BIOLOGICAL CONTROL of Imperata Cylindrica in WEST AFRICA USING FUNGAL PATHOGENS

Home , Bipolaris sacchari, Imperata, Imperata brevifolia, Imperata cylindrica, Kudzu, Saccharum arundinaceum

BIOLOGICAL CONTROL OF Imperata cylindrica IN WEST AFRICA USING FUNGAL PATHOGENS

ALANA DEN BREEYEN

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

2007

To my parents, Henk and Pathia Den Breeÿen

ACKNOWLEDGMENTS

I thank my major professor, Dr. Raghavan Charudattan, for his belief in me even at times

when it didn’t seem justified and for giving me the opportunities to grow professionally. I thank

Dr. Fen Beed, for helping me realize that if I’m not part of the solution then I am part of the problem and for his generous hospitality and friendship during my time at IITA, Cotonou, Benin.

I also thank my committee members, Dr Jeffrey Rollins, Dr. Greg MacDonald, and Dr. Fredy

Altpeter for their time, interest and support. My thanks go to the many people who have helped

me with my research: Jim DeValerio, for always being a willing ’driver’ and one of my first

friends in Gainesville, Mark Elliott, whose love and friendship has made life here well worth the

journey, to Gabriella Maia, for her love and friendship and willingness to sit in the midday sun

harvesting cogongrass rhizomes, and to Adolphe Avocanh who willing shared his office,

expertise and friendship during my time at IITA, Benin. My thanks go to Eldon Philman and

Herman Brown for the help they provided during the course of my experiments. Thanks go to

my former labmates, Jennifer Cook and Jonathan Horrell. My special thanks go to my lab mate,

Abby Guerra, who often put up with more than he should have and whose special friendship has

brightened my days. To the students and staff at IITA, Benin who were always willing to help

with my research and for reminding me to slow down, merci beaucoup! Thanks to Ms. Gail

Harris, Ms. Donna Perry, Ms. Lauretta Rahmes, Ms. Jan Sapp, Ms. Sherri Mizell and Ms. Dana

Lecuyer who helped me in their own special way. My thanks go to Mark and Maria Minno for

allowing me access to their cogongrass collection and Meghan Brenan for help with my

statistical data analysis. My special thanks to Sarah Clark Selke, Linley Smith, Tara Taranowski,

Heidi HansPetersen, and Amanda Watson, the amazingly talented women who have become my

friends.

My special thanks go to my Mom and Dad, who made the realization of the dream that much easier, and to my brother and sister-in-law, Ashley and Izelle for their love and support.

To all my friends, back home and around the world, thank you for your special blend of friendship.

TABLE OF CONTENTS

page

ACKNOWLEDGMENTS ...... 4

LIST OF TABLES...... 8

LIST OF FIGURES ...... 10

ABSTRACT...... 12

CHAPTER

1 INTRODUCTION ...... 14

Biology and Morphology...... 15 Cogongrass in the United States...... 16 Cogongrass in West Africa...... 18 Biological Control Strategies for Forest Weeds ...... 20 Biological Control Strategies for Cogongrass ...... 26 Potential Insect and Nematode Biocontrol Agents...... 27 Potential Plant Pathogenic Biocontrol Agents ...... 28 Mycoherbicides ...... 30 Genetic Variation in Invasive Weeds: The Implication for Biological Control...... 31 Biological Control and Genetic Markers...... 33

2 AN EXAMINATION OF GENETIC DIVERSITY OF Imperata cylindrica BASED ON INTER-SIMPLE SEQUENCE REPEATS AND SEQUENCE ANALYSIS OF INTERNAL TRANSCRIBED AND CHLOROPLAST TRNL-F SPACER DNA ...... 36

Introduction...... 36 Materials and Methods ...... 40 Population Sampling ...... 40 Extraction of Genomic DNA and PCR Amplification of ISSR Population Markers...... 40 Results...... 44 Analysis of ISSR Markers...... 44 Analysis of ITS...... 45 Analysis of trnL-F ...... 47 Discussion...... 47

3 EVALUATION OF FUNGAL PATHOGENS AS BIOLOGICAL CONTROL AGENTS OF Imperata cylindrica (L.) BEAUV. FROM THE USA AND BENIN...... 60

Introduction...... 60 Materials and Methods ...... 62 Propagation of Test Plants...... 62 Inoculum Production ...... 63

Imperata cylindrica Inoculations ...... 64 Efficacy of Bipolaris sacchari and Drechslera gigantea Applied Singly or as a Mixture...... 64 Statistical Analysis ...... 65 Results...... 65 Efficacy of Bipolaris sacchari and Drechslera gigantea Applied Singly or as a Mixture...... 65 Florida Imperata cylindrica Accessions and Fungal Isolates...... 65 Benin Imperata cylindrica Accessions and Fungal Isolates...... 68 Discussion...... 69

4 EVALUATION OF OBLIGATE BIOTROPHIC FUNGI ASSOCIATED WITH Imperata cylindrica IN SOUTH AFRICA AS POTENTIAL BIOCONTROL AGENTS FOR COGONGRASS IN WEST AFRICA...... 92

Introduction...... 92 Materials and Methods ...... 95 Propagation of Test Plants...... 95 Source of Pathogens and Inoculum ...... 95 Treatments and Experimental Design ...... 96 Results...... 96 Discussion...... 97

5 SUMMARY AND CONCLUSIONS...... 103

APPENDIX

A FIELD EVALUATION OF THE INDIGENOUS FUNGUS, Colletotrichum caudatum, ASSOCIATED WITH COGONGRASS IN BENIN ...... 108

Introduction...... 108 Materials and Methods ...... 111 Inoculum Production ...... 111 Inoculation...... 111 Disease Assessment...... 112 Statistical Analysis ...... 112 Results...... 112 Discussion...... 113

B ACCESSION AND GENBANK DATA FOR THE USA Imperata cylindrica AND CLOSELY RELATED SPECIES...... 119

C ACCESSION AND GENBANK DATA FOR THE WEST AND SOUTH AFRICAN Imperata cylindrica ...... 121

LIST OF REFERENCES...... 122

BIOGRAPHICAL SKETCH ...... 140

LIST OF TABLES

Table page

2-1 Genetic variability of all single sampled populations of Imperata cylindrica and closely related species with the ISSR primers...... 53

2-2 Comparison of genetic diversity measures for all sampled populations of Imperata cylindrica and closely related species as determined using ISSR population markers. ....54

2-3 Summary of analysis of molecular variance (AMOVA) based on ISSR genotypes of the African and USA Imperata cylindrica accessions...... 55

3-1 Pairwise comparison of least square means of disease severity (DS) from different treatments on Florida accessions of Imperata cylindrica at 7 days after inoculation...... 75

3-2 Pairwise comparison of least square means of disease severity (DS) from different treatments on Florida accessions of Imperata cylindrica at 21 days after inoculation...... 75

3-3 Disease severity statistical classes of Bipolaris sacchari, Drechslera gigantea, and their mixture on the Florida Imperata cylindrica accessions at 7 days after inoculation...... 76

3-4 Disease severity statistical classes of Bipolaris sacchari, Drechslera gigantea, and their mixture on the Florida Imperata cylindrica accessions at 21 days after inoculation...... 77

3-5 Pairwise comparison of least square means of disease severity (DS) from different treatments on the Benin accession of Imperata cylindrica at 7 days after inoculation (Trial 1)...... 78

3-6 Pairwise comparison of least square means of disease severity (DS) from different treatments on the Benin accession of Imperata cylindrica at 7 days after inoculation (Trial 2)...... 79

3-7 Pairwise comparison of least square means of disease severity (DS) from different treatments on the Benin accession of Imperata cylindrica at 21 days after inoculation (Trial 1)...... 80

3-8 Pairwise comparison of least square means of disease severity (DS) from different treatments on the Benin accession of Imperata cylindrica at 21 days after inoculation (Trial 2)...... 81

4-1 Collections of obligate fungi associated with Imperata cylindrica in South Africa...... 100

4-2 Number of Imperata cylindrica inflorescences infected with the head smut, Sporisorium schweinfurthianum...... 100

A-1 Analysis of variance for disease incidence of Colletotrichum caudatum at 7 days after inoculation...... 116

B-1 Imperata cylindrica and closely related species sampled in USA for population structure and virulence studies...... 119

C-1 Imperata cylindrica and closely related species sampled in West and South Africa for population structure and virulence studies...... 121

LIST OF FIGURES

Figure page

2-1 Photographs of ISSR gels...... 56

2-2 UPGMA dendrogram of all Imperata cylindrica accessions and closely related species sampled based on ISSR data...... 57

2-3 Parsimony tree inferred from ITS sequence data for all Imperata cylindrica accessions and closely related species...... 58

2-4 Neighbor-joining tree of distances derived from ITS sequence data for all Imperata cylindrica accessions and closely related species...... 59

3-1 Disease symptoms and disease severity on Imperata cylindrica Florida accessions at 21 days after inoculation...... 82

3-2 Disease symptoms and diseases severity on Imperata cylindrica Benin accession at 21 days after inoculation...... 83

3-3 Mean disease ratings of Bipolaris sacchari on Imperata cylindrica Florida accessions (Trial 1)...... 84

3-4 Mean disease ratings of Drechslera gigantea on Imperata cylindrica Florida accessions (Trial 1)...... 85

3-5 Mean disease ratings of the Mixture on Imperata cylindrica Florida accessions (Trial 1)...... 86

3-6 Mean disease ratings of Bipolaris sacchari on Imperata cylindrica Florida accessions (Trial 2)...... 87

3-7 Mean disease ratings of Drechslera gigantea on Imperata cylindrica Florida accessions (Trial 2)...... 88

3-8 Mean disease ratings of the Mixture on Imperata cylindrica Florida accessions (Trial 2)...... 89

3-9 Disease severity of Florida (Fl) and Benin (B) isolates of Bipolaris sacchari, Drechslera gigantea, and their mixture on the Benin accession of Imperata cylindrica...... 90

4-1 Imperata cylindrica seedlings prior to inoculation with a 1 x 108 sori ml-1 Sporisorium schweinfurthianum soil drench treatment...... 101

4-2 Imperata cylindrica inflorescence infected with the head smut, Sporisorium schweinfurthianum four months after a soil drench treatment...... 102

A-1 Field trial site with reemerging Imperata cylindrica plants at IITA, Cotonou, Benin. ...117

A-2 Field inoculation of Imperata cylindrica Benin using an amended Colletotrichum caudatum formulation...... 118

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

BIOLOGICAL CONTROL OF Imperata cylindrica IN WEST AFRICA USING FUNGAL PATHOGENS

Alana Den Breeyen

December 2007

Chair: Raghavan Charudattan Major: Plant Pathology

Imperata cylindrica (cogongrass, speargrass) is a noxious, rhizomatous grass with a pan- tropical distribution. It is ranked as one of the top 10 invasive weeds in the world. The plant has several survival traits that include an extensive rhizome system, wind disseminated seeds and high genetic plasticity. In the southeastern USA, it is a very serious invasive species and a federally designated noxious weed. In West Africa, cogongrass represents one of the most serious constraints to crop production and poverty alleviation. Within the genus Imperata, cogongrass is the most variable species and specimens have been found ranging from the western

Mediterranean to South Africa, throughout Australia, India and Southeast Asia (including the

Pacific Islands). Included in the study were other Imperata taxa including I. brevifolia, I. brasiliensis, and the ornamental variety I. cylindrica var. rubra. Due to possible hybridization, introgression, and overlapping morphological characters, there is often taxonomic confusion between cogongrass and I. brasiliensis, especially in North America. Biological control would be an economical and environmentally friendly method for controlling this invasive weed and the objective of this study was to determine the suitability of the West African cogongrass as a host to two pathogens, Bipolaris sacchari and Drechslera gigantea, found to be effective biological control agents in the southeastern USA. Understanding the extent of genetic

variation can be of importance in the detection of ecotypes or races of a species as they may

differ considerably in their susceptibility to predators, parasites, or herbicides. In the course of

this study, USA and West African I. cylindrica accessions were found to be geographically and

genetically distinct. These differences were defined by three distinct groups. Within the Florida

cogongrass accessions, genetic variation was high confirming that I. cylindrica was introduced in

the USA multiple times. Imperata brasiliensis was not genetically distinct from the USA I.

cylindrica population forming a sister species to the Florida I. cylindrica accessions. In addition,

Imperata cylindrica var. rubra was more closely related to the African accessions. Imperata

brevifolia was found to be genetically distinct from all the Imperata accessions within the ISSR

study. Bipolaris sacchari and D. gigantea isolates from the USA and West Africa were

evaluated in greenhouse trials to determine their efficacy as biological control agents on the West

African I. cylindrica. There were no significant differences in disease incidence and disease

severity between the Florida and Benin isolates inoculated on the Benin cogongrass. The Florida

isolates of B. sacchari and D. gigantea are not suitable for use in a formulated mycoherbicide

product for cogongrass control in West Africa because of low levels of disease and lack of weed

suppression. An indigenous fungus Colletotrichum caudatum, evaluated in conjunction with the

B. sacchari and D. gigantea isolates, was consistently found to cause similar or greater disease

symptoms on the Benin accessions than the latter. Puccinia imperatae, P. fragosoana and,

Sporisorium schweinfurthianum, associated with cogongrass in South Africa were evaluated for

control of cogongrass in West Africa. Neither Puccinia spp. produced symptoms on I. cylindrica

Benin, whereas S. schweinfurthianum replaced the inflorescences with a teliospore-producing stroma, causing floral sterility. The amended C. caudatum spore formulation elicited more

severe symptoms compared with the mycelial inoculum and control treatments.

CHAPTER 1 INTRODUCTION

Invasive exotic species are a threat to global biodiversity second only to habit loss and

landscape fragmentation, and the economic impact of these species is a serious concern worldwide (Allendorf and Lundquist, 2003). An estimated 50,000 non-indigenous species

established in the United States cause major environmental damage and economic losses totaling over an estimated $125 billion, with 79 of these exotic species causing over $97 billion in damages from 1906 to 1991 (Pimental et al, 2005; Allendorf and Lundquist, 2003). In crop

systems, more than 500 introduced species are considered serious weeds while over 5,000

introduced species are found in natural ecosystems (Pimentel et al., 1989, Morin, 1995). In

1999-2000, over $41 million was spent in Florida for the management of invasive terrestrial and

aquatic plant species occurring in waterways, roadsides, rights-of-way, forests, and other natural

areas (Florida DEP, 2000). As a result of altering ecosystem processes, invasive plant species

displace native plant and animal species, support exotic animals and pathogens, and subsequently

alter gene pools by hybridizing with native species (Randall, 1996).

In the United States, invasive weeds cause an approximate 12% reduction in crop yields.

Agriculture, the mainstay of most national economies in Africa, provides 60% of all employment

and generates more than 40% of the continent’s foreign exchange earnings (GISP Brochure,

2007). In rural communities, agriculture supports 80% of the population and together with food

security, is critical to the livelihoods and survival of individuals, communities and countries in

Africa. According to McGinley (2007), Africa spends an estimated $60 million annually to

control invasive alien species. The threats posed by invasive species consequently have a greater

impact in developing countries, whose economies rely heavily on agriculture, forestry and

fishing (GISP Brochure, 2007). Rural communities in developing countries, whose livelihoods

are exclusively based on these economic sectors, are the most at risk while the poorest section of

these communities are wholly dependent on biodiversity-based products for food, fuel and

construction material.

Biology and Morphology

A warm-season, rhizomatous, grass species, cogongrass [Imperata cylindrica (L.) Beauv.] is found throughout tropical and subtropical regions of the world (Holm et al., 1977). It employs several survival strategies that include an extensive rhizome system, adaption to poor soil, drought tolerance, adaptability to fire, wind disseminated seeds and high genetic plasticity

(MacDonald, 2004). Cogongrass is a perennial, C4 plant species which allows it to

photosynthesize more efficiently under high light intensities and temperatures, and under

conditions of limited moisture availability (Groves, 1991).

Growing in loose -to- compact tufts, cogongrass plants can reach up to 1 meter in height

(Clewell, 1985). Considered stemless except for the flowering stalk, one to eight basal leaves

(4– to 10- mm wide; 15– to 150- cm long) originate from the rhizomes. The linear-lanceolate leaves, with a characteristic off-center white midrib, have high silicon content (Coile and

Shilling, 1993) that results in coarse leaf texture and serrated leaf margins. Cogongrass rhizomes, comprising of greater than 60% total plant biomass, are covered with tough scale leaves (cataphylls) that are thought to protect the rhizomes from desiccation, physical degradation and fire (Holm et al., 1997). The rhizomes have short internodes and form extensive, dense underground mats. The rhizomes can develop in two stages: primary seedling rhizomes, and secondary rhizomes that sprout from seedling rhizomes (Gabel, 1982). In central

Florida, Gaffney (1996) found cogongrass rhizomes were restricted to the top 10 to15 cm of soil on a phosphate mine site, but grew down to 80 cm below ground on a clay settling pond site. In

Southeast Asia, rhizomes typically occur 10 to 40 cm below ground and form dense, extensive

layers with some rhizomes growing as deep as 1 m (Ayeni and Duke, 1985; Murniati, 2002).

The cylindrical inflorescence is a dense panicle of paired spikelets (3 to 20 cm long, 0.5 to 2.5

cm). The spikelets are unawned with long silky hairs (Clewell, 1985) producing small seeds (1

to 1.3 mm long) (King and Grace, 2000a, b).

Worldwide, cogongrass shows variable phenological development depending upon climate and population genetics. In its native Japan, cogongrass flowers from May to June (Ohwi, 1965) while populations growing in the Mediterranean tend to flower during the spring and summer

(October to March). In tropical and subtropical areas, including Florida, cogongrass tends to

flower year-round (Bryson and Carter, 1993). Flowering is often caused by stress due to

mowing, burning, grazing, cool temperatures, defoliation, and nitrogen fertilization (Holm et al.,

1997; Coile and Shilling, 1993).

Cogongrass reproduces by seed, with a single plant producing several thousand which can

be wind-dispersed over long distances. Although seed movement is usually limited to 15 m or

less, seeds were reported to travel up to 24 km when in open country (Hubbard, 1944). Seed dormancy is absent and mature seeds germinate within a week of shedding (Santiago, 1965) and

remain viable for up to one year. Imperata cylindrica also reproduces vegetatively through the

rhizomes allowing the plant to rapidly colonize new areas. Eussen (1980) reported that a single

rhizome node could lead to the production of 350 shoots in 6 weeks and ground cover of 4 m2 in

11 weeks.

Cogongrass in the United States

Cogongrass, ranked as one of the world’s top ten invasive weeds (Holm et al., 1977),

infests over 500 million ha worldwide including 200 million ha in Asia and an estimated 100,000 hectares in southeastern USA (Schmitz and Brown, 1994). Several known introductions occurred in the United States during the last century with the variety major accidentally

introduced into Alabama (1912) from Japan while cogongrass from the Philippines was

purposefully introduced to Mississippi in 1921 as potential forage (MacDonald, 2004).

Additional forage trials were conducted in Florida, Texas (where the planting died out in the first

year) and Alabama with the consequence that by the late 1940s over 1000 acres had become established in central and northwest Florida. By 2004, greater than 500,000 hectares with some level of cogongrass infestation was present in Florida (MacDonald, 2004). Cogongrass infests ditch banks, pastures, road sides/right of ways, golf courses, lawns, forests and recreational and natural areas throughout Florida with the reclaimed phosphate mining areas of Central Florida supporting hundreds of acres of cogongrass monoculture (MacDonald et al., 2002).

Cogongrass infestations, in areas other than closed canopy forests, plantations and heavily cultivated lands, are managed using relatively expensive, laborious and repetitive control measures. In the United States, the most effective management strategies require an integrated approach utilizing mechanical (discing, mowing), cultural (burning), chemical (herbicide applications of glyphosate and imazapyr) and revegetation methods (Van Loan et al., 2002).

Effective control strategies may be hampered by the existence of genetic variation and potentially adapted genotypes due to fact that there were at least two separate cogongrass introductions and the occurrence of Imperata brasiliensis Trin. (Brazilian satintail) in the United

States. According to McDonald et al. (1996), crosses of I. cylindrica and I. brasiliensis produced viable seeds. This observation raises the question of whether these two species should be considered forms of the same species. Hall (1983, 1998) has considered these to be members fo the same species due to the evidence of frequent hybridization (Gabel, 1982). The ‘ornamental’ variety I. cylindrica var. rubra (Red Baron or Japanese Blood Grass), sold in nurseries throughout the United States, is extremely cold-tolerant (-23ºC) unlike I. cylindrica var. major.

In Alabama, cogongrass rhizomes were able to survive winter temperatures of -14ºC (Wilcut et

al., 1988), but were unable to survive winters in Mississippi when the temperature reached -8 ºC

(Hubbard, 1944). The creation of new potentially cold-tolerant biotypes through the

hybridization of I. cylindrica with var. rubra could possibly extend the range of this invasive

weed north and westward (Shilling et al., 1997).

Cogongrass in West Africa

Imperata cylindrica var. africana (cogongrass and/or speargrass) ranges from Senegal to

Cameroon and from the south coast to the arid Sudan zone in the north. It is a serious weed in

the moist savanna and forest zones especially in areas under intensive agricultural practices

including recurrent fires, tillage, weeding and other farm activities (Chikoye et al., 2001).

Although cogongrass occurs in South Africa, Madagascar and East Africa, it has greater invasive

significance in Western Africa than either Eastern or Southern Africa (Ivens, 1983). According

to Ivens (1983), I. cylindrica var. africana differs morphologically from the var. major of

Southeast Asia (and consequently of the USA) with generally narrower leaf width, longer

ligules, slightly longer spikelets with longer hairs and anthers. However, it has similar growth

habits and ecological requirements as var. major.

Cogongrass infestations place major constraints on the production of coconut, oil-palm,

rubber, cereals, legumes, root and tuber crops (Chikoye et al., 1999). Annual crop yields are

drastically reduced when grown in competition with I. cylindrica often causing losses of 80% to

100%. When crops were grown in slash plots without additional weeding, complete crop failure usually occurred (Koch et al., 1990; Udensi et al., 1999). Koch et al. (1990) reported that I. cylindrica infestations of cassava fields intercropped with maize, caused yield losses greater than

90% while physical injuries caused by the sharp cogongrass rhizomes to roots and tubers of cassava and yam increased fungal disease incidences resulting in additional crop loss. In

addition, I. cylindrica increases the risk of fires in perennial crops, plantations, and forest reserves and recurring fires result in soil degradation due to significant loss of organic matter

(Holm et al., 1977, Terry et al., 1997). Cogongrass is a serious invasive weed that can decimate crop yields and demands the commitment of significant proportions of resources, such as capital, labor, and herbicides, for effective control. West African farmers perceived cogongrass to be their most serious weed (Chikoye et al., 1999). Labor-intensive farming practices such as hoe- weeding, slashing and fallowing are the most common control practices. Long-term land abandonment has been the traditional way to manage severe infestations of cogongrass in West

Africa. This is because long-term land abandonment of 10 to 15 years results in extensive natural bush regrowth that effectively smothers the weed (Chikoye et al., 1999). However, long- term fallow cropping is no longer feasible due to increasing population pressure on limited arable land. Glyphosate or paraquat are the only two herbicides available if farmers chose to utilize chemcical control as part of their management strategy. However, the high cost of herbicides, lack of awareness, lack of capital, and the non-availability of effective herbicides were cited as reasons for the low utilization of herbicides (Chikoye et al., 2000). Farmers can combine several control practices in an integrated management strategy for enhanced suppression of the weed.

Lack of capital resources however, was the most widely stated reason for the non-development of sustainable management strategies. Other reasons included: 1) the lack of better management strategies, 2) the lack of adequate labor, 3) the lack of equipment, and 4) health problems.

Chikoye et al. (1999) concluded that research to develop a comprehensive management strategy for cogongrass in West Africa must become a priority as this weed threatens the livelihood of over 200 million people in the region.

Biological Control Strategies for Forest Weeds

Invasive weeds cause significant economic and ecological losses in both commercial and native forest ecosystems. These weeds are highly successfully and compete for resources such as light, nutrients and water which results in the suppression of young or naturally regenerating trees (Green, 2003). According to Campbell (1998), an estimated 138 alien tree and shrub species including salt cedar (Tamarix pendantra Pallas), eucalyptus (Eucalyptus spp.), Brazilian pepper (Schinus terebinthifolius Raddi), and Australian melaleuca (Melaleuca quinquenervia

(Cav.) S.T. Blake), have invaded native U.S. forest and shrub ecosystems (Pimental et al., 2005).

Melaleuca quinquenervia is spreading at a rate of 11,000 ha/year throughout the forest and grassland ecosystems of the Florida Everglades causing damage to the natural vegetation and wildlife (Pimental et al., 2005).

Acacia saligna – Uromycladium tepperianum. One of the most successful classical biocontrol programs documented is the control of Acacia saligna (Labill.) H. Wendl. in South

Africa, by the Australian rust fungus Uromycladium tepperianum (Sacc.) McAlpine (reviewed by Charudattan, 2001). As one of the most serious invasive weeds in South Africa, A. saligna

(Port Jackson willow) forms dense stands that replace indigenous vegetation and interfere with agricultural practices (Morris, 1997). As the weed is difficult and costly to control mechanically and chemically, it was an excellent candidate for biological control. In 1987, a gall-forming rust fungus, U. tepperianum, was imported into South Africa after extensive host-range testing undertaken in Australia confirmed its specificity (Morris, 1987). Morris (1987) studied the teliospore germination, early stages of host infection, host specialization and the reactions of some African Acacia and Albizia species to inoculation with U. tepperianum in Australia. Cross- inoculation of teliospores isolated from A. implexa Benth., A. saligna and Paraserienthes lophantha spp. lophantha (Willd.) Nielson with the same three species suggested that distinct

genotypes of the rust occur on these species. As the reactions were distinguishable at the host-

species level, they should, according to Anikster (1984), be termed formae speciales. Morris

(1987) showed that normal galls only developed on the species from which the teliospores were

collected even though several known U. tepperianum host species were included in the study.

Although the results indicated that these rust genotypes may be specific to a single host species,

Morris (1987) recommended that a larger range of recorded host species be tested prior to

formally designating the formae speciales. Since the initial releases in the late 1980s, U.

tepperianum, has subsequently spread and established throughout the range of A. saligna

(Morris, 1997). Fifteen years of monitoring (1991–2005) showed that tree density declined by

between 87% and 98% compared with data recorded prior to the U. tepperianum release (Wood

and Morris, 2007). The introduction of the seed-feeding weevil, Melanterius compactus Lea,

should accelerate a continuous decline in stand density over time. According to Wood and

Morris (2007), U. tepperianum remains a highly effective biological control agent against the

invasive tree A. saligna in South Africa.

Passiflora tarminiana-Septoria passiflorae. Passiflora tarminiana Coopens (banana

passion flower, banana poka vine), native to the South American Andes, is an aggressive and

invasive tropical vine in Hawaii. It invades disturbed areas and its effects include a loss of

biodiversity, smothering of trees and the encouragement of other invasive species such as feral

pigs that feed on the fruit (ISSG Database, 2005). By 1983, more than 50,000 ha of wet and

mesic forests in Hawaii were infested with P. tarminiana costing the Division of Forestry and

Wildlife $90,000 annually in the Hilo Forest Reserve (Trujillo, 2005). In 1991, Septoria passiflorae Syd. was isolated from infected Hawaiian banana poka seedlings growing in

Colombia. Subsequent to the pathogen’s importation to Hawaii, host-range tests, completed in

1994, confirmed its specificity to banana poka vine (Trujillo et al., 1994). Septoria passiflorae was approved as a biocontrol agent for banana poka vine in 1995. In 1996/1997, field inoculations using spore suspensions resulted in significant control of banana poka vines on the islands of Hawaii, Maui and Kauai (Trujillo, 2005). In the Hilo Forest Reserve, Hawaii, banana poka vine biomass was reduced between 40 to 60% annually after the first inoculations. Four years later, a 95% biomass reduction was observed (Trujillo, 2005). The introduction of S. passiflorae has resulted in the preservation of endangered species and the regeneration of the indigenous Acacia koa Gray forest (the source of the most valuable timber species in Hawaii) while saving millions of dollars in weed control and forest revitalization in Kauai, Maui and

Hawaii.

Ageratina riparia-Entyloma ageratinae. Ageratina riparia (Regel) R. King &

H.Robinson (mist flower, Hamakua ‘Pa-makani’), a low-growing perennial with tiny white daisy-like flowers, was accidentally introduced to Hawaii in 1925. By 1972, mist flower infestations extended over 100,000 ha of range and forest lands on the Hawaiian Islands (Morin et al., 1997; Trujillo, 1985). The foliar smut fungus, Entyloma ageratinae Barreto and Evans was introduced to Hawaii from Jamaica in 1975 to attack this aggressive weed in the mid-1970s.

Initally misidentified as a Cercosporella sp., the pathogen was renamed E. ageratinae sp. nov.

Barreto and Evans (Barreto and Evans, 1988), based on physiological host reactions rather than morphological differences. However, Trujillo (2005) renamed the fungus as E. compositarum f. sp. ageratinae. After extensive host range studies confirmed that the pathogen was specific to mist flower, field inoculations were made at invaded sites throughout the Hawaiian Islands

(Trujillo, 2005). Total control of the plant was achieved in less than seven months in wet areas and within three to eight years in dry areas. Biological control of mist flower in Hawaii has been

an outstanding success in Hawaii with extensive rehabilitation of the Hawaiian rangelands.

Entyloma ageratinae was subsequently released in New Zealand in 1998 to suppress mist flower.

Within two years, it had become established at all the initial release sites and was found up to 80

km from the nearest release site (Barton et al., 2007). Within 5.5 years it had spread to the entire

North Island mist flower sites with the most distant site found 440 km from the nearest release.

Results showed that almost 60% infection of live leaves was quickly reached and maximum

plant height declined significantly. As the mist flower cover significantly decreased due to the

introduction of the fungus, the species richness and mean percentage cover of native plants

increased. It is interesting to note that no significant change was observed in the species richness

and mean percentage cover of exotic plants (excluding mist flower). According to Barton et al.

(2007) the introduction of E. ageratinae as a biological control agent of mist flower in New

Zealand has been very successful.

Broad-leaved tree species-Chondrostereum purpureum and Cylindrobasdium laeve.

The use of endemic wood-rotting fungi as bioherbicides to control invasive woody weed species has been successfully implemented in the Netherlands and South Africa (Green, 2003). Applied directly on freshly cut stump surfaces, these fungi prevent re-sprouting of inoculated tree stumps.

Chondrostereum purpureum (Pers:Fr.) Pouzar was developed as the mycoherbicide BIOCHON in the Netherlands, where it was used to control the introduced Prunus serotina Ehrh. (de Jong et al., 1990). BIOCHON was commercially available for several years as a method for combating the invasive North American Prunus and Populus spp. in natural and commercial forests in the

Netherlands (de Jong, 2000). In Canada, two stump treatment products, Myco-tech™ and

Chontrol™, containing native isolates of C. purpureum have been developed and registered.

Extensive epidemiological evidence showed that C. purpureum posed no threat to susceptible

crops grown further than 500 m from the application site (de Jong et al., 1990). The use of this

control method will have only a limited impact on non-target trees since the spores of this fungus

are ubiquitous and healthy trees are resistant to attack. However, Setliff (2002) showed that C.

purpureum is an important pathogen with epidemic potential in forest tree species especially in

the Betulaceae (Betula and Alnus) and the Saliaceae (Populus and Salix) families. A thorough

investigation into the ecological impacts of C. purpureum for each geographical region should be

undertaken with the purpose of identifying low disease risk areas where C. purpureum can be

applied with the minimum environmental impact (Setliff, 2002). In South Africa, a

basidiomycete Cylindrobasidium laeve (Pers.) Chamuris was registered as the bioherbicide,

Stumpout. It is applied directly to freshly cut stumps of several invasive Acacia spp.. According

to Morris et al. (1999), this bioherbicide kills 95 to 100% of resprouting stumps.

Clidemia hirta-Colletotrichum gloeosporioides f.sp. clidemiae. Clidemia hirta (L.) D.

Don (Koster’s curse), a perennial shrub native to Central and South America, and the islands of

the West Indies (Wester and Wood, 1977, Steyermark and Huber, 1978 in Peters, 2001) is

invasive in tropical forest understories particularly in the Hawaiian Islands. The negative effect

on the native ecosystem has led to the fear that similar effects will be felt in the other introduced

regions including the islands of Seychelles, the Comoros Island, Réunion, Mauritius, the

Malaysian Peninsula and parts of Micronesia (Palau) (Gerlach, 2006). First reported in Oahu in

1941 (Anon, 1954), Clidemia populations reached significant proportions by 1984 with more than 60,000 ha infested forested areas on Kauai, Molokai, Maui, and Hawaii (Trujillo, 2005).

All management strategies implemented to date including physical removal and herbicide

treatments implemented have not proven practical and generally failed especially when initiated

after first fruit set. A fungus, Colletotrichum gloeosporioides (Penz) Sacc. f. sp. clidemiae was

isolated from diseased clidemia leaves collected in Panama in 1985 (Trujillo, 2005). Following

the confirmation of specificity of this pathogen to C. hirta, permission was granted in 1986 by

the Hawaii Board of Agriculture to use C. gloeosporioides f. sp. clidemiae to control clidemia.

Repeated field inoculations undertaken from 1986 until 1992 at Aiea State Park, Oahu resulted in

significant control of the weed (Trujillo, 2005). After several spore formulations were tested for

field applications, a 5 x 105 per ml spore suspension amended with 0.5% gelatin and 2.0%

sucrose that caused severe defoliation was selected. Annual sprays directed at clidemia in Aiea

State Park, Oahu using this spore formulation resulted in effective control of the weed and in the

regeneration of several native species such as Acacia koa and fern species (Trujillo, 2005).

Hedychium gardnerianum-Pseudomonas solanacearum. Native to India, Hedychium

gardnerianum Ker-Gawl (wild ginger or kahili ginger) is widespread throughout the tropics and

is invasive in many forest ecosystems including Hawaii, New Zealand, Réunion, South Africa and Jamaica (Anderson and Gardner, 1999). In 1954, H. gardnerianum was brought into Hawaii as an ornamental plant where it subsequently escaped cultivation and is currently considered a naturalized plant species. According to Anderson and Gardner (1999), kahili ginger has invaded

500 ha of Hawaii Volcanoes National Park forests (1000-1300 m elevation) and is found on all islands in Hawaii (Smith, 1985). Displacing native plants in forest ecosystems, kahili ginger forms dense stands that smother the native understory. Further spread is facilitated by rhizomes which make this weed difficult to control (ISSG Database, 2006). Due to the fact that kahili ginger is widely distributed throughout the Hawaiian national parks, environmental concerns regarding herbicide use to control kahili ginger has resulted in biological control being considered as the only practical approach for long-term management of this invasive weed in native forests. A Hawaiian Ralstonia (=Pseudomonas) solanacerum (E.F. Smith) Yabuuchi et

al. strain was isolated from both edible (Zingiber officinale Roscoe) and ornamental gingers.

The isolate from edible ginger was less virulent than the isolate from ornamental ginger

(Anderson and Gardner, 1999). Ralstonia solanacerum systemically infects edible and

ornamental gingers causing decay and wilting of the infected tissue. Host-specificity tests

showed that this strain did not infect the tested native and cultivated solanaceous species. All

inoculated H. gardnerianum plants including rhizomes developed symptoms within 3 to 4 weeks

after inoculation, with most inoculated plants completely dead within 4 months (Anderson and

Gardner, 1999). While limited infection occurred close to the inoculation sites on H. coronarium

J. König, Z. zerumbet (L.) Sm., Heliconia latispatha Benth., and Musa sapientum Gros Michel,

no further systemic infection was observed in these species and the plants continued to grow

normally (Anderson and Gardner, 1999). Despite concerns about the potential negative impacts

of the bacterium on the edible ginger Z. officinale, Anderson and Gardner (1999) concluded that

kahili ginger infestations were often remote enough to limit the risk of contamination to edible

ginger plants. Ralstonia solanacerum should contribute to the control of H. gardnerianum,

however, possible contamination of agricultural lands from runoff water from treatment areas

should be considered when using the R. solanacearum as a biocontrol agent (Anderson and

Gardner, 1999).

Biological Control Strategies for Cogongrass

The use of one or more biotic organisms to maintain another pest organism at a level where it is no longer a problem is known as biological control. Agents are selected on the basis of their specificity and ability to cause significant damage to the target organism (Evans, 1991).

Biological control of weeds with plant pathogens can adopt either a classical, augmentative or an inundative (bioherbicide) approach. The classical biocontrol approach involves the introduction of natural enemies from the center of origin to an area where it does not occur to control the

target weed in exotic environment, and augmentative approach involves periodic releases of the

control agent to augment an existing biocontrol population, while an inundative approach utilizes

the application of indigenous biocontrol agents in massive doses to create an epidemic within the

target weed population. During the 1970s, when weed biocontrol was still the domain of

entomologists, there was a rapid increase in interest by scientists in exploiting pathogens,

particularly fungi, for weed control (Freeman, 1981). Until fairly recently there has been a

dearth of information on the biological control of grassy weeds (Evans, 1991) with information

on the natural enemy complex. The effect of this complex on the population dynamic is largely

unknown and a general lack of basic botanical information on the centers of origin and diversity.

Implementing biological control strategies to manage cogongrass is especially pertinent in

natural ecosystems and low maintenance areas where the costs of using chemical and mechanical

control methods would be prohibitive. Based on literature records and on-line databases, there

are a number of potential biocontrol agents, including plant pathogens, arthropods and other

invertebrates associated with cogongrass throughout the world (Van Loan et al., 2002). This list

includes 24 fungi, 51 insects, six nematodes, four mites and a parasitic plant in the USA (Minno

and Minno, 1998).

Potential Insect and Nematode Biocontrol Agents

In Malaysia, Vayssiere (1957) reported several polyphagous insect species that caused

damage on cogongrass as well as other cultivated cereals. Although Bryson (1985, 1987) reported that three species of the North American skipper butterfly (Lepidoptera: Hesperiidae) fed on cogongrass, they were deemed to be unsuitable as biocontrol agents due to their lack of host specificity. The gall midge, Orseolia javanica Kieffer and van Leeuwen-Reijnvaan reportedly host specific to I. cylindrica (Mangoendihardjo, 1980; Soenarjo, 1986) and destroys the apical shoot meristem particularly in areas where I. cylindrica is regularly cut or slashed. By

reducing photosynthesis due to leaf blade reduction, heavy infestations could ultimately lead to

lower rhizome carbohydrate reserves. Unfortunately, gall midges are known to be highly parasitized by generalist parasitoids (Van Loan, 2002) thereby limiting their effectiveness as biocontrol agents.

Other potential agents include the nematode Heterodera sinensis Chen, Zheng and Peng,

the mite Aceria imperata Zaher and Abou-Awad and two unidentified dipteran stem borers (Van

Loan, 2002). An extensive survey of herbivorous arthropods and pathogens of cogongrass

throughout Florida (Minno and Minno, 1998) showed that the insect fauna associated with

cogongrass in Florida was similar to that associated with cogongrass in Africa and Asia. Several fungi, nematodes, mites, elaterid and scarab beetles, skipper butterfly and moth caterpillars, and grasshoppers were shown to cause damage.

Potential Plant Pathogenic Biocontrol Agents

Pathogens associated with cogongrass include Myrellina imperatae Sankaran and Sutton

which causes leaf spots and chlorosis and Giberella imperatae C. Booth and Prior which

originates from from Papua New Guinea and Australia and causes dieback of cogongrass

(Sankaran and Sutton, 1992; Booth and Prior, 1984). Xanthomonas albilineans (Ashby)

Dowson, the causal agent of leaf scald of sugarcane, reportedly caused the appearance of

discolored lines along the leaf veins of cogongrass in Australia (Persley, 1973). Other reported

pathogens include Puccinia rufipes Diet., Claviceps imperatae Tanda and Kawatani, Monodisma

fragilis Alcorn, Deightoniella africana Hughes, Mycosphaerella imperatae Sawada, Bipolaris

maydis (Y. Nisik.) Shoemaker, Colletotrichum caudatum (Sacc.) Peck, C. graminicola (Ces.)

G.W. Wilson, Aschochyta sp., Didymaria sp., Dinemasporium sp., Chaetomium fusiforme

Chivers and Helminthosporium, Curvularia and Fusarium species (Chadrasrikul, 1962; Chase et

al., 1996; Caunter, 1996).

In the late 1980s, Caunter and Wong (1988) surveyed for foliar pathogens of I. cylindrica throughout Malaysia and, while they generally found low disease incidence, several candidate fungi were isolated including Colletotrichum caudatum, Aschochyta sp., P. rufipes, C. graminicola, Didymaria sp., and Dinemasporium sp. An investigation into the bioherbicide potential of the pathogen C. caudatum in Malaysia showed that greenhouse inoculations caused necrotic streaks along the leaf edges but no plant death was observed (Caunter, 1996). C. caudatum was recently identified on cogongrass in Florida (Minno and Minno, 1998).

Chase et al. (1996) evaluated the pathogenicity of several fungi isolated from diseased cogongrass collected at various sites in Florida. Research showed that four of the six isolates were pathogenic to cogongrass causing up to 100% disease incidence. There was no record of weed mortality. Chaetomium fusiforme and an isolate from the Helminthosporium species group were found to be the most promising based on the level of disease incidence recorded in the greenhouse. In 2001, Yandoc (2001) screened both fungal isolates collected from cogongrass stands in north central Florida and fungal pathogens previously isolated from various diseased grasses collected throughout Florida for pathogenicity and biocontrol potential. The study revealed that an isolate of B. sacchari from cogongrass and D. gigantea from large crabgrass had potential as biocontrol agents. Several obligate biotrophs including Puccinia imperatae Poirault,

P. fragosoana F. Beltran and P. rufipes Diet. should be considered as potential classical biocontrol agents for cogongrass. A smut, Sphacelotheca schweinfurthiana (Thum.) Sacc., is well represented on cogongrass in the Old World, mainly Africa (Evans, 1991) and recently,

Minno and Minno (1998) found the smut on cogongrass in Florida. It is interesting to note that these obligate parasites are common in the Mediterranean region where I. cylindrica is not a serious weed (Evans, 1991; Holm et al., 1977).

Mycoherbicides

Mycoherbicides utilize indigenous fungi isolated from weeds and cultured to produce large

quantities of spores or mycelia that can be applied at rates that will cause high levels of infection.

The eventual goal of the mycoherbicide approach is the suppression of the target weed before

economic losses are incurred (Templeton, 1982). One to several applications is generally

required annually as the pathogens do not maintain sufficient inoculum to initiate new epidemics

on new weed infestations (Charudattan, 1988; TeBeest et al., 1992). According to Yandoc-Ables

et al. (2007), 200 or more plant pathogens, including foliar and soil-borne fungi, bacteria and

deleterious rhizobacteria have been or are currently under evaluation as potential bioherbicides.

Worldwide, 10 bioherbicides have been registered since 1980; in South Africa, Stumpout®

(Cylindrobasidium laeve) and Hakatak (Colleotrichum acutatum J.H. Simmonds); in the USA,

Lockdown®[formerly Collego®] (Colletotrichum gloeosporioides (Penz.) Penz. and Sacc. in

Penz. f. sp. aeschynomene), Dr. BioSedge® [Puccinia canaliculata (Schwein.) Lagerh], DeVine®

[Phytophthora palmivora (E.J. Butler) E.J. Butler], Smolder® (Alternaria destruens E.G.

Simmons); in Canada, Mallet WP [formerly BioMal] (Colletotrichum gloeosporioides f. sp.

malvae); in Japan, Camperico™ (Xanthomonas campestris Migula pv. poae); in China, Lubao

(Colletotrichum gloeosporioides f.sp. cuscutae); and the Netherlands, BioChon®

(Chondrostereum purpureum) (Charudattan, 2001) were or currently are registered as

bioherbicides in the USA and Canada. Limited availability of many other bioherbicides can be

ascribed to the continued lack of commercial backing, high costs of mass production, limited

markets, resistant weed biotypes and new chemical herbicide chemistries (Charudattan, 2001).

A major constraint in bioherbicide development is the requirement for appropriate formulations

for successful establishment of the bioherbicide in the field to overcome the dew requirement

that exists for several of them (Auld, 1992; Greaves and McQueen, 1990). Various invert and

vegetable oil emulsion formulations have eliminated or greatly reduced the free moisture requirements and increased efficacy of spore inoculum (Boyette et al., 1999). Environmental conditions are among the primary factors that influence formulation performance of bioherbicides as inoculum production is dependent on sporulation of the formulation. One likely limitation on the practical efficacy of using D. gigantea as a mycoherbicide is the fact it does not readily sporulate in liquid culture (Chandramohan, 1999). Yandoc and Charudattan (1998) were able to culture D. gigantea and B. sacchari readily on several grains. Chandramohan (1999) showed that by using a biphasic culturing system, where the blended fungal cultures, derived from the liquid phase, are spread on trays and allowed to dry gradually. This system allows for multiple spore harvests with up to three years viability when stored at 4ºC in sterile soil or as dry spores. However, spore viability of D. gigantea and B. sacchari on solid substrates was reduced even after a short-term storage. The dew period and temperature requirements for both fungi were determined (Chandramohan, 1999; Yandoc et al., 2005) and these two factors should not be a limiting factor in their practical application as bioherbicides. Chandramohan (1999) showed that a 12 h dew period at 28 ºC was optimal for D. gigantea and within a range that occurs normally under field conditions. Chandramohan et al. (2002) demonstrated that several pathogens could control different weedy grasses under field conditions using an emulsion based inoculum preparation either singly or in combination. Yandoc (2001) showed that the efficacy of D. gigantea and B. sacchari could be improved by formulating their spores in an oil emulsion and that the formulation caused higher levels of disease and damage with as little as a 4 h dew period in the field.

Genetic Variation in Invasive Weeds: The Implication for Biological Control

Knowledge of the taxonomy of a selected target weed is a prerequisite for undertaking extensive and often costly explorations for potential biological control agents in the native range

(Morin et al., 2006). Although host specific biological control agents are most likely to be found

within the native range of the target weed, agents have been used from the introduced range, and

from congeners when host-specific requirements were broad (Van Klinken and Julien, 2003;

Wapshere et al., 1989). Biological control agents are likely to be more diverse and hosts-specific

in the centre of diversity for the genus and these regions should always be considered for initial

exploration (Wapshere et al., 1989; Müller-Schärer et al., 1991). Determining the native host

range starts with an extensive literature search and consultations with herbaria worldwide, with

additional taxonomic and/or molecular work to fine tune the distribution and occurrence of

genotypes (Goolsby et al., 2006). The importance of identifying the origin and/or the genetic

diversity of invasive weed populations has been essential to implementing successful biocontrol

program. Within the following biological control program examples, Salvinia (Forno and

Harley, 1979); Chondrilla (Chaboudez, 1994); Rubus (Evans et al., 1998); Onopordum

(O’Hanlon et al., 2000a), Chromolaena (Ye et al., 2004) and Ligustrum (Milne and Abbott,

2004), all incorporated molecular studies to determine the genetic diversity of related native species or populations to determine the putative centre of origin (Wardill et al., 2005). The genetic variation of invasive plant populations, through the selection of resistant genotypes and differential efficacy resulting from genotypic variation, can be challenging for weed management strategies (Sterling et al., 2004). Weed populations vary significantly in size and morphological

character expression as a consequence of phenotypic plasticity as individual genotypes respond

to changes in environmental factors by changing their growth and development. Genetic

differentiation, between and within populations, causes variation in weed species and if weed

populations are persistent in an area long enough to enable local environment adaptation, then

genetic differentiation is favored (Barrett, 1982). Increasing evidence suggests that rapid

adaptive evolution of invasive plants is occurring during the process of range expansion (Müller-

Schärer et al., 2004). Whether this evolutionary change affects plant-antagonist interactions will have significant implications for biological control of invasive plant programs. Practical implications of understanding the population genetic structure of a target invasive weed include the reconstruction of the historical process of migration and colonization. Insights into the ecological persistence and evolutionary potential of populations in new habitats ultimately lead to a better understanding of weedy adaptation (Shilling et al., 1997). According to Shilling et al.

(1997), Florida cogongrass is highly outcrossed, as indicated by morphological, RAPDs and mode-of-reproduction studies. However, evidence of clonal populations was also observed.

Cogongrass populations develop greater genetic diversity by reproducing through seed and vegetatively propagating rhizomes, allowing the weed greater adaptive changes to the environment.

Biological Control and Genetic Markers

Molecular genotyping of the target weed using DNA–based molecular markers may provide information that allows biocontrol researchers to pinpoint the origin of an invasive species or population (Nissen et al., 1995). Different DNA-based marker techniques, including chloroplast DNA restriction fragment length polymorphisms (cpDNA RFLP) and random amplified polymorphic DNA (RAPD) analysis, provide estimates of genetic variation, identifying species and other locally adapted forms, and define genetic relationships. These tools are well-established and widely used in weed research (Nissen et al., 1995; O’Hanlon et al.,

2000b). A study by Wardill et al. (2005) used internal transcribed spacer regions (ITS1) and non-coding chloroplast trnL DNA fragments to identify eight of nine described subspecies of

Acacia nilotica (L.) Delile and reported a previously unknown Pakistan genotype. The study verified that the Australian populations were mostly comprised of A. nilotica ssp. indica (Benth.)

Brenan and suggested that surveys for biocontrol agents should concentrate on the Indian

subcontinent, A. nilotica ssp. indica’s native range.

Inter simple sequence repeat (ISSR) techniques employ an approach similar to RAPD

except that ISSR primer sequences are designed from microsatellite regions dispersed throughout

the various nuclear genomes and the annealing temperatures used are higher than those used for

RAPD markers. The ISSR method is an intermediate technique between RAPD-PCR and more

advanced techniques including simple sequence repeats (SSR) and amplified fragment length

polymorphisms (AFLP). Inter simple sequence repeat analysis involves PCR amplification of

genomic DNA using a single primer that targets the repeat, with 1 to 3 bases that anchor the

primer at the 3’ or 5 ‘ end. In addition to freedom from the necessity of obtaining flanking

genomic sequence information, ISSR analysis is technically simpler than many other marker

systems. The method provides highly reproducible results and generates abundant

polymorphisms in many systems. The first studies employing ISSR markers were published in

1994 (Zietkiewicz et al., 1994; Gupta et al., 1994) with the initial studies focusing on cultivated

species and the hyper variable nature of ISSR markers was demonstrated (Wolfe and Liston,

1998). To test the utility of the method in natural populations, Wolfe et al. (1998a, b) reexamined a known hybrid complex of four species of Penstemon for which three other molecular data sets were available. Their results clearly demonstrated the utility of ISSR markers for addressing questions of hybridization and diploid hybrid speciation. Based on the published, unpublished and in-progress studies conducted using ISSR markers, it is clear that

ISSR markers have great potential for studies of natural populations (Wolfe et al., 1998b).

O’Hanlon et al. (2000b) classified and discussed the goals of genetic weed research into

three areas of investigation: (i) patterns of genetic diversity within invading weeds; (ii) the

taxonomic identity of weeds; and (iii) determining the origin(s) of introduced weeds.

Widespread concerns that weed species with higher levels of genetic diversity will exhibit

considerable potential for adaptation; thereby reducing the effectiveness of weed control has led

to the prioritization of determining the extent of variation in weed species (Dekker, 1997). There

is, despite these efforts, a lack of evidence correlating genetic diversity with physiological,

morphological or other ecological adaptations, including adaptations to biocontrol agents

(Chaboudez and Sheppard, 1995).

The objectives of this study were to: 1) determine the genetic variation between

Southeastern USA and African cogongrass using populations markers and DNA sequence data to interpret the relation with respect to implementing biological control strategies using two fungi,

Bipolaris sacchari and Drechslera gigantea, in Benin, West Africa; 2) determine the applicability of using the B. sacchari and D. gigantea associated with cogongrass in Florida and

Benin as bioherbicides on the West African cogongrass by assessing the disease severity of the pathogens, individually and as a mixture, on cogongrass populations from Southeastern USA and

West Africa; 3) assess the effectiveness of biotrophic fungi as potential classical biological control agents on the West African cogongrass; and 4) implement a field trial using an amended oil Colletotrichum caudatum formulation on West African cogongrass.

CHAPTER 2 AN EXAMINATION OF GENETIC DIVERSITY OF Imperata cylindrica BASED ON INTER- SIMPLE SEQUENCE REPEATS AND SEQUENCE ANALYSIS OF INTERNAL TRANSCRIBED AND CHLOROPLAST TRNL-F SPACER DNA

Introduction

The genetic variability of weedy plant populations presents significant challenges for weed management, primarily because this variability may influence the effectiveness of control methods. Generally, exotic invasive plants as colonizing species have fairly low levels of genetic variation due to genetic bottlenecks resulting from a single introduction (Barrett, 1988;

Warwick, 1990). Despite low levels of variability, many exotic weeds colonize new areas extremely well. Multiple founder events and hybridization within the founder populations may lead to increased diversity, and unique genetic combinations could yield novel phenotypes

(Ellstrand and Schierenbeck, 2000). As with many invasive weeds, Imperata cylindrica (L.)

Beauv., has been introduced multiple times in the USA and has been present long enough to have

potentially evolved unique ecotypes. Some of this evolution may create genotypic variability

within a host that can influence the efficacy of management approaches (Sterling et al., 2004).

Understanding the interplay between host and pathogen with respect to pathogen life history and

genetic factors that affect pathogen establishment and spread is central to the choice of biological

control agents (Thrall and Burdon, 2004). Genetic diversity studies should reveal both the

underlying genetic diversity and the divergent evolution of target plant species. Highly diverse

populations might harbor differing levels of resistance to pathogens and ultimately prove more

difficult to control (Charudattan, 2005).

Imperata is a member of the family Poaceae within the subtribe Saccharinae and tribe

Andropogoneae (Panicoideae). There are nine Imperata species with a worldwide distribution in

the warm regions of both hemispheres. The genus is economically significant due to the weedy

characteristics of I. cylindrica [cogongrass] (Gabel, 1982). According to Gabel (1982), Imperata

cylindrica is the most variable species within this genus, with a wide geographic range. As such,

it would be expected to possess a high degree of morphological variability. This variability is

thought to be responsible for the multitude of scientific names for the species. Hubbard (1944)

and Santiago (1980) separated the species into five major varieties based on growth, geographic

origin, morphological characteristics and karyotype. Imperata cylindrica var. major, is native to

Asia, Australia, the Pacific Islands and East Africa and it is characterized by 2n = 20

chromosomes. Imperata cylindrica var. europa, found in North Africa, the Mediterranean, and

east to Afghanistan and is characterized by 2n=40 chromosomes. Imperata cylindrica var.

africana found in Southern, Central and West Africa has a chromosome count of 2n=60.

Imperata cylindrica var. condensata and I. cylindrica var. latifolia are present in Chile, South

America and Northern India respectively but chromosome numbers for these species have not

been reported. Hubbard (1944) and Santiago (1980) also reported differences in floral and leaf

morphology that distinguishes the different varieties within I. cylindrica. However, Gabel

(1982) cited that the variability between the varieties of I. cylindrica was greater than between

the species of Imperata as a whole and does not recognize the varietal distinctions as described

by Hubbard (1944).

In the United States, the potential for genetic variability and the development of adapted

genotypes exist because there were at least two separate cogongrass introductions into the U.S.

Southeast during the last century (Shilling et al., 1997). The variety major was accidentally

introduced into Alabama from Japan in 1912 and purposefully introduced to Mississippi in 1921

from the Philippines as a potential forage crop. Forage trials were conducted in Florida, Texas

(where the planting died out in the first year) and Alabama (Dickens, 1974; Hall, 1983; Tabor,

1949; Tanner and Werner, 1986), leading to the establishment of populations. Adding to the

genetic variability is I. brasiliensis Trin. (Brazilian satintail), which is native to North, Central

and South America and overlaps in its distribution with cogongrass in Florida and possibly elsewhere in the Southeast. The Brazilian satintail and cogongrass are morphologically and genetically very similar, and their hybrids produce fertile offspring (Shilling et al., 1997).

Hybridization, introgression, and overlapping morphological characters often cause taxonomic confusion between the two species, especially in North America. Both are both nonnative, rhizomatous perennial grasses in the USA, similar in appearance and are primarily distinguished by stamen numbers. The Brazilian satintail usually has one stamen per flower while cogongrass has two stamens per flower (Gabel, 1982; Hubbard, 1944). Other distinguishing characteristics include Brazilian satintail's relatively shorter spikelets (<3.5 mm) and narrower culm leaves (<5 mm) compared to cogongrass (Allen and Thomas, 1991). These characteristics overlap (Gabel,

1982; Lippincott, 1997) and it is likely that the two grasses have been misidentified in the

Southeast (Lippincott, 1997). Identification is further confounded by Brazilian satintail × cogongrass hybridization in the Southeast, the extent of which is unknown (Willard et al., 1988).

Some systematists consider the two species synonymous. Hall (1978) suggests that Brazilian satintail be classified as an infrataxon within I. cylindrica, while Gabel (1982) separates the taxa as two distinct species based upon continents of origin and morphological, cytological and genetic attributes. At present, it is unlikely that I. cylindrica will spread outside of the USA Gulf

Coast States due to the plant’s lack of low-temperature tolerance. An ornamental variety I. cylindrica var. rubra (Red Baron/Japanese Blood Grass) is being sold in nurseries in the United

States, which is more cold tolerant than var. major (Wilcut et al., 1988). If this variety were introduced into the southeastern USA and it hybridizes with var. major, the resulting offspring

may be more cold hardy and potentially extend its invasive range north and westward (Shilling et

al., 1997).

Molecular characterization could potentially fill in the taxonomic gaps that exist within I.

cylindrica. Within cogongrass populations in Florida, RAPD analyses, coupled with

morphological characterization studies, showed that cogongrass was highly outcrossed but clonal

populations were also observed. This suggested that cogongrass spread in Florida was through

both seeds and rhizomes (Shilling et al., 1997). Using RAPD analyses, Cheng and Chou (1997)

identified distinct ecotypes within six cogongrass sites selected from diverse habitats in Taiwan.

The I. cylindrica phenotypes varied with habitat and environment, being affected by soil conditions, temperature, and other climatic factors. Genetic variation, determined by restriction

fragment-length polymorphisms (RFLPs) of the intergenic spacer region (IGS) of ribosomal

DNA (rDNA) in inter- and intra-specific populations of cogongrass in Taiwan revealed two

major clusters. One population (Chuwei) was found to be distinctly different from the remaining

14 populations studied based on the phylogenetic study (Chou and Tsai, 1999).

DNA sequence analysis is used extensively for phylogenetic studies at many different

taxonomic levels (Chase et al., 1993; Soltis et al., 1999; Salamin et al., 2002). The internal transcribed spacer (ITS) region has often been utilized in phylogenetic studies of flowering plants and grasses (Baldwin et al., 1995; Hodkinson et al., 2002a; Hodkinson et al., 2002b). The chloroplast genome has also been used in plant systematics for phylogeny reconstruction (Souza-

Chies et al., 1997). Some noncoding regions of chloroplast DNA such as the trnL-intron and trnL-F spacer (Taberlet et al., 1991; Chase et al., 2000) have been used for evolutionary studies of closely related taxa.

The objective of my study was to characterize the genetic variation between the

Southeastern USA and West African cogongrass using inter-simple sequence repeat (ISSR)

markers and sequence analysis of the internal transcribed spacer region (ITS) and the noncoding

chloroplast trnL-trnF spacer DNA. The data were interpreted in relation to the implementation

of biological control strategies using the fungi, Bipolaris sacchari (E. Butler) Shoem. and

Drechslera gigantea (Heald and F.A. Wolf) Ito, in Benin West Africa.

Materials and Methods

Population Sampling

Florida Imperata cylindrica accessions. Imperata cylindrica accessions were collected

from 32 counties throughout Florida, USA during summer 2004 (Appendix C). Two sites were

sampled within each county, and the sites were at least 20 miles apart. Five individual plants

were collected from every site; each sampled at least 10 m apart to minimize progeny sampling.

Samples were brought back to Gainesville and established under greenhouse conditions in a

commercial potting mix (Metromix 300; Scott’s Sierra Horticultural Products Co., Marysville,

Ohio).

West African Imperata cylindrica accessions. West African I. cylindrica accessions were

collected from infested cogongrass sites during the summer of 2005 and 2006. Rhizomes were

collected from natural and invasive cogongrass sites in Cameroon, The Guinea, Benin and

Nigeria and transported to the International Institute of Tropical Agriculture (IITA), Benin

experimental station (Appendix D). The rhizomes were cut into 4-cm segments and planted in

plastic tubs containing field soil.

Extraction of Genomic DNA and PCR Amplification of ISSR Population Markers

Genomic DNA was extracted from fresh leaf samples (50 mg) using the DNeasy plant

mini kit according to the manufacturer’s directions (QIAGEN, Chatsworth, California).

Inter Simple Sequence Repeats. Three ISSR primers used in the study, UBC-825

([AC]8T), UBC-830 ([TG]8G) and UBC-841 ([GA]8YC; Y = C or T), were obtained from the

University of British Columbia (UBC). The PCR amplifications were carried out in volumes of

25 µl containing 18 µl of ddH2O, 2.5 µl of 10X buffer (Invitrogen), 1.5 µl of 50 mM MgCl2

(Invitrogen, Carlsbad, California), 0.5 µl of an equal mixture of 10 mM dNTPs (Invitrogen), 1 unit of Taq DNA polymerase, and 2.0 µl template DNA. DNA amplification was performed in an MJ Research PTC-100 thermocycler (Watertown, Massachusetts) and the initial denaturation step was setup to run for 1.5 min at 94 ºC, followed by 34 cycles of 40 s at 94 ºC, 45 s at 48 ºC

(UBC-825), and 50 ºC (UBC-830; UBC- 841), 40 s at 72 ºC, 45 s at 94 ºC, 45 s at 47 ºC (UBC-

825) and 49 ºC (UBC-830 and UBC-841) (modified from Wolfe et al., 1998b). Different annealing temperatures were obtained for optimization of the West and South African I. cylindrica accessions with 45 s at 51.7 ºC (UBC-825), 52.6 ºC (UBC-830), and 48.4 ºC (UBC-

841). The final extension step was run for 5 min at 72 ºC. PCR products were seperated by electrophoresis at 70V for 2.5 h in 1.2% agarose gels prepared in 1X TAE buffer stained with

2% ethidium bromide. Band sizes were estimated using 100bp ladder (Invitrogen; Biolabs).

Negative controls, lacking template DNA, were included in each PCR reaction and all sample/primer combinations were repeated at least once. Replicate DNA extractions were carried out on a subset of the samples and the arbitrary ‘fingerprint’ of these was compared with the original. In all cases, an identical profile was obtained. In studies using ISSR primers in natural populations (Wolfe et al., 1998a, b), three primers were generally sufficient to genotype every individual in the study. The ISSR primer sequences, obtained from the University of

British Columbia Biotechnology Laboratory, were selected as they yielded polymorphisms in a

preliminary genetic diversity study (The Soltis Lab, Florida Museum of Natural History,

Dickinson Hall, University of Florida, Gainesville, Florida).

Bands were scored for presence or absence for each DNA sample (See Figure 2-1 A, B for

examples of gel photographs). The binary data matrix was inputted into POPGENE (Yeh et al.,

1997), assuming all loci were dominant and in Hardy-Weinberg equilibrium. The following

indices were used to quantify the amount of genetic diversity within the Imperata populations examined: the number and percentage of polymorphic loci, Nei’s genetic identity (h) (Nei,

1973), and Shannon index of phenotypic diversity (I) (King and Schaal, 1989). Genetic differentiation among populations was estimated by Nei’s gene diversity statistics. Pairwise genetic distance among populations was calculated to construct an unweighted pair-group method arithmetic average (UPGMA) dendrogram. An analysis of molecular variance

(AMOVA; Excoffier et al., 1992) computed with Arlequin 1.1. (Schneider et al., 1997), which

assumes these data are haplotypic, was conducted to calculate variance components and their

significance levels for variation between cogongrass accessions from Africa and the USA,

among accessions within a region, and within populations.

DNA sequence analysis. The ITS region was amplified by PCR using a forward primer

(F=CTAGGGCGTCAAGGAACACTTCTATTGC) and the reverse primer

(R=CGATGCGCTGCGGTGCTCGATGGGTCCT) designed from the I. cylindrica genotype

122 [AY116297] (Hodkinson et al., 2002b). The plastid trnL intron and the trnL-F intergenic

spacer for noncoding chloroplast DNA were amplified as one unit using the Universal C and F

primers described by Taberlet et al. (1991). For ITS, the thermal cycling (MJ Research PTC-

100, Watertown, Massachusetts) comprised 30 cycles, each with 3 min denaturation at 94ºC, 30 s

annealing at 57ºC, and an extension of 45 s at 72ºC. A final extension of 5 min at 72ºC was also

included. For trnL-F, the thermal cycling comprised 30 cycles, each with 1 min denaturation at

94ºC, 1 min annealing at 53ºC and 55 ºC for the USA and West Africa samples respectively, and

an extension of 3 min at 72ºC, including a final extension of 5 min at 72ºC. Amplified DNA

fragments were purified using Qiagen PCR purification kit.

Nucleotide sequencing of the DNA samples was carried out at the University of Florida

DNA Sequencing Core Laboratory using ABI Prism BigDye Terminator cycle sequencing

protocols (part number 4303153) developed by Applied Biosystems (Perkin-Elmer Corp., Foster

City, CA). The excess dye-labeled terminators were removed using MultiScreen® 96-well

filtration system (Millipore, Bedford, MA, USA). The purified extension products were dried in

SpeedVac® ((ThermoSavant, Holbrook, NY, USA) and then suspended in Hi-diformamide.

Sequencing reactions were performed using POP-7 sieving matrix on 50-cm capillaries in an

ABI Prism® 3130 Genetic Analyzer (Applied Biosystems, Foster City, CA, USA) and were

analyzed by ABI Sequencing Analysis software v. 5.2 and KB Basecaller.

Forward and reverse DNA sequences were aligned with CLUSTAL X (Thompson et al.,

1997) with subsequent manual correction using MacClade 4 (Maddison and Maddison, 2000).

Gaps were coded as missing data. Phylogenetic analysis was by a distance method (neighbor

joining) and parsimony (strict consensus). Support for internal branches was assessed by using

1,000 bootstrapped data sets but saving no more than 10 trees per replicate to reduce time spent

swapping on large numbers of trees (Felsenstein, 1985). Parsimony and distance analyses were

conducted with the computer software program PAUP [PAUP: Phylogenetic Analysis Using

Parsimony, version 4.0b8] (Swofford, 1998). Heuristic searches included 1,000 replicates of step-wise addition sequence with tree-bisection reconnection (TBR). Four supplementary I. cylindrica accessions [I. cylindrica var. major AF345653, AF345653; AY116297; AF190764]

were included in the analyses after a BLAST search was performed (Altschul et al., 1997).

Erianthus rockii Keng (AY345217), Miscanthus sinensis Andersson (EF211950; EF211950),

Saccharum arundinaceum Retz. (AY116295), Pennisetum purpureum Schumach. (AF325232),

Setaria parviflora (Poir.) Kerguélen (AF109831) and Echinochloa crus-galli (L.) Beauv.

(AJ113708) were chosen as outgroups for the analyses from GenBank (as described by

Hodkinson et al., 2002b).

Voucher specimens of the USA and African cogongrass accessions treated in the ITS study

are listed in Appendix A and Appendix B.

Results

Analysis of ISSR Markers

The three primers, UBC-825 ([AC]8T), UBC-830 ([TG]8G) and UBC-841 ([GA]8YC; Y =

C or T), produced multiple, clear and reproducible bands with heterogeneity within/between/among populations (Figure 2-1 A, B). The three primers yielded 36 putative loci among all I. cylindrica accessions and the three closely related species, I. brasiliensis, I. brevifolia Vasey and I. cylindrica var. rubra. At the species level, the Imperata cylindrica from

North West Florida had the highest number of polymorphic loci, 19 (h = 0.1702, I = 0.2594).

Imperata cylindrica from Benin and Nigeria followed closely with 18 polymorphic loci (Benin: h = 0.1667, I = 0.2524; Nigeria: h = 0.1535; I = 0.2388), respectively (Table 2-1). While I. cylindrica var. rubra had only 4 polymorphic loci (h = 0.0448, I = 0.0656), I. brasiliensis, I. brevifolia and I. cylindrica (The Guinea) showed no polymorphisms and this is possibly due to the small sample size (1- 4 samples) analyzed. At the population level, the percentage of polymorphic loci (P), Nei’s gene diversity (h), and Shannon’s index (I) were 94.44%, 0.2662 and

0.4121, respectively (Table 2-1). The coefficient of gene differentiation was high (GST =

0.6542). All three parameters indicated that genetic diversity within populations was

comparatively equal. When populations were grouped according to geographic region, genetic diversity of the African Imperata cylindrica was higher than the USA I. cylindrica and closely related Imperata species as revealed by the percentage polymorphic loci and total gene diversity

(Table 2-2). The African populations exhibited higher genetic differentiation (GST = 0.5683) than those in USA (GST = 0.3842) and closely related Imperata species (GST = 0.5777).

A UPGMA dendrogram based on pairwise genetic differences identified three strongly supported geographic clusters [80%] (Figure 2-2) was generated. However, not all the African accessions grouped together within the African cluster. The Nigerian accessions, with the exception of accession 11, form a separate, strongly supported cluster (82%) and are genetically different from all the Imperata accessions. One var. rubra accession (CGS66_1) is included within the Nigerian cluster. It is interesting to note that a second var. rubra accession

(CGS66_2) is closely related to the I. cylindrica Florida accessions. The Guinea accession fell within the I. cylindrica Florida accessions, forming a sister group with the I. cylindrica Florida accessions, CGS52 and CGS64. The I. cylindrica Florida accessions grouped according to the area from which they were sampled. The Imperata brevifolia formed a sister group with the

African accessions. Imperata brasiliensis groups together with the North West Florida accessions.

AMOVA analysis indicated that more than a third of the total variation (33.63%) was accounted for by differentiation among the African and USA Imperata groups, with a further

21.91% accounting for variation among populations within groups, and the remainder (44.45%) was partitioned within the populations (Table 2-3).

Analysis of ITS

The aligned ITS matrix was 472 base pairs (bp) long yielding 296 constant, 71 uninformative and 105 parsimony-informative characters. Figure 2-3 shows one of 9,840

equally most parsimonious trees for the ITS sequence data with groups consistent in all shortest

trees marked as solid lines. The tree has 276 steps, with a consistency index (CI) of 0.8188 and a

homoplasy index (HI) of 0.1812.

The parsimony tree showed three main clusters, including the majority of I. cylindrica accessions sampled from the USA and Africa (Figure 2-3). There was strong support for the grouping (100%) containing the Florida I. cylindrica accessions (CGS27, CGS33, CGS21,

CGS38, CGS5, CGS41, CGS29, CGS56, CGS1, CGS20, CGS18, CGS19, CGS26, CGS30,

CGS35, CGS36, CGS37) including the I. brasiliensis accession. There was another strong grouping (93% bootstrap support) for the remaining Florida I. cylindrica accessions (CGS42,

CGS52, CGS43, CGS48, CGS51, CGS53). There was no partitioning of accessions according to the regions sampled within Florida (Appendix B). In addition, there was medium support (72%) for the group containing all the West and South African accessions with the exception of a single accession from Nigeria (CGN1) and Cameroon (CGC6) that were more closely associated with the first group of Florida I. cylindrica accessions. There was strong support for the two I. cylindrica var. rubra accessions (89%) while the I. brevifolia was strongly grouped with Florida

I. cylindrica accession CGS63 (100% bootstrap support) and forms a sister group to the I. cylindrica var. rubra.

The neighbor-joining tree (NJ) showed a more branched topology with different groupings

than those inferred from the parsimony tree (Figure 2-4). Compared with the parsimony tree, there was a strong support (98%) for the group containing the Florida I. cylindrica accessions

CGS21, CGS5, CGS33, CGS38 and CGS41 and low support (67%) for the remaining accessions

in the clade (CGS29, CGS56, CGS18, CGS30, CGS1, CGS20, CGS37, CGS19, CGS26, CGS36

and CGS35). As with the parsimony tree, there was strong support (88%) for the group

containing the I. cylindrica Florida accessions CGS42, CGS52, CGS43, CGS48, CGS53 and

CGS51and the distance analyses supported CGS42, CGS52, CGS43 and CGS48, CGS53,

CGS51 as sister groups. The NJ tree grouped the West and South African I. cylindrica

accessions together (98% bootstrap support) with the exception of single accessions from Nigeria

(CGN1) and Cameroon (CGC6) as seen in the parsimony tree. However, one Benin accession

(CGB2) was not included within the West/South Africa group and was more closely associated

with the Florida I. cylindrica. There was strong support for the two I. cylindrica var. rubra

accessions (86%) while the I. brevifolia was strongly grouped with Florida I. cylindrica

accession CGS63 (94%) and is a sister group to the var. rubra.

The supplementary GenBank I. cylindrica sequences grouped within the I. cylindrica

accessions. The one exception was I. cylindrica AF190764 which grouped within outgroups. A possible reason for this is that the inputted sequence contained poor quality sequence information.

Analysis of trnL-F

The aligned trnL-F matrix was 816 bp and yielded 796 constant, 11 uninformative and 9

parsimony-informative characters of which 5 were found in the two outgroups. Due to the low

variability of the locus, no additional analyses were undertaken (data not shown).

Discussion

In this study, the genetic variation in Imperata cylindrica using polymorphism among the

ISSR regions within the genome was assessed. The ISSR method proved to be robust and

reproducible and provided a relatively simple method of examining genetic variation with the

plant accessions.

Accurate estimates of genetic diversity within exotic invasive weeds are useful for

optimizing the application of both classical and inundative fungal biological control agents.

Several population-level studies of invasive weed species have used ISSR markers. In Phalaris

minor Retz., a diploid, predominantly inbreeding, annual grass weed found in India, relatively

low levels of polymorphisms were found within and between populations in comparison with

other grassy weed species (McRoberts et al., 1999). These findinga are indicative of the

predominantly inbreeding behavior of P. minor (McRoberts et al., 2005).

Carthamus lanatus L. (saffron thistle), a widespread, invasive weed of crops and pastures in Australia, is targeted for both inundative and classical biological control (Ash et al., 2003).

They identified two distinct groups of C. lanatus that correlated with location (northern and southern regions). Within-group genetic diversity was lower in the northern group than the southern group and the difference in these groups might have implications for the biological control using exotic fungi.

Knowledge of the genetic variation in both native and invasive stands of Pueraria lobata

(Willd.) Ohwi [kudzu] could potentially lead to effective biological control measures (Sun et al.,

2005). ISSR studies of USA and Chinese populations of P. lobata and four closely related

cogeneric species, P. edulis Pampanini, P. montana (Lour.) Merr., P. phaseoloides (Roxb.)

Benth. and P. thomsoni Benth. revealed clear separation of the five species. Pueraria lobata

populations from the USA were genetically closer to the Chinese P. lobata populations than the

other congeneric populations. High genetic differentiation was observed in the Chinese samples

of P. lobata, P. montana, and P. thomsoni, while within the USA P. lobata population, a high

genetic diversity and low population differentiation was observed. Sun et al. (2005) stated that

these results supported the hypothesis of multiple introductions into the USA from different

sources in Japan or China, followed by subsequent gene exchange and recombination. The

results would suggest that the USA P. lobata population may not respond uniformly to classical

biocontrol agents.

A study of the genetic variation of Chromolaena odorata (L.) R.M. King & H. Robinson

across its invasive range in China using ISSR markers revealed low levels of variation with low

genetic population differentiation (Ye et al., 2004). The genetically homogeneous C. odorata

populations in China suggest that if a suitable biological control agent were effective for control,

there could potentially be a uniform host response to this biocontrol agent.

In the present study, all three genetic diversity indices (P = 94.44%, h = 0.2662 and I =

0.4121) revealed that the genetic diversity of I. cylindrica is high. Genetic diversity within the

USA accessions was found to be higher than within the African accessions and this outcome

could be due to ecogeographic variation since the USA accesseions were located at different

latitudes and subtropical regions than the African accessions.

A high GST value (0.6542) indicated distinct genetic differentiation among the Imperata accessions studied. Forty-four percent of the genetic variation in the samples can be attributed to variation within populations (i.e., accessions). This differs from a study of RAPD and ISSR techniques to detect genetic diversity of Monochoria vaginalis (Burman f.) C. Presl, a common aquatic weed found in rice fields in southern China (Li et al., 2005). Despite slightly different results between the RAPD and ISSR data, the study indicated low levels of genetic variation within populations (23.30%) and a high degree of genetic variation (76.60%) among populations of M. vaginalis in southern China. In a recent study of the population genetic variation and structure of the invasive weed Mikania micrantha H.B.K., high levels of genetic variation and differentiation were detected in the introduced M. micrantha populations (Wang et al., 2007).

Although multiple introductions mitigated the loss of genetic variation associated with

bottlenecks, the authors found that the bottlenecks enhanced the population differentiation. An

AMOVA analysis indicated that 67.74% of the variation was partitioned within populations, with

28.83% attributed to differences among populations within regions, and 8.43% of the variation due to regional differences.

An AMOVA analysis revealed that differentiation between the African and USA

accessions was 33.63% of the total gene variance while 44.45% of the variation was attributed to variation within the accessions. This differentiation could possibly be explained by the geographical isolation between the USA and Africa. Compared with population genetic estimates based on ISSR analyses of other invasive weed species (Ye et al., 2004; Chapman et al., 2004; Sun et al., 2005), substantial amounts of intraspecific and within-population variation was revealed across the introduced range of I. cylindrica. By transforming among-population variation from different geographical sources into within-population variation, high genetic variation in introduced populations could be created through multiple introductions (Kolbe et al.,

2004). In this study, multiple introductions are inferred among the populations of I. cylindrica and the results substantiate the fact that multiple introductions played a role in the spread of cogongrass in the USA and possibly West Africa. Imperata cylindrica accessions that are genetically more similar appear to maintain an equivalent level of genetic variation (Figure 2-2;

Table 2-1). Imperata cylindrica propagates by seeds and vegetatively through rhizomes. The

high genetic variation within populations and lower genetic differentiation among populations

suggests that the I. cylindrica accessions sampled are outcrossing with the I. cylindrica Florida

accessions seemingly more outcrossed compared to the African accessions. RAPD analyses of

10 Florida I. cylindrica populations revealed that cogongrass was highly outcrossed (Shilling et

al., 1997) though the data also showed evidence of clonal populations.

The results from this study indicats that the I. cylindrica accessions and the closely related

species I examined formed three distinct clusters, with the African and USA I. cylindrica

accessions geographically and genetically separated. Imperata cylindrica var. rubra formed a

sister group to the African accessions. The Imperata brasiliensis was not genetically distinct

from the I. cylindrica Florida accessions and are closely related.

The ITS phylogenetic analyses generated different tree topologies, with the African I.

cylindrica accessions distinct from the USA Imperata accessions. Within the Florida I.

cylindrica, the accessions sampled from West and East Florida did not group together. The I.

brasiliensis grouped within the USA I. cylindrica accessions. However, the I. brevifolia was not

genetically distinct from all Imperata populations as seen in the ISSR dendrogram and formed a

strong group with a single Florida I. cylindrica accession (CGS63). Imperata cylindrica var. rubra grouped together with and was genetically more similar to the African accessions, as reflected in the ISSR data.

In summary, I. cylindrica maintains high levels of diversity throughout the USA and West

Africa. Multiple introductions, sexual reproduction allowing gene flow within populations, and

possible multiple founding populations by more than one genetically distinct individual are

factors that influence the high levels of diversity. The population structure of I. cylindrica

revealed in this study has implications for cogongrass biological control management strategies

in USA and West Africa. The genetic difference between the USA and African I. cylindrica accessions suggests that using West African accession would not be a suitable host for fungi associated with cogongrass in the USA. It is interesting to note that the levels of disease incidence and disease severity of the Florida isolates of B. sacchari and D. gigantea on the Benin

I. cylindrica accession were not significantly different when compared with the Benin isolates of

the same fungi. It is probable that, under an inundative biocontrol strategy, high inoculum levels

would negate any differences in host response. It is also important that multiple accessions of I.

cylindrica be included for pathogencity screening with other candidate fungal biological control agents.

Understanding weed population genetics can help biocontrol researchers determine

whether different genotypes exist within the target weeds exotic and native ranges. Variation in

response to disease within and among populations may be due to a number of causes: a) genetic

variation among plants for traits influencing their susceptibility to attack; b) local variation in

density or in genotype of pathogens; and c) local variation in the abiotic and biotic environment

factors (de Nooij and van Damme, 1988). Most weed species grow over a broad range of

environmental conditions and display large differences in size and morphology (Barrett, 1982).

Unfortunately, management aimed at the control of weeds is often developed without

consideration of the genetic diversity of the target populations, and studies on the genetic

variation of weeds on relevant spatial scales are still limited. The work reported here should

guide us to consider the level of genetic variation of selected weed targets, in conjunction with

other biotic and abiotic environmental conditions, when considering the implemention of

biological control programs.

Table 2-1. Genetic variability of all single sampled populations of Imperata cylindrica and closely related species with the ISSR primers. No. of Shannon Accessions polymorphic loci P (%)a Nei’s diversity index 36 loci total (h) (I) South Africa 14 38.89 0.1454 0.2137 Benin 18 50.00 0.1667 0.2524 Cameroon 8 22.22 0.0899 0.1305 Nigeria 18 50.00 0.1535 0.2388 Central Florida 14 38.89 0.1325 0.1993 Central East Florida 9 25.00 0.1114 0.1587 Central West Florida 10 27.78 0.1141 0.1659 North Central Florida 15 41.67 0.1347 0.2056 North East Florida 12 33.33 0.1124 0.1695 North West Florida 19 52.78 0.1702 0.2594 I. cylindrica var. rubra 4 11.11 0.0448 0.0656 Total 34 94.44 0.2672 0.4159 Note: a Percentage polymorphic loci.

Table 2-2. Comparison of genetic diversity measures for all sampled populations of Imperata cylindrica and closely related species as determined using ISSR population markers. Parameters P (%)a Total gene Within Coefficient of

diversity (Ht) population (HS) gene differentiation (GST)

Total 94.44 0.2784 ± 0.0226 0.0963 ± 0.0037 0.6542

Africa 83.33 0.2574 ± 0.0378 0.1111 ± 0.0099 0.5683

USA 58.33 0.2017 ± 0.0438 0.1246 ± 0.0171 0.3824 USA and related species 72.22 0.2085 ± 0.0341 0.0880 ± 0.0080 0.5777 a Percentage polymorphic loci.

Table 2-3. Summary of analysis of molecular variance (AMOVA) based on ISSR genotypes of the African and USA Imperata cylindrica accessions. Variance Percentage Source of variation SSb components variation Among groups 132.839 0.62830 33.63 Among populations within groups 40.0170 0.40996 21.91 Within populations 287.165 1.25219 45.45 Total 460.021 a DF = Degrees of freedom, b SS = Sum of squares,

Central Florida accessions North West Florida Accessions A M 5 25 26 27 28 29 30 31 32 8 9 11 12 13 14 15 16 17 18 58 59 M

1500

1000

500

M 1 2 3 4 5 6 7 8 1 234 5678 1 234 5 6 7 8 B

1500 1000

500

Figure 2-1. Photographs of ISSR gels. A) Imperata cylindrica Florida accessions showing bands generated with primer UBC-841. B) I. cylindrica Benin accessions showing bands generated with all primers individually. Each lane is the result of an independent PCR reaction. M is a 100-bp ladder.

I. cylindrica South Africa (CGSA1) I. cylindrica Nigeria (CGN2) I. cylindrica Benin (CGB1) I. cylindrica Benin (CGB2) I. cylindrica Benin (CGB5) I. cylindrica Benin (CGB6) 75 I. cylindrica Cameroon (CGC1) I. cylindrica Cameroon (CGC2) I. cylindrica Cameroon (CGC3) 78 I. cylindrica Cameroon (CGC4) I. cylindrica Benin (CGB3) 79 I. cylindrica Benin (CGB4) 77 I. cylindrica Nigeria (CGN11) I. brevifolia California (CGS6) I. cylindrica North East Florida (CGS61) I. cylindrica The Guinea (CGG1) I. cylindrica Central Florida (CGS64) I. cylindrica Central East Florida (CGS52) I. cylindrica Central Florida (CGS2) I. cylindrica Central Florida (CGS32) I. cylindrica Central Florida (CGS36) I. cylindrica Central East Florida (CGS51) I. cylindrica North Central Florida (CGS24) I. cylindrica North Central Florida (CGS35) I. cylindrica Central Florida (CGS33) I. cylindrica Central Florida (CGS55) I. cylindrica Central Florida (CGS54) I. cylindrica Central Florida (CGS56) I. cylindrica Central Florida (CGS34) I. cylindrica Central East Florida (CGS47) I. cylindrica Central Florida (CGS53) I. cylindrica North Central Florida (CGS22) I. cylindrica North Central Florida (CGS23) I. cylindrica North Central Florida (CGS47) I. cylindrica North Central Florida (CGS4) I. cylindrica North Central Florida (CGS19) 80 I. cylindrica North Central Florida (CGS21) I. cylindrica North East Florida (CGS41) I. cylindrica North East Florida (CGS42) I. cylindrica North East Florida (CGS61) I. cylindrica North East Florida (CGGS43) I. cylindrica Central Florida (CGGS64) I. cylindrica North Central Florida (CGS60) I. cylindrica Central East Florida (CGS45) I. cylindrica Central East Florida (CGS49) 71 I. cylindrica Central East Florida (CGS46) I. cylindrica Central East Florida (CGS50) I. cylindrica Central East Florida (CGS48) I. cylindrica Central Florida (CGS63) I. cylindrica North Central Florida (CGS20) 70 I. cylindrica North West Florida (CGS9) I. cylindrica North West Florida (CGS11) 72 I. cylindrica North West Florida (CGS12) I. cylindrica North West Florida (CGS14) I. cylindrica North West Florida (CGS58) I. brasiliensis South East Florida (CGS3) I. cylindrica Central West Florida (CGS5) I. cylindrica Central West Florida (CGS25) I. cylindrica Central West Florida (CGS26) I. cylindrica Central West Florida (CGS27) I. cylindrica Central West Florida (CGS29) 74 I. cylindrica Central West Florida (CGS30) I. cylindrica Central West Florida (CGS31) I. cylindrica Central West Florida (CGS57) I. cylindrica North West Florida (CGS59) I. cylindrica South West Florida (CGS65) I. cylindrica Central West Florida (CGS28) I. cylindrica North West Florida (CGS15) 82 I. cylindrica North West Florida (CGS8) I. cylindrica North West Florida (CGS17) I. cylindrica North West Florida (CGS13) I. cylindrica North West Florida (CGS16) 73 I. cylindrica North West Florida (CGS18) I. cylindrica North East Florida (CGS38) I. cylindrica North East Florida (CGS40) I. cylindrica North East Florida (CGS62) I. cylindrica North East Florida (CGS39) I. cylindrica var. rubra (CGS66_1) I. cylindrica Nigeria (CGN6) 76 I. cylindrica var. rubra (CGS66_2) I. cylindrica Nigeria (CGN8) I. cylindrica Nigeria (CGN9) I. cylindrica Nigeria (CGN10)

Figure 2-2. UPGMA dendrogram of all Imperata cylindrica accessions and closely related species sampled based on ISSR data. Numbers above branches are bootstrap percentages above 70%.

Strict I.CGS27 cylindrica Central West Florida (CGS27) I.CGS33 cylindrica Central Florida (CGS33) I.CGS21 cylindrica North Central Florida (CGS21) I.CGS38 cylindrica North East Florida (CGS38) I.CGS5 cylindrica Central West Florida (CGS5) I.CGS41 cylindrica North East Florida (CGS41) I.CGS29 cylindrica Central West Florida (CGS29) I.CGS56 cylindrica Central Florida (CGS56) I.CGS1 cylindrica Central Florida (CGS1) I.CGS20 cylindrica North Central Florida (CGS20) I.CGS18 cylindrica North West Florida (CGS18) I.CGS19 cylindrica North Central Florida (CGS19) I.CGS26 cylindrica Central West Florida (CGS26) 100 I.CGS30 cylindrica Central West Florida (CGS30)

I.CGS3 brasiliensis South East Florida (CGS3)

I.CGS35 cylindrica North Central Florida (CGS35) 100 I.CGS36 cylindrica Central Florida (CGS36)

I.CGS37 cylindrica North Central Florida (CGS37) I.CGN1 cylindrica Nigeria (CGN1) I.CGC6 cylindrica Cameroon (CGC6) I.CGS42 cylindrica North East Florida (CGS42) I.CGS52 cylindrica Central East Florida (CGS52) 93 I.CGS43 cylindrica North East Florida (CGS43) I.CGS48 cylindrica Central East Florida (CGS48) I.CGS51 cylindrica Central East Florida (CGS51)

I.CGS53 cylindrica Central Florida (CGS53) 96 I.CGS6 brevifolia California I.CGS63 cylindrica Central Florida (CGS63) 100 I.AF092512 cylindrica (AF092512) I.AF345653 cylindrica (AF345653) I.CGB5A cylindrica Benin (CGB5A) I.CGC11 cylindrica Cameroon (CGC11) I.CGB1 cylindrica Benin (CGB1)

I.CGB3A cylindrica Benin (CGB3A) I.CGN2 cylindrica Nigeria (CGN2) I.CGN4 cylindrica Nigeria (CGN4) I.CGB2 cylindrica Benin (CGB2) I.CGC2 cylindrica Cameroon (CGC2) 98 I.CGN3 cylindrica Nigeria (CGN3) I.CGB5 cylindrica Benin (CGB5) I.CGN3A cylindrica Nigeria (CGN3A)

I.CGN4A cylindrica Nigeria (CGN4A) I.CGN2A cylindrica Nigeria (CGN 2A) I.CGS66 cylindrica 1 var. rubra (CGS66) I.CGS66 cylindrica 2 var. rubra (CGS66) I.CGSA1 cylindrica South Africa (CGSA1) I.AY cylindrica 11 62 9 7 (AY116297) Sacc.AY 11 62arundinaceum 9 5.1 (AY116295) 100 98 MiscanthusEF211951.1 sinensis (EF211951) MiscanthusEF211950.1 sinensis (EF211950) ErianthusAF345217 rockii (AY345217) I.AF190764 cylindrica (AF190764) 89 PennisetumAF345232 purpureum (AF345232) SetariaAF019831.1 parviflora (AF019831) EchinochloaAJ133707.1 crus-galli (AJ113708)

Figure 2-3. Parsimony tree inferred from ITS sequence data for all Imperata cylindrica accessions and closely related species. Numbers above branches are bootstrap percentages above 80%. Numbers in parentheses represent the GENBANK accession numbers.

NJ I.CGS21 cylindrica North Central Florida (CGS21) I.CGS5 cylindrica Central West Florida (CGS5) I.CGS33 cylindrica Central Florida (CGS33) I.CGS38 cylindrica North East Florida (CGS38) 98 CGS27I. cylindrica Central West Florida (CGS27) CGS41I. cylindrica North East Florida (CGS41) I.CGS29 cylindrica Central West Florida (CGS29) CGS56I. cylindrica Central Florida (CGS56)

CGS18I. cylindrica North West Florida (CGS18) I.CGS30 cylindrica North West Florida (CGS30) I.CGS1 cylindrica Central Florida (CGS1) CGS20I. cylindrica North Central Florida (CGS20) I.CGS37 cylindrica North Central Florida (CGS37) I.CGS19 cylindrica North Central Florida (CGS19) CGS3I. brasiliensis South East Florida (CGS3) CGS26I. cylindrica Central West Florida (CGS26) 100 I.CGS36 cylindrica Central Florida (CGS36) 100 I.CGS35 cylindrica North Central Florida (CGS36) 97 I.CGN1 cylindrica Nigeria (CGN1) I.CGC6 cylindrica Cameroon (CGC6) I.CGB2 cylindrica Benin (CGB2)

I.CGB3A cylindrica Benin (CGB3A) I.CGN4A cylindrica Nigeria (CGN4A) I.CGN4 cylindrica Nigeria (CGN4) I.CGB5 cylindrica Benin (CGB5) 92 I.CGN3A cylindrica Nigeria (CGN3A) I.CGN2 cylindrica Nigeria (CGN2) I.CGN3 cylindrica Nigeria (CGN2) I.CGC11 cylindrica Cameroon (CGC11)

CGSA1I. cylindrica South Africa (CGSA1) CGN2AI. cylindrica Nigeria (CGN2A) I.CGC2 cylindrica Cameroon (CGC2) I.CGB1 cylindrica Benin (CGB1)

I.CGB5A cylindrica Benin (CGB1) 94 I.CGS6 brevifolia I.CGS63 cylindrica Central Florida (CGS63) I. cylindrica var. rubra (CGS66) 94 CGS66 1 CGS66I. cylindrica 2 var. rubra (CGS66)

I.AF092512 cylindrica (AF092512)

I.AF345653 cylindrica (AF092512) I.CGS42 cylindrica North East Florida (CGS42) I.CGS52 cylindrica Central East Florida (CGS52)

I.CGS43 cylindrica North East Florida (CGS43) 88 I.CGS48 cylindrica Central East Florida (CGS48) I.CGS53 cylindrica Central Florida (CGS53) CGS51I. cylindrica Central East Florida (CGS51) I. AY116297cylindrica (AY116297) 100 Sacc.AY116295.1 arundinaceum (AY116295)

MiscanthusEF211951.1 sinensis (EF211951)

EF211950.1Miscanthus sinensis (EF211950) ErianthusAF345217 rockii (AY345217) I.AF190764 cylindrica (AF190764) 97 PennisetumAF345232 purpureum (AF345232) SetariaAF019831.1 parviflora (AF019831)

EchinochloaAJ133707.1 crus-galli (AJ113708) 0.005 substitutions/site Figure 2-4. Neighbor-joining tree of distances derived from ITS sequence data for all Imperata cylindrica accessions and closely related species. Numbers above branches are bootstrap percentages above 80%. Numbers in parentheses represent the GENBANK accession numbers.

CHAPTER 3 EVALUATION OF FUNGAL PATHOGENS AS BIOLOGICAL CONTROL AGENTS OF Imperata cylindrica (L.) BEAUV. FROM THE USA AND BENIN

Introduction

Biological control is defined as the selective process in which a pathogen or insect is used

against a targeted weed without damaging nontarget species such as native plants and crops

(Ghosheh, 2005). The use of biological control agents in locations requiring low-maintenance and natural areas, where high-input weed control is neither feasible nor practical, is the biocontrol ideal. Classical biological control agents selected should have the biological characteristics to ensure successful establishment and continued development in locations where suppression of the target weed population is desired (Morin et al., 2006).

Ranked as one of the 10 worst invasive weeds in tropical and subtropical regions of the

world (Holm et al., 1977), Imperata cylindrica is recorded as a weed of 35 economically

important crops in 73 countries in the tropics. It infests more than 500 million ha worldwide,

including 200 million ha in Asia and Africa. The potential of biological control using plant

pathogens to manage cogongrass infestations in the USA has been considered (Van Loan et al.,

2002). In the southeastern USA, two fungi Bipolaris sacchari (Butler) Shoem., isolated from

cogongrass, and Drechslera gigantea (Heald & F.A. Wolf) Ito, isolated from large crabgrass

[Digitaria sanguinalis (L.) Scop.], were shown to be pathogenic on a single Florida Imperata

cylindrica accession collected from the Lake Alice area on the University of Florida campus

(Yandoc, 2001). Yandoc et al. (2005) evaluated the efficacy of B. sacchari and D. gigantea in

various formulations in greenhouse and miniplot trials. Amended spore suspensions of B.

sacchari and D. gigantea, containing 105 spores ml-1 in a 1% aqueous gelatin solution, caused

disease symptoms on cogongrass under greenhouse conditions with disease severity ranging

from 42 to 49% (Yandoc et al., 2005). An amended spore in oil emulsion suspension (4%

horticultural oil, 10% light mineral oil, and 86% water) yielded higher levels of disease severity

than spores amended in 1% gelatin. Higher levels of disease severity were achieved when a

mixture of both pathogens were applied under field conditions. Foliar injury ranged from 30 to

86% for B. sacchari and from 9 to 70% for D. gigantea. Despite the promising levels of disease severity and weed mortality recorded for these fungi, the regenerative potential of the cogongrass rhizomes allowed the plants to outgrow the effects of these pathogens.

In West Africa, cogongrass is the most dominant and difficult weed to control and thereby it limits yields of maize, cassava, yam, cowpea, groundnut, cotton, sorghum and plantation crops

(Chikoye et al., 1999, 2000, 2001; Ivens, 1980). Several control strategies, including mechanical

removal, herbicide applications, solar drying of rhizomes after cultivation and the use of cover

crops have not been found effective to prevent the unrelenting spread of cogongrass in West

Africa. As a result, increasing land areas have become unsuitable for agricultural cultivation and

thus I. cylindrica is considered the most serious constraint to crop production and poverty

alleviation in West Africa (Holm et al., 1977; Ivens, 1980; Udensi, 1994; Terry et al., 1997).

Bipolaris sacchari and D. gigantea isolates associated with cogongrass in Benin and

Florida were evaluated for pathogenicity and virulence using detached cogongrass leaf material

(Adolphe Avocanh, 2006; unpublished data). Results showed that the Florida isolates of both

fungi were pathogenic and highly virulent on the Benin cogongrass accession while the Benin

isolates were moderately virulent. Weed populations with limited amounts of genetic variation

are likely to be better candidates for biological control agents (Burdon and Marshall, 1981) and

therefore understanding the extent of genetic variation can be of importance in the detection of

ecotypes or races of a species as they may differ considerably in their susceptibility to predators,

parasites, or herbicides. Achieving a precise genetic match between the targeted weed and

pathogen genotype during the selection process is often a necessary criterion for the success of

biological control programs (Charudattan, 2005). Determining whether intraspecific variation in

resistance exists within invasive weed species can be determined by challenging various accessions or genotype(s) of the weed with a range of pathogen strains (Hasan et al., 1995;

Morin and Syrett, 1996; Den Breeÿen and Morris, 2003; Ellison et al., 2004).

With the goal of seeking suitable biological control agent for cogongrass in West Africa

the objectives of this study were to: 1) determine the virulence of fungi, B. sacchari and D.

gigantea, associated with I. cylindrica from the USA, on the USA I. cylindrica as well as I.

brasiliensis, I. brevifolia and I. cylindrica var. rubra; 2) determine the virulence of fungi

associated with I. cylindrica from West Africa and USA by inoculating West African Benin I.

cylindrica with B. sacchari, D. gigantea, and Colletotrichum caudatum isolates; and 3) compile

a virulence ranking system.

Materials and Methods

Propagation of Test Plants

Florida Imperata cylindrica accessions. Cogongrass rhizomes, collected from 32

counties throughout Florida, USA, were brought back to Gainesville and established under

greenhouse conditions in Metromix 300 (Scott’s Sierra Horticultural Products Co., Marysville,

Ohio) and fertilized with 14-14-14 Osmocote (Scott’s Sierra Horticultural Products Co.

Marysville, OH). The rhizomes were cut into 4-cm segments and planted in plastic trays (13 x

17 x 56 cm). After 3–4 wk, rhizomes with healthy shoots were transplanted to 3-inch diameter plastic pots containing Metromix 300.

Benin Imperata cylindrica accessions. Cogongrass rhizomes were collected from infested

cogongrass sites in Benin during the summer of 2005 and 2006. Collections were taken from

natural and invasive cogongrass populations in the field and transported to the experimental field

site at the International Institute of Tropical Agriculture (IITA), Benin. The rhizomes were cut into 4-cm segments and planted in plastic tubs containing field soil. After 4-6 wk, rhizomes with healthy shoots were transplanted to 3-inch diameter plastic pots containing field soil.

Inoculum Production

Florida fungal isolates. Pure cultures of B. sacchari and D. gigantea were obtained from

Dr. R. Charudattan’s collection (Plant Pathology Department, University of Florida, Gainesville,

FL). These pathogens were previously isolated from diseased grasses collected in Florida and their pathogenicity on several weedy grass species was previously evaluated (Chandramohan,

1999; Yandoc, 2001). Inoculum was produced by transferring mycelial blocks, from stored slant cultures, to fresh potato dextrose agar (PDA) plates and incubating the plates at 28 ºC for 8 d.

Spores were harvested by adding 10 ml sterile distilled water to the cultures, gently scraping the surface with a rubber policeman, and transferring the suspension to PDA medium in two liter

Erlenmeyer flasks with resulting spore suspension. The PDA flasks were incubated at 28ºC for

14 d (12h light/12h dark) until the agar was fully colonized.

Benin fungal isolates. West African cultures of B. sacchari, D. gigantea, and C. caudatum (Sacc.) Peck were obtained from Dr. F. Beed’s culture collection (IITA, Cotonou,

Benin). These pathogens were isolated from diseased grasses collected in Benin and their pathogenicity had been previously evaluated on several weedy grass species (Adolphe Avocanh,

2006; unpublished data). Inoculum was produced by transferring mycelial blocks from stored slant cultures to PDA plates and allowing cultures to grow at 25°C (24h/24h) for 8 d. Spores were harvested by adding 10 ml sterile distilled water amended with 0.1% Tween 80 and the spores were released from culture plates by gently scraping the surface with a sterile glass rod.

Tween 80 was used to improve spore adhesion on the leaves during inoculation. The spore concentration for both experiments was determined with the aid of a haemocytometer.

Imperata cylindrica Inoculations

University of Florida, USA. Inoculations were carried out in a greenhouse 6 wk after the rhizomes were planted. Spores suspensions, 1 x 106 ml-1, were applied until run-off with hand-

pumped sprayers (Figure 1-1). All inoculated and nontreated control plants were incubated in a

dew chamber for 24 h (27 ± 1ºC, RH>90%) and then transferred to the greenhouse for

observation.

International Institute of Tropical Agriculture, Benin. Inoculations were carried out in

a shade house 8 wk after the rhizomes were planted. Spore suspensions, 1 x 106 ml-1, were

applied, until run-off, with hand-pumped sprayers. All inoculated and nontreated control plants

were incubated in a dew chamber for 24 h (27 ± 1ºC, RH >90%) and then transferred to the

shade house for observation.

All experiments consisted of 9 replications per treatment with 3 replications for the

nontreated control plants and were repeated. Koch’s postulates were completed by isolating the

fungi from inoculated cogongrass plants and comparing them with the original isolates used in

the inoculation for cultural characteristics and morphology.

Efficacy of Bipolaris sacchari and Drechslera gigantea Applied Singly or as a Mixture

The spore concentrations of B. sacchari, D. gigantea, and their mixture were 1 x 106

spores ml-1respectively. Spores were suspended in 250 ml sterilized distilled water amended

with 0.1% Tween 80. The Tween 80 was used to improve spore adhesion on the leaves during

inoculation. Control plants were sprayed with sterile distilled water containing 0.1% Tween 80.

The efficacy of the two isolates and their mixture was measured in terms of the percentage

disease severity (DS) and rated with the aid of pictorial key for southern corn leaf blight (James,

1971) and gray leaf blotch (Smith, 1989), with 50% as the maximum level of disease.

Assessments were made at 7 and 21 d after inoculation (DAI). The experiment had a completely

randomized design with nine replications per treatment.

Statistical Analysis

All statistical analyses were performed using SAS statistical software (Version 9.1, SAS

Institute, Cary, NC). A repeated-measure analysis (PROC MIXED) was used to determine the

statistical significance of the main effects (treatments and/or accessions) and pairwise

comparisons between levels of significant effects were performed using the least significant

differences. PROC GLM was used to determine the analysis of variance for the virulence and

accession data. Data were not pooled due to the significant interaction within treatment and

accessions. Means and standard errors were calculated using the PROC MEANS procedure and

are reported for each treatment x time combination. All reported differences among treatments

are significant at P < 0.05, unless otherwise stated.

Results

Efficacy of Bipolaris sacchari and Drechslera gigantea Applied Singly or as a Mixture

Florida Imperata cylindrica Accessions and Fungal Isolates

Discrete leaf lesions were visible within 24 h after inoculation with each treatment. At the end of the monitoring period (21 DAI), the infected leaves were chlorotic, the lesions had coalesced and turned necrotic (Figure 3-1). However, no secondary disease spread was observed on the plants in the greenhouse, and the infection was confined to the sprayed foliage. New emerging leaves from inoculated plants were disease-free, thus decreasing the level of DS.

Disease severity means were scored between 0 (= no disease present - no lesions observed) and 3

(= 15 – 30% of leaf blade diseased; many longitudinal lesions on leaf blade) at 7 and 21 DAI.

In the first trial, the fixed-effect analysis showed a significant treatment effect on the

Florida accessions at 7 DAI (F-value = 355.49; Pr>F = <0.0001) and 21 DAI (F-value = 345.34;

Pr>F = <0.0001) for B. sacchari and D. gigantea applied singly or as a mixture. There were

significant differences between the B. sacchari and the D. gigantea applied singly and their mixture at 7 DAI (Table 3-1). However, the mixture treatments with both fungi were more damaging to cogongrass than with the individual treatments. At 21 DAI, there were no significant differences between any of the treatments applied singly or as a mixture (Figure 3-2).

The control treatment was significantly different from the B. sacchari and D. gigantea applied singly or as a mixture at 7 and 21 DAI.

In the second trial, the fixed-effect analysis showed a significant treatment effect on the

Florida accessions treated at 7 DAI (F-value = 109.97; Pr>F = 0.0003) and 21 DAI (F-value =

211.64; Pr>F = <0.0001) for the B. sacchari, D. gigantea, and their mixture. There were no

significant differences between treatments applied singly or as a mixture at 7 and 21 DAI (Table

3-1, Table 3-2). The control treatment was significantly different from the B. sacchari and D.

gigantea applied singly or as a mixture at 7 and 21 DAI, which confirms that both fungi are

virulent pathogens of cogongrass.

Different cogongrass accessions were susceptible to B. sacchari and D. gigantea applied

singly or in a mixture, over two trials at 7 and 21 DAI (Figure 3-3 - Figure 3-8). Trial 1 at 7

DAI, the B. sacchari was most damaging on accession 22 (mean = 2.3333) and least virulent on

accession 32 (mean = 0.6667) [Figure 3-3 A]. Drechslera gigantea was most damaging on

accessions 49, 32 and 39 (mean = 2.0) and least damaging on accession 23 (mean = 0.4444)

[Figure 3-4 A]. The mixture treatment was most damaging on accession 59 (mean = 3.111) and

least virulent on accession 12 (mean = 0.444) [Figure 3-5 A]. Whereas at 21 DAI, the B.

sacchari was most damaging on accession 59 (mean = 2.1667) and least damaging on accessions

12 and 52 (mean = 0.7778) [Figure 3-3 B]. Drechslera gigantea was most damaging on

accession 49 (mean = 2.25) and least damaging on accessions 12, 38, and 57 (mean = 0.6667)

[Figure 3-4 B]. The mixture was most damaging on accession 44 (mean = 2.2222) and least

damaging on accession 12 (mean = 0.3333) [Figure 3-5 B].

For trial two, the B. sacchari was most damaging on accession 46 (mean = 1.6667) and

least damaging on accession 35 (mean = 0.5) [Figure 3-6 A]. Drechslera gigantea was most damaging on accessions 59 and 21 (mean = 1.1111) and not at all damaging on I. brasiliensis, I. brevifolia Vasey and I. cylindrica var. rubra (mean = 0.000) [Figure 3-7 A]. The mixture was most damaging on accession 56 (mean = 2.0) and least damaging on accession 40 (mean = 0.00)

[Figure 3-8 A] at 7 DAI. While at 21 DAI, the B. sacchari was most damaging on accession 53

(mean = 2.0) and least damaging on I. brasiliensis Trin. (mean = 0.3333) [Figure 3-6 B]. The D. gigantea treatment was most damaging on accession 56 (mean = 2.0) and not damaging at all on

I. brasiliensis and I. cylindrica var. rubra (mean = 0.000) [Figure 3-7 B]. The mixture was most damaging on accession 43 (mean = 1.6667) and not damaging at all on I. cylindrica var. rubra and accession 39 (mean = 0.000) [Figure 3-8 B]. Disease severity of the three treatments was observed to be more staggered over all accessions for trial one (Figure 3-3 – Figure 3-5) when compared with the more uniform response to all treatments for trial two (Figure 3-6 – Figure 3-

8). No consistent response to the different treatments was observed over the two trials at 7 or 21

DAI for all accessions treated.

The Florida accessions were grouped according to their statistical variance in disease

severity to B. sacchari and D. gigantea applied singly or in a mixture (Table 3-7, Table 3-8).

Over the two trials, there were no significant differences in treatment effect between the majority

of accessions inoculated at 7 and 21 DAI compared with the control treatment.

Benin Imperata cylindrica Accessions and Fungal Isolates

In contrast to the Florida accessions, discrete leaf lesions became visible after 72 h for each treatment on the Benin cogongrass accession (Figure 3-1 and Figure 3-2). By the end of the

monitoring period, the inoculated plants were slightly chlorotic and lesions were coalescing and

turned necrotic (Figure 3-2).

For the first trial, the fixed-effect analysis showed a significant treatment effect on the

Benin accession treated at 7 DAI (F-value = 2.94; Pr>F = 0.0444) and at 21 DAI (F-value = 5.05;

Pr>F = 0.0059) for the B. sacchari and D. gigantea applied singly or as a mixture. Results of the

first trial showed a low disease incidence score of between 0 (= no disease present - no lesions

observed) and 1 (= 0.1 – 0.5% of leaf blade diseased; only a few longitudinal lesions with or

without chlorotic margins) for all the isolates tested at 7 and 21 DAI. When the data sets were

analyzed separately, there were significant differences between the B. sacchari isolates from

Florida and Benin, the Benin mixture and the controls at 7 DAI (Figure 3-9 A). At 21 DAI, there

was a significant difference between the treatments and control but no significant difference

between any of the treatments (Figure 3-9 C).

Pairwise comparisons between all the treatments showed a significant difference between

the B. sacchari isolates from Benin and Florida and the B. sacchari isolate from Florida and the

mixture from Florida and Benin at 7 DAI (Table 3-5). The Bipolaris sacchari from Florida and

C. caudatum from Benin were significantly different from the control at 7 DAI. At 21 DAI, the

only significant differences were recorded for the control and fungi applied singly while the

Benin mixture was significantly different from the control (Table 3-7).

For the second trial, the fixed-effect analysis showed no significant treatment effect on the

Benin accessions (F-value = 5.01; Pr>F = 0.0059). There was a significant treatment effect at 21

DAI (F-value = 3.2; Pr>F = 0.0304) for the treatment B. sacchari, D. gigantea, and their mixture.

Results for the second trials showed a higher disease incidence score of between 0 and 4 (= 45 -

55% leaf blade diseased; many lesions with some coalescing on leaf blade) at 7 and 21 DAI

compared with the first trial (Figure 3-9 C, Figure 3-9 D). There was a significant treatment

difference between the D. gigantea Benin and D. gigantea Florida isolates, B. sacchari Benin and C. caudatum isolates compared to the control at 7 DAI (Figure 3-9 B). There was no significant difference among treatments and the control at 7 DAI (Figure 3-9 C). At 21 DAI, the

C. caudatum, B. sacchari Florida and D. gigantea Benin isolates were significantly different

from the control (Figure 3-9 D).

Pairwise comparisons for all treatments at 7 DAI showed a significant difference between

the B. sacchari Benin and C. caudatum Benin, B. sacchari Benin and D. gigantea Florida

isolates and between the Benin mixture and C. caudatum Benin, D. gigantea Benin and D.

gigantea Florida isolates (Table 3-6). All treatments, with the exclusion of the B. sacchari Benin

isolate, were significantly different from the control. At 21 DAI, significant differences were

observed between the B. sacchari Florida and the Florida mixture and C. caudatum and the

Florida mixture (Table 3-8). All treatments, with the exclusion of B. sacchari Benin and the

Benin mixture, were significantly different from the control treatment.

Discussion

The biocontrol potential of four isolates of the fungal pathogens B. sacchari and D.

gigantea, applied singly or as a mixture, on different I. cylindrica accessions and closely related

species (I. brasiliensis, I. brevifolia, and I. cylindrica var. rubra) was confirmed in repeated

greenhouse trials on cogongrass accessions in Florida, USA and in shade house trials in Benin,

West Africa. In addition, an indigenous C. caudatum isolate was included in the Benin study.

Gene-for-gene type resistance-pathogenicity relationships are well documented for

pathogen-host interactions in weed systems (Espiau et al., 1998; Ellison et al., 2004). The

assumption that intraspecific variation in host plants pertaining to disease resistance is present in

all weed-pathogen systems is incorrect. Morin et al. (1993) showed that four different, but

closely related Xanthium species were highly susceptible to the same strain of the rust Puccinia

xanthii Schwein. The detection of intraspecific variation in resistance has significant

implications for biological control programs. Firstly, it is possible that more than a single strain

of the pathogen will be required to control all genotypes of the exotic weed as shown with the

rust fungi used against Chondrilla juncea L. [skeletonweed] (Hassan et al., 1995) and Mikania

micrantha Kunth. [mile-a-minute weed] (Ellison et al., 2004), the hemibiotrophic pathogen,

Mycovellosiella lantanae (Chupp) Deighton var. lantanae on Lantana camara (L.) in South

Africa (Den Breeÿen and Morris, 2003; Den Breeÿen, 2004). Additionally, finding the most virulent strain of the pathogen may require exact host-matching in the native range for the genotypes of the exotic weed found in the introduced country (Morin et al., 2006). Pathogenicity testing of strains of the bridal creeper rust, Puccinia myrsiphylli (Thuem.) Winter from two different rainfall regions in South Africa, revealed that Australian bridal creeper [Asparagus asparagoides (L.) Druce] populations were only susceptible to strains from the winter rainfall areas where the Australian bridal creeper originated (Morin and Edwards, 2006). As found with the rust fungi on mile-a-minute weed (Ellison et al., 2004), skeletonweed (Hasan, 1985) and blackberry (Bruzzese and Hasan, 1986), further extensive testing of strains and weed accessions might be required depending on the genetics of the weed and pathogen. While different pathogen genotypes (i.e., races) can vary in their ability to cause disease in a host, it is equally true that different host genotypes can be affected by particular pathogen races (Cousens and

Croft, 2000). Examples of race and host-pathogen specificity in biological control systems include Puccinia chondrillina Bubak & Syd., a rust pathogen introduced to control Chondrilla

juncea in Australia and the rust, Phragmidium violaceum (Schultz) Winter to control Rubus spp.

Only one of the three distinct morphological forms of C. juncea in Australia (wide, medium, and narrow leaved forms) was susceptible to the single pathotype of P. chondrillina released which

resulted in an increase in abundance of the other C. juncea forms. In 1980, a second race that

attacked only the intermediate form was introduced, which subsequently resulted in an increase

in the third morphological form (Thrall and Burdon, 2004). Bruzzese and Hasan (1986) showed

that of 15 isolates of Ph. violaceum (Schultz) Winter tested for host specificity, one that was

most pathogenic on two Rubus spp. was only slightly pathogenic on another Rubus and caused

no symptoms on a fourth species. Biocontrol is therefore more likely to succeed if the pathogen

is relatively nonspecific to host genotypes or if multiple races of pathogen were introduced

together. When considering the release of a classical biological control agent, we should be

concerned with more than the climatic match of the pathogen and its host specifity. Given the

high cost of biocontrol programs, we need to maximize the success rate of releases, and this

should logically involve an understanding of the ecology of the host. The epidemiology of a

pathogen is likely to be intimately related to the ecology of its host in a dynamic weed

population as well as the effect of the ecology of interacting organisms affected by their abiotic

environment (Cousens and Croft, 2000). Genetic diversity and the size, abundance, and spatial

distribution of host plants will most likely affect the pathogen’s ability to reproduce and spread.

Disease severity induced by the pathogen will ultimately affect the host size and vigor. As a

result, disease severity will affect the weed population dynamics and interactions with other

components of the ecosystem.

Chandramohan and Charudattan (1996) developed a multiple-pathogen strategy that

targeted several weedy grass species simultaneously using three or more host-specific pathogens

combined and applied inundatively. The research showed that the pathogen mixture was either superior or comparable to the single pathogen application in biocontrol efficacy (Chandramohan,

1999). However, they concluded that using a pathogen mixture did not necessarily result in significant disease severity on most of the grass species tested. The possibility of a multiple- pathogen strategy for I. cylindrica in West Africa should only be considered if the pathogens effectively affect different tissues of plant such as leaf (Bipolaris, rusts, and others) and seed pathogens (smuts), and fungi associated with the rhizome (Rhizoctonia sp.). All pathogens should be evaluated to determine their overall affect on plant fitness. Applying a single, host- specific agent should be preferable for cogongrass control within field crop situations, although the scope of my cogongrass project was to assess the control of a single target weed with two or more pathogens. However, results from this study indicate that applying the B. sacchari and D. gigantea as a mixture had no enhanced effects on the cogongrass in Florida and Benin. Using oil formulations, such as the oil emulsions used by Yandoc (2001) is one way to increase the level of damage from these pathogens, but the cost of such formulations is likely to be disincentive for their use in West Africa.

Genetic variation is high within the the USA and African I. cylindrica accessions (Chapter

2) with the two sets of accessions genetically distinct. This would suggest that the West African

accessions would be an incompatible host for the USA B. sacchari and D. gigantea isolates. The

indigenous fungus C. caudatum associated with cogongrass in Benin, evaluated in conjunction

with the B. sacchari and D. gigantea, was found to cause similar or greater disease symptoms on the Benin accessions. Although the isolates were able to cause foliar disease, disease severity on

all accessions was consistently lower than observed in previous studies (Yandoc et al., 2005).

The slight decrease in disease severity noted in several accessions at 21 DAI could be due to the

emergence of new healthy leaves and new healthy shoots from the soil. No plant mortality was observed over the entire experimental period.

Tropical African agriculture is dominated by smallholdings. In Nigeria, more than 80% of small-scale farms were found to range from 0.1 to less than 6.0 ha (Olayide, 1980). The situation is similar in other West African countries, where most holdings are less than 2 ha and

96% cover less than 10 ha (Harrison, 1987, cited by Ker, 1995). Small-scale farmers are most affected by cogongrass infestations as they rely on manual weeding which requires massive labor input and is not effective on underground rhizomes (Chikoye et al., 2000). To design, develop, and implement an inundative biocontrol strategy for cogongrass in West Africa that has a high probability of acceptance by users, research programs should be participatory and address the priorities and potential socio-economic constraints of the farmers. Technology transfer, preferably, using locally available and low-cost materials, will play an important if not a pivotal role in the success of the selected mycoherbicides. Small-scale farming communities will be directly involved in the application and evaluation of the mycoherbicide. The practical effectiveness of mycoherbicide applications in West Africa is most affected by the limited availability and the prohibitive costs of materials to produce sufficient spore quantities and limited availability of appropriate spray equipment.

Currently there is no single management strategy that effectively controls cogongrass infestations in a sustainable way throughout the world and management strategies need to incorporate integrated approaches including preventive, cultural, biological, mechanical, and chemical control methods. Evaluating the management requirements for each cogongrass infestation and implementing the most appropriate and therefore the most effective control method would go a long way in addressing the apparent ineffectiveness of current control

methods. The prospect of implementing native fungal biological control agents as part of an integrated management strategy for cogongrass in West Africa appears to be promising.

Investigations into fungal species associated with closely related Imperata species (I. brevifolia and I. brasiliensis) might yield additional, potentially effective biocontrol agents. Undertaking a comprehensive international survey for virulent pathogens should be seriously considered given the apparent lack of effective biocontrol agents to date.

Table 3-1. Pairwise comparison of least square means of disease severity (DS) from different treatments on Florida accessions of Imperata cylindrica at 7 days after inoculation. Treatment 1a Treatment 2a DS Estimate Pr > t Trial 1 B. sacchari D. gigantea 0.1088 0.1555 B. sacchari Mixture -0.4811 0.0003 Mixture D. gigantea 0.5900 0.0001 B. sacchari Control 1.2769 <0.0001 Control D. gigantea -1.1680 <0.0001 Mixture Control 1.7580 <0.0001

Trial 2 B. sacchari D. gigantea 0.04916 0.3038 B. sacchari Mixture 0.03465 0.4596 Mixture D. gigantea 0.01451 0.7492 B. sacchari Control 0.9631 <0.0001 Control D. gigantea -0.9139 <0.0001 Mixture Control 0.9284 <0.0001 Note: All reported differences among treatments are significant at P < 0.05. a Treatments were compared pairwise by Treatment 1 vs. Treatment 2.

Table 3-2. Pairwise comparison of least square means of disease severity (DS) from different treatments on Florida accessions of Imperata cylindrica at 21 days after inoculation. Treatment 1a Treatment 2a DS Estimate Pr > t Trial 1 B. sacchari D. gigantea 0.1387 0.0502 B. sacchari Mixture 0.1325 0.0559 Mixture D. gigantea -0.006197 0.9146 B. sacchari Control 1.3336 <0.0001 Control D. gigantea -1.1949 <0.0001 Mixture Control 1.2011 <0.0001

Trial 2 B. sacchari D. gigantea 0.1248 0.0192 B. sacchari Mixture 0.1847 0.0053 Mixture D. gigantea -0.05994 0.1493 B. sacchari Control 1.0639 <0.0001 Control D. gigantea -0.9392 <0.0001 Mixture Control 0.8792 <0.0001 Note: All reported differences among treatments are significant at P < 0.05. a Treatments were compared pairwise by Treatment 1 vs. Treatment 2.

Table 3-3. Disease severity statistical classes of Bipolaris sacchari, Drechslera gigantea, and their mixture on the Florida Imperata cylindrica accessions at 7 days after inoculation. Accession number B. sacchari D. gigantea Mixture Control Trial 1 8, 14, 16, 39, 40, 52, 54, 61 11, 23, 29, 33, 34, 37, 41, 43, 44, 46, 47, 48 A A A B 56, 58, 59, 60, 65 B B A C 38 A B A C 12, 15, 22 A B B C 49 AB A B C 45, 53, 57 B C A D 32, 51 C B A D

Trial 2 3, 4, 5, 7, 8, 9, 15, 16, 21, 22, 23, 24, 26, 27 29, 30, 31, 32, 33, 34, 37, 38, 41, 43, 47, 48 49, 50, 51, 52, 54, 55, 57, 58, 59, 60, 61 A A A B 56 B B A C 6, 46 A B A C 53 A B B C 66 A A A A 11 A A B C 36, 40, 42 A A B B 65 A AB AB B 35 AB A AB B 25, 45 AB B A C

Note: All reported differences among treatments are significant at P < 0.05. See Figure 3-1 and Figures 3-3 – 3-7.

Table 3-4. Disease severity statistical classes of Bipolaris sacchari, Drechslera gigantea, and their mixture on the Florida Imperata cylindrica accessions at 21 days after inoculation. Accession number B. sacchari D. gigantea Mixture Control Trial 1 11, 15, 16, 22, 29, 32, 37, 39, 41, 43, 46, 48 51, 52, 54, 58, 59, 60, 65 A A A B 44, 57 B B A C 33, 53, 34 A B A C 8, 38, 45 A B B C 14, 40, 47 A A B C 61 A AB B C 12 A AB BC C 49 A B C C 56 B A AB C

Trial 2 5, 6, 7, 8, 9, 11, 12, 13, 15, 16, 21, 23, 24 25, 26, 27, 29, 30, 31, 32, 34, 37, 38, 42, 47 48, 49, 50, 51, 52, 55, 58, 60, 61 A A A B 3 B B A C 22, 46 A B A C 43, 54, 57, 59 A B B C 33 A B AB B 53 A B C D 35 B A B C 4 B A AB C 56 AB A AB B 45 AB B AB C Note: All reported differences among treatments are significant at P < 0.05. See Figure 3-1 and Figures 3-3 – 3-7.

Table 3-5. Pairwise comparison of least square means of disease severity (DS) from different treatments on the Benin accession of Imperata cylindrica at 7 days after inoculation (Trial 1). Treatment 1a Treatment 2a DS Estimate Pr> t B. sacchari (Benin) B. sacchari (Florida) -0.7778 0.0049 B. sacchari (Benin) Mixture (Benin) -0.0556 0.8324 B. sacchari (Benin) Mixture (Florida) -0.2761 0.2329 B. sacchari (Benin) Control (Benin) 0.1111 0.6373 B. sacchari (Benin) C. caudatumb (Benin) -0.4444 0.0756 B. sacchari (Benin) D. gigantea (Benin) -0.3333 0.1713 B. sacchari (Benin) D. gigantea (Florida) -0.3333 0.1713 B. sacchari (Florida) Mixture (Benin) 0.7222 0.0148 B. sacchari (Florida) Mixture (Florida) 0.5016 0.0406 B. sacchari (Florida) Control (Benin) 0.8889 0.0020 B. sacchari (Florida) C. caudatum (Benin) 0.3333 0.1713 B. sacchari (Florida) D. gigantea (Benin) 0.4444 0.0756 B. sacchari (Florida) D. gigantea (Florida) 0.4444 0.0756 Mixture (Benin) Mixture (Florida) -0.2206 0.3916 Mixture (Benin) Control (Benin) 0.1667 0.5285 Mixture (Benin) C. caudatum (Benin) -0.3889 0.1547 Mixture (Benin) D. gigantea (Benin) -0.2778 0.3000 Mixture (Benin) D. gigantea (Florida) -0.2778 0.3000 Mixture (Florida) Control (Benin) 0.3872 0.1028 Mixture (Florida) C. caudatum (Benin) -0.1683 0.4593 Mixture (Florida) D. gigantea (Benin) -0.0572 0.7995 Mixture (Florida) D. gigantea (Florida) -0.0572 0.7995 Control (Benin) C. caudatum (Benin) -0.5556 0.0313 Control (Benin) D. gigantea (Benin) -0.4444 0.0756 Control (Benin) D. gigantea (Florida) -0.4444 0.0756 C. caudatum (Benin) D. gigantea (Benin) 0.1111 0.6373 C. caudatum (Benin) D. gigantea (Florida) 0.1111 0.6373 D. gigantea (Benin) D. gigantea (Florida) 0.0000 1.0000 Note: All reported differences among treatments are significant at P < 0.05. aTreatments were compared pairwise by Treatment 1 vs. Treatment 2. bAn indigenous Colletotrichum caudatum isolate was included in the study.

Table 3-6. Pairwise comparison of least square means of disease severity (DS) from different treatments on the Benin accession of Imperata cylindrica at 7 days after inoculation (Trial 2). Treatment 1a Treatment 2a DS Estimate Pr > t B. sacchari (Benin) B. sacchari (Florida) -1.1095 0.1350 B. sacchari (Benin) Mixture (Benin) 0.00304 0.9967 B. sacchari (Benin) Mixture (Florida) -0.5540 0.4416 B. sacchari (Benin) Control (Benin) 1.0016 0.1741 B. sacchari (Benin) C. caudatumb (Benin) -1.9434 0.0167 B. sacchari (Benin) D. gigantea (Benin) -1.8541 0.0214 B. sacchari (Benin) D. gigantea (Florida) -1.8541 0.0214 B. sacchari (Florida) Mixture (Benin) 1.1125 0.1239 B. sacchari (Florida) Mixture (Florida) 0.5556 0.4159 B. sacchari (Florida) Control (Benin) 2.1111 0.0066 B. sacchari (Florida) C. caudatum (Benin) -0.8339 0.2400 B. sacchari (Florida) D. gigantea (Benin) -0.7446 0.2917 B. sacchari (Florida) D. gigantea (Florida) -0.7446 0.2917 Mixture (Benin) Mixture (Florida) -0.5570 0.4261 Mixture (Benin) Control (Benin) 0.9986 0.1638 Mixture (Benin) C. caudatum (Benin) -1.9465 0.0143 Mixture (Benin) D. gigantea (Benin) -1.8571 0.0183 Mixture (Benin) D. gigantea (Florida) -1.8571 0.0183 Mixture (Florida) Control (Benin) 1.5556 0.0341 Mixture (Florida) C. caudatum (Benin) -1.3895 0.0602 Mixture (Florida) D. gigantea (Benin) -1.3001 0.0764 Mixture (Florida) D. gigantea (Florida) -1.3001 0.0764 Control (Benin) C. caudatum (Benin) -2.9450 0.0007 Control (Benin) D. gigantea (Benin) -2.8557 0.0009 Control (Benin) D. gigantea (Florida) -2.8557 0.0009 C. caudatum (Benin) D. gigantea (Benin) 0.8933 0.8997 C. caudatum (Benin) D. gigantea (Florida) 0.8933 0.8997 D. gigantea (Benin) D. gigantea (Florida) 0.0000 1.0000 Note: All reported differences among treatments are significant at P < 0.05. aTreatments were compared pairwise by Treatment 1 vs. Treatment 2. bAn indigenous Colletotrichum caudatum isolate was included in the study.

Table 3-7. Pairwise comparison of least square means of disease severity (DS) from different treatments on the Benin accession of Imperata cylindrica at 21 days after inoculation (Trial 1). Treatment 1a Treatement 2a DS Estimate Pr > t B. sacchari (Benin) B. sacchari (Florida) 0.08333 0.6334 B. sacchari (Benin) Mixture (Benin) -0.2222 0.2448 B. sacchari (Benin) Mixture (Florida) -0.3333 0.1261 B. sacchari (Benin) Control (Benin) 0.5556 0.0094 B. sacchari (Benin) C. caudatumb (Benin) -0.2222 0.2448 B. sacchari (Benin) D. gigantea (Benin) -0.3333 0.1261 B. sacchari (Benin) D. gigantea (Florida) 0.0000 1.0000 B. sacchari (Florida) Mixture (Benin) -0.3056 0.0966 B. sacchari (Florida) Mixture (Florida) -0.4167 0.0506 B. sacchari (Florida) Control (Benin) 0.4722 0.0160 B. sacchari (Florida) C. caudatum (Benin) -0.3056 0.0966 B. sacchari (Florida) D. gigantea (Benin) -0.4167 0.0297 B. sacchari (Florida) D. gigantea (Florida) -0.0833 0.6334 Mixture (Benin) Mixture (Florida) -0.1111 0.5951 Mixture (Benin) Control (Benin) 0.7778 0.0009 Mixture (Benin) C. caudatum (Benin) 0.0000 1.0000 Mixture (Benin) D. gigantea (Benin) -0.1111 0.5529 Mixture (Benin) D. gigantea (Florida) 0.2222 0.2448 Mixture (Florida) Control (Benin) 0.8889 0.0008 Mixture (Florida) C. caudatum (Benin) 0.1111 0.5951 Mixture (Florida) D. gigantea (Benin) 0.0000 1.0000 Mixture (Florida) D. gigantea (Florida) 0.3333 0.1261 Control (Benin) C. caudatum (Benin) -0.7778 0.0009 Control (Benin) D. gigantea (Benin) -0.8889 0.0003 Control (Benin) D. gigantea (Florida) -0.5556 0.0094 C. caudatum (Benin) D. gigantea (Benin) -0.1111 0.5529 C. caudatum (Benin) D. gigantea (Florida) 0.2222 0.2448 D. gigantea (Benin) D. gigantea (Florida) 0.3333 0.0907 Note: All reported differences among treatments are significant at P < 0.05. aTreatments were compared pairwise by Treatment 1 vs. Treatment 2. bAn indigenous Colletotrichum caudatum isolate was included in the study.

Table 3-8. Pairwise comparison of least square means of disease severity (DS) from different treatments on the Benin accession of Imperata cylindrica at 21 days after inoculation (Trial 2). Treatment 1a Treatmenta DS Estimate Pr > t B. sacchari (Benin) B. sacchari (FL) -1.4390 0.0975 B. sacchari (Benin) Mixture (Benin) -0.0411 0.9611 B. sacchari (Benin) Mixture (Florida) 0.3388 0.6822 B. sacchari (Benin) Control (Benin) 1.4499 0.0952 B. sacchari (Benin) C. caudatumb (Benin) -1.6612 0.0596 B. sacchari (Benin) D. gigantea (Benin) -0.8361 0.3292 B. sacchari (Benin) D. gigantea (Florida) -0.4886 0.5640 B. sacchari (Florida) Mixture (Benin) 1.3979 0.1065 B. sacchari (Florida) Mixture (Florida) 1.7778 0.0418 B. sacchari (Florida) Control (Benin) 2.8889 0.0027 B. sacchari (Florida) C. caudatum (Benin) -0.2222 0.7836 B. sacchari (Florida) D. gigantea (Benin) 0.6028 0.4692 B. sacchari (Florida) D. gigantea (Florida) 0.9503 0.2605 Mixture (Benin) Mixture (Florida) 0.3799 0.6464 Mixture (Benin) Control (Benin) 1.4910 0.0871 Mixture (Benin) C. caudatum (Benin) -1.6201 0.0654 Mixture (Benin) D. gigantea (Benin) -0.7951 0.3523 Mixture (Benin) D. gigantea (Florida) -0.4476 0.5969 Mixture (Florida) Control (Benin) 1.1111 0.1832 Mixture (Florida) C. caudatum (Benin) -2.0000 0.0245 Mixture (Florida) D. gigantea (Benin) -1.1750 0.1691 Mixture (Florida) D. gigantea (Florida) -0.8275 0.3245 Control (Benin) C. caudatum (Benin) -3.1111 0.0015 Control (Benin) D. gigantea (Benin) -2.2861 0.0136 Control (Benin) D. gigantea (Florida) -1.9386 0.0313 C. caudatum (Benin) D. gigantea (Benin) 0.8250 0.3259 C. caudatum (Benin) D. gigantea (Florida) 1.1725 0.1699 D. gigantea (Benin) D. gigantea (Florida) 0.3475 0.6807 Note: All reported differences among treatments are significant at P < 0.05. aTreatments were compared pairwise by Treatment 1 vs. Treatment 2. bAn indigenous Colletotrichum caudatum isolate was included in the study.

A B

C D

Figure 3-1. Disease symptoms and disease severity on Imperata cylindrica Florida accessions at 21 days after inoculation. A) B. sacchari. B) D. gigantea. C) Mixture. D) Control.

A B

C D

E F

Figure 3-2. Disease symptoms and diseases severity on Imperata cylindrica Benin accession at 21 days after inoculation. A) B. sacchari Florida. B) B. sacchari Benin. C) D. gigantea Florida. D) D. gigantea Benin. E) Mixture Florida. F) Mixture Benin. G) Control.

Figure 3-3. Mean disease ratings of Bipolaris sacchari on Imperata cylindrica Florida accessions (Trial 1). A) 7 days after inoculation. B) 21 days after inoculation.

Figure 3-4. Mean disease ratings of Drechslera gigantea on Imperata cylindrica Florida accessions (Trial 1). A) 7 days after inoculation. B) 21 days after inoculation.

Figure 3-5. Mean disease ratings of the Mixture on Imperata cylindrica Florida accessions (Trial 1). A) 7 days after inoculation. B) 21 days after inoculation.

Figure 3-6. Mean disease ratings of Bipolaris sacchari on Imperata cylindrica Florida accessions (Trial 2). A) 7 days after inoculation. B) 21 days after inoculation.

Figure 3-7. Mean disease ratings of Drechslera gigantea on Imperata cylindrica Florida accessions (Trial 2). A) 7 days after inoculation. B) 21 days after inoculation.

Figure 3-8. Mean disease ratings of the Mixture on Imperata cylindrica Florida accessions (Trial 2). A) 7 days after inoculation. B) 21 days after inoculation.

Disease means

Treatments B

Disease means

Treatments

Figure 3-9. Disease severity of Florida (Fl) and Benin (B) isolates of Bipolaris sacchari, Drechslera gigantea, and their mixture on the Benin accession of Imperata cylindrica. Included in the trial was a Colletotrichum caudatum, indigenous to Benin. A) Trial 1 at 7 days after inoculation (DAI). B) Trial 2 at 7 DAI. C) Trial 2 at 21 DAI. D) Trial 2 at 21 DAI. Bars with the same letters are not significantly different.

Disease means

Treatments

Disease means

Treatments Figure 3-9. Continued.

CHAPTER 4 EVALUATION OF OBLIGATE BIOTROPHIC FUNGI ASSOCIATED WITH Imperata cylindrica IN SOUTH AFRICA AS POTENTIAL BIOCONTROL AGENTS FOR COGONGRASS IN WEST AFRICA

Introduction

The classical approach to biological control of weeds involves the introduction,

establishment and self sustenance of herbivores or pathogens from the native range of the target

weed into the area where the weed has naturalized and become a problem (Briese, 2000). The

aim is to achieve a long-term equilibrium between the population of natural enemies and the

weed, and in the process trigger a reduction of the weed population below the economic or

ecological threshold. In a rangeland, pasture, or natural areas, a reduction in weed density will

open up niches for other desirable native or introduced plant species to invade, thereby restoring

a level of balance in the ecosystem. A strategic land management approach, however,

combining classical biocontrol with other techniques such as herbicides, fertilizers, and

revegetation, is recommended to control the whole spectrum of weed species present in the system. Until recently, there has been a dearth of information on the biological control of grassy weeds (Evans, 1991). Information on the natural enemy complex and the effect of this complex

on the population dynamics is largely unknown with a general lack of basic botanical

information on the centers of origin and diversity of a majority of the invasive grasses. This has

resulted in a general absence of attempted, and thus of successful, biological control projects

against grasses (Waterhouse, 1999).

Despite the importance of the problems caused by cogongrass throughout the tropical areas

of the world, biological control efforts have been few and rather piecemeal (Caunter, 1996).

Other complicating factors include existence of closely related grasses of economic and ecologic

value (Holm et al., 1977) and potential conflict of interest with groups that value cogongrass

(Evans, 1991). Similarly, little information exists on the pathogens of cogongrass and their potential as biological control agents (Evans, 1991) and it is likely that fungi associated with cogongrass are more diverse and abundant than indicated by herbarium records (Evans, 1991;

Charudattan, 1997; Minno and Minno, 1998).

The rust fungi are particularly well suited as biocontrol agents as they are typically wind dispersed, may infect directly through the host epidermis or through stomata and do not require wounds, are virulent, and highly host specific (Quimby, 1982). Rust fungi have complex life cycles that vary among genera and species. A complete life cycle (macrocyclic) contains five successive spore stages (basidiospores, pycniospores, aeciospores, urediniospores, and teliospores) but most rusts are microcyclic and lack one or more of the stages. Autoecious rusts are capable of completing their entire life cycle on one host species while heteroecious rusts require two hosts in two different plant families to complete their life cycle (Littlefield, 1981).

Infection is usually local, forming individual colonies in leaves or other aerial parts of the host and dependent on reinfection each year. Infection is sometimes systemic and persistent in the plant.

Smut fungi present a rather uniform life cycle with a saprophytic haploid phase and a parasitic dikaryophase. The haploid phase usually commences with the formation of basidiospores in the basidium and ends with the production of dikaryotic, parasitic hyphae. The dikaryotic phase ends with the production of basidia. The young basidium becomes a thick- walled teliospore and separates at maturity from the sorus and functions as a dispersal unit.

Most of the Ustilaginomycetes are dimorphic, producing a yeast or yeast-like phase in the haploid state (Bauer et al., 1997). Studies have been conducted that aim to develop a classical biological control for itch-grass, Rottboellia cochinchinensis (Lour.) Clayton using the head

smut, Sporisorium ophiuri (P. Henn.) Vánky (Ellison and Evans, 1990), which has recently been

approved for release in Costa Rica. This is the first time that a true smut fungus has been used in

a classical biocontrol program.

Evans (1987, 1991) suggested that some of the known fungal pathogens of cogongrass should be considered for introduction as classical biological control agents on this invasive weed.

Promising species include the rust fungi, Puccinia fragosoana Beltran, P. rufipes Diet. [USDA -

ARS, 2001] and P. imperatae Poirault (Evans, 1987); a smut fungus, Sporisorium

(=Sphacelotheca) schweinfurthianum (Thuem.) Sacc. (Evans, 1987) that is well represented on cogongrass in the Old World, mainly Africa (Evans, 1991) and the hemibiotrophic fungus

Colletotrichum caudatum (Sacc.) Peck (Caunter, 1996). It is interesting to note that the three rust species and smut fungus are common in the Mediterranean region where I. cylindrica is not a serious weed (Evans, 1991; Holm et al., 1977). The aecial stages of P. imperatae and P. fragosoana are unknown while P. rufipes has a known aecial stage on Thunbergia, an alternate host (Cummins, 1971). The presence of an economically important plant as an alternate host potentially excludes this rust from further evaluation as a biocontrol agent due to potential nontarget effects. Besides, Caunter (1996) cited that although P. rufipes was present in

Malaysia, it was having little effect on cogongrass.

The objective of this study were to: 1) determine the efficacy of two rust pathogens,

Puccinia fragosoana and P. imperatae and a head smut fungus, Sporisorium schweinfurthianum

(associated with the South African Imperata cylindrica) on the West African Benin I. cylindrica, and 2) determine the inoculation method required for successful infection of the biotrophic fungi on the West African Benin I. cylindrica.

Materials and Methods

Propagation of Test Plants

Viable Imperata cylindrica seeds were germinated by placing on water-soaked cotton wool in petri dishes and leaving the dishes in a shade house for 14 to 21 d. Germinated seedlings were subsequently replanted in 8-cm-diameter pots containing sterilized field soil and allowed to establish for 25 d (Figure 4-1).

Source of Pathogens and Inoculum

Collections of I. cylindrica leaves infected with the rusts, P. fragosoana and P. imperatae and inflorescences infected with the head smut, S. schweinfurthianum, all from South Africa, were stored at 4°C until required. The geographic source of these pathogens are given in Table

4-1.

Puccinia fragosoana and P. imperatae. For the first inoculation, uredospores of P. fragosoana and P. imperatae were harvested by placing rust infected I. cylindrica leaf tissue in

200 ml distilled water amended with 0.1% Tween 80 and agitating manually for 5 min. For the second inoculation, uredospores of P. fragosoana and P. imperatae were harvested by scraping a sharp bent needle over the respective open uredinia (pustules) and collecting the spores which were then stored separately in plastic vials at -20°C until inoculation. Both spore suspensions

were adjusted to a final suspension of 1 x 106 spores ml-1 using a haemocytometer.

Sporisorium schweinfurthianum. The smut sori were harvested by placing the infected inflorescences in a 250 ml Erlenmeyer beaker filled with 150 ml distilled H2O amended with

0.1% Tween 80 with a final suspension of 1 x 108 sori ml-1. Control plants for both experiments were inoculated with distilled water amended with 0.1% Tween 80 only.

Treatments and Experimental Design

Puccinia fragosoana and P. imperatae. Eight-week-old I. cylindrica plants (5-blade stage) per treatment were inoculated with a 1 x 106 ml-1 spore suspension amended with 0.1%

Tween 80, using either a small hand-held sprayer, until the liquid started to drip (i.e., runoff) or by applying the spore suspension directly onto the leaves using a small, sterilized paint brush.

The inoculated plants were lightly misted with water before being covered with black plastic sheeting for 24 h at 28ºC to facilitate uredospore germination. After 24 h, the black plastic sheeting was removed and the inoculated plants were left in the shade house and monitored weekly for lesion and pustule development for 30 d.

Sporisorium schweinfurthianum. There were three treatments including the controls. A

1 x 108 sori ml-1 suspension amended with 0.1% Tween 80 was applied to 15 I. cylindrica seedlings as a soil drench and leaf inoculation, respectively. For the soil drench treatment, 15 potted seedlings were drenched until through-flow and the pots remained submersed in the spore suspension for 24 h. The seedlings for the soil drench treatment were not watered for the first 7 days to maintain the level of spore concentration. For the leaf inoculation, 15 potted seedlings were inoculated with a 1 x 108 ml-1spore suspension amended with 0.1% Tween 80, using a hand held sprayer, until runoff. Seedlings in both treatments were incubated at 28°C (RH<80%) for

24 h.

The inoculated seedlings were placed in the shade house for 14 d and subsequently moved to the field plot for up to 6 months. Distilled water amended with 0.1% Tween 80 was applied as both soil drench and leaf inoculation control treatments. All experiments were repeated twice.

Results

Puccinia fragosoana and P. imperatae. Four weeks after inoculation, no disease symptoms were observed for either P. fragosoana or P. imperatae. Neither inoculum method

was successful because no pustules developed. Possible reasons for the lack of infection could possibly be attributed to a decrease in spore viability due to long-term storage and the low

temperature germination requirements of Puccinia fragosoana.

Sporisorium schweinfurthianum. In trial one, typical smut symptoms appeared after

flowering 4 months after inoculation with a 1 x 108 sori ml-1 soil drench treatment on I.

cylindrica Benin plants. Five of the 15 plants inoculated produced a single inflorescence. Four of the five inflorescences were replaced with a brown, powdery mass of teliospores (Figure 4-2;

Table 4-2). No disease symptoms were observed on the cogongrass plants treated with the leaf inoculation. For trial two, no smut symptoms were observed on the single inflorescence produced after four months. No disease symptoms were observed on the cogongrass plants treated with the leaf inoculation. No inflorescences emerged in the control treatment and it was therefore not possible to compare the efficacy of the three treatments statistically.

Discussion

For any classical biocontrol program to be implemented, the researchers need to address the level of pathogenicity (i.e., the biocontrol potential of the agent) and determine the host range

(the agent’s safety to nontarget plants). Consideration of ecological principles in the selection and evaluation process for potential biocontrol agents should reduce future risks (Louda and

Arnett, 1999). Biological control of weeds using imported plant pathogens is safe,

environmentally sound, and cost effective (Barton, 2004). There are several successful classical

biocontrol programs using rust fungi to control invasive weeds. The control of Acacia saligna

(Labill.) H. Wendl. in South Africa, by the rust fungus Uromycladium tepperianum (Sacc.)

McAlpine, which is indigenous to Australia, is one of the most successful classical biocontrol

programs documented (Morris, 1999). In Australia, the rust fungus Puccinia chondrillina Bubák

& Syd. has been successfully established on skeletonweed (Chondrilla juncea L.), formerly a

major weed of wheat throughout the eastern states of Australia. The weed has been reduced to

tolerable densities. Puccina carduorum Jacky, a rust imported from Turkey and released into the

USA to control musk thistle, Carduus thoermeri J.A Weinm., has spread widely from its original

introduction in the northeastern United States to Wyoming and California in the west (Bruckart et al., 1996; Luster et al., 1999). Baudoin et al. (1993) showed that P. carduorum can significantly reduce musk thistle density. In Australia, release of rust fungi for control of blackberries (Rubus spp.), Parthenium, Mimosa pigra L. and rubber vine [Cryptostegia grandiflora (Roxb. ex R. Br.) R. Br.] has been moderately successful while the bridal creeper rust fungus P. myrsiphylli, released in 2000, is showing promise of becoming a spectacular success (Morin et al., 2002).

Head smut fungi are extremely host specific (Valverde et al., 1999) and the host range of the genus Sporisorium is restricted to the Poaceae. A head smut, Sporisorium ophiuri (P. Henn)

Vánky (Ustilaginales), has been thoroughly studied on itchgrass, Rottboellia cochinchinensis, an

erect, strongly tufted, C4, annual grass. An important weed in several crops including maize,

sugarcane, upland and rain-fed rice, beans, sorghum, and some perennials, such as citrus and oil

palm at early growth stages, itchgrass has been reported as a weed of many crops in several

countries (Holm et al. 1977, Ellison, 1987, 1993; Reeder et al. 1996). The smut is a soilborne

pathogen, infecting itchgrass seedlings before they emerge from the soil, thereby causing a

systemic infection that leads to seed sterility. Sporisorium ophiuri has been recorded as a head

smut of itchgrass in Africa and Asia and is restricted to the Old World (Reeder and Ellison,

1999). Reeder et al. (1996) screened S. ophiuri on graminaceous species to which the genus

Sporisorium is restricted. The inflorescence of infected plants emerges covered with a mass of

dark brown teliospores that are shed back to the soil to start a new infection in the next seedling

generation of R. cochinchinensis (Reeder et al., 1996). Infection with head smut should

effectively reduce seed rain and reduce R. cochinchinensis density in subsequent crops. There is

no plant-to-plant spread within a generation, as the smut infects only young seedlings. The smut

is also capable of growth on artificial media and produces large numbers of sporidia in liquid

fermentation. Sporidia are recognized as the infective propagule for a number of smut species,

including many belonging to the genus Sporisorium. Therefore, the possibility exists that

sporidia could eventually form the basis of a mycoherbicide formulation that is applied

inundatively. This would allow for a much faster spread of the biocontrol agent, and thus hasten

the impact of the smut on the weed population (Reeder et al., 1996).

A single cogongrass plant can produce up to 3000 seeds per inflorescence; implementing a classical biological control program with the head smut fungus, S. schweinfurthianum, could

effectively reduce the number of seeds produced leading to a reduction in cogongrass density.

The value of implementing S. schweinfurthianum as a classical biocontrol agent, in combination

with the foliar pathogens, such as the rust fungus, P. imperatae, and C. caudatum, B. sacchari,

and D.gigantea (see Chapter 3) could provide an effective means to manage this invasive weed.

The lack of infection following the inoculation of two rust species, P. fragosoana and P.

imperatae, makes it impossible to reach any conclusions with regard to the biocontrol potential

of these pathogens. The success with respect to the head smut, S. schweinfurthianum, on the

Benin cogongrass, albeit limited, offers hope that this pathogen may have potential as a classical

as well as an inundative biocontrol agent. The reason for the presenting these results are to have

a record of the research completed to date. Although a relatively small piece of the puzzle,

results from this research will benefit other researchers embarking on follow-up work.

Table 4-1. Collections of obligate fungi associated with Imperata cylindrica in South Africa. Fungi and life stages Locality and GPS coordinates (in Roman numerals East North East of Utrecht, Central North KwaZulu-Natal (27°32.834’S 30°28.690’E) P. fragosoana, II and III Church St East, Pretoria, Gauteng P. fragosoana, II and III (25°44.349’S 28°15.659’E) S. schweinfurthianum South of Sabie, Mpumalanga (25°09.929’S 30°46.164’E) Puccinia imperatae II

Table 4-2. Number of Imperata cylindrica inflorescences infected with the head smut, Sporisorium schweinfurthianum. Number of emerged Number of inflorescences infected Treatment (N=15)a inflorescences Soil drench 5 4 Leaf inoculation 1 0 Control 0 0 a Total number of seedlings inoculated per treatment.

100

Figure 4-1. Imperata cylindrica seedlings prior to inoculation with a 1 x 108 sori ml-1 Sporisorium schweinfurthianum soil drench treatment.

101

Figure 4-2. Imperata cylindrica inflorescence infected with the head smut, Sporisorium schweinfurthianum four months after a soil drench treatment.

102

CHAPTER 5 SUMMARY AND CONCLUSIONS

Imperata cylindrica (L.) Beauv. (cogongrass, speargrass) is a noxious, rhizomatous grass with a pan tropical distribution. It is ranked as one of the top 10 invasive weeds in the world.

Several survival strategies that include an extensive rhizome system, wind-disseminated seeds and high genetic plasticity enable this plant to be a highly invasive colonizer (MacDonald, 2004).

Within the genus Imperata, cogongrass is the most variable species and specimens have been found ranging from the western Mediterranean to South Africa, throughout Australia, India, and

Southeast Asia (including the Pacific Islands). In a previous study, it was demonstrated that two fungal pathogens, Bipolaris sacchari (Butler) Shoem. and Drechslera gigantea (Heald & F.A.

Wolf) Ito, were pathogenic and had potential as biological control agents of cogongrass in the

USA.

The genetic diversity between the USA and African cogongrass was assessed using ISSR markers and DNA sequence analysis of the ITS. DNA-based molecular marker techniques offer novel approaches to quantify genetic diversity of weeds in native and introduced habitats; they were employed in this study to better understand the relationship between the USA and West

African cogongrass. Included in the study were three closely related Imperata species: Imperata brasiliensis Trin., is native to North, Central and South America and overlaps in its distribution with cogongrass in Florida and possibly elsewhere in the Southeast, and is morphologically and genetically very similar; Imperata brevifolia Vasey is native to California, and I. cylindrica var. rubra, an ornamental variety sold in nurseries throughout the USA.

Weed populations with limited amounts of genetic variation are likely to be better targets for biological control agents (Burdon and Marshall, 1981; Chaboudez and Sheppard, 1995).

Understanding the extent of genetic variation can be of importance in the detection of ecotypes

103

or races of a species as they may differ considerably in their susceptibility to predators, parasites, or herbicides. In this study, the ISSR analyses showed that the USA and African I. cylindrica accessions, which were defined by three distinct groups, were genetically separated. Within the

USA cogongrass, there was a distinction between cogongrass accessions collected throughout the state of Florida and this study confirms that cogongrass was introduced multiple time into the

USA.

Imperata brasiliensis grouped with the Florida I. cylindrica and forms a sister species to the Florida cogongrass. In addition, I. cylindrica var. rubra was more closely related to the

African accessions. Imperata brevifolia was not genetically distinct from all the Imperata accessions treated within this study. Different tree topologies were generated with the ITS sequence analyses, which further confirmed that the African I. cylindrica was distinct from the

USA Imperata species. The Florida I. cylindrica accessions formed two distinct groups within the tree, one representing the majority of accessions colleted in West and Central Florida and one representing accessions collected in East Florida. Imperata brasiliensis grouped within the USA

I. cylindrica accessions. The Imperata cylindrica var. rubra accessions grouped together and were genetically more similar to the African cogongrass than to the USA cogongrass. The

African cogongrass exhibited higher genetic differentiation than those in USA and closely related Imperata species.

These findings have implications for implementing a biological control approach to manage cogongrass. The hypothesis that fungal biological control within the USA and African cogongrass would result in differential responses across the different accessions was not supported by this study as the West African (Benin accession) was found to be equally susceptible to the Florida and Benin isolates of B. sacchari and D. gigantea. However, it is

104

recommended that samples from different groups of I. cylindrica be included when screening

other candidate biological control pathogens in the future. Low genetic variation among the

different West African cogongrass would suggest that the Florida isolates of B. sacchari and D.

gigantea could potentially control the West African variety of cogongrass if employed in a

mycoherbicide formulation.

Virulence of the fungal pathogens, B. sacchari and D. gigantea, applied singly or as a mixture associated with I. cylindrica accessions in Florida and Benin, and closely related species,

I. brasiliensis, I. brevifolia and I. cylindrica var. rubra, was confirmed in repeated greenhouse trials on cogongrass accessions in Florida, USA and in shade house trials in Benin, West Africa.

An indigenous Colletotrichum caudatum (Sacc.) Peck isolate was included in the Benin study.

No consistent response was observed over the two trials to the different treatments on the Florida cogongrass accessions over the monitoring period. The majority of the Florida cogongrass accessions exhibited no significant differences in their reaction to the treatments compared with the control treatment over the monitoring period, and there was no uniformity in the variance of disease severity over accessions. Testing in Benin revealed that there were no significant differences between the Florida and Benin isolates when inoculated on the Benin cogongrass.

When compared to the B. sacchari and the D. gigantea isolates, the indigenous fungus C. caudatum consistently caused similar or greater disease symptoms on the Benin accession. The

slight decrease in disease severity noted can be attributed to the regrowth i.e., emergence of new leaves and shoots. At no time during experimental period was plant mortality observed. From these results, the potential to use B. sacchari and D. sacchari as inundative biocontrol agents to control cogongrass in the USA and West Africa appears limited. Follow-up research, with respect to the indigenous Colletotrichum caudatum, should be undertaken in West Africa. These

105

should include extensive host range testing to exclude the potential for nontarget effects.

Additional field trials, using available local resources, should be further examined.

Herbaceous weeds have been targeted for classical biological control strategies the most

frequently, while grassy weeds have not, until recently, been considered as suitable targets.

Despite the importance of the problems caused by cogongrass throughout the tropical areas of the world, biological control efforts have been very limited (Caunter, 1996) and only scant

information exists on the pathogens of cogongrass and their potential as biological control agents

(Evans, 1991). Evans (1987, 1991) suggested that some of the known fungal pathogens of

cogongrass should be considered for introduction as classical biological control agents on this

invasive weed. Promising species include the rust fungi, Puccinia fragosoana Beltran, P. rufipes

Diet. and P. imperatae Poirault and a smut fungus, Sporisorium schweinfurthianum (Thuem.)

Sacc. that is well represented on cogongrass in the Old World, mainly Africa, and the

hemibiotrophic fungus Colletotrichum caudatum. In this study, neither of the two rust

pathogens, P. fragosoana and P. imperatae was successful in causing infection on Benin

cogongrass in glasshouse trials. Two reasons for the negative response are that the rust spores

where possibly not viable after long-term storage and the uredosporeos of P. fragosoana requires

low temperature for germination. Typical smut symptoms appeared after flowering 4 months

after inoculation with a soil drench treatment on the Benin accession while no diseased

inflorescences were observed on the leaf-inoculated cogongrass. It is interesting to note that the

control plants did not flower within the experimental time-frame. The assumption would be that

the smut caused the inoculated plants to flower earlier, but more studies are needed to test this

hypothesis. The efficacy of the C. caudatum under field conditions was improved by

106

formulating the spores in an oil emulsion. The oil sformulation caused high levels of disease incidence and disease severity even with no period of dew after inoculation.

The detection of intraspecific variation in resistance has significant implications biological

control programs as more than one strain of the pathogen might be required to control all

genotypes of the exotic weed. Finding the most virulent strain of the pathogen may require exact

host-matching in the native range for the specific genotypes of the exotic weed targeted for

control. Determining the level of genetic variation within weed species and populations can be

employed as an essential tool in the selection of the most suitable, and ultimately the most effective, biological control agents. Implementing an amended C. caudatum formulation could

potentially play a role in an augmentative biocontrol strategy for long-term control of cogongrass

in West Africa. A classical biological control program using the head smut fungus, S.

schweinfurthianum might effectively reduce the number of seeds produced, leading to a

reduction in cogongrass density over time and to the success of biocontrol of this invasive weed.

107

APPENDIX A FIELD EVALUATION OF THE INDIGENOUS FUNGUS, Colletotrichum caudatum, ASSOCIATED WITH COGONGRASS IN BENIN

Introduction

The use of indigenous plant pathogens, isolated from weeds, cultured to produce effective levels of infective propagules and used at inundative levels of inoculum is known as the bioherbicide or approach to biological control of invasive weeds (Charudattan, 2001). Applying infective propagules at rates high enough to cause extensive infection suppresses the target weed before economic losses occur. One or several applications are required annually as these pathogens generally do not reproduce in sufficient numbers between growing seasons to initiate new epidemics on new weed infestations (Smith, 1982; Templeton, 1982; Charudattan, 1988).

Of the nine fungi that were developed as practical mycoherbicides or were considered to have excellent prospects of being commercially available for use in the future, five are Colletotrichum strains (Templeton, 1992). Collego, a formulation of the endemic anthracnose fungus C. gleosporiodes (Penz.) Penz. & Sacc. in Penz. f. sp. aeschynomene, was developed to control northern joint vetch (Aeschynomene virginica (L.) B.S.P.) in rice and soybean fields. Registered in 1982, the active ingredient contains 15% C. gleosporiodes f. sp. aeschynomene conidia as a dry powder formulation with shelf-life of 18 months. It was the first commercially available mycoherbicide for use in annual weed in annual crops giving 90% control efficiency (TeBeest and Templeton, 1985). The successful development of Collego led to the discovery of another

Collectotrichum-based mycoherbicide, by Philom Bios Inc., Canada. It contains conidia of C. gleosporiodes (Penz.) Penz. & Sacc. in Penz. f. sp. malvae to control Malva pusilla Sm. (round- leaved mallow) in Canada and USA. The most effective period of application of a

108

mycoherbicide is at the early growth stage of the weed, although it can be effective at any stage

of weed growth (Makowski and Mortensen, 1992).

One of the possible reasons why Colletotrichum spp. are more successful than others are

that their pathogenicity strategies which include a hemibiotrophic phase that often precedes the necrotrophic phase. Many Colletotrichum species produce large quantities of extracellular lytic

enzymes as well as toxins, while other species are known to produce local epidemics. While the relative contribution of each property is unclear, it is most likely that a combination of these factors make these fungi superior weed biological control agents (Sharon, 1998).

Hemibiotrophic pathogens are characterized by their initial infection of living host cells

followed by their necrotrophic phase and typical sporulation on dead tissues (Luttrell, 1974).

Hemibiotrophs are divided into biotrophic hemibiotrophs, species in which colonization of host

tissues is entirely biotrophic, with the death of host cells delayed until about the time the

pathogen begins to sporulate (Parbery, 1996). Necrotrophic hemibiotrophs have a short biotrophic phase that allows them to take possession of host tissue prior to killing and invading the tissue. Examples of necrotrophic hemibiotrophs include various species of Colletotrichum

(Latunde-Dada et al., 1996; O’Connell et al., 1993; Wei et al., 1997), including C. lindemuthianum (Sacc. & Magnus) Lams.-Scrib. (O’Connell and Bailey, 1991). Colletotrichum

species produce two major sources of inoculum; conidia produced in acervuli and ascospores produced in and released from perithecia (Bailey et al., 1992). Dissemination of spores from young acervuli occurs in drops of water while wind distributes the dry spore masses from older acervuli and ascospores from perithecia. Successful Colletotrichum mycoherbicides products

such as Collego™ are easily isolated and cultured on various solid and liquid media. All require

relatively long dew periods of between 12 and 24 h at optimal temperatures of between 20–30ºC.

109

Colletotrichum acutatum (Sacc.) Peck, incorrectly identified as C. gloeosporioides, utilized for

mycoherbicidal use on Hakea sericea Schrad. & J.C.Wendl. in South Africa was cultured on

sterilized bran and effectively used on hakea seedlings under field conditions (Morris, 1989).

In the late 1980s, Caunter and Wong (1988) surveyed for foliar pathogens of I. cylindrica throughout Malaysia and, while they generally found low disease incidence, several candidate fungi were isolated including a tentatively identified isolate of C. caudatum. The pathogen’s widespread distribution and previous confirmation of host specificity made it a suitable candidate for investigation into its potential as a bioherbicide (Caunter, 1996). The conidia germinated via

a single germ tube within 12 h after inoculation and appressoria formed within 24 h with symptoms visible after 48 h (Caunter, 1996). Initial symptoms were chlorotic spots on the leaf tips which developed into necrotic streaks along the leaf edges. At 34 DAI, only 40% of the leaves had died and no plant death was observed. No effect was observed on regenerative ability of the rhizomes. The addition of adjuvants and amendments, which affect spore germination, virulence or environmental requirement, can influence disease severity. Caunter (1996) showed that amendments of 0.1 – 5.0% sucrose, glucose, and yeast-extract had little or no effect on spore germination but had an influence on appressorial formation and possibly plays a role in increasing disease severity.

During a survey for pathogens and insects associated with cogongrass in the southeastern

USA, C. caudatum was identified on cogongrass in Florida (Minno and Minno, 1998). In

November 2002, C. caudatum was observed on infected I. cylindrica plants in the field in Benin

(J. Hotegni, personal communication) and an isolate (COCBEN8) was confirmed to be pathogenic and highly virulent on I. cylindrica in Benin (Adolphe Avocanh, 2006, unpublished

110

data). The objective of this study was to evaluate the efficacy of this indigenous leaf-spot

causing fungus isolate from Benin on the I. cylindrica Benin accession under field conditions.

Materials and Methods

Inoculum Production

A pure culture of Colletotrichum caudatum (isolate COCBEN8) was obtained from Dr. F.

Beed’s culture collection (IITA, Cotonou, Benin). Initial isolations were made from diseased cogongrass plants growing at the IITA Station, Cotonou, Benin and their pathogenicity confirmed (Adolphe Avocanh, 2006, unpublished data). The inoculum was produced by transferring mycelial blocks, from stored slant cultures, to 150 ml potato salts broth. The broth was prepared by steeping 200 g grated potato in distilled water for 1 hour, boiled for 5 min, filtered through cheese cloth and the filtrate made up to 1 liter. To the filtrate was added 20 g sucrose, 10 g KNO3, 5 g KH2PO4 2.5 g MgSO4.7 H2O, 0.02 g FeCl2 up to. The pH was adjusted to 6 using NaOH and the medium was autoclaved at 121°C for 15 min. The inoculated flasks were placed on a rotary shaker at 100 rpm and the fungua allowed to grow at 25°C for 8 d under laboratory conditions. Spores and mycelia were harvested by filtering the resulting suspension through sterile cheese cloth. The spore concentration was determined with the aid of a haemocytometer.

Inoculation

A field site was selected on the IITA-Benin station and all existing cogongrass plants were removed from the site (Figure A-1). The cogongrass plants were allowed to regenerate naturally.

Sixteen 0.50 m2 plots with 0.25 m2 clearing between plots were marked in a randomized

complete block design. A 1 x 106 spores ml-1 amended with 1% paraffin oil and finely ground kaolin (0.006 ml of 1 x 1010 spore ml-1 suspension and 6 g of kaolin made up to 60 ml) and a

mycelial suspension amended with 1% paraffin oil and kaolin (6 g dried mycelia and 6 g of

111

kaolin) were applied respectively at a rate of 100 ml amended spore or mycelial suspension per

plot at 15 s per plot. The two control treatments included a 1% kaolin in water and 1% paraffin

oil and 1% kaolin in water. Field plots were inoculated in the late afternoon to limit the effect of

high temperature on the outcome of the experiment. Due to the fact that C. caudatum is

ubiquitous throughout Benin, an initial disease assessment was made prior to inoculation (0 days after inoculation [DAI]).

Disease Assessment

Disease was assessed as disease incidence (DI) and disease severity (DS). Disease incidence was based on the number of plants infected among the total number of plants inoculated, expressed as a percentage of diseased plants (Horsfall and Cowling, 1978). Disease severity rating was based on the Horsfall-Barrett scale (Horsfall and Barrett, 1945). The disease rating scale consisted of 12 class-values representing the percentage of disease severity as: 0 =

0%; 1 = 0 - 3%; 2 = 3 – 6%; 3 = 6 – 12%; 4 = 12 – 25%; 5 = 25 – 50%; 6 = 50 – 75%; 7 = 75 –

88%; 8 = 88 – 94%; 9 = 94 – 97%; 10 = 97 – 100%; 11 = 100%. The mean class value was used to determine the final disease severity value.

Statistical Analysis

Percentage data were transformed using an arcsine square root transformation. Analysis of

variance (ANOVA) was performed using the General Linear Models Procedure (SAS Institute,

Cary, NC). Means were compared using the Student-Newman-Keuls test at P<0.05.

Results

Due to the fact that the fungus, Colletotrichum caudatum is ubiquitous in the field in

Benin, an initial disease incidence assessment was made prior to inoculation at 0 DAI. Results showed that after an initial DI rating of 30% for the mycelial and 53.33% for the spore suspension treatments respectively, DI was observed at 100% for both treatments for the

112

experiment duration (Figure A-2A). There were no significant differences in DI at any of the

observation dates (Table A-1). Both control treatments had equal or higher DI ratings compared with the fungal treatments at 0 DAI. Similarly, DI increased to 100% after 7 DAI (Figure A-

2A). Disease severity (DS) ratings ranged from one (0-3%) to six (50-75%) between treatments

(Figure B-2B). The two control treatments had similar DS ratings to the mycelial treatment, ranging from 1 to 4. However, the DS rating for the spore suspension treatment ranged from 1 to

6 for the same observation period. Compared with the mycelial and control treatments, the spore

suspension treatment caused the most disease severity starting at 7 DAI and continuing up to 35

DAI.

Discussion

Many fungi have a requirement for free water, such as dew, for germination and infection

(Dhingra and Sinclair, 1985). While these requirements often limit the efficacy and therefore the commercial potential of bioherbicides (Auld, 1993), they can be overcome by formulation, application technology and careful timing of application (Walker and Boyette, 1986; Makowski and Mortensen, 1990). Fungal propagules formulated in invert oil emulsions (water in oil) and

suspension oil emulsions (oil in water) have been employed for the enhancement of bioherbicide

fungal agents as a result of their ability overcome the limitation of insufficient free water and spore protection (Amsellem et al., 1990; Daigle et al., 1990; Auld, 1993; Boyette et al., 1993;

Womack and Burge, 1993, Chandramohan et al, 2002; Yandoc et al., 2005). Due to their high

viscosity, invert emulsions are considered to be difficult to prepare and apply, and can be

fungitoxic (Womack et al., 1996), phytotoxic (Auld, 1993; Womack and Burge, 1993; Yang et al., 1993) and cause nontarget plant damage (Daigle et al., 1990; Amsellem et al., 1991).

Suspension oil emulsions were developed in an effort to reduce the high and costly oil content of invert emulsions and showed promise in reducing the free water limitation in two weed-pathogen

113

systems (Auld, 1993; Boyette, 1994) and reduced the spray volume and number of conidia of C. truncatum (Schwein.) Andrus & W.D. Moore required for control of Sesbania exaltata (Raf.)

Cory. Boyette et al. (2007) showed, in greenhouse and field experiments, that an oil in water

emulsion of unrefined corn oil and a surfactant, Silwet L-77 increased the biological control

efficacy of C. truncatum for S. exaltata by stimulating the germination and appressoria formation in vivo and in vitro and delaying the required dew period. Soybean yields were significantly

higher in the weed-controlled plots. Oil suspension emulsions of the potential mycoherbicide, C. orbiculare (Berk. & Mont.) Arx, which causes anthracnose of Bathurst burr, Xanthium spinosum

L., were evaluated in field trials (Klein et al., 1995). Spores of C. orbiculare in kaolin were mixed with different rates of vegetable and mineral oils and applied at several spore concentrations at field sites over two seasons. Results varied, with improved mycoherbicide activity in the kaolin formulation compared to an aqueous spore suspension treatment in the first season, but the improvement was not repeatable at sites sprayed for the second season (Klein et al., 1995). The effective control of seven weed species using three pathogens formulated with an oil emulsion of Sunspray 6E and paraffin oil (4:1 ratio) was demonstrated by Chandramohan

(1999). A similar oil emulsion was used by Yandoc (2001) to improve the level of biocontrol efficacy of B. sacchari (Butler) Shoem. and D. gigantea (Heald & F.A. Wolf) Ito on cogongrass.

Greenhouse trials showed that an oil emulsion containing 4% horticultural oil, 10% light mineral oil, and 86% water resulted in higher levels of foliar blight with no dew exposure or shorter dew periods (Yandoc et al., 2005). Field trials showed that both fungi, when applied at 105 spores ml-

1 in an oil emulsion, caused foliar injury while phytotoxic damage was also observed from the oil

emulsion.

114

The efficacy of the C. caudatum may be improved by formulating the spores in an oil emulsion. The results show that using a properly formulated C. caudatum spore suspension could be applied as an augmentative biocontrol strategy for long-term control of cogongrass in

West Africa.

115

Table A-1. Analysis of variance for disease incidence of Colletotrichum caudatum at 7 days after inoculation. Degrees of Source freedom Sum of squares Mean Square F value Pr > F Model 3 1366.66667 455.555556 0.52 0.6799 Error 8 7000.00 875.000 Corrected total 11 8366.667

116

Figure A-1. Field trial site with reemerging Imperata cylindrica plants at IITA, Cotonou, Benin.

117

) % ( Disease incidence Disease incidence

Time (weeks after inoculation)

Mycelium Spore suspension Control1 Control 2

B Disease severity (mean rating) Disease severity

Time (weeks after inoculation)

Figure A-2. Field inoculation of Imperata cylindrica Benin using an amended Colletotrichum caudatum formulation. A) Disease incidence. B) Disease severity.

118

APPENDIX B ACCESSION AND GENBANK DATA FOR THE USA Imperata cylindrica AND CLOSELY RELATED SPECIES

Table B-1. Imperata cylindrica and closely related species sampled in USA for population structure and virulence studies. GenBank Plant accesssion Collection sites accessiona GPS coordinates USA: CGS1 Marion EU267045 N 29°15.608' W 82°09.896' CGS2 Marion N 29°18.310' W 82°09.091' CGS4 Leon N 30°40.254' W 84°16.961' CGS5 Hernanado EU267064 N 28°27.875' W 82°26.706' CGS7 Alabama N 30°40.326' W 87°54.130' CGS8 Walton N 30°43.586' W 86°07.776' CGS9 Walton N 30°47.492' W 86°11.174' CGS11 Santa Rosa N 30°44.110' W 86°44.576' CGS12 Santa Rosa N 30°42.725' W 86°52.698' CGS13 Santa Rosa N 30°37.8 32' W 86°5 9.453' CGS14 Escambia N 30°28.580' W 87°19.427' CGS15 Escambia N 30°18.206' W 86°25.652' CGS16 Santa Rosa N 30°23.318' W 87°04.563' CGS17 Okaloosa N 30°24.688' W 86°45.406' CGS18 Holmes EU267046 N 30°43.224' W 85°55.624' CGS19 Lafayette EU267047 N 30°08.252' W 83°13.608' CGS20 Hamilton EU267048 N 30°28.398' W 83°13.629' CGS21 Gilchrist EU267049 N 29°37.373' W 82°45.162' CGS22 Gilchrist N 29°36.435' W 82°52.418' CGS23 Levy N 29°33.715' W 82°54.492' CGS24 Levy N 29°21.127' W 82°44.324' CGS25 Citrus N 28°58.153' W 82°38.117' CGS26 Citrus EU267050 N 28°48.924' W 82°34.672' CGS27 Hernando EU267051 N 28°26.977' W 82°38.017' CGS28 Pasco N 28°21.359' W 82°41.964' CGS29 Pasco EU267052 N 28°21.426' W 82°30.263' CGS30 Hillsborough EU267054 N 28°08.527' W 82°09.039' CGS31 Hillsborough N 28°04.824' W 82°03.870' CGS32 Polk N 28°07.627' W 81°58.434' CGS33 Polk EU267055 N 28°14.921' W 82°03.338' CGS34 Sumter N 28°50.072' W 82°02.738' CGS35 Bradford EU267056 N 29°54.220' W 82°08.143' CGS36 Marion EU267057

119

Table B-1. Continued. GenBank Plant accessions Collection sites accession GPS coordinates CGS37 Alachua EU267058 N 29°42.983' W 82°08.340' CGS38 Clay EU267059 N 29°58.670' W 81°52.670' CGS39 Clay N 29°59.020' W 81°43.750' CGS40 St Johns N 29°59.381' W 81°36.600' CGS41 St Johns EU267060 N 30°07.422' W 81°37.500' CGS42 Nassau EU267061 N 30°32.016' W 81°51.466' CGS43 Putnam EU267062 N 29°36.377' W 82°01.835' CGS44 Flagler N 29°28.762' W 81°26.834' CGS45 Volusia N 29°08.601' W 81°00.838' CGS46 Volusia N 29°01.097' W 80°59.512' CGS47 Brevard N 28°45.072' W 80°53.307' CGS48 Brevard EU267063 N 28°01.671' W 80°39.093' CGS49 St Lucie N 27°22.254' W 80°29.867' CGS50 St Lucie N 27°22.467' W 80°27.920' CGS51 Okeechobee EU267065 N 27°15.923' W 80°51.903' CGS52 Okeechobee EU267066 N 27°21.853' W 80°58.177' CGS53 Highlands EU267067 N 27°22.229' W 81°04.776' CGS54 Highlands N 27°24.754' W 81°30.792' CGS55 Hardee N 27°29.580' W 81°46.557' CGS56 Hardee EU267068 N 27°28.953' W 81°54.951' CGS57 Manatee N 27°28.038' W 82°04.612' CGS58 Osceola N 27°46.315' W 81°06.769' CGS59 Osceola N 27°41.826' W 80°53.968' CGS60 Union CGS61 Putnam N 29°43.141' W 81°48.750' CGS62 Putnam N 29°43.141' W 81°48.750' CGS63 Lake EU267070 CGS64 Lake CGS65 Lee N 26°43.765' W 81°39.222' Imperata brasiliensis CGS3 Miami-Dade EU267053 N 25°29.344' W 80°27.305' Imperata brevifolia CGS6 California N 34°30.154' W 119°20.768' Imperata cylindrica var. rubra CGS66 (1) Nursery plants EU267072 CGS66 (2) Nursery plants EU267072 a GenBank accession numbers given to the ITS sequences fro m the present study.

120

APPENDIX C ACCESSION AND GENBANK DATA FOR THE WE ST AND SOUTH AFRICAN Imperata cylindrica

Table C-1. Imperata cylindrica and closely related species sa mpled in West and South Africa for population structure and virulence studies. Plant GenBank accessions Collection sites accessionsa GPS coordinates Cameroon: CGC2 Nkom pipeline - invasive EU267036 N 4º 4.725' E 11º34.320' CGC6 Alae - invasive EU267037 N 4º 5.076' E 11º35.212' CGC8 Eteme nkom - invasive N 4º 5.642' E 11º35.574' CGC9 Essong MF Savane - natural N 4º 6.065' E 11º36.214' CGC10 Ateba - invasive N 4º 5.505' E 11º35.521' CGC11 Etengue - invasive EU267035 Nigeria: CGN1 Alabata 1 EU267038 CGN2 Alabata 2 EU267039 CGN3 Ijaiya 1 EU267041 CGN4 Ijaiya 2 EU267043 Guinea: CGG1 Benin: CGB1 IITA Research Station EU267030 N 6º25.252' E 2º19.720' CGB2 IITA Research Station EU267031 N 6º25.252' E 2º19.720' CGB3 Savalou 1 EU267032 CGB4 Savalou 2 CGB5 Sirarou 1 EU267033 CGB6 Sirarou 2 CGB7 IITA Research Station N 6º25.252' E 2º19.720' South Africa: CGSA1 KwaZulu Natal EU267073 S 27°32.834' E 30°28.690' CGSA2 KwaZulu Natal S 27°32.834' E 30°28.690' CGSA3 KwaZulu N atal S 27°32.834' E 30°28.690' CGSA4 KwaZulu Natal S 27°32.834' E 30°28.690' CGSA5 KwaZulu N atal S 27°32.834' E 30°28.690' CGSA6 KwaZulu Natal S 27°32.834' E 30°28.690' a GenBank accession number s given to the ITS sequences from the present study.

121

LIST OF REFERENCES

Allen, C.M., Thomas, R.D., 1991. Brachiaria plantaginea, Imperata cylindrica, and Panicum maximum: three grasses (Poaceae) new to Louisiana and a range extension for Rottboellia cochinchinensis. Sida, Contributions to Botany 14, 613.

Allendorf, F.W., Lundquist, L.L., 2003. Introduction: Population biology, evolution and control of invasive sp ecies. Conservation Biology 17(1), 24-30.

Altschul, S. F., Madden, T. L., Schaffer, A. A., Zhang, J., Zhang, Z., Miller, W., Lipman, D. J., 1997. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Research 25, 3389-3402.

Amsellem , Z., Sharon, A., Gressel, J., Quimby, P.C. Jr., 1990. Complete abolition of high inoculum threshold of two mycoherbicides (Alternaria cassiae and A. crassa) when applied in invert emulsion. Phytopathology 80, 925-929.

Anderson, R.C., Gardner, E.D., 1999. An evaluation of the wilt-causing bacterium Ralstonia solancearum as a potential biological control agent for the alien kahili ginger (Hedychium garderianum) in Hawaiian forests. Biological Control 15, 8 9-96.

Anikster, Y., 1984. The formae speciales. In: Bushnell, W.R., Ro elfs, A.P. (Eds.), The Cereal Rusts Vol. I. Academic Press. Orlando, pp. 115-130.

Anon., 1954. Notes and exhibitions. In: Proceedings of the Hawaiian Entomological Society (Honolulu). Volume 15, pp 263-265; 279-281; 282-283.

Ash, G.J., Raman, R., Crump, N.S., 2003. An investigation of gen etic variation in Carthamus lanatus in New South Wales, Australia, using intersimple se quence repeats (ISSR) analysis. Weed Researc h 43, 208-213.

Auld, B.A., 1992. Development and commercialization of biocontrol agents. In: Proceedings of First International Weed Control Congress. Agmedia, Melbourne, Australia, pp. 269-272.

Auld, B.A., 1993. Vegetable oil suspension emulsions reduce dew dependence of a mycoherbicide. Crop Protection 12, 477-479.

Ayeni, A. O., Duke, W.B., 1985. The influence of rhizome features on subsequent regenerative capacity in speargrass [Imperata cylindrica (L.) Beauv.]. Agriculture, Ecosystems and Environment 13(3/4), 309-317.

Bailey, J.A ., O'Connell, R.J., Pring, R.J., Nash, C., 1992. Infection strategies of Colletotrichum species. In: Bailey J.A., Jeger, M.J. (Eds.), Colletotrichum: Biology, Pathology and Control. C.A.B. International: Wallingford, U.K., pp. 88-120.

122

Baldwin, B.G., Sanderson, M.J., Porter, J.M., Wojciechowski, M.F., Cambell, C.S., Donoghue, M.J., 1995. The ITS region of nuclear ribosomal DNA: a valuable source of evidence on angiosperm phylogeny. Annals of the Missouri Botanical Garden 82, 247–277.

Barreto, R.B., Evans, H.C., 1988. Taxonomy of a fungus introduced into Hawaii for biological control of Ageratina riparia (Eupatoriae; Compositae) with observations on related weed pathogens. Transactions of the British Mycological Society 91, 81-97.

Barrett, S.C.B., 1982. Genetic variation in weeds. In: Charudattan, R., Walker H.L. (Eds.), Biological Control of Weeds with Plant Pathogens. John Wiley & Sons, New York, pp. 78- 112.

Barrett, S.C.B., 1988. Genetics and evolution of agricultural weeds. In: Altieri, M.A., Liebman, M. (Eds.), Weed Management in Agroecosystems: Ecological Approaches. Boca Raton, FL: CRC, pp. 57–75.

Barton, J., 2004. How good are we at predicting the field host-range of fungal pathogens used for classical biological control of weeds? Biological Control 31, 99–122.

Barton, J., Fowler, S.V., Gianotti, A.F., Winks, C.J., de Beurs, M., Arnold, G.C., Forrester, G., 2007. Successful biological control of mist flower (Ageratina riparia) in New Zealand: Agent establishment, impact and benefits to the native flora. Biological Control 40, 370- 385.

Baudoin, A.B.A.M., Abad, R.G., Kok, L.T., Bruckart, W.L., 1993. Field evaluation of Puccinia carduorum for biological control of musk thistle. Biological Control 3, 53–60.

Bauer, R., Begerow, D., Oberwinkler, F., 1997. Ustilaginomycotina. Ustilaginomycetes. The true smut fungi. Version 30 November 1997. The Tree of Life Project, http://tolweb.org/Ustilaginomycetes/20530/1997.11.

Booth, C., Prior, C., 1984. A new Giberella species on Imperata (Graminae). Transactions of the British Mycological Society 82, 180-182.

Boyette, C.D., 1994. Unrefined corn oil improved the mycoherbicidal activity of Colletotrichum truncatum for hemp sesbania (Sesbania exaltata ) control. Weed Technology 8, 526-529.

Boyette, C.D., Hoagland, R.E., Weaver, M.A., 2007. Biocontrol efficacy of Colletotrichum truncatum for hemp sesbania (Sesbania exaltata) is enhanced with unrefined corn oil and surfactant. Weed Biology and Management 7, 70-76.

Boyette, C.D., Jackson, M.A., Quimby, P.C. Jr., Connick, W.J. Jr., Zidak, N.K. Abbas, H.K., 1999. Biological control of the weed hemp sesbania with Colletotrichum truncatum. In: Abstracts of the 10th International Symposium on Biological Control of Weeds, Montana State University, Bozeman, MT, USA.

123

Boyette, C.D., Quimby, P.C. Jr, Bryson, C.T., Egley, G.H., Fulgham, F.E., 1993. Biological control of hemp sesbania (Sesbania exaltata) under field conditions with Colletotrichum truncatum formulated in an invert emulsion. Weed Science 41, 497-500.

Briese, D.T., 2000. Classical biological control. In: Sindel, B.M. (Ed), Australian Weed Management Systems. R.G. & F.J. Richardson Publishers, Melbourne, Victoria, pp. 161– 192.

Bruckart, W.L., Politis, D.J., Defago, G., Rosenthal, S.S., Supkoff, D.M., 1996. Susceptibility of Carduus, Cirsium, and Cynara species artificially inoculated with Puccinia carduorum from musk thistle. Biological Control 6, 215–221.

Bruzzese, E., Hasan, S., 1986. The collection and selection in Europe of isolates of Phragmidium violaceum (Uredinales) pathogenic to species of European blackberry naturalised in Australia. Annals of Applied Biology 108, 527-533.

Bryson, C.T., 1985. A new food plant record for Atalopedes campestris (Boisduival) Hesperiidae. Journal of the Lepidopterists’ Society 39 (4), 335.

Bryson, C.T., 1987. Native butterflies accepted cogongrass (Imperata cylindrica [L.] Beauv.) as host plant. Journal of the Mississippi Academic Society 3, 1.

Bryson, C.T., Carter, R., 1993. Cogongrass, Imperata cylindrica, in the United States. Weed Technology 7(4), 1005-1009.

Burdon, J. J., Marshall, D.R., 1981. Biological control and the reproductive mode of weeds. Journal of Applied Ecology 18, 649-658.

Campbell, F.T., 1998. Worst Invasive Plant Species in the Conterminous United States. Report, Western Ancient Forest Campaign, Springfield, VA.

Caunter, I.G., 1996. Colletotrichum caudatum, a potential bioherbicide for control of Imperata cylindrica. In: Moran, V.C., Hoffmann, J.H. (Eds.), Proceedings of the IX International Symposium on Biological Control of Weeds. University of Cape Town, Cape Town, South Africa, pp. 525-527.

Caunter, I.G., Wong, S.H., 1988. Foliar diseases of Lalang in Penang: biocontrol potential. MAPPS Newsletter12, 6.

Chaboudez, P., 1994. Patterns of clonal variation in skeleton weed (Chondrilla juncea), an apomictic species. Australian Journal of Botany 42, 283-295.

Chaboudez, P., Sheppard, A.W., 1995. Are particular weeds more amenable to biological control? – A reanalysis of mode of reproduction and life history. In: Delfosse, E.S., Scott, R.R. (Eds.), Proceedings of the Eighth International Symposium on Biological Control of Weeds, Lincoln, New Zealand, pp. 95-102.

124

Chandramohan, S., 1999. Multiple-Pathogen Strategy for Bioherbicidal Control of Several Weeds. PhD Dissertation, University of Florida, Gainesville FL, 191p.

Chandramohan, S., Charudattan, R., 1996. Multiple-pathogen strategy for bioherbicidal control of several weeds. Weed Science Society of America Abstract 36, 49.

Chandramo han, S., Charudattan, R., Sonoda, R.M., Singh, M., 2002. Field evaluation of a fungal pathogen mixture for the control of seven weedy grasses. Weed Science 50, 204-213.

Charudattan, R., 1988. Inundative control of weeds with indigenous fungal plant pathogens. In: Burge, M. N. (Ed), Fungi in Biological Control Systems. Manchester University Press, Manchester U.K., pp. 86-110.

Charudattan, R., 1997. Development of biological control for noxious plant species – Progress report May 15 – November 1997, University of Florida, Gainesville, Florida, USA.

Charudattan, R., 2001. Biological control of weeds by means of plant pathogens: Significance for integrated weed management in modern agro-ecology. BioControl 46, 229-260.

Charudattan, R., 2005. Ecological, practical, and political inputs into selections of weed targets: what makes a good biological control agent? Biological Control 35, 183-196.

Chase, M.W., Soltis, D.E., Olmstead, R.G., Morgan, D., Les, D.H., Mishler, B.D., Duvall, M.R., Price, R.A., Hills, H.G., Qiu, Y.L., Kron, K.A., Rettig, J.H., Conti, E., Palmer, J.D., Manhart, J.R., Sytsma, K.J., Michaels, H.J., Kress, W.J., Karol, K.G., Clark, W.D., Hedren, M., Gaut, B.S., Jansen, R.K., Kim, K.J., Wimpee, C.F., Smith, J.F., Furnier, G.R., Strauss, S.H., Xiang, Q.Y., Plunkett, G.M., Soltis, P.S., Swensen, S.M., Williams, S.E., Gadek, P.A., Quinn, C.J., Eguiarte, L.E., Golenberg, E., Learn, G.H., Graham, S.W., Barrett, S.C.H., Dayanandan, S., Albert, V.A., 1993. Phylogenetics of seed plants: an analysis of nucleotide sequences from the plastid gene rbcL. Annals of the Missouri Botanical Garden 80, 528–580.

Chase, M.W., de Bruijn, A.Y., Reeves, G., Cox, A.V., Rudall, P.J., Johnson, M.A.T., Eguiarte, L.E., 2000. Phylogenetics of Asphodelaceae (Asparagales): an analysis of plastid rbcL and trnL-F DNA sequences. Annals of Botany (London) 86, 935–956.

Chase, C.A., Shilling, D.G., Bewick, T.A., Charudattan, R., 1996. Fungal isolates with potential for the biological control of cogongrass [Imperata cylindrica (L.) Beauv.]. Weed Science Society of America Abstract 160, 49.

Cheng, K.T., Chou, C.H., 1997. Ecotypic variation of Imperata cylindrica populations in Taiwan: I. Morphological and molecular evidences. Botanical Bulletin of Academia Sinica 38, 215-223.

Chikoye, D., Ekeleme, F., Ambe, J.T., 1999. Survey of distribution and farmers’ perceptions of speargrass [Imperata cylindrica (L.) Raeuschel] in cassava-based systems in West Africa. International Journal of Pest Management 45 (4), 305-311.

125

Chikoye, D., Ekeleme, F., Udensi, U.E., 2001. Cogongrass suppression by intercropping cover crops in corn/cassava systems. Weed Science 49, 658-667.

Chikoye, D., Manyong, V.M., Ekeleme, F., 2000. Characteristics of speargrass (Imperata cylindrica) dominated in West Africa: crops, soil properties, farmer perceptions and management strategies. Crop Protection 19, 481-487.

Chou, C.H., Tsai, C.C., 1999. Genetic variation in the intergenic spacer of ribosomal DNA of Imperata cylindrica (L.) Beauv. var. major (Cogongrass) populations in Taiwan. Botanical Bulletin of Academia Sinica 40, 319-327.

Clewell, A.F., 1985. Guide to the vascular plants of the Florida Panhandle. Tallahassee, FL: Florida State University Press, 605p.

Coile, N.C., Shilling, D.G., 1993. Cogongrass, Imperata cylindrica (L.) Beauv.: a good grass gone bad! Botany Circular No. 28. Gainesville, FL: Florida Department of Agriculture and Consumer Services, Division of Plant Industry, 3p.

Cousens, R., Croft, A.M., 2000. Weed populations and pathogens. Weed Research 40, 63-82.

Cummins, G.B., 1971. The rust fungi of cereals, grasses, and bamboos. Springer-Verlag, New York.

Daigle, D.J., Connick, W.J. Jr, Quimby, P.C. Jr, Evans, J., Trask-Morrell, B., Fulgham, F.E., 1990. Invert emulsion: carrier and water source for the mycoherbicide, Alternaria cassiae. Weed Technology 4, 327-331. de Jong, M.D., 2000. The BioChon story: deployment of Chondrostereum purpureum to suppress stump sprouting in hardwoods. Mycologist 14, 58-62. de Jong, M.D., Scheepens, P.C., Zadoks, J.C., 1990. Risk analysis for biological control: a Dutch case study in biocontrol of Prunus serotina by the fungus Chondrostereum purpureum. Plant Disease 74, 189-194. de Nooij, M.P., van Damme, J.M.M., 1988. Variation in host susceptibility among and within populations of Plantago lanceolata L. infected by the fungus Phomopsis subordinaria (Desm.) Trav. Oecologia 75, 535-538.

Dekker, J., 1997. Weed diversity and weed management. Weed Science 45, 357-363.

Den Breeÿen, A., 2004. Release strategies for the establishment of the leaf spot pathogen, Mycovellosiella lantanae var. lantanae, on Lantana camara in South Africa. In: Cullen, J.M., Briese, D.T., Kriticos, D.J., Lonsdale, W.M., Morin, L., Scott, J.K. (Eds.), Proceedings of the XI International Symposium on Biological Control of Weeds. CSIRO Entomology, Canberra, Australia, pp. 386-388.

126

Den Breeÿen, A., Morris, M.J., 2003. Pathogenicity and host specificity of Mycovellosiella lantanae var. lantanae, a potential biocontrol agent for Lantana camara in South Africa. Biocontrol Science and Technology 13, 313-322.

Dhingra, O.D., Sinclair, J.B., 1985. Basic plant pathology methods. CRC Press Inc., Boca Raton, FL.

Dickens, R., 1974. Cogongrass in Alabama after sixty years. Weed Science 22, 177-179.

Ellison, C.A., 1987. Preliminary studies to assess the potential of fungal pathogens for biological control agents of the graminaceous weed Rottboellia cochinchinensis (Lour.) Clayton. MSc thesis. Imperial College, University of London.

Ellison, C.A., 1993. An evaluation of fungal pathogens for biological control of the tropical graminaceous weed Rottboellia cochinchinensis. PhD thesis, University of London.

Ellison, C.A., Evans, H.C., 1990. Preliminary studies to assess the potential of fungal pathogens as biological control agents for Rottboellia cochinchinensis (Graminae). In: Delfosse, E.S. (Ed), Proceedings of the VIII International Symposium on Biological Control of Weeds, Istituto Sperimentale per la Patologia Vegetale, Ministero dell'Agricoltura e delle Foreste, Rome, Italy, pp. 477-482.

Ellison, C.A., Evans, H.C., Ineson, J., 2004. The significance of intraspecifies pathogenicity in the selection of a rust pathotype for the classical biological control of Mikania micrantha (mile-a-minute weed) in Southeast Asia. In: Cullen, J.M., Briese, D.T., Kriticos, D.J., Lonsdale, W.M., Morin, L., Scott, J.K. (Eds.), Proceedings of the XI International Symposium on Biological Control of Weeds. CSIRO Entomology, Canberra, Australia, pp. 103-107.

Ellstrand, N.C., Schierenbeck, K.A., 2000. Hybridization as a stimulus for the evolution of invasiveness in plants? Proceedings of the National Academy of Sciences USA 97, 7043– 7050.

Espiau, C., Riviere, D., Burdon, J.J., Gartner, S., Daclinat, B., Hasan, S., Chaboudez, P., 1998. Host-pathogen diversity in a wild system: Chondrilla juncea-Puccinia chondrillina. Oecologia 113, 133-139.

Eussen, J.H.H., 1980. Biological and ecological aspects of alang-alang [Imperata cylindrica (L.) Beauv.)]. In: Proceedings of BIOTROP workshop on alang-alang in Bogor, 27-29 July 1976. Biotropica Special Publication No. 5, Bogor, Indonesia, pp. 15-22.

Evans, H.C., 1987. Fungal pathogens of some subtropical and tropical weeds and the possibilities for biological control. Biocontrol News and Information 8, 7-30.

Evans, H.C., 1991. Biological control of tropical grassy weeds. In: Baker, F.W.G., Terry, P.J. (Eds.), Tropical Grassy Weeds. CAB International, Wallingford, UK, pp. 52-72.

127

Evans, H.C., Symon, D.E., Roush, R.T., Groves, R.H., Williams, J., Corey, S., 1998. Taxonomy and genotypes of the Rubus fructicosus L. aggregate in Austalia. Plant Protection Quarterly 13, 152-156.

Excoffier, L., Smouse, P.E., Quattro, J.M., 1992. Analysis of molecular variance inferred from metric distance among DNA haplotypes: applications to human mitochondrial DNA restriction data. Genetics 131, 479-491.

Felsenstein, J., 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39, 783-791.

Forno, I.W., Harley, K.L.S., 1979. Occurrence of Salvinia molesta in Brazil. Aquatic Botany 6, 185-187.

Freeman, T.E., 1981. Use of conidial fungi in biological control. In: Cole, G.T., Kendrick, B. (Eds.), Biology of Conidial Fungi,Vol. 2. Academic Press, New York, USA, pp. 143-165.

Gabel, M.L., 1982. A biosystematic study of the genus Imperata (Graminae: Andropogoneae). PhD. dissertation, Iowa State University, Ames, Iowa, USA.

Gaffney, J.F., 1996. Ecophysiological and technological factors influencing the management of cogongrass (Imperata cylindrica). PhD Dissertation, Gainesville, FL: University of Florida, 114p.

Gerlach, J., 2006. Invasive Species Specialist Group Database, Ecology of Clidemia hirta. http://www.issg.org/database/species/ecology.asp.

Ghosheh, H.Z., 2005. Constraints in implementing biological weed control: A review. Weed Biology and Management 5, 83-92.

Global Invasive Species Programme (GISP): Invasive Species and Poverty: Exploring the Links. www.gisp.org/publications/brochures/invasivesandpoverty.pdf.

Goolsby, J.A., Van Klinken, R.D., Palmer, W.A., 2006. Maximising the contribution of native- range studies towards the identification and prioritisation of weed biocontrol agents. Australian Journal of Entomology 45, 276-286.

Greaves, M.P., McQueen, M.D., 1990. The use of mycoherbicides in the field. Aspects of Applied Biology 23, 163-168.

Green, S., 2003. A review of the potential use of bioherbicides to control forest weeds in the UK. Forestry 76(3), 285-298.

Groves, R.H., 1991. The physiology of grassy weeds in tropical agriculture. In: Baker, F.W.G., Terry, P.J. (Eds.), Tropical Grassy Weeds. CAB International, United Kingdom, pp. 39-51.

128

Gupta, M., Chyi, Y-S., Romero-Severson, J., Owen, J.L., 1994. Amplification of DNA markers from evolutionarily diverse genomes using single primers of simple-sequence repeats. Theoretical and Applied Genetics 89, 998-1006.

Hall, D.W., 1978. The grasses of Florida. PhD. Dissertation, Gainesville, FL: University of Florida, 498p.

Hall, D.W., 1983. Weed watch – Cogongrass. Florida Weed Science Society Newsletter 5, 1-3.

Hall, D.W., 1998. Is cogongrass really an exotic? Wildland Weeds 1, 14.

Hasan, S., 1985. Search in Greece and Turkey for Puccinia chondrillina strains suitable to Australian forms of skeleton weed. In: Delfosse, E.S. (Ed.), Proceedings of the VI International Symposium on Biological Control of Weeds. Agriculture Canada, Ottawa, Canada, pp. 625-632.

Hasan, S., Chaboudez, P., Espiau, C., 1995. Isozyme patterns and susceptibility of North American forms of Chondrilla juncea to European strains of the rust fungus, Puccinia chondrillina. In: Delfosse, E.S., Scott, R.R. (Eds.), Proceedings of the VIII International Symposium on Biological Control of Weeds. CSIRO Publishing, Melbourne, Australia, pp. 367-373.

Hodkinson, T.R., Chase, M.W., Takahashi, C., Leitch, I.J., Bennett, M.D., Renvoize, S.A., 2002a. The use of DNA sequencing (ITS and trnL-F), AFLP and fluorescent in-situ hybridisation to study allopolyploid Miscanthus (Poaceae). American Journal of Botany 89, 279–286.

Hodkinson, T.R., Chase, M.W., Lledó, M.D., Salamin, N., Renvoize, S.A., 2002b. Phylogenetics of Miscanthus, Saccharum and related genera (Saccharinae, Andropogoneae, Poaceae) based on DNA sequences from ITS nuclear ribosomal DNA and plastid trnL intron and trnL-F intergenic spacers. Journal of Plant Research 115, 381-392.

Holm, L.G., Plucknett, D.L., Pancho, J.V., Herberger, J.P., 1977. The World’s Worst Weeds: Distribution and Biology. Honolulu,University Press of Hawaii. 609p.

Horsfall, J.G., Cowling, E.B., 1978. Pathometry: The measurement of plant disease. In: Horsfall, J.G., Cowling, E.B. (Eds.), Plant Diseases: An Advance Treatise, Vol. 2. Academic Press, New York, NY, pp. 119-136.

Horsfall, J.G., Barratt, R.W., 1945. An improved grading system for measuring plant disease. Phytopathology Abstracts 35, 655.

Hubbard, C.E., 1944. Imperata cylindrica. Taxonomy, Distribution, Economic Significance, and Control. Imperial Agriculture Bureau Joint Publication No. 7, Imperial Bureau Pastures and Forage Crops, Aberystwyth, Wales, UK.

Invasive Species Specialist Group Database [ISSG], 2005. Ecology of Passiflora tarminiana. http://www.issg.org/database/species/ecology.asp.

129

Invasive Species Specialist Group Database [ISSG], 2006. Ecology of Hedychium gardnerianum. http://www.issg.org/database/species/ecology.asp.

Ivens, G.W., 1983. The natural control of Imperata cylindrica Nigeria and Northern Thailand. Mountain Research and Development 3(4), 372-377.

James, C., 1971. A Manual of Assessment Keys for Plant Diseases. Canada Department of Agriculture, Ottawa, Canada. American Phytopathological Society, St. Paul, MN, 78p.

Ker, A., 1995. Farming systems of the African savanna: a continent in crisis. International Development Research Centre, Ottawa, ON, Canada, 167p.

King, L.M., Schaal, B.A., 1989. Ribosomal-DNA variation and distribution in Rudbeckia missouriensis. Evolution 43, 1117-1119.

King, S.E., Grace, J.B., 2000a. The effects of gap size and disturbance type on invasion of wet pine savanna by cogongrass, Imperata cylindrica (Poaceae). American Journal of Botany 87(9), 1279-1286.

King, S.E., Grace, J.B., 2000b. The effects of soil flooding on the establishment of cogongrass (Imperata cylindrica), a nonindigenous invader of the southeastern United States. Wetlands 20(2), 300-306.

Klein, T.A., Auld, B.A., Fang, W., 1995. Evaluation of oil suspension emulsion of Colletotrichum orbiculare as a mycoherbicide in field trials. Crop Protection 14 (3), 193- 197.

Koch, W., Grobmann, F., Weber, A., Lutzeyer, H.J., Akobundu, I.O., 1990. Weeds as components of maize/cassava cropping systems. In: Standortgemaesse Landwirtschaft in West Africa (IITA, Ibadan and University of Hohenheim, Stuttgard, Germany), pp. 219- 244.

Kolbe, J.J., Glor, R.E., Schettino, L.R., Lara, A.C., Larson, A., Losos, J.B., 2004. Genetic variation increases during biological invasion by a Cuban lizard. Nature 431, 177-181.

Latunde-Dada, A.O., O'Connell, R.J., Nash, C., Pring, R.J., Lucas, J.A., Bailey, J.A., 1996. Infection process and identity of the hemibiotrophic anthracnose fungus (Colletotrichum destructivum) from cowpea (Vigna unguiculata). Mycological Research 100, 1133-1141.

Li, W.G., Shen, J.J., Wang, J.B., 2005. Genetic diversity of the annual weed Monochoria vaginalis in southern China detected by random amplified polymorphic DNA and inter- simple sequence repeat analysis. Weed Research 45, 424-430.

Lippincott, C.L., 1997. Ecological consequences of Imperata cylindrica (cogongrass) invasion in Florida sandhill. PhD. Dissertation, Gainesville, FL: University of Florida, 165p.

Littlefield, L.J., 1981. Biology of the Plant Rusts: An Introduction. Ames: Iowa State University Press, 103p.

130

Louda S.M., Arnett A.E., 1999. In: Spencer, N.R. (Ed). Predicting Non-Target Ecological Effects of Biological Control Agents: Evidence from Rhinocyllus conicus, Proceedings of the X International Symposium on Biological Control of Weeds, 4-14 July 1999. Montana State University, Bozeman, Montana, USA, pp. 551-567.

Luster, D.G., Berthier, Y.T., Bruckart, W.L., Hack, M.A., 1999. In: Spencer, N.R. (Ed), Post- Release Spread of Musk Thistle Rust Monitored from Virginia to California Using DNA Sequence Information, Program Abstracts, Proceedings of the X International Symposium on Biological Control of Weeds, 4-14 July 1999. Montana State University, Bozeman, Montana, USA, p. 75.

Luttrell, E.S., 1974. Parasitism of fungi on vascular plants. Mycologia 66 (1), 1-15.

MacDonald, G.E., 2004. Cogongrass (Imperata cylindrica) – Biology, ecology, and management. Critical Reviews in Plant Sciences 23(5), 367-380.

MacDonald, G.E., Shilling, D.G., Brecke, B.J., Gaffney, J.F., Langeland, K.A., Ducar, T.A., 2002. Weeds in the sunshine: Cogongrass [Imperata cylindrica (L.) Beauv.] biology, ecology and management in Florida. University of Florida IFAS Extension Publication SS- AGR-52.

Maddison, D., Maddison, W., 2000. MacClade 4: Analysis of phylogeny and character analysis. CD-ROM, Sinauer Associates, Inc.

Makowski, R.M.D., Mortensen, K., 1990. Colletotrichum gloeosporioides f. sp. malvae as a bioherbicide for roundleaved mallow (Malva pusilla): conditions for successful control in the field. In: Delfosse, E.S. (Ed.), Proceedings of VII International Symposium on Biological Control ofWeeds, Rome: Istituto Sperimentale per la Patologia Vegetale, Ministero dell’Agricoltura e delle Foreste, pp. 513-522.

Makowski, R.M.D., Mortensen, K., 1992. The first mycoherbicide in Canada: Colletotrichum gloeosporioides f. sp. malvae for round-leaved mallow control. Proceedings of the First International Weed Control Congress 2, 298-300.

Mangoendihardjo, S., 1980. Some notes on the natural enemies of Alang-alang [Imperata cylindrica (L.). Beauv.] in Java. In: Soewardi, B. (Ed), Proceedings of BIOTROP Workshop on Alang-alang, Bogor, Indonesia. BIOTROP SEAMO Regional Center for Tropical Biology, Bogor, Indonesia, pp. 47-55.

McDonald, S.K., Shilling, D.G., Bewick, T.A., Gordon, D., Hall, D., Smith, R., 1996. Factors influencing cogongrass, Imperata cylindrica (L.) Beauv., dispersion, establishment, and persistence. Weed Science Society of America Abstracts 36, 46.

McGinley, M., 2007. Invasive alien species and development challenges in Africa. In: Cleveland, C.J. (Ed.), Encyclopedia of Earth. Washington, D.C.: Environmental Information Coalition, National Council for Science and the Environment. http://www.eoearth.org/article/Invasive_alien_species_and_development_challenges_in_A rica

131

McRoberts, N., Finch, R.P., Sinclair, W., 1999. Assessing the significance of molecular diversity data in natural plant populations. Journal of Experimental Botany 50, 1635-1645.

McRoberts, N., Sinclair, W., McPherson, A., Franke, A.C., Saharan, R.P., Malik, R.K., Singh, S., Marshall, G., 2005. An assessment of genetic diversity within and between populations of Phalaris minor using ISSR markers. Weed Research 45, 431-439.

Miller, M.P., 1997. Tools for population genetics analysis (TFPGA) 1.3: a Windows program for the analysis of allozyme and molecular population genetic data. Computer software distributed by author; http://bioweb.usu.edu/mpmbio/tfpga.htm.

Milne, R.I., Abbott, R.T., 2004. Geographic origin and taxonomic status of the invasive Privet, Ligustrum robustum (Oleaceae), in the Mascarene Islands, determined by chloroplast DNA and RAPDs. Heredity 92, 78-87.

Minno, M.C., Minno, M., 1998. Survey of herbivorous arthropods and pathogens of cogongrass in Florida. Final Report to D. Shilling, University of Florida, Jay, FL and J. Meeker, FDACS, Gainesville FL.

Morin, L., Auld, B.A., Brown, J.F., 1993. Host-range of Puccinia xanthii and post-penetration development on Xanthium occidentale. Canadian Journal of Botany 71, 959-965.

Morin, L., Edwards, P.B., 2006. Selection of biological control agents for bridal creeper – a retrospective review. Australian Journal of Entomology 45, 286-290.

Morin, L., Evans, K.J., Sheppard, A.W., 2006. Selection of pathogen agents in weed biological control: critical issues and peculiarities in relation to arthropod agents. Australian Journal of Entomology 45, 349-365.

Morin, L., Hill, R.L., Matayoshi, S., 1997. Hawaii’s successful biological control strategy for mist flower (Ageratina riparia) - Can it be transferred to New Zealand? Biocontrol News and Information 18, 77N-88N.

Morin, L., Syrett, P., 1996. Prospects for biological control of Hieracium pilosella with the rust Puccinia hieracii var. piloselloidarum in New Zealand. In: Moran, V.C., Hoffmann, J.H. (Eds.), Proceedings of the IX International Symposium on Biological Control of Weeds. University of Cape Town, Cape Town, South Africa, pp. 199-204.

Morin, L., Willis, A.J., Armstrong, A., Kriticos, D., 2002. Spread, epidemic development and impact of the bridal creeper rust in Australia: summary of results. In: Spafford Jacob, H., Dodd, J., Moore, J.H. (Eds.), Proceedings of the 13th Australian Weeds Conference. Plant Protection Society of WA Inc., Perth, W.A., pp. 385–388.

Morin, M., 1995. Vascular plants of the United States. In: La Roe, E.T., Farris, G.S., Puckett, C.E., Doran, P.D., Mac, M.J. (Eds.), Our Living Resources: A Report to the Nation on Distribution, Abundance, and Health of U.S. Plants, Animals, and Ecosystems. U.S. Department of the Interior, National Biological Service, Washington, D.C., pp. 200-205.

132

Morris, M.J., 1987. Biology of the Acacia gall rust, Uromycladium tepperianum. Plant Pathology 36, 100–106.

Morris, M.J., 1989. A method for controlling Hakea sericea Schrad. seedlings using the fungus Colletotrichum gloeosporioides (Penz.) Sacc. Weed Research 29 (6), 449-454.

Morris, M.J., 1997. Impact of the gall-forming rust fungus Uromycladium tepperianum on the invasive tree Acacia saligna in South Africa. Biological Control 10, 75-82.

Morris, M.J., 1999. The contribution of the gall-forming rust fungus Uromycladium tepperianum (Sacc.) McAlp. to the biological control of Acacia saligna (Labill.) Wendl. (Fabaceae) in South Africa. African Entomology Memoir 1, 125–128.

Morris, M.J., Wood, A.R., den Breeÿen, A., 1999. Plant pathogens and biological controlof weeds in South Africa: a review of projects and progress during the last decade. African Entomology Memoir 1, 129–137.

Müller-Schärer, H., Lewisohn, T.M., Lawton, J.H. 1991. Searching for weed biocontrol agents – when to move on? Biocontrol Science and Technology 1, 271-280.

Müller-Schärer, H., Schaffner, U., Steinger, T., 2004. Evolution in invasive plants: implications for biological control. Trends in Ecology and Evolution 19(8), 417-422.

Murniati, G., 2002. From Imperata cylindrica grasslands to productive agroforestry. PhD. Dissertation, Aula, Wageningen: Wageningen University, 172p.

Nei, M., 1978. Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics 89, 583-590.

Nissen, S.J., Masters, R.A., Lee, D.J., Rowe, M.L., 1995. DNA-based marker systems to determine genetic diversity of weedy species and their application to biocontrol. Weed Science 43, 504-513.

O’Connell, R.J., Bailey, J.A., 1991. Hemibiotrophy in Colletotrichum lindemuthianum. In: Mendgen, K., Leseman, D.E. (Eds.), Electron Microscopy of Plant Pathogens. Springer- Verlag: New York, pp. 211-222.

O’Hanlon, P.C., Briese, D.T., Peakall, R., 2000a. Know your enemy: the use of molecular ecology in the Onopordum biological control project. In: Spencer, N.R. (Ed.), Proceedings of the X International Symposium of Biological Control of Weeds. Montana State University, Montana, pp. 281-288.

O’Hanlon, P.C., Peakall, R., Briese, D.T., 2000b. A review of new PCR-based genetic markers and their utility to weed ecology. Weed Research 40, 239-254.

O’Connell, R.J., Uronu, A.B., Waksman, G., Nash, C., Keon, J.P.R., Bailey, J.A., 1993. Hemibiotrophic infection of Pisum sativum by Colletotrichum truncatum. Plant Pathology 42, 774-783.

133

Ohwi, J., 1965. Flora of Japan. Washington, DC: Smithsonian Institution. 1067p.

Olayide, S.O., 1980. Characteristics, problems and significance of farmers. In: Olayide, S.O., Eweka, J.A., Bello-Osagie, V.E. (Eds.), Nigerian small farmers' problems and prospects in integrated rural development, Centre for Agricultural Rural and Development, University of Ibadan, Ibadan, Nigeria. pp. 2–9.

Persley, G.J., 1973. Naturally occurring alternative hosts of Xanthomonas albilineans in Queensland. Plant Disease Reporter 57, 1040-1042.

Peters, H.A., 2001. Clidemia hirta invasion at the Pasoh Forest Reserve: an unexpected plant invasion in an undisturbed tropical forest. Biotropica 33 (1), 60–68.

Pimental, D., Zuniga, R., Morrison, D., 2005. Update on the environmental and economic costs associated with alien-invasive species in the United States. Ecological Economics 52, 273- 288.

Quimby, P.C., Jr., 1982. Impact of diseases on plant populations. In: Charudattan, R., Walker, H.L., (Eds.), Biological control of weeds with plant pathogens. John Wiley & Sons: New York, pp. 47-60.

Randall, J.M., 1996. Weed control for the preservation of biological diversity. Weed Technology 10, 370-383.

Reeder, R.H., Ellison, C.A., 1999. Estado actual de la investigación en control biológico clásico de la Rottboellia cochinchinensis con el carbón Sporisorium ophiuri: potencial y riesgos.. In: Sánchez, V., (Ed), Control biológico de Rottboellia cochinchinensis. Centro Agronómico Tropical de Enseñanza (CATIE), Informe Técnico No. 308, Turrialba, Costa Rica, pp. 101-135.

Reeder, R.H., Ellison, C.A., Thomas, M.B., 1996. Population dynamic aspects of the interaction between the week Rottobellia cocinchinensis (itch grass) and the potential biological control agent Sporisorium ophiuri (head smut). In: Moran, V. C., Hoffman, J. H. (Eds.), Proceedings of the IX International Symposium on Biological Control of Weeds. University of Cape Town, Rondebosch, South Africa, pp. 205–211.

Salamin, N., Hodkinson, T.R., Savolainen, V., 2002. Building supertrees: an empirical assessment using the grass family (Poaceae). Systematic Biology 51, 136–150.

Sankaran, K.V., Sutton, B.C., 1991. Myrellina imperatae sp. Nov. on Imperata from Papua New Guinea and Australia. Mycological Research 95, 1021-1022.

Santiago, A., 1965. Studies on the autoecology of Imperata cylindrica (L.) Beauv. In: Proceedings of the Ninth International Grassland Congress, San Paulo, Brazil, pp. 499- 502.

Santiago, A., 1980. Gene ecological aspects for the Imperata weed and practical implications. BIOTROP 5, 23-34.

134

Schmitz, D.C., Brown, T.C., 1994. As assessment of invasive non-indigenous species in Florida’s public lands. Technical Report No. TSS-94-100. Florida Department of Environmental Protection, Tallahassee, Florida, USA.

Schneider, S., Kueffer, J.M., Roessli, D., Excoffier, L. 1997. Arlequin: a software for population genetic data analysis, version 1.1. Genetics and Biometry Laboratory Department of Anthropology, University of Geneva, Geneva, Switerzerland.

Sharon, A., 1998. The potential of weed-pathogenic Colletotrichum species as mycoherbicides (abstract). Phytoparasitica 26, 4.

Shilling, D.G., Bewick, T.A., Gaffney, J.F., McDonald, S.K., Chase, C.A., Johnson, E.R.R.L., 1997. Ecology, Physiology, and Management of Cogongrass (Imperata cylindrica). Final Report. Florida Institute of Phosphate Research, IFAS, University of Florida, Gainesville FL. 128p.

Smith, K.L., 1989. Epidemiology of gray leaf spot of field corn caused by Cercospera zeae- maydis. PhD. dissertation, University of Maryland.

Smith, R.J., Jr., 1982. Integration of microbial herbicides with existing pest management programs. In: Charudattan, R., Walker, H.L. (Eds.), Biological Control of Weeds with Plant Pathogens. John Wiley & Sons, New York, pp. 189-203.

Soenarjo, E., 1986. Alang-alang gall midge potential as an alternate host for parasitoids. International Rice Research Newsletter 11(5), 22-23.

Soltis, P.S., Soltis, D.E., Chase, M.W., 1999. Angiosperm phylogeny inferred from multiple genes as a tool for comparative biology. Nature 402, 402–404.

Sterling, T.M., Thompson, D.C., Abbott, L.B., 2004. Implications of invasive plant variation for weed management. Weed Technology 18, 1319-1324.

Sun, J.H., Li, Z-C., Jewett, D.K., Britton, K.O., Ye, W.H., Ge, X-J., 2005. Genetic diversity of Pueraria lobata (kudzu) and closely repeated taxa as revealed by inter-simple sequence repeat analysis. Weed Research 45, 255-260.

Swofford, D.L., 1998. PAUP*: Phylogenetic analysis using parsimony (*and other methods) version 4.0. Sinauer Associates, Inc.

Taberlet, P., Gielly, L., Pautou, G., Bouvet, J., 1991. Universal primers for amplification of three non-coding regions of chloroplast DNA. Plant Molecular Biology 17, 1105-1109.

Tabor, P., 1949. Cogongrass, Imperata cylindrica (L.) Beauv. in the southeastern United States. Agronomy Journal 41, 270.

Tanner, G.W., Werner, M.R., 1986. Cogongrass in Florida - an encroaching problem. Florida Cooperative Extension Service Fact Sheet WRS-5, 4p.

135

TeBeest, D.O., Templeton, G.E., 1985. Mycoherbicides: progress in the biological control of weeds. Plant Disease 69, 6-10.

TeBeest, D.O., Yang, X., Cisar, C.R., 1992. The status of biological control of weeds with fungal pathogens. Annual Review of Phytopathology 30, 637-657.

Templeton, G.E., 1982. Status of weed control with plant pathogens. In: Charudattan, R., Walker, H.L. (Eds), Biological Control of Weeds with Plant Pathogens. John Wiley & Sons, Inc., New York, pp. 29-44.

Templeton, G.E., 1992. Use of Colletotrichum strains as mycoherbicides. In: Colletotrichum: Biology, Pathology and Control. British Society of Plant Pathology, UK, pp. 358 – 380.

Terry, P.J., Adjers, G, Akobundu, I.O., Anoka, A.U., Drilling, M.E., Tjitrosemito, S., Utomo, M., 1997. Herbicides and mechanical control of Imperata cylindrica as a first step in grassland rehabilitation. Agroforestry Systems 36, 151-179.

Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F., Higgins, D.G., 1997. The CLUSTAL_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Research 25, 4876–4882.

Thrall, P.H., Burdon, J.J., 2004. Host-pathogen life history interactions affect biological control success. Weed Technology 18 (5), 1269-1274.

Trujillo, E.E., 1985. Biological control of Hamakua pa-makani with Cercosporella sp. in Hawaii. In: Delfosse, E.S. (Ed.), Proceedings of the VI International Symposium on Biological Control of Weeds. Agriculture Canada, Ottawa, Canada, pp.661-671.

Trujillo, E.E., 2005. History and success of plant pathogens for biological control of introduced weeds in Hawaii. Biological Control 33, 113-122.

Trujillo, E.E., Norman, D.J., Killgore, E.M., 1994. Septoria leaf spot, a potential biocontrol control of banana poke vine in forests of Hawaii. Plant Disease 78, 883-885.

Udensi, E.U., 1994. Mucuna (Mucuna pruriens var utilis [Wight] Burck) for control of speargrass (Imperata cylindrica [L.] Raeuschel) in a derived savannah environment. MSc thesis, IITA.

Udensi, E.U., Akobundu, I.O., Ayeni, A.O, Chikoye, D., 1999. Management of Imperata cylindrica with velvetbean (Macuna pruriens var. utilis) and herbicides. Weed Technology 13, 201-208.

USDA-ARS., 2001. U.S. Department of Agriculture, ARS National Programs, Crop Protection and Quarantine. http://www.nps.ars.usda.gov/programs.htm?npnumber=304

136

Valverde, B.E., Merayo, A., Reeder, R., Riches, C.R., 1999. Integrated management of itchgrass (Rottboellia cochinchinensis) in maize in seasonally dry Central America: Facts and perspectives. In: Proceedings of the Brighton Crop Protection Conference—Weeds, Farnham, United Kingdom: British Crop Protection Council, pp. 131–140.

Van Klinken, R.D., Julien, M.H., 2003. Learning from past attempts: does classical biological control of Noogoora burr (Asteraceae: Xanthium occidentale) have a promising future? Biocontrol Science and Technology 13, 139-153.

Van Loan, A.N., Meeker, J.R., Minno, M.C., 2002. Cogon Grass. In: Van Driesche, R. et al.(Eds.), Biological Control of Invasive Plants in the Eastern United States. USDA Forest Service Publication FHTET-2002-04, 413p.

Vayssiere, P., 1957. Les mauvaises herbes en Indo-Malasie. Journal d’ Agriculture Tropical et de Botanique appliqué 4 (9-10), 392-401.

Walker, H.L., Boyette, C.D., 1986. Influence of sequential dew periods on biocontrol of sicklepod (Cassia obtusifolia) by Alternaria cassiae. Plant Disease 70, 962-963.

Wang, T., Su, Y., Chen, G., 2007. Population genetic variation and structure of the invasive weed Mikania micrantha in Southern China: consequences of rapid range expansion. Journal of Heredity Advanced Access, September 28, 2007. doi:10.1093/jhered/esm080.

Wapshere, A.J., Delfosse, E.S, Cullen, J.M., 1989. Recent developments in biological control of weeds. Crop Protection 8, 227-250.

Wardill, T.J., Graham, G.C., Zalucki, M., Palmer, W.A., Payford, J., Scott, K.D., 2005. The importance of species identity in the biocontrol process: identifying the subspecies of Acacia nilotica (Leguminosae:Mimosoideae) by genetic distance and the implications for biological control. Journal of Biogeography 32, 2145-2159.

Warwick, S.I., 1990. Genetic variation in weeds - with particular reference to Canadian agricultural weeds. In: Kawano, S. (Ed.), Biological Approaches and Evolutionary Trends in Plants. London: Academic Press, pp. 3–18.

Waterhouse, D.F., 1999. When is classical biological control the preferred option for exotic weeds? In: Proceedings of the Symposium of Biological Control in the Tropics, MARDI Training Centre, Serdang, Selangor, 18-19 March 1999, Malaysia. CABI Publishing, Kuala Lumpur, Malaysia, pp. 52-53.

Wei, Y.D., Byer, K.N., Goodwin, P.H., 1997. Hemibiotrophic infection of round-leaved mallow by Colletotrichum gloeosporioides f. sp. malvae in relation to leaf senescence and reducing reagents. Mycological Research 101, 357-364.

Wester, L.L., Wood, H.B., 1977. Koster’s curse (Clidemia hirta), a weed pest in Hawaiian Forest. Environmental Conservation 4, 35-41.

137

Wilcut, J.W., Dute, R.R., Truelove, B., Davis, D.E., 1988. Factors limiting the distribution of cogongrass, Imperata cylindrica, and torpedograss, Panicum repens. Weed Science. 36, 577-582.

Wolfe, A.D., Liston, A., 1998. Contributions of PCR-based methods to plant systematics and evolutionary biology. In: Soltis D.E., Soltis, P.S., Doyle, J.J. (Eds.), Molecular systematics of plants. II. DNA sequencing. Chapman and Hall:New York, pp. 43-86.

Wolfe, A.D., Xiang, Q-Y, Kephart, S.R., 1998a. Diploid hybrid speciation in Penstemon (Scrophulariaceae). Proceedings of the National Academy of Sciences of the USA 95, 5112-5115.

Wolfe, A.D., Xiang, Q-Y, Kephart, S.R., 1998b. Assessing hybridization in natural populations of Penstemon (Scrophulariaceae) using hypervariable inter-simple sequence repeat markers. Molecular Ecology 7, 1107-1125.

Womack, J.G., Burge, M.N., 1993. Mycoherbicide formulation and the potential for bracken control. Pesticide Science 37, 337-341.

Womack, J.G., Eccleston, G.M., Burge, M.N., 1996. A vegetable oil-based invert emulsion for mycoherbicide delivery. Biological Control 6, 23-28.

Wood, A.W., Morris, M.J., 2007. Impact of the gall-forming rust fungus Uromycladium tepperianum on the invasive tree Acacia saligna in South Africa: 15 years of monitoring. Biological Control 41, 68-77.

Yandoc, C.B., 2001. Biological Control of Cogongrass, Imperata cylindrica (L.) Beauv. PhD Dissertation, Gainesville, FL: University of Florida, 194p.

Yandoc, C.B., Charudattan, R., 1998. Use of natural substrates for the production of fungal inoculum for biological weed control studies. Weed Science Society of America Abstracts 38, 80.

Yandoc, C.B., Charudattan, R., Shilling, D.G., 2005. Evaluation of fungal pathogens as biological control agents for cogongrass (Imperata cylindrica). Weed Technology 19, 19- 26.

Yandoc-Ables, C.B., Rosskopf, E.N., Charudattan, R., 2006. Plant pathogens at work: progress and possibilities for weed biocontrol, Part 1: Classical vs. bioherbicidal approach. APSNET online feature, August 2006, http://apsnet.org/online/feature/weed1.

Yang, S., Johnson, D.R., Dowler, W.M., Connick, W.J. Jr., 1993. Infection of leafy spurge by Alternaria alternata and A. angustiovoidea in the absence of dew. Phytopathology 83, 953- 958.

Ye, W.H., Mu, H.P., Cao, H.L., Ge, X.J., 2004. Genetic structure of the invasive Chromolaena odorata in China. Weed Research 44, 129-135.

138

Zietkiewicz, E., Rafalski, A, Labuda, D., 1994. Genome fingerprinting by simple sequence repeat (SSR)-anchored polymerase chain reaction amplification. Genomics 20, 176-183.

139

BIOGRAPHICAL SKETCH

Alana Den Breeÿen was born in Cape Town, South Africa, in 1965. She entered the

University of Stellenbosch at Stellenbosch in 1987, majoring in plant pathology and botany and

received her Bachelor of Science degree in natural sciences in 1990.

In 1991, she started a master’s program at the Department of Plant Pathology of the

University of Stellenbosch and graduated in March 1994. She worked as a research assistant for

Dr. Pedro Crous at the University of Stellenbosch where she was involved in mycological research and taught an Introductory Mycology to undergraduate agriculture students. She worked as a researcher at the Agricultural Research Council – Plant Protection Research Institute based in Stellenbosch from 1997 to 2003. She worked on several fungal biocontrol projects using classical and bioherbicide control strategies on invasive weed species including Lantana camara, Chromolaena odorata and Eichhornia crassipes (waterhyacinth). Her research involved traveling to exotic places in Brazil, Jamaica, Cuba, and China to survey for and collect fungal pathogens associated with the above-mentioned invasive weeds.

In 2004, she embarked on a new adventure when she started on a doctoral program at the

Plant Pathology Department at the University of Florida. She won the Outstanding Graduate

Student Award in 2004 and the Davidson Travel Scholarship in 2007. She served as president of the Plant Pathology Graduate Student Association from August 2005 to June 2006. She is a student member of the American Phytopathological Society (APS) and the Florida and

International Weed Science Societies. She also served on the APS Joint Committee of Women in Plant Pathology and Cultural Diversity in 2005-2006.

140