Biological Control of Invasive Plants in the Eastern United States

IQ22ƒ€y„„ih2uxe€‡iih FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF ITW IR22viep‰2ƒ€…‚qi FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF IVI IS22g‰€‚iƒƒ2ƒ€…‚qi FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF IWS IT22ƒ‡evvy‡E‡y‚„ƒ FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF PHW IU2gexehe2„rsƒ„vi FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF PIU IV22w…ƒu2„rsƒ„vi2@xyhhsxq2„rsƒ„viA FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF PPW IW22f…vv2„rsƒ„vi2@ƒ€ie‚2„rsƒ„viA FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF PRU PH22ƒvixhi‚pvy‡i‚2„rsƒ„vi2@‡sxqih2ƒvixhi‚2y‚2ƒieƒshi2„rsƒ„viA FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF PSI PI22€v…wiviƒƒ2„rsƒ„vi2@g…‚vih2„rsƒ„viD2f‚sƒ„v‰2„rsƒ„viA FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF PSS ƒig„syx2s†X2‡iihƒ2yp2yvh2psivhƒ2exh2€eƒ„…‚iƒ FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF PTQ PP22w…v„spvy‚e2‚yƒi FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF PTS PQ22„‚y€sgev2ƒyhe2e€€viD2‡i„vexh2xsqr„ƒrehiD2exh2„…‚ui‰2fi‚‚‰ FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF PWQ PR22f‚esvsex2€i€€i‚„‚ii FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF QII ƒig„syx2†X2‡iihƒ2yp2py‚iƒ„ƒ FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF QPQ PS22u…h… FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF QPS PT22wsviEeEwsx…„i2‡iih FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF QQI PU22ƒu…xu2†sxi FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF QRQ PV22gyqyx2q‚eƒƒ FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF QSQ PW22qe‚vsg2w…ƒ„e‚h FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF QTS iii Biological Control of Invasive Plants in the Eastern United States iv ƒig„syx2†sX2gyxgv…ƒsyx FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF QUQ QH22ƒivig„syx2yp2e€€‚y€‚se„i2p…„…‚i2„e‚qi„2‡iihƒ2py‚2fsyvyqsgev2gyx„‚yv FFFFFFFFFFFFFFFFFFFFFFFFFF QUS Native Weeds vs. Introduced Weeds .................................................................................... 375 Selecting Targets to Minimize Risk to Non-Target Organisms......................................... 376 Selecting Target Weeds to Promote Success ........................................................................ 377 Predicting Successful Biological Control .............................................................................. 378 How Targets Are Selected ...................................................................................................... 379 Future Targets for Biological Control of Weeds .................................................................. 379 References ................................................................................................................................. 385 e…„ry‚ƒ FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF QVU sxhiˆ FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF QWI I NTRODUCTION Invasive plants are species that, after they have been moved from their native habitat to a new location, spread on their own. Some invasive plants reach high densities and cause economic or environmental harm or harm to humans (National Invasive Species Coun- cil, 2001). Problems caused by invasive species have in- creased dramatically in recent decades due, in part, to an increasing human population (e.g., increased international travel, globalization of world trade). In response, the President issued Executive Order 13112 on Invasive Species in February 1999. The Order established the National Invasive Species Council to provide national leadership to prevent the introduc- tion of invasive species and provide for their control and to minimize the economic, ecological, and hu- man health impacts that invasive species cause. The Weed Science Society of America recog- nizes about 2,100 invasive plant species (i.e., noxious or weedy plants) in the United States and Canada. Currently, 94 kinds of invasive plant species are offi- cially recognized as Federal Noxious Weeds and many more species are designated on State noxious weed lists. In the United States, invasive plant spe- cies comprise from 8 to 47 percent of the total flora of most States. In 1994, the economic impact of weeds on the United States economy was estimated to be $20 billion annually (Westbrooks, 1998). Once an invasive plant species becomes estab- lished it is not easily suppressed nor eliminated as these species often possess characteristics that favor their population increase, such as early maturation, profuse reproduction by seeds and/or vegetative structures, long life of seeds in the soil, adaptation for spread, and production of biological toxins that suppress the growth of other plants. In addition, many invasive plants are free of attack in their in- vaded range by specialized insects or plant pathogens, allowing plant resources to be shifted from defense to growth and reproduction. Integrated invasive plant management relies on a combination of control technologies. These include biological, mechanical, chemical, and cultural appli- cations. Before the mid-1950s, chemical and mechani- cal applications were the main tactics used to sup- press invasive plants in the continental United States. In the 1940s, classical biological plant control efforts were initiated and significantly increased in the United States and since then, biological control has become the most widely used tactic for weed sup- pression. Biological control is the science of reconnect- ing invasive plants with the specialized natural en- emies that often limit their density in their native ranges. This process consists of surveys in the plant’s area of origin to discover candidate natural enemies, studies on their biology and host specificity and re- lease and evaluation of their impacts on the target plant. The U.S. Department of Agriculture’s Animal and Plant Health Inspection Service (APHIS) is re- sponsible for controlling introductions of species brought into the United States for biological control of plants, in accordance with the requirements of sev- eral plant quarantine laws, the National Environmen- tal Policy Act, and the Endangered Species Act. Peti- tions for release of plant biological control agents are judged by a Technical Advisory Committee (TAG), which represents the interests of a diverse set of fed- eral and other agencies. Biological weed control has been most success- ful outside of crop lands, primarily in rangeland, pas- tures, and water bodies. Many projects have been conducted on grazing lands in the semiarid western United States. In the eastern United States, projects have been targeted against aquatic, pasture, and for- est weeds. Projects in the western United States have been summarized previously by the W-84 project (Nechols et al., 1995). No such compilation has yet been done for the eastern United States. The purpose of this book is to provide a refer- ence guide for field workers and land mangers con- cerning the historical and current status of the bio- logical control of invasive plant species in the eastern United States. Weeds associated with lakes, ponds and rivers (Section I); wetlands (Section II); prairies and grasslands (Section III); old fields and pastures (Section IV); and forests (Section V) are discussed. Authors are leaders in research on biological control of the plant species they discuss. Each chapter com- piles published articles, unpublished reports and per- sonal experiences of the authors, and provides the most up-to-date and accurate information concern- ing biological control of each invasive plant species. The choice of plant species included in this book was based on information found in Julien and Griffiths’ World Catalog of Agents and Their Target 1 Biological Control of Invasive Plants in the Eastern United States 2 Weeds (1998) and 5 years (1995-1999) of programs from National Meetings of the Entomological Soci- ety of America. This initial list was reviewed by lead- ing weed biological control scientists (Bernd Blossey, Gary Buckingham, Alex McClay, Loke Kok and Jack DeLoach) before settling on the 31 invasive plant species included here. We provide this information to assist in the plan- ning and execution of weed biological control projects in the region. We believe that weed biologi- cal control projects will increasingly be seen as an essential approach to protecting natural areas, waterbodies, forests, and pastures in the region. REFERENCES Julien, M. H. and M. W. Griffiths (eds.). 1998. Biological Control of Weeds: A World Catalog of Agents and Their Target Weeds. 4th ed. CABI-Bioscience, Wallingford, United Kingdom. National Invasive Species Council. 2001. Meeting the Invasive Species Challenge: Management Plan. Washington, D.C. Nechols,, J. R., L. A. Andres, J W. Beardsley, R. D. Goeden, and C. G. Jackson (eds.). 1995. Biological Control in the Western United States: Accomplish- ments and Benefits of Regional Project W-84, 1964- 1989. Division of Agriculture and Natural Resources, Publication 3361, University of California, Oakland, California. Westbrooks, R. G. 1998. Invasive Plants, Changing the Landscape of America: Fact Book. Federal Inter- agency Committee for the Management of Noxious Weeds, Washington, D.C. SECTION I: WEEDS OF L AKES , P ONDS , AND RIVERS Alligatorweed Floating Fern Water Chestnut Waterhyacinth Waterlettuce Eurasian Watermilfoil Hydrilla 1 ALLIGATORWEED G. R. Buckingham U. S. Department of Agriculture, Agricultural Research Service, Invasive Plants Research Laboratory, Gainesville, Florida, USA PEST STATUS OF WEED Alligatorweed (Alternanthera philoxeroides [Mart.] Griseb.) is a South American immigrant that has in- vaded waterways in the United States, primarily in the southeastern states. It also is a weed in tropical and mild temperate regions around the world. Alligatorweed roots readily along waterways and then grows over the water surface as an anchored floating plant. It also grows terrestrially during dry periods. Alligatorweed is a federal noxious weed and a prohibited or noxious plant in Arizona, California, Florida, and South Carolina (USDA, NRCS, 1999). Nature of Damage Economic damage. Alligatorweed disrupts many eco- nomic uses of water (Anonymous, 1987; Holm et al., 1997). Thick mats prevent drainage canals, ditches, streams, and other small waterways from emptying rapidly during periods of heavy water load, thus caus- ing flooding (Fig. 1). If mats break loose, they create obstructions by piling up against bridges, dams, and sharp bends in waterways. Thick mats also increase mosquito habitat. Navigation of small waterways is obstructed, as is shoreline navigation in large water- ways. Efficiency of irrigation systems is decreased. Fishing and swimming can be affected, although a small fringe of alligatorweed probably benefits fish- ing. A perusal of various commercial Internet sites in April, 2001 indicated that costs would be approxi- mately $170 to $370/ha for control of alligatorweed with the herbicides glyphosate and fluoridone. Ecological damage. Alligatorweed, like many other invasive aquatic plants, displaces native plants in ditches, along banks, and in shallow water (Holm et al., 1997). Vogt et al., (1992) discussed competi- tion with native plants before and after insect biological control agents were released. Alligatorweed disrupts water flow causing increased sedimentation, and it shades submersed plants and animals causing reduced oxygen levels beneath the mat (Quimby and Kay, 1976). Extent of losses. Current data on the extent of infestation and control costs are lacking. At the be- ginning of the biological control program in 1963, there were an estimated 65,723 ha of water infested in eight southern states and 26,933 ha of plants in 1970 (Coulson, 1977). The largest infestation, 22,700 ha, was in Louisiana and the smallest, 21 ha, in Mis- sissippi. In 1981, the infestation in the southern states was estimated to have increased, but only because of increases in Texas and Louisiana (Cofrancesco, 1988). All other states reported a decrease. Even though the infestation estimate had tripled in Louisiana, state officials considered biological control to be satisfac- tory. Much of the increase was due to terrestrial in- vasion by alligatorweed. p2IF22e— 2 —2  2  2 — 2 2™ 2  2 — —D2 ˜  2 — 2 2— 2™—2 F @€ —2™  2…ƒheD2e‚ƒ2˜2qF2‚F f™"— FA 5 Biological Control of Invasive Plants in the Eastern United States 6 p2PF2e— 2—2 D2E ™™ 2 — F2@€ —2™   …ƒheD2e‚ƒFA Geographical Distribution Alligatorweed, a South American native, grows in the coastal plain from Virginia, ca. 37º N, to southern Florida, ca. 25º N, and westward along coastal areas to Texas. It is also found in southern California (Reed, 1970). A distribution map provided by Reed (1970) indicates that the northern limit inland is at about the middle of Alabama, Georgia, and South Caro- lina, ca. 33.5º N, with an extension slightly further north in the warmer Mississippi Valley, ca. 35º N. However, both southwestern Kentucky, ca. 36.5º N, and Tennessee are included within its range on the USDA PLANTS Database on the Internet (USDA, NRCS, 1999). BACKGROUND INFORMATION ON PEST PLANT Taxonomy Alligatorweed is in the tribe Gomphreneae, subfam- ily Gomphrenoideae, family Amaranthaceae (Mabberley, 1997), order Caryophyllales, subclass Caryophyllidae (Cronquist, 1988). There are an es- timated 170 species of Alternanthera in the Western Hemisphere with 120 species in South America alone (Vogt et al., 1979). Less than 5% of the species in South America are amphibious with most being me- sophytic or xerophytic. Kartesz (1994) listed 15 spe- cies of Alternanthera, including ornamentals and im- migrants, in the United States and Canada. There are a few species in Asia, mostly introduced from South America. Engler (1934) included A. philoxeroides in the subgenus Telanthera, section 1. Alternanthera can be differentiated from related aquatic species by the opposite, non-succulent leaves; white flowers in short, headlike spikes; and by the presence of a style (Figs. 2, 3). Wain et al. (1984) reported two diverse forms of alligatorweed – one with slender stems and short, rounded leaves, and the other with broad stems and long, slender leaves. Their isozyme study indi- cated that the genetic difference between the forms was similar to the distances reported between sub- species in other plant studies. The importance of these forms in plant invasion and in control efforts has not been investigated. Julien and Broadbent (1980) listed the synonymy for A. philoxeroides. p2QF2„ 2  2—— 2 2— — — 22—" D2 D2 — " 2" D — 2— 2—2 F2@€ —2™   …ƒheD2e‚ƒ2˜2qF2‚F2f™"— FA Biology Alligatorweed initially roots in wet soil on banks or in shallow water along shorelines and then grows out into waterways. Penfound (1940) reported that be- ginning in March in Alabama, shoots grew to 38 cm in 1.5 months, to 145 cm in 2.5 months., and to 508 cm in 5.5 months. By September, the mat extended up to 4.6 m away from shore. Alligatorweed is a pe- rennial with hollow stems that buoy the shoots. Float- ing mats expand over surfaces of all types of water- ways and are practically impenetrable. If a water- way dries, alligatorweed changes to a terrestrial form with smaller, tougher leaves and stems. Only veg- etative growth has been reported in the United States, although viable seeds have been found in the United Alligatorweed 7 States (Holm et al., 1997). Roots develop at closely spaced nodes along stems. When the stems break, floating sections are able to establish readily on moist soil. Alligatorweed has been reported to reproduce by seeds in South America (Holm et al., 1997). Analysis of Related Native Plants in the Eastern United States There are no native Alternanthera in the aquatic habi- tats of the United States. A second introduced spe- cies, Alternanthera sessilis (L.) R. Br. ex DC., which is pantropical, is reported to be naturalized in the Florida panhandle (Godfrey and Wooten, 1981). As the name implies, the flowers of A. sessilis are sessile compared with flowers of alligatorweed, which are stalked. According to Vogt et al. (1979), the South American Alternanthera pungens Kunth also is es- tablished in the United States. Three additional gen- era in the Amaranthaceae are associated with aquatic habitats. Amaranthus has six species in the range of alligatorweed in the southeastern United States, Iresine has one species, and Blutaparon has one spe- cies, Blutaparon (=Philoxerus) vermiculare (L.) Mears or silverhead, which occurs in maritime habitats. Corell and Corell (1972) placed some of the wetland species of Amaranthus in the genus Acnida. One Amaranthus in the eastern United States, Amaranthus pumilus Raf. (dwarf or seabeach amaranth) is feder- ally listed as threatened. Its range might overlap with the distribution of alligatorweed biological control agents along the coast of South Carolina, but it is found mostly north of their ranges (Godfrey and Wooten, 1981). HISTORY OF BIOLOGICAL CONTROL EFFORTS IN THE EASTERN UNITED STATES In 1959, the U.S. Army Corps of Engineers requested that the U.S. Department of Agriculture, Agricultural Research Service evaluate the potential for biological control of alligatorweed (Zeiger, 1967; Buckingham, 1994). Consequently, field surveys and studies of bi- ologies and host ranges of potential biological con- trol agents were conducted in South America (Buckingham, 1996; Coulson et al., 2000). Ultimately, three insect species, Agasicles hygrophila Selman and Vogt (Coleoptera: Chrysomelidae), Amynothrips andersoni O’Neill (Thysanaptera: Phlaeothripidae), and Arcola (as Vogtia) malloi (Pastrana) (Lepidoptera: Pyralidae, Phycitinae), were introduced into the United States (Coulson, 1977). Area of Origin of Weed Alligatorweed is native along the coast of South America from Venezuela to Buenos Aires Province in Argentina (Vogt et al., 1979). It also was reported from the upper and middle Amazon River basin and the Paraná River basin (Vogt et al., 1979). Sites with alligatorweed were most common in Paraguay, Uru- guay, and northern Argentina, but this might reflect more intensive surveys for biological control agents in those areas. Areas Surveyed for Natural Enemies Annual surveys from 1960 to 1962 covered eastern and northern South America from Argentina to Ven- ezuela, including Trinidad (Coulson, 1977). Addi- tional surveys were conducted in Argentina and Uru- guay (Coulson et al., 2000). Surveys also were con- ducted in the southeastern United States (Coulson, 1977). Natural Enemies Found As many as 40 insect species (not all could be identi- fied to species level) were recorded on alligatorweed (Vogt, 1973). Five of the 40 species were considered to suppress alligatorweed (Vogt, 1973). These five were: A. hygrophila, A. andersoni, A. malloi, Herpetogramma bipunctalis (F.) (Lepidoptera: Pyralidae), and Prodenia sp. (Lepidoptera: Noctuidae) (Maddox et al., 1971; Vogt, 1973). The flea beetle Disonycha argentinensis Jacoby (Co- leoptera: Chrysomelidae) was later considered a po- tential agent for control of terrestrial alligatorweed (Cordo et al., 1984). No species able to suppress the weed were discovered in the United States, but na- tive species of Disonycha and the moth H. bipunctalis were commonly found on alligatorweed in the United States (Vogt et al., 1992). A fungus native to the United States, Nimbya (=Alternaria) alternantherae (Holcomb and Antonopoulus) Simmons and Alcorn (Hyphomycetes), causes purplish leaf spots and can defoliate plants. However, damage is rarely severe (Holcomb, 1978). In Brazil where it also is native, N. alternantherae was not particularly damaging, but Biological Control of Invasive Plants in the Eastern United States 8 preliminary studies indicated possible potential as a mycoherbicide (Barreto et al., 2000). Damage by a second Brazilian fungus, Cercospora alternantherae Ellis and Langlois, was occasionally severe and this species might have potential as a biocontrol agent. (Barreto et al., 2000). A virus-like disease that stunts alligatorweed in Florida has not been studied (Hill and Zettler, 1973). Host Range Tests and Results Four insect species were tested in host range experi- ments in Argentina and Uruguay and in quarantine in Albany, California. Maddox et al. (1971) reported that 14 plant species were tested with the alligatorweed flea beetle, but they did not list the spe- cies. Buckingham (1996) reported that the 14 species were apparently in 12 genera of eight families. These families were Amaranthaceae, Brassicaceae, Chenopodiaceae, Cucurbitaceae, Malvaceae, Nymphaeaceae, Poaceae, and Polygonaceae. For the flea beetle, A. hygrophila, slight adult feeding was found on apical leaves of Chenopodium ambrosioides L. and larval feeding and development on Atriplex patula var. hastata (L.) Gray, but only one malformed adult emerged (Maddox and Resnik, 1969; Maddox et al., 1971). Field observations in South America also played a role in obtaining approval of the flea beetle for introduction into the United States (Anderson, 1963). At least 14 species of aquatic or related plant species in proximity to damaged alligatorweed in Argentina were examined for flea beetles, as were crop plants in the vicinity. No beetles or damage were found. The alligatorweed thrips, A. andersoni, was tested on 21 species in 13 genera of six families (Maddox, 1973). Families were the same as those tested with the flea beetle except Cucurbitaceae and Malvaceae were not tested with the thrips. No devel- opment took place in the no-choice and choice ex- periments except on alligatorweed. Vogt found thrips on the native Alternanthera hassleriana Chod. in Ar- gentina (Maddox et al., 1971), but that species has not been reported as naturalized in the United States. Field examinations in Argentina of 46 other plant species in 26 genera of 11 families yielded no A. andersoni or its damage (Maddox, 1973). The alligatorweed stem borer, A. malloi, was tested in choice and no-choice tests on 30 plant spe- cies in 17 genera of the six families tested with the thrips (Maddox and Hennessey, 1970). Although there was some feeding on test plants in no-choice tests, development of third or younger instars was restricted to alligatorweed. A few older larvae fin- ished their development on five species in the same amaranth tribe as alligatorweed, Gomphrenae. Field examinations of 51 plant species in Argentina from 1962 through 1967 discovered this moth only on alligatorweed (Maddox and Hennessey, 1970). In South American surveys, moths were reared from A. hassleriana and from the closely related Philoxerus portulacoides St. Hil. (Vogt et al., 1992). After the moth was released in the United States, it was reared from the native species Blutaparon vermiculare, col- lected in Louisiana and Texas and from subsequent cage tests (Vogt et al., 1992). However, the numbers reared from Philoxerus and Blutaparon, which are closely related, were small. Pemberton (2000) re- ported rearing this species from Alternanthera flavescens Kunth., a native of coastal hammocks in Florida. The flea beetle, D. argentinensis, was tested on 54 species in 38 genera of 19 families in no-choice larval tests (Cordo et al., 1984). All of the eight fami- lies used in tests with A. hygrophila were included along with additional families containing crop and ornamental species. Flea beetle larvae fed moderately on four species of Amaranthaceae and two species of Chenopodiaceae, but development to adults was re- stricted to Alternanthera paronychioides St. Hilaire (6.4%) and Beta vulgaris L. (3.0%), while 44% com- pleted development on alligatorweed. The develop- ment on beets, B. vulgaris, prevented release of this species in the United States. Interestingly, Australia tested this species with 36 species in 31 genera of 18 families (Sands et al., 1982). In those tests, no devel- opment was observed on beets, but the smaller num- ber of larvae used in the tests (52 versus 234 in the American tests) could have accounted for the slight difference between the two studies. Based upon the American test results (3% development), only one adult would have been expected in the Australian tests and, indeed, one of the 52 larvae did develop to the last instar. Both Australia and New Zealand released this species, but it did not establish (Julien and Griffiths, 1998). No host range tests were conducted with the other two major biotic suppressants listed by Vogt (1973), H. bipunctalis and Prodenia sp. The former, the southern beet webworm, also is native to North Alligatorweed 9 America. The latter pupates in the soil, which pre- vented its use for control of aquatic alligatorweed that was the target of the biological control program (Maddox et al., 1971). If there is future interest in controlling terrestrial alligatorweed, which is com- monly eaten by cattle, the Prodenia sp. could be stud- ied further although it also attacked Amaranthus sp. Releases Made Field-collected alligatorweed flea beetles from Argen- tina were processed through quarantine and released in 1964 in California and South Carolina and in 1965 in Florida (Coulson, 1977). Beetles from Uruguay also were released in South Carolina and a mixed quarantine colony started with beetles from both Ar- gentina and Uruguay was released in Mississippi (1965). Later, beetles collected at release sites, mostly in Florida, were redistributed in Alabama (1967), Arkansas (1969), Georgia (1966), Louisiana (1970), North Carolina (1967), Tennessee (1968), and Texas (1967). Beetles from a quarantine colony held in Gainesville, Florida (of insects originating from Necochea, Argentina) were released in 1979 in Ala- bama, Florida, North Carolina, and South Carolina (Buckingham et al., 1983). A quarantine colony of alligatorweed thrips from Argentina was released in Alabama (1968), Cali- fornia (1967), Florida (1967), Georgia (1967), Mis- sissippi (1968), South Carolina (1967), and Texas (1968) (Coulson, 1977). Eggs from alligatorweed stem borer females collected as larvae in Argentina and held in quaran- tine were first released in Florida and Georgia in 1971 (Coulson, 1977). Eggs from females collected at Necochea, Argentina, and held in quarantine were released in Georgia and South Carolina in 1972 in an attempt to establish more cold-tolerant populations. Most of the other releases in 1971 and 1972 were from quarantine (Albany) or greenhouse (Gainesville) colonies started with part of the eggs collected in Necochea. Releases were made in the preceding states and in Alabama (1972) and North Carolina (1971). A handbook that provides instructions for re- lease of the alligatorweed agents was prepared by the U.S. Army Engineer Waterways Experiment Station, based upon the successes of these releases (Anony- mous, 1981). BIOLOGY AND ECOLOGY OF KEY NATURAL ENEMIES Alligatorweed flea beetle, Agasicles hygrophila Selman and Vogt (Coleoptera: Chrysomelidae) Adults are small (4 to 6 mm long), black and yellow striped beetles that jump when disturbed (Selman and Vogt, 1971). Feeding causes “shot holes” in the leaves, but with heavy adult and larval feeding the leaves are completely eaten, as are upper portions of the stems (Fig. 4). Females deposited clusters of 12 to 54 eggs in two contiguous rows on the underside of apical leaves (Maddox, 1968). The yellowish eggs hatched in four days at 20 to 30 ºC. Dark colored larvae ate leaf tissue but often left one epidermis intact, creat- ing a window in the leaf. The three instars developed in eight days at 20 to 30 ºC. Mature larvae pupated one to two days after entering stems. The pupal pe- riod lasted five days and a premating and preovipo- sition period lasted about six days. The total life cycle from egg to egg was completed in 25 days at 20 to 30 ºC. Females lived about 48 days and usually depos- ited only one egg cluster per day for an average of 1,127 total eggs (Maddox, 1968). Beetles were multi- voltine near Buenos Aires, Argentina, producing five generations per year (Maddox, 1968) and probably four to six generations in Florida and the lower Mis- sissippi River Valley (Coulson, 1977; Vogt et al.,1992). p2RF22e 2— 2— — 22 2——  —2˜  D2e—™ 2   —2ƒ  —2— †D2  2 — 2— 2 2 2  F2@€ —2™  2…ƒheD2e‚ƒ2˜ qF2‚F2f™"— FA Biological Control of Invasive Plants in the Eastern United States 10 Most feeding and oviposition by A. hygrophila is on aquatic alligatorweed. Flea beetles, especially larvae, rarely attack plants rooted on shore or in moist ditches. What appears to be typical feeding damage is occasionally observed on terrestrial plants, but it is usually nocturnal feeding by native Disonycha flea beetles. In laboratory experiments in Argentina, fe- males oviposited equally on aquatic and terrestrial alligatorweed (Maddox, 1968), but in my experience, females stopped ovipositing almost immediately when fed terrestrial plants. Beetles are specific to alligatorweed and have not been reported on other host plants in the United States even after almost 40 years. A flavone feeding stimulant, 7-a-L-rhamnosyl- 6-methoxyluteolin (I), has been isolated from alligatorweed (Zielske et al., 1972) and may be the basis for this specificity. Alligatorweed thrips, Amynothrips andersoni O’Neill (Thysanaptera: Phlaeothripidae) Adults are small (ca. 2 mm long), black elongate in- sects with fringed wings (O’Neill, 1968). Larvae, in contrast, are bright orange (Fig. 5). Both feed in the tips of stems where they cause leaf deformation and stunting of the plant (Fig. 6). Often, the edges of leaves curl inwards which provides excellent shelter. Fe- males had a four-day preoviposition period after which they deposited a mean of 201 eggs on hairs in the nodes of the apical leaves (Maddox and Mayfield, 1979). The elongated oval eggs were amber colored. Larval development took eight to 13 days at 24 ºC and the whole life cycle from egg to egg required 28 days on average. There were two larval stages, fol- lowed by a resting pupal stage on the plant. Unmated females produced only males, but fertilized females produced equal numbers of males and females. Maddox and Mayfield (1972) reported a method for rearing and studying the thrips in the laboratory. In Argentina, larvae were most abundant in the spring and declined through late summer (Maddox and Mayfield, 1979). Adults were the predominant overwintering stage, although small numbers of lar- vae and eggs also were present. There were four or five generations, with no reproductive diapause. Pre- dation by spiders and hemipterous insects appeared to be an important regulating factor in Argentina, especially for pupae (Maddox et al., 1971). Dispersal is limited by wing length. Short-winged adults were present in Argentina at most times, but long-winged, dispersing adults were present in the spring (Maddox and Mayfield, 1979). Long-winged forms were be- lieved to be absent in the United States (Coulson, 1977) but were later reported (Buckingham, 1989; Vogt et al., 1992). Unlike the flea beetle, the thrips attacks both aquatic and terrestrial plants, although Maddox et al. (1971) reported that it preferred ter- restrial plants in Argentina. Alligatorweed stem borer, Arcola malloi (Pastrana) (Lepidoptera: Pyralidae, Phycitinae) This inconspicuous, light tan moth has a 20 mm wing- span and rests with its folded wings curled partly around its body (Fig. 7). Wing tips lie against the plant, but the head is held aloft with the body at an angle to the plant. Pastrana (1961) provides a more p2SF2v— — 22 2—— 2 D e   2—  2y)x D2— 2˜   — F2@€ —2™  2…ƒheD2e‚ƒFA p2TF2v —2  2— 222 —22™— —™ ™22 2˜2—  — 2— — 22 2—— 2 D e   2—  2y)x F2@€ — ™  2…ƒheD2e‚ƒ2˜2qF2‚F2f™"— FA Alligatorweed 11 chewed exit holes to the outside epidermis, which was left intact as escape hatches for the emerging moths. Amber colored pupae darkened as they developed inside silken cocoons. The life cycle from egg to egg required about 39 days at 23 ºC. There were three to four generations per year near Buenos Aires (Maddox, 1970). The moth was multivoltine in the lower Mississippi River Valley, but the number of generations was not determined (Vogt et al., 1992). Brown and Spencer (1973) reported parasitism by Trichogramma sp. (Hymenoptera: Trichogrammatidae) on eggs and by Gambrus spp. (Hymenoptera: Ichneumonidae) on larvae in newly established populations in Florida. Both aquatic and terrestrial alligatorweed plants are attacked by A. malloi. Stems collapse, turn yel- low and die, and heavily damaged mats eventually rot and sink (Brown and Spencer, 1973). Leaves re- main on damaged stems, distinguishing stem borer damage from that caused by flea beetles. Vogt et al. (1992) discussed the migratory behavior of A. malloi in the Mississippi River Valley, where it flew in spring and summer from winter refuges near the Louisiana coast north to Arkansas and northern Mississippi, up to 900 to 1000 km. EVALUATION OF PROJECT OUTCOMES Establishment and Spread of Agents Alligatorweed flea beetles did not establish in Arkan- sas, California, North Carolina, or Tennessee. It was hoped that the population from Necochea, Argen- tina, might be more cold tolerant than the first beetles released, but there have been no reports of an increase in the flea beetle’s range after those 1979 releases. Langeland (1986) reported that releases of the Necochea population were unsuccessful at two study sites in North Carolina. Flea beetles survive mostly in coastal areas or where the mean January tempera- ture is 11.1 ºC or higher (Coulson, 1977; Vogt et al., 1992). Coulson (1977) reported the establishment of alligatorweed thrips in Florida, Georgia, and South Carolina. In 1981, thrips were still present in South Carolina (Buckingham, unpublished data) and in 1982 they were reported in Alabama, Florida, Louisiana, Mississippi, and Texas (Cofrancesco, 1988). p2UF2e 22 2—— 2 2˜ D e ™ —2— 2@€— ——AD2— 2™™ 2—2 2—2 2 2 2  ™ 22 2˜ F2@€ —2™   …ƒheD2e‚ƒ2˜2‡ 2gF2h FA p2VF2ƒ 2 — — 2 —2˜2— — 2  2—— 2 2˜ D2e ™ —2—  @€— ——AD2 D2 2  D2— 2  F @€ —2™  2…ƒheD2e‚ƒFA complete description. Females deposited single white eggs on the undersides of apical leaves (Maddox, 1970). After a preoviposition period of less than 24 hours, moths laid an average of 267 eggs over six to eight days. The eggs hatched in three to four days. Newly hatched larvae tunneled into tips of stems and bored downwards. As they matured, they exited the stems and dropped down on silken threads to bore into other stems. Damaged tips quickly wilted and heavily damaged stems turned yellow and died (Fig. 8). Whitish larvae have wavy, tan, longitudinal stripes. There are five instars that developed in about 24 days (Maddox, 1970). Mature larvae bored through nodes and sealed the holes with masticated tissue apparently to protect against water intrusion. Larvae then Biological Control of Invasive Plants in the Eastern United States 12 The alligatorweed stem borer successfully es- tablished at release sites in all states except Alabama and North Carolina (Coulson, 1977). Later, Cofrancesco (1988) reported it in Alabama and North Carolina as well as in Louisiana, Mississippi, and Texas. Vogt et al. (1992) discussed this moth’s long distance dispersal ability and noted that in summer it is present in Arkansas. Suppression of Target Weed Alligatorweed flea beetle damage was spectacular in the early phases of the program. Vast areas were de- foliated (Fig. 9). Mats attacked by the stem borer turned yellow and died (Fig. 10). These two agents have suppressed alligatorweed in much of the warmer parts of its range, so that control efforts are needed only sporadically. In the Carolinas, Tennessee, and the northern regions of the Gulf Coast states, the plant is usually not controlled biologically unless flea beetles are released early in the season from field col- lections made in Florida or other southern sites. In the Mississippi River Valley, moths and/or flea beetles often migrate north early enough to provide local control (Vogt et al., 1992). Fortunately, alligatorweed is not as invasive at the margin of its range as it is further south. Cofrancesco (1988) surveyed aquatic plant man- agers in 1981 about the importance of alligatorweed in ten southern states. None reported that it was a major problem, although some reported locally seri- ous problems, and none reported chemical control efforts directed specifically at it. Chemical controls were usually incidental to waterhyacinth control ef- forts. Recovery of Native Plant Communities There was relatively little evaluation of the biologi- cal control of alligatorweed program, mostly because of the speed of the control and the desire to quickly begin programs for biological control of waterhyacinth (Eichhornia crassipes [Martius] Solms- Laub.) and hydrilla (Hydrilla verticillata [L. f.] Royle). However, Vogt et al. (1992) did conduct long term studies in the lower Mississippi River Valley. They reported observations of native plant popula- tions increasing after alligatorweed was suppressed and included tables of the species involved; however, they did not collect quantitative data. Economic Benefits Long-term economic benefits of alligatorweed con- trol have not been estimated. The fact that most con- trol efforts are now incidental to waterhyacinth con- trol instead of directed at alligatorweed (Cofrancesco, 1988) suggests substantial benefit from reduced chemical control costs. Undoubtedly, there also are indirect cost savings from reduced ditch and canal clearing and from reduced local flooding. Andres (1977) discussed the costs and benefits of the alligatorweed program, including a 76% reduction p2WF2e— 2 —2 2— ™   2 — 2˜2 2——  —2˜  D2e—™ 2   —2ƒ  —2— †F2‡ 2h 2™ ™ 2 —2˜   ˜2- —"22 — .2  2 2— —˜ 2— 22 2  — F @€ —2™  2…ƒheD2e‚ƒFA p2IHF2e— 2 —2" 2˜2 —— 2 2˜ D2e ™ —2—  @€— ——AD2 2  2˜22 —  — D22™ —22 2  2  — 2˜2 2—— 2 —2˜  D e—™ 2   —2ƒ  —2— 2†F @€ —2™  2…ƒheD2e‚ƒFA Alligatorweed 13 in the hectares treated by the U.S. Army Corps of Engineers and a 92% reduction in weed control costs at one lake in Georgia. However, I am unaware of any newer studies on the costs or benefits of alligatorweed control. RECOMMENDATIONS FOR FUTURE WORK Alligatorweed has invaded regions in the United States with climates colder than the native regions in South America. Therefore, there might not be natu- ral enemies suitable for use in the northern parts of alligatorweed’s range in the United States. However, more complete control in the warmer areas of the range might be possible by using some of the origi- nally discovered agents that were not pursued or new, as yet, undiscovered agents. Recent surveys in the Amazon River drainage for waterhyacinth insects have located several species that had not been found during earlier surveys (DeQuattro, 2000). A similar situation might be true for alligatorweed. Also, patho- gens, both in South America and in the United States, should be more carefully evaluated, especially for their potential to complement insect damage. Vogt et al. (1992) suggested that Alternanthera species in Asia should be examined as sources of bio- logical control agents for alligatorweed. Herbivorous insects and pathogens on Asian plants in this genus, if specific both to Alternanthera and to aquatic habi- tats, should be safe for use in the United States and might be more damaging than co-adapted agents from the target plant. The terrestrial South American flea beetle D. argentinensis, which has been released in Australia, should be re-evaluated for its safety and potential use in the United States, if there are no conflicts over con- trol of terrestrial alligatorweed, as there have been in the past (Coulson, 1977). The flea beetle might re- duce the invasion of ditches, canals, and shallow ponds when water returns after a drought. Additional attempts could be made to establish the alligatorweed flea beetle and the stem borer in California, where they did not establish. In the east- ern United States, these two species have probably established in all areas where the long-term climate allows. Annual importation from overwintering sites in Florida or coastal areas will be necessary in north- ern areas of the range, as has been done with alligatorweed flea beetles by the U.S. Army Corps of Engineers (Zattau, 1989). A supply of these insects should be created for use by agencies and individuals other than the Corps of Engineers. REFERENCES Anderson, W. H. 1963. Status of research on biological control of alligatorweed with insects, unpublished report. U.S. Department of Agriculture, Agriculture Research Service, Beltsville, Maryland, USA; available on request from the Biological Documenta- tion Center, National Agricultural Library, 4 th Floor, 10301 Baltimore Boulevard, Beltsville, Maryland, 20705-0000). Andres, L. A. 1977. The economics of biological control of weeds. Aquatic Botany 3: 111-123. Anonymous. 1981. The use of insects to manage alligatorweed. Instruction Report A-81-1. U.S. Army Engineer Waterways Experiment Station, Vicksburg, Mississippi, USA. Anonymous. 1987. Alligatorweed - a monster of a problem. Landscape Management 25: 94, 108. Barreto, R., R. Charudattan, A. Pomella, and R. Hanada. 2000. Biological control of neotropical aquatic weeds with fungi. Crop Protection 19: 697-703. Brown, J. L. and N. R. Spencer. 1973. Vogtia malloi, a newly introduced phycitine moth (Lepidoptera: Pyralidae) to control alligatorweed. Environmental Entomology 2: 519-523. Buckingham, G. R. 1989. Macropterous adults of alligatorweed thrips, Amynothrips andersoni O’Neill, found in Florida. Florida Entomologist 72: 221-223. Buckingham, G. R. 1994. Biological control of aquatic weeds, pp. 413-479. In Rosen, D., F. D. Bennett, and J. L. Capinera (eds.). Pest Management in the Subtropics: Biological Control – the Florida Experi- ence. Intercept, Andover, Hampshire, United Kingdom. Buckingham, G. R. 1996. Biological control of alligatorweed, Alternanthera philoxeroides, the world’s first aquatic weed success story. Castanea 61: 231-243. Buckingham, G. R., D. Boucias, and R. F. Theriot. 1983. Reintroduction of the alligatorweed flea beetle (Agasicles hygrophila Selman & Vogt) into the United States from Argentina. Journal of Aquatic Plant Management 21: 101-102. Cofrancesco, A. F., Jr. 1988. Alligatorweed survey of ten southern states. Miscellaneous Paper A-88-3. U.S. Army Engineer Waterways Experiment Station, Vicksburg, Mississippi, USA. Biological Control of Invasive Plants in the Eastern United States 14 Cordo, H. A., C. J. DeLoach, and M. Ferrer. 1984. Biology and larval host range of the flea beetle Disonycha argentinensis (Coleoptera: Chrysomelidae) on alligatorweed in Argentina. Annals of the Entomological Society America 77: 134- 141. Corell, D. S. and H. B. Corell. 1972. Aquatic and Wet- land Plants of Southwestern United States. Vol. II. Stanford University Press, Stanford, California, USA. Coulson, J. R. 1977. Biological control of alligatorweed, 1959-1972. A review and evaluation. Technical Bulletin 1547. U.S. Department of Agriculture, Agricultural Research Service, Coulson, J. R., P. V. Vail, M. E. Dix, D. A. Nordlund, and W. C. Kaufmann (eds.). 2000. 110 Years of Biological Control Research and Development in the United States Department of Agriculture: 1883-1993. U.S. Department of Agriculture, Agricultural Research Service. Washington, D. C. Cronquist, A. 1988. The Evolution and Classification of Flowering Plants, 2 nd ed. New York Botanical Garden, Bronx, New York. DeQuattro, J. 2000. Watch out water-hyacinth! new jungle enemies are coming. Agricultural Research 48: 10-12. Engler, A. 1934. Die natürlichen Pflanzenfamilien. Vol. 16C, Angiospermae, Riehe Centospermae, 2nd ed. Verlag von Wilhelm Engelmann, Leipzig, Germany. Godfrey, R. K. and J. W. Wooten. 1981. Aquatic and Wetland Plants of Southeastern United States. Dicotyledons. University of Georgia Press, Athens, Georgia, USA. Hill, H. R. and F. W. Zettler. 1973. A virus-like stunting disease of alligatorweed from Florida. Southern Abstracts 443. Holcomb, G. E. 1978. Alternaria alternanthereae from alligatorweed is also pathogenic on ornamental Amaranthaceae species. Phytopathology 68: 265-266. Holm, L., J. Doll, E. Holm, J. Pancho, and J. Herberger. 1997. World Weeds: Natural Histories and Distribu- tion. John Wiley and Sons, New York. Julien, M. H. and J. E. Broadbent. 1980. The biology of Australian weeds. 3. Alternanthera philoxeroides (Mart.) Griseb. Journal Australian Institute of Agricultural Science 46: 150-155. Julien, M. H. and M. W. Griffiths (eds.). 1998. Biological Control of Weeds: A World Catalogue of Agents and Their Target Weeds, 4 th ed. CAB International, Wallingford, United Kingdom. Kartesz, J. T. 1994. A Synonymized Checklist of the Vascular Flora of the United States, Canada, and Greenland. Vol. 1. Checklist. Timber Press, Portland, Oregon, USA. Langeland, K. A. 1986. Management program for alligatorweed in North Carolina. UNC-WRRI-86- 224. Water Resources Research Institute, University of North Carolina, Raleigh, North Carolina, USA. Mabberley, D. J. 1997. The Plant Book, 2nd ed. Cam- bridge University Press, Cambridge, United King- dom. Maddox, D. M. 1968. Bionomics of an alligatorweed flea beetle Agasicles sp., in Argentina. Annals of the Entomological Society of America 61: 1299-1305. Maddox, D. M. 1970. The bionomics of a stem borer, Vogtia malloi (Lepidoptera: Phycitidae), on alligatorweed in Argentina. Annals of the Entomo- logical Society of America 63: 1267-1273. Maddox, D. M. 1973. Amynothrips andersoni (Thysanoptera: Phlaeothripidae), a thrips for the biological control of alligatorweed. l. host specificity studies. Environmental Entomology 2: 30-37. Maddox, D. M. and R. D. Hennessey. 1970. The biology and host range of Vogtia malloi Pastrana, unpub- lished report. U.S. Department of Agriculture, Agricultural Research Service, Entomological Research, Entomological Research Division, Albany, California, USA; available on request from the Biological Documentation Center, National Agricul- tural Library, 4 th Floor, 10301 Baltimore Boulevard, Beltsville, Maryland, 20705-0000) Maddox, D. M. and A. Mayfield. 1972. A method for rearing and studying Amynothrips andersoni in the laboratory. Journal of Economic Entomology 65: 1521-1523. Maddox, D. M. and A. Mayfield. 1979. Biology and life history of Amynothrips andersoni, a thrip for the biological control of alligatorweed. Annals of the Entomological Society of America 72: 136-140. Maddox, D. M. and M. E. Resnik. 1969. Determination of host specificity of the alligatorweed flea beetle, Agasicles n. sp., with radioisotopes. Journal of Economic Entomology 62: 996-999. Maddox, D. M., L. A. Andres, R. D. Hennessey, R. D. Blackburn, and N. R. Spencer. 1971. Insects to control alligatorweed: an invader of aquatic ecosys- tems in the United States. Bioscience 21: 985-991. O’Neill, K. 1968. Amynothrips andersoni, a new genus and species injurious to alligatorweed. Proceedings of the Entomological Society of Washington 70: 175-183. Pastrana, J. A. 1961. Una nueva Phycitidae (Lep.) parasito de la “lagunilla. Revista de Investigaciones Agricolas 15: 265-272. Pemberton, R. W. 2000. Predictable risk to native plants in weed biological control. Oecologia 125: 489-494. Penfound, W. T. 1940. The biology of Achyranthes philoxeroides (Mart.) Standley. American Midland Naturalist 24: 248-252. Alligatorweed 15 Quimby, P. C., Jr. and S. H. Kay. 1976. Alligatorweed and water quality in two oxbow lakes of the Yazoo River basin. Journal of the Mississippi Academy of Science 21 (supplement): 13. Reed, C. F. 1970. Selected Weeds of the United States. Agriculture Handbook 366. U.S. Department of Agriculture, Agricultural Research Service. Sands, D. P. A., R. C. Kassulke, and K. L. S. Harley. 1982. Host specificity of Disonycha argentinensis [Col: Chrysomelidae], an agent for the biological control of Alternanthera philoxeroides (alligator weed) in Australia. Entomophaga 27: 163-172. Selman, B. J. and G. B. Vogt. 1971. Lectotype designa- tions in the South American genus Agasicles (Co- leoptera: Chrysomelidae), with description of a new species important as a suppressant of alligatorweed. Annals of the Entomological Society of America 64: 1016-1020. USDA, NRCS. 1999. United States Department of Agriculture, Natural Resources Conservation Service. Plants Database http://plants.usda.gov (accessed January 2001). Vogt, G. B. 1973. Exploration for natural enemies of alligator weed and related plants in South America, appendix B, pp. 1-66. In Gangstad, E. O., R. A. Scott, Jr., and R. G. Cason (eds.). Biological Control of Alligatorweed. Technical Report 3. U.S. Army Engineer Waterways Experiment Station, Aquatic Plant Control Program. Vicksburg, Mississippi, USA. Vogt, G. B., J. U. McGuire, Jr., and A. D. Cushman. 1979. Probable evolution and morphological varia- tion in South American disonychine flea beetles (Coleoptera: Chrysomelidae) and their amaranthaceous hosts. Technical Bulletin 1593. U.S. Department of Agriculture, Agricultural Research Service, Washington, D.C. Vogt, G. B., P. C. Quimby, Jr., and S. H. Kay. 1992. Effects of weather on the biological control of alligatorweed in the lower Mississippi Valley region, 1973-83. Technical Bulletin 1766. U.S. Department of Agriculture, Agricultural Research Service, Washington, D.C. Wain, R. P., W. T. Haller, and D. F. Martin. 1984. Genetic relationship among two forms of alligatorweed. Journal Aquatic Plant Management 22: 104-105. Zattau, W. C. 1989. Aquatic plant control operations support center, pp. 304-306. Proceedings 23rd Annual Meeting Aquatic Plant Control Research Program, November 14-17, 1988, West Palm Beach, Florida. Miscellaneous Paper A-89-1. U.S. Army Engineer Waterways Experiment Station, Vicksburg, Mississippi, USA. Zielske, A. G., J. N. Simons, and R. M. Silverstein. 1972. A flavone feeding stimulant in alligatorweed. Phytochemistry 11: 393-396. Zeiger, C. F. 1967. Biological control of alligatorweed with Agasicles n. sp. in Florida. Hyacinth Control J. 6: 31-34. PEST STATUS OF WEED Salvinia molesta D. S. Mitchell is a floating fern na- tive to South America that in the last half of the twen- tieth century spread widely throughout the tropics and subtropics, moved in part by the trade in orna- mental plants for fish tanks and ponds. It forms dense mats over lakes and slow moving rivers and causes large economic losses and a wide range of ecological problems to native species and communities. It is of interest in the United States because of its recent es- tablishment in east Texas. Nature of Damage Economic damage. Mats of S. molesta (referred to hereafter as salvinia) impede access to and use of wa- terways for commercial and recreational purposes and degrade waterside aesthetics (Fig. 1). Mats re- duce habitats for game birds, limit access to fishing areas, and probably alter fisheries, all with negative economic consequences. Salvinia can clog water in- takes and interfere with agricultural irrigation, water supply, and electrical generation. It provides habi- tats for vectors of human disease with serious socio- economic impacts. In developing countries, the impact of salvinia can be devastating because weed mats block the use of waterways for transportation, cutting off access to important services, farm lands, and hunting grounds. The harm from salvinia mats to fisheries also can be very significant to communities dependent on fish for local consumption (sometimes as the main source of protein) or in areas where fish sales are the main source of cash income (Bennett, 1966; Thomas and Room, 1986). Salvinia also is a weed of paddy rice that reduces production by competing for wa- ter, nutrients and space (Anon., 1987). Ecological damage. The ability to grow very quickly (Cary and Weerts, 1983; Mitchell and Tur, 1975; Mitchell, 1978/9; Room, 1986) and blanket wa- ter bodies makes salvinia an aggressive and competi- tive weed (Fig. 2). Initially, salvinia forms a single layer over water, but with continued growth the mats become multi-layered and can reach up to 1 m in thickness (Thomas and Room, 1986). Thick mats sup- port other colonizing plants, and the high biomass 2 F LOATING FERN (SALVINIA) M. H. Julien, 1 T. D. Center,2 and P. W. Tipping2 1 CSIRO Entomology, Indooroopilly, Australia 2 U.S. Department of Agriculture, Agriculture Research Service, Fort Lauderdale, Florida, USA p2IF2ƒ——2  —2hF2ƒF2w™ 2™  —2—2 22„ —2 ™2™  ™— —2 ™ — — 2 2—2 —  —  ™F2@€  — 2˜2„F2g  FA p2PF2ƒ——2  —2hF2ƒF2w™ 2™  —2 — ˜ 2—2 2 2  2   2 —2 ™ 22u—"—2x— — €—"D2e— —F2@€  — 2˜2wF2t  FA 17 Biological Control of Invasive Plants in the Eastern United States 18 and stability of such mats make them difficult to dis- lodge and destroy (Storrs and Julien, 1996). Plants and animals dependent on open water to gain sunlight, oxygen, and space for sustenance and growth, or for landing, fishing, nest building, or mat- ing, are displaced by dense salvinia infestations. Wa- ter under mats of salvinia has a lower oxygen con- centration (due to reduced surface area of water avail- able for oxygenation, inhibition of photosynthesis by submerged plants, and consumption of dissolved oxygen by decaying salvinia), higher carbon dioxide and hydrogen sulphide concentrations, lower pH, and higher temperatures than nearby open water (Mitchell, 1978; Thomas and Room, 1986). Through high growth rates and slow decom- position rates, salvinia reduces the concentration of nutrients that would otherwise be available to pri- mary producers and organisms that depend on them (Sharma and Goel, 1986; Storrs and Julien, 1996). Mats of salvinia provide ideal habitat for Man- sonia mosquitoes, a principal vector of rural elephan- tiasis in Sri Lanka (Pancho and Soerjani, 1978), and for other mosquito species involved in the transmis- sion of encephalitis, dengue fever, and malaria (Creagh, 1991/92). Two species of Mansonia that oc- cur in the United States, Mansonia dyari Belkin and Mansonia titillans (Walker), have been implicated in the transmission of St. Louis encephalitis and Ven- ezuelan equine encephalitis, respectively (Lounibos et al., 1990). Extent of losses. The most detailed assessment of costs caused by salvinia was conducted in Sri Lanka using 1987 as the base year (Doeleman, 1989). Paddy rice losses, fishing losses, other losses (power gen- eration, transport, washing and bathing, etc.), health costs, abatement costs, and economic benefits were considered. No environmental costs were included, but they were recognized as important. There were no identified benefits from salvinia. Total costs asso- ciated with salvinia were estimated to be between 24.7 million and 56.7 million rupees (in Australian dol- lars, between 0.9 and 2.1 million) for 1987. This in- formation was used to determine the benefits from biological control over the following 25 years. The benefits were 53 rupees or dollars per rupee or dollar invested, or 1,673 man-hours per man-hour invested. Using this information as a guide, Room and Julien (1995) estimated that the annual benefits gained from successful biological control of salvinia world- wide were approximately $150 million U.S. Geographical Distribution The native range of salvinia is an area in southeastern Brazil (Forno and Harley, 1979). Its first recorded exotic establishment was in Sri Lanka in 1939 (Will- iams, 1956). It has since become established in India (Cook and Gut, 1971), Australia (Room and Julien, 1995), Papua New Guinea (Mitchell, 1979), Cuba, Trinidad, Guyana, Columbia (Holm et al., 1979), South Africa (Cilliers, 1991), Botswana (Edwards and Thomas, 1977), Kenya, Zambia (Mitchell and Tur, 1975), Namibia (Forno and Smith, 1999), Madagas- car (Room and Julien, 1995), Ghana and Cote D’Ivoire (M. Julien, pers. obs.), Indonesia (Java, Borneo, Sulawesi), Malaysia (mainland Sabah, Sarawak) (R. Chan, pers. comm.), the Philippines (Pablico et al., 1989), Fiji (Kamath, 1979), and New Zealand (Randall, 1996). Salvinia was first reported outside of cultivation in the United States in 1995 at a pond in southeast- ern South Carolina (Johnson, 1995). It was eradicated before spread occurred. It was next found in Hous- ton, Texas, in May 1998, and then at other sites in Texas and in Louisiana during 1998. During 1999 it was found in ponds and rivers in Alabama, Arizona, California, Florida, Georgia, Hawaii, Mississippi, and Oklahoma (Jacono et al., 2000; see also Jacono’s web site). Salvinia is readily available for purchase in the United States, particularly through the Internet. BACKGROUND INFORMATION ON PEST PLANT Taxonomy The aquatic fern family Salviniaceae is placed within the order Hydropteridales and consists of a single genus, Salvinia. Ten species of Salvinia occur world- wide (Herzog, 1935; de la Sota, 1962, 1963, 1964, 1982; Mitchell, 1972). None are native to the United States (Jacono et al., 2000) although seven species originate in the Americas (de la Sota, 1976). Salvinia molesta was given recognition as a species in 1972 (Mitchell, 1972) and is grouped within the Salvinia auriculata complex, together with Salvinia auriculata Aublet, Salvinia biloba Raddi, and Salvinia herzogii de la Sota (Mitchell and Thomas, 1972). Species within this com- plex are characterized by the presence of divided but apically joined “basket” hairs on the abaxial surface, which produce an “egg-beater-like” appearance Floating Fern (Salvinia) 19 (Fig. 3a) (de la Sota, 1962; Mitchell and Thomas, 1972; Forno, 1983). Salvinia molesta can be distinguished from species within the S. auriculata complex by the arrangement of sporangia, the shape of sporocarps (Mitchell and Thomas, 1972; Mitchell, 1972), and by the pattern of leaf venation (Forno, 1983). The accepted common name is salvinia, but it also is called Kariba weed, water fern, or African pyle (in Africa); giant azolla or Australian azolla (in the Philippines); and giant salvinia, water spangles, or floating fern (in the United States). Salvinia minima Baker, the only other Salvinia species present in the United States also is exotic and can be distinguished by the presence of divided hairs on the abaxial leaf surface that are spreading and free at the tips (Fig. 3b). Biology Plant form. Salvinia is a free-floating aquatic fern with a horizontal rhizome just beneath the water surface (Bonnet, 1955; Room, 1983). Each plant is a colony of ramets. Each ramet comprises an internode, a node, a pair of floating leaves, the submerged ‘root,’ and associated buds. The ‘root’ is a modified leaf that looks and functions like a root (Croxdale 1978, 1979, 1981). Salvinia is morphologically variable, primarily in response to the level of crowding and availability of nutrients. These two factors are largely indepen- dent of one another. There are three growth forms, with a continuum among them, that are associated with the degree of crowding experienced by the plant (Mitchell and Tur, 1975). The primary form (Fig. 4a) occurs as isolated plants in the initial ‘invading’ stage of an infestation. This form has small, oval leaves less than 15 mm wide that lie flat on the water surface. The secondary form (Fig. 4b) occurs when plants have been growing over open water for some time, either freely or on the edge of stable mats. Intern- odes are longer, with larger, boat-shaped (slightly keeled) leaves that have rounded apices and are vari- able in size, but are normally between 20 mm and 50 mm wide. The entire lower leaf surface is in contact with the water. The tertiary form (Fig. 4c) occurs when plants are growing in crowded mat conditions associated with mature infestations. Internodes are short with large heart-shaped, or oblong and deeply keeled leaves up to 60 mm in width when fully opened. The undersides of adjacent leaves are in contact with each other. Growth and reproduction. Salvinia is pentap- loid, has a chromosome number of 45, and is inca- pable of sexual reproduction (Loyal and Grewal, 1966). Each node bears a series of up to three axillary buds that develop successively under normal grow- ing conditions (Room, 1988), and up to six in response to damage (Julien and Bourne, 1986). The number of axillary buds that grow, the rate of growth, and plant size are largely dependent on available nutrients. Growth is apically dominant and progresses by ex- pansion of apical and axillary buds, the latter form- ing branches. New plants form when older plants break apart due to senescence or damage (Room, 1983). p2Q—F p2Q˜F p2QF2ƒ——2 ™ 22 2… 2ƒ—  ™—2˜ 2 2˜2 2 —2 —F2@—A s2ƒ——2  —2 2 —2 2—2) E ˜ — +2 — 2—2 2D2  22ƒ—— —2@˜A2 2 —2— 2 —— 2—2 2F @€  — 2˜2wF2t  2‘—“2—2„F2g   ‘˜“FA Biological Control of Invasive Plants in the Eastern United States 20 Factors affecting growth. Salvinia is a perennial plant with no seasonal periodicity, although changes in growth may be related to seasonal variations such as changes in temperature. Salvinia is well adapted to growth in low nutrient waters and can take up nutri- ents quickly when they become available (Room and Thomas, 1986). The proportion of axillary buds that develop is correlated with the nitrogen content of the plant (Room, 1983; Julien and Bourne, 1986), and the ni- trogen content increases following removal of buds by insects or other agents (Room and Thomas, 1985; Julien and Bourne, 1986; Forno and Semple, 1987). At low levels of nitrogen leaves are larger, ‘roots’ longer, sporocarps occur more frequently, and rhi- zome branching is reduced (Room, 1983; Julien and Bourne, 1986; Room, 1988; Room and Julien, 1995). The nitrogen content of salvinia ranges from 0.6 to 4.0% dry weight (Room and Thomas, 1986). The maximum rate of nitrogen uptake, calculated from rates of growth, is near 8 mg nitrogen/g dry weight of salvinia/day or about 6,000 kg nitrogen/ha / year (Room, 1986). Actual measurements at a sewage treat- ment lagoon indicated an uptake of 1,580 kg nitro- gen/ha/year (Finlayson et al., 1982). The optimum temperature for growth is 30°C. Room (1986) described the effect of temperature, above and below 30°C, on relative growth rates and predicted no growth below 10°C and above 40°C. Temperature does not affect the proportion of axil- lary buds that expand to initiate new branches (Room, 1988). Exposure to temperatures below –3°C or above 43°C for more than two hours kills salvinia (Whiteman and Room, 1991). Leaf temperatures that exceeded 40°C and sometime approached 50°C for the hottest parts of days did not obviously affect growth, but water temperatures remained below 40°C and probably acted as a heat sink for the plants (Storrs and Julien, 1996). Plants may be killed by frost but protected parts and unfrozen buds survive. Water bodies are normally cooler than the air in summer and warmer in winter due to thermal inertia. This helps protect salvinia from temperature extremes. Mats of salvinia can grow in water bodies with conductivities ranging from 100 μS/cm to 1,400 μS/ cm (Mitchell et al., 1980; Room and Gill, 1985). In water with 10% of the salinity of seawater (4,800 μS/ cm), growth was reduced by 25% (Divakaran et al, 1980); at 20% salinity, growth was very slow; while at 30%, plants died after 30 minutes exposure (Room and Julien, 1995). Optimum pH for growth is 6.0 (Cary and Weerts, 1984). In the field the plant grows at pH values from 5.2 to 9.5 (Holm et al., 1977; Mitchell et al., 1980). p2R—F p2R˜F p2R™F p2RF2„ 2  2  2 2 2ƒ——   —2hF2ƒF2w™ X2@—A2 2—2 D @˜A2 2 ™ —2 D2—2@™A2 2 —  F2@€  — 2˜2wF2t  FA Floating Fern (Salvinia) 21 Salvinia compensates for the destruction of buds by initiating growth of dormant buds. Complete compensation occurs only when high levels of nitro- gen are available (Julien and Bourne, 1986; Julien et al., 1987). Destruction of leaves (Julien and Bourne, 1988) and rhizomes (Julien and Bourne, 1986) does not induce compensatory growth. Growth rates and density. On Lake Kariba, Zimbabwe, numbers of leaves (ramets) doubled in eight to 17 days (Gaudet, 1973; Mitchell and Tur, 1975). In the Kakadu National Park, Australia, dry weight doubled in five to 30 days (Storrs and Julien, 1996). Under ideal growth conditions, biomass and numbers of ramets typically double in two to three days (Mitchell and Tur, 1975; Cary and Weerts, 1983). Densities from as high as 2,500 large tertiary form ramets per m 2 (in nutrient-poor water) to 30,000 small tertiary form ramets (in nutrient-rich waters) have been noted. At these densities, natality is equaled by mortality (Room and Julien, 1995). Salvinia is 95% water by weight and biomass of living shoots can exceed 600 g/m2 of dry weight, while biomass of liv- ing and dead shoots and ‘roots’ may exceed 1,600 g/ m2 of dry weight or 400 t/ha of fresh weight (Room and Julien, 1995). Fresh weight biomass in Texas var- ied through the year, reaching a high in October 1999 of 248 t/ha and a low of 84 t/ha in January 2000 (P. Tipping, unpub. data). Spread. Salvinia is spread within and between aquatic systems mainly by people. It is spread acci- dentally when equipment or boats are moved and de- liberately when it is used as a pond, aquarium, or water-garden plant or as a biological weapon (Gewertz, 1983). It is carried on animals as they move from infested water bodies (Forno and Smith, 1999). Dispersal within a water body or catchment is by wind and water currents (Room and Julien, 1995). Currents and floods wash mats away and growth is best in still or slow moving water. In its native range in southeastern Brazil, salvinia is a component of floating and emergent plant com- munities. Salvinia supports a variety of natural en- emies (Forno and Bourne, 1984), and it normally does not form the extensive mats prevalent in its exotic range. Analysis of Related Native Plants in the Eastern United States The Salviniaceae are included within a mono- phyletic clade of heterosporous genera that also en- compasses the Azollaceae and Marsileaceae (Pryer and Smith, 1998; Pryer et al., 1995; Hasebe et al., 1995), all aquatic leptosporangiate ferns. This clade includes ten species in the North American flora. The Azollaceae once were included within the Salviniaceae, but the relationship is not close and they have since been separated (Lumpkin, 1993). Azollaceae consists of the single genus Azolla. There are three species of Azolla – Azolla caroliniana Willdenow, Azolla mexicana C. Presl, and Azolla filiculoides Lamarck – that are native to North America (Lumpkin, 1993). The Marsileaceae includes two genera, Marsilea and Pilularia, both of which occur in North America. Seven species within these two genera are part of the North American flora (Johnson, 1993): Marsilea quadrifolia Linnaeus, Marsilea ancylopoda A. Braun, Marsilea oligospora Gooding, Marsilea mollis B. L. Robinson and Fernald, Marsilea macropoda Engelmann, Marsilea vestita Hooker and Greville, and Pilularia americana A. Braun. Marsilea quadrifolia, a native of Europe and Asia, is introduced and M. ancylopoda is extinct, so only eight native species remain. HISTORY OF BIOLOGICAL CONTROL EFFORTS Area of Origin of Weed The native range of Salvinia molesta includes a rela- tively small area (20,000 km 2 ) in southeastern Brazil, including the states of Sao Paulo, Paraná, Santa Catarina and Rio Grande do Sul. It occurs between the latitudes 24 0 05’ S and 32 0 05’ S; at altitudes 0 to 500 meters; and up to 200 km inland (Map 1). Salvinia occurs in natural lagoons, artificial dams, swamps, drainage canals, and along margins of rivers (Forno and Harley, 1979). Biological Control of Invasive Plants in the Eastern United States 22 Areas Surveyed for Natural Enemies The first surveys for potential biological control agents for S. molesta were conducted in Trinidad, Guyana and northeastern Brazil from 1961 to 1963 (Bennett, 1966), and in Argentina prior to 1975 (Bennett, 1975), where species in the S. auriculata complex other than S. molesta occur. At this time the true identity and the native range of S. molesta were not known. Surveys for natural enemies were conducted in Trinidad, Venezuela, Guyana, Uruguay, Paraguay, Brazil, and Argentina during 1978 to 1981. In 1978, the previously unknown range of S. molesta was iden- tified (Forno and Harley, 1979), permitting surveys to focus on the relatively small native range of the target weed rather than the larger range of the S. auriculata complex (Forno and Bourne, 1984). Natural Enemies Found The natural enemies of S. molesta and the related spe- cies in the S. auriculata complex are listed in Forno and Bourne (1984), including species collected by Bennett (1975). Twenty-five phytophagous or pos- sibly phytophagous species have been recorded from S. molesta, compared to 49 species from the four spe- cies of the S. auriculata complex. Four of these spe- cies have been used as biological control agents against S. molesta. The first three, Cyrtobagous singularis Hustache, Paulinia acuminata (De Geer), and Samea multiplicalis (Guenée) (identified during the early exploration [Bennett, 1966]) have not been successful control agents. The fourth, Cyrtobagous salviniae Calder and Sands, was found during later work (Sands, 1983) and has been extremely success- ful. Origin of Salvinia molesta and Cyrtobagous salviniae Salvinia molesta under biological control Salvinia molesta not under control ? ? Status of releases unknown ? w—2IF2g  2  2ƒ——2  —2 —2˜ 2 22 2—2  2 ˜ 2—2™  2  2˜ ™— ™  2˜2 2  2g  ˜—  2 —— 2 —2˜ 2™™  F Floating Fern (Salvinia) 23 Host Range Tests and Results Host range studies on the three unsuccessful species are summarized in Bennett (1966), Sankaran and Ramaseshiah (1973), Sands and Kassulke (1984, 1986), and Knopf and Habeck (1976). Host range tests to assess feeding by C. salviniae were carried out in Australia on 46 species from six families of Pteridophyta (ferns), eleven families of Monocotyledons, and sixteen families of Dicotyle- dons (Forno et al., 1983). (This weevil was later found to be a new species and subsequently descrbed as C. salviniae Calder and Sands). Test plants were exposed to mature weevils in three replicates in choice tests. Adult feeding occurred on Pistia stratiotes L., but the insect was unable to reproduce on that species. Mi- nor leaf feeding was observed in choice tests on Ip- omea batatas (L.) Lam. (sweet potato) when the leaves were held in contact with water, an abnormal condi- tion. Adults failed to feed on I. batatas in no-choice tests in a non-aquatic situation and died within seven days. Host specificity tests indicated that this weevil was restricted to S. molesta. It has never been observed attacking plants other than Salvinia species in the field in South America, including those that grew in asso- ciation with S. molesta such as water fern (Azolla sp.), waterhyacinth (Eichhornia crassipes [Mart.] Solms- Laubach), and waterlettuce (P. stratiotes) (Forno et al., 1983). Importantly, this weevil has not been found to attack any other plants even when huge popula- tions were starving following population crashes of salvinia. Releases Made The grasshopper P. acuminata, collected from S. auriculata in Trinidad, was released in Zimbabwe (in 1969 and again in 1971), Kenya and Zambia (1970), Botswana (1971 and 1975), Sri Lanka (1973 and 1978), India (1994), and Fiji (1975). It failed to establish in Botswana, Kenya, and Sri Lanka and does not pro- vide control in the countries where it established. The weevil C. singularis, collected from S. auriculata in Trinidad, was released in Botswana (in 1971 and 1976), Zambia (1971), and Fiji (1976). It is established in each country but is not providing control. The moth S. multiplicalis, also collected from S. auriculata in Trinidad, was released in Zambia (1970), Botswana (1972), and Fiji (1976). It failed to establish in Fiji and Zambia and does not control the weed in Botswana (Julien and Griffiths, 1998). It was later collected from S. molesta in Brazil and released in Australia during 1981, where it established widely but failed to provide control (Room et al., 1984; Forno, 1987). Releases of C. salviniae from S. molesta in south- eastern Brazil were made first in Australia in 1980. This weevil now controls the weed in most tropical and subtropical areas (Fig. 5) and in some temperate climates (Fig. 6). It has been released in 15 countries and controls the weed in at least 12 of these (Table 1). In Cote D’Ivoire, insufficient time has elapsed to measure success. Information on project results is not available for the Philippines or Indonesia. Map 1 shows the countries that have or have had serious salvinia problems and those where biological control has been successful. p2SF2@—A2v—" 2w ——D2e— —2™    2ƒ——2  —2hF2ƒF2w™ 2˜    2 2 2˜ ™— 2™  Y2@˜A2w  2 — WH72 2— —2 2v—" 2w ——2 —    2  2IR2  2˜2 2— —  D2g ˜—  2 —— 2g—  2— ƒ—F2@€  — 2˜2€F2‚ FA p2S—F p2S˜F Biological Control of Invasive Plants in the Eastern United States 24 HISTORY OF BIOLOGICAL CONTROL EFFORTS IN THE EASTERN UNITED STATES A weevil found attacking S. minima in Florida was identified as C. singularis (Kissinger, 1966), but Calder and Sands (1985) listed Cyrtobagous specimens from Florida as C. salviniae and did not consider the range of C. singularis to include North America. When S. molesta was found in the United States and biologi- cal control was considered, further morphological ex- amination of weevils from Florida suggested that they were C. salviniae (C. O’Brien, pers. comm.). This weevil was collected from S. minima in Florida by scientists from the USDA, ARS Invasive Plant Re- search Laboratory and released at sites in Liberty, Bridge City, and Toledo Bend Reservoir in Texas, and at Salter Creek in Louisiana during 1999. When another USDA laboratory planned to re- lease an Australian population of C. salviniae, a mo- lecular comparison of the D2 gene was made between Florida and Australian material to provide a means of distinguishing between weevils from the two sources. Several base pair differences were found but the taxonomic significance of this is not yet clear. Further studies utilizing molecular, morphological and bioassay methods are planned. Until these dif- ferences are explained, further release of the Florida „—˜2IF2„ 2ƒ—2 2‚ — 2 2g ˜—  ƒ—— 2g—  2—2ƒ—2 2i—™ 2g  —2 2h— 2 2s— 2‚ — 2@w    2t  2—2q D2IWWVA Country Initial Release Date Status Australia 1980 Control in tropical and subtropical areas; some control in temperate areas Botswana Spread from Namibia Control in 1 to 5 years Cote D’Ivoire 1998 Established and spreading Fiji 1991 Successful control Ghana 1996 Control India 1983 Control at Bangalore and Kerala Indonesia 1997 Status is unknown Kenya 1990 Control except where affected by herbicide Malaysia 1989 Control where released. Needs redistribution Namibia 1984 Good control Papua New Guinea 1982 Good control Philippines 1989 Established on Panay. Impact unknown Republic of South Africa 1985 Successful control within 2 years Sri Lanka 1986 Successful control Zambia 1990 Excellent control Zimbabwe 1992 Good control within 2 years p2TF2@—A2e2™ "22  — 2e— —D2 — ƒ D2  2  2ƒ——2  —2hF2ƒF w™ D2—2@˜A2  2 —2 — 2— 2 — —2  2g ˜—  2 —— 2 —  — F2@€  — 2˜2wF2t  FA p2T—F p2T˜F Floating Fern (Salvinia) 25 population will be suspended. No release of Austra- lian material has been done yet. BIOLOGY AND ECOLOGY OF KEY NATURAL ENEMIES Salvinia weevils, Cyrtobagous salviniae and C. singularis (Coleoptera: Curculionidae) The genus Cyrtobagous Hustache was originally thought to be monotypic, containing only the spe- cies C. singularis. This weevil was known to feed on various South American Salvinia species and had been used unsuccessfully as a biological control agent for salvinia during the 1970s (Julien and Griffiths, 1998). Discovery of the native range of S. molesta in the late 1970s allowed surveys for natural enemies to con- centrate on the target weed. A weevil thought to be C. singularis was collected from S. molesta and sub- sequently released in Australia, where it controlled the weed. Comparative studies determined that this was a new species, subsequently named C. salviniae (Fig. 7). Differences in the biology between the two species explained why one species failed to cause sig- nificant damage to the weed while the other proved to be an excellent control agent. Larvae of C. salviniae tunnel within the rhizomes causing them to disinte- grate. Larvae also tunnel in the buds and adults eat buds, thus suppressing growth and vegetative propa- gation of this sterile weed. Larvae and adults of C. singularis feed on leaves and other tissues but don’t affect the rhizomes or meristems. This research pro- vided a classic example of the importance of careful taxonomic study of both the weed and the insects for successful biological control (Thomas and Room, 1986). Other important differences between the two species are that C. salviniae has a higher intrinsic rate of increase, lays seven times more eggs, and ovipos- its with greater frequency. Furthermore, this species’ oviposition is less affected by changes in the nutrient status of the host plant, and larval and pupal survival rates are higher (Sands et al., 1986). C. salviniae. The adult male of C. salviniae (1.8 x 0.9 mm) is slightly smaller than the female (2.2 x 1.2 mm). Newly emerged adults are brown, darken- ing to black in about five days. Detailed descriptions are given in Calder and Sands (1985) of the features that distinguish this species from C. singularis. Adults are found on or beneath young leaves, on or inside the developing leaves or among ‘roots’. When under water, adults respire by means of an air bubble (called a plastron) that adheres to their ventral surface (Forno et al., 1983). Multiple matings occur five to 26 days after emergence. At 25.5ºC, oviposition begins after six to 14 days. Eggs (0.5 x 0.24 mm) are laid singly in cavities excavated by adults in lower leaves, develop- ing leaves, rhizomes, and ‘roots.’ At 25.5ºC, females lay one egg every two to five days for at least 60 days (Forno et al., 1983). Eggs hatch in 10 days at this tem- perature. Newly emerged larvae (1 mm) are white. They feed initially on ‘roots’ in or on the small buds, and later inside rhizomes, completing three instars in approximately 23 days (Forno et al., 1983). Devel- opment rate is dependent on temperature and the nutrient status of the host plant, larval development taking 13.4 days at 31ºC on ‘high’ nitrogen plants. Larvae prefer to tunnel in young rhizomes and more tunneling occurs if plants are low in nitrogen. Larvae do not survive below 16.3o C (Sands et al., 1983). Pu- pation occurs in a cocoon (2 x 2.6 mm), which is woven from ‘root hairs’ and attached underwater to the ‘roots,’ rhizomes or leaf bases. At 25.5ºC, pupae require 12.6 days for full development. Pupal dura- tion is not affected by plant quality (Forno et al., 1983; Sands et al., 1983). Oviposition does not occur be- low 21ºC,and eggs fail to hatch below 20ºC or above 36ºC. Adults feed between 13ºC and 33ºC (Forno et al., 1983). The lower lethal temperature at which 50% of the adult population would be expected to die is - 5.2ºC (Reaney 1999). Population densities of C. salviniae are capable of reaching 300 adults and 900 larvae per m2 , levels estimated as necessary for control (Room, 1988, 1990; Room and Thomas, 1985). Weevils in the genus Cyrtobagous were first re- corded from the United States in Florida at the p2UF2„ 2— —2  2g ˜—  —— 2g—  2—2ƒ—F2@€  — 2˜ ‚F2g —FA Biological Control of Invasive Plants in the Eastern United States 26 Archbold Biological Station (Highlands County) in 1962 (Kissinger, 1966). It is assumed that these wee- vils were accidentally introduced from South America, because of the lack of any earlier U.S. records and the adventive status of its host plant, S. minima. Kissinger (1966) considered the Florida wee- vils to be C. singularis, but this was before C. salviniae was recognized as a separate species. Calder and Sands (1985) later classified the Florida specimens as C. salviniae, but noted that the C. salviniae from S. minima in Florida were significantly smaller than those from S. molesta in Brazil. Based solely on mor- phological features, the weevils from Florida (later released in Texas) seem to be C. salviniae (C. O’Brien, pers. comm.). However, recent DNA assessments suggest that that the Florida material differs from C. salviniae from Australia in some respects (Goolsby et al., 2000). Whether these differences imply sepa- rate species status is under investigation. C. singularis. The biology of C. singularis is less well known and it is presumably similar to C. salviniae. For morphological differences between adults of this species and those of C. salviniae see Calder and Sands (1985), and for larvae, May and Sands (1986). Feeding differences between the spe- cies are outlined in Sands and Schotz (1985), and other differences in life history and intrinsic rates of in- crease are discussed in Sands et al. (1986). Adults are small (2 to 3 mm), black, sub-aquatic weevils that re- side on or beneath leaves. While under water they respire using a plastron. Adults preferentially feed on apical leaves but also on the second to fifth pair of leaves, buds, and petioles (Sands and Schotz, 1985). Eggs are laid singly in cavities made by females in leaves. Unlike C. salviniae, whose larvae feed inter- nally, those of C. singularis feed on the outer sur- faces of submerged buds, rhizomes, and petioles. Feeding results in bud destruction, but not rhizome disintegration, and plants retain their capacity for regrowth (Sands and Schotz, 1985). Field population densities of C. singularis do not exceed 50 adults per m2 (Schlettwein, 1985), a level that is insufficient to significantly damage salvinia (Room, 1990). This in- sect has not been a useful biological control agent (Julien and Griffiths, 1998). Waterlettuce moth, Samea multiplicalis (Lepidoptera: Pyralidae) Larvae and adults of S. multiplicalis are very similar to the closely related species, Niphograpta (Sameodes) albiguttalis (Warren) (waterhyacinth moth, Pyralidae). Center et al. (1982) give characters to sepa- rate larvae of these species. Samea multiplicalis was originally described from Brazil where it was ob- served feeding on waterhyacinth. It is widely distrib- uted throughout warmer regions of North and South America. In Florida, it is most commonly found on waterlettuce but also is present on A. caroliniana and S. minima. It is occasionally abundant on small waterhyacinth plants, feeding within inflated leaf petioles (Knopf and Habeck, 1976). This species was introduced into Australia for biological control of both S. molesta and P. stratiotes (Sands and Kassulke, 1984). Adults (Fig. 8) are 6.5 to 10.5 mm long (Sands and Kassulke, 1984), tan, with brown and cream markings on both fore and hind wings. Females lay approximately 300 eggs during their four-to-seven- day life span (Knopf and Habeck, 1976; Sands and Kassulke, 1984; Taylor, 1984). Moths prefer to lay eggs on undamaged salvinia plants with high nitro- gen content (Taylor and Forno, 1987). Most eggs are laid singly among the epidermal plant hairs on the lower surfaces of waterlettuce leaves, on the upper surfaces of Salvinia leaves, or lodged between the leaves of Azolla species. Eggs hatch in about four days at 26ºC. Larvae (Fig. 9) may feed from within a ref- ugium (made of silk and plant hair) attached to the external leaf surface or, for waterlettuce, within gal- leries in the leaves. Larvae periodically extend the refugium to reach fresh leaf material (Knopf and Habeck, 1976). Larger larvae feed on the buds of plants, often killing the growing apex. Larvae also will eat mature waterlettuce fruits and consequently destroy enclosed seeds. On salvinia, temperature and plant quality in- teract to determine rates of insect growth, number of larval instars, fecundity, and survival (Taylor and Sands, 1986; Taylor, 1984, 1988, 1989). Optimum temperature for development for all stages is 28 to 30ºC. Fecundity is greatest at 20 to 22ºC and egg sur- vival highest at 25 to 26ºC (Taylor, 1988). Larvae complete development in 17 to 35 days on salvinia (Sands and Kassulke, 1984; Taylor, 1984). Develop- ment on waterlettuce and other salvinia species is described in Bennett (1966) and Knopf and Habeck (1976). Pupation occurs within a silken cocoon. On waterlettuce, this cocoon is usually formed within the spongy portion of a leaf but on S. molesta it is Floating Fern (Salvinia) 27 constructed among old leaves. Pupal development re- quires four to seven days at 28ºC on waterlettuce and S. minima (Knopf and Habeck, 1976) and eight to nine days at 26ºC on S. molesta (Sands and Kassulke, 1984). Total development (egg to adult) requires 24.6 (Knopf and Habeck, 1976) to 42 days (Sands and Kassulke, 1984). Intrinsic rates of increase are highest in autumn and decline in summer and winter (Room et al., 1984; Taylor, 1988). Levels of parasitism (24%) and dis- ease in Australian populations on salvinia do not ex- plain the seasonal variation in population growth rates (Semple and Forno, 1987); rather, field popula- tion densities are strongly determined by tempera- ture and the nutritional quality of the plant (Taylor, 1988). Salvinia grasshopper, Paulinia acuminata (Orthoptera: Pauliniidae) This semi-aquatic grasshopper (Fig. 10) is adapted to living on floating mats of host plants where humid- ity is constantly high. Paulinia acuminata can com- plete its development on Salvinia spp., P. stratiotes, Azolla sp., and Hydromystria sp. and will feed on a range of other plant species (Bennett, 1966; Sands and Kassulke, 1986). It feeds on all species in the S. auriculata complex from Trinidad to northern Ar- gentina and Uruguay (Bennett, 1966). Eggs of P. acuminata, in an ootheca, are attached underwater to the undersides of leaves and hatch in 19 to 21 days. Out of water, eggs fail to hatch (Bennett, 1966). There are five or six nymphal stages, six being common when temperatures are lower (Thomas, 1980). Eggs take 17 to 20 days to hatch; six nymphal instars com- plete development after 47 days; the pre-ovipositional period takes eight to 10 days; and duration from egg to adult is 67 days on average (Sands and Kassulke, 1986). p2VF2e 2ƒ— —2 ™— 2@q : AF @€  — 2˜2‚F2g —FA p2WF2v— —2 2ƒ— —2 ™— 2@q : AF @€  — 2˜2€F2‚ FA p2IHF2„ 2— —2—  2€—— —™— —2@h 2q AF2@€  — 2˜2€F ‚ FA Females live 50 days or more and lay 200 or more eggs. Adults usually are brachypterous but mac- ropterous forms occasionally occur in the field and are frequent in laboratory cultures (Bennett, 1966). Adults and nymphs feed on all plant parts above the water level when food is scarce. Grasshoppers prefer to feed on new growth and oviposition is reduced when the weed is matted (Mitchell and Rose, 1979). Under extremely hot conditions adults shelter in the water with their heads exposed (Thomas, 1980). In the field at Lake Kariba, where the mean annual tem- perature was 24 to 28ºC, it was estimated that P. acuminata could complete three generations per year (Thomas, 1980). A significant reduction in salvinia on Lake Kariba occurred in 1972/3, following the release of P. acuminata in 1970. However, the decline of the weed at this location has been attributed to other fac- tors such as nutrient stress on the weed (Mitchell and Rose, 1979; Marshall and Junor, 1981). Chisholm Biological Control of Invasive Plants in the Eastern United States 28 (1979) demonstrated that densities of more than 85 P. acuminata per m2 feeding for 24 days were required to reduce production of new leaves. On Lake Kariba during the period when salvinia was a problem the maximum grasshopper density recorded was 27 per square meter (Marshall and Junor, 1981), suggesting that the insect might not have been a primary cause of decline in the weed. Sands and Kassulke (1986) noted that field populations in Africa of 45 to 54 P. acuminata per m 2 grazed up to 87% of leaves with- out affecting apical or lateral growth or killing plants. EVALUATION OF PROJECT OUTCOMES In the United States, S. molesta is a new problem so evaluations are premature. As mentioned previously, C. salviniae, a naturalized species on S. minima in Florida, was released in Texas in 1999. It will take another year to determine if establishment has oc- curred. RECOMMENDATIONS FOR FUTURE WORK Currently, three release sites and three insect-free control sites are being monitored monthly in water bodies in eastern Texas and western Louisiana in- fested with S. molesta using standard protocols. In addition, molecular techniques are being utilized to identify and compare at least six salvinia species, in- cluding those outside of the S. auriculata complex. This will provide a means of identifying Salvinia spe- cies should any new invasions occur in the United States or elsewhere. Similar molecular testing of Cyrtobagous sp. collected from different salvinia spe- cies may provide insights to the taxonomy of this group and their associated host ranges. We anticipate that the same favorable results as obtained in at least 12 countries on three continents can be repeated in the United States, namely, reduc- tions in the density and abundance of S. molesta to acceptable levels. Reductions may range from local extinctions to maintenance of small populations of the weed along the fringes of ponds, lakes, and rivers. ACKNOWLEDGEMENTS We thank Drs. Tim Heard and Peter Room for re- viewing the manuscript and Dr. Peter Room for pre- paring the world map. We also thank Dr. Peter Room and Mr. Richard Chan for permission to use their photographs. REFERENCES Anonymous. 1987. Indonesia girds to battle Salvinia molesta. Weedwatcher 2: 1-2. (Publication of Southeast Asian Weed Information Centre, Bogor, Indonesia.) Bennett, F. D. 1966. Investigations on the insects attack- ing the aquatic ferns, Salvinia spp. in Trinidad and northern South America. Proceedings of the Southern Weed Conference 19: 497-504. Bennett, F. D. 1975. Insects and plant pathogens for the control of Salvinia and Pistia, pp. 28-35. In Brezonik, P. L. and J. L. Fox (eds.). Pr