assigns closer values more wei ght to the predicted value than those that are further away. Results Survival and Developmenta l Tim e Calculations Survival of neonate larvae to the adult stage of E. utilis was obtained at 15, 20, 25, and 30C, but all larvae died at 10, 33, and 35C (Fig. 4-1). The highe st percent surv ival (62%) was recorded at 20C followed by 25 and 30C (44%), and the lowest was obtained at 15C (2%) (Fig.4-1). Time of development of each insect st age decreased with increasing temperatures from 15 to 30C (eggs: F4,1083 = 9625.2, P < 0.0001; larvae: F3,96 = 2289.9, P < 0.0001; pupae: F3,77 =

PAGE 73

73 268.5, P < 0.0001; Table 4-1), and duration of developm ent to adult was longest (136.6 days) at 15oC and shortest (30.94 days) at 30oC. Temperature Thresholds and Degree-day Calculations The linear regression p arameters for each insect stage and all immature stages combined for E. utilis are shown in Table 4-2. The lower temper ature thresholds vari ed from 8.17C for eggs to 11.46C for pupae, and for all stages combined (egg-adult) was 9.6oC (Table 4-2). The degree-day requirements for E. utilis varied from 94.34 for eggs to 384.61 for all larval stages combined, and the degree-day requirement from egg to adult was 588.23 (Table 4-2). The parameter estimates for the Logan nonlin ear model for each insect stage and all immature stages combined of E. utilis are shown in Table 4-3. The upper temperature threshold varied from 32C for the larval stage and 35 C for the egg stage; the upper threshold for complete development was predicted to be 33C (Table 4-3). The rate of development increased with temperature until the curve reached an optimum between 28 and 31oC, followed by a rapid decrease as temperatures reached the upper temperature threshold (Fig. 4-2). Cold Tolerance of E. utilis Survival of all insect stages of E. utilis was reduced when exposed to cold tem peratures (Table 4-4; Fig. 4-3). Temperat ure and exposure time, and their interaction, had a significant effect on survival of eggs, pupae and adults (eggs: temperature F2,45 = 489.7, P < 0.0001; exposure time F4,45 = 44.1, P < 0.0001; temperature x time F8,45 = 11.6, P < 0.0001; pupae: F2,79 = 47.4, P < 0.0001; F9,79 = 48.3, P < 0.0001; F8,79 = 5.9, P < 0.0001; adults: F2,45 = 105.2, P < 0.0001; F4,45 = 46.8, P < 0.0001; F8,45 = 19.1, P < 0.0001 respectively; Table 4-4; Fig. 4-3). Survival of larvae decreased with exposure ti me but no differences were detected between temperatures (temperature F2,45 = 1.7, P = 0.2; exposure time F4,45 = 65.6, P < 0.0001; temperature x time F8,45 = 2.3, P = 0.03; Table 4-4). The egg stage was the most sensitive to cold

PAGE 74

74 temperatures, since all eggs died after 2 days exposure to 0C (Table 4-4). The pupal stage was the most tolerant to cold temperatures; therefore, this stage was considered the most plausible over-wintering stage and was used for further anal ysis. Pupal survival was high after 8 days of exposure at 5 and 10C, with an average of 68.7 and 87.5% respectiv ely (Fig. 4-3). In contrast, survival at 0oC decreased rapidly after 4 days with an average survival of 18.7% after 8 days (Fig. 4-3). The pupal lethal times at Ltime50 at 0C was 5.38 days and the Ltime90 was 8.82 days, while at 5C, the Ltime50 and Ltime90 were 10.44 and 27.65 days, respectively (Fig. 4-3). GIS Maps to Predict Number of Generations of E. utilis in the USA GIS m aps were generated to predict th e number of generations per year of E. utilis in Florida, Texas, Arizona, California and Hawaii (Fig. 4-4). Because Brazilian peppertree is present only in southern parts of Texas, Arizona, and California, we focused on these areas for further analysis. The number of generations fo llowed a temperature gr adient with 9 to 6 generations per year in warmer areas (e.g., southern Florida, Texa s, and Arizona) compared to 1 to 4 generations in colder areas (e.g., northern Florida, south-east Arizona, and south-central California) (Fig. 4-4). The predicted number of generations per year of E. utilis varied from 0.5 to 9.8 in the Hawaiian Islands, where the insect has been releas ed and established (Fig. 4-4). Permanent establishment will not occur in areas with < 1 generation per year (Fig. 4-4). According to the isothermal lines (Ltime50 and Ltime90 at 0 and 5oC), potential establishment may occur throughout Florida and southern California, and in southern parts of Texas and Arizona (Fig. 4-5). No limitations due to cold temperatures occur in Hawaii since all the isothermal lines are no rth of the Islands. Discussion Climate (e.g., temperature, humidity) is know n to influence insect distribution and population growth, and therefore, the ability of a species to establish in new areas (Crawley

PAGE 75

75 1987, Sutherst 2000, Byrne et al. 2002). Thus, unde rstanding how temperature ranges affect insect development and survival is important fo r predicting areas of establishment and potential distribution of biologica l control agents in proposed areas of introduction. In addition, the mapping of the predicted distribution may help to assess the proximity to native species that may be at risk of non-target effects. This study show ed that temperature aff ected the development of E. utilis a candidate biocontrol agent of Brazilian peppertree in Florida. Developmental times decreased with increasing temperatures, ranging from 136.6 days at 15C to 30.9 at 30C, and no survival to adulthood was observed at 10, 33, and 35C. Previous studies reported that time of development from egg to adult of E. utilis was 43.6 days at 22C (Mar tin et al. 2004), which is close to the 51.2 days recorded here at 20C. However, this is the first study to evaluate development of E. utilis at different temper atures, providing a bette r understanding of its environmental requirements. Insect development may be different at fluctuating temperatures found in natural environments (Worner 1992, Brakefield and Mazzo tta 1995). However, biol ogical interpretations obtained from linear and non-linear models using constant temper atures have been shown to approximate fluctuating temperature conditions (Worner 1992, Doerr et al. 2002, Liu et al. 2002). Linear and non-linear models were used to determine lower (9.6C) and upper (33C) temperature limits for E. utilis which may be close to the true critical temperatures thresholds for this insect. The lower threshold predicted by the linear model may be underestimated since no survival was obtained at 10C. If we consid er both laboratory data and model predictions together, the true lower threshold may range between 9.6 and 15C, and the upper threshold between 30 and 33C. Temperatures such as 9.6 or 33C are present only for several hours a day in Florida (depending on the location), which most likely will allow E. utilis to survive at these

PAGE 76

76 extremes. Furthermore, during times with unfa vorable weather conditions, insects can locate microclimates that are suitable for survival (Ferro and Southwick 1984, Wilhoit et al. 1991, Chown and Crafford 1992). Simila r studies have been conducted on other lepidopteran species (e.g., Kim et al. 2001, Doerr et al. 2002, Liu et al 2002, Nabeta et al. 2005). For example, the lower (8.4C) and upper thresholds (34C) for de velopment and degree-days from egg to adult (500) were calculated for the tortricid moth Endopiza viteana (Clemens) in the laboratory (Tobin et al. 2001). GIS models that employ both laboratory develo pmental data and long-term climate records can be useful for predicting favorable areas for permanent establishmen t of biological control agents (Pilkington and Hoddle 2006, Diaz et al. in press). In this study, the number of generations per year for E. utilis varied from 0.5 to 9.8 depend ing on the location. Clearly, in areas where <1 generation per ye ar is possible, permanent esta blishment is precluded. Brazilian peppertree is mainly present in south and central Florida, but recent reports indicate that this plant is now found in more north ern parts of the state (norther n peninsula near Georgia border and Florida Panhandle) (Meise nberg 2007). The models pred icted a range of 5.8 to 9.7 generations per year for E. utilis throughout Florida, which indicat es the existence of favorable conditions for establishment of this agent. If a release permit is eventually issued by APHIS, the potential dispersal and spread of E. utilis to states other than Flor ida should be considered. Although Brazilian peppertree is still grown as an ornamental in other states (e.g., California, Texas, and Arizona), this plant has been recogn ized as an invasive species in Hawaii (Morton 1978, Randall 1993), California (Randall 2000), a nd Texas (Gonzalez and Christoffersen 2006). Therefore, the establishment of E. utilis in other states may help to control Brazilian peppertree in these locations. During periods of unfavorab le conditions (e.g., winter ), many lepidopteran

PAGE 77

77 species enter diapause (Nagarkatti et al. 2001, Goehring and Oberhauser 2002, Sims 2007). Therefore, if E. utilis does diapause, the predicted numbers of generations obtained here may be overestimated. However, there is no evidence of diapause in Hawaii or the native range. In addition, the pre-ovi position period of E. utilis at different temperatur es was not evaluated, and including this variable in the model may improve the predicti on. According to Martin et al. (2004), the peak of egg production for E. utilis at 22C occurred 2 days after adult emergence; thus, the pre-oviposition period for this species is quite short. Nevertheless, the calculation of number of generations provides an indicator of potenti al efficacy of the biocontrol agent at a specific location (Hart et al. 2002, Mills 2005, Pi lkington and Hoddle 2006). Therefore, those locations capable of produc ing many generations of E. utilis (e.g., south Florida, south Texas) may experience greater control of the target weed compared to those areas with fewer generations (e.g., south-east Arizona south-central California). In Hawaii, the predicted number of generations of E. utilis ranged from 0.5 to 9.8, depending prim arily on altitude. According to Yoshioka and Markin (1991), this insect species is widely distributed on Brazilian peppertree in the Hawaiian Islands, but the population dynamics (e.g., generations per year) have not been evaluated. In addition to temperature-dependent developm ent studies, the ability of an insect to tolerate unfavorable conditions such as low te mperatures also should be considered when predicting establishment into new areas (MacD onald et al. 1999, Bale et al. 2002, Chen and Kang 2005, Coetzee et al. 2007, Lapointe et al. 2007). Exposure to low temperatures (5 and 0C) affected the survival of all life stages of E. utilis and relative susceptibi lity was eggs > adults> larvae > pupae. Eggs were the most sensitive st age with 100% mortality after 2 days at 0C, while pupae were the most cold-tolerant stage wi th 100% mortality occurr ing after 12 days at

PAGE 78

78 0C. Therefore, we considered pup ae as the over-wintering stage of E. utilis and used the pupal lethal times (Ltime50 and Ltime90) to predict the northern limit of the species distribution. The Ltime50 was 5 days and Ltime90 was 9 days at 0C, while at 5C Ltime50 and Ltime90 were 10 and 28 days, respectively. Based on the isotherm al lines, potential establishment of E. utilis may occur throughout Florida and southern parts of California, Texas and Arizona. In addition, favorable conditions are present throughout the Hawaiian archipelago, which permitted the establishment of E. utilis. The result for Hawaii has been vali dated since the insect is known to be widely established there (Goeden 1977, Yo shioka and Markin 1991, Julien and Griffiths 1998). Climate matching between the region of orig in and region of introduction is important when selecting candidate agents for biological control (Sutherst and Maywald 1985, Byrne et al. 2003, Senaratne et al. 2006). Florida is located betw een the latitudes 25 and 30N, and, therefore, may provide suitable conditions for E. utilis establishment since this insect was collected between the same range of latitudes in the southern hemisphere (Martin et al. 2004). However, models (e.g., CLIMEX) that employ solely a co mbination of climatic variables to predict potential distributions will identif y climatic homologues without reference to a particular species (Davis et al. 1998a, 1998b, Baker et al. 2000). Ther efore, more accurate predictions may be achieved by selecting climatic variables (e.g., te mperature) and their thresholds according to biological responses of a particular species determined by experimentation (Sharpe and deMichele 1977, Baker et al. 2000, Pilkington and Hoddle 2006). Results from this study provide a better understanding of the temperature requirements of E. utilis In addition, GIS mapping is useful for predicting the population dynamics a nd performance of this species throughout the introduced range of Brazilian peppertree in th e USA. However, complementary field studies

PAGE 79

79 should be conducted to determine other key envi ronmental factors (e.g., relative humidity) that may affect survival of E. utilis and incorporate this informa tion into the GIS mapping program to improve the prediction for establishment of this species in the USA.

PAGE 80

80 Table 4-1. Mean ( SE) developmental time (days) of immature stages of E. utilis at five constant temperatures. Different letters in the same row indicate statistical differences P < 0.05. 15C 20C 25C 30C 35C Eggs 16.55 0.09 a N=220 7.93 0.04 b N=244 4.95 0.03 c N=222 4.09 0.01 d N=232 4.18 0.03 d N=170 Larvae (1-5) 85.57 1.34 a N=7 26.85 0.37 b N=35 20.00 0.28 c N=31 19.40 0.29 c N=27 Pupae 34.50 0.50 a N=2 16.41 0.33 b N=31 9.34 0.37 c N=26 7.45 0.21 d N=22 Egg to adult 136.62 51.19 34.29 30.94

PAGE 81

81 Table 4-2. Linear regression parameters es timates describing the relationship between temperature and developmental rate (1/D) of E. utilis stages Stage Intercept Slope R2 N Lower Threshold (C) Degree Days Eggs -0.0866 0.0106 0.94 1087 8.17 94.34 Larvae (1-5) -0.0219 0.0026 0.85 99 8.42 384.61 Pupae -0.0871 0.0076 0.99 80 11.46 131.58 Egg to adult -0.0164 0.0017 0.95 80 9.6 588.23

PAGE 82

82 Table 4-3. Logan non-linear model parameters for developmental rate of E. utilis = maximum developmental rate, = empirical constant, TL= u pper temperature threshold, T= temperature range at which physiologi cal breakdown becomes the overriding influence (Logan 1976). Stage Parameter Estimate ( SE) R2 Egg 0.021 (0.011) 0.979 0.086 (0.022) TL 35.005 (0.650) T 1.716 (0.619) Larvae 0.002 (0.003) 0.966 0.133 (0.088) TL 32.005 (0.200) T 2.497 (1.919) Pupae 0.002 (0.002) 0.988 0.1702 (0.071) TL 33.002 (0.098) 2.869 (1.491) Egg to adult 0.0008 (0.0006) 0.989 0.169 (0.178) TL 33.003 (0.142) T 3.895 (4.425)

PAGE 83

83 Table 4-4. Mean ( SE) percentage surv ival of eggs, larvae, and adults of E. utilis at different temp eratures and exposure times (days) Exposure time (days) Eggs 10C 5C 0C Larvae V 10C 5C 0C Adults 10C 5C 0C 0.5 94.6 2.0 57.3 8.6 34.2 4.1 93.7 6.2 93.7 6.2 93.7 6.2 100 0 100 0 87.5 12.5 1 89.5 3.4 61.8 3.1 9.0 3.2 75.0 .2 75.0 10.2 93.7 6.2 100 0 100 0 81.2 6.2 2 93.2 2.9 21.9 9.1 0 0 75.0 0 68.7 6.2 68.7 6.2 100 0 93.7 6.2 68.7 6.2 4 77.4 4.7 2.1 2.1 0 0 68.7 11.9 68.7 6.2 56.2 6.2 93.7 6.2 93.7 6.2 0 0 8 81.6 3.4 0 0 0 0 31.2 6.2 0 0 0 0 100 0 31.2 11.9 0 0

PAGE 84

84 0 10 20 30 40 50 60 70 80 90 1001918273671120 days % survival 10 15 20 25 30 33 35 Fig. 4-1. Percentage survival from larvae to adult of E. utilis at seven constant temperatures.

PAGE 85

85 0 0.05 0.1 0.15 0.2 0.25 0.3 91419242934 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 1015202530 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 1015202530 0 0.01 0.02 0.03 0.04 1015202530 Figure 4-2. Developmental rates (Days -1) of each immature stage of E. utilis A) eggs, B) larvae, C) pupae, (D) total immature stages. Single dots are observed values and solid lines are Logan non-linear model predictions. Develo p mental rate ( Da y s-1 ) Temperatures ( C) (A) (B) (C) (D)

PAGE 86

86 Exposure time (days) 0510152025 0 20 40 60 80 100 T=0oC Percent pupal survival 0 20 40 60 80 100 0 20 40 60 80 100 T=5oC T=10oC Figure 4-3. Pupal survival of E. utilis at different exposure times and temperatures (10, 5 and 0 C). Single dots are observed values and lines are the expected value of the logistic regression. The data obtained at 10 C did not follow a logistic regression so no line was plotted.

PAGE 87

87 Figure 4-4. Predicted number of generations per year for E. utilis in the USA. A) Florida, B) California-Arizona, C) Texas, D) Hawaii. (A) (D) (B) (C)

PAGE 88

88 Figure 4-5. Map showing the isothermal lines (Ltime50 and Ltime90) at 0 and 5C for E. utilis in the USA. 5C: Ltime50 (10 d) 5C: Ltime90 (28 d) 0C: Ltime50 (5 d) 0C: Ltime90 (9 d)

PAGE 89

89 CHAPTER 5 EFFECT OF HERBIVORY ON GROWTH A ND BIOM ASS OF BRAZILIAN PEPPERTREE SEEDLINGS Introduction Brazilian peppertree, Schinus terebinthifolius R addi (Anacardiaceae) is native to South America (Barkley 1944, 1957), and was introduced in to Florida as an ornamental plant in the 1840s (Mack 1991, Cuda et al. 2006). In the USA, Brazilian peppertree occurs in Hawaii, California, Arizona, Texas and Florida (Habeck et al. 1994, Cuda et al. 2006). This noxious weed invades a variety of habitat types in Flor ida, including disturbed sites (e.g., canals, fallow farmlands) and natural communities (e.g., pi nelands, hardwood hammocks and mangrove forests) (Morton 1978, Cuda et al 2006). Brazilian peppertree is characterized by having a large numbers of fruit per female plant, an effective dispersal mechanism by birds (Panetta and McKee 1997), tolerance to shade (Ewel 1978), fire (Dor en et al. 1991), saline conditions (Ewe 2001, Ewe and Sternberg 2002) and dr ought (Nilson and Mulle r 1980b), and allelopathic effect on neighboring plants (Gogue et al. 1974, Nilson and Muller 1980a Morgan and Overholt 2006). Herbicides and mechanical methods (e.g., cutt ing, burning and flooding) are routinely used for controlling existing Brazilian peppertree stands (Gioeli and Langeland 1997, Cuda et al. 2006). However, these methods are labor intensive and not cost effective especially for large infestations. A classical biological control program was initiated in the 1980s, and the leaflet roller moth Episimus utilis Zimmerman (Lepidoptera: Tortricida e) was selected as a potential biocontrol agent (Martin et al. 2004, Cuda et al. 2006). The larval stage of E. utilis has five instars but occasionally six; larvae feed on plan t leaflets and can completely defoliate small plants (Martin et al. 2004). Adults are small, grayish brown in color and have distinct wing patterns that separate males from females (Zimmerman 1978). The entire life cycle of E. utilis occurs in the canopy of Brazilian peppertree, which may be a favorable characteristic for

PAGE 90

90 establishment of the insect in areas subjected to regular fl ooding in south Florida. Although E. utilis was released and established in Hawaii in the 1950s, control of Brazilian peppertree was not achieved (Goeden 1977, Yoshioka and Markin 1991). Several biotic and perhaps abiotic factors may explain this outcome. For example, high levels of larval parasitism and predation were reported in the field soon after the release of E. utilis in Hawaii, which undoubtedly affected its performance (Davis 1959, Krau ss 1963, Goeden 1977, Julien and Griffiths 1998). During the process of agent selection, several factors are usually considered such as host specificity (Follett and Duan 1999, van Driesche et al. 2000, Sheppard et al. 2003), climatic adaptability (Wapshere 1983, Byrne et al. 2003, Sena ratne et al. 2006), and impact on the target weed (McEvoy and Coombs 1999, Briese et al. 2002, van Klinken and Raghu 2006). More recently, several authors have emphasized the impor tance of measuring the herbivores ability to suppress the target weed in or der to select the most effec tive biocontrol agent (McEvoy and Coombs 1999, Balcianus 2000, Pratt et al. 2005, van Klinken and Raghu 2006). According to Treadwell and Cuda (2007), multiple artificial defolia tions significantly affected growth and fruit production of Brazilian peppertree. Artificial defo liation has been frequently used to measure plant responses to herbivor y (Dhileepan et al. 2000, Broughton 2003, Wirf 2006, Treadwell and Cuda 2007). However, plants may respond differently to actual insect damage (Lehtila and Boalt 2004, Schat and Blossey 2005). The objective of this study was to examine th e effect of different levels of herbivore damage on growth and biomass allocation of Brazil ian peppertree seedlings in the laboratory. In addition, plant responses to herbivory were meas ured immediately following insect damage and after a period of 2 months. This information provides a better understand ing of the interaction

PAGE 91

91 between E. utilis and its host plant, especia lly its potential effectivene ss as a biocontrol agent of Brazilian peppertree in Florida. Materials and Methods Plants and Insects Laboratory experim ents were conducted at the Biologi cal Control Research and Containment Laboratory (BCRCL) located at the Indian River Research and Education Center (IRREC) of the University of Florida, Fort Pi erce, FL, USA. Brazilian peppertree plants were grown from seeds collected from the west coast of Florida (haplotype A and hybrid A genotypes). Plants were grown in nursery pots (18 cm height, 17 cm diameter) using potting soil (Fafard #3B mix), and all plants were pla ced in the greenhouse at BCRCL and watered as needed. Plants were fertilized once with 15 g of Osmocote (a slow release fertilizer 15-9-12, NK-P), and 400 ml per pot of liquid ferti lizer (Miracle Grow 24-8-16) monthly. A colony of E. utilis was initiated in August 2006 at the BCRCL, and insects were reared on Brazilian peppertree potted pl ants (all Florida genotypes) inside environmental growth chambers (25 2C, 60-70% RH, 14L: 10D photoperi od) (for more details on rearing see Martin et al. 2004). Insects were originally collected in 2003 in the vicinity of Curitiba located in Parana Province, southern Brazil, and maintained at th e quarantine facility in Gainesville, FL. Voucher specimens of E. utilis were deposited in the Florida St ate Collection of Arthropods, Florida Department of Agriculture and Consumer Services, Gainesville, Florida, USA. Experimental Procedure Brazilian peppertree seedlings (5 m onths old) were used in the experiment. Six plants were harvested, oven-dried at 70C for 1 w eek, and weighed to measure initial aboveground biomass (leaflets, stems). Three treatments were established (12 replicat es per treatment): 1) control (no herbivory), 2) lo w herbivory (15 larvae/plant) and 3) high herbivory (30

PAGE 92

92 larvae/plant). Neonate larvae (0, 15, or 30) of E. utilis were placed on each treatment plant, and covered by a clear acrylic cylinder (45 cm height, 15 cm diameter) with six holes (6 cm diameter) and tops covered by a fine mesh to allow air circulation. Pl ants were assigned randomly to each treatment. Experiments were conducted inside an environmental growth chamber (25 2C, 60-70% RH, 14L: 10D photope riod) at the BCRCL. U pon adult emergence, all insects were removed from the cylinders and si x plants from each treatment were transplanted to larger pots (22 cm height, 22 cm diameter pots) to allow continued growth. The total number of adults emerged from each treatment was recorded in order to corroborate final insect densities per treatment. Plant parameters were measured at two different times during the experiment: 1) immediately after adults emerged (six plants per treatment), and 2) two months following exposure to herbivory (six plants per treatment). The following plant parameters were measured: 1) number of leaflets, 2) height (cm), 3) basal stem diameter (cm), 4) biomass (g) (roots, stems, and leaflets), and 5) shoot: root ratio (g g-1). For the biomass parameter, plants were harvested, oven-dried at 70C for 1 week, and weighed (separately for total leaflets, stems and roots) at each time interval. Relative plant growth rate (RGR) wa s separately calculated for leaflets and stems using the following index (Wirf 2006): RGR= [log final biomass log initial biomass] / months Data Analysis Plant parameters (e.g., height, biom ass, etc) were compared among herbivory treatments (control, low, and high) using one-way analysis of variance (ANOVA) (SAS Institute 1999). The shoot: root ratio was log transformed prior to th e statistical analysis (Zar 1999). The relative growth rate (RGR) was compared between herbi vory treatment and time interval (immediately after herbivory and 2-months following herb ivory) using two-way ANOVA (SAS Institute

PAGE 93

93 1999). Since there was a significant inte raction between treatment and time ( P < 0.05), one-way ANOVA was used to compare plant treatments sepa rately for each time interval (immediately herbivory and 2-months following herbivory). The total number of E. utilis adults emerging from low and high herbivory treatments was analyz ed using one-way ANOVA (SAS Institute 1999). Means were separated using the post-hoc Stude nt-Neuman-Keuls (SNK) test (SAS Institute 1999). A significance level of = 0.05 was used for all statistical analyses. Results The total number of adults recovered at the end of the experiment differed between herbivory treatm ents (F1,23 = 94.91, P < 0.0001). Whereas an average of 3.2 0.7 adults emerged in the low herbivory treatment, 12.1 0.6 adult moths were recovered from the high herbivory treatment. These data corroborate that two different herbivory leve ls were established. Brazilian peppertree plants exposed to the high herbivory treatments had fewer leaflets compared to the low herbivory and control treatment immediately following the herbivory event (Table 5-1). In addition, plant height and number of leaflets were significantly lower on low and high herbivory treatments than on the control two months follo wing insect damage (Table 5-1). Basal stem diameter, however, was not affect ed by herbivory (Table 5-1). Biomass allocation of Brazilian peppertree al so differed between herbivory treatments (Table 5-2). Plants exposed to high herbivory tr eatments had a lower leaflet biomass and a lower shoot: root ratio compared to the low herbivor y and control treatments (Table 5-2). However, stem biomass was not affected by herbivory (T able 5-2). Two months after insect damage occurred, plant biomass also differed between tr eatments; leaflet biomass and shoot: root ratio were lower in the high herbivory treatment compared to the cont rol (Table 5-2). In addition, the relative growth rate (RGR) of Brazilian pepp ertree was negatively affected by herbivory (Fig. 5-1). Differences were detect ed for leaflet RGR between treatment and time,

PAGE 94

94 and also its interaction (treatment: F2,35 = 16.88, P < 0.0001; time: F1,35 = 9.15, P = 0.005; treatment x time: F2,35 = 8.29, P = 0.0014). High levels of defoliati on resulted in negative values of RGR for both leaflets (-0.5 0.05 g/month) and stems (0.02 0.04 g/month) immediately after herbivory occurred (Fig. 51). Moreover, Brazilian peppertree plants exposed to high levels of defoliation had lower leaflet RGR compared to control plants (no he rbivory) immediately after herbivory and 2 months fo llowing insect damage (Fig. 5-1a). Even though no differences were detected in the stem RGR between treatment and time (treatment: F2,35 = 2.95, P = 0.06; time: F1,35 = 0.06, P = 0.8; treatment x time: F2,35 = 0.99, P = 0.38), slightly higher RGR was detected on plants exposed to lo w herbivory immediately followi ng insect damage (Fig. 5-1b). Discussion Classic al biological control has been successful in controlling many invasive weeds in both aquatic and terrestrial ecosystems (Harley 1990, Julien and Griffith 1998, Grevstad 2006, Barton et al. 2007). However, biological control programs often have been criticized for their lack of predictability in terms of ag ent establishment and success (Ehl er 1990, Harris 1998). Therefore, evaluating the effectiveness of biocontrol agents against the target weed is important during the process of agent selection to maximize the su ccess of biological control programs (Balciunas 2000, Raghu and Dhileepan 2005, van Klinken and Ra ghu 2006). This study showed that growth and biomass allocation of Brazilian peppertree seedlings were negatively affected by herbivory caused by E. utilis Plants exposed to high levels of defoliation (30 larvae/plant) suffered a significant reduction in number of leaflets, plant height, foliar bi omass, and relative growth rate (RGR) compared to control plants (no he rbivory). Because aboveground architecture (e.g., number of leaflets) is important for determini ng the photosynthetic capacity and growth rates of plants (Pearcy et al. 1987), de foliation by the biocontrol agent E. utilis clearly affected the performance of Brazilian peppertree in the labora tory. The decreased plant growth may translate

PAGE 95

95 into a reduced competitive ability of Brazilian peppertree after herbivory, which may allow reestablishment of native plant species in the field. Terrestrial plants have evolved different strategies to cope with herbivory (Rosenthal and Kotanen 1994, Strauss and Agrawal 1999). Plants may compensate for the removal of plant tissue by changing biomass allo cation patterns following herb ivore damage (Belsky 1986, Schierenbeck et al. 1994, Gavloski and Lamb 2000). For example, Melaleuca quinquenervia (Cav.) S.T. Blake partially compensated for herbivory by producing new stems and replacing foliage, but reproduction was still reduced (Pratt et al. 2005). There is usually a balance between plant biomass invested in the shoots and that invested in the roots (Pearsall 1927, Wilson 1988), and many plants can restore the shoot: root ratio after herbivory (Chapin 1980, Paige and Whitnam 1987). Results obtained here showed that the shoot: root ra tio of the invasive Brazilian peppertree was reduced after exposed to high levels of defoliation, and this ratio was not restored after 2 months. In addition, foliar RGR also decreased in the high herbivory treatment compared to control plants (no herbivory). Even though plant reproduction was not measured here, the reduced growth of Brazilian peppertree seedlings ma y translate into lower resource allocation to fruit production and decreased plant fitness. This indirect effect of herbivory on Brazilian peppertree reproduction was osbser ved by Treadwell and Cuda (2007). Plant compensatory growth may vary with the amount and timing of insect damage (Maschinski and Whitman 1989, Thomson et al. 2003, Schooler and McEvoy 2006). In this case, low levels of defoliation (15 larvae/plant) did not affect Brazilian pepper tree growth and biomass allocation in the laboratory. Th erefore, the effectiveness of E. utilis in suppressing this weed will vary in relation to insect densities pres ent in the field. Similarly, the damage on Lythrum salicaria L. was proportional to densi ties of the biocontrol agent Galerucella pusilla Duftschmid

PAGE 96

96 (Schooler and McEvoy 2006). According to Tr eadwell and Cuda (2007), mature Brazilian peppertree plants were able to compensate fo r a single artificial defoliation, but multiple defoliations significantly reduced fruit production in the field. Young plants are usually more vulnerable to insect damage, and this may be the case for Brazilian peppertree seedlings that suffer a reduction of growth and biomass fo llowing one defoliation event (high herbivory treatments). However, the environmental conditions (biotic and abiotic) present in the introduced range also will influence the plants ability to tolerate herbivory (Cox and McEvoy 1983, Belsky 1986, Osier and Lindroth 2004). For example, Br azilian peppertree plants growing in high fertility soils (e.g., abandoned farm lands in the Everglades) may be able to recover faster from defoliation by increasing resource a llocation towards aboveground biomass. In summary, Brazilian peppertree seedlings were negatively affected by high levels of defoliation in the laboratory. Plant parameters su ch as height, foliar biomass, RGR, and shoot: root ratio were reduced immediat ely after herbivory o ccurred. In addition, Brazilian peppertree seedlings were not able to compensate for in sect damage 2 months following herbivory. The ability of plants to compensate for herbivory has direct implications for biological control, and should be considered when examining effectivene ss of potential agents. The leaflet roller moth E. utilis may be an effective biocontrol agent agai nst Brazilian peppertree in Florida. Further studies are needed to provide a better understanding of th e effect of this agent on its host-plant, in particular the potential to reduce plant populations in th e field (Halpern and Underwood 2006). Thus, not only individual plan t parameters such as those m easured in this study, but also population parameters (e.g., fruit production, seedli ng recruitment) should be evaluated in order to help predict the outcome of biocontrol programs.

PAGE 97

97 Table 5-1. Morphometric plant parameters (mean SE) of Brazilian peppertree exposed to different levels of herbivory. Control= no herbivory, low=15 larvae/plant, high=30 larvae/plant. Different letters in the same row indicate statistical differences, df = 2, 17 ( P < 0.05). Control Low High F P After herbivory Number of leaflets Plant height (cm) Basal stem diameter (cm) 66.16 2.9 a 55.5 2.85 0.64 0.07 55.83 6.2 a 57.33 1.62 0.62 0.02 20.83 4.3 b 51.5 2.48 0.49 0.03 25.7 1.59 3.15 <0.0001 0.24 0.07 Two-months after herbivory Number of leaflets Plant height (cm) Basal stem diameter (cm) 119 7.15 a 75.67 2.58 a 0.68 0.03 89.83 12.1 b 67.67 1.67 b 0.65 0.04 73.16 7.82 b 66.5 3.22 b 0.59 0.03 6.25 3.77 1.14 0.01 0.04 0.34

PAGE 98

98 Table 5-2. Biomass allocation (g) (means SE) of Brazilian peppertree exposed to different levels of herbivory. Control=no herbivory, low=15 larvae/plant, high=30 larvae/plant. Different letters in the same row indicate statistical differences, df = 2, 17 ( P < 0.05). Control Low High F P After herbivory Total leaflets Total stems Shoot: root ratio 3.02 0.61 a 2.25 0.45 3.35 0.35 a 3.35 0.43 a 2.52 0.21 3.05 0.26 a 1.01 0.17 b 1.79 0.18 2.31 0.14 b 8.26 1.46 4.11 0.003 0.26 0.03 Two-months after herbivory Total leaflets Total stems Shoot: root ratio 5.23 0.7 a 4.19 0.64 6.44 0.78 a 3.32 0.9 ab 3.28 0.63 4.69 0.47 ab 2.11 0.49 b 2.13 0.42 3.77 0.44 b 4.82 3.24 5.15 0.02 0.06 0.01

PAGE 99

99 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2A fte r h er b i v or y Tw o -mont h s a fterLeaflet RGR (g / month ) C L H b a a ab b -0.1 -0.05 0 0.05 0.1 0.15 0.2Aft e r her b i v or y Tw o -mon th s afterStem RGR (g / month ) C L H a a a a a Figure 5-1. Relative growth rate (RGR) (means SE) of leaflets and stems of Brazilian peppertree exposed to different levels of he rbivory. A) leaflets RGR B) stems RGR. C=control (no herbivory), L=low herbivory (15 larvae/plant), H=high herbivory (30 larvae/plant). Different letters indicate st atistical differences between treatments ( P < 0.05). A B a a

PAGE 100

100 CHAPTER 6 CONCLUDING REMARKS Brazilian peppertree ( Schinus terebinthifolius ) is recognized as one of the m ost widespread exotic plants in Florida (Cuda et al. 2006). This noxious weed form s large monocultures and displaces native vegetation in a variety of habitats throughout central and south Florida (Cuda et al. 2006). Classical biological control has been successful in suppressing many invasive weeds in both terrestrial and aquatic eco systems (e.g., Julien and Griffith 1998, Grevstad 2006, Barton et al. 2007). However, many authors have emphasized the importance of selecting effective biocontrol agents to avoid the release of herbivor es that will fail to suppress the target weed in the introduced range (McEvoy a nd Coombs 1999, Balciunas 2000, Pr att et al. 2005, van Klinken and Raghu 2006). The purpose of this study was to examine the performance of two candidate biocontrol agents of Brazilian peppertree, mo re specifically, a) to determine life-history parameters of the leaflet roller moth Episimus utilis and the thrips Pseudophilothrips ichini on different plant genotypes (chapter 2), b) to evaluate the eff ect of host-plant quality on the development of E. utilis (chapter 3), c) to examine the te mperature requirements and potential distribution of E. utilis in Florida (chapter 4), and d) to measure the impact of E. utilis on growth and biomass allocation of Brazilian peppertree seedlings (chapter 5). Chloroplast DNA (cpDNA) and nuclear micros atellite analyses indicated that two different populations of Brazilian peppertree were introduced sepa rately on the east and west coasts of Florida (Williams et al. 2005, 2007). Ha plotype A is more common on the west coast whereas haplotype B is more common on the ea st coast, and extensive hybridization has occurred between these two types of plants si nce arriving in Florida (Williams et al. 2005, 2007). Results from this study revealed that Br azilian peppertree genotype did not affect the performance of the leaflet roller E. utilis but had a strong effect on the performance of two

PAGE 101

101 populations of the thrips P. ichini (chapter 2). Whereas poor su rvival to adulthood (0-4%) was obtained for the haplotype 5 thrips on all Florida genotypes, higher survival (~50%) was observed when the haplotype 2-3 thrips was fed these Florida genotypes. Moreover, thrips haplotype 5 were unable to lay eg gs on Florida genotypes, whereas thrips haplotype 2-3 laid eggs on all plants tested except for Brazil haplotype D. Therefore, th e thrips haplotype 2-3 that was originally collected from Braz il haplotype A are well adapted to Florida genotypes and should be considered as a potential bioc ontrol agent of Brazilian peppe rtree in Florida. (Note: P. ichini was recently recommended for release by the federa l interagency Technical Advisory Group (TAG) for Biological Control Agents of Weeds (J.P. Cuda, unpubl. rept.)). Because Brazilian peppertree invades different habitats in Florida (e.g., mangrove forests, pinelands, abandoned farms), the environmenta l conditions encountered in these different habitats may influence the survival and effectiven ess of potential biocontrol agents. Results from this study showed that the performance of E. utilis was influenced by host-plant quality (chapter 3). Survival to adulthood (36-55%) was similar when insects were rear ed either on Brazilian peppertree irrigated with freshwater (0 ppt) or with water with a low salinity (6 ppt), but lower survival (6%) was obtained in high salinity-environment (12 ppt). Therefore, Brazilian peppertree plants growing in hi gh-salinity environments (e.g., mangrove forests) may not support high populations of E. utilis Differences in insect performanc e also were detected on Brazilian peppertree plants exposed to diffe rent nutrient levels (c hapter 3). Increased survival to adulthood was observed under medium (25%) and high (40%) nutrient levels compared to low nutrient levels (1.25%). Moreover, faster development (34 d) and higher fertility (88% egg hatch) occurred in the high nutrient treatments, which contained higher leaflet nutrient content (N, P, K), lower toughness and higher moisture conten t of newly expanded leaflets. This study

PAGE 102

102 indicates that Brazilian peppert rees growing in high-fertilit y soils (e.g., disturbed sites, abandoned farmlands; Li and Norland 2001) will provide high-quality hosts for E. utilis development and survival, and insect establishment may occur more rapidly at those sites compared to low-fertility soils. Climate (e.g., temperature) is known to infl uence insect distribution and population growth, and therefore, th e ability of a species to establish in new areas (Crawley 1987, Sutherst 2000, Byrne et al. 2002). According to this st udy, the number of generations per year for E. utilis varied from 0.5 to 9.8 across the entire range of Brazilian peppertree in USA (chapter 4). The models predicted a range of 5.8 to 9.7 generations per year for E. utilis in Florida, which suggests the existence of favorable temperature conditions for establ ishment of this agent. Those locations capable of produc ing many generations of E. utilis (e.g., south Florida, south Texas) may experience greater control of the target weed compared to those areas with fewer generations (e.g., south-east Ariz ona, south-central California). In addition, exposure to low temperatures (5 and 0C) affected the survival of all life stages of E. utilis, and pupae were the most cold-tolerant stage (chapter 4). Based on the isothermal lin es of the pupal lethal exposure times for 50 and 90% of the population, establishment of E. utilis may occur throughout Florida and southern California, but only in the sout hern parts of Texas and Arizona. Moreover, favorable conditions are present throughout the Ha waiian archipelago, which has been validated since the insect is known to be widely estab lished there (Yoshioka and Markin 1991, Julien and Griffiths 1998). During the process of agent selection, eval uating the ability of biocontrol agents to suppress the target weed is important to maxi mize the success of biological control programs (Raghu and Dhileepan 2005, van Klinken and Ra ghu 2006). This study showed that growth and

PAGE 103

103 biomass allocation of Brazilian peppertree seed lings were negatively affected by herbivory (chapter 5). Plants exposed to high levels of defoliation as a result of E. utilis feeding damage (30 larvae/plant) suffered a significant reduction in number of leaflets, plant height, foliar biomass, shoot: root ratio, and relative growth rate compared to control plants (no herbivory). In addition, Brazilian peppertree seedli ngs were not able to compensate for insect damage 2 months after herbivory. However, low levels of defolia tion (15 larvae/plant) di d not affect Brazilian peppertree growth and biomass allocation in th e laboratory. Therefore, the effectiveness of E. utilis to suppress this weed will vary in relation to insect densities present in the field. Overall, this dissertation (c hapters 2-5) contributes to the body of knowledge on weed biological control by providing a better understanding of the interactions between two biocontrol agents and Brazilian peppertree. However, many questions are still unanswered and should be evaluated in the near future. For example, diffe rences in chemical profiles between Brazilian peppertree genotypes may help explain the population genetic structure of P. ichini found in the native range. Even though P. ichini was has been approved to rele ase by TAG, host-specificity of the thrips haplotypes 2-3 may differ from the first population and should be evaluated. In addition, further studies are needed to determine whether th e leaflet roller E. utilis is safe to release in Florida. Long-term field studies shoul d be conducted if these ag ents are released in Florida in order to monitor insect establishmen t and impact on Brazilian peppertree populations and changes in plant community structure over time. The major findings of this study are summarized below: 1. Two candidate biological control agents differe d in their ability to develop on different Brazilian peppertree genot ypes. The leaflet roller E. utilis performed well on all host plant genotypes, whereas two populations of the thrips P. ichini varied in their ability to utilize their host-plant. 2. Different populations of the thrips P. ichini are locally adapted to different Brazilian peppertree genotypes in the native range.

PAGE 104

104 3. The thrips haplotype 2-3, originally collect ed from Brazilian peppertree haplotype A, is well adapted to all Florida ge notypes and should be considered as a potential biocontrol agent of Brazilian peppertree in Florida. 4. Brazilian peppertrees growing in fresh or low saline environments, and also plants present in high-fertility soils, wi ll provide high quality hosts for E. utilis establishment. 5. The best sites for field releases of E. utilis in Florida are fresh upland communities (e.g., pineland forests, along canals ), and high-nutrient environm ents (e.g. disturbed sites, abandoned farmlands that occur in the Everglades). 6. The predicted number of generations of E. utilis will follow a temperature gradient, with more generations per year in warmer areas. 7. The existence of favorable temp erature conditions in Florida for E. utilis development suggests that this agent could successfu lly establish throughout the state. 8. High levels of defoliation caused by E. utilis feeding damage negatively affected growth and biomass allocation of Brazilian peppertree seedlings. 9. The leaflet roller E. utilis may be an effective biologica l control agent of Brazilian peppertree in Florida if the results of host range testing indicate that it is safe to release.

PAGE 105

105 LIST OF REFERENCES Agrawal, A. A., and P. M. Kotanen. 2003. Herbivores and the success of exotic plants: a phylogenetically controlled experi ment. Ecol. Lett. 6: 712-715. Balciunas, J. K. 2000. Code of best practices for classi cal biological control of weeds, p.435. In N. R. Spencer (ed.), Proceedings of the X International Symposium on Biological Control of Weeds. Montana, USA Bale, J. S., G. J., Masters, and I. D. Hodkinson. 2002. Herbivory in global climate change research: direct effects of ri sing temperature on insect herb ivores. Glob. Change Biol. 8: 1. Baker, R. H. A., C. E. Sansford, C. H. Jarvis, R. J. C. Cannon, A. MacLeod, and K. F. A. Walters. 2000. The role of climatic mapping in predicting the potential geographical distribution of non indigenous pe sts under current and future climates. Ecosyst. Environ. 82: 57-71. Bakker, J. D., and S. D. Wilson. 2004. Using ecological restoration to constrain biological invasion. J. Appl. Ecol. 41: 1058-1064. Barkley, F. A. 1944. Schinus L. Brittonia 5, 160-198. Barkley, F. A. 1957. A study of Schinus L., tomo 8. Lilloa Revista de Botanica, Universidad Nacional del Tucuman, Argentina. Barton, J., S. V. Fowler, A. F. Gianotti, C. J. Winks, M. de Beurs, G. C. Arnold, and G. Forrester. 2007. Successful biological control of mist Xower ( Ageratina riparia ) in New Zealand: Agent establishment, impact and bene fits to the native flora. Biol. Control 40: 370-385. Baskerville, G. L., and P. Emin. 1969. Rapid estimation of heat accumulation from maximum and minimum temperatures. Ecology 50: 514-517. Belsky, A. J. 1986. Does herbivory benefit plants? A revi ew of the evidence. Am. Nat. 127: 870892. Bennett, F. D., and D. H. Habeck. 1991. Brazilian peppertree prosp ects for biological control in Florida, pp. 23-33. In T. Center, R. F. Doren, R. L. Hofstetter, R. L. Myers, L. D. Whiteaker (eds.), Proceedings, Symposium of Exotic Pest Plants, 2-4 November 1988, Miami, Florida. Bennett, F. D., L. Crestana, D. H. Habeck, and E. Berti-Filho. 1990. Brazilian peppertree prospects for biologic al control, pp. 293-297. In Proceedings, VII. International Symposium on Biological Cont rol of Weeds, 6-11 March 1988, Rome, Italy. Ministero delAgriculture e delle Foreste, Rome/CSIRO, Melbourne, Australia.

PAGE 106

106 Blair, A. C., and L. M. Wolfe. 2004. The evolution of an invasive plant: an experimental study with Silene latifolia Ecology 85: 3035-3042. Blossey, B., and R. Notzold. 1995. Evolution of increased competitive ability in invasive nonindigenous plants: a hypothesis. J. Ecol. 83: 887-89. Bodle, M. J., A. P. Ferriter, and D. D. Thayer. 1994. The biology, distribution, and ecological consequences of Melaleuca quinquenervia in the Everglades, pp. 341-355. In S. M. Davis, and J. C. Ogden (eds), Ever glades: The Ecosystem and Its Restoration. St. Lucie Press, Delray Beach, Florida. Boecklen, W. J., and S. Mopper. 1998. Local adaptation in specialis t herbivores: theory and evidence, pp. 64-88. In S. Mopper and S. Y. Strauss (eds.), Genetic Structure and Local Adaptation in Natural Insect Populations: Effects of Ecology, Life History, and Behavior. Chapman and Hall, New York. Borchert, D, and R. Magarey. 2005. User manual for NAPPFAST. [cited 2007 Sept 18] ( http://www.nappfast.org/us erm anual/nappfast-manual.pdf ). Bowdish, T. I., and P. Stiling. 1998. The influence of salt and nitr ogen on herbivore abundance: direct and indirect eff ects. Oecologia 113: 400-405. Brakefield, P. M., and V. Mazzotta. 1995. Matching field and laborat ory environments: effects of neglecting daily temperature variation on insect reaction no rms. J. Evol. Biol. 8: 559573. Briere, J. F., P. Pracros, A. Y. Le Roux, and J. S. Pierre. 1999. A novel rate model of temperature-dependent development for Arthropods. Environ. Entomol. 28: 22-29. Briese, D.T., W. Pettit, A. Swirepik, and A. Walker. 2002. A strategy for the biological control of Onopordum spp. thistles in south-easter n Australia. Biocontrol Sci. Techn. 12: 121-137. Briese, D.T., and A. Walker. 2002. A new perspective on the selection of test plants for evaluating the host-specificity of weed biological c ontrol agents: the case of Deuterocampta quadrijuga a potential insect control agent of Heliotropium amplexicaule. Biol. Control 25: 273-287. Broughton, S. 2003. Effect of artificial defoli ation on growth and biomass of Lantana camara L. (Verbenaceae). Plant Protect. Q. 18: 110-115. Bryant, J. P., F. S. Chapin, and D. R. Klein. 1983. Carbon nutrient balance of boreal plants in relation to vertebrate herbivory. Oikos 40: 357-368. Byrne, M. J., S. Currin, and M. P. Hill. 2002. The influence of climate on the establishment and success of the biocontrol agent Gratiana spadicea released on Solanum sisymbriifolium in South Africa. Biol. Control 24: 128-134.

PAGE 107

107 Byrne, M. J., J. Coetzee, A. J. McConnachie, W. Parasram, and M. P. Hill. 2003. Predicting climate compatibility of biol ogical control agents in thei r region of introduction, pp. 2835. In J. M. Cullen, D. T. Briese, D. J. Kr iticos, W.M. Londsdale, L. and Morin, J. K. Scott. (eds.), Proceedings, XI Internationa l Symposium on Biological Control of Weeds, Australia. Campbell, A, B. D. Frazer, N. Gilbert, A. P. Gutierrez, and M. Mackauer. 1974. Temperature requirements of some aphids and their parasites. J. Appl. Ecol. 11: 431. Cappuccino, N., and D. Carpenter. 2005. Invasive exotic plants suffer less herbivory than noninvasive exotic plants. Biol. Lett. 1: 435-438. [Cal-IPC] California Invasive Plant Council. 2006. Invasive plant inventory. ( http://portal.cal-ipc.org/weedlist ) Chapin, F. S. 1980. The mineral nutrition of wild pl ants. Annu. Rev. Ecol. Syst. 11: 233-260. Chen, B., and L. Kang. 2005. Implication of pupal cold tolerance for the northern overwintering range limit of the leafminer Liriomyza sativae (Diptera: Agromyzidae) in China. Appl. Entomol. Zool. 40: 437-446. Chown, S. L., and J. E. Crafford. 1992. Microclimate temperatur es at Marion Island (4654 S, 3745 E). S. Afr. J. Antarct. Res. 22: 37. Cock, M. J. W. 1982. The biology and host specificity of Liothrips mikaniae (Priesner) (Thysanoptera: Phlaeothripidae), a po tential biological control agent of Mikania micrantha ( Compositae). Bull. Entomol. Res. 72: 523-533. Coetzee, J. A., M. J. Byrne, and M. P. Hill. 2007. Predicting the distribution of Eccritotarsus catarinensis a natural enemy released on water hy acinth in South Africa. Entomol. Exp. Appl. 125: 237-247. Cox, G.W. 1999. Alien species in North America and Hawaii: impacts on natural ecosystems. Island Press, Washington, D.C. Cox, C. S., and P.B. McEvoy. 1983. Effect of summer moisture stress on the capacity of tansy ragwort to compensate for defoliation by cinna bar moth (Tyria jacobaeae). J. Appl. Ecol. 20: 225-234. Crawley, M. J. 1987. What makes a community invasi ble? Colonization, Succession and Stability, pp. 429. In Proceedings, 26th Symposium of The British Ecological Society Symposium held Jointly w ith the Linnean Society of London. Cuda, J. P. 2004. Biological control of weeds, pp. 304-308. In J. L. Capinera (ed.) Encyclopedia of Entomology, Kluwer Academic Publishers. Dordrecht, Netherlands.

PAGE 108

108 Cuda, J. P., J. C. Medal, D. H. Habeck, J. H. Pedrosa-Macedo, and M. Vitorino. 1999. Classical biological control of Brazilian peppertree ( Schinus terebinthifolius ) in Florida. ENY-820, University of Flor ida, Cooperative Extension Service. ( http://edis.ifas.ufl. edu/pdffiles/IN/IN11400.pdf ) Cuda, J. P., J. C. Medal, J. H. Pedrosa-Macedo, and D. H. Habeck. 2002. Request for field release of a nonindigenous thrips Pseudophilothrips ichini (Thysanoptera: Phlaeothripidae) for classical biolog ical control of Brazilian peppertree, Schinus terebinthifolius (Anacardiaceae), in Florida (Submitted to IFAS and TAG in October 2002). Cuda, J.P., D.H. Habeck, S.D. Hight, J.C. Medal and J.H. Pedrosa-Macedo. 2004. Brazilian Peppertree, Schinus terebinthifolius : Sumac Family-Anacardiaceae, pp. 439-441. In E.M. Coombs, J.K. Clark, G.L. Piper, and A.F. Co francesco, Jr. (eds.) Bi ological Control of Invasive Plants in the United States. Ore gon State University Press, Corvallis, OR. Cuda, J. P., A. P. Ferriter, V. Manrique, and J. C. Medal. 2006. Floridas Brazilian Peppertree Management Plan. Recommendations from the Brazilian Peppertree task force Florida Exotic Pest Plant Council, 2nd Edition. ( http://www.fleppc.org/Manage_Plans/2006BPm anagePlan5.pdf ) Davies J. T, J. E Ireson, and G. R. Allen. 2005. The impact of gorse thrips, ryegrass competition, and simulated grazing on gorse seedling performance in a controlled environment. Biol. Control 32: 280-286. Davis, C. J. 1959. Recent introductions for biological control in Hawaii-IV. Proc. Hawaiian Entomol. Soc. 17: 62-66. Davis, M.A., and K. Thompson. 2000. Eight ways to be a colonizer; two ways to be an invader: a proposed nomenclature scheme for invasion ecology. Bull. Ecol. Soc. Am 81: 226-30. Davis, A. J., J. H. Lawton, B. Shorrocks, and L. S. Jenkinson. 1998a. Individualistic species responses invalidate simple physiological models of community dynamics under global environmental change. J. Anim. Ecol. 67: 600. Davis, A. J., L. S. Jenkinson, J. H. Lawton, B. Shorrocks, and S. Wood. 1998b. Making mistakes when predicting shifts in species range in response to global warming. Nature (Lond.) 391: 783. DeLoach, C.J. 1997. Biological control of weeds in the United States and Canada, pp.172-194. In J. O. Luken, and J.W. Thieret (eds.) Asse ssment and management of plant invasions. Springer, New York, USA.

PAGE 109

109 Dhileepan, K., and R. E. C. McFadyen. 2001. Effects of gall damage by the introduced biocontrol agent Epiblema strenuana (Lep., Tortricidae) on the weed Parthenium hysterophorus (Asteraceae) J. Appl. Ent. 125: 1-8. Dhileepan, K., S. D. Setter, R. E. McFadyen. 2000. Response of the weed Parthenium hysterophorus (Asteraceae) to defoliation by the introduced biocontrol agent Zygogramma bicolorata (Coleoptera: Chrysomelid ae). Biol. Control 19: 9-16. Diaz, R., W. A. Overholt, J. P. Cuda, P. D. Pratt, and A. Fox. 2008. Temperature-dependent development, survival and potential distribution of Ischnodemus variegatus (Hemiptera: Blissidae), an herbivore of West Indian marsh grass ( Hymenachne amplexicaulis ). Ann. Entomol. Soc. Am. (in press). Dodd, A.P. 1940. The biological campaign against pr ickly pear. Commonwealth Prickly Pear Board, Brisbane. Doerr, M. D., J. F. Brunner, and V. P. Jones. 2002. Temperature-dependent development of Lacanobia subjuncta (Lepidoptera: Noctuidae). Environ. Entomol. 31: 995-999. Doren, R. F., and D. T. Jones. 1997. Management in Everglades National Park, pp. 275286. In D. Simberloff, D.C. Schmitz, and T.C. Brown (eds.), Strangers in Paradise: Impact and Management of N onindigenous Species in Florida. Island Press, Washington, D.C. Doren, R.F., L.D. Whiteaker, and A.M. LaRosa. 1991. Evaluation of fire as a management tool for controlling Schinus terebinthifolius as secondary successional growth on abandoned agricultural land. E nviron. Manage.15: 121-129. Ebert, D. 1994. Virulence and local adaptation of a horizontally transmitted parasite. Science 265: 1084-1086. Edmunds, G. F., and D. N. Alstad. 1978. Coevolution in insect he rbivores and conifers. Science 199: 941-945. Ehler, D. 1990. Introduction strategies in biological control of insects, pp. 111-130. In M. Mackauer, L. E. Ehler, and J. Roland (eds .), Critical Issues in Biological Control. Intercept, Andover. Ellstrand, N. C., and K. A. Schierenbeck. 2000. Hybridization as a stimulus for the evolution of invasiveness in plants ? Proc. Nat. Acad. Sci. 97: 7043-7050. Elton, C.S. 1958. The ecology of invasions by animals and plants. Methuen, London. Emery, S. M., and K. L. Gross. 2005. Effect of timing of pr escribed fire on the demography of an invasive plant, spotted knapweed Centaurea maculosa J. Appl. Ecol. 42: 60-69.

PAGE 110

110 Ewe, S. M. L. 2001. Ecophysiology of Schinus terebinthifolius contrasted with native species in two south Florida ecosystems. Ph.D Dissertation, University of Miami, Florida. Ewe, S. M. L., and S. L. Sternberg. 2002. Seasonal water-use by the invasive exotic, Schinus terebinthifolius in native and disturbed communities. Oecology 133: 441-448. Ewel, J. J. 1978. Ecology of Schinus, pp. 7-21. In Proceedings of techniques for control of Schinus in South Florida, December 2. The Sanibel Captiva Conservation Foundation, Inc., Sanibel, Florida. Ewel, J., D. Ojima, D. Karl, and W. Debusk. 1982. Schinus in successional ecosystems of Everglades National Park. South Florida Re s. Cent. Rep. T-676. Everglades National Park, National Park Service, Homestead, Florida. Feeny, P. P. 1970. Seasonal changes in oak leaf tannins and nutrients as a cause of spring feeding by winter moth cat erpillars. Ecology 51: 565-581. Ferro, D. N., and E. E. Southwick. 1984. Microclimates of small arthropods: estimating humidity within the leaf boundary layer. Environ. Entomol. 15: 999. Follett, P. A., and J. J. Duan. 1999. Nontarget effects of biol ogical control. Kluwer Academic Publishers, Do rdrecht, Netherlands. Gandon, S., and P. A. van Zandt. 1998. Local adaptation and hostparasite interactions. Trend Ecol. Evol. 13: 214-216. Garcia, C.A. 1977. Biologia e aspectos da ecologia e do comportamento defensive comparada de Liothrips ichini Hood 1949 (Thysanoptera Tubulifera). M.S. Thesis, Universidade Federal do Parana, Brazil. Garcia-Barros, E. 2000. Body size, egg size, and interspecifi c relationships with ecological and life history traits in butterflie s (Lepidoptera: Papilionoidae, He sperioidae). Biol. J. Linn. Soc. 70: 251-284. Gavloski, J. E., and R. J. Lamb. 2000. Compensation by cruciferous pl ants is specific to the type of simulated herbivory. Environ. Entomol. 29: 1273-1282. Genton, B. J., P. M. Kotanen, P. O. Cheptou, C. Adolphe, and J. A. Shykoff. 2005. Enemy release but not evolutionary loss of defense in a plant invasion: an intercontinental reciprocal transplant experiment. Oecologia 146: 404-414. Gioeli, K and K. Langeland. 1997 (revised 2003). Brazilian pepper-tree cont rol. University of Florida, Cooperative Extension Service. Inst itute of Food and Agricultural Sciences, SSAGR-17.

PAGE 111

111 Goeden, R. D. 1977. Biological control of weeds, pp.357-414. In Introduced Parasites and Predators of Arthropod Pests and Weeds: A World Review. US. Dep. Agric. Agric. Handb. 480: 1-551. Goehring, L., and K. S. Oberhauser. 2002. Effects of photoperiod, temp erature, and host plant age on induction of reproductive diapause and development time in Danaus plexippus Ecol. Entomol. 27: 674-685. Gogue, G. J., C. J. Hurst, and L. Bancroft. 1974. Growth inhibition by Schinus terebinthifolius. HortScience 9: 301. Gonzalez, L., and B. Christoffersen. 2006. The quiet invasion: A guide to invasive plants of the Galveston Bay area. Galveston Bay Estuary Program, Houston, TX. ( http://www.galvbayinvasives.org ) Goolsby, J. A., P. J. de Barro, J. R. Makins on, R. W. Pemberton, D. M. Hartley, D. R. Frohlich. 2006. Matching the origin of an invasive weed for selection of a herbivore haplotype for a biological control programme. Mol. Ecol. 15: 287-297. Goranson, C. E., and C. K. Ho. 2004. Environmental gradients and herbivore feeding preferences in coas tal salt marshes. O ecologia 140: 591-600. Gordon, D. R. 1998. Effects of Invasive, Non-Indige nous Plant Species on Ecosystem Processes: Lessons from Florida. Ecol. Appl. 8: 975-989. Grevstad, F. S. 2006. Ten-year impact of the biological control agents Galerucella pusilla and G. calmariensis (Coleoptera: Chrysomelid ae) on purple loosestrife ( Lythrum salicaria ) in Central New York State. Biol. Control 39: 1-8. Habeck, D. H., F. D. Bennett, and J. K. Balciunas. 1994. Biological control of terrestrial and wetlandweeds, pp. 523-547. In D. Rosen, F. D. Bennett, and J. L. Capinera (Eds.), Pest Management in the Subtropics: Biological Controla Florida Perspective, Intercept, Andover, UK. Halpern, S. L., and N. Underwood. 2006. Approaches for testing herb ivore effects on plant population dynamics. J. Appl. Ecol. 43: 922-929. Hartley, K. L. S. 1990. The role of biological control in the management of water hyacinth, Eichhornia crassipes. Biocontrol News Inf. 11: 11-22. Hartley, K. L. S., and I. W. Forno. 1992. Biological control of weeds: a handbook for practitioners and students. Melbourne, Inkata.

PAGE 112

112 Harmuch, D. A., J. H. Pedrosa-Macedo, J. P. Cuda and M. D. Vitorino. 2001. Biological aspects of Pseudophilothrips ichini (Hood, 1949) (Thysanoptera, Tubulifera: Phlaeothripidae) in Schinus terebinthifolius Raddi, p. 30. In Abstracts Book and Official Program, VII Symposio de Control Biologico (SICONBIOL), 3-7 June, Pocos de Caldas, Minas Gerais, Brazil. Harris, P. 1991. Classical biocontrol of weeds: its definition, selection of effective agents, and administrative-political pr oblems. Can. Entomol. 123: 827-49. Harris, P. 1998. Evolution of classical weed biocontro l: meeting survival challenges. B. Entomol. Soc. Can. 30: 134-143. Hart, A. J., A. G. Tullett, J. S. Bale, and K. F. A. Walters. 2002. Effects of temperature on the establishment potential in the UK of the nonnative glasshouse biocontrol agent Macrolophus caliginosus Physiol. Entomol. 27: 112. Hemminga, M. A., and J. van Soelen. 1988. Estuarine gradients and the growth and development of Agapanthia villosoviridescens (Coleoptera), a stem-borer of the salt marsh halophyte Aster tripolium Oecologia 77: 307-312. Herms, D. A., and W. J. Mattson. 1992. The dilemma of plants: to grow or defend. Q. Rev. Biol. 67: 283-335. Herrera, A. M., D. D. Dahlsten, N. Tomic-Carruthers, and R. I. Carruthers. 2005. Estimating temperature-dependent developmental rates of Diorhabda elongata (Coleoptera: Chrysomelidae), a biolog ical control agent of saltcedar ( Tamarix spp.). Environ. Entomol. 34: 775-784. Hill RL, GP Markin, AH Gourlay, SV Fowler, and E Yoshioka. 2001. Host range, release, and establishment of Sericothrips staphylinus Haliday (Thysanoptera: Thripidae) as a biological control agent for gorse, Ulex europaeus L. (Fabaceae), in New Zealand and Hawaii. Biol. Control 21:63-74. Hinz, H. L., and H. Muller-Scharer. 2000. Influence of host condition on the performance of Rhopalomyia n. sp. (Diptera: Cecido myiidae), a biological cont rol agent for scentless chamomile, Tripleurospermum perforatum. Biol. Control 18: 147-156. Hobbs, R. J., and S. E. Humphries. 1995. An integrated approach to the ecology and management of plant invasi ons. Conserv. Biol. 9: 761-770. Honek, A. 1993. Intraspecific variation in body size a nd fecundity in inse cts: a general relationship. Oikos 66: 483-492. Hood, J. D. 1949. Brasilian Thysanoptera I. Rev Entomol. 20: 3-88. Hosking, J. R., R. E. McFadyen, and N. D. Murray 1988. Distribution and biological control of cactus species in easte rn Australia. Plant Prot. Quart. 3: 115-123.

PAGE 113

113 Huberty, A. F., and R. F. Denno. 2004. Plant water stress and its c onsequences for herbivorous insects: a new synthesis. Ecology 85: 1383-1398. Hufbauer, R. A., and G. K. Roderick. 2005. Microevolution in biological control: mechanisms, patterns, and proce sses. Biol. Control 35: 227-239. Huffaker, C. B., and C. E. Kennett. 1959. A ten-year study of vegetational changes associated with biological control of Klamath weed. J. Range Manage. 12: 69-82. Hulme, P. E. 2006. Beyond control: wider implications fo r the management of biological invasions. J. Appl Ecol. 43: 835-847. Hunter, M. D. 2001. Multiple approaches to estimating the relative importance of topdown and bottom-up forces on insect populations: experiments, life tables, and timeseries analysis. Basic Appl Ecol 4: 293-310. Inderjit. 2005. Invasive plants: ecological and agricultural aspects. Birkhauser Verlag, Switzerland. Joshi, J., and K. Vrieling. 2005. The enemy release and EICA hypothesis revisited: incorporating the fundamental difference betw een specialist and generalist herbivores. Ecol. Lett. 8: 704-714. Julien, M. H. 1992. Biological control of weeds: A worl d catalogue of agents and their target weeds, 3rd edition. CAB International, Wallingford, UK. Julien, M. H., and M. W. Griffiths. 1998. Biological control of weeds a world catalogue of agents and their target weeds, 4th edition. CABI Publishing, Wallingford, UK. Karban, R. 1989. Fine-scale adaptation of herbivorous thrips to individual host plants. Nature 340: 60-61. Karban, R., and S. Y. Strauss. 1994. Colonization of new host plant individuals by locally adapted thrips. Ecography 17: 82-87. Karban, R., and I. T. Baldwin. 1997. Induced responses to herbi vory. University of Chicago Press, Chicago. Keane, R. M., and M. J. Crawley. 2002. Exotic plant invasions and the enemy release hypothesis. Trends Ecol Evol. 17: 164-170. Kim, D. S., J. E. Lee, and M. S. Yiem. 2001. Temperature-dependent development of Carposina sasakii (Lepidoptera: Carposinidae) and its stage emergence models. Environ. Entomol. 30: 298-305.

PAGE 114

114 Krauss, N. L. 1963. Biological control investigations on Christmas berry ( Schinus terebinthifolius) and Emex ( Emex spp.). Proc. Hawaiian Entomol. Soc. 18: 281-287. Lapointe, S. L., D. M. Borchert, and D. G. Hall. 2007. Effect of low temperatures on mortality and oviposition in conjunction with climate mapping to predict spread of the root weevil Diaprepes abbreviatus and introduced natural en emies. Environ. Entomol. 36: 73-82. Leather, S. R. 1988. Size, reproductive potential and fecundi ty in insects: things aren't as simple as they seem. Oikos 51: 386-389. Lehtila, K. and E. Boalt. 2004. The use and usefulness of artificial herbivory in plantherbivore studies, pp. 258-275. In W. W. Weisser and E. Siemann (eds.), Insects and Ecosystems Function, Springer, Heidelberg. Levine, J., S. Hacker, C. Harley, and M. Bertness. 1998. Nitrogen effects on an interaction chain in a salt marsh community. Oecologia 117: 266-272. Lewis, P. A., C. J. DeLoach, A. E. K nutson, J. L. Tracy, and T. O. Robbins. 2003. Biology of Diorhabda elongate deserticola (Coleoptera: Chrysomelidae), an Asian leaf beetle for biological control of saltcedar s (Tamarix spp.) in the Unite d States. Biol. Control 27: 101-116. Li, Y., and M. Nordland. 2001. The role of soil fertility in invasion of Brazilian pepper ( Schinus terebinthifolius ) in Everglades National Park, Florida. Soil Sci. 166: 400-405. Liu, S. S., F. Z. Chen, and M. P. Zalucki. 2002. Development and survival of the diamondback moth (Lepidoptera: Plutellidae) at constant and altern ating temperatures. Environ. Entomol. 31: 221-231. Logan, J. A., D. J. Wollkind, S. C. Hoyt, and L. K. Tanigoshi. 1976. An analytic model for description of temperature depe ndent rate phenomena in Arthropods. Environ. Entomol. 5: 1133-1140. Mabberley, D.J. 1997. The plant book: a portable dictionary of the vascular plants utilizing Kubitzkis the families and genera of vascul ar plants (1990), Cronqui sts An integrated system of classification of flowering plants (1981), and current botanical literature arranged largely on the principles of ed itions 1-6 (1896/971931) of Williss A dictionary of the flower ing plants and ferns, 2nd edition. Cambridge University Press, UK. MacDonald, J. R., J. S. Bale, and F. A.Walters. 1999. Temperature, development and establishment potential of Thrips palmi (Thysanoptera: Thripidae) in the United Kingdom. Eur. J. Entomol. 96: 169.

PAGE 115

115 Mack, R. N. 1991. The commercial seed trade: an early di sperser of weeds in the United States. Econ. Bot. 45: 257-273. Mack, R. N., D. Simberloff, W. M. Lonsdale, H. Evans, M. Clout, and F. A. Bazzaz. 2000. Biotic invasions: causes, epidemiology, global consequences, and control. Ecol. Appl. 10: 689-710. Maddox, D. M. 1973. Amynothrips andersoni (Thysanoptera: Phlaeothripidae), a thrips for the biological control of alligatorweed. 1. host specificity studies. Environ. Entomol. 2: 30-37. Martin, C. G., J. P. Cuda, K. D. Awadzi, J. C. Medal, D. H. Habeck, and J. H. Pedrosa-Macedo. 2004. Biology and laboratory rearing of Episimus utilis (Lepidoptera: Tortricidae), a candidate for classical bi ological control of Br azilian peppertree, Schinus terebinthifolius (Anacardiaceae), in Florida. Environ. Entomol. 33: 1351-1361. Maschinski, J., and T. G. Whitham. 1989. The continuum of plant responses to herbivory: the influence of plant associa tion, nutrient availability, and timing. Am. Nat. 134: 1-19. Mattson, W. J. Jr. 1980. Herbivory in relation to plant nitrogen conten t. Annu. Rev. Ecol. Syst. 11:119-161. McClay, A. S., and J. K. Balciunas. 2005. The role of pre-release efficacy assessment in selecting classical biological control agents for weeds applying the Anna Karenina principle. Biol. Control 35: 197-207. McEvoy, P.B., and E. M. Coombs. 1999. Biological control of pl ant invaders: regional patterns, field experiments, and structur ed population models. Ecol. Appl. 9: 387-401. McFadyen, R. E. 1992. Biological control against Parthenium weed in Australia. Crop Prot. 24: 400-407. McFadyen, R.E. 2003. Does ecology help in the selection of biocontrol agents?, pp. 5-9. In H. S. Jacob, D. T. Briese, D.T. (eds.), Improving the Selection, Testing and Evaluation of Weed Biological Control Agents. CRC for Australian Weed Management, Glen Osmond, Australia. Meisenberg, M. 2007. Panhandlers beware! Wildland Weeds 11: 22. Mills, N. 2005. Selecting effective parasitoids for biolog ical control introduc tions: Codling moth as a case study. Biol Control 34: 274. Moeri, O. 2007. Application of the F1 Sterile Insect Technique (F1SIT) for field host range testing of Episimus utilis Zimmerman (Lepidoptera: Tortri cidae), a candidate biological control agent of Brazilian peppertree. Master Thesis, University of Florida, Florida.

PAGE 116

116 Moon, D., and P. Stiling. 2000. Relative importance of abiotically induced direct and indirect effects on a salt march herbivore. Ecology 81: 470-481. Moon, D., and P. Stiling. 2002. The effect of salinity and nutri ents on a tritrophic salt-marsh system. Ecology 83: 2465-2476. Mooney, H.A., and R. J. Hobbs. 2000. Invasive species in a changing world, Island Press. Morgan, E. C., and W. A. Overholt. 2005. Potential allelopathic effects of Brazilian pepper ( Schinus terebinthifolius Raddi, Anacardiaceae) aqueous extract on germination and growth of selected Flor ida native plants. J. Torr ey Bot. Soc. 132: 11-15. Morton, J. F. 1978. Brazilian pepper its impact on people, animals and the environment. Econ. Bot. 32: 353-359. Morton, J. F. 1980. The Australian pine or beefwood ( Casuarina equisetifolia L.), an invasive "weed" tree in Florida. P. Fl. St. Hortic. Soc. 93: 87-95. Mound, L. A. 2005. Thysanoptera: diversity and interactions. Annu. Rev. Entomol. 50: 247-69. Mound, L.A., and R. Marullo. 1996. The thrips of Central and S outh America: an introduction (Insecta: Thysanoptera). Mem. Entomol. Int. 6: 1-487. Myers, J. H. 1987. Population outbreaks of introduced inse cts: lessons from biological control of weeds. In P. Barbosa, J. C. Schultz (eds), Insect Outbreaks. Academic Press, New York. Myers, J. H., and D. R. Bazely. 2003. Ecology and control of introduced plants. Cambridge University Press, Cambridge, UK. Myrtinger, L., and G. B. Williamson. 1986. The invasion of Schinus into saline communities of Everglades National Park. Florida Scientists 49: 57-61. Nabeta, F. H., M. Nakai, and Y. Kunimi. 2005. Effects of temperature and photoperiod on the development and reproduction of Adoxophyes honmai (Lepidoptera: Tortricidae). Appl. Entomol. Zool. 40: 231-238. Nagarkatti, S., P. C. Tobin, and M. C. Saunders. 2001. Diapause induction in the grape berry moth (Lepidoptera: Tortricidae) Environ. Entomol. 30: 540-544. Nauman, C. E., and D. F. Austin. 1978. Spread of the exotic fern Lygodium microphyllum in Florida. Am. Fern J. 68: 65-66. Nilson, E. T., and W. H. Muller. 1980a. A comparison of the relative naturalization ability of two Schinus species in southern California. I. S eed germination. Bull. Torrey Bot. Club 107: 51-56.

PAGE 117

117 Nilson, E. T., and W. H. Muller. 1980b. A comparision of the relative naturalizing ability of two Schinus species (Anacardiaceae) in sout hern California. II. Seedling establishment. Bull. Torrey Bot. Club 107: 232-237. Nissen, S. J., R. A. Masters, D. J. Lee, and M. L Rowe. 1995. DNA-based marker systems to determine genetic diversity of weedy species and their application to biocontrol. Weed Sci. 43: 504-513. Nordlund, D. A. 1996. Biological control, integrated pe st management and conceptual models. Biocontrol News Inf 17: 35-44. Osier, T. L., and R. L. Lindroth. 2004. Long-term effects of defoliation on quaking aspen in relation to genotype and nutrient availability: plant growth, phytochemistry and insect performance. Oecologia 139: 55-65. Paige, K. N., and T. G. Whitham. 1987. Overcompensation in response to mammalian herbivory: the advantage of be ing eaten. Am. Nat. 129: 407-416. Panetta, F.D., and J. McKee. 1997. Recruitment of the invasive ornamental, Schinus terebinthifolius, is dependent upon frugivores. Aust. J. Ecol. 22: 432-438. Parker, I.M., D. Simberloff, W. M. Lonsdale, K. Goodell, and M. Wonham. 1999. Impact: toward a framework for understanding the ecological effects of invaders. Biol. Invasions 1:3-19. Pearcy, R. W., 0. Bjorkman, M. M. Caldwell, J. E. Keeley, R. K. Monson, and B. R. Strain. 1987. Carbon gain by plants in natural environments. BioScience 37: 21-29. Pearsall, W. H. 1927. Growth studies. VI. On the relative sizes of growing plant organs. Ann. Bot. 41: 549. Pemberton, R. W., and A. P. Ferriter. 1998. Old world climbing fern ( Lygodium microphyllum ), a dangerous invasive weed in Florida. Am. Fern J. 8: 165-175. Pilkington, L. J., M. S. Hoddle. 2006. Use of life table statistics and degree-day values to predict the invasion success of Gonatocerus ashmeadi (Hymenoptera: Mymaridae), an egg parasitoid of Homalodisca coagulata (Hemiptera: Cicadellidae), in California. Biol. Control 37: 276-283. Pimentel, D. R. Zuniga, and D. Morrison. 2005. Update on the environmental and economic costs associated with alieninvasive species in the Unite d States. Ecol. Econ. 52: 273-288. Pratt, P. D., M. B. Rayamajhi, T. K. Van, T. D. Center, and P. W. Tipping. 2005. Herbivory alters resource allocation and compensation in the invasive tree Melaleuca quinquenervia Ecol. Entomol. 30: 316-326.

PAGE 118

118 Price, P. W. 2000. Host plant quality, insect herb ivores and biocontrol, pp. 583-590. In Proceedings of the X International Symposiu m on Biological Control of Weeds, Montana State University, Bozeman, Montana, USA. Raghu, S., and K. Dhileepan. 2005. The value of simulating herb ivory in selecting effective weed biological control ag ents. Biol. Control 34: 265-273. Randall, J. M. 1993. Exotic weeds in North American and Hawaiian natural areas: The Nature Conservancys plan of attack, pp. 159. In B. N. McKnight (ed.), Biological pollution: the control an d impact of invasive exotic species, Indiana Academy Science, Indianapolis. Randall, J. M. 2000. Schinus terebinthifolius Raddi, pp. 282-287. In C. C. Bossard, J. M. Randall, M. C. Hoshovsky (eds.), Invasive plants of californias wildlands, University of California Press, Berkeley, California. Raupp, M. J. 1985. Effect of leaf toughness on mandibular wear of the leaf beetle, Plagiodera versicolor Ecol. Entomol. 10: 73-79. Rhainds, M., and L. Shipp. 2003. Dispersal of adult western flower thrips (Thysanoptera: Thripidae) on chrysanthemu m plants: impact of feeding-induced senescence of inflorescences. Environ. Entomol. 32: 1056-1065. Rhainds, M., L. Shipp, L. Woo ddrow, and D. Anderson. 2005. Density, dispersal, and feeding impact of western flower thrips (Thysanoptera: Thripidae) on flowering chrysanthemum at different spatia l scales. Ecol. Entomol. 30: 96-104. Richardson, D. M., and P. Pysek. 2008. Fifty years of invasion eco logy the legacy of Charles Elton. Divers ity Distrib. 14: 161-168. Rice, W.R. 1983. Sexual reproduction: an adaptation reducing parent-offspring contagion. Evolution 37: 1317-1320. Rosenthal, J. P., and P. M. Kotanen. 1994. Terrestrial plant tolera nce to herbivory. Trends Ecol. Evol. 9: 145-148. Roy M, J. Brodeur, and C. Cloutier. 2002. Relationship between temperature and developmental rate of Stethorus punctillum (Coleoptera: Coccinellidae) and its prey Tetranychus mcdanieli (Acarina: Tetranychidae). Environ. Entomol. 31: 177-187. SAS Institute. 1999. SAS/STAT Users guide. SAS Institute, Cary, NC. Schat, M., and B. Blossey. 2005. Influence of natural and simulated leaf beetle herbivory on biomass allocation and plan t architecture of purple loosestrife ( Lythrum salicaria L.). Environ. Entomol. 34: 906-914.

PAGE 119

119 Schile, L., and S. Mopper. 2006. The deleterious effects of sa linity stress on leafminers and their freshwater host. Ecol. Entomol. 31: 345-351. Schierenbeck, K. A., R. N. Ma ck, and R. R. Sharitz. 1994. Effects of herbivory on growth and biomass allocation in native and introduced species of Lonicera Ecology 75: 1661-1672. Schooler, S. S., and P. B. McEvoy. 2006. Relationship between insect density and plant damage for the golden loosestrife beetle, Galerucella pusilla on purple loosestrife ( Lythrum salicaria ). Biol. Control 36: 100-105. Schoonhoven L. M., J. J. A. van Loon, and M. Dicke. 2005. Plant as insect food: not the ideal, pp. 99-134. In Insect-Plant Biology, Oxford Univ ersity Press Inc., New York. Schwab, L. K., and S. Raghu. 2006. Nutrient composition of soil and plants may predict the distribution and abundance of specialist insect herbivores: implications for agent selection in weed biological cont rol. Aust. J. Entomol. 45: 345-348. Scriber, J. M. 1979. Effects if leaf-water supplemen tation upon post-ingestive nutritional indices of forb-, shrub-, vine-, and tree feeding Lepidoptera. Entomol. Exp. Appl. 25: 240-55. Senaratne, K. A. D.W., W. A. Palmer, and R. W. Sutherst. 2006. Use of CLIMEX modeling to identify prospective areas for exploration to find new biological control agents for prickly acacia. Aust. J. Entomol. 45: 298-302. Sharpe, P. J. H., and D. W. deMichele. 1977. Reaction kinetics of poikilotherm development. J. Theor. Biol. 64: 649. Sheppard, A.W. 2003. Prioritizing agents based on predicted efficacy: beyond the lottery approach, pp. 11-21. In H. S. Jacob, D. T. Briese (eds.), Improving the selection, testing and evaluation of weed biological control agents. CRC for Australian Weed Management, Glen Osmond, Australia. Sheppard, A. W., R. Hill, R. A. DeClerck-Floate, A. McClay, T. Olckers, P. C. Jr. Quimby, H. G. Zimmermann. 2003. A global review of risk-benefit-cost analysis for the introduction of classical biologi cal control agents against weed s: a crisis in the making? Biocontrol News Inf. 24: 91N-108N. Showers, W. B., R. L. Hellmich, M. E. Derrick-Robinson, and W. H. Hendrix. 2001. Aggregation and dispersal behavior of ma rked and released european corn borer (Lepidoptera: Crambidae) adults Environ. Entomol. 30: 700-710. Siemann, E., and W. E. Rogers. 2001. Genetic differences in grow th of an invasive tree species. Ecol. Lett. 4: 514-518.

PAGE 120

120 Simberloff, D. 1996. Impacts of introduced species in the United States. Consequences: Nat. Implic. Environ. Change 2: 13-22. Sims, S. R. 2007. Diapause dynamics, seasonal phenology, and pupal color dimorphism of Papilio polyxenes in southern Florida, USA. Entomol. Exp. Appl. 123: 239-245. Slansky, F. Jr., and P. Feeny. 1977. Stabilization of the rate of nitrogen accumulation by larvae of the cabbage butterfly on wild and cultivated food plants. Ecol. Mongr. 47: 209-228. Stiling, P, and D. C. Moon. 2005. Quality or quantity: the direct and indirect effects of host plants on herbivores and their natural enemies. Oecologia 142:413-420. Strauss, S. Y., and A. A. Agrawal. 1999. The ecology and evolution of plant tolerance to herbivory. Trends Ecol. Evol 14: 179-185. Strong, D.R., and R. W. Pemberton. 2000. Biological control of i nvading species: risk and reform. Science 288: 1969-1970. Suckling, D. M., J. F. Brunner, G. M. Burnip, and J. T. S. Walker. 1994. Dispersal of Epiphyas postvittana (Walker) and Planotortrix octo Dugdale (Lepidoptera: Tortricidae) at a Canterbury, New Zealand orchar d. New Zeal. J.Crop Hort. 22: 225-234. Sutherst, R. W. 2000. Change and invasive species: A conceptual framework, pp. 211 240. In H. A. Mooney, R. J. Hobbs. (eds.), Inva sive species in a changing world. Island Press, Washington, DC. Sutherst, R. W., and G. E. Maywald. 1985. A computerized system for matching climates in ecology. Agr. Ecosyst. Environ. 13: 281-299. Tilman, D. 1982. Resource competition and community stru cture, Princeton University Press. Thomas, P. A., and P. M. Room. 1986. Towards biological control of Salvinia in Papua New Guinea. In E. S. Delfosse (ed), Proceedings of the VI International Symposium on Biological Control of Weeds.Vancouver, Canada. Thomson, V. P., S. A. Cunningham, M. C. Ball, and A. B. Nicotra. 2003. Compensation for herbivory by Cucumis sativus through increased photosynthe tic capacity and efficiency. Oecologia 134: 167. Tobin, P. C., S. Nagarkatti, M. C. Saunders. 2001. Modeling development in grape berry moth (Lepidoptera: Tortricidae) Environ. Entomol. 30: 692-699. Treadwell, L. W. and J. P. Cuda. 2007. Effects of defoliation on gr owth and reproduction of Brazilian peppertree ( Schinus terebinthifolius ). Weed Sci. 55: 137-142.

PAGE 121

121 Turner, C. E., T. D. Center, D. W. Burrows, and G. R. Buckingham. 1998. Ecology and management of Melaleuca quinquenervia an invader of wetlands in Florida, U.S.A. Wetlands Ecol. Manage. 5: 165-178. van Driesche, R. G., and T. S. Bellows. 1996. Biological control. Chapman and Hall, New York, US. van Driesche, R. G., T. Heard, A. McClay, and R. Reardon. 2000. Host-specificity Testing of Exotic Arthropod Biological C ontrol Agents: The Biological Basis for Improvement of Safety. USDA Forest Service, Morgantown, USA. van Klinken, R. D., and S. Raghu. 2006. A scientific approach to agent selection. Aust. J. Entomol. 45: 253-258. Wang, Y. 2002. Ecological, physiological, a nd molecular responses of Iris hexagona to salinity stress. PhD Dissertation, University of Louisiana, Lafayette, USA. Wang, Y., S. Mopper, and K. H. Hasenstein. 2001. Effects of salinity on endogenous ABA, IAA, JA, and SA in Iris hexagona J. Chem. Ecol 27: 327-342. Wapshere, A. J. 1983. Discovery and testing of a clim atically adapted strain of Longitarsus jacobaeae (Col: Chrysomelidae) for Australia. Entomophaga 28: 27-32. Waring, G. L., and N. S. Cobb. 1992. The impact of plant stress on herbivore population dynamics. In E. Bernays (ed), Insect-plant inte ractions vol 4, CRC Press, Boca Raton, Florida. Wheeler, G. 2001. Host plant quality factors that in fluence the growth and development of Oxyops vitiosa, a biological control agent of Melaleuca quinquenervia Biol. Control 22: 256-264. Wheeler, G., and T. D. Center. 1996. The influence of hydrilla l eaf quality on larval growth and development of the biological control agent Hydrellia pakistanae (Diptera: Ephydridae). Biol Control 7: 1-9. Wilhoit, L. R., R. C. Axtell, R. E. Stinner. 1991. Estimating manure temperatures from air temperatures and results of its use in models of filth fly (Diptera: Muscidae) development. Environ. Entomol. 20: 634. Williams, J. R. 1954. The biological cont rol of weeds, pp. 95-98. In Report of the Sixth Commonwealth Entomological Congress, London, UK. Williams, D. A., D. S. L. Sternberg, C. R. Hughes. 2002. Characterization of polymorphic microsatellite loci in the i nvasive Brazilian peppertree, Schinus terebinthifolius Mol. Ecol. Notes 2: 231-232.

PAGE 122

122 Williams, D. A., W. A. Overholt, J. P. Cuda, and C. R. Hughes. 2005. Chloroplast and microsatellite DNA diversities reveal the introduction history of Brazilian peppertree ( Schinus terebinthifolius ) in Florida. Mol. Ecol. 14: 3643-3656. Williams, D. A., E. Muchugu, E., W. A. Overholt, and J. P. Cuda. 2007. Colonization patterns of the invasive Brazilian peppertree, Schinus terebinthifolius in Florida. Heredity 98: 284-293. Williamson, M. 1996. Biological invasions. Chapman and Hall, London, UK. Willis, A. J., J. Memmott, and R. I. Forrester. 2000. Is there evidence for the postinvasion evolution of increased size among invasive plant sp ecies? Ecol. Lett. 3: 275283. Wilson, J. B. 1988. A review of the evidence on the contro l of shoot:root ratio, in relation to models. Ann. Bot. 61: 433. Wirf, L. 2006. Using simulated herbivory to predict the efficacy of a biocontrol agent: the effect of manual defoliation and Macaria pallidata Warren (Lepidoptera: Geometridae) herbivory on Mimosa pigra seedlings. Aust. J. Entomol. 45: 324-326. Woodall, S. 1979. Physiology of Schinus, pp. 3-6. In R. Workman (ed.), SchinusProceedings of techniques for control of Schinus in South Florida, December 2. The Sanibel Captiva Conservati on Foundation, Inc., Sanibel, FL. Worner, S. P. 1992. Performance of phonological models unde r variable temperature regimes: consequences of the Kaufman or rate su mmation effect. Environ. Entomol. 21: 689-699. Yoshioka, E. R., and G. P. Markin. 1991. Efforts of biological control of Christmas berry Schinus terebinthifolius in Hawaii, pp. 377-385. In T. Center, R. F. Doren, R. L. Hofstetter, R. L. Myers, and L. D. Whiteak er (eds.), Proceedings, Symposium of Exotic Pest Plants, 2-4 November 1988, Miami, Florida. Zar, J. H. 1999. Biostatistical analysis, 4th editio n. Prentice-Hall, Inc., Upper Saddle River, New Jersey. Zimmerman, E. C. 1978. Insects of Hawaii, vol. 9, Microlep idoptera, Part I. Monotrysia, Tineoidea, Tortricoidea, Graci llaroidea, Yponomeutoidea, an d Alucitoidea. University of Hawaii, Honolulu. Zwolfer, H., and M. Preiss. 1983. Host selection and oviposition behavior in West-European ecotypes of Rhinocyllus conicus Froel. (Col., Curculionidae) Z. Angew. Entomol. 95: 113-122.

PAGE 123

123 BIOGRAPHICAL SKETCH Veronica Manrique was born in Buenos Aires, Argentina in 1972. She obtained her B. S. in Biology in May 2000 at the University of Bue nos Aires, Argentina. Her initial experience working in entomology was at the USDA-ARS S outh American Biological Control Laboratory, where she conducted exploratory surveys of bi ocontrol agents of cr op pests in northeast Argentina, and was responsible of the maintenance of several insect colonies. Since then, she has developed a deep interest in classical biological control, pl ant-insect interactions, and management of invasive species. In 2000, Veroni ca worked as a foreign research associate for six months at the USDA-ARS Southern Insect Management Reseach Unit in Stoneville, MS. After that, she started her graduate studies at Texas A&M University, and she obtained her Master degree in Entomology in December 2003. Her masters thesis was entitled Host habitat location mediated by olfactory stimuli in Anaphes iole (Hymenoptera: My maridae), an egg parasitoid of Lygus hesperus (Hemiptera: Miridae). Veronica entered the graduate program at the Department of Entomology & Nematology of the University of Florida in January 2004. During her research, she conducted se veral studies to evaluate the effectiveness of two biocontrol agents of the invasive Brazilian peppertree in Florida. She completed her Dissertation and was granted the degree of Doctor of Philosophy in May 2008.