Field Host Range, Foraging Depth, and Impact Of

FIELD HOST RANGE, FORAGING DEPTH, AND IMPACT OF CRICOTOPUS LEBETIS SUBLETTE (DIPTERA: CHIRONOMIDAE), A BIOLOGICAL CONTROL AGENT OF HYDRILLA VERTICILLATA (L.F.) ROYLE (HYDROCHARITACEAE)

EUTYCHUS MUKURE KARIUKI

A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2017

To my loving family

ACKNOWLEDGMENTS

I would like to thank my Major Advisor, Dr. Raymond L. Hix, and my Co-Advisor,

Dr. James P. Cuda, for their support and guidance during my Ph.D. program. I am also thankful to my committee members, Dr. Stephen D. Hight for his invaluable support and mentorship during the course of my research; Dr. Jennifer Gillett-Kaufman for her constant support, especially through the writing process of my dissertation; and Dr. Lyn

Gettys for always being available to help with questions.

I am grateful to all others who provided their assistance, including Dr. Edzard van

Santen (University of Florida), Dr. Lazarus Mramba (University of Florida), Dr. Emma

Weeks (University of Florida), John Mass (United States Department of Agriculture

(USDA), Tallahassee, Florida), Kelle Sullivan (Florida Fish and Wildlife Conservation

Commission), Dr. Lamberth Kanga (Florida A&M University), and Dr. Muhammad

Haseeb (Florida A&M University). I am thankful to all my colleagues and lab mates at the University of Florida who reviewed this manuscript and offered valuable comments and suggestions.

I am equally thankful to the USDA for providing funding to this study through the

Hydrilla Integrated Pest Management Risk Avoidance and Mitigation Project (IPM

RAMP) grant 2010-02825 and the National Institute of Food and Agriculture Crop

Protection and Pest Management (NIFA CPPM) grant 2014-70006-22517.

TABLE OF CONTENTS

page

ACKNOWLEDGMENTS ...... 4

LIST OF TABLES ...... 8

LIST OF FIGURES ...... 10

LIST OF ABBREVIATIONS ...... 13

ABSTRACT ...... 14

CHAPTER

1 LITERATURE REVIEW ...... 16

Hydrilla ...... 16 Origin and Invasion History of Hydrilla ...... 16 Distribution of Hydrilla in the United States ...... 17 Description of Hydrilla ...... 18 Adaptations of Hydrilla...... 19 Importance of Hydrilla ...... 23 Ecological damage ...... 23 Economic damage ...... 25 Potential benefits ...... 25 Summary...... 27 Management of Hydrilla...... 27 Chemical control ...... 27 Mechanical control ...... 30 Biological control ...... 30 History, Benefits, and Approaches of Biological Control of Aquatic Weeds ...... 35 Cricotopus lebetis as a Biological Control Agent of Hydrilla ...... 37 Background ...... 37 Description ...... 38 Distribution ...... 39 Biology and Feeding Damage ...... 39 Host Range ...... 41 Need for Research ...... 42

2 FIELD HOST SPECIFICITY OF A POTENTIAL HYDRILLA BIOLOGICAL CONTROL AGENT, CRICOTOPUS LEBETIS SUBLETTE (DIPTERA: CHIRONOMIDAE) IN LAKE ISTOKPOGA, FLORIDA ...... 49

Introduction ...... 49 Materials and Methods ...... 51 Study Area ...... 51

Sampling ...... 51 Plant Sample Processing ...... 53 Cricotopus lebetis Rearing ...... 53 Data Analysis ...... 54 Results ...... 54 Discussion ...... 55

3 ASSESSMENT OF MAXIMUM WATER DEPTH CRICOTOPUS LEBETIS SUBLETTE LARVAE CAN LOCATE HYDRILLA VERTICILLATA (L.F.) ROYLE ... 66

Introduction ...... 66 Materials and Methods ...... 69 Greenhouse Experiments ...... 69 Experiment I: Inoculation of neonate Cricotopus lebetis on the water surface ...... 72 Experiment II: Inoculation of Cricotopus lebetis egg masses on the water surface ...... 73 Data collection ...... 73 Data analyses ...... 74 Field Experiments ...... 74 Experiment I: Use of sentinel stems to determine the vertical foraging depth of Cricotopus lebetis ...... 75 Experiment II: Survey of naturally growing stems to determine the vertical foraging range of Cricotopus lebetis ...... 75 Data collection ...... 76 Data analyses ...... 76 Results ...... 77 Greenhouse Experiments ...... 77 Experiment I: Inoculation of neonate Cricotopus lebetis on the water surface ...... 77 Experiment II: Inoculation of Cricotopus lebetis egg masses on the water surface ...... 78 Field Experiments ...... 79 Experiment I: Use of sentinel stems to determine the vertical foraging range of Cricotopus lebetis ...... 79 Experiment II: Survey of naturally growing stems to determine the vertical foraging range of Cricotopus lebetis ...... 79 Discussion ...... 80

4 COMPETITIVE INTERACTIONS BETWEEN HYDRILLA VERTICILLATA AND THE NATIVE VALLISNERIA AMERICANA AS INFLUENCED BY CRICOTOPUS LEBETIS ...... 97

Introduction ...... 97 Materials and Methods ...... 100 Study Site and Plant Source ...... 100 Set Up and Experimental Design ...... 100

Planting ...... 101 Introduction and monitoring of Cricotopus lebetis ...... 102 Monitoring of shading effect and temperature ...... 102 Plant harvest ...... 102 Data Analyses ...... 103 Results ...... 103 Discussion ...... 106

5 CONCLUSIONS ...... 127

LIST OF REFERENCES ...... 130

BIOGRAPHICAL SKETCH ...... 147

LIST OF TABLES

Table page

1-1 List of herbicides approved by U.S. Environmental Protection Agency for management of Hydrilla verticillata (L.f.) Royle...... 46

1-2 List of four biological control agents approved by the USDA for management of Hydrilla verticillata (L.f.) Royle...... 47

3-1 Description of the treatments in a greenhouse experiment designed to evaluate how deep Cricotopus lebetis Sublette neonates can swim to locate Hydrilla verticillata (L.f.) Royle at four different depth treatments in artificial water columns...... 84

3-2 Summary of the logistic regression examining the effect of water depth on the proportion of Cricotopus lebetis Sublette adult emergence from Plexiglas tubes inoculated with C. lebetis neonates in greenhouse experiment I...... 85

3-3 Sex ratio (mean proportion males) of Cricotopus lebetis Sublette that emerged from Plexiglas tubes containing Hydrilla verticillata (L.f.) Royle at four depths. Plexiglas tubes were inoculated with C. lebetis neonates in greenhouse experiment I...... 86

3-4 Summary of the logistic regression examining the effect of water depth on the proportion of Cricotopus lebetis Sublette adult emergence from Plexiglas tubes inoculated with C. lebetis eggs in greenhouse experiment II...... 90

3-5 Sex ratio (mean proportion males) of Cricotopus lebetis Sublette that emerged from Plexiglas tubes containing Hydrilla verticillata (L.f.) Royle at one of four depths. Plexiglas tubes were inoculated with C. lebetis eggs in greenhouse experiment II...... 91

3-6 Summary of the logistic regression analysis examining the effects of water depth on the proportion (mean ± 95% CI) of damaged apical meristems and number (mean ± 95% CI) of Cricotopus lebetis Sublette larvae per apical meristems of Hydrilla verticillata (L.f.) Royle on sentinel bouquets placed at 0 m, 0.45 m, and 0.9 m below the water surface in Lake Istokpoga, Florida, 2017...... 95

3-7 Summary of the logistic regression examining the effects of water depth on the proportion (mean ± 95% CI) of damaged apical meristems and number (mean ± 95% CI) of Cricotopus lebetis Sublette larvae per apical meristem of Hydrilla verticillata (L.f.) Royle stems sampled from the surface level, mid- depth, and bottom of Lake Istokpoga, Florida, 2017...... 96

4-1 Description of the treatments in an experiment designed to evaluate the impact of Cricotopus lebetis Sublette on the competitive interaction between Hydrilla verticillata (L.f.) Royle and Vallisneria americana Michx...... 111

4-2 Two-way analysis of variance examining the effects of herbivory by Cricotopus lebetis Sublette and interspecific competition by Vallisneria americana Michx. on the ability of Hydrilla verticillata (L.f.) Royle to produce biomass and turions, elongate, and form surface cover...... 118

LIST OF FIGURES

Figure page

1-1 Map showing distribution of Hydrilla verticillata (L.f.) Royle in natural drainage systems within the United States as of 2016...... 44

1-2 Morphological characteristics of Hydrilla verticillata (L.f.) Royle...... 45

1-3 Cricotopus lebetis Sublette larva exhibiting a diagnostic blue band around the second and third thoracic segments...... 48

2-1 Apical meristem of Hydrilla verticillata (L.f.) Royle as viewed under a dissecting microscope...... 61

2-2 Damage on native aquatic plants caused by larvae of Cricotopus lebetis Sublette...... 62

2-3 Number (mean ± SE) of adults matching the description of Cricotopus lebetis Sublette reared from stem tips of plant samples collected in Lake Istokpoga, Florida, U.S...... 63

2-4 Number (mean ± SE) of larvae of Cricotopus lebetis Sublette per 100 stem tips (or leaves for Vallisneria americana Michx.) recorded from plant material collected at Lake Istokpoga, Florida, U.S...... 64

2-5 The percent (mean ± SE) damaged stem tips by Cricotopus lebetis Sublette from plant material collected at Lake Istokpoga, Florida. U.S...... 65

3-1 Impact of Cricotopus lebetis Sublette on Hydrilla verticillata (L.f.) Royle grown at different water depth treatments in greenhouse experiment I...... 87

3-2 Number of egg masses (mean ± 95% CI) produced by Cricotopus lebetis Sublette that emerged from Hydrilla verticillata (L.f.) Royle grown at different water depth treatments in greenhouse experiment I...... 88

3-3 Cumulative percentage emergence pattern of adult Cricotopus lebetis Sublette following development of the insects in Plexiglas tubes on Hydrilla verticillata (L.f.) Royle at different depths...... 89

3-4 Impact of Cricotopus lebetis Sublette on Hydrilla verticillata (L.f.) Royle grown at different water depth treatments in greenhouse experiment II...... 92

3-5 Number of egg masses (mean ± 95% CI) produced by Cricotopus lebetis Sublette that emerged from Hydrilla verticillata (L.f.) Royle grown at different water depth treatments in greenhouse experiment II...... 93

3-6 Cumulative percentage emergence pattern of adult Cricotopus lebetis Sublette following development of the insect in Plexiglas tubes on Hydrilla verticillata (L.f.) Royle at different depths...... 94

4-1 Morphological characteristics of Vallisneria americana Michx. and Hydrilla verticillata (L.f.) Royle...... 112

4-2 Planting details for monoculture and mixed culture treatments in outdoor concrete tanks...... 113

4-3 Experimental design for competition and competition-herbivory treatments.. ... 114

4-4 Hydrilla verticillata (L.f.) Royle with a dense surface canopy in one of the tanks covered with 70% shade cloth...... 115

4-5 Average weekly water temperature recorded by an underwater sensor, onset® HOBO® Pendant® Data Logger...... 116

4-6 Relationship between the light intensity (lx) (measured at the soil level in tanks of the control, competition, herbivory, and herbivory-competition treatments) and time after the initiation of the treatments...... 117

4-7 Comparison of dry biomass (means ± SE) of Hydrilla verticillata (L.f.) Royle produced in treatments with and without Cricotopus lebetis Sublette...... 119

4-8 Comparison of turions (means ± SE) of Hydrilla verticillata (L.f.) Royle produced in treatments with and without Cricotopus lebetis Sublette...... 120

4-9 Comparison of stem length (means ± SE) of Hydrilla verticillata (L.f.) Royle grown in treatments with and without Cricotopus lebetis Sublette...... 121

4-10 Difference in percentage of damaged apical meristems (means ± SE) of Hydrilla verticillata (L.f.) Royle grown in treatments with and without Cricotopus lebetis Sublette...... 122

4-11 Comparison of surface cover (means ± SE) of Hydrilla verticillata (L.f.) Royle grown in treatments with and without Cricotopus lebetis Sublette...... 123

4-12 Difference in the biomass of American eelgrass, Vallisneria americana Michx., grown in mixed culture treatments with and without Cricotopus lebetis Sublette (herbivory)...... 124

4-13 Comparison of characteristics of Vallisneria americana Michx. interplanted with Hydrilla verticillata (L.f.) Royle in mixed culture treatments with Cricotopus lebetis Sublette (herbivory) and without C. lebetis (no herbivory).. . 125

4-14 Comparison of biomass of Hydrilla verticillata that expanded into pots originally planted with Vallisneria americana Michx. in treatments with and without Cricotopus lebetis Sublette...... 126

LIST OF ABBREVIATIONS

FAMU Florida A&M University

IFAS Institute of Food and Agricultural Sciences

UF University of Florida

U.S. United States

USDA United States Department of Agriculture

USEPA United States Environmental Protection Agency

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

FIELD HOST RANGE, FORAGING DEPTH, AND IMPACT OF CRICOTOPUS LEBETIS SUBLETTE (DIPTERA: CHIRONOMIDAE), A BIOLOGICAL CONTROL AGENT OF HYDRILLA VERTICILLATA (L.F.) ROYLE (HYDROCHARITACEAE)

Eutychus Mukure Kariuki

December 2017 Chair: Raymond L. Hix Cochair: James P. Cuda Major: Entomology and Nematology

Hydrilla [Hydrilla verticillata (L.f.) Royle] is among the most destructive invasive aquatic plants in the United States. Management of the plant is complicated by numerous challenges that include herbicide resistance, non-target effects and high costs of mechanical harvesters, and ineffectiveness of previously released insect biological control agents. Search for new biological control agents has been proposed by scientists to complement the existing management tools. A potential candidate is a stem-mining midge (Cricotopus lebetis Sublette) discovered in 1992 attacking the apical meristems of hydrilla in Kings Bay, Citrus Co., Florida.

An increasing body of evidence demonstrates C. lebetis can suppress hydrilla infestations and may have value as an augmentative biological control agent for hydrilla.

However, to better predict the efficacy of C. lebetis as a hydrilla biological control agent, more information is needed. Thus, this study investigated the field host range of C. lebetis; the effect of C. lebetis herbivory on competition between hydrilla and a native plants species, Vallisneria americana Michx.; and the effect of water depth on the establishment and survival of C. lebetis.

Field studies revealed larvae of C. lebetis infested only hydrilla stems and not the stem or leaf tissue of native plants, indicating hydrilla was its favored field host.

Mesocosm experiments demonstrated that damage by larvae of C. lebetis significantly reduced hydrilla biomass and turion production, and in the process, provided competitive advantage to the native plant, V. americana. Greenhouse and field studies revealed the larvae of C. lebetis can locate, attack, and complete development in submersed hydrilla at depths ranging from 0 m to at least 2.7 m, providing evidence that hydrilla in Florida’s shallow waterbodies grows within the vertical foraging range of the insect.

Overall, the results in this dissertation provide definitive evidence for the augmentative biological control potential of C. lebetis and provides new information that can be used to better evaluate the suitability of C. lebetis in hydrilla management programs.

CHAPTER 1 LITERATURE REVIEW

Hydrilla

Origin and Invasion History of Hydrilla

Hydrilla [Hydrilla verticillata (L.f.) Royle (Hydrocharitaceae)] is a submersed aquatic plant endemic to Asia, the Pacific islands of Fiji and Guam, Australia, portions of

Europe, in and around central African lakes, and the Indian Ocean islands of Mauritius and Reunion (Cook and Lüönd 1982, Balciunas et al. 2002). Its exact center of origin is uncertain, but most botanists speculate it is somewhere in the warmer areas of Asia, ranging from Iran, Afghanistan, Pakistan, and India, to Southeast Asia (Cook and Lüönd

1982).

In the New World, hydrilla was first reported in the U.S. near Tampa, Florida in the 1950s (Schmitz et al. 1991). New introductions of hydrilla have since been reported in many other regions of the world. For example, hydrilla was first reported in New

Zealand in1963 (Hofstra et al. 2000), Jamaica in 1977 (Cook and Lüönd 1982), Brazil in

2005 (Sousa et al. 2009), and South Africa in 2006 (Madeira et al. 2007). The plant now occurs on all continents, except Antarctica (Balciunas et al. 2002, Sousa et al. 2009).

Most studies suggest the aquarium plant trade is the main introduction pathway of hydrilla between countries (Schmitz et al. 1991, Madeira et al. 2007).

In the U.S., the hydrilla population consists of monoecious and female dioecious forms (Langeland 1996, Madeira et al. 2004). The monoecious form has both staminate

[male] and pistillate [female] flowers on the same plant, whereas the dioecious form has staminate and pistillate flowers on separate plants (Madeira et al. 2000). The female dioecious hydrilla was the form first introduced in the U.S. near Tampa, Florida (Schmitz

et al. 1991). About two decades later, the monoecious form was introduced into the northeastern U.S. The monoecious form was first reported in 1976 in Delaware and in

1982 in the Potomac River Basin (Steward et al. 1984). Madeira et al. (1999) concluded that the U.S. dioecious hydrilla originated from Bangalore, India, and the monoecious form from Seoul, South Korea, a conclusion based on analysis of random amplified polymorphic DNA (RAPD) of hydrilla samples from around the world. The two forms of hydrilla in the U.S. are dispersed between water bodies mainly via stem fragments entangled on recreational and fishing boats (Balciunas et al. 2002).

Distribution of Hydrilla in the United States

Dioecious hydrilla has since spread to fresh water bodies throughout Florida and many other states in the U.S. (Figure 1-1). As of 2016, hydrilla had infested over

162,000 ha of Florida public lakes and waterways (Schardt 2010, FWC 2016), a sevenfold increase from the 23,081 ha reported in 1990 (Schardt and Schmitz 1991), and a 30-fold spike from the 5,314 ha in 1982 (Schardt and Nall 1983). In Florida, dioecious hydrilla populations increase from a low density in the spring to high density in the fall (November) (Bowes et al. 1979).

Similarly, monoecious hydrilla has spread from its initial introduction site in

Delaware to other states such as Connecticut, Massachusetts, Maine, New York, and as far south as Georgia and Alabama. In the western U.S., monoecious hydrilla has been documented in Washington and California (Figure 1-1).

The distribution of monoecious and dioecious hydrilla is reported to overlap in

Lake Gaston, at the Virginia-North Carolina border (Ryan et al. 1995), in some water bodies of Georgia, South Carolina, and as far west as California (Madeira et al. 2000).

Description of Hydrilla

Hydrilla was described in detail by Cook and Lüönd (1982), Yeo et al. (1984), and Langeland (1996). The plant grows submersed in water primarily as a single stem in depths as much as 15 m (Yeo et al. 1984, Langeland 1996). Under these conditions, branching is usually sparse but becomes profuse as it grows nears the water surface, forming a dense canopy (Langeland 1996). Occasionally, hydrilla stems break loose and survive in a free-floating state (Langeland 1996, Owens et al. 2001).

Hydrilla has three types of stems: erect, horizontal, and subterranean. The horizontal stems originate from the basal nodes of a new plant and grow along a horizontal plane. The erect stems form on the nodes of the horizontal stems and grow vertically towards the water surface. The erect stems bear the roots, branches, leaves and reproductive structures—turions and flowers. The subterranean stems form on the basal nodes of erect stems, during winter, and grow downwards into the hydrosoil, where they produce tubers (subterranean turions) at their terminal nodes (Yeo et al.

1984).

The roots develop at the base of the erect stems (Yeo et al. 1984). Yeo et al.

(1984) reported that the roots are long, simple, and adventitious. Langeland (1996) described the color of the roots as glossy white, but turn reddish brown when growing in highly organic sediments or green when exposed to light.

Hydrilla leaves occur in whorls of three to eight, and each range from 2 to 4 mm wide and 6 to 22 mm long. The leaf margin is strongly serrated and the midrib, which often is red in color, often has visible tooth-like projections (Cook and Lüönd 1982,

Langeland 1996).

Hydrilla flowers are attached at leaf axils and clustered around the stems tips.

The floral parts (sepals, petals, and pistil or stamen) occur in a set of three each.

According to Madeira et al. (2004), the monoecious and dioecious hydrilla appear similar and can only be definitively distinguished based on the floral morphology or molecular characteristics. Female flowers have whitish sepals and translucent petals, and they usually float on the water surface while still attached to the leaf axil (Langeland

1996). In contrast, the male flowers have whitish red or brown sepals and whitish or reddish petals, and they usually detach from hydrilla stems just before maturity and float free on the water surface (Langeland 1996).

Hydrilla produces two specialized vegetative propagules: turions in leaf axils and tubers in terminal nodes of the subterranean stems (Yeo at al. 1984). Turions are 5 to 8 mm long, spiny, and usually dark green. In contrast, tubers (subterranean turions) are 5 to 15 mm in length, not spiny, and usually off-white to yellow but turn green when exposed to ambient light (Bowes et al. 1977, Yeo et al. 1984, Langeland 1996). The main morphological characteristics of hydrilla are illustrated in Figure 1-2.

Adaptations of Hydrilla

Hydrilla has been referred to as ‘the perfect aquatic weed’ because of its advanced morphological, physiological, and reproductive adaptations that allow it to invade and dominate exotic aquatic habitats (Langeland 1996, Dayan and Netherland

2005). Hydrilla reproduces prolifically via clones, grows rapidly, and efficiently assimilates carbon dioxide (CO2) and light, the two major resources that limit the photosynthetic capacity of submersed plants (Bowes and Salvucci 1989, Madsen and

Sand-Jensen 1991, Langeland 1996, Santamaria 2002).

Hydrilla grows rapidly and forms dense surface mats, a growth habit that allows the plant to outcompete native plants for light (Van et al. 1976). The plant can grow and elongate at a rate of up to 2.5 to10 cm per day (Glomski and Netherland 2012), until it nears the water surface, where it branches profusely, forming a dense mat (Langeland

1996). Because hydrilla biomass contains a high amount of water (approximately 90%), the plant can form the dense surface mats from a limited supply of nutrients and in a short period (Van et al. 1976). The dense surface mats allow hydrilla to intercept sunlight more efficiently and block light availability for other submersed species, limiting their photosynthetic capacity (Langeland 1996).

Hydrilla can photosynthesize unimpeded under relatively low concentrations of light and CO2 (Bowes and Salvucci 1989, Madsen and Sand-Jensen 1991). Studies have demonstrated that hydrilla has lower light compensation and saturation points than most native plants and can photosynthesize at levels below 1% of full light (Van et al.

1976, Bowes et al. 1977). The low light requirement allows hydrilla to initiate photosynthesis earlier in the morning than other plant species, gaining a competitive advantage in accessing a scarce supply of free CO2 (Van et al. 1976, Madsen and

Sand-Jensen 1991). The low light requirement also allows hydrilla to colonize deeper water levels than most submersed plant species (Van et al. 1976, Madsen and Sand-

Jensen 1991).

Hydrilla has evolved two main carbon-concentrating mechanisms that allow the plant to photosynthesize under CO2-limiting conditions (Salvucci and Bowes 1983, Rao et al. 2002). First, hydrilla can assimilate photosynthetic carbon from bicarbonate ions

(HCO3-), a source exploited by only about 50% of submersed plants (Salvucci and

Bowes 1983, Madsen and Sand-Jensen 1991). Second, hydrilla can switch its carbon

fixation from a C3 to C4 pathway, an adaptation that is rare among freshwater

macrophytes (Holaday and Bowes 1980, Reiskind et al. 1997, Rao et al. 2002).

Compared to the C3 pathway, the C4 pathway consumes more energy, but fixes carbon unimpeded in conditions that are inhibitory to the C3 pathway (e.g., low CO2

concentration, high dissolved oxygen concentration, high temperature, and alkaline

conditions) (Gowik and Westhoff 2011). These inhibitory conditions often occur within

hydrilla surface mats during daytime, especially in summer months (Van et al. 1976,

Bowes et al. 1979). Optimum levels of CO2 concentration triggers hydrilla to revert to C3

pathway.

Hydrilla has evolved the ability to reproduce and disperse through four main

methods: turions, tubers, stem fragments, and seeds (Haller et al. 1976, Owens et al.

2001), although hydrilla in the U.S. is not known to reproduce via seeds; it reproduces

asexually via tubers, turions, and stem fragments (Langeland 1996). Tubers can be

located to a depth of 20 cm in the hydrosoil, where they remain dormant for up to four

years (Haller et al. 1976, Van and Steward 1990, Netherland 1997, Doyle and Smart

2001). Similarly, mature turions abscise from the plant and drop to the hydrosoil (Yeo et

al. 1984), where they remain dormant for as long as one year (Van and Steward 1990).

Dormant turions and tubers can persist through adverse conditions—such as

drawdowns, drought, and herbicide treatments—and sprout when conditions are

favorable (Netherland 1997, Doyle and Smart 2001). The stem fragments allow hydrilla

to disperse because they are easily spread via factors such as boating and flowing

water (Doyle and Smart 2001, Owens et al. 2001, Baniszewski et al. 2016). Owens et al. (2001) reported that up to 83% of stem fragments can sprout into new plants.

Hydrilla also exhibits polyploidy, occurring in diploid (2n = 2×= 16), triploid (3n =

2× = 24), or tetraploid forms (2n = 2× = 32) (Cook and Lüönd 1982). Some botanists speculate that the multiple ploidy levels in hydrilla create genetically diverse clones adapted to a diverse environmental conditions and habitats (Arias et al. 2005, te Beest et al. 2011). Additionally, the broad genetic diversity of clones allows hydrilla to develop biotypes with adaptations facilitating invasion success (Arias et al. 2005, te Beest et al.

2011) or herbicide resistance. For example, studies by Michel et al. (2004) and Puri et al. (2007) reported that several hydrilla biotypes in Florida have developed herbicide resistance through selection of somatic mutations.

The extensive physiological and morphological adaptations exhibited by hydrilla allow it to not only colonize diverse habitats but also tolerate hydrological shifts within habitats. For example, hydrilla has been reported growing in waters with a salinity of up to 7% (Haller et al. 1974), lentic (still) to lotic (rapidly flowing) waters (Colon-Gaud et al.

2004), and acidic waters down to pH 4 (Van et al. 1976), but performs optimally at pH 7

(Steward 1991a). Studies in Florida have documented hydrilla flourishing in hypereutrophic lakes that were uninhabitable to native aquatic plants (Moxley and

Langford 1982) and in depths of up to 15 m, for example, in the Crystal River

(Langeland 1996). In addition, hydrilla flourishes in both temperate and tropical climates

(Cook and Lüönd 1982, Langeland 1996, Madeira et al. 2007). Ecologists speculate that the extensive adaptations of hydrilla coupled with the absence of its co-evolved herbivores and pathogens, as posited by natural enemy hypothesis, have allowed

hydrilla to aggressively colonize and dominate aquatic habitats in the U.S., making the

plant the most destructive aquatic weed in the U.S. (Langeland 1996, Dayan and

Netherland 2005).

Importance of Hydrilla

Ecological damage

Hydrilla invasions pose a major threat to biodiversity and ecosystem functioning

in U.S. freshwater ecosystems. Thus, hydrilla is listed both on the federal list of noxious

weeds (USDA-APHIS 2017) and on the FDACS (Florida Department of Agriculture and

Consumer Services) list of class I prohibited aquatic plants (FDACS 2017). Studies

have revealed that after hydrilla invades an exotic ecosystem, it spreads rapidly and

displaces ecologically important native plant species (van Dijk 1985, Chambers et al.

1993, Colon-Gaud et al. 2004). The invasion process reduces macrophytes diversity and associated phytophilous organisms, eventually, narrowing the foundation of native food webs (Kelly and Hawes 2005). Cases of hydrilla displacing native plants have been documented in Atchafalaya River basin in Louisiana (Colon-Gaud et al. 2004), Lake

Seminole in Georgia/Florida (Chambers et al. 1993), and the tidal Potomac River in

Virginia (Carter and Rybicki 1986).

The dense monotypic stands formed by hydrilla have been reported to imperil native fishes. Many studies have demonstrated that dense stands increase the habitat

spatial complexity. Thus, they impede foraging efficiency of fish (Colle and Shireman

1980, Theel and Dibble 2008); create hypoxic conditions especially in summer months

(dissolved oxygen < 2.0 mgL-1) (Van et al. 1976, Miranda and Hodges 2000, Burleson et

al. 2001); reduce diversity of fish prey (Schultz and Dibble 2012); and facilitate invasion

of exotic fish species (Nico and Muench 2004). Field studies have revealed that dense

hydrilla stands also reduce productivity of native sport fishes such as the popular bluegill (Lepomis macrochirus Rafinesque) (Colle and Shireman 1980, Theel and Dibble

2008) and largemouth bass (Micropterus salmoides Lacepede) (Valley and Bremigan

2002). In addition, hydrilla, together with other invasive plants, has been reported to increase the extinction rate of native fish species in the North America (Dextrase and

Mandrake 2006, Schultz and Dibble 2012).

Dense hydrilla stands have been reported to cause drastic diurnal changes in water quality, impacting native organisms (Van et al. 1976). During the day, especially in summer, photosynthetic activity by dense hydrilla stands has been shown to deplete

- CO2 and HCO3 and elevate O2 concentration to over 200% in the surrounding water.

These changes, coupled with high temperature, intensifies photorespiration losses in

C3-type native flora (Van et al. 1976, Bowes et al. 1979). During the night, respiration by

hydrilla depletes O2 and increases CO2 concentration in the surrounding water, creating hypoxic conditions that are detrimental to native aquatic fauna (Van et al. 1976, Bowes et al. 1979, Frodge et al. 1990, Colon-Gaud et al. 2004, Bradshaw et al. 2014).

Hydrilla surface mats have been reported to increase the invasibility of U.S. aquatic ecosystems, especially in the southeastern U.S. For example, studies in Florida water bodies have revealed that hydrilla surface mats provide a substrate for an invasive epiphytic cyanobacterium, Aetokthonos hydrillicola gen. et sp. Nov (Wilde et al.

2014), that causes avian vacuolar myelinopathy (AVM)—a newly discovered disease

that kills herbivorous waterfowl and their avian predators such as bald eagles

(Haliaeetus leucocephalus L.). The cyanobacterium secretes a neurotoxin that kills birds

when ingested (Bidigare et al. 2009). In addition, hydrilla beds have been found to serve

as preferred nesting sites for an invasive South American catfish [Hoplosternum littorale

(Hancock, 1828)], first reported in 1995 in Florida lakes (Nico and Muench 2004).

Ecologists describe these types of associations as ‘invasional meltdown’, where one invasive species facilitates another’s invasion—a process that accelerates damage to native ecosystems (Simberloff and von Holle 1999).

Economic damage

Dense stands of hydrilla also interfere with recreational and commercial water- based activities. Severe hydrilla infestations can ensnare swimmers, entangle angling gear, and foul boat propellers, as well as clog water intake systems for power generation and water supplies (Balciunas et al. 2002). These intrusions in water-use activities constrain regional economies (Henderson 1992, Newroth 1985). For example, in 1991, dense stands of hydrilla in Lake Moultrie, South Carolina, damaged a power generating plant and stopped its operation for seven weeks, costing an estimated $2.6 million in repairs and fish loss (Balciunas et al. 2002). Similarly, the fishing industry across Georgia and Florida loses millions of dollars annually because the mats reduce angler effort (Haller et al. 1980, Brown and Maceina 2002). The efforts of government agencies to manage hydrilla cost additional millions of dollars. For example, during the last 35 years, the state of Florida spent approximately $280 million to control hydrilla in the state’s public water bodies (Schardt 2010, FWC 2016).

Potential benefits

Amid the aforementioned negative consequences of hydrilla invasion in U.S. ecosystems, a few benefits have been reported in eutrophic ecosystems (Moxley and

Langford 1982, Taheruzzaman and Kushari 1989, Evans and Wilkie 2010). According to

Brönmark and Weisner (1992), eutrophic conditions support the irruption of

phytoplankton populations, leading to competitive displacement of ecologically

important native macrophytes. However, because hydrilla is highly adaptable, it flourishes in eutrophic waters and provides some of the same ecological functions of the displaced macrophytes (Moxley and Langford 1982, Evans et al. 2007, Rybicki and

Landwehr 2007). These ecological functions include compartmentalization of nutrients,

a process that limits eutrophication and increases water clarity (Moxley and Langford

1982, Langeland 1996). For example, a field study by Moxley and Langford (1982)

revealed that after hydrilla invaded two eutrophic lakes in Florida, it partly restored the

water quality and the production of forage, sportfish, and invertebrates in the two lakes.

Similarly, Rybicki and Landwehr (2007) reported that the expansion of a hydrilla

infestation in the then eutrophic Potomac Estuary between 1982 and 2001 restored

waterfowl populations. However, long-term studies suggest that the restorative benefits

by hydrilla are minor and the best way to restore native fauna and flora in eutrophic

lakes is by reducing anthropogenic nutrient inputs (Ruhl and Rybicki 2010).

A few researchers have proposed that dense hydrilla biomass in eutrophic water

bodies can be harvested and used as feedstock to produce biogas and compost

(Taheruzzaman and Kushari 1989, Evans and Wilkie 2010). Evans and Wilkie (2010)

estimated that, combined with benefits accrued from nutrient remediation, hydrilla

biomass can yield a net energy benefit ratio (ratio of energy returned on energy

invested) of 1.54 to 3.94 for biogas production, and 1.32 to 6.34 for compost production.

However, the study recommended that exploitation of hydrilla biomass might not be

economically feasible, unless novel harvesting technologies that will not disrupt fish and

other wildlife populations are developed (Taheruzzaman and Kushari 1989, Evans and

Wilkie 2010).

Summary

Taken together, most scientists agree that hydrilla is destructive to U.S. fresh water ecosystems. The plant not only causes ecological and economic losses that far outweigh its potential benefits, but also presents challenges in maintaining it below its economic threshold (Joyce et al. 1992). Most ecologists agree that hydrilla becomes destructive only after its surface canopy exceeds 20% (intermediate level) of a water body, and theoretically, a surface canopy of less than 20% may be ecologically beneficial (Joyce et al. 1992). However, current management tools in the U.S. have been ineffective in controlling hydrilla to levels below 20% surface cover, mainly because of the plant’s superior morphological, physiological, and reproductive adaptations (Moxley and Langford 1982, Dibble et al. 1996, Kirk and Henderson 2006).

Management of Hydrilla

Past and current management efforts of hydrilla in the U.S. have focused primarily on the use of herbicides (Haller et al. 1990, Fox et al. 1994), mechanical removal (Langeland 1996), and biological control agents (Sutton and Vandiver 1986,

Leslie et al. 1987).

Chemical control

Herbicide options currently available for management of aquatic weeds in the

U.S. are limited, with even fewer available for management of hydrilla (Getsinger et al.

2008, Netherland 2014). These herbicides are registered for aquatic use by U.S.

Environmental Protection Agency (USEPA) and are broadly classified based on their

mode of action into two groups, contact and systemic.

The contact herbicides that are currently used to manage hydrilla were registered more than five decades ago and include copper, diquat, and endothall (Table 1-1,

Getsinger et al. 2008, Netherland 2014). These contact herbicides are effective in controlling hydrilla infestations in small water bodies and in specific zones such as along shorelines (Langeland 1996, Netherland 2014). However, these contact herbicides are broad spectrum and have a short residual effect, requiring frequent applications

(Richardson 2008). Additionally, contact herbicides often pose irrigation restrictions and fail to adequately control perennial weeds such as hydrilla (Richardson 2008). Two additional contact herbicides, carfentrazone and flumioxazin, have been registered for aquatic use within the last two decades but are not widely used in the management of hydrilla (Netherland 2014). More research is required and use patterns need to be developed before the two newer contact herbicides are incorporated into the hydrilla management programs (Netherland 2014).

On the other hand, until early 2000s, fluridone was the only systemic herbicide approved by USEPA for controlling hydrilla in large water bodies (Michel et al. 2004).

Fluridone was registered in 1986 (Netherland 2014). The herbicide is enzyme-specific and selective, has a longer residual effect, and requires fewer applications than most contact herbicides (Netherland et al. 1993, Fox et al. 1994). In addition, fluridone kills target plants slowly, avoiding sudden mass death of plants and the resulting decomposition, which may create hypoxic conditions. Fluridone was widely used to manage hydrilla in Florida in the late 1980s and in the 1990s (Netherland and Jones

2015). It was reported to be effective at rates as low as 12 ug L-1 of lake water, lasting for at least 60 days (Netherland et al. 1993).

Newer systemic herbicides (penoxsulam, bispyribac-sodium, imazamox, and topramezone) were registered for aquatic use within the last decade (Netherland 2014).

The first three herbicides target the plant-specific enzyme acetolactate synthase (ALS), whereas topramezone targets a plant-specific enzyme 4-hydroxyphenyl-pyruvate dioxygenase (Netherland 2014). Studies have revealed that repeated use of herbicides containing the same mode of action may cause hydrilla to develop herbicide resistant biotypes (Michel et al. 2004, Puri et al. 2009).

Michel et al. (2004) reported the extensive and repeated use of fluridone in

Florida caused some hydrilla populations to develop resistance to this herbicide. The hydrilla resistance to fluridone, a phytoene desaturase inhibitor, was the first known natural case in higher plants (Michel et al. 2004). A few years later, Berger and

MacDonald (2011) discovered hydrilla biotypes in Florida that have developed resistance to a contact herbicide, endothall. Herbicide resistant biotypes persist for many years. For example, a field study in Kissimmee Chain of Lakes (KCOL) system in

Florida by Netherland and Jones (2015) revealed the fluridone-resistant biotypes remained dominant at the lakes despite the absence of fluridone selection pressure for eight years. Most botanists now believe the herbicide-resistant hydrilla biotypes will inevitably spread to other U.S. rivers and lakes (Getsinger et al. 2008, Puri et al. 2009,

Netherland and Jones 2015). There are additional concerns that the ability of hydrilla to develop herbicide-resistant biotypes may impede the effectiveness of future chemical control of hydrilla in the U.S. (Michel et al. 2004, Netherland and Jones 2015). Taken together, these developments have generated more interest in developing new

management tools, including biological control agents, to complement the existing chemical control options (Cuda et al. 2002, Netherland and Jones 2015).

Mechanical control

Mechanical control involves specialized harvesters that remove hydrilla from the water. This method is unpopular because it is non-selective, expensive, and difficult to implement, especially in large water bodies (Langeland 1996). Furthermore, hydrilla can disperse and reproduce from the fragments created by harvesters (Owens et al. 2001,

Hershner and Havens 2008). Thus, mechanical control is usually limited to critical areas that require immediate removal of hydrilla such as those around intake points for domestic water supplies and in high use areas such as boat ramps (Langeland 1996).

Biological control

The search for insect biological agents to control hydrilla in the U.S. began in

1968, with opportunistic surveys for natural enemies in India (Balciunas et al. 2002).

Comprehensive surveys began in 1971 when scientists from the Commonwealth

Institute of Biological Control (CIBC) were funded by USDA, in a 5-year project, to search for natural enemies of hydrilla in Pakistan (Balciunas et al. 2002). Worldwide searches for natural enemies of hydrilla were initiated in 1981 (Balciunas et al. 2002).

Many potential candidates were revealed by these exploration efforts. Two leaf-mining

Hydrellia flies (Ephydridae) and two Bagous weevil species (Curculionidae) were eventually approved and released in the U.S. (Buckingham 1994). In addition to insect agents, the potential of other organisms to manage hydrilla have been explored. An exotic herbivorous fish, in the family Cyprinidae, is currently being used to control hydrilla in the U.S. (Sutton and Vandiver 1986), and an adventive fungal pathogen is being investigated for its potential to control hydrilla (Cuda et al. 2008).

Insects:

The four approved insect agents released for classical biological control of hydrilla are the tuber-attacking weevil Bagous affinis Hustache (Coleoptera:

Curculionidae), stem-mining weevil B. hydrillae O’Brien (Coleoptera: Curculionidae),

Indian leaf-mining fly Hydrellia pakistanae Deonier (Diptera: Ephydridae), and Australian leaf-mining fly H. balciunasi Bock (Diptera: Ephydridae) (Table 1-2) (Buckingham 1994).

The tuber-attacking weevil, B. affinis, was the first host-specific insect imported from India and released into Florida in 1987 (Buckingham 1988). The adults feed on hydrilla stems and leaves exposed during dry seasons or drawdowns, whereas the larvae feed on the exposed tubers (Buckingham and Bennett 1998). The weevil, however, failed to establish because Florida lakes and rivers rarely undergo drawdown periods that expose the tubers (Godfrey et al. 1994).

The stem-mining weevil, B. hydrillae, is a semi-aquatic insect that was imported from Australia into Florida in 1991 (Balciunas and Purcell 1991). Its establishment in the

U.S. was not confirmed until 2009, when adult B. hydrillae were collected in southern

Louisiana, at least 580 km from the nearest release site (Center et al. 2013). Adults feed around the leaf nodes, whereas the larvae mine the stem and feed within the stem tissue (Balciunas and Purcell 1991). According to Balciunas and Purcell (1991), the larval feeding damage weaken the hydrilla stems and cause them to break-off. The stem fragments eventually float to the shoreline where the larva pupates (Balciunas and

Purcell 1991, Center et al. 2013). Balciunas et al. (2002) reported that, in its native range in Australia, sometimes the insect destroyed nearly all ‘topped out’ hydrilla surface mats. A recent study by Center et al. (2013) revealed the insect is established

and widely dispersed in the southeastern U.S. but reported there is no evidence the insect is having impact on hydrilla (Center et al. 2013).

The Indian leaf-mining fly, H. pakistanae, was imported from Pakistan and

introduced into Florida in 1987. The insect was later released in Louisiana, Alabama,

Georgia, Texas, and California (Center et al. 1997). Larvae mine and feed within the

leaf tissue, reducing the photosynthetic capacity of the plant. In laboratory tests, one

larva destroyed up to 12 leaves during its development, which takes between 18 to 30

days (Baloch and Sana-Ullah 1974). However, field studies revealed the established

populations of H. pakistanae failed to reach the effective damage threshold, and the

impact of the insect on hydrilla was negligible (Wheeler and Center 2001). The poor

establishment of the insect was linked to a range of stress factors including nitrogen-

deficient hydrilla, attacks by native parasitoids (Coon et al. 2014), and temperature

extremes in cold winters and hot summers (Buckingham 1994, Wheeler and Center

2001, Cuda et al. 2008).

The Australian leaf-mining fly, H. balciunasi, was imported from Australia and

released in Florida and Texas in 1989 (Grodowitz et al. 1997). The larvae destroy

hydrilla by mining and feeding on the internal leaf tissue (Baloch and Sana-Ullah 1973,

Buckingham et al. 1991). The insect failed to establish in Florida, but a small population

of this insect is established in eastern Texas (Grodowitz et al. 1997). The failure was

attributed to the same factors that affected H. pakistanae: poor host quality, parasitism

by native parasitoids, climate incompatibility, and inbreeding depression (Grodowitz et

al. 1997).

A major threat to the establishment of H. pakistanae and H. balciunasi is the parasitoid of native Hydrellia species, Trichopria columbiana Ashmead (Hymenoptera:

Diapriidae) (Coon et al. 2014). The parasitoid has been reported to parasitize populations of H. pakistanae and H. balciunasi in the U.S. and suppress the population growth and spread of these two biological control agents (Wheeler and Center 2001,

Doyle et al. 2002, Coon et al. 2014).

As aforementioned, the four released agents have failed to control hydrilla to levels below the economic threshold (Cuda et al. 2008). As a result, scientists are still exploring for new potential biological control agents of hydrilla (Cuda et al. 2008,

Copeland et al. 2011). Part of these exploration efforts revealed a potential candidate, a stem mining midge Cricotopus lebetis Sublette (Diptera: Chironomidae), which was discovered in 1992 attacking the apical meristem of hydrilla in Kings Bay, Citrus Co.,

Florida (Cuda et al. 2002). Cricotopus lebetis is described in detail below in the section titled ‘Cricotopus lebetis as Biological Control Agent of Hydrilla.’

Fish:

The grass carp [Ctenopharyngodon idella Val. (Cypriniformes: Cyprinidae)] is an herbivorous fish native to China and Russia (Sutton and Vandiver 1986). The fish was introduced into the U.S. in the 1960s for control of aquatic weeds in small water bodies

(Sutton 1977). Grass carp are voracious feeders, and field studies have demonstrated

24 individuals can destroy one hectare of hydrilla in about two years (Shireman and

Maceina 1981). However, the use of grass carp has generated concerns among aquatic ecosystem stakeholders because the fish is a generalist and can eliminate desirable macrophytes along with the invasive hydrilla (Sutton 1977, Sutton and Vandiver 1986,

Pipalova 2006). In addition, it is difficult to remove grass carp from water bodies once they are established; they are long lived (for about 10 years), tolerates diverse environmental conditions, and can displace native fish (Sutton and Vandiver 1986,

Pipalova 2006). Some scientists believe the grass carp is not an ideal biological control agent (Kirkagac and Demir 2004). The state of Florida requires a permit for use or possession of grass carp, and allows only the use of certified triploid (sterile) grass carp in aquatic weed management programs (Sutton and Vandiver 1986). As a precaution, use of grass carp is restricted to water bodies such as water reservoirs, ponds, and canals, with control structures that will prevent escape of the fish, and where the non- selective elimination of macrophytes is acceptable (Sutton and Vandiver 1986,

McKnight and Hepp 1995).

Pathogens:

Mycoleptodiscus terrestris (Gerd.) Ostazeski is a native fungal pathogen that attacks the lower epidermal layer of hydrilla leaves. A strain of M. terrestris was first discovered in 1983, attacking Myriophyllum spicatum L. (Haloragaceae) in

Massachusetts and Alabama (Gunner 1983). Subsequent greenhouse studies revealed infected hydrilla plants disintegrate and die within two weeks of infection (Shearer

1996), but attempts to inoculate wild hydrilla populations with the fungus failed (Shearer

1996). The failure of the fungus to establish in the field was attributed to extreme environmental conditions associated with dense hydrilla surface mats, including the drastic diurnal changes of water chemistry and the high water temperature during summer months, coupled with lack of an effective field application method (Bowes et al.

1979, Shearer 1996, Shearer et al. 2010). The fungus is still being evaluated for its

potential as an inundative biological control agent for the management of hydrilla

(Shearer et al. 2010, Cuda et al. 2016).

History, Benefits, and Approaches of Biological Control of Aquatic Weeds

Biological control of invasive weeds using insect agents began in terrestrial systems over a century ago, with many success stories reported worldwide (McFadyen

2000, Winston et al. 2014). In aquatic systems, insect agents were first used to manage invasive weeds in 1964 when the alligatorweed flea beetle [Agasicles hygrophila

Selman and Vogt (Coleoptera: Chrysomelidae)] was introduced into the U.S. to control alligatorweed [Alternanthera philoxeroides (Mart.) Griseb. (Amaranthaceae)]

(Buckingham 1996). The alligatorweed project became the world’s first success story of biological control of aquatic weeds. Many other insect agents have since been successfully used in tropical and sub-tropical regions to control aquatic weeds

(McFadyen 2000). For example, the South American weevil [Neohydronomus affinis

Hustache (Coleoptera: Curculionidae)] was used to effectively control water lettuce

[Pistia stratiotes L. (Araceae)] in Australia (Harley et al. 1990) and South Africa (Hill

2003). Another example is Neochetina bruchi (Hustache) (Coleoptera: Curculionidae), which was reported to have successfully controlled water hyacinth [Eichhornia crassipes

(Mart.) Solms (Pontederiaceae)] in many continents, including Africa, Asia, and North

America (Winston et al. 2014). Finally, the salvinia weevil [Cyrtobagous salviniae Calder and Sands (Coleoptera: Curculionidae)] was used to successfully control salvinia

[Salvinia molesta Mitchell (Salviniaceae)] in Australia, in various countries in Africa and

Asia, and in the U.S. (Winston et al. 2014). Biological control has become a major tool in the management of weeds in both terrestrial and aquatic environments in the U.S.

(Wapshere 1989, U.S. Congress Office of Technology Assessment 1993, Cuda et al.

2008, Winston et al. 2014).

The use of biological control agents to control invasive weeds has many benefits

over conventional methods, especially at the landscape level (McFadyen 1998, Culliney

2005). Biological control agents often provide a long-term control of the target weed

species, self-perpetuate (except fish), pose less risk to non-target plants, and are cost-

effective in the long run (Cullen et al. 2008). Modern biological control programs pose

limited risks to non-target organisms, especially in the U.S., due to stringent risk mitigation protocols that guide the importation and release of biological control agents

(Howarth 1991). Pemberton (2000) reported that out of 112 biological control agents of weeds that are established in the U.S. and Caribbean, only one organism has been documented attacking a native plant species unrelated to the target weed species.

When data from native plants that are closely related to the target weeds were considered, only 13.4% of established biological control organisms were found attacking the non-target native plants (Pemberton 2000). Ideally, biological agents must be host specific and damaging to target weed performance, which ultimately impacts their population dynamics (Howarth 1991).

Biological control programs involve three approaches: classical (importation), augmentation, and conservation (habitat manipulation) (Julien and White 1997,

McFadyen 1998, Goeden and Andres 1999, Cuda et al. 2008). Classical biological control is the predominant approach, and it involves the importation and release of exotic natural enemies to provide long-term control of invasive plants (McFadyen 1998).

Augmentation approach involves mass rearing of native or non-native weed biological

control agents, and periodically releasing them en masse on the target weeds species.

Conservation involves identifying and manipulating factors that enhance the populations and the actions of the existing natural enemies on target weed species (Cuda et al.

2008).

Cricotopus lebetis as a Biological Control Agent of Hydrilla

Background

Cricotopus lebetis belongs to the family Chironomidae, the most abundant and species-rich insect family in freshwater habitats worldwide (Oliver 1971, Pinder 1986).

Typically, the larval stages of chironomids are aquatic whereas adults are terrestrial

(Ferrington 2008). Chironomid larvae—most of which are detritivores and a few are either parasitic or phytophagous—constitute a large proportion of the insect fauna on many macrophytes (Oliver 1971, Pinder 1986, Cerba et al. 2010, 2011).

Known phytophagous species belong to the genera Cricotopus van der Wulp,

Glyptotendipes Kieffer, Polypedilum Kieffer, and Endochironomus Kieffer. Species in the genera Glyptotendipes, Polypedilum, and Endochironomus are not ideal biological control candidates because they are facultative phytophages; they feed on plants only when they are extending their tubes (Berg 1950, Oliver 1971). Some Glyptotendipes spp. and Endochironomus spp. are known to occasionally attack hydrilla in Florida,

Georgia, and Texas (Balciunas and Minno 1985). However, some Cricotopus species are obligate phytophages and are considered potential biological control agents of invasive macrophytes (Oliver 1971).

Specifically, two Cricotopus spp., C. myriophylli L. and C. lebetis, have been studied extensively because they attack two of the three most destructive aquatic weeds in the U.S., the invasive Eurasian watermilfoil (Myriophyllum spicatum L.) and

hydrilla, respectively (Oliver 1971, Kangasniemi and Oliver 1983, MacRae et al. 1990,

Cuda et al. 2002). The larval stages of the two species feed and develop within the meristematic tissue of their host plants. Field studies have shown the growth of the damaged plants is stunted and remains well below the water surface (Oliver 1971,

Kangasniemi and Oliver 1983, Cuda et al. 2002, 2011). A few other Cricotopus spp. are known to mine and feed on the leaf tissue of native Potamogeton spp. in Michigan and are not considered potential biological control agents (Berg 1950).

The first record of C. lebetis in the U.S. was from specimens collected in

Louisiana from 1957 to 1959 (Sublette 1964). The native range of the insect is unknown

(Epler et al. 2000). Recent field studies in Florida revealed that C. lebetis is present in many of the state’s lakes where hydrilla is present (Stratman et al. 2013a).

Description

The taxonomy and morphology of C. lebetis were re-described in detail by Epler et al. (2000). In general, an egg mass of C. lebetis is linear in shape and contains one or two rows of eggs arranged diagonally inside a gelatinous matrix (Cuda et al. 2002).

Newly laid eggs are white; but after 24 hours, the fertilized eggs turn grayish-brown, while the unfertilized eggs remain white and eventually disintegrate (Cuda et al. 2002).

Neonates are small and translucent. After molting into the second instar, the larvae turn greenish and acquire a diagnostic blue band around the second and third thoracic segments (Figure 1-3) (Epler et al. 2000). Adults are 3 to 4 mm in length. Forewings are clear with light brown veins, and halteres (modified, hind wing of a dipteran) are pale.

The first, fourth, and seventh abdominal tergites (dorsal sclerite of an insect body) are mostly stramineous (straw-colored); the rest of the tergites are mostly brown (Epler et al. 2000). Males and females can be easily distinguished using antennal and abdominal 38

features: males have long plumose antennae and a slender tapering abdomen, whereas

females have a short antenna and an abdomen as wide as the thorax (Epler et al.

2000). Because adults of Cricotopus spp. share most morphological features, C. lebetis

specimens can be misidentified as either C. tricinctus (Meigen), C. nitens (Kieffer), or C. taiwanus Tokunaga (Epler et al. 2000).

Distribution

Stratman et al. (2014) studied the thermal tolerance and distribution of C. lebetis in Florida. The field survey found C. lebetis was widely distributed in Florida waterbodies. Thermal tolerance tests revealed larvae of C. lebetis develop optimally at

temperatures ranging from 20 °C to 30 °C, with a lower development threshold of 15 °C

and a maximum of 32 °C (Stratman et al. 2014). Input of the thermal tolerance data into

a Plant Pest Forecasting System (NAPPFAST) model predicted that much of the southeastern U.S. is suitable for establishment of C. lebetis (Stratman et al. 2014).

Coincidentally, the predicted distribution range of C. lebetis in the U.S. matched

the current distribution of dioecious hydrilla in the U.S. (Stratman et al. 2014), suggesting C. lebetis can successfully establish and help suppress dioecious and monoecious hydrilla in southeastern U.S..

Biology and Feeding Damage

The biology of C. lebetis was described in detail by Cuda et al. (2002). Cuda et

al. (2002) also developed a method for rearing C. lebetis under laboratory conditions.

Baniszewski et al. (2016) reported the adventive hydrilla leafcutter moth Parapoynx

diminutalis Snellen (Lepidoptera: Crambidae) occasionally infests the hydrilla stem tips

used to rear C. lebetis and becomes a pest in laboratory colonies. The study

demonstrated the moth could be controlled using Bacillus thuringiensis (subspecies

kurstaki) (Btk), at a rate of 0.2 mL Btk per 3.8 L of well water. Mitchell et al. (2017) reported neonates starved for 2 days or more did not survive to adult stage and showed the survival rate of C. lebetis was enhanced by placing neonates on hydrilla immediately after hatching.

Under laboratory conditions, with a constant temperature of 25 °C, C. lebetis develops from egg to adult in 12 to 26 days (Cuda et al. 2002). Temperature andquality of the host plant influence the development time of the insect (Cuda et al. 2002).

Females oviposit on or below the water surface an egg string containing an average of

151 ± 8.4 eggs. Eggs hatch in approximately 2 days, and neonates emerge simultaneously.

Larvae of C. lebetis tunnel into the apical meristem and damage the meristematic tissue of the plant. They develop through four instars in about 14 days to reach the pupal stage (Cuda et al. 2002). The fourth instar, just before molting into a pupa, constructs a pupation site by making an exit hole from the tunnel and sealing it with cellulose fibers mined from the stem walls. The pupation site provides refuge for the fourth instar to molt into a pupa and complete its development.

Pupae complete development inside the pupation site, located within the hydrilla stem tip, in approximately 24 to 48 hours. Pharate adults (adult insects prior to emergence from a pupa) exit the pupation site through the cellulose-covered exit hole and actively swim to the water surface aided by an air bubble trapped in the pupal skin.

Adults are terrestrial, do not feed, and live for 1 to 2 days. Their sole function is reproduction. Shortly after emergence, males and females mate. The females then oviposit on or below the water surface, and both sexes die soon after (Cuda et al.

2002). The short development time for pupal and adult stages, and the lack of feeding in the adult stage, are common life history traits among chironomid species (Oliver 1971).

Cuda et al. (2002) studied the seasonal abundance of C. lebetis in Kings Bay watershed, Florida, in 1997 and 1998 and found that C. lebetis is a multivoltine species, prevalent throughout the summer and fall season. In addition, Cuda et al. (2002) reported the field abundance of C. lebetis peaked twice, in early summer (May and

June) and in fall (October and November)—a trend similar to that observed in the laboratory-reared population, which peaked after approximately every four generations.

The fall peak of C. lebetis population coincided with that of hydrilla (Bowes et al. 1979,

Cuda et al. 2002).

In the wild, C. lebetis has been noted as causing substantial damage to hydrilla

(Cuda et al. 2002, Cuda et al. 2011). A field study in Crystal River, Florida, revealed that

C. lebetis damaged up to 70% of the hydrilla apical meristems under field conditions

(Cuda et al. 2002). A replicated greenhouse study indicated that C. lebetis larval feeding can reduce hydrilla biomass by more than 99% (Cuda et al. 2011). The feeding damage prevented hydrilla from growing to the surface, suggesting that C. lebetis can reduce the invasiveness of hydrilla (Cuda et al. 2002, Cuda et al. 2011). These findings generated more interest in exploiting C. lebetis as a potential biological control agent of hydrilla

(Cuda et al. 2002).

Host Range

Available field records indicate that hydrilla is the only field host of C. lebetis.

However, these records were based on field surveys that focused more on hydrilla than on other submersed plants, and therefore, may be inconclusive (Stratman et al. 2013b).

Laboratory host-range tests by Stratman et al. (2013b) found that C. lebetis has a broad 41

fundamental (or laboratory) host range. The insect completed development on plants

from seven aquatic plant families: Hydrocharitaceae, Najadaceae, Potamogetonaceae,

Ceratophyllaceae, Characeae, Cyperaceae, and Alismataceae. However, evidence

from other biological control studies has shown that the fundamental host range of

insects often is significantly broader than the realized (or field) host range: an insect

species attacking multiple hosts under laboratory setting may be host specific, or nearly

so, under natural conditions (Balciunas et al.1996). Biotic factors such as host limitation,

competition, and parasitism by acquired natural enemies may limit the range of field

hosts attacked by an insect herbivore (Balciunas et al. 1996). Therefore, to better assess the suitability of C. lebetis as a biological control agent of hydrilla, its field host range needs to be investigated (Stratman et al. 2013b).

Need for Research

As mentioned earlier management of hydrilla in Florida, especially in large water bodies, poses significant challenges. The challenges include herbicide resistance

(Michel et al. 2004, Berger and MacDonald 2011, Netherland and Jones 2015), restricted use of grass carp (Sutton and Vandiver 1986), non-target effects of mechanical harvesters (Haller et al. 1980), and failure of established classical insect biological control agents to achieve the desired level of control (Cuda et al. 2008). Thus, hydrilla continues to cause significant ecological and economic damage to freshwater ecosystems in the U.S. (Bidigare et al. 2009, Pimentel et al. 2005).

To solve these problems, scientists have proposed searching for new biological control candidates to complement existing management tools (Wheeler and Center

2001, Michel et al. 2004, Overholt and Cuda 2005). One such tool for managing hydrilla could be the stem mining midge C. lebetis. However, more studies are needed to fill the 42

existing knowledge gaps on field host specificity of C. lebetis, its foraging depth, and its effect on hydrilla’s competitiveness (Cuda 2017). Thus, the main objectives of this study were to investigate the field host range of C. lebetis, larval foraging depth, and impact of the insect on competitiveness of hydrilla.

Figure 1-1. Map showing distribution of Hydrilla verticillata (L.f.) Royle in natural drainage systems within the United States as of 2016. Mapping indicates recorded presence of H. verticillata from at least one location within a designated drainage, but does not necessarily imply the frequency of occurrence in that drainage. Source: https://nas.er.usgs.gov/XIMAGESERVERX/2016/20161027161554.jpg. Accessed October 11, 2017.

Figure 1-2. Morphological characteristics of Hydrilla verticillata (L.f.) Royle. Source: IFAS, Center for Aquatic and Invasive Plants, University of Florida, Gainesville. Available at http://plants.ifas.ufl.edu/wp- content/uploads/images/hydver/hydverLD.jpg. Accessed June 12, 2017.

Table 1-1. List of herbicides approved by U.S. Environmental Protection Agency for management of Hydrilla verticillata (L.f.) Royle. The mention of a trade or generic chemical names does not constitute an endorsement or approval of the product by the author. Common name Mode of action Year registered Herbicide resistance reported (Vencill 2002, Bajsa et al. 2012, Netherland (Netherland 2014) 2014) Copper Unknown 1950s No Diquat Photosystem 1 disrupter 1962 No Endothall Serine/threonine protein phosphatases inhibitor 1960 Yes Fluridone Phytoene desaturase inhibitor 1986 Yes Carfentrazone- Protoporphyrinogen oxidase (PPO) inhibitor 2004 No ethyl Penoxsulam Acetolactate synthase (ALS) inhibitor 2007 No Imazamox ALS inhibitor 2008 No Flumioxazin PPO inhibitor 2010 No Bispyribac-sodium ALS inhibitor 2012 No Topramezone 4-hydroxyphenyl-pyruvate dioxygenase 2013 No inhibitor

Table 1-2. List of four biological control agents approved by the USDA for management of Hydrilla verticillata (L.f.) Royle. Order: Family species Origin Year States introduced Remarks Coleoptera: Bagous Australia 1991 Florida Failed to Curculionidae hydrillae establish O’Brien Bagous affinis India 1987 Florida Low Hustache population Diptera: Hydrellia Pakistan 1987 Florida, Louisiana, Ineffective Ephydridae pakistanae Alabama, Georgia, Deonier Texas, and California Hydrellia Australia 1989 Florida, Texas Low balciunasi population Bock

Figure 1-3. Cricotopus lebetis Sublette larva exhibiting a diagnostic blue band around the second and third thoracic segments. Photo courtesy of Jerry F. Butler, University of Florida.

CHAPTER 2 FIELD HOST SPECIFICITY OF A POTENTIAL HYDRILLA BIOLOGICAL CONTROL AGENT, CRICOTOPUS LEBETIS SUBLETTE (DIPTERA: CHIRONOMIDAE) IN LAKE ISTOKPOGA, FLORIDA

Introduction

Hydrilla stem mining midge, Cricotopus lebetis Sublette (Diptera: Chironomidae),

was discovered in 1992 attacking the apical meristems of hydrilla in Kings Bay, Citrus

Co., Florida (Cuda et al. 2002). Subsequent studies revealed that larvae of C. lebetis

mine hydrilla apical meristems, causing damage that stunts plant growth and prevents hydrilla from forming a canopy at the water surface, and ultimately reducing the economic and ecological damage associated with the hydrilla surface mats (Cuda et al.

2002, 2011). In addition, surveys of Florida water bodies revealed the insect is widely

distributed in Florida (Stratman et al. 2013a). These findings suggested C. lebetis has

the potential, if found to be host specific under field conditions, to be used as an

augmentative biological control agent for hydrilla (Cuda et al. 2002, 2011, Stratman et

al. 2013a).

However, a study by Stratman et al. (2013b) determined that C. lebetis has a

broad fundamental or physiological host range under laboratory conditions (Stratman et

al. 2013b). The larval stage of the insect completed development on multiple non-target

plant species belonging to diverse plant families, but had a higher larval survival rate in

plant species belonging to the family Hydrocharitaceae (Stratman et al. 2013b). Native

plants that were attacked by C. lebetis included Elodea canadensis Michx.

(Hydrocharitaceae), Vallisneria americana Michx. (Hydrocharitaceae), and Najas

guadalupensis (Spreng.) Magnus (Najadaceae) (Stratman et al. 2013b).

Although the lack of host specificity under laboratory conditions is undesired, it does not disqualify C. lebetis as a potential biological control agent for hydrilla. Many biological control practitioners recommend that plant species attacked under laboratory conditions be screened under field conditions to avoid false rejections of potentially effective candidate biological control agents (Cullen 1990, Balciunas et al. 1996, Smith et al. 2009). Previous studies have reported that the laboratory host range (fundamental host range) is often broader than field (realized host range) (e.g., Cullen 1990,

Balciunas et al. 1996, Smith et al. 2009).

This discrepancy between laboratory and field host ranges has been attributed to the artificial conditions in laboratories, which often restrict the natural host finding behavior of insects, forcing the insects to oviposit and even develop on hosts that they would not use under field conditions (Dunn 1978, Wapshere 1989, Clement and

Cristofaro 1995, Balciunas et al. 1996). In contrast, insects in natural conditions usually employ complex behavioral sequences to locate and select their hosts. Choice of a host by insect herbivores is further influenced by biotic factors such as host quality, competition, and parasitism by acquired parasitoids (Balciunas et al. 1996, Coon et al.

2014). Therefore, an insect species attacking multiple hosts under a laboratory setting may be host specific, or nearly so, under field conditions (Balciunas et al. 1996). For example, Balciunas et al. (1996) reported that a stem boring weevil, Bagous hydrillae

O’Brien (Coleoptera: Curculionidae), a biological control agent for hydrilla, had a laboratory host range of more than 16 feeding hosts and 11 oviposition hosts. However, an extensive field survey revealed the weevil had a narrower host range in the field, where the weevil was found on only seven other species, in addition to hydrilla. Based

on the field surveys, the weevil was considered safe and approved for release in

Florida.

Host specificity of C. lebetis under field conditions is currently unknown.

Therefore, this study was conducted to determine the host range of C. lebetis under

field conditions. This information will provide a better prediction of the potential effect and safety of augmentative releases of C. lebetis for control of hydrilla.

Materials and Methods

Study Area

To determine the field host range of C. lebetis, H. verticillata and four native macrophyte species were sampled from Lake Istokpoga (27.351679°N, 81.288061°W),

located in Highlands Co., Florida. Lake Istokpoga is a eutrophic floodplain lake and

among the largest lakes in Florida (Bachmann et al. 2000). Its surface area is

approximately 116 km2, with an average depth of 1.2 m (Bachmann et al. 2000). Lake

Istokpoga was chosen for this study because it had a reported presence of hydrilla

populations infested by C. lebetis (Stratman et al. 2013a).

Sampling

This study was conducted on arbitrary dates in 2015, 2016, and 2017. The

surveyed plants included the invasive weed H. verticillata and four native submersed

plants Najas guadalupensis (Spreng.) Magnus, Ceratophyllum demersum L., Vallisneria

americana Michx., and Potamogeton illinoensis Morong. These native plants were

selected for this study because they occurred naturally in Lake Istokpoga and have

been reported to serve as host plants for C. lebetis under laboratory conditions

(Stratman et al. 2013b).

Hydrilla and non-target submersed plant species were surveyed from a boat by

visual inspection and by using a rake method, following a technique modified from

Johnson and Newman (2011). After identifying a submersed plant bed, a rake with a 2

m long handle was lowered into the water, briefly dragged along the water column to

capture the submersed plants, and lifted back into the boat along a vertical plane, while twisting the handle 180° around its axis to avert the loss of the collected plants

(Johnson and Newman 2011). Plants were removed from the rake, separated by

species, and bagged in separate 16 L heavy-duty zipper seal plastic bags (61 cm length

× 46 cm width).

This sampling process was repeated within an area of approximately 10 m radius

until the sample bag of each plant species was approximately two-thirds full. On those

occasions where heterogeneous submersed vegetation beds were encountered, more

than one plant species was sampled from the same site. The bags holding the samples

were partially filled with water from the corresponding sample site to prevent mortality of

associated midges that may have been dislodged from their host plants. The bags were

stored in a portable ice chest, and transported to the Weed Biological Control

Laboratory at the University of Florida, Gainesville, Florida, for further processing within

24 hours of collection. Several additional liters of water were collected from the

sampling site so that plants presumably containing larvae of C. lebetis could be

maintained without killing the larvae due to osmotic imbalance. Three samples of each

available submersed plant species were collected from sampling sites that were at least

100 m apart. The coordinates of each sampling site were recorded using a global

positioning system (GPS) device.

Plant Sample Processing

To determine larval density of C. lebetis and percentage of apical meristems damaged by C. lebetis, 100 stem tips were selected randomly from each plant sample, and examined under a compound microscope for signs of feeding damage and for presence of larvae of C. lebetis (Figure 2-1). For V. americana samples, 100 whole leaves from each sample were inspected for signs of feeding damage by C. lebetis and for the presence of larvae (Figure 2-2). The number of larvae found on each sample of

100 stem tips and the number of stems exhibiting signs of C. lebetis feeding damage were recorded. Feeding damage caused by larvae of C. lebetis was determined based on a description reported by Cuda et al. (2002) and by comparison to photographs of plants damaged by C. lebetis feeding under laboratory conditions (Figures 2-1 and 2-2).

Larvae of C. lebetis were identified by the presence of the diagnostic blue band (Epler et al. 2000, Epler 2001).

Cricotopus lebetis Rearing

To collect emerging adults of C. lebetis, approximately 1kg wet weight of each plant sample was placed in an 11.4 L Sterilite® plastic tray (39.7 cm length × 31.4 cm width × 15.2 cm height; Sterilite, Townsend, MA) containing water collected from the corresponding sample site. The trays were placed in a Bugdorm-2120 ® emergence cage, (Megaview Science Co., Taichung, Taiwan) and aerated with an aquarium air pump, Aqua Culture ®. Emergence cages were monitored daily, for 14 days, for adult C. lebetis emergence. This rearing method was similar to that described by Copeland et al.

(2012a) and Stratman et al. (2013a). Emerged adults were collected daily by a mouth aspirator, sexed, and counted.

Data Analyses

Because the morphology of plants varied widely, for example, a stem tip of hydrilla and a whole leaf of V. americana, the plant samples were considered distinctly different. As a result, data on adult emergence and number of larvae are reported as mean ± SE (e.g., the results of field studies reported by Blossey et al. 1994, Smith et al.

2009, Copeland et al. 2011). The proportion of stem tips (or leaf tissue of V. americana) damaged were analyzed with a generalized linear models procedure using a binomial distribution family with a logit link. Pairwise comparison of the margins was done using a

Chi-square test with a Bonferroni correction for multiple comparison (v. 14; StataCorp.,

College Station, Texas, U.S.).

Results

Most submersed plants were found around an island in Lake Istokpoga known as

‘Bumble Bee Island’ (27.351679°N, 81.288061°W). Hydrilla was the dominant submersed plant species at the lake, and constituted approximately 90% of the standing vegetation within the submersed plant communities: C. demersum, P. illinoensis, and V. americana constituted the majority of the rest. Alam et al. (1996) similarly reported that hydrilla was abundant in the Lake Istokpoga and P. illinoensis and V. americana were less common.

Adults matching the morphological descriptions of C. lebetis were reared from the invasive plant H. verticillata and from the native plants N. guadalupensis, C. demersum, P. illinoensis, and V. americana (Figure 2-3). However, stem dissections confirmed the presence of C. lebetis larvae within the stem tissues of hydrilla but did not reveal any larvae of C. lebetis within the stem tissue of the native plant species (Figure

2-4). The number of larvae of C. lebetis per 100 stem tips of hydrilla ranged from 1.3 ±

1.3 to 8.0 ± 2.0 among the sampling dates (Figure 2-4). Larval infestation by C. lebetis

in hydrilla samples was detected on all sampling occasions (Figure 2-4).

Analysis of the stem and leaf tissue of the sampled plants revealed damage associated with C. lebetis within the apical meristems of H. verticillata and N. guadalupensis (Figure 2-5). Samples of H. verticillata had significantly higher damage than those of N. guadalupensis (Figure 2-5). No damage associated with C. lebetis were found on the stem or leaf tissue of the other native plants, C. demersum, V. americana, and C. demersum (Figure 2-5). The mean percentage of apical meristems exhibiting damage associated with C. lebetis among the sampling dates ranged from

8.8% ± 1.1% to 53.3% ± 0.9% for hydrilla, and 0.0% ± 0.0% to 2.0% ± 0.0% for N.

guadalupensis (Figure 2-5). Larval damage in hydrilla apical meristems associated with

C. lebetis was detected on all sampling occasions (Figure 2-5).

Discussion

Prior to this study, information about the host range of C. lebetis was based

solely on laboratory studies (Stratman et al. 2013b) and on anecdotal evidence from

field studies (Cuda et al. 2002). The present study provided the first empirical evidence

demonstrating hydrilla was the favored field host of C. lebetis. The study further

demonstrated the field host range of C. lebetis was narrower than its reported laboratory

range (Stratman et al. 2013b). The results agreed with previous studies that have

reported fundamental host range of insect herbivores is often wider than their field

(realized) host range (e.g., Cullen 1990, Balciunas et al. 1996, Smith et al. 2009). For

example, Balciunas et al. (1996) reported that Bagous hydrillae had a broad laboratory

host range but was approved for release in Florida after extensive field surveys

revealed B. hydrillae had a narrow field host range. Similarly, in another example in a 55

terrestrial system, Smith et al. (2009) reported that Aceria salsolae de Lillo and Sobhian

(Acari: Eriophyidae), a potential biological control agent of Salsola species in North

America, had a wide laboratory host range, but a field study revealed the mite had a narrow field host range and posed no significant risks to non-targets.

Preference of C. lebetis for hydrilla was evidenced by a combination of the consistent and higher levels of adult midges that were reared from hydrilla compared to native plants sampled, and the results of stem dissections that revealed a consistent presence of larvae of C. lebetis and significantly higher levels of associated feeding damage within the apical meristem of hydrilla. The results in the present study showing

9% to 53% percent of hydrilla apical meristems had damage associated with C. lebetis was consistent with the findings by Cuda et al. (2002). A field survey by Cuda et al.

(2002) revealed the percentage of hydrilla apical meristems damaged by C. lebetis in

Crystal River, Citrus Co., Florida, ranged from 15% to 57% in 1997 and 0% to 73% in

1998. Likewise, results showing the 1.3 to 8 larvae of C. lebetis per 100 stems of hydrilla in the present study was slightly higher than the average of 0.3 larvae per 100 stems of hydrilla collected in the wild reported by Stratman et al. (2013a), which was based on a survey of multiple lakes and rivers in Florida. Thus, the percentage of damaged apical meristems of hydrilla and the abundance of C. lebetis observed in the present study was consistent with the findings of previous studies and indicated hydrilla was an important host of C. lebetis under field conditions.

The observed absence of larvae of C. lebetis and the associated larval damage within the stem or leaf tissues of C. demersum, P. illinoensis, and V. americana indicated C. lebetis was not utilizing these non-targets as field hosts. These

observations paralleled, to some extent, findings by Stratman et al. (2013b), which demonstrated these non-targets were relatively poor hosts of C. lebetis. Stratman et al.

(2013b) reported the survival rate of C. lebetis in dioecious hydrilla was 57%, compared to 30% in P. illinoensis, 20% in C. demersum, and 7% in V. americana.

However, the results showing the relatively low percentage of damaged meristems of N. guadalupensis stems (0% to 2%) and the absence of larvae of C. lebetis within the stem tissue of N. guadalupensis stems contrasted with findings by

Stratman et al. (2013b). Stratman et al. (2013b) reported, under laboratory conditions,

N. guadalupensis was a more suitable host to C. lebetis than hydrilla and demonstrated the insect survived at a higher rate in N. guadalupensis (83%) than in dioecious hydrilla

(57%). The results of the present study indicated hydrilla (8% to 53% apical meristem damaged and 1.3 to 8 larvae of C. lebetis per 100 stems) was the most suitable host of

C. lebetis under field condition. Together, these findings indicated, under field conditions, C. lebetis predominantly utilized hydrilla as a host and was unlikely to substantially attack the populations of these non-target aquatic plants.

The discrepancy between the laboratory host and the field host ranges of C. lebetis could be explained by the reported egg mass properties, oviposition behavior of

C. lebetis, and by the growth habit of hydrilla (Yeo et al. 1984, Langeland 1996). Cuda et al. (2002) reported the egg masses of C. lebetis are sticky and are oviposited on the water surface, and Stratman et al. (2013b) showed females of C. lebetis chose to oviposit near potential host plants and hydrilla was among their most preferred oviposition hosts. Because hydrilla forms dense monotypic surface mats and often overshadows the canopies of existing native plants (Yeo et al. 1984, Langeland 1996),

and C. lebetis inherently prefers hydrilla as oviposition host (Stratman et al. 2013b), it is likely the females of C. lebetis searching for an oviposition site will encounter hydrilla

surface mats, and not the canopies of the native plants. The oviposited egg masses of

C. lebetis will likely to adhere to hydrilla surface mats, allowing the subsequent larval

stage to disproportionately colonize hydrilla compared to the native plants. This

assumption is further supported by the resource concentration hypothesis (Root 1973),

which posits that herbivorous insects are more likely to find and stay in more dense

stands than on small patches of their potential host plants (Root 1973, Long et al.

2003). In Lake Istokpoga, where this study was conducted, hydrilla was the dominant

macrophyte species.

In the present study, the chironomid rearing process associated the adults that

matched the morphological characteristics of C. lebetis with all the non-targets host

plants that were collected from Lake Istokpoga. However, stem dissections

demonstrated neither the larvae of C. lebetis nor the associated larval damage on the

stem and leaf tissue of the same non-target plant species, except for the less than 2%

stem tip damage on N. guadalupensis. The discrepancy can be explained by several

reasons. First, native macrophytes occurred in mixed stands dominated by hydrilla, and

it is likely the pupal stages of C. lebetis contaminated the non-targets during the

disturbances caused by the raking. Pupae of C. lebetis have been reported to exit from

stem tissue of hydrilla and swim to the surface where the adults eclose (emergence of

the adult insect at the terminal molt) (Cuda et al. 2002). Copeland et al. (2012a,b)

reported that immatures of chironomids could be transferred between plant species

during the collection of the plant materials. Colon-Gaud et al. (2004) and Bogut et al.

(2010) suggested the native plant C. demersum, which supported a consistently higher level of adult emergence in the present study, is a free-floating plant with firm and dissected branches, giving it a bottlebrush shape and making it efficient in trapping aquatic organisms such as immature chironomids. Second, taxonomists have reported a major limitation associated with the morphology-based identification to species of adult members of the Cricotopus sylvestris species-group (which includes C. lebetis), which are common and widespread throughout the eastern U.S. (Epler et al. 2000,

Epler 2001, Gresens et al. 2012). Members of this group are characterized by broad intraspecific variations that often obscure interspecific boundaries, and some phenotypes of another member of C. sylvestris species-group may be inadvertently misidentified as C. lebetis (Epler et al. 2000, Epler 2001, Gresens et al. 2012). To overcome this limitation, previous studies have recommended, morphological-based identification of adults of the members of C. sylvestris species-group should be confirmed by examining the associated larval or pupal stages of the insect (Epler 2001) or using DNA barcoding (Ekrem et al. 2010, Gresens et al. 2012). In the case of C. lebetis, the larvae are easy to identify based on setal tufts on abdominal segments I-VII and diagnostic blue bands on the second and third thoracic segments (Epler 2001). The discrepancy between the results of adult emergence and stem dissection demonstrated the importance of basing the identification of field-collected members of C. sylvestris species-group on life stages with definitive diagnostic features.

Overall, the narrow field host range of C. lebetis demonstrated by the results of the present study supported the recommendation by previous studies (e.g., Cuda et al.

2011, Stratman et al. 2013b) that C. lebetis could have value as an augmentative

biological control agent of hydrilla. During the course of the present study, the mats of hydrilla at the study site at Lake Istokpoga rarely topped out. Apical meristem damage caused by larvae of C. lebetis modifies the architecture of the hydrilla by stunting the

growth of the plant and preventing its canopy from reaching the water surface (Cuda et

al. 2011). The discrepancy between the laboratory host range and the field host range

revealed by the present study supported the necessity of conducting field host

specificity studies for biological control agents to avoid the risk of incorrectly rejecting a

potentially effective agent. It is worth noting that the present study was limited to the

range of submersed plant species that occurred at Lake Istokpoga and reported by

Stratman et al. (2013b) as being part of the fundamental host range of C. lebetis.

Additionally, choice of a host plant species by insect herbivores is influenced by biotic

factors such as host quality, competition, and parasitism by acquired parasitoids (Dunn

1978, Wapshere 1989, Clement and Cristofaro 1995, Balciunas et al. 1996). Therefore,

to further enhance our understanding of the field host range of C. lebetis, future

research should extend the investigation of host range of C. lebetis to other lakes with

different biogeographical properties. For example, nearby lakes without hydrilla can be

sampled to see if C. lebetis is present.

A B

Figure 2-1. Apical meristem of Hydrilla verticillata (L.f.) Royle as viewed under a dissecting microscope. A) Apical meristem of H. verticillata damaged by Cricotopus lebetis Sublette. B) Undamaged apical meristem of H. verticillata about to be attacked by larva of C. lebetis. Photographs by Eutychus Kariuki.

A B

Figure 2-2. Damage on native aquatic plants caused by larvae of Cricotopus lebetis Sublette. A) Damage on a leaf of Vallisneria americana Michx.. B) Damage on a stem tip of Najas guadalupensis (Spreng.) Magnus. The plants were damaged under laboratory conditions. Photographs by Eutychus Kariuki.

Hv Ng Cd 30 Pi Va

20 Adults per Sample

27-Jan-15 24-Jun-15 9-Dec-15 8-Sep-16 3-Nov-16 13-Jan-17

Adult emergence

Figure 2-3. Number (mean ± SE) of adults matching the description of Cricotopus lebetis Sublette reared from stem tips of plant samples collected in Lake Istokpoga, Florida, U.S. Hv, Hydrilla verticillata (L.f.) Royle; Ng, Najas guadalupensis (Spreng.) Magnus; Cd, Ceratophyllum demersum L.; Pi, Potamogeton illinoensis Morong; Va, Vallisneria americana Michx.

Hv Ng 10 Cd Pi Va 8

4 Larvaeper Stem100 Tips (n) 2

7-Apr-15 23-Apr-15 24-Jun-15 9-Dec-15 8-Sep-16 3-Nov-16 13-Jan-17

Sampling Date

Figure 2-4. Number (mean ± SE) of larvae of Cricotopus lebetis Sublette per 100 stem tips (or leaves for Vallisneria americana Michx.) recorded from plant material collected at Lake Istokpoga, Florida, U.S. Hv, Hydrilla verticillata (L.f.) Royle; Ng, Najas guadalupensis (Spreng.) Magnus; Cd, Ceratophyllum demersum L.; Pi, Potamogeton illinoensis Morong; Va, V. americana.

Hv 80 Ng *** Cd Pi Va 60 *** ***

*** DamagedStem Tips (%) 20 ***

* ***

7-Apr-15 23-Apr-15 24-Jun-15 9-Dec-15 8-Sep-16 3-Nov-16 13-Jan-17

Sampling Date

Figure 2-5. The percent (mean ± SE) damaged stem tips by Cricotopus lebetis Sublette from plant material collected at Lake Istokpoga, Florida. U.S. Hv, Hydrilla verticillata (L.f.) Royle; Ng, Najas guadalupensis (Spreng.) Magnus; Cd, Ceratophyllum demersum L.; Pi, Potamogeton illinoensis Morong; Va, Vallisneria americana Michx. Asterisk denote a statistically difference between stem samples of H. verticillata and N. guadalupensis (*P < 0.05; ***P < 0.001). No damage associated with C. lebetis were found on the samples of C. demersum, P. illinoensis, and V. americana.

CHAPTER 3 ASSESSMENT OF MAXIMUM WATER DEPTH CRICOTOPUS LEBETIS SUBLETTE LARVAE CAN LOCATE HYDRILLA VERTICILLATA (L.F.) ROYLE

Introduction

Host finding behavior of biological control agents for invasive weeds has attracted interest from biological control scientists (Wan and Harris 1996, Marko et al.

2005, Reeves et al. 2009, Stratman et al. 2013b, Catton et al. 2014). Understanding this behavior can improve the predictability for the efficacy of biological control agents

(Zwölfer and Harris 1971, Heard 2000, Reeves et al. 2009, Catton et al. 2014). Several studies have revealed that host-finding ability among phytophagous insects varies widely and can be conceptualized as a continuum with two extremes (Barton-Browne

1977, Fenemore 1988, Wan and Harris 1996, Marko et al. 2005, Reeves et al. 2009,

Stratman et al. 2013b, Catton et al. 2014). Using biological control agents as examples, on one end of the continuum are species that can detect their host plants from a distance using visual and/or olfactory sensory cues, such as the milfoil weevil

[Euhrychiopsis lecontei Dietz (Coleoptera: Curculionidae)], a potential control agent of

Eurasian water milfoil (Myriophyllum spicatum L.) (Reeves et al. 2009). At the other extreme, are species that detect their hosts based on gustatory (taste) or contact cues, such as the Canada thistle flea beetle [Altica carduorum Guer. (Coleoptera:

Chrysomelidae)] (Wan and Harris 1996). In addition, the life stage of the insect responsible for host finding may differ by species (Mayhew 1997, Cuda et al. 2002,

Stratman et al. 2013b). In many species, immatures have relatively limited mobility, cannot choose their development location, and depend on the female parental stage to choose an oviposition site near or on a suitable host plant (Mayhew 1997).

Host-finding in phytophagous insects follows a sequence of behaviors governed

by both internal (idiothetic) and external (allothetic) cues (Bernays and Chapman 1994).

In the absence of host-related cues, most phytophagous insects orient themselves and move based on idiothetic control (Visser 1988). Encounter of a host stimulus triggers the control of insect movement to transition from idiothetic to allothetic (Visser 1988), and the pattern of movement changes from straight (ranging) to convoluted (local

search) (Fenemore 1988, Visser 1988, and Reeves et al. 2009). Specific host-related stimuli eventually cue the insect to its host plant and to the specific part of the host plant that supports its feeding, oviposition, and/or pupation (Bernays and Chapman 1994).

Unlike terrestrial ecosystems, host-finding behavior of phytophagous insects in aquatic ecosystems has not been widely studied (for a review see Hay and Steinberg

1992). However, host-finding behaviors of aquatic insects that have been studied extensively include the hydrilla stem mining midge Cricotopus lebetis Sublette (Diptera:

Chironomidae), a potential biological control agent for the invasive weed hydrilla

[Hydrilla verticillata (L.f.) Royle (Hydrocharitaceae)] (Stratman et al. 2013b), and the milfoil weevil (Marko et al. 2005, Reeves et al. 2009, Reeves and Lorch 2011).

Research on these two potential biological control agents revealed that phytophagous insects in aquatic ecosystems can locate their host plants using olfaction and visual cues (Marko et al. 2005, Reeves et al. 2009, Stratman et al. 2013b). The milfoil weevil can locate its host (M. spicatum) by cuing in on the plant exudates, glycol and uracil

(Marko et al. 2005), and on the visual cues from as far away as 17.5 cm (Reeves et al.

2009, Reeves and Lorch 2011). In addition, milfoil weevils can visually discriminate

plant species underwater (Reeves et al. 2009).

Because field and laboratory studies have shown that C. lebetis has the potential

to attack and suppress the invasive weed hydrilla, researchers have been interested in

understanding this insect’s host finding behavior (Stratman et al. 2013b). To date,

several studies have provided some insights on how C. lebetis detects and locates its

host plant (Cuda et al. 2002, Lerner et al. 2008, Stratman et al. 2013b). The adults are

terrestrial, whereas the larvae and pupa are aquatic. Females choose where to oviposit

on the water surface (Cuda et al. 2002). Lerner et al. (2008) explained that chironomid

females, such as C. lebetis, identify water surfaces used for oviposition by using

polarized light. An oviposition study by Stratman et al. (2013b) discovered that C. lebetis

females do not oviposit on random sites; rather, they select sites near potential host

plants and, in the absence of potential host plants, near available substrates. Cuda et

al. (2002) reported that the egg masses of C. lebetis are sticky and denser than water,

and explained that as the egg masses sink they attach either to host plants or

substrates within the water, or continue to sink to the bottom of the water column. Taken

together, female oviposition behavior and the egg mass properties increase the

chances of neonates hatching near potential host plants. Larvae are nektonic; they

swim or drift in the water column until they encounter hydrilla (Cuda et al. 2002,

Stratman et al. 2013b). Chironomid larvae use three active modes of locomotion to

move within the water column—swimming, crawling, and whole-body respiratory

undulation, a sinusoidal wave action of the body bending in a head-to-tail direction

(Brackenbury 2000). Because they have negative buoyancy, they free fall along the

water column if they stop swimming (Brackenbury 2000, Brackenbury 2003). Behavioral

assays by Stratman et al. (2013b) revealed that C. lebetis neonates locate hydrilla

randomly, but as the larvae matured, they acquired the ability to visually locate their host plant, hydrilla.

A gap in knowledge exists in how deep neonates of C. lebetis can swim or fall in the water column to locate and attack hydrilla. Hydrilla can grow to a depth of 15 m

(Langeland 1996). Approximately half of the hydrilla standing crop is composed of profusely branched stems that form a dense mat of vegetation, occurring within the upper 0.5 m of the water column (Haller and Sutton 1975). Therefore, information on the vertical foraging depth of C. lebetis can be used to better predict the efficacy of C. lebetis in controlling hydrilla populations that occur at various water depths. Studies in other systems have revealed that some chironomid species occur at relatively great depths (Linevich 1971). For example, a study in the world's deepest fresh water lake,

Lake Baikal in Siberia, Russia, found a chironomid species, Sergentia koschowi

Linevich (Diptera: Chironomidae), living at a record depth of 1360 m (Linevich 1971).

The objective of this study was to determine the foraging depth of C. lebetis in the water column and the extent to which water depth limited the establishment and survival of C. lebetis. Experiments were conducted in controlled greenhouse studies on hydrilla placed at known depths and in a south Florida lake under natural conditions.

Materials and Methods

Greenhouse Experiments

To test how deep larvae of C. lebetis swim to locate hydrilla and if water depth was a limiting factor in the establishment/survival of C. lebetis, two greenhouse experiments were conducted between October 2016 and April 2017 at UF/IFAS

Entomology & Nematology Department facilities (29.633882°N, 82.366733°W), located in Gainesville, Alachua Co., Florida.

Hydrilla stem tips used in this study were collected from the UF/IFAS Center for

Aquatic and Invasive Plants’ (CAIP) facilities (29.726796°N, 82.415235°W), located in

Gainesville, Florida. Approximately 15 cm long apical portions of healthy hydrilla stems were obtained and rinsed thoroughly with well water to remove any unwanted organisms from the stems

Eggs and neonates of C. lebetis used in this study were obtained from a laboratory colony, which was founded from insects collected from Lake Istokpoga

(27.351679°N, 81.288061°W), Highlands Co., Florida, and reared following procedures described by Cuda et al. (2002). The colony of C. lebetis was housed in laboratory facilities of the UF/IFAS Entomology & Nematology Department located in Gainesville,

Florida.

Experiments were performed in extruded acrylic tubes, also known as Plexiglas tubes (15.24 cm diameter; estreetplastics.com, Royse City, Texas), which were obtained in four different lengths: 0.3 m, 0.9 m, 1.8 m, and 2.7 m (Table 3-1). Plexiglas containers have been used to rear hydrilla and C. lebetis in previous studies and are not known to be toxic to developing larvae (Cuda et al. 2011). Before starting new experiments, the tubes were thoroughly cleaned with well water and air-dried. Clean tubes were then secured upright, using 0.9 m long bungee cords, along the edges of metal mesh greenhouse benches and filled with well water, leaving an approximately 5 cm column of air space at the top.

Bouquets of hydrilla were formed by pairing two bundles of clean hydrilla stems, each bundle containing 60 shoot tips. Unpublished data revealed that making the bouquet with the two bundles of 60 shoot tips, instead of one bundle of 120 shoot tips,

ensured that bouquets remained healthy during the experiment. The bundles were paired by tying their bases on the opposite ends of a ribbon (40 cm length × 2 cm width)

made from a black polyethylene shade cloth (Sunblocker® Premium, Farmtek.,

Dyersville, IA). The bases of the two bundles were spaced 2 cm apart whereas the tops

remained untied and spread out, forming a cone shaped bouquet. A twisted nylon twine

was tied to the base of the complete 120 stem tip bouquet and used to lower it to the

bottom of each of the extruded acrylic tubes (treatment units), except for the 0 m

treatment. In the 0 m treatments, the stem tips were positioned just below the water

surface in Plexiglas tubes measuring 0.3 m in height. In the rest of the treatments, 0.9

m, 1.8 m, and 2.7 m, the bouquets of hydrilla were placed at the bottom of the 0.9 m,

1.8 m, and 2.7 m tall tubes, respectively. An additional treatment of 0 m, replicated four

times, was set up in an environmental growth chamber maintained at 25 °C and a 14:10

(L:D) photoperiod to provide basis for detecting any potential anomalies in C. lebetis

performance that may be caused by the greenhouse conditions. To confirm that the

clean stem tips used in the experiment were free of any insect contamination, an

additional set of two 0.9 m tubes received bouquets of hydrilla, but did not receive any

insect treatment.

In the first of two experiments, each tube was inoculated with 185 ± 5 C. lebetis

neonates. In the second experiment, each tube was inoculated with 185 ± 5 C. lebetis

eggs. A laboratory study by Cuda et al. (2002) revealed that C. lebetis females

oviposited on the water surface, and the gelatinous egg masses would either attach to

substrates on the water surface or sink to the bottom of the water column. Therefore,

under natural conditions, larvae may eclose on the water surface, if an egg mass is

attached to a substrate on the water surface, or at the bottom of a waterbody, if an egg mass sinks to the bottom. The site of larval eclosion may influence the distance the larvae must swim to locate submersed hydrilla. Thus, in this study, the tubes were inoculated with neonates in the first experiment and egg stage in the second experiment. After inoculation, tubes were covered with a fine mesh cloth that allowed airflow, but confined emerging adults within the tubes. Natural light in the greenhouse was supplemented by incandescent bulbs set at 14:10 L:D photoperiod. Each greenhouse experiment was replicated six times: twice in space and thrice in time.

Experiment I: Inoculation of neonate Cricotopus lebetis on the water surface

The main objectives of this experiment were to determine the ability of neonates of C. lebetis to swim from the water surface and locate host plants, mine into hydrilla, and impact plant growth at various water depths. Neonates were inoculated into each of the 16 treatment units described above (Plexiglas tubes containing hydrilla stems at specific depths: 0 m, 0.9m, 1.8 m, and 2.7 m). Because neonates of C. lebetis are active swimmers and difficult to count under a microscope, counting was done at the egg stage. Several egg masses of C. lebetis, 24 hours old, were randomly selected from the laboratory colony and analyzed for fertility and egg count under a dissecting microscope (8×; Zeiss Stemi DV4; Carl Zeiss, Berlin, Germany). Egg masses of C. lebetis contain 50 to 250 eggs, and fertilized eggs turn grayish brown 24 hours after oviposition, whereas the unfertilized eggs remain white (Cuda et al. 2002). Two or three egg masses that contained a total of 185 ± 5 fertilized eggs were selected and placed in

35 mL Fisherbrand® culture tubes (Fisher Scientific, Hampton, NH) that contained 25 mL of well water. Egg masses were stored in an insect rearing room, set at 24 °C and

14:10 L:D photoperiod, until larval eclosion. Once the eggs hatched, the culture tubes 72

containing the neonates were immediately transferred to the greenhouse where the

neonates were gently emptied onto the water surface of the Plexiglas tubes. The

Plexiglas tube tops were then covered with the fine mesh cloth, secured in place using

Dixon® worm gear clamps (Dixon Valve & Coupling Co., Chestertown, Maryland). Each of the 16 Plexiglas tubes was considered a treatment unit.

Experiment II: Inoculation of Cricotopus lebetis egg masses on the water surface

The main objective of the second experiment was to determine the effect on egg

masses of C. lebetis that sank to the bottom of water bodies; whether the resultant

larval stage would locate, develop in, and damage hydrilla stem shoots staged at the

aforementioned depth treatments. The experiment was conducted using the same

procedure followed in the first experiment except in this case, the Plexiglas tubes

holding the four treatments were each inoculated with fertilized eggs of C. lebetis (n =

185 ± 5). The eggs were examined and counted under a dissecting microscope, placed

in 35 mL culture tubes, and transferred to the greenhouse, where they were deposited

on the water surface of the treatment units with a medicine dropper. To contain the

introduced C. lebetis in the experimental tubes, tube tops were covered with a fine

mesh cloth and secured with a Dixon® worm gear clamp.

Data collection

In both experiments, the Plexiglas tubes were monitored daily for adult

emergence from day 12 until the end of adult emergence. Adult C. lebetis has been

reported to eclose beginning on the 13th day after oviposition (Cuda et al. 2013).

Emerged adults from each tube were collected daily using a mouth aspirator, sexed,

and counted. Adult males and females isolated by treatment tube were placed in

oviposition chambers that consisted of 35 mL glass culture tubes containing 25 to 30

mL of water, and capped with a perforated plastic cap (Cuda et al. 2002). Oviposition chambers were checked after 24 hours to recover oviposited egg masses. Fertile and infertile eggs in the egg masses were counted and recorded. After the last adult emerged in a depth tube, bouquets of hydrilla were recovered from the experimental

tube. In the laboratory, bouquets were untied and the stem tips examined under a

dissecting microscope for signs of feeding damage caused by C. lebetis. Stems

exhibiting feeding damage were counted and recorded. Length of each stem was

measured using a meter stick to an accuracy of 0.01 m (Johnson Level & Tool Co.,

Mequon, Wisconsin).

Data analyses

Data were analyzed with a generalized linear models procedure, SAS PROC

GLIMMIX, using a distribution function appropriate for the response variable in question,

i.e., binomial for proportions, Poisson for count data, and normal for plant height (V. 9.2;

SAS Institute Inc., Cary, NC, U.S.). Over dispersion was deemed not to be an issue for

binomial and count data because the χ2/df ratio was less than 1.10 (Warton and Hui

2011). Estimated means and 95% confidence intervals were back transformed using the

ilink option, and P-values for multiple comparisons were adjusted using simulation

option in the above-named procedure.

Field Experiments

Two field experiments were conducted between January and March in 2017 at

Lake Istokpoga. The objective of the two field experiments was to determine the vertical

distribution of C. lebetis in the water column under natural conditions. The study site

was approximately 0.9 m deep, and was chosen because it had established populations

of both hydrilla and C. lebetis (Stratman et al. 2013a).

Experiment I: Use of sentinel stems to determine the vertical foraging depth of Cricotopus lebetis

Sentinel bouquets consisting of 120 hydrilla stem tips, each free of C. lebetis and approximately 15 cm in length, were placed at three depths; 1) 0 m, just below the water surface; 2) 0.45 m, mid-depth; and 3) 0.9 m, at the sediment level. The sentinel bouquet of hydrilla (hereafter referred to as bouquet) was made from healthy, thoroughly washed

hydrilla stem tips previously collected from Lake Istokpoga. A replicate consisted of a

heavy-duty, plastic-coated steel garden stake (184 cm in height by 1.5 cm diameter;

Gardener's Blue Ribbon®, Lititz, PA), hammered 0.3 m into the lake sediment. Three

hydrilla bouquets were secured to the garden stake at the three different depths using

self-locking nylon cable ties (2 mm × 100 mm). During each setup of the experiment,

100 stem tips were randomly collected from hydrilla growing around the replicate site

and, later in the laboratory, examined for the presence of larval C. lebetis and for stem

tip damage. Two weeks after initiating the experiment, the sentinel bouquets were

retrieved and transported to the laboratory for further processing. In the laboratory,

bouquets were untied to free the stem tips, which were examined under a dissecting

microscope for presence of C. lebetis larvae and for signs of feeding damage. The

experiment was replicated five times in space and three times in time. The replicates

were spaced at least 100 m apart and repeated three times at two-week intervals.

Experiment II: Survey of naturally growing stems to determine the vertical foraging range of Cricotopus lebetis

The second field experiment was conducted from January to February 2017 at

the same study site as the first experiment. The objective was to determine the vertical

distribution of naturally occurring C. lebetis within the water column in Lake Istokpoga.

One hundred stem tips were sampled from hydrilla plants growing within the following

depths: 1) just at or below the water surface, 0 to 0.3 m; 2) mid-depth 0.3 to 0.6 m; and

3) bottom, 0.6 to 0.9 m. The sampling was done by visual inspection to locate plants at the appropriate depth using a boat and by using the rake method, following a technique modified from Johnson and Newman (2011). On each sampling day, 100 stems of hydrilla were sampled from each depth section (surface, mid-depth, and bottom), from three randomly selected locations that were at least 100 m apart from one another and

from stationary treatment sites in Field Experiment I. This experiment was repeated

biweekly, four times.

Data collection

In the first field experiment, bouquets of hydrilla were retrieved on day 14. The

average development period for the larval stage is about 14 days (Cuda et al. 2002).

Therefore, theoretically, the 14-days period allowed the bouquets of hydrilla to be

exposed to attack by all larval instars. The bouquets were bagged in separate 3.8 L

plastic zipper-seal bags (Uline Inc., Pleasant Prairie, WI). Plant samples obtained from

the second experiment were similarly bagged in separate 3.8 L plastic zipper-seal bags.

The bagged samples were placed in a portable ice chest, and transported to the Weed

Biological control laboratory at the University of Florida, Gainesville, Florida, for further

processing. In the laboratory, 100 stems were randomly selected from each depth and

examined under a dissecting microscope for C. lebetis larvae and mining damage.

Data analyses

Data on the percentage of hydrilla meristems damaged by larvae of C. lebetis

and on the abundance of C. lebetis per 100 hydrilla meristems were analyzed with a

generalized linear models procedure using a binomial distribution family with a logit link

(v. 14; StataCorp., College Station, Texas, U.S.). Results were presented in terms of

odd ratios because odd ratios provide direction and magnitude of differences between

proportions (Rita and Komonen 2008, Warton and Hui 2011). For all data analyses, the

level of significance was held at α = 0.05.

Results

Greenhouse Experiments

Experiment I: Inoculation of neonate Cricotopus lebetis on the water surface

Adult emergence (Table 3-2) and larval feeding damage on the apical meristems of

hydrilla (Figure 3-1) occurred at all four treatment depths. Logistic regression analysis

showed survival of neonates to adult stage, expressed as the proportion of adult

emergence, differed statistically among the depth treatments (Table 3-2). The likelihood

of adult emergence (survival) was 4.5 times greater in 0.9 m treatments (P=0.0038) and

2.4 times greater in 1.8 m treatments (P=0.0459) when compared to adult emergence in

0 m treatments. The probability of adult emergence in 2.7 m and 0 m treatments did not

differ statistically (P=0.805). Sex ratio of the emerged adults, proportion of males, did

not differ statistically among depth treatments (Table 3-3). The sex ratios (male: female)

were 1:1.33, 1:1.00, 1:1.08, and 1:1.04 for treatments 0 m, 0.9 m, 1.8 m, and 2.7 m,

respectively. Egg mass production of the collected adults by treatment tube was

significantly higher in 0.9 m treatment, but not statistically different among the rest of the

treatments (Figure 3-2). The number of eggs per mass (mean ± SE) was 138.27 ± 9.32,

135.42 ± 5.30, 137.29 ± 7.67, 153.00 ± 21.08 in treatments 0 m, 0.9 m, 1.8 m, and 2.7

m, respectively. Adult emergence commenced on day 16 in all treatments, and on day

14 in the growth chamber treatment (Figure 3-3).

Mean percentage of apical meristems damaged by C. lebetis larvae exceeded

85% in all treatments, no matter the depth of the host plant (Figure 3-1A). Feeding

damage was highest in 1.8 m treatment (P=0.0398), followed by 0.9 m (P=0.4332), but no statistical differences were detected between the means of any depth treatments

(Figure 3-1A). However, mean length of the stems was significantly longer at 0.9 m than

stems at 0 m or 1.8 m, but similar to 2.7 m treatments. Stems at the 0 m treatment level

were significantly shorter than all other treatment depths, while stems at 1.8 m were

intermediate in length (Figure 3-1B).

Experiment II: Inoculation of Cricotopus lebetis egg masses on the water surface

At all depth treatments, egg masses of C. lebetis inoculated on the water surface

were observed to sink to the bottom of the Plexiglas tubes, where the larvae hatched

and attacked hydrilla. Results in this experiment exhibited trends similar to those in

greenhouse experiment I. Logistic regression analysis showed the likelihood of adult

emergence was significantly 2.1 times higher in the 0.9 m depth treatment compared to

the 0 m treatment (P=0.0058) but not statistically different among the 0 m, 1.8 m, and

2.7 m depth treatments (Table 3-4). The sex ratio of the insect did not differ among

depth treatments (Table 3-5). The numerical value sex ratio (male: female) was 1:1.17;

1:1.08; 1:1.22; and 1:0.92 in treatments 0 m, 0.9 m, 1.8 m, and 2.7 m, respectively. The

production of egg masses by females that emerged from different depths was higher in

0.9 m and 1.8 m treatments and differed statistically among treatments (Figure 3-5).

The number of eggs per mass (mean ± SE) was 145.25 ± 13.13; 162.67 ± 6.28; 151.27

± 7.39; and 167.13 ± 12.69 in treatments 0 m, 0.9 m, 1.8 m, and 2.7 m, respectively.

Damage to the apical meristem of the bouquets did not differ statistically among

the four depth treatments (Figure 3-4a). Compared to 0 m treatment, the stem height

was longest in 0.9 m and 2.7 m treatments. Stem length in 2.7 m and 0 m treatments

did not differ. The initiation of adult emergence in the greenhouse experiment ranged 78

from 16 days in all depth tubes except 0 m depth, which was one day later, and lasted as long as 31 days for the 0 m depth (Figure 3-6).

Field Experiments

Experiment I: Use of sentinel stems to determine the vertical foraging range of Cricotopus lebetis

Apical meristems of the hydrilla sentinel stems, at all treatment depths, were damaged and infested by larvae of C. lebetis (Table 3-6). The larvae of C. lebetis damaged 15% of the sentinel stems at 0 m depth treatment, 23% of stems placed at

0.45 m treatment, and 14% of stem at 0.9 m treatment. Logistic regression analysis showed the likelihood of larval damage on hydrilla was 74% higher at 0.45 m treatments

(P=0.018), compared to the 0 m treatment. However, the likelihood of larval damage on the apical meristems of the stems at 0.9 m and 0 m treatments did not differ statistically

(P=0.847). Similarly, the odds of larval presence did not differ among treatments (Table

3-6).

Experiment II: Survey of naturally growing stems to determine the vertical foraging range of Cricotopus lebetis

Larval damage to apical meristems of hydrilla and larvae of C. lebetis occurred on naturally growing plants collected at all treatment depths (Table 3-7). Logistic regression analysis detected no statistical difference in the proportion of damaged hydrilla apical meristems growing at the surface (39%), mid-depth (42%), and bottom sections (45%) of the water column. Similarly, the odds of larval presence, expressed as the number of larvae per apical meristem of hydrilla, did not differ among the sampled sections of the water column (Table 3-7).

Discussion

Results from the greenhouse study provided the first empirical evidence of larvae of C. lebetis attacking submersed hydrilla at depths ranging from 0 m to at least 2.7 m.

Field studies conducted at Lake Istokopoga provided further evidence of the ability of C. lebetis to attack hydrilla ranging from the water surface level to the hydrosoil, which occurred at a depth of approximately 0.9 m. Previous reports of the vertical foraging depth of C. lebetis relied on anecdotal evidence (Cuda et al. 2002, 2011). For example,

Cuda et al. (2011) speculated that, in Crystal River, Florida, the established population of C. lebetis was attacking hydrilla to depths of up to 0.7 m below the water surface.

Results from the greenhouse experiments demonstrated that egg masses can sink to depths ranging from 0 m to at least 2.7 m, where the neonates can successfully hatch, locate and attack their host plant, and complete their development. In addition, results from the greenhouse study demonstrated that larvae hatching at the water surface, for example, in cases where the sticky egg masses are attached to substrates floating on the water surface, are able to swim or drift, locate, and attack hydrilla to depths of at least 2.7m.

In Florida, hydrilla commonly occurs at depths of up to 3 m (Langeland 1996) with half of the plant’s biomass occurring within the upper 0.5 m of the water column

(Haller and Sutton 1975). This growth habit suggests that most of the hydrilla in Florida grows within a depth range accessible to larvae of C. lebetis. Intense larval feeding damage to the apical meristem of hydrilla (over 85%) was observed in all greenhouse depth treatments. This finding indicated that C. lebetis can attack hydrilla not only within the upper 0.5 m of the water column where the dense surface mats occur, but also the sprouting hydrilla at the hydrosoil level. Higher rate of adult emergence and egg mass 80

production observed at 0.9 m demonstrated the insect can sustain a higher population

level in a depth range that has been reported to contain half of the hydrilla biomass

(Langeland 1996). The results from the field study in Lake Istokpoga provided additional

evidence of the ability of C. lebetis to attack hydrilla wherever it occurred in the water

column.

Percent adult emergence of C. lebetis observed in the greenhouse study (range,

16% to 51%) was comparable to that reported in previous laboratory studies. For

instance, Cuda et al. (2002) reported less than 30% adult emergence of C. lebetis

reared on the Florida strain of hydrilla whereas Stratman et al. (2013a) reported 56% adult emergence of C. lebetis reared on a similar strain of hydrilla. Although the proportion of adult emergence in the greenhouse study varied extensively among

treatments (range, 16% to 51% in experiment I and 23% to 40% in experiment II), the

proportion of damage to the apical meristem did not vary (range, 87% to 94% in

experiment I and 92% to 94% in experiment II). Larval density is positively correlated

with the proportion of apical meristem damage (Cuda et al. 2011). Therefore, the high

level of apical meristem damage observed in greenhouse treatments indicated that

early instars occurred in high densities in all treatments but, as the insects developed,

mortality occurred at varying rates among the treatments. A similar survival trend,

relatively high and equal larval densities among treatments of different levels of abiotic

stresses but varying rates of adult eclosion, was observed by Baniszewski et al. (2015).

This study reported that although subjecting the egg stage to refrigeration for two days

did not impact the resultant larval stage, it disproportionately increased mortality of the

pupal stage and significantly reduced adult eclosion.

The relatively low survival of C. lebetis in the 0 m treatments was not surprising.

Photosynthetic and respiration activities by the hydrilla surface mats have been reported to cause drastic diurnal fluctuations of water quality, which imperil fauna inhabiting the impacted water (Van et al. 1976). Equally important, quality of hydrilla apical meristems in the 0 m treatment was likely poor due to the air/water interphase and lack of room for vertical elongation. This was evidenced by the 0 m treatment having the least stem elongation during the course of the study. Other aquatic insects, such as the rice water weevil [Lissorhoptrus oryzophilus Kuschel (Coleoptera: Curculionidae)], avoid host plants growing at the surface level and prefer plants growing in deeper waters (Stout et al. 2002, Tindall et al. 2013). In the studies conducted at Lake Istokpoga, the proportion of damaged hydrilla apical meristems likely provided the better estimate of larval activity and abundance than the number of larvae observed in the stem samples. Previous studies hypothesize that sampling procedures of stem samples often captures the older larval instars but excludes most of the other life stages of C. lebetis and their resultant damage to the plants (Cuda et al. 2002).

Results in this study showed the vertical foraging depth of C. lebetis was higher than the reported depths of other biological control agents of hydrilla. For example,

Hydrellia pakistanae Deonier (Diptera: Ephydridae) attacks the top 20 cm of the hydrilla canopy (Wheeler and Center 2001), and Bagous affinis Hustache (Coleoptera:

Curculionidae) only attacks hydrilla exposed during dry seasons or drawdowns (Bennett and Buckingham 1991). Field observations in Australia suggested that Bagous hydrillae

O’Brien attacked hydrilla within the upper 1 m of the water surface (Balciunas and

Purcell 1991).

Knowing the foraging range of C. lebetis will be useful in the mass rearing of C. lebetis, which has been recommended for augmentative control of hydrilla (Cuda et al.

2002, Stratman et al. 2013b). In addition, this information can be used in designing field sampling and monitoring techniques, mass rearing facilities, and in predicting the efficacy of the insect in the management of hydrilla. Additionally, results from greenhouse experiments confirm that both neonates and the egg stage can be used to effectively inoculate release sites.

Although this work does not elucidate the maximum foraging depth of the insect, it does provide empirical evidence that C. lebetis has the capacity to reach and attack hydrilla along the entire water column of the shallow lakes and rivers in which it is growing in Florida. Additional research should focus on the effect of larval feeding damage on the growth pattern of hydrilla under field conditions and on the potential impact of larval feeding on hydrilla sprouting from tubers or turions.

Table 3-1. Description of the treatments in a greenhouse experiment designed to evaluate how deep Cricotopus lebetis Sublette neonates can swim to locate Hydrilla verticillata (L.f.) Royle at four different depth treatments in artificial water columns. Treatment Treatment Description number (depth, cm) 1 0 A bouquet of hydrilla was placed in 30 cm tubes, ensuring the hydrilla tips were just below the water surface 2 90 A bouquet of hydrilla was placed at the bottom of a 90 cm deep tube 3 180 A bouquet of hydrilla was placed at the bottom of a 180 cm deep tube 4 270 A bouquet of hydrilla was placed at the bottom of a 270 cm deep tube

Table 3-2. Summary of the logistic regression examining the effect of water depth on the proportion of Cricotopus lebetis Sublette adult emergence from Plexiglas tubes inoculated with C. lebetis neonates in greenhouse experiment I. Depth Mean 95% CI for the P-Value Odds 95% CI for the (m) Proportion Mean Ratio Odds Ratio Survival Lower Upper Lower Upper 0 0.1856 0.1063 0.3039 0.9 0.5072 0.3363 0.6765 0.0038 4.517 1.726 11.819 1.8 0.3505 0.2393 0.4805 0.0459 2.368 1.017 5.510 2.7 0.1550 0.1202 0.1976 0.5327 0.805 0.394 1.643

Table 3-3. Sex ratio (mean proportion males) of Cricotopus lebetis Sublette that emerged from Plexiglas tubes containing Hydrilla verticillata (L.f.) Royle at four depths. Plexiglas tubes were inoculated with C. lebetis neonates in greenhouse experiment I. Depth Mean 95% CI for the P-Value Odds 95% CI for the (m) Proportion Mean Ratio Odds Ratio Males Lower Upper Lower Upper 0 0.4272 0.3712 0.4850 0.9 0.5009 0.4039 0.5978 0.1904 1.346 0.852 2.125 1.8 0.4807 0.4451 0.5165 0.1155 1.241 0.944 1.633 2.7 0.4942 0.3954 0.5934 0.2392 1.310 0.823 2.084

100 A

DamagedMeristem (%) 85

80 0 0.9 1.8 2.7

B a ab 35

b 30

25 c Stem Height Stem (cm)

0 0.9 1.8 2.7 Depth (m)

Figure 3-1. Impact of Cricotopus lebetis Sublette on Hydrilla verticillata (L.f.) Royle grown at different water depth treatments in greenhouse experiment I. A) Percentage (mean ± 95% CI) of damaged hydrilla meristems. B) Length (mean ± 95% CI) of H. verticillata stems. Different letters indicate means are statistically different at α = 0.05.

Figure 3-2. Number of egg masses (mean ± 95% CI) produced by Cricotopus lebetis Sublette that emerged from Hydrilla verticillata (L.f.) Royle grown at different water depth treatments in greenhouse experiment I. Treatments were initially inoculated with C. lebetis neonates. Different letters indicate means are statistically different at α = 0.05.

Figure 3-3. Cumulative percentage emergence pattern of adult Cricotopus lebetis Sublette following development of the insects in Plexiglas tubes on Hydrilla verticillata (L.f.) Royle at different depths. The tubes were inoculated with C. lebetis neonates in greenhouse experiment I. Error bars were not included to increase clarity of the graphical trends.

Table 3-4. Summary of the logistic regression examining the effect of water depth on the proportion of Cricotopus lebetis Sublette adult emergence from Plexiglas tubes inoculated with C. lebetis eggs in greenhouse experiment II. Depth Mean 95% CI for the P-Value Odds 95% CI for the (m) Proportion Mean Ratio Odds Ratio Emergence Lower Upper Lower Upper 0 0.2315 0.1745 0.3004 0.9 0.3892 0.3075 0.4776 0.0058 2.115 1.275 3.508 1.8 0.2937 0.2236 0.3751 0.2029 1.380 0.828 2.299 2.7 0.2559 0.2026 0.3175 0.5609 1.141 0.716 1.818

Table 3-5. Sex ratio (mean proportion males) of Cricotopus lebetis Sublette that emerged from Plexiglas tubes containing Hydrilla verticillata (L.f.) Royle at one of four depths. Plexiglas tubes were inoculated with C. lebetis eggs in greenhouse experiment II. Depth Mean 95% CI for the P-Value Odds 95% CI for the (m) Proportion Mean Ratio Odds Ratio Males Lower Upper Lower Upper 0 0.4553 0.3616 0.5522 0.9 0.4769 0.4251 0.5291 0.6862 1.091 0.701 1.696 1.8 0.4540 0.3950 0.5143 0.9817 0.995 0.629 1.573 2.7 0.5176 0.4921 0.5430 0.2097 1.284 0.859 1.920

100 A

90 MeristemDamage (%) 85

80 0 0.9 1.8 2.7

40 B ab

30 b

b 25 a Stem Height Stem (cm) 20

0 0.9 1.8 2.7 Depth (m)

Figure 3-4. Impact of Cricotopus lebetis Sublette on Hydrilla verticillata (L.f.) Royle grown at different water depth treatments in greenhouse experiment II. A) Percentage (mean ± 95% CI) of damaged hydrilla meristems. B) Length (mean ± 95% CI) of H. verticillata stems. Different letters indicate means are statistically different at α = 0.05.

Figure 3-5. Number of egg masses (mean ± 95% CI) produced by Cricotopus lebetis Sublette that emerged from Hydrilla verticillata (L.f.) Royle grown at different water depth treatments in greenhouse experiment II. Treatments were initially inoculated with C. lebetis eggs. Different letters indicate means are significantly different at α = 0.05.

Figure 3-6. Cumulative percentage emergence pattern of adult Cricotopus lebetis Sublette following development of the insect in Plexiglas tubes on Hydrilla verticillata (L.f.) Royle at different depths. The tubes were inoculated with C. lebetis eggs in greenhouse experiment II. Error bars were not included to increase clarity of the graphical trends.

Table 3-6. Summary of the logistic regression analysis examining the effects of water depth on the proportion (mean ± 95% CI) of damaged apical meristems and number (mean ± 95% CI) of Cricotopus lebetis Sublette larvae per apical meristems of Hydrilla verticillata (L.f.) Royle on sentinel bouquets placed at 0 m, 0.45 m, and 0.9 m below the water surface in Lake Istokpoga, Florida, 2017. Variable Depth Mean 95% CI for the P-Value Odds 95% CI for the Odds Mean Ratio Ratio Lower Upper Lower Upper 0.00 0.152 0.098 0.207 Proportion damaged meristems 0.45 0.231 0.154 0.309 0.0180 1.7450 1.1009 2.7657 0.90 0.144 0.057 0.230 0.8470 0.9309 0.4505 1.9238

0.00 0.003 0.001 0.006 Number larvae 0.45 0.004 0.000 0.008 0.709 1.2514 0.3856 4.0618 0.90 0.002 0.000 0.004 0.403 0.6167 0.1984 1.9167

Table 3-7. Summary of the logistic regression examining the effects of water depth on the proportion (mean ± 95% CI) of damaged apical meristems and number (mean ± 95% CI) of Cricotopus lebetis Sublette larvae per apical meristem of Hydrilla verticillata (L.f.) Royle stems sampled from the surface level, mid-depth, and bottom of Lake Istokpoga, Florida, 2017. Variable Depth Mean 95% CI for the P- Odds 95% CI for the Odds Mean Value Ratio Ratio Lower Upper Lower Upper Proportion damaged Surface 0.39 0.30 0.48 meristem Mid-Depth 0.42 0.34 0.50 0.5700 1.1490 0.7119 1.8540 Bottom 0.45 0.38 0.52 0.1640 1.3578 0.8824 2.0894

Number larvae Surface 0.01 0.004 0.019 Mid-Depth 0.02 0.007 0.024 0.3810 1.3574 0.6850 2.6899 Bottom 0.02 0.007 0.025 0.3950 1.3599 0.6693 2.7629

CHAPTER 4 COMPETITIVE INTERACTIONS BETWEEN HYDRILLA VERTICILLATA AND THE NATIVE VALLISNERIA AMERICANA AS INFLUENCED BY CRICOTOPUS LEBETIS

Introduction

Competition and herbivory are two major factors among numerous interspecific interactions that regulate the structure, function, and dynamics of aquatic plant communities (Gopal and Goel 1993). Although insect herbivores rarely kill their hosts, they usually damage host tissue and predispose the impacted host to disease infections; thus, impeding their competitive ability (Crawley 1989, Louda et al. 1990).

Under constant herbivore pressure, unpalatable or tolerant species gain a competitive advantage over plant species more susceptible to herbivory in accessing limited resources (Crawley 1989, Hulme 1996). In the native range, endemic specialist herbivores naturally regulate native plant species (Williams 1954, Keane and Crawley

2002). However, competitive release of an invasive species, such as hydrilla, occurs in

the adventive range, because they are liberated from natural regulation by their co- evolved specialist herbivores and pathogens (Williams 1954, Elton 1958). Generalists, on the other hand, are non-selective and have been reported to play a critical role in providing community resistance to invasion by exotic plant species (Strong et al. 1984,

Parker and Hay 2005).

Before the 1990s, the role of herbivory in the regulation of aquatic communities

was largely underestimated by ecologists (Gregory 1983, Lodge 1991, Newman 1991,

Wood et al. 2016). However, a mounting body of evidence, especially from classical

biological control studies, has since shown that herbivory is one of the major factors that

regulates macrophyte communities (Wood et al. 2016). To date, biological control

practitioners have identified many aquatic insect species that selectively attack invasive

plant species. These specialists weaken the fitness of targeted invasive plants, impeding their ability to compete with native plant species (Van et al. 1998, Grodowitz et al. 2000, Center et al. 2005). For example, an outdoor tank study by Center et al. (2005) revealed that herbivory on water hyacinth [Eichhornia crassipes (Mart.) Solms], by specialist weevils Neochetina bruchi (Hustache) and N. eichhorniae (Warner)

(Coleoptera: Curculionidae) reduced the fitness of water hyacinth, and shifted the competitive balance of the plant in favor of a weaker competitor, water lettuce Pistia stratiotes L. (Araceae). Similarly, Van et al. (1998) demonstrated that damage by the specialist Hydrellia pakistanae Deonier (Diptera: Ephydridae) on hydrilla [Hydrilla verticillata (L.f.) Royle] reduced the competitive ability of hydrilla in favor of a weaker competitor, American eelgrass, Vallisneria americana Michx. In the same study, Van et al. (1998) reported that the Australian weevil Bagous hydrillae O'Brien (Coleoptera:

Curculionidae) damaged hydrilla and similarly reduced its competitive ability. As a result, manipulation of herbivory and competition has attracted interest as a potential tool for managing hydrilla.

Hydrilla is considered one of the most invasive plants in U.S. freshwater ecosystems, and has been recognized for its ability to competitively displacing native plants (Moxley and Langford 1982, Chambers et al. 1993, Posey et al. 1993). Many field studies in the U.S. have revealed cases of hydrilla outcompeting and displacing native aquatic plants. For example, Chambers et al. (1993) reported that hydrilla in Lake

Seminole, Florida, competitively displaced several native plants, which included coontail

(Ceratophyllum demersum L.), fanwort (Cabomba caroliniana Gray), and Illinois pondweed (Potamogeton illinoensis Morong). Similar cases have been reported in other

hydrilla-infested waterbodies in the U.S. (Moxley and Langford 1982, Posey et al. 1993).

Enemy release hypothesis explains that invasive plants, such as hydrilla, are stronger competitors in introduced ranges because, in their new range, they are liberated from natural regulation from their co-evolved specialist herbivores and pathogens (William

1954, Elton 1958, Gopal and Goel 1993, Keane and Crawley 2002). As a result, biological control practitioners generally agree that the release and establishment of specialist herbivores of hydrilla in invaded water bodies will likely suppress the competitive edge that hydrilla has over the native plant species.

Several research teams have evaluated the impact of insect herbivory and competition on hydrilla, using the native plant V. americana as an interspecific competitor and biological control agents as herbivores (Van et al. 1998, Doyle et al.

2007). They reported that in the absence of herbivores, hydrilla readily outcompeted V. americana for light and nutrients, but attack on hydrilla by specialist herbivores shifted the competitive advantage to V. americana (Van et al. 1998). Hydrilla and V. americana belong to the same family (Hydrocharitaceae), but have different growth habits. Unlike hydrilla, V. americana rarely grows up to the surface, does not form a canopy, and relies on light availability near the hydrosoil for photosynthesis (Figure 4-1, Titus and Adams

1979).

The aim of this study was to determine if another potential biological control agent of hydrilla, Cricotopus lebetis Sublette (Diptera: Chironomidae), can reduce the ability of hydrilla to outcompete a native plant, using the native V. americana as a model interspecific competitor. Cricotopus lebetis was discovered in 1992 attacking the apical meristems of hydrilla in Kings Bay, Citrus Co., Florida. Studies have since confirmed

that C. lebetis can reduce the ability of hydrilla to form a dense surface mat (Cuda et al.

2002, Cuda et al. 2011), but there are no reports of how C. lebetis affects the competitiveness of hydrilla.

Materials and Methods

Study Site and Plant Source

The study was conducted in 16 outdoor concrete tanks (approx. 900 L capacity,

2.22 m length × 0.81 m width) located at Bivens Arm Research Center, University of

Florida, Gainesville, Alachua Co., Florida (29.628144°N, 82.353459°W). Hydrilla and V. americana propagules were collected from ponds located at the University of Florida's

Institute of Food and Agricultural Sciences, Center for Aquatic and Invasive Plants

(UF/IFAS CAIP) (29.726796°N, 82.415235°W). Cricotopus lebetis was obtained from

Lake Istokpoga and maintained in a colony in a laboratory facility of the UF/IFAS

Entomology & Nematology Department, Gainesville, Florida (29.633882°N,

82.366733°W), following a procedure described by Cuda et al. (2002).

Set Up and Experimental Design

The study was initiated in August 2015, whereas the second experiment was initiated in September 2016. Initiating the experiments after the month of August, instead of the summer months, avoided the summer heat. In summer months, temperatures in dense surface mats of hydrilla has been reported to reach 45 °C

(Wheeler and Center 2001), a lethal temperature level for C. lebetis (Stratman et al.

2014). In addition, a survey by Cuda et al. (2002) showed field abundance of C. lebetis was high beginning in September, peaking in November.

The study was conducted using a 2 × 2 factorial design, with two levels of C. lebetis herbivory (present and absent) and two levels of plant competition (monoculture

100

and mixed culture), a design similar to that described by Doyle et al. (2007). The study had four treatments: control, competition, herbivory, and competition-herbivory (Table 4-

1). For each field season, treatments were replicated four times in space and conducted over a span of 18 weeks.

Planting

In preparation for planting, each of the 16 tanks randomly received 30 plastic pots (5.1 L, 21.6 cm diameter × 21.6 cm depth) (Figures 4-1, 4-2). Each pot contained approximately 15 cm of Peat Moss BM2 germinating mix (Berger, St. Modestede,

Quebec) mixed with a slow release fertilizer, Osmocote® Plus 15:9:12 (Scotts Miracle-

Gro, Marysville, OH) at 2g L-1 of soil (Gettys personal comm.) and topped with 5 cm of sand (Cuda et al. 2002). The pots were arranged in three rows of 10 pots (Figure 4-2A,

B). After the placement of the pots, all the tanks were filled with well water.

Eight tanks were assigned randomly to monoculture treatments (no interspecific competition) (Figure 4-2A). The remaining eight tanks were assigned to mixed culture treatments (interspecific competition) (Figure 4-2B, Figure 4-3). In mixed culture tanks,

15 pots were each planted with a 15 cm long apical shoot of hydrilla, and the remaining

15 pots were planted with a 15 cm long V. americana ramet (Figure 4-2B). Pots planted with hydrilla and those planted with V. americana were arranged in an alternating pattern both lengthwise and widthwise (Figure 4-2B, 4-3). In monoculture tanks, all pots were planted with 15 cm long apical shoots of hydrilla (Figure 4-2A).

After planting, each tank was covered by a 70% shade cloth (Sunblocker®

Premium, Farmtek., Dyersville, IA) (Figure 4-4) to avoid excessive solar heating within tank cultures (Webb et al. 2012) and infestation of the treatments by unwanted insects; later in the experiment, to confine the inoculated C. lebetis within herbivory and

101

competition-herbivory tanks. The planting dates between replicates were offset by three days.

Introduction and monitoring of Cricotopus lebetis

Plants, in both the monoculture and mixed culture tanks, were allowed to grow for approximately 30 days. At this time, hydrilla and V. americana had grown to the water surface, and hydrilla had formed a surface canopy. For monoculture tanks, four randomly selected tanks were each inoculated with 1,200 larvae of C. lebetis (herbivory treatment); the remaining four monoculture tanks were not inoculated with insects

(control treatments). Similarly, for mixed culture tanks, four randomly selected tanks were each inoculated with 1,200 larvae of C. lebetis (competition-herbivory treatment); the remaining four tanks were not inoculated with insects (plant competition only treatment). Each tank was considered a replicate.

Monitoring of shading effect and temperature

Light intensity and temperature in each treatment tank were measured in 15 minute intervals at the sediment level using an underwater sensor (onset® HOBO®

Pendant® Temperature/Light Data Logger). Light intensity measurements were used to estimate impact of herbivory on hydrilla surface mats.

Plant harvest

Plants were harvested 18 weeks after the inoculation of the tanks with C. lebetis.

Plants in each pot were harvested separately, by cutting them at the sediment level with a Corona®, hand pruner (model number BP 4314D; Corona, California, U.S.).

Harvested plant materials from each pot were sorted by species and placed in approximately 3.8 L zipper seal plastic bags (28 cm length × 28 cm width). The bags were labelled with tank and pot numbers and transferred for further processing in

102

portable ice chests to the Weed Biological control laboratory at the University of Florida,

Gainesville, Florida.

In the laboratory, each bagged sample was processed separately. The plant

samples were removed from the plastic bags and washed with well water to remove

accumulated sediments and debris (Madsen 1993). Each washed plant sample was

individually placed on a tray and processed. For hydrilla samples, 15 stems from each

tank were sampled and their height measured. In addition, the turions in each pot were

counted and recorded. For V. americana, 15 rosettes were sampled from each tank that

contained treatments with V. americana. In addition, the rosettes and leaves in each pot

were counted and recorded. Biomasses of both hydrilla and V. americana were

obtained by placing the plant tissues in heavy-duty kraft paper bags (19.7 cm × 12.1 cm

× 40.6 cm; Uline Inc., Pleasant Prairie, WI) and oven drying the bagged samples at 60

°C in a Blue M Electric forced draft convection oven (TPS, New Columbia, PA), to a constant weight (Madsen 1993).

Data Analyses

Data were analyzed with a generalized linear models procedure, SAS PROC

GLIMMIX, using a distribution function for the appropriate response variable, i.e., for

hydrilla data, lognormal for biomass, plant height, surface area, tip damage, and

Poisson for count data; for V. americana data, normal for biomass, and plant height.

Light intensity was analyzed using linear regression (SAS Version 9.4). Year was

treated as a random effect in all analyses.

Results

Average weekly water temperature in the treatment tanks was 22.9 ºC for the first

trial period initiated in 2015, and 18.5 ºC for the last 13 weeks of the second trial period

103

initiated in 2016 (Figure 4-5). Regression analysis of the light intensity recorded at the

soil surface over time showed the average weekly light intensity remained constant in

the herbivory treatment (df=339, t=339, P=0.0944) and competition-herbivory treatment

(df=339, t=1.16, P=0.2451), the two treatments that included the midge C. lebetis. In contrast, light intensity decreased over time in the control treatment (df =339, t=-2.19,

P=0.0295) and competition treatment (df=339, t=-3.14, P=0.0018), the two treatments without C. lebetis (Figure 4-6). The P-values in the regression analysis tested the null hypothesis that the slope of regression line is equal to zero (no effect).

Two-way ANOVA did not detect an interaction effect between the competition and herbivory factors (Table 4-2). One-way ANOVA revealed herbivory by C. lebetis impacted all the parameters that were measured in hydrilla (Table 4-2). Biomass of hydrilla per pot was significantly lower in treatments with C. lebetis compared to treatments without (ANOVA; F1, 26=34.50, P=0.000, Table 4-2). Herbivory by C. lebetis reduced the biomass of hydrilla by about 70% in monoculture treatments and 51% in mixed culture treatments (Figure 4-7). Turion production also was significantly less in treatments with herbivory by C. lebetis compared to treatments without C. lebetis

(ANOVA; F1, 26=11.83, P=0.002; Figure 4-8). Length of hydrilla stems was significantly

shorter in herbivory treatments compared to non-herbivory treatments (ANOVA; F1,

26=37.33, P=0.000, Figure 4-9). The effect of herbivory reduced height of hydrilla stems by 58% in mixed cultures and by 68% in monoculture treatments (Figure 4-9). Similarly,

damage by C. lebetis on the apical meristem of hydrilla occurred exclusively, as

expected, and ranged from 58% to 67% on hydrilla growing in treatments exposed to C.

lebetis (ANOVA; F1, 26=1069.20, P=0.000; Figure 4-10); no damage was detected on V.

104

americana plants and on hydrilla plants growing in control treatments. Percent cover of

hydrilla canopy was significantly higher in treatment tanks without C. lebetis, and was

significantly lower in treatment tanks with C. lebetis (ANOVA; F1, 26=37.38, P=0.000;

Figure 4-11).

Interspecific competition between hydrilla and V. americana generated varying responses in hydrilla. Hydrilla produced more biomass per pot in mixed culture treatments than in monoculture treatments (ANOVA; F1, 26=37.38, P=0.050; Table 4-2).

Percent cover of the hydrilla canopy was significantly higher in monoculture than in

mixed culture treatments (ANOVA; F1, 26= 7.43, P=0.011). However, as shown on Table

4-2, no differences were detected in stem length (ANOVA; F1, 26=0.34, P=0.563), turion

production (ANOVA; F1, 26 =2.25, P=0.145), and percent meristem damage by C. lebetis

(ANOVA; F1, 26=1.36, P=0.255) between hydrilla in mixed culture and monoculture treatments. In contrast with what was observed with hydrilla, V. americana did not

invade pots originally planted with hydrilla.

In mixed culture treatments, attack on hydrilla by C. lebetis benefited V. americana. Biomass of V. americana was significantly higher in plants in mixed culture treatments with herbivory than in treatments without herbivory (P=0.030; Figure 4-12).

V. americana biomass was 2.35 times more in treatments with herbivory than in

treatments without herbivory. However, no statistical differences were detected between

rosette height (P=0.908; Figure 4-13A), number of rosettes per pot (P=0.428; Figure 4-

13B) and number of leaves per pot (P=0.428; Figure 4-13C) of plants in treatments with

and without herbivory. Similarly, no statistical differences were detected between the

105

biomass of hydrilla that expanded into pots originally planted with V. americana in treatments with C. lebetis and treatments without C. lebetis (P=0.159; Figure 4-14).

Discussion

Our study provided the first empirical evidence of the effectiveness of C. lebetis in attacking and weakening the fitness of hydrilla in a mesocosm environment. Our results demonstrated the attack on the apical meristem of hydrilla by larvae of C. lebetis impeded the ability of the plant to produce biomass and turions, grow in height, and form surface mats. Readings from the underwater data loggers further revealed that in the absence of herbivory, plant growth reduced the amount of light penetrating the water column over time. However, in the presence of herbivory, the results revealed the amount of light penetrating the water column remained constant over time, indicating the attack by C. lebetis prevented the surface mats of hydrilla from thickening and increasing their shading capacity. Prior to this study, empirical evidence of larvae of C. lebetis restricting production of biomass and growth of the stem in hydrilla were based entirely on laboratory studies (Cuda et al. 2011, 2016).

Additionally, our study demonstrated that selective attack on the meristems of hydrilla by an insect herbivore could shift the competitive balance between hydrilla and native macrophyte species, in favor of the latter. This was evidenced by V. americana producing more biomass in mixed culture treatments containing C. lebetis than in mixed culture without C. lebetis. Lack of evidence of larval feeding damage on V. americana was not surprising. A laboratory study by Stratman et al. (2013b) reported that V. americana was a poor host of C. lebetis, and explained that the survival rate of larvae of

C. lebetis feeding and developing on the plant was less than 7%. The shift in competitive balance between an invasive plant and a native plant, observed in the

106

present study, was consistent with a report by Van et al. (1998), who similarly demonstrated in an outdoor study that attack by H. pakistanae, weakened the fitness of hydrilla and transferred the competitive advantage to V. americana.

The suppressed production of hydrilla biomass (51% to 70% reduction), by C. lebetis observed in the present study was consistent with the laboratory findings by

Cuda et al. (2011), who reported that larval feeding by C. lebetis reduced hydrilla biomass up to 99%. The demonstrated effects of C. lebetis herbivory on hydrilla, which included reduced stem length, surface cover, and biomass, are desirable in the control of hydrilla because the resultant reduced shading at the water surface allows more photosynthetic active radiation to filter through the water column and reach native macrophyte species. In the present study, these effects were manifested by the higher biomass production of V. americana in mixed culture treatments containing C. lebetis than in treatments without C. lebetis. Equally important, the suppressed growth of hydrilla by C. lebetis will likely ease other ecological challenges such as the reported

‘invasional meltdown’, where the dense surface mats of hydrilla facilitate invasion by other exotic species, such as South American catfish [Hoplosternum littorale (Hancock,

1828)] (Simberloff and von Holle 1999, Nico and Muench 2004) and the epiphytic cyanobacterium that causes avian vacuolar myelinopathy (AVM) (Bidigare et al. 2009) and economic problems, such as impediment of water-based recreational and navigational activities (Langeland 1996).

Although, herbivory in the present study, did not restrict the spread of hydrilla into pots originally planted with V. americana, it suppressed stem growth and production of the turions. Because hydrilla reproduces and spreads vegetatively via turions, tubers,

107

and stem fragments and via the growth of horizontal stems (Yeo et al. 1984, Langeland

1996), insect herbivore attack on hydrilla on only one part of the hydrilla may slow but not stop localized expansion of the plant. However, the cumulative effect of herbivory on hydrilla over time, as predicted by the enemy release hypothesis, may restrict the abundance of the plant in invaded macrophyte communities. In the present study, larval feeding damage on the hydrilla stems may have directly hindered the production of turions because the turions are formed in leaf axils of the stems (Yeo at al. 1984).

Grodowitz et al. (2007) reported that attack by another dipteran herbivore, H. pakistanae, on hydrilla similarly reduced turion production.

The higher production of hydrilla biomass per pot in mixed culture than in monoculture treatments suggested intraspecific competition among hydrilla plants was stronger than the interspecific competition. Grodowitz et al. (2007) and Van et al. (1998) reported similar findings in studies conducted outdoors. Van et al. (1998, 1999),

Steward (1991b), and Mony et al. (2007) explained that where nutrient availability was not limiting, hydrilla readily outcompeted V. americana. Van et al. (1998 and1999) demonstrated one hydrilla plant was competitively equivalent to about seven V. americana plants. However, contrasting findings have been reported in nutrient deficient environments, such as pure sand in laboratory conditions, where hydrilla was reportedly readily outcompeted by V. americana, without the influence of herbivory (Steward

1991b, Van et al. 1999, Mony et al. 2007). In natural waterbodies, however, carbon and light are the major limiting resources for macrophyte species and play a critical role in interspecific competition (Madsen and Sand-Jensen 1991, Santamaria 2002). Hydrilla has advanced carbon-concentrating mechanisms (Salvucci and Bowes 1983, Rao et al.

108

2002), canopy-forming growth habit (Yeo et al. 1984), and lower light compensation and saturation points (Van et al. 1976, Bowes et al. 1977) that allow it to outcompete native plants for these two limiting resources. These previous reports, coupled with the results in the present study demonstrating the higher production of hydrilla biomass in mixed culture than in monoculture, confirms that in the absence of natural enemies (e.g., C. lebetis) hydrilla readily outcompetes native plants and eventually dominates macrophyte communities in introduced ranges.

The lack of interaction between herbivory and interspecific competition observed in the present study was consistent with studies by Doyle et al. (2007) and Grodowitz et al. (2007), which similarly showed that herbivory by H. pakistanae on hydrilla and interspecific competition between V. americana and hydrilla acted independently, without synergy or antagonism. Synergy have been reported by Cuda et al. (2016) between effects of C. lebetis and fungal pathogens Mycoleptodiscus terrestris (Mt), and low doses of acetolactate synthase-inhibiting herbicide, imazamox, on hydrilla. Absence of reported antagonism in the present and previous studies is encouraging as it indicated C. lebetis could be included in integrated weed management programs.

The average weekly water temperature in the present study (22.9 ºC for the first trial and 18.5 ºC, which represented only the last 13 weeks of second trial) was within the reported minimum and maximum thresholds (15 and 32 °C) that are conducive for optimal development of C. lebetis (Stratman et al. 2014). Equally important, the 58% to

68% apical meristem damage observed in this study was in the same range as the peak damage rates of 60% to 70% reported under field conditions, suggesting the results of

109

the present study closely paralleled the impact of C. lebetis observed under field conditions (Cuda et al. 2002).

The successful inoculation of C. lebetis in the outdoor tanks and its subsequent establishment and attack on hydrilla supported a recommendation by Cuda et al. (2011) and Stratman et al. (2013a) that C. lebetis could be mass reared in laboratories and released en masse (augmentative biological control) in waterbodies infested by hydrilla for controlling the plant. In addition, our results not only demonstrated the ability of C. lebetis to attack hydrilla and shift competitive balance in favor of native plants, but provided further support to the emerging body of evidence over the last few decades

(Wood et al. 2016) that insect herbivory plays a critical role in structuring macrophyte communities.

However, it is important to note that diverse biotic and abiotic factors influence the ultimate efficacy of biological control agents of hydrilla in natural environments.

Previous studies have identified such factors as attacks by native parasitoids (Coon et al. 2014), climate suitability and quality of the host plants (Grodowitz et al. 1997), and anthropogenic actions such as nutrient inputs into the aquatic systems (Van et al. 1998,

Ruhl and Rybicki 2010). Therefore, future research should investigate how these factors impact the role of C. lebetis in mediating interspecific competition. Specifically, factors reported to have additive, antagonistic, or synergistic effects on insect herbivory (e.g., quality of host plants, eutrophication levels, and predators such as fish).

110

Table 4-1. Description of the treatments in an experiment designed to evaluate the impact of Cricotopus lebetis Sublette on the competitive interaction between Hydrilla verticillata (L.f.) Royle and Vallisneria americana Michx. Treatment Competition Herbivory Treatment description factor factor Control monoculture absent Hydrilla only Competition mixed culture absent Hydrilla, V. americana Herbivory monoculture present Hydrilla and Cricotopus lebetis Competition- mixed culture present Hydrilla, V. americana, and herbivory Cricotopus lebetis

111

A B

Figure 4-1. Morphological characteristics of Vallisneria americana Michx. and Hydrilla verticillata (L.f.) Royle. A) Vallisneria americana. B) Hydrilla verticillata. Source: IFAS, Center for Aquatic and Invasive Plants, University of Florida, Gainesville. Available at http://plants.ifas.ufl.edu/plants-by-common-name/. Accessed November 12, 2017.

112

Figure 4-2. Planting details for monoculture and mixed culture treatments in outdoor concrete tanks (approx. 900 L capacity, 2.22 m length × 0.81 m width). A) Monoculture treatments: control and herbivory treatments. B) Mixed culture treatments: competition and competition-herbivory treatments. Circles without patterns represent pots planted with Hydrilla verticillata (L.f.) Royle and circles with stripes represent pots planted with Vallisneria americana Michx.

113

Figure 4-3. Experimental design for competition and competition-herbivory treatments. Each tank contained 15 pots planted with one 15 cm long apical shoot of Hydrilla verticillata (L.f.) Royle and additional 15 pots, each planted with a 15 cm long Vallisneria americana Michx. ramet. Photograph by Eutychus Kariuki.

114

Figure 4-4. Hydrilla verticillata (L.f.) Royle with a dense surface canopy in one of the tanks covered with 70% shade cloth (Sunblocker ®Premium, Farmtek., Dyersville, IA). Photograph by Eutychus Kariuki.

115

30 A 2015/2016 28

18 Average Weekly Water Temp (°C) Temp Water Weekly Average 16

14 0 2 4 6 8 10 12 14 16 18 20 Weeks After Treatment

24 B 2016/2017 22

Average Weekly Water Temp (°C) Temp Water Weekly Average 14

12 2 4 6 8 10 12 14 16 18 20 Weeks After Treatment

Figure 4-5. Average weekly water temperature recorded by an underwater sensor, onset® HOBO® Pendant® Data Logger. A) Water temperature data for experiment conducted between 2015 and 2016. B) water temperature data for experiment conducted between 2016 and 2017. In the second experiment, data recording began on the 6th week.

116

Control Competition 800

P=0.0295 P=0.0018 P=0.0295 P=0.0018 600

400 Light Intensity (lx) Intensity Light

200

1400 Herbivory Herbivory + Competition

1200

1000

800 P=0.0944 P=0.2451 Light Intensity (lx) Intensity Light

600 P=0.0944 P=0.2451

400 0 2 4 6 8 10 12 14 16 180 2 4 6 8 10 12 14 16 18 Weeks After Treatment Weeks After Treatment

Figure 4-6. Relationship between the light intensity (lx) (measured at the soil level in tanks of the control, competition, herbivory, and herbivory-competition treatments) and time after the initiation of the treatments. The graphs are based on data back-transformed from a log scale. The P-value tested the null hypothesis that the slope of regression line is equal to zero (no effect).

117

Table 4-2. Two-way analysis of variance examining the effects of herbivory by Cricotopus lebetis Sublette and interspecific competition by Vallisneria americana Michx. on the ability of Hydrilla verticillata (L.f.) Royle to produce biomass and turions, elongate, and form surface cover. Hydrilla Source of Variation Num DF Den DF F Value P>F Biomass per pot Herbivory 1 26 34.5 0.000 Competition 1 26 4.24 0.05 Herbivory × Competition 1 26 2.18 0.152 Stem length Herbivory 1 26 37.33 0.000 Competition 1 26 0.34 0.563 Herbivory × Competition 1 26 1.03 0.32 Surface cover Herbivory 1 26 37.38 0.000 Competition 1 26 7.43 0.011 Herbivory × Competition 1 26 1.11 0.301 Meristem damage Herbivory 1 26 11069.2 0.000 Competition 1 26 1.36 0.255 Herbivory × Competition 1 26 1.36 0.255 Turion per pot Herbivory 1 26 11.83 0.002 Competition 1 26 2.25 0.145 Herbivory × Competition 1 26 4 0.056

118

10 No Herbivory Herbivory

6 ** 4 ***

Hydrilla Dry Biomass (g per pot) Biomass Dry Hydrilla 2

0 No Competition Competition

Figure 4-7. Comparison of dry biomass (means ± SE) of Hydrilla verticillata (L.f.) Royle produced in treatments with and without Cricotopus lebetis Sublette. Asterisk denote a statistically difference between herbivory and non-herbivory treatments (**P < 0.01; ***P < 0.001). Interaction effect was not significant (F1, 26=2.18, P=0.152).

119

1.6

1.4 No Herbivory Herbivory 1.2

1.0

0.8

0.6 Turions per Pot

0.4

0.2

0.0 No Competition Competition

Figure 4-8. Comparison of turions (means ± SE) of Hydrilla verticillata (L.f.) Royle produced in treatments with and without Cricotopus lebetis Sublette. Interaction effect was not significant (F1, 26=4.00, P=0.056).

120

50 No Herbivory Herbivory

30 *** **

20 Hydrilla Stem Length Stem (cm) Hydrilla 10

0 No Competition Competition

Figure 4-9. Comparison of stem length (means ± SE) of Hydrilla verticillata (L.f.) Royle grown in treatments with and without Cricotopus lebetis Sublette. Asterisk denote a statistical difference between herbivory and non-herbivory treatments (**P < 0.01; ***P < 0.001). Interaction effects were not significant (F1, 26=1.03, P=0.320).

121

No Herbivory 80 Herbivory *** 60 ***

40 DamagedMeristem (%) 20

0 No Competition Competition

Figure 4-10. Difference in percentage of damaged apical meristems (means ± SE) of Hydrilla verticillata (L.f.) Royle grown in treatments with and without Cricotopus lebetis Sublette. Asterisk denote a significant difference between herbivory and non-herbivory treatments (***P < 0.001). Interaction effects were not significant (F1, 26=1.36, P=0.255).

122

1.0 No Herbivory Herbivory

0.8

*** 0.6 *** 0.4 Proportionof Surface Cover 0.2

0.0 No Competition Competition

Figure 4-11. Comparison of surface cover (means ± SE) of Hydrilla verticillata (L.f.) Royle grown in treatments with and without Cricotopus lebetis Sublette. Asterisk denotes a statistical difference between herbivory and non-herbivory treatments (***P < 0.001). Interaction effects were not significant (F1, 26=1.11, P=0.301).

123

* 6

2 American Eelgrass American Biomass (g per pot)

0 No Herbivory Herbivory

Figure 4-12. Difference in the biomass of American eelgrass, Vallisneria americana Michx., grown in mixed culture treatments with and without Cricotopus lebetis Sublette (herbivory). Asterisk denotes a statistical difference between herbivory and non-herbivory treatments (*P < 0.05).

124

40 AA a nsa 30

10 Rossette LengthRossette (cm)

0 B nsa B a 15

Rossette(count) 5

0 C nsa 200

150 a

100 Leaves (count) 50

0 No Herbivory Herbivory

Figure 4-13. Comparison of characteristics of Vallisneria americana Michx. interplanted with Hydrilla verticillata (L.f.) Royle in mixed culture treatments with Cricotopus lebetis Sublette (herbivory) and without C. lebetis (no herbivory). A) rosette length B) rosette count, and C) leaf count. The letters ‘ns’ indicate the means for non-herbivory and herbivory treatments are not statistically different at α = 0.05.

125

Hydrilla Expansion (gper Expansion pot) Hydrilla 2

0 No Herbivory Herbivory

Figure 4-14. Comparison of biomass of Hydrilla verticillata that expanded into pots originally planted with Vallisneria americana Michx. in treatments with and without Cricotopus lebetis Sublette. The letters ‘ns’ indicate the means for non-herbivory and herbivory treatments are not statistically different at α = 0.05.

126

CHAPTER 5 CONCLUSIONS

Management of hydrilla in the U.S. is problematic due to numerous challenges such as development of herbicide-resistant biotypes of hydrilla, high operational costs and non-target effects of mechanical harvesters, and ineffectiveness of previously released insect biological control agents. Proposed solutions include searching for new biological control agents to complement the existing management tools. As part of the proposed efforts, a hydrilla stem mining midge, Cricotopus lebetis Sublette (Diptera:

Chironomidae), has been identified as a promising biological control agent of hydrilla.

The insect was discovered in 1992 attacking the apical meristem of hydrilla in Kings

Bay, Florida (Cuda et al. 2002). Evidence from previous laboratory and field studies has shown the larval stage of the insect mines the apical meristem of hydrilla. The feeding damage caused by the larvae prevents the hydrilla canopy from reaching the water surface, a desirable effect that ameliorates the economic and ecological damage associated with the dense surface mats of hydrilla.

Several aspects of C. lebetis have been investigated by researchers to determine the potential of the insect as a biological control agent of hydrilla. Epler et al. (2000) re- described the insect. Cuda et al. (2002) described the biology of the insect and developed a method for rearing the insect in laboratory conditions. Cuda et al. (2011) also demonstrated that the larval stage of C. lebetis can suppress the growth of hydrilla.

Furthermore, Stratman et al. (2013ab) reported the insect is already widespread in

Florida waterbodies, and determined the fundamental (laboratory) host range of the insect. Finally, Stratman et al. (2014) studied the thermal tolerances of the insect and predicted that much of the southeastern U.S. is climatically suitable for development of

127

insect. Based on these past studies, there is a consensus the insect may have value as an augmentative biological agent for hydrilla in Florida. However, further research is required to better determine the efficacy and suitability of C. lebetis as a biological control agent, (Cuda et al. 2011, Stratman et al. 2013b). Thus, the present study addresses some of the more critical issues regarding the suitability of the midge as a biological control agent, e.g., the field host range of C. lebetis, its impact on the competitiveness of hydrilla to native aquatic vegetation, and the foraging depth of the insect.

Generally, the screening and approval processes of a classical biological control agent require decades to complete and input from many researchers. Information generated in the present study contributes to this extensive process, especially for other countries that may want to import C. lebetis for biological control of hydrilla. Specifically, our results demonstrated that C. lebetis favored hydrilla as a host plant under field conditions in Florida, revealed the vertical foraging range of C. lebetis ranged from a depth of 0 m to at least 2.7 m, and provided empirical evidence that demonstrated an attack on hydrilla by C. lebetis could shift the competitive balance between plant species in favor of native plants. Overall, the present study demonstrated that C. lebetis may have value as an augmentative biological control agent for hydrilla in Florida and other southeastern states impacted by hydrilla.

Based on the findings of the present study, we recommend that future research focus on enhancing our understanding of C. lebetis by extending the field host range studies of C. lebetis to waterbodies with different biogeographical properties; examining the impact of C. lebetis on the growth pattern of hydrilla under field conditions and on

128

the germination of hydrilla tubers or turions recovering from adverse conditions such as drawdowns or herbicide treatments; and exploring the impact of biotic and abiotic factors known to influence the role of insect herbivores, such as C. lebetis, in mediating the interspecific competitions between plants. Specifically, factors reported to have additive, antagonistic, or synergistic effects on insect herbivory include host plant quality, eutrophication levels, and predators such as fish.

129

LIST OF REFERENCES

Alam SK, Ager LA, Rosegger TM, Lange TR. 1996. The effects of mechanical harvesting of floating plant tussock communities on water quality in Lake Istokpoga, Florida. Lake Reserv. Manage. 12:455–461.

Arias RS, Netherland MD, Scheffler BE, Puri A, Dayan FE. 2005. Molecular evolution of herbicide resistance to phytoene desaturase inhibitors in Hydrilla verticillata and its potential use to generate herbicide-resistant crops. Pest Manag. Sci. 61:258– 268.

Bachmann RW, Hoyer MV, Canfield DE, Jr. 2000. The potential for wave disturbance in shallow Florida lakes. Lake Reserv. Manage. 16:281–291.

Bajsa J, Pan Z, Dayan FE, Owens DK, Duke SO. 2012. Validation of serine/threonine protein phosphatase as the herbicide target site of endothall. Pestic. Biochem. Physiol. 102:38–44.

Balciunas JK, Minno MC. 1985. Insects damaging hydrilla in the USA. J. Aquat. Plant Manage. 23:77–83.

Balciunas JK, Purcell MF. 1991. Distribution and biology of a new Bagous weevil (Coleoptera: Curculionidae) which feeds on the aquatic weed Hydrilla verticillata. J. Aust. Entomol. Soc. 30:333–338.

Balciunas JK, Burrows DW, Purcell MF. 1996. Comparison of the physiological and realized host-ranges of a biological control agent from Australia for the control of the aquatic weed, Hydrilla verticillata. Biol. Control. 7:148–158.

Balciunas JK, Grodowitz MJ, Cofrancesco AF, Shearer JF. 2002. Hydrilla, pp. 91–114. In: R. Van Dreische, S. Lyon, B. Blossey, M. Hoddle, and R. Reardon (eds.). Biological control of invasive plants in the eastern United States. Publication FHTET-2002-04, USDA Forest Service, Morgantown, WV.

Baloch GM, Sana-Ullah. 1974. Insects and other organisms associated with Hydrilla verticillata (L.f.) L.C. (Hydrocharitaceae) in Pakistan, pp. 61–66. In: A. J. Wapshere (ed.). Proceedings of the VII International Symposium on Biological Control of Weeds. Commonw. Inst. Biol. Control, Misc. Publ. Montpelier, Fr.

Baniszewski J, Weeks ENI, Cuda JP. 2015. Impact of refrigeration on eggs of the hydrilla tip mining midge Cricotopus lebetis (Diptera: Chironomidae): Larval hatch rate and subsequent development. J. Aquat. Plant Manage. 53:209–215.

Baniszewski J, Cuda JP, Gezan SA, Sharma S, Weeks ENI. 2016. Stem fragment regrowth of Hydrilla verticillata following desiccation. J. Aquat. Plant Manage. 54:53–60.

130

Barton-Browne L. 1977. Host-related responses and their suppression: Some behavioral considerations, pp.117–127. In: H. H. Shorey and J. J. McKelvey (eds.). Chemical control of insect behavior: Theory and application. Wiley, New York, NY.

Bennett CA, Buckingham GR. 1991. Laboratory biologies of Bagous affinis and B. laevigatus (Coleoptera: Curculionidae) attacking tubers of Hydrilla verticillata (Hydrocharitaceae). Ann. Entomol. Soc. Am. 84:420–428.

Berg CO. 1950. Biology of certain Chironomidae reared from Potamogeton. Ecol. Monogr. 20:83–101.

Berger S, MacDonald G. 2011. Suspected endothall tolerant hydrilla in Florida. Proc. Southern Weed Sci. Soc. 64:331. [Abstract]

Bernays EA, Chapman RF. 1994. Host plant selection by phytophagous insects (contemporary topics in entomology). Springer, New York, NY. 312 pp.

Bidigare RR, Christensen SJ, Wilde SB, Banack SA. 2009. Cyanobacteria and BMAA: Possible linkage with avian vacuolar myelinopathy (AVM) in the southeastern United States. Amyotroph. Lateral Scler. 10:71–73.

Blossey B, Schroeder D, Hight SD, Malecki RA. 1994. Host specificity and environmental impact of two leaf beetles (Galerucella calmariensis and G. pusilla) for biological control of purple loosestrife (Lythrum salicaria). Weed Sci. 42:134–140.

Bogut I, Čerba D, Vidaković J, Gvozdić V. 2010. Interactions of weed-bed invertebrates and Ceratophyllum demersum stands in a floodplain lake. Biologia. 65:113–121.

Bowes G, Salvucci ME. 1989. Plasticity in the photosynthetic carbon metabolism of submersed aquatic macrophytes. Aquat. Bot. 34:233–266.

Bowes G, Van TK, Garrard A, Hailer WT. 1977. Adaptation to low light levels by hydrilla. J. Aquat. Plant Manage. 15:32–35.

Bowes G, Holaday AS, Haller WT. 1979. Seasonal variation in the biomass, tuber density, and photosynthetic metabolism of hydrilla in three Florida lakes. J. Aquat. Plant Manage. 17:61–65.

Brackenbury J. 2000. Locomotory modes in the larva and pupa of Chironomus plumosus (Diptera, Chironomidae). J. Insect Physiol. 46:1517–1527.

Brackenbury J. 2003. Swimming kinematics and wake elements in a worm-like insect: The larva of the midge Chironomus plumosus (Diptera). J. Zool. 260:195–201.

131

Bradshaw EL, Allen MS, Netherland M. 2014. Spatial and temporal occurrence of hypoxia influences fish habitat quality in dense Hydrilla verticillata. J. Freshw. Ecol. 30:491–502.

Brönmark C, Weisner S. 1992. Indirect effects of fish community structure on submerged vegetation in shallow eutrophic lakes: An alternative mechanism. Hydrobiologia. 243:293–301.

Brown SJ, Maceina MJ. 2002. The influence of disparate levels of submersed aquatic vegetation on largemouth bass population characteristics in a Georgia reservoir. J. Aquat. Plant Manage. 40:28–35.

Buckingham GR. 1988. Reunion in Florida - hydrilla, a weevil, and a fly. Aquatics. 10:19–25.

Buckingham GR. 1994. Biological control of aquatic weeds, pp. 413–480. In: D. Rosen, F. D. Bennett, and J. L. Capinera (eds.). Pest management in the subtropics: Biological control—A Florida Perspective. Intercept, Andover, UK.

Buckingham GR. 1996. Biological control of alligatorweed, Alternanthera philoxeroides, the world's first aquatic weed success story. Castanea. 61:232–243.

Buckingham GR, Bennett CA. 1998. Host range studies with Bagous affinis (Coleoptera: Curculionidae), an Indian weevil that feeds on hydrilla tubers. Environ. Entomol. 27:469–478.

Buckingham GR, Okrah EA, Christian-Meier M. 1991. Laboratory biology and host range of Hydrellia balciunasi [Diptera: Ephydridae]. Entomophaga. 36:575–586.

Burleson ML, Wilhelm DR, Smatresk NJ. 2001. The influence of fish size on the avoidance of hypoxia and oxygen selection by largemouth bass. J. Fish Biol. 59:1336–1349.

Carter V, Rybicki N, Jr. 1986. Resurgence of submersed aquatic macrophytes in the tidal Potomac River, Maryland, Virginia, and the District of Columbia. Estuaries. 9:368–375.

Catton HA, Lalonde RG, Clerck-Floate D, Rosemarie A. 2014. Differential host-finding abilities by a weed biocontrol insect create within-patch spatial refuges for nontarget plants. Environ. Entomol. 43:1333–1344.

Center TD, Grodowitz MJ, Cofrancesco AF, Jubinsky G, Snoddy E, Freedman JE. 1997. Establishment of Hydrellia pakistanae (Diptera: Ephydridae) for the biological control of the submersed aquatic plant Hydrilla verticillata (Hydrocharitaceae) in the southeastern United States. Biol. Control. 8:65–73.

132

Center TD, Van TK, Dray FA, Jr., Franks SJ, Rebelo MT, Pratt PD, Rayamajhi MB. 2005. Herbivory alters competitive interactions between two invasive aquatic plants. Biol. Control. 33:173–185.

Center TD, Parys K, Grodowitz M, Wheeler GS, Dray FA, O'Brien CW, Johnson S, Cofrancesco A. 2013. Evidence of establishment of Bagous hydrillae (Coleoptera: Curculionidae), a biological control agent of Hydrilla verticillata (Hydrocharitales: Hydrocharitaceae) in North America? Fla. Entomol. 96:180– 186.

Cerba D, Mihaljevic Z, Vidakovic J. 2010. Colonisation of temporary macrophyte substratum by midges (Chironomidae: Diptera). Ann. Limnol. Int. J. Lim. 46:181– 190.

Cerba D, Mihaljevic Z, Vidakovic J. 2011. Colonisation trends, community and trophic structure of chironomid larvae (Chironomidae: Diptera) in a temporal phytophylous assemblage. Fund. Appl. Limnol. 179:203–214.

Chambers PA, Barko JW, Smith CS. 1993. Workshop summaries evaluation of invasions and declines of submersed aquatic macrophytes. J. Aquat. Plant Manage. 31:218–220.

Clement SL, Cristofaro M. 1995. Open-field tests in host specificity determination of insects for biological control of weeds. Biocontrol Sci. Techn. 5:395–406.

Colle DE, Shireman JV. 1980. Coefficients of condition for largemouth bass, bluegill, and redear sunfish in hydrilla-infested lakes. Trans. Am. Fish. Soc. 109:521–531.

Colon-Gaud JC, Kelso WE, Rutherford DA. 2004. Spatial distribution of macroinvertebrates inhabiting hydrilla and coontail beds in the Atchafalaya Basin, Louisiana. J. Aquat. Plant Manage. 42:85–91.

Cook CDK, Lüönd R. 1982. A revision of the genus Hydrilla (Hydrocharitaceae). Aquat. Bot. 13:485–504.

Coon BR, Harms NE, Cuda JP, Grodowitz MJ. 2014. Laboratory biology and field population dynamics of Trichopria columbiana (Hymenoptera: Diapriidae), an acquired parasitoid of two hydrilla biological control agents. Biocontrol Sci. Techn. 24:1243–1264.

Copeland RS, Nkubaye E, Nzigidahera B, Cuda JP, Overholt WA. 2011. The African burrowing mayfly, Povilla adusta (Ephemeroptera: Polymitarcyidae), damages Hydrilla verticillata (Alismatales: Hydrocharitaceae) in Lake Tanganyika. Fla. Entomol. 3:669–676.

Copeland RS, Nkubaye E, Nzigidahera B, Epler JH, Cuda JP, Overholt WA. 2012a. The diversity of Chironomidae (Diptera) associated with Hydrilla verticillata

133

(Alismatales: Hydrocharitaceae) and other aquatic macrophytes in Lake Tanganyika, Burundi. Ann. Entomol. Soc. Am. 105:206–224.

Copeland RS, Gidudu B, Wanda F, Epler JH, Cuda JP, Overholt WA. 2012b. Chironomidae (Insecta: Diptera) collected from Hydrilla verticillata (Hydrocharitaceae) and other submersed aquatic macrophytes in Lake Bisina and other Ugandan Lakes. J. East Afr. Nat. Hist. 101:29–66.

Crawley MJ. 1989. Insect herbivores and plant population dynamics. Annu. Rev. Entomol. 34:531–564.

Cuda JP. 2017. Incorporating biocontrol agents into an integrated management plan: practical considerations. Research Methods in Aquatic Plant Management. J. Aquat. Plant Manage. (accepted 8/2/17).

Cuda JP, Coon BR, Dao YM, Center TD. 2002. Biology and laboratory rearing of Cricotopus lebetis (Diptera: Chironomidae), a natural enemy of the aquatic weed hydrilla (Hydrocharitaceae). Ann. Entomol. Soc. Am. 95:587–596.

Cuda JP, Charudattan R, Grodowitz MJ, Newman RM, Shearer JF, Tamayo ML, Villegas B. 2008. Recent advances in biological control of submersed aquatic weeds. J. Aquat. Plant Manage. 46:15–32.

Cuda JP, Coon BR, Dao YM, Center TD. 2011. Effect of an herbivorous stem-mining midge on the growth of hydrilla. J. Aquat. Plant Manage. 49:83–89.

Cuda JP, Shearer JF, Weeks EN, Kariuki E, Baniszewski J, Giurcanu M. 2016. Compatibility of an insect, a fungus, and a herbicide for integrated pest management of dioecious hydrilla. J. Aquat. Plant Manage. 54:20–25.

Cullen JM. 1990. Current problems in host-specificity screening, pp. 27–36. In: E. S. Delfosse (ed.). Proceedings of the VII International Symposium on Biological Control of Weeds. Istituto Sperimentale per la Patologia Vegetale, Rome, Italy.

Cullen R, Warner KD, Jonsson M, Wratten SD. 2008. Economics and adoption of conservation biological control. Biol. Control. 31:272–280.

Culliney TW. 2005. Benefits of classical biological control for managing invasive plants. CRC Crit. Rev. Plant Sci. 24:131–150.

Dayan FE, Netherland MD. 2005. Hydrilla, the perfect aquatic weed, becomes more noxious than ever. Outlooks on Pest Manage. 16:277–282.

Dextrase AJ, Mandrake NE. 2006. Impacts of alien invasive species on freshwater fauna at risk in Canada. Biol. Invasions. 8:13–24.

Dibble ED, Killgore KJ, Dick GO. 1996. Measurement of plant architecture in seven aquatic plants. J. Freshw. Ecol. 11:311–318.

134

Doyle RD, Smart RM. 2001. Effects of drawdowns and dessication on tubers of hydrilla, an exotic aquatic weed. Weed Sci. 49:135–140.

Doyle RD, Grodowitz MJ, Smart RM, Owens C. 2002. Impact of herbivory by Hydrellia pakistanae (Diptera: Ephydridae) on growth and photosynthetic potential of Hydrilla verticillata. Biol. Control. 24:221–229.

Doyle R, Grodowitz M, Smart M, Owens C. 2007. Separate and interactive effects of competition and herbivory on the growth, expansion, and tuber formation of Hydrilla verticillata. Biol. Control. 41:327–338.

Dunn PH. 1978. Shortcomings in the Classic Tests of Candidate Insects for the Biocontrol of Weeds, pp. 51–56. In: T. E. Freeman (ed.). Proceedings of the IV International Symposium on Biological Control of Weeds. University of Florida, Gainesville, FL.

Ekrem T, Stur E, Hebert PD. 2010. Females do count: Documenting Chironomidae (Diptera) species diversity using DNA barcoding. Org. Divers. Evol. 10:397–408.

Elton CS. 1958. The ecology of invasions by animals and plants. Methuen, London, UK. 181 pp.

Epler JH. 2001. Identification manual for the larval Chironomidae (Diptera) of North and South Carolina. A guide to the taxonomy of the midges of the southeastern United States, including Florida. Special Publication SJ2001-SP13. North Carolina Department of Environment and Natural Resources, Raleigh, NC, and St. Johns River Water Management District, Palatka, FL. 526 pp.

Epler JH, Cuda JP, Center TD. 2000. Redescription of Cricotopus lebetis (Diptera: Chironomidae), a potential biocontrol agent of the aquatic weed hydrilla (Hydrocharitaceae). Fla. Entomol. 83:171–180.

Evans JM, Wilkie AC. 2010. Life cycle assessment of nutrient remediation and bioenergy production potential from the harvest of hydrilla (Hydrilla verticillata). J. Environ. Manage. 91:2626–2631.

Evans JM, Wilkie AC, Burkhardt J, Haynes RP. 2007. Rethinking exotic plants: Using citizen observations in a restoration proposal for Kings Bay, Florida. Ecol. Restoration. 25:199–210.

[FDACS] Florida Department of Agriculture and Consumer Services. 2017. State of Florida prohibited aquatic plants list (effective as of June 30, 2008). Available at https://www.flrules.org/gateway/RuleNo.asp?id=5B-64.011. Accessed November 10, 2017.

Fenemore PG. 1988. Host-plant location and selection by adult potato moth, Phthorimaea operculella (Lepidoptera: Gelechiidae): A review. J. Insect Physiol. 34:175–177.

135

Ferrington LC, Jr. 2008. Global diversity of non-biting midges (Chironomidae; Insecta- Diptera) in freshwater. Hydrobiologia. 595:447–455.

Fox AM, Haller WT, Shilling DG. 1994. Use of fluridone for hydrilla management in the Withlacoochee River, Florida. J. Aquat. Plant Manage. 32:47–55.

Frodge JD, Thamas GL, Pauley GB. 1990. Effects of canopy formation by floating and submergent aquatic macrophytes on the water quality of two shallow Pacific Northwest lakes. Aquat. Bot. 38:231–248.

[FWC] Florida Fish and Wildlife Conservation Commission. 2016. Annual report of activities conducted under the cooperative aquatic plant control program in Florida public waters for fiscal year 2015–2016. Florida Fish and Wildlife Conservation Commission Invasive Plant Management Section. Available at http://myfwc.com/media/4137470/annualreport15-16.pdf. Accessed September 29, 2017.

Getsinger KD, Netherland MD, Grue C, Koschnick TJ. 2008. Recent improvements in the use of aquatic herbicides and establishment of future research directions. J. Aquat. Plant Manage. 46:32–41.

Glomski LM, Netherland MD. 2012. Does hydrilla grow an inch per day? Measuring short-term changes in shoot length to describe invasive potential. J. Aquat. Plant Manage. 50:54–57.

Godfrey KE, Anderson LWJ, Perry SD, Dechoretz N. 1994. Overwintering and establishment potential of Bagous affinis (Coleoptera: Curculionidae) on Hydrilla verticillata (Hydrocharitaceae) in northern California. Fla. Entomol. 77:221–230.

Goeden RD, Andres LA. 1999. Biological control of weeds in terrestrial and aquatic environment, pp. 871–890. In: T. S. Bellows, T. W. Fisher, L. E. Caltagirone, D. L. Dahlsten, C. Huffaker, and G. Gordh (eds.). Handbook of biological control: Principles and applications of biological control. Academic Press, San Diego, CA.

Gopal B, Goel U. 1993. Competition and allelopathy in aquatic plant communities. Bot. Rev. 59:155–210.

Gowik U, Westhoff P. 2011. The path from C3 to C4 photosynthesis. Plant Physiol. 155:56–63.

Gregory SV. 1983. Plant-Herbivore Interactions in stream systems, pp. 157–189. In: J. R. Barnes and G. W. Minshall (eds.). Stream ecology. Application and testing of general ecological theory. Plenum Press, NY.

Gresens S, Stur E, Ekrem T. 2012. Phenotypic and genetic variation within the Cricotopus sylvestris species-group (Diptera, Chironomidae), across a Nearctic- Palaearctic gradient. Fauna Norv. 31:137–149.

136

Grodowitz MJ, Center TD, Cofrancesco AF, Freedman JE. 1997. Release and establishment of Hydrellia balciunasi (Diptera: Ephydridae) for the biological control of the submersed aquatic plant Hydrilla verticillata (Hydrocharitaceae) in the United States. Biol. Control. 9:15–23.

Grodowitz MJ, Doyle RD, Smart RM. 2000. Potential use of insect biocontrol agents for reducing the competitive ability of Hydrilla verticillata. ERCD/EL SR-00-1. U.S. Army Engineer Research and Development Center, Vicksburg, MS.

Grodowitz MJ, Owens CS, Smart RM, Nachtrieb JG. 2007. Impact of herbivory and plant competition on the growth of hydrilla in small ponds. ERDC/TN-APCRP-BC- 08. U.S. Army Engineer Research and Development Center, Vicksburg, MS.

Gunner HB. 1983. Biological control technology development. Microbial control of Eurasian watermilfoil, pp. 86–100. In: Proceedings, 17th Annual Meeting, Aquatic Plant Control Research Program. Miscellaneous Paper A-83-3. U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS.

Haller WT, Sutton DL. 1975. Community structure and competition between Hydrilla and Vallisneria. Hyacinth Contr. J. 13:48–50.

Haller WT, Sutton DL, Barlowe WC. 1974. Effects of salinity on growth of several aquatic macrophytes. Ecology. 55:891–894.

Haller WT, Miller JL, Garrard LA. 1976. Seasonal production and germination of hydrilla vegetative propagules. J. Aquat. Plant Manage. 14:26–29.

Haller WT, Shireman JV, Durant DF. 1980. Fish harvest resulting from mechanical control of hydrilla. Trans. Am. Fish. Soc. 109:517–520.

Haller WT, Fox AM, Shilling DG. 1990. Hydrilla control program in the upper St. Johns River, Florida, USA. Proc. European Weed Soc. Symp. on Aquat. Weeds. 8:111– 116.

Harley KLS, Kassulke RC, Sands DPA, Day MD. 1990. Biological control of water lettuce, Pistia stratiotes [Araceae] by Neohydronomus affinis [Coleoptera: Curculionidae]. Entomophaga. 35:363–374.

Hay ME, Steinberg PD. 1992. The chemical ecology of plant–herbivore interactions in marine vs. terrestrial communities, pp. 372–408. In: G. A. Rosenthal and M. R. Berenbaum (eds.). Herbivores: Their interactions with secondary plant metabolites. Volume II. Ecological and evolutionary processes. Academic Press, San Diego, CA.

Heard TA. 2000. Concepts in insect host-plant selection behaviour and their application to host specificity testing, pp. 1–10. In: R. G. Van Driesche, T. A. Heard, A. McClay, and R. Reardon (eds.). Proceedings of session: Host-Specificity Testing of Exotic Arthropod Biological Control Agents—The Biological Basis for

137

Improvement in Safety. USDA Forest Service, Forest Health Technology Enterprise Team, Morgantown, WV.

Henderson JE. 1992. Economics and aquatic plant control, pp. 6–17. In: Proceedings, 26th Annual Meeting, Aquatic Plant Control Research Program. Miscellaneous Paper A-92–2. U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS.

Hershner C, Havens KJ. 2008. Managing invasive aquatic plants in a changing system: Strategic consideration of ecosystem services. Conserv. Biol. 22:544–550.

Hill MP. 2003. The impact and control of alien aquatic vegetation in South African aquatic ecosystems. Afr. J. Aquat. Sci. 28:19–24.

Hofstra DE, Clayton J, Green JD, Adam KD. 2000. RAPD profiling and isozymes analysis of New Zealand Hydrilla verticillata. Aquat. Bot. 66:153–166.

Holaday AS, Bowes G. 1980. C4 acid metabolism and dark CO2 fixation in a submersed aquatic macrophyte (Hydrilla verticillata). Plant Physiol. 65:331–335.

Howarth FG. 1991. Environmental impacts of classical biological control. Ann. Rev. Entomol. 36:485–509.

Hulme PE. 1996. Herbivory, plant regeneration and species co-exstence. J. Ecol. 84:609–616.

Johnson JA, Newman RM. 2011. A comparison of two methods for sampling biomass of aquatic plants. J. Aquat. Plant Manage. 49:1–8.

Joyce JC, Langeland KA, Van TK, Vandiver VV. 1992. Organic sedimentation associated with hydrilla management. J. Aquat. Plant Manage. 30:20–23.

Julien M, White G. 1997. Biological control of weeds: Theory and practical application. ACIAR Monograph No. 49. ACIAR, Canberra, Australia.190 pp.

Kangasniemi BJ, Oliver DR. 1983. Chironomidae (Diptera) associated with Myriophyllum spicatum in Okanagan Valley lakes, British Columbia. Can. Entomol. 115:1545–1546.

Keane RM, Crawley MJ. 2002. Exotic plant invasions and the enemy release hypothesis. Trends Ecol. Evol. 17:164–170.

Kelly DJ, Hawes I. 2005. Effects of invasive macrophytes on littoral zone productivity and food web dynamics in a New Zealand high-country lake. J. North Am. Benthol. Soc. 24:300–320.

138

Kirk JP, Henderson JE. 2006. Management of hydrilla in the Santee Cooper Reservoirs, South Carolina: Experiences from 1982 to 2004. J. Aquat. Plant Manage. 44:98– 103.

Kirkagac M, Demir N. 2004. The effects of grass carp on aquatic plants, plankton and benthos in ponds. J. Aquat. Plant Manage. 42:32–39.

Langeland KA. 1996. Hydrilla verticillata (L.f.) Royle (Hydrocharitaceae), "The perfect aquatic weed". Castanea. 61:293–304.

Lerner A, Meltser N, Sapir N, Erlick C, Shashar N, Broza M. 2008. Reflected polarization guides chironomid females to oviposition sites. J. Exp. Biol. 211:3536–3543.

Leslie AJ, Dyke JMV, Hestand RS, Thompson BZ. 1987. Management of aquatic plants in multi-use lakes with grass carp (Ctenopharyngodon idella). Lake Reserv. Manage. 3:266–276.

Linevich AA. 1971. The Chironomidae of Lake Baikal. Limnologica (Berlin). 8:51–52.

Lodge DM. 1991. Herbivory on freshwater macrophytes. Aquat. Bot. 41:195–224.

Long ZT, Mohler CL, Carson WP. 2003. Extending the resource concentration hypothesis to plant communities: Effects of litter and herbivores. Ecology. 3:652– 665.

Louda SM, Keeler KH, Holt RD. 1990. Herbivore influences on plant performance and competitive interactions, pp. 413–444. In: J. Grace and D. Tilman (eds.). Perspectives on plant competition. Academic Press, San Diego, CA.

MacRae IV, Winchester NN, Ring RA. 1990. Feeding activity and host preference of the milfoil midge, Cricotopus myriophylli Oliver (Diptera: Chironomidae). J. Aquat. Plant Manage. 28:89–92.

Madeira PT, Van TK, Center TD. 1999. Integration of five Southeast Asian accessions into the world-wide phenetic relationships of Hydrilla verticillata as elucidated by random amplified polymorphic DNA analysis. Aquat. Bot. 63:161–167.

Madeira PT, Jacono CC, Van TK. 2000. Monitoring hydrilla using two RAPD procedures and the nonindigenous aquatic species database. J. Aquat. Plant Manage. 38:33–40.

Madeira PT, Van TK, Center TD. 2004. An improved molecular tool for distinguishing monoecious and dioecious hydrilla. J. Aquat. Plant Manage. 42:28–32.

Madeira PT, Coetzee JA, Center TD, White EE, Tipping PW. 2007. The origin of Hydrilla verticillata recently discovered at a South African dam. Aquat. Bot. 87:176–180.

139

Madsen JD. 1993. Biomass techniques for monitoring and assessing control of aquatic vegetation. Lake Reserv. Manage. 7:141–154.

Madsen TV, Sand-Jensen K. 1991. Photosynthetic carbon assimilation in aquatic macrophytes. Aquat. Bot. 41:5–40.

Marko M, Newman R, Gleason F. 2005. Chemically mediated host-plant selection by the milfoil weevil: A freshwater insect–plant interaction. J. Chem. Ecol. 31:2857– 2876.

Mayhew PJ. 1997. Adaptive patterns of host-plant selection by phytophagous insects. Oikos. 79:417–428.

McFadyen RE. 1998. Biological control of weeds. Ann. Rev. Entomol. 43:369–393.

McFadyen RE. 2000. Successes in the biological control of weeds, pp. 3–14. In: N. R. Spencer (ed.). Proceedings of the X International Symposium on Biological Control of Weeds. Montana State University, Bozeman, MT.

McKnight SK, Hepp GR. 1995. Potential effects of grass carp herbivory on waterfowl foods. J. Wildl. Manage. 59:720–727.

Michel A, Arias RS, Scheffler BE, Duke SO, Netherland M, Dayan FE. 2004. Somatic mutation-mediated evolution of herbicide resistance in the nonindigenous invasive plant hydrilla (Hydrilla verticillata). Mol. Ecol. 13:3229–3237.

Miranda LE, Hodges KB. 2000. Role of aquatic vegetation coverage on hypoxia and sunfish abundance in bays of a eutrophic reservoir. Hydrobiologia. 427:51–57.

Mitchell A, Berro A, Cuda JP, Weeks ENI. 2017. Impact of food deprivation on hydrilla tip mining midge survival and subsequent development. Fla. Entomol. (accepted 9/21/17).

Mony C, Koschnick TJ, Haller WT, Muller S. 2007. Competition between two invasive Hydrocharitaceae (Hydrilla verticillata (Lf)(Royle) and Egeria densa (Planch)) as influenced by sediment fertility and season. Aquat. Botany. 86:236–242.

Moxley D, Langford F. 1982. Beneficial effects of hydrilla on two eutrophic lakes in central Florida. Proc. Annu. Conf. Southeast. Assoc. Fish and Wildl. Agencies. 36:280–286.

Netherland MD. 1997. Turion ecology of hydrilla. J. Aquat. Plant Manage. 35:1–10.

Netherland MD. 2014. Chemical control of aquatic weeds, pp. 71–88. In: L. A. Gettys, W. T. Haller, and D. G. Petty (eds.). Biology and control of aquatic plants: A best management practices handbook, 3rd ed. Aquatic Ecosystem Restoration Foundation, Marietta, GA.

140

Netherland MD, Jones D. 2015. Fluridone-resistant hydrilla (Hydrilla verticillata) is still dominant in the Kissimmee chain of lakes, FL. Invasive Plant Sci. Manage. 8:212–218.

Netherland MD, Getsinger KD, Turner EG. 1993. Fluridone concentration and exposure time requirements for control of hydrilla and Eurasian watermilfoil. J. Aquat. Plant Manage. 32:189–194.

Newman RM. 1991. Herbivory and detritivory on freshwater macrophytes by invertebrates: A review. J. North Am. Benth. Soc. 10:89–114.

Newroth PR. 1985. A review of Eurasian water milfoil impacts and management in British Columbia, pp. 139–153. In: L. W. J. Anderson (ed.). Proceedings of the First International Symposium on the Watermilfoil (Myriophyllum spicatum) and Related Haloragaceae Species. Aquat. Plant Manage. Soc., Washington, DC.

Nico LG, Muench AM. 2004. Nests and nest habitats of the invasive catfish Hoplosternum littorale in lake Tohopekaliga, Florida: A novel association with non-native Hydrilla verticillata. Southeast. Nat. 3:451–466.

Oliver DR. 1971. Life history of the Chironomidae. Ann. Rev. Entomol. 16:211–230.

Overholt WA, Cuda JP. 2005. University of Florida scientists explore Africa for natural enemies of hydrilla. Aquatics. 27:18–20.

Owens CS, Madsen JD, Smart RM, Stewart RM. 2001. Dispersal of native and nonnative aquatic plant species in the San Marcos River, Texas. J. Aquat. Plant Manage. 39:75–79.

Parker JD, Hay ME. 2005. Biotic resistance to plant invasions? Native herbivores prefer non-native plants. Ecol. Lett. 8:959–967.

Pemberton RW. 2000. Predictable risk to native plants in weed biological control. Oecologia. 125:489–494.

Pimentel D, Zuniga R, Morrison D. 2005. Update on the environmental and economic costs associated with alien-invasive species in the United States. Ecol. Econ. 52:273–288.

Pinder LCV. 1986. Biology of freshwater Chironomidae. Annu. Rev. Entomol. 31:1–23

Pipalova I. 2006. A review of grass carp use for aquatic weed control and its impact on water bodies. J. Aquat. Plant Manage. 44:1–12.

Posey MH, Wigland C, Stevenson JC. 1993. Effects of an introduced aquatic plant, Hydrilla verticillata, on benthic communities in the upper Chesapeake Bay. Est. Coast. Shelf Sci. 37:539–555.

141

Puri A, MacDonald GE, Haller WT, Singh M. 2007. Growth and reproductive physiology of fluridone-susceptible and resistant hydrilla (Hydrilla verticillata) biotypes. Weed Sci. 55:441–445.

Puri A, Haller WT, Netherland MD. 2009. Cross-resistance in fluridone-resistant hydrilla to other bleaching herbicides. Weed Sci. 57:482–488.

Rao SK, Magnin NC, Reiskind JB, Bowes G. 2002. Photosynthetic and other phosphoenolpyruvate carboxylase isoforms in the single-cell, facultative C4 system of Hydrilla verticillata. Plant Physiol. 130:876–886.

Reeves JL, Lorch PD. 2011. Visual active space of the milfoil weevil, Euhrychiopsis lecontei Dietz (Coleoptera: Curculionidae). J. Insect Behav. 24:264–273.

Reeves J, Lorch P, Kershner M. 2009. Vision is important for plant location by the phytophagous aquatic specialist Euhrychiopsis lecontei Dietz (Coleoptera: Curculionidae). J. Insect Behav. 22:54–64.

Reiskind JB, Madsen TV, Ginkel LV, Bowes G. 1997. Evidence that inducible C4-type photosynthesis is a chloroplastic CO2-concentrating mechanism in Hydrilla, a submersed monocot. Plant Cell Environ. 20:211–220.

Richardson RJ. 2008. Aquatic plant management and the impact of emerging herbicide resistance issues. Weed Technol. 22:8–15.

Rita H, Komonen A. 2008. Odds ratio: An ecological sound tool to compare proportions. Ann. Zool. Fenn. 45:66–72.

Root RB. 1973. Organization of a plant-arthropod association in simple and diverse habitats: The fauna of collards (Brassica oleracea). Ecol. Monogr. 43:95–124.

Ruhl H, Rybicki N. 2010. Long-term reductions in anthropogenic nutrients link to improvements in Chesapeake Bay habitat. Proc. Natl. Acad. Sci. USA. 107:16566–16570.

Ryan FJ, Coley CR, Kay SH. 1995. Coexistence of monoecious and dioecious hydrilla in Lake Gaston, North Carolina and Virginia. J. Aquat. Plant Manage. 33:8–12.

Rybicki NB, Landwehr JM. 2007. Long-term changes in abundance and diversity of macrophyte and waterfowl populations in an estuary with exotic macrophytes and improving water quality. Limn. Ocean. 52:1195–1207.

Salvucci ME, Bowes G. 1983. Two photosynthetic mechanisms mediating the low photorespiratory state in submersed aquatic angiosperms. Plant Physiol. 73:488– 496.

142

Santamaría L. 2002. Why are most aquatic plants widely distributed? Dispersal, clonal growth and small-scale heterogeneity in a stressful environment. Acta Oecol. 23:137–154.

Schardt JD. 2010. Adapting to evolving hydrilla management ssues in Florida. Program and Abstracts, 50th Annual Meeting of the Aquatic Plant Management Society, Bonita Springs, FL. 53 pp.

Schardt JD, Nall LE. 1983. 1982 Aquatic flora of Florida survey report. Bureau of aquatic plant research and control. Florida Dept. of Natural Resources, Tallahassee, FL. 116 pp.

Schardt JD, Schmitz DC. 1991. 1990 Florida aquatic plant survey. Technical Report No. 91-CGA. Florida Department of Natural Resources, Tallahassee, FL. 89 pp.

Schmitz DC, Nelson BV, Nall LE, Schardt JD. 1991. Exotic aquatic plants in Florida: A historical perspective and review of the present aquatic plant regulation program, pp. 303–326. In: T. D. Center, R. F. Doren, R. L. Hofstetter, R. L. Myers, and L. D. Whiteaker (eds.). Proceedings of the Symposium on Exotic Pest Plants. Technical Report NPS/NREVER/NRTR-91/06. U.S. Department of the Interior, National Park Service, Denver, CO.

Schultz R, Dibble E. 2012. Effects of invasive macrophytes on freshwater fish and macroinvertebrate communities: The role of invasive plant traits. Hydrobiologia. 684:1–14.

Shearer JF. 1996. Field and laboratory studies of the fungus Mycoleptodiscus terrestris as a potential agent for management of the submersed aquatic macrophyte Hydrilla verticillata. Technical Report A-96-3. U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS.

Shearer JF, Grodowitz MJ. 2010. Effect of dried Mycoleptodiscus terrestris inoculum on fluridone-susceptible and fluridone-resistant hydrilla biotypes. APCRP Technical Notes Collection. ERDC/TN APCRP-BC-19. U.S. Army Engineer Research and Development Center, Vicksburg, MS.

Shireman JV, Maceina MJ. 1981. The utilization of grass carp, Ctenopharyngodon idella Val., for hydrilla control in Lake Baldwin, Florida. J. Fish Biol. 19:629–636.

Simberloff D, Von Holle B. 1999. Positive interactions of nonindigenous species: Invasional meltdown? Biol. Invasions. 1:21–32.

Smith L, Cristofaro M, de Lillo E, Monfreda R, Paolini A. 2009. Field assessment of host plant specificity and potential effectiveness of a prospective biological control agent, Aceria salsolae, of Russian thistle, Salsola tragus. Biol. Control 48:237– 243.

143

Sousa WT, Thomaz SM, Murphy KJ, Silveira MJ, Mormul RP. 2009. Environmental predictors of the occurrence of exotic Hydrilla verticillata (Lf) Royle and native Egeria najas Planch. in a sub-tropical river floodplain: The Upper River Paraná, Brazil. Hydrobiologia. 632:65–78.

Steward KK. 1991a. Growth of various hydrilla races in waters of differing pH. Fla. Sci. 54:117–125.

Steward KK. 1991b. Competitive interaction between monoecious hydrilla and American eelgrass on sediments of varying fertility. Fla. Sci. 54:135–147.

Steward KK, Van TK, Carter V, Pieterse AH. 1984. Hydrilla invades Washington, DC and the Potomac. Am. J. Bot. 71:162–163.

Stout MJ, Riggio MR, Zou L, Roberts R. 2002. Flooding influences ovipositional and feeding behavior of the rice water weevil (Coleoptera: Curculionidae). J. Econ. Entomol. 95:715–721.

Stratman K, Overholt WA, Cuda JP, Netherland MD, Wilson PC. 2013a. The diversity of Chironomidae associated with hydrilla in Florida, with special reference to Cricotopus lebetis (Diptera: Chironomidae). Fla. Entomol. 96:654–657.

Stratman KN, Overholt WA, Cuda JP, Netherland MD, Wilson PC. 2013b. Host range and searching behaviour of Cricotopus lebetis (Diptera: Chironomidae), a tip miner of Hydrilla verticillata (Hydrocharitaceae). Biocontrol Sci. Techn. 23:317– 334.

Stratman KN, Overholt WA, Cuda JP, Mukherjee A, Diaz R, Netherland MD, Wilson PC. 2014. Temperature-dependent development, cold tolerance, and potential distribution of Cricotopus lebetis (Diptera: Chironomidae), a tip miner of Hydrilla verticillata (Hydrocharitaceae). J. Insect Sci.14:153.

Strong DR, Lawton JH, Southwood RES. 1984. Insects on plants. Community patterns and mechanisms. Blackwell Scientific Publications, Oxford, UK. 313 pp.

Sublette JE. 1964. Chironomidae (Diptera) of Louisiana. I. Systematics and immature stages of some lentic chironomids of west-central Louisiana. Tulane Stud. Zool. 11:109–150.

Sutton DL. 1977. Grass carp Ctenopharyngodon idella Val. in North America. Aquat. Bot. 3:157–164.

Sutton DL, Vandiver VV, Jr. 1986. Grass carp: A fish for biological management of hydrilla and other aquatic weeds in Florida. UF/IFAS Agric. Expt. Sta. Bull. No. 867. University of Florida, Gainesville, FL. 13 pp.

144

Taheruzzaman Q, Kushari DP. 1989. Evaluation of some common aquatic macrophytes cultivated in enriched water as possible source of protein and biogas. Hydrobiologia. 23:207–212. te Beest M, Le Roux JJ, Richardson DM, Brysting AK, Suda J, Kubešová M, Pyšek P. 2011. The more the better? The role of polyploidy in facilitating plant invasions. Ann. Bot. 109:19–45.

Theel HJ, Dibble ED. 2008. An experimental simulation of an exotic aquatic macrophyte invasion and its influence on foraging behavior of bluegill. J. Freshw. Ecol. 23:79–89.

Tindall KV, Bernhardt JL, Stout MJ, Beighley DH. 2013. Effect of depth of flooding on the rice water weevil, Lissorhoptrus oryzophilus, and yield of rice. J. Insect Sci. 13:62.

Titus JE, Adams MS. 1979. Coexistence and the comparative light relations of the submersed macrophytes Myriophyllum spicatum L. and Vallisneria americana Michx. Oecologia. 40:273–286.

U.S. Congress, Office of Technology Assessment. 1993. Harmful non-indigenous species in the United States. OTA-F-565. Congress Government Printing Office. Washington DC, U.S. 391 pp.

[USDA-APHIS] United States Department of Agriculture Animal and Plant Health Inspection Service. 2017. Federal Noxious Weed List (effective as of December 10, 2010). Available at https://www.aphis.usda.gov/plant_health/plant_pest_info/weeds/downloads/weed list.pdf. Accessed June 8, 2017.

Valley RD, Bremigan MT. 2002. Effects of macrophyte bed architecture on largemouth bass foraging: Implications of exotic macrophyte invasions. Am. Fish. Soc. 131:234–244. van Dijk G. 1985. Vallisneria and its interactions with other species. Aquatics. 7:6–10.

Van TK, Steward KK. 1990. Longevity of monoecious Hydrilla propagules. J. Aquat. Plant Manage. 28:74–76.

Van TK, Haller WT, Bowes G. 1976. Comparison of the photosynthetic characteristics of three submersed aquatic plants. Plant Physiol. 58:761–768.

Van TK, Haller WT, Bowes G, Garrard LA. 1977. Effects of light quality on growth and chlorophyll composition in Hydrilla. J. Aquat. Plant Manage. 15:29–31.

Van TK, Wheeler GS, Center TD. 1998. Competitive interactions between hydrilla (Hydrilla verticillata) and vallisneria (Vallisneria americana) as influenced by insect herbivory. Biol. Control. 11:185–192.

145

Van TK, Wheeler GS, Center TD. 1999. Competition between Hydrilla verticillata and Vallisneria americana as influenced by soil fertility. Aquat. Bot. 62:225–233.

Vencill WK. (ed.). 2002. Herbicide handbook, 8th ed. Weed Sci. Soc. Am., Lawrence, KS. 493 pp.

Visser JH. 1988. Host-plant ﬁnding by insects: Orientation, sensory input and search patterns. J. Insect Physiol. 34:259–268.

Wan FH, Harris P. 1996. Host finding and recognition by Altica carduorum, a defoliator of Cirsium arvense. Entomol. Exp. Appl. 80:491–496.

Wapshere AJ. 1989. A testing sequence for reducing rejection of potential biological control agents for weeds. Ann. Appl. Biol. 114:515–526.

Warton DI, Hui FK. 2011. The arcsine is asinine: The analysis of proportions in ecology. Ecology. 92:3–10.

Webb MA, Ott RA, Bonds CC, Smart RM, Dick GO, Dodd L. 2012. Propagation and establishment of native aquatic plants in reservoirs. Management data series. No. 273. Texas Parks and Wildlife Department, Austin, TX. 53 pp.

Wheeler GS, Center TD. 2001. Impact of the biological control agent Hydrellia pakistanae (Diptera: Ephydridae) on the submersed aquatic weed Hydrilla verticillata (Hydrocharitaceae). Biol. Control. 21:168–181.

Wilde SB, Johansen JR, Wilde HD, Jiang P, Bartelme B. Haynie RS. 2014. Aetokthonos hydrillicola gen. et sp. nov.: Epiphytic cyanobacteria on invasive aquatic plants implicated in Avian Vacuolar Myelinopathy. Phytotaxa. 181:243–260.

Williams JR. 1954. The biological control of weeds, pp. 95–98. In: Report of the Sixth Commonwealth Entomological Congress, London, U.K.

Winston RL, Schwarzländer M, Hinz HL, Day MD, Cock MJW, Julien MH. 2014. Biological control of weeds: A world catalogue of agents and their target seeds, 5th ed. FHTET-2014–04. USDA Forest Service Publication, Morgantown, WV. 838 pp.

Wood K, O'Hare M, McDonald C, Searle K, Daunt F, Stillman R. 2016. Herbivore regulation of plant abundance in aquatic ecosystems. Biol. Rev. 92:1128–1141.

Yeo RR, Falk RH, Thurston JR. 1984. The morphology of hydrilla (Hydrilla verticillata (L.f.) Royle). J. Aquat. Plant Manage. 22:1–17.

Zwölfer H, Harris P. 1971. Host specificity determination of insects for biological control of weeds. Ann. Rev. Entomol. 16:159–178.

146

BIOGRAPHICAL SKETCH

Eutychus Mukure Kariuki was born in Kenya. He attended Egerton University,

Kenya, where he graduated in 2006 with a Bachelor of Science degree in agriculture

(first class honors). Eutychus then moved to the United States, where he attended

Florida A&M University (FAMU). He graduated in 2010 with a Master of Science degree

in agricultural sciences, majoring in entomology. His master’s research focused on the

evaluation of Gratiana boliviana as a biological control agent of tropical soda apple, an

invasive weed found in pastures in the United States. In 2011, Eutychus began working

toward a Doctor of Philosophy degree in entomology in the cooperative program

between the University of Florida and Florida A&M University. As a graduate student,

Eutychus participated in several professional conferences. In 2011, Eutychus was a

recipient of the Outstanding M.S. Student Display Presentation Award from

Entomological Society of America (ESA)-Southeastern Branch (SEB). In 2012,

Eutychus was a participant in the FAMU debate team, which was a recipient of an

Overall Runner-up award in the ESA national student debate competitions. Eutychus also served as the FAMU representative to the ESA-SEB, Student Affairs Committee.

Eutychus completed his Doctor of Philosophy degree in entomology and nematology in

2017. His goal is to pursue a career in public service.

147