Establishment of Hydrellia pakistanae (Diptera ... - PubAg - USDA

33 downloads 1246 Views 194KB Size Report
Florida, quarantine facility during May to August 1985. ... The Gainesville quarantine facility received a .... altering rearing, release, and establishment tech-.
8, 65–73 (1997) BC960491

BIOLOGICAL CONTROL ARTICLE NO.

Establishment of Hydrellia pakistanae (Diptera: Ephydridae) for the Biological Control of the Submersed Aquatic Plant Hydrilla verticillata (Hydrocharitaceae) in the Southeastern United States TED D. CENTER,* MICHAEL J. GRODOWITZ,† ALFRED F. COFRANCESCO,† GREG JUBINSKY,‡ EDWARD SNODDY,§ AND JAN E. FREEDMAN† *USDA-ARS, Aquatic Plant Control Research Unit, 3205 College Avenue, Fort Lauderdale, Florida 33314; †USAE Waterways Experiment Station, 3909 Halls Ferry Road, Vicksburg, Mississippi; ‡Florida Department of Environmental Protection, Bureau of Aquatic Plant Control, 3917 Commonwealth Boulevard, Tallahassee, Florida 32399; §Tennessee Valley Authority, OSA 1B, Muscle Shoals, Alabama 35611 Received May 24, 1996; accepted October 8, 1996

Control of hydrilla is currently achieved primarily with herbicides or by the release of exotic herbivorous fish (Sutton and Vandiver, 1986), but both methods are expensive for widespread use. Development of alternative control methods has been assigned high priority by various federal and state agencies. One possible management alternative involves ‘‘inoculative’’ or ‘‘classical’’ biological control, i.e., the introduction of hostspecific plant-feeding insects obtained from within the native range of hydrilla. Searches for potential biological control agents began during the 1970s in India, Pakistan, and Africa and resumed during the 1980s in other areas of the world (Baloch and Sana-Ullah, 1974; Baloch et al., 1980; Sankaran and Rao, 1972; Pemberton, 1980; Center et al., 1990). The resultant surveys revealed the presence of several biological control candidates including an undescribed Hydrellia species later named H. pakistanae Deonier (Deonier, 1978). Hydrellia pakistanae is a small, ephydrid fly whose larvae mine hydrilla leaves. Eggs are laid on floating leaves and the larvae either enter the leaves directly or descend into the water column before entering the plant. Females apparently oviposit indiscriminately, but the larvae are quite selective (Buckingham et al., 1989; Wheeler and Center, 1996). Each larva mines about 12 leaves during the course of its development and then pupates within an air-filled puparium that is attached in a leaf axil (Baloch and Sana-Ullah, 1974; Buckingham et al., 1989). Total generation time is 18 to 30 days (Baloch and Sana-Ullah, 1974). Laboratory studies done in Pakistan (Baloch et al., 1980) and later in a U.S. quarantine facility located in Gainesville, Florida (Buckingham et al., 1989), verified that H. pakistanae was host-specific to hydrilla. Accordingly, Dr. Gary Buckingham petitioned the USDA Animal and Plant Health Inspection Service—Plant Protec-

The Indian leaf-mining fly Hydrellia pakistanae Deonier was released in the United States during October 1987 as a potential agent for the biological control of the submersed aquatic plant, Hydrilla verticillata (L.f.) Royle. Since then, nearly 3 million individuals have been released in four states at over 25 separate locations. Hydrellia pakistanae is now permanently established at many locations in Florida as well as other portions of the southeastern United States. Despite early failures, populations established from at least 70% of the release attempts. Modifications of rearing and release procedures that considered the biology of the agent in relation to various habitat characteristics facilitated successful establishment. r 1997 Academic Press KEY WORDS: aquatic weeds; hydrilla control; biological control of weeds; Hydrellia pakistanae; Hydrilla verticillata; phytophagous insects; bioagent release techniques; Ephydridae; Hydrocharitaceae.

INTRODUCTION

Hydrilla, Hydrilla verticillata (L.f.) Royle (Hydrocharitaceae), is a submersed, leafy-stemmed vascular hydrophyte that is widely distributed in the Old World. It was introduced into the United States through the aquarium trade in the early 1950s (Schmitz et al., 1991). Currently, infestations of this invasive plant constitute the most severe aquatic plant problem in the southern United States. Hydrilla infested 22,000 ha in Florida by 1988 and the state was spending about $7 million annually in attempts to control about 6000 ha (Schmitz et al., 1991). During this same period it rapidly expanded its range and now occurs as far west as California (Sonder, 1979). 65

1049-9644/97 $25.00 Copyright r 1997 by Academic Press All rights of reproduction in any form reserved.

66

CENTER ET AL.

tion and Quarantine (USDA-APHIS-PPQ) in June 1987 for permission to release this species. Permission was obtained after review by the Technical Advisory Group on Biological Control of Weeds (Coulson, 1992) and this insect was first released on 29 October 1987 (Buckingham, 1988). In this paper, we document the results of this and later releases of H. pakistanae and the subsequent establishment of self-perpetuating field populations. This project represents the first attempt to use an insect to control a weed in the submersed aquatic environment by implementation of the ‘‘classical’’ approach to biological control. MATERIALS AND METHODS

Insects. Buckingham et al. (1989) documented the collection of the original H. pakistanae stock from India and its receipt and disposition at the Gainesville, Florida, quarantine facility during May to August 1985. This initial stock was used for all host specificity testing and provided progeny for the initial U.S. releases made during 1987 to 1990. However, because of the risk associated with excessive inbreeding of the original stock, fresh material was obtained from overseas. The Gainesville quarantine facility received a fresh consignment of field-collected H. pakistanae from Kumbaleagodu, Bangalore, Karnataka State, India, on 25 August 1990 and one from Rawalpindi, Pakistan, on 28 August 1990. These insects were reared in quarantine for less than three generations prior to field release to confirm identifications and to eliminate parasites. After permission was granted for release of H. pakistanae, a portion of the original quarantine colony was used for the first release, a portion was retained in quarantine, and a portion was transferred to the U.S. Department of Agriculture, Agricultural Research Service (ARS), Aquatic Plant Control Research Unit in Fort Lauderdale, Florida. Insects reared at ARS were later provided to the Waterways Experiment Station (WES) of the U.S. Army Corps of Engineers in Vicksburg, Mississippi, the Tennessee Valley Authority (TVA) in Muscles Shoals, Alabama, and to the Florida Department of Natural Resources (FDNR, now the Department of Environmental Protection) Bureau of Aquatic Plant Control in Tallahassee, Florida. Rearing. Two basic techniques were used to rear H. pakistanae. The primary method was modified slightly from that described by Buckingham et al. (1989). Hydrilla sprigs containing freshly oviposited eggs were placed in 2- to 4-liter water-filled containers containing approximately 100 g of hydrilla. Adults were removed after emergence and transferred to oviposition chambers. The oviposition chambers consisted of wooden boxes covered with either plastic or glass tops and fitted with cloth sleeves covering entry ports (Fig. 1). Adults were provided sugar water and yeast hydrolysate ei-

ther separately or in mixture (4 g yeast hydrolysate plus 7 g sucrose in 7 ml of water). Partially submersed hydrilla sprigs placed in shallow dishes within the chambers provided the oviposition substrate. Eggladen sprigs were transferred to new water-filled containers after 1 to 3 days to perpetuate the rearing process. The second rearing method consisted of placing freshly harvested hydrilla in large tanks. Either adults or hydrilla sprigs containing immatures were placed on the floating hydrilla. Fly populations were allowed to increase in size within the tanks without additional handling. The hydrilla was then removed and transferred to field sites. Releases. The basic release method involved the placement of insects directly on hydrilla infestations either within open hydrilla beds or in protective cages (Fig. 2). The cages were constructed of two frames made of 2.5 cm2 (cross section), channeled aluminum extrusion rods. The frames were attached to one another with a piano hinge. The tops of the upper frames were covered with fine-mesh screen (different materials with varying mesh sizes were used but 60% greenhouse shade cloth usually sufficed). Floatation was provided by filling the hollow aluminum extrusion rods with polystyrene. The cages were placed at each site on the water surface with the screened side up. To prevent lateral movement, the cage was anchored by passing poles through U-bolts attached to the outside of the lower frame. The ends of the poles were forced into the bottom soil. The hinged upper frame provided access into the cage for fly releases and for later assessments. Numerous variations of the basic release technique were employed. For example, individuals from both laboratory colonies and from previously established field sites were released. Most of the individuals from laboratory colonies were progeny of the original quarantine colony. However, some freshly acquired material from overseas was released at selected sites. All life stages were released although the majority were eggs or larvae. Immature stages were transferred on or within the plant tissue to minimize mortality. The number of releases made and the total number of individuals released varied among sites. In some cases, predetermined quantities were released. In other instances, releases were made repeatedly and were discontinued only after evidence of establishment was obtained. The brief growing season limited the numbers released at more northern sites (i.e., Alabama). Assessment of establishment. Sites were checked for the presence of H. pakistanae by a variety of qualitative sampling methods. These included hand collection of adults from the water surface, microscopic examination of randomly selected hydrilla sprigs for the presence of immatures and associated damage, extraction of immatures from hydrilla samples via Berlese funnels, and

RELEASE AND ESTABLISHMENT OF Hydrellia pakistanae

67

FIG. 1. A typical sleeve cage used as an egging chamber for Hydrellia pakistanae. Females oviposited on exposed leaves of partially submersed hydrilla sprigs. The resultant progeny were released at field sites.

collection of emerging adults from hydrilla samples held in containers for at least 30 days. We considered H. pakistanae to be established when we consistently recovered adults, using any of the above sampling methods, from release sites for five generations (i.e., about 4 months) after the last release. Voucher specimens were prepared and the identity of the first specimens were verified by Dr. Richard L. Deonier (Gainesville, FL), a taxonomic expert on the genus Hydrellia. Voucher specimens were also deposited in the Florida State Collection of Arthropods (Gainesville, FL). RESULTS AND DISCUSSION

Hydrellia pakistanae is now clearly established at several of the release sites in the southeastern United States, especially in southern Florida. This is indicated by the fact that populations persisted from at least 70%

of the release attempts; i.e., H. pakistanae adults were collected at least 4 months after releases were discontinued (Tables 1–3). Hydrellia pakistanae has now been present at some of these sites for over 6 years. Populations at many of these sites became large enough to support redistributions of H. pakistanae to other locations (Table 3). For example, over 4000 adults were collected at the Miami Canal in south Florida during a single week in June 1993 for incorporation into colonies at TVA and WES. A large population of H. pakistanae existed at the Miami Canal even though H. pakistanae was never released there, the closest release site being over 15 km away. We also collected an estimated 2 million immatures from a pond in Muscle Shoals, Alabama, which were later moved to Lake Seminole, Georgia, and Lake Boeuf, Louisiana. These high numbers resulted from the release of only 50,000 individuals in the area. Thriving H. pakistanae populations now also occur at many sites where they were not released.

68

CENTER ET AL.

FIG. 2. Hydrellia pakistanae release cages were constructed of hollow, channeled aluminum extrusion bars filled with polystyrene for floatation. U-bolts on the sides encircled poles anchored into the bottom soil to hold the cage at the surface while allowing it to rise or fall with changing water levels.

For example, during November 1991, H. pakistanae adults were recovered near Naples, Florida, at a site at least 120 km from the nearest established release site. This rapid dispersal demonstrates the vitality of these populations and supports the conclusion that H. pakistanae is permanently established. It is now difficult to find hydrilla beds in southern Florida that do not support H. pakistanae populations. This represents the first successful establishment of an introduced insect biological control agent for the management of a weed growing in the submersed aquatic environment. While these results substantiate our conclusion that H. pakistanae is well established in the southeastern United States, early releases failed to produce persistent field populations (Table 1). Success came only after altering rearing, release, and establishment techniques. Modifications included the rearing and subsequent release of larger numbers of individuals, the use of individuals not subjected to intensive, long-term laboratory rearing, as well as alterations to actual release procedures and site selection based on recognition of mitigating factors such as plant quality (Wheeler and Center, 1996). These changes, taken together, apparently facilitated establishment of H. pakistanae. The initial supply of insects for release originated from 2-year-old colonies maintained at the quarantine facility. The first releases were necessarily small and infrequent because of the limited numbers of insects available (Table 1). For example, the releases of fewer than 10,000 individuals at Lake Patrick (aka Lake

Leonore) and Rodman Reservoir did not result in establishment. We therefore released larger numbers over shorter time intervals. The first successful establishment of H. pakistanae occurred nearly simultaneously at the Palm Beach Airport and Hacienda Village pond sites (Table 2). This required an intensive effort. For example, establishment of a persistent population at the Hacienda Village site required nine successive releases made at weekly intervals that totaled nearly 30,000 individuals. This suggested that single releases and the small numbers previously released inhibited establishment and dictated the strategy of frequently releasing large numbers over shorter time intervals (Table 2). To meet the associated demand for higher numbers of individuals, we increased rearing capabilities at the Fort Lauderdale ARS laboratory and later at TVA facilities in Muscle Shoals, Alabama, and WES facilities in Vicksburg, Mississippi. These added facilities supported release efforts in Georgia, Alabama, Louisiana, and Texas as well as supplemental efforts in Florida (Table 3). In addition to laboratory-reared individuals, the releases were supplemented with fieldcollected adult flies to maximize the numbers released. We eventually released nearly 3 million individuals from laboratory and field sources. We also identified several problems that were likely impediments to establishment. These included possible inbreeding and other adverse effects of laboratory rearing. Selective breeding induced by laboratoryrearing is known to reduce competitive ability in re-

69

RELEASE AND ESTABLISHMENT OF Hydrellia pakistanae

TABLE 1 Early Hydrellia pakistanae Releases Made from October 1987 to April 1989 Release series

State

County

Florida Broward

Location

Everglades Holiday Park Florida Broward North New River Canal Florida Glades Intersection US 27 and SR 78 Florida Glades Lake Okeechobee, Big Bear Beach Florida Glades Lake Okeechobee, Fisheating Bay Florida Glades Lake Hicpochee Florida Glades Lake Hicpochee Florida Marion Rodman Reservoir Florida Osceola Lake Tohopekaliga Florida Palm Beach Lake Osborne Florida Polk Lake Patrick (Lake Lenore) Florida Polk Sears Lake Totals

First date

Quantity released

No. Eggs Duration No. site and Late Infested Source Release (days) releases visits neonates instarts Adults plants a code b method c Result d

26-Feb-88

123

2

2

6,089

0

0

1

C

O

NE

23-Jan-89



1

1

0

0

0

1

C

O

NE

28-Apr-89



2

1

609

0

0

2

C

O

NE

18-Nov-88

109

7

7

10,351

0

0

2

C

O

NE

11-Aug-88



1

1

0

0

0

1

C

O

NE

17-Mar-88 16-Mar-89 9-Nov-87

— 53 —

1 2 1

1 2 1

0 1,138 1,000

0 0 0

0 0 0

1 2 2

C C C

O O O

NE NE NE

1-Mar-89



7

1

5,055

18

110

2

C

O

NE

23-May-88 29-Oct-87

— —

1 3

1 1

3,870 8,600

0 0

0 600

2 1

C C

O O

NE NE

2-Mar-89

42

2 30

2 21

3,266 37,761

0 18

0 710

2

C

O

NE

a

Infested plants containing unknown quantities of larvae and/or eggs. C, from laboratory cultures. c O, released in open hydrilla beds. d NE, no evidence of establishment. b

leased populations (Dame and Birkenmeyer, 1975; Bush and Neck, 1976; Chambers et al., 1983; and Grodowitz et al., 1992). As a case in point, observations of adult H. pakistanae behavioral changes suggested that the original individuals had adapted to laboratory conditions and were therefore poorly suited to natural habitats. This was deduced from a progressive reduction of adult activity and reproduction. We therefore imported new colonies from India and Pakistan. About 2000 immatures from these new colonies were released during November of 1990 in southern Florida at two small catchment ponds in northern Broward County (designated the Lakeview sites in Table 1). The Indian strain was released at the north pond (about 1700 immatures) and the Pakistani strain at the south pond (about 300 immatures). Two months later, H. pakistanae was not found at the north site but 15 individuals were collected at the south site at the point of their release. Hence, it appears that the Pakistani strain of H. pakistanae colonized the south site despite the small numbers released. This suggests that insects released immediately after importation directly in the field may indeed establish more readily. By the time we received

this fresh stock, though, we had established several field populations at other locations using insects from old cultures, so either approach can be effective. More research is needed on the effects of laboratory rearing on field establishment and the subsequent performance of biological control agents. Another possible problem was associated with the life stage released. Earliest releases of H. pakistanae were composed mainly of eggs (Table 1). Egg-laden sprigs were transferred directly from oviposition chambers to field sites. These infested sprigs were usually submersed among leafy stems in the upper layer of the existing hydrilla beds. However, eggs are usually laid at the water surface in protected sites on the plant tissue. With no establishment resulting from the exclusive release of eggs, we suspected that poor survivorship resulted from submerging the eggs too deeply. Moreover, we noted that eggs were easily dislodged from the hydrilla sprigs. This prompted us to release larvae rather than eggs. We believed that second and third instars would survive better because, in laboratory experiments, we observed higher mortality rates associated with first instars. Eggs were obtained as

70

CENTER ET AL.

TABLE 2 Releases of Hydrellia pakistanae Made from November 1989 to March 1992 Release series

State

County

Location

First date

No. Eggs Duration No. site and Late Infested (days) releases visits neonates instars Adults plants a

Source Release code b method c Result d

Alabama Jackson

49

12

12

0

29,512

0

2

C

C

NE

Alabama Alabama

? ?

? ?

? ?

0 0

25,000 12,633

0 0

2 2

C C

C C

NE E?

1,245 9,670 0 1,300

0 0 5,500 0

2 2 2 1

C C F, C C

C C O O, C

E? E E E

Alabama Alabama Florida Florida

Florida

Florida

Florida Florida Florida Florida

Florida Florida Florida Florida Florida Florida Georgia Georgia Georgia Totals

Comer 29-Aug-90 Bridge Jackson Chisenhall Jul-91 Lauderdale Muscle 1-Aug-91 Shoals Pond Facility Marshall Murphy Hill 1-Sep-90 Marshall Murphy Hill Aug-91 Brevard Stick Marsh 12-Nov-91 Broward Lake View, 5-Nov-90 North Pond Broward Lake View, 14-Nov-90 South Pond Glades Lake 6-Nov-90 Okeechobee, Fisheating Bay Jefferson Wacissa 31-Oct-90 River Jefferson Wacissa 22-Feb-91 River Lake Lake Robin 20-Dec-90 Hood Okeechobee Lake 21-Nov-90 Okeechobee, Eagle Bay Osceola Lake Toho13-Dec-90 pekaliga Osceola Lake Toho12-Feb-92 pekaliga Polk Cattle Farm 10-Feb-92 Lake Broward Hacienda 8-Jan-90 Village Glades Lake Hicpo- 2-Mar-90 chee Palm Beach Palm Beach 3-Nov-89 Airport Decatur Lake Semi26-Jun-90 nole Decatur Lake Semi2-May-91 nole Decatur Lake Semi17-Mar-92 nole

Quantity released

— ? 49 9

1 4 3

4 2

0 0 1,969 0



2

1

0

200

0

1

C

C

E



1

1

0

0

1,030

2

C

O

NE

7

9

2

10,430

3,514

3,834

1

FR, C

O

TE

52

4

4

46,390

0

0

1

FR, C

O

TE



1

1

0

0

1,200

2

F

O

NE

121

13

10

147,984

0

3,000

1

C, F

O, C

E

7

2

2

0

0

3,300

2

F

O

NE



1

1

0

0

0

1

C

O

NE

81

3

3

0

0

0

1

C

C

NE

44

10

9

0

28,831

0

2

C

C

E

189

33

17

0

31,904

0

2

C

C

E

70

11

11

5,826

11,610

0

2

C

C

E

112

21

8

20,110

16,916

3,244

2

F, FR, C

C

NE

106

23

18

96,305

51,281

300

2

F, FR

C

TE

177

21

16

1,804

93,620

C

E?

175

123

?

1 ?

0 2,000,000 F

235,141 132,946 22,668

Note. Predominant method used shown in boldface. a Infested plants containing unknown quantities of larvae and/or eggs. b C, from laboratory cultures; F, field collected (usually adults); FR, field collected, then partially reared (mostly eggs of field-collected adults). c O, released in open hydrilla beds; C, released in cage. d E, permanently established; E?, weakly established; TE, temporarily established; NE, no evidence of establishment.

71

RELEASE AND ESTABLISHMENT OF Hydrellia pakistanae

TABLE 3 Redistributions of Hydrellia pakistanae from Field-Established Colonies (August 1991 to August 1995) Release series

State Alabama

County/ parish Lauderdale

Location

Muscle Shoals Pond Facility Alabama Lauderdale Muscle Shoals Pond Facility Alabama Lauderdale Muscle Shoals Pond Facility Alabama Lauderdale Muscle Shoals Pond Facility Alabama Lauderdale Muscle Shoals Pond Facility Alabama Jackson Sublet Ferry Alabama Jackson Sublet Ferry New Spruce California e Imperial Canal California e Imperial Sumac Main Canal California e Imperial Thistle Main Canal Florida Marion Blue Run Florida Hillsborough Hillsborough River Florida Polk Lake Fannie Florida Hernando Weeki Wachee River Florida Broward Orange Brook Golf Course Florida Palm Beach West Jupiter Louisiana LaFourche Lake Bouef Louisiana LaFourche Lake Bouef Louisiana LaFourche Lake Bouef Louisiana LaFourche Lake Bouef Texas Goliad Coleto Creek Reservoir Texas Live Oak Choke Canyon Texas Live Oak Choke Canyon Texas Live Oak Choke Canyon Totals

First date

Quantity released

No. Eggs Duration No. site and Late Infested Source Release (days) releases visits neonates instars Adults plants a code b method c Result d

Aug-91

?

?

?

0

12,633

0

2

C

C

E?

4-Aug-92

17

5

5

0

33,259

0

2

C

C

E

18-Aug-93

71

49

14

52,075

51,179

0

2

C

C

E

27-Sep-94



1

1

0

528

0

2

C

C

E

27-Apr-95

126

5

2

8,277

6,188

0

2

C

C

E

30-Jun-93 7-Jul-94 5-Apr-94

90 126 —

18 22 1

18 22 1

0 43,372 0 101,671 0 3,000

0 0 0

2 2 2

C C FR

C C O

E E TE

5-Apr-94



1

1

0

3,100

0

2

FR

O

TE

5-Apr-94



1

1

0

3,125

0

2

FR

O

TE

10-Sep-92 8-Apr-94

28 26

3 3

2 2

0 0

0 1,100 0 14,481

1 2

F, FR F

O O

E E

19-Aug-94 30-Aug-94

77 —

4 1

3 1

17,100 4,000

0 0

2,700 0

2 2

FR, F FR

C O

NE E

6-Sep-94

49

3

3

16,500

0

2,410

2

F, FR

O, C

SE

27-Mar-95 15-Aug-91 3-Sep-92 1-Jan-93 20-Oct-93 2-June-94

43 117 — 25 — 27

10 14 2 4 1 8

10 14 1 3 1 4

0 0 0 0 0 0

C C C O O O

E NE NE NE NE E

20-Jan-93 23-Oct-93 26-Sep-94

— 89 179

1 9 19 185

1 8 9 127

C O O

NE TE NE

0 66,500 2 FR 71,051 0 1 C 7,160 0 2 C 8,128 0 250,000 F 3,870 0 2 C 19,584 0 2 C, FR

0 3,693 0 53,178 35,753 0 0 76,693 0 151,130 484,023 87,191

2 2 2

C C, FR C

Note. Predominant method used shown in boldface. a Infested plants containing unknown quantities of larvae and/or eggs. Numbers, when given, are estimates derived from subsamples. b C, from laboratory cultures; F, field collected (usually adults); FR, field collected, then partially reared (mostly eggs of field-collected adults). c O, released in open hydrilla beds; C, released in cage. d E, permanently established; E?, weakly established; TE, temporarily established; NE, no establishment; SE, supplemental release at established site. e California releases made by (and data courtesy of) Drs. C. Turner, C. Godfrey, and L. Anderson from insects collected in Florida.

before, but the egg-laden sprigs were held for as long as 14 days prior to release. We also suspected other problems. The few adults that emerged might have been dispersing too quickly, thereby reducing mating encounters. Predation, particu-

larly by damselflies, might have further reduced the reproductive population. Additionally, we observed coots and moorhens selectively eating the infested hydrilla sprigs placed at the sites, thereby reducing or eliminating the released individuals. This prompted us to

72

CENTER ET AL.

institute the use of shallow, floating cages to protect the plant material harboring the larvae as well as the emerging adults (Fig. 2). The cages served several functions. They did not confine the insects completely, but merely detained them, thereby increasing the likelihood of mating encounters among emerging adults. The cages also clearly demarcated the point of release, thus facilitating recovery of successive generations of flies to verify establishment. The cages were shallow (with only a 2 cm interior space above the water surface), limiting maneuverability by flying predators within the cages and presumably decreasing predation and improving adult survival. The cages also protected the plant material from larger herbivores such as coots and moorhens. The first successful establishment attempts made at the Hacienda Village and Palm Beach Airport sites involved the use of these cages (Table 2). All modifications of the original rearing and release procedures apparently contributed to the ultimate establishment of H. pakistanae. All of these corrective measures were instituted simultaneously, so it is not possible to isolate any one change that seemed more significant than the others. However, the data indicated that establishment occurred only after all of these corrective measures were implemented. Most of the early open releases using mainly small numbers of eggs failed to establish (e.g., Everglades Holiday Park) but later caged releases using large numbers of late instar larvae were quickly successful (e.g., Hacienda Village). This underscores the importance of understanding the biology of the biocontrol agent as well as the biotic and abiotic conditions associated with recipient habitats when designing release protocols. Hydrellia pakistanae populations are now well established across a broad geographical range. Populations at some sites have repeatedly failed to establish, or persisted at very low levels, despite massive releases (Tables 2 and 3). These include sites in Georgia (i.e., Lake Seminole), Florida (i.e., the Wacissa River), and Louisiana (i.e., Lake Boeuf). For example, we estimated that over 2 million insects were released throughout Lake Seminole but only very weak establishment resulted in one area (the Fireman’s Cut). While the causes of these failures are unknown, they are most likely associated with site and environmental conditions. Lack of establishment or poor establishment cannot be attributed to the release methodology or the quality of the insects released because successes were attained with even less effort at many other locations using the same techniques and insect stock. Also, the colder winters and associated senescence of hydrilla at northern sites could not account for the lack of establishment because verifiable establishment has occurred at the northernmost sites in northeastern Alabama (i.e., Muscle Shoals and Guntersville, Tables 2 and 3). In

addition, one female H. pakistanae was recovered in Lake Seminole at Fireman’s Cut during July 1993, 1 year after the last release at that specific area and 8 months after the last release (i.e., December 1992) anywhere in the reservoir. Hence, the H. pakistanae population at Lake Seminole successfully overwintered, although the resultant establishment seemed tenuous. Variation in plant quality was probably the most important factor limiting establishment at some sites. For example, Wheeler and Center (1996) found that leaf texture and tissue nitrogen concentrations affect larval survival, duration of larval stadia, and ultimate adult size of H. pakistanae. They reared H. pakistanae on field-collected plants from a variety of conditions. Coarse plants with low tissue-nitrogen concentrations were less suitable than soft plants with higher nitrogen levels. Hence, establishment of populations on low quality hydrilla might be impossible, or could require more inoculum to overcome low survival in the founding population. In conclusion, permanently established H. pakistanae populations now occur from northern Alabama to the southern tip of Florida and further releases are continuing throughout the region. Meanwhile, additional sites are being colonized by naturally dispersing populations. This success was apparently based on modifications of initial rearing and release procedures. Modifications made were based on our increased understanding of both the insect biological control agent and its associated habitats. While success has been extensive, it has not been universal. Research is needed into the causes of failed attempts (Wheeler and Center, 1996), but it is clear that success can be attained even when initial attempts fail. Harris (1988) noted that about a third of all insects released as weed biological control agents fail to establish but that half of those that do establish consume appreciable amounts of the targeted weed. A higher percentage of projects might be successful if more persistent effort was directed toward ensuring agent establishment. ACKNOWLEDGMENTS The data presented herein, unless otherwise noted, were obtained from research funded by the U.S. Army Corps of Engineers Waterways Experiment Station, Aquatic Plant Control Research Program; Florida Department of Natural Resources, Bureau of Aquatic Plant Control; and the USDA, Agricultural Research Service. Permission was granted by the Chief of Engineers to publish this information. We gratefully acknowledge the persistent efforts put forth by Mr. W. C. Durden, Mr. F. A. Dray, Mr. Boudanath Maharajh, and Ms. D. Niehaus to accomplish the objectives of this project. We are also grateful to the many student assistants who assisted with field work and rearing of the insects. We are especially indebted to Dr. Gary Buckingham and Ms. Christine Bennett who graciously provided stock cultures of H. pakistanae, processed foreign consignments of insects through quarantine, and provided invaluable advice and suggestions. This project would have been much more difficult

RELEASE AND ESTABLISHMENT OF Hydrellia pakistanae without their valuable insight. We also thank Dr. Charley Turner, Dr. Cris Godfrey, and Dr. Lars Anderson, who provided information on the release of Hydrellia pakistanae in California.

REFERENCES Baloch, G. M., and Sana-Ullah 1974. Insects and other organisms associated with Hydrilla verticillata (L.f.) L.C. (Hydrocharitaceae) in Pakistan. In ‘‘Proceedings of the III International Symposium on Biological Control of Weeds, 10–14 September 1973, Montpellier, France’’ (A.J. Wapshere, Ed.), Misc. Publ. 8, pp. 61–66. Commonwealth Institute of Biological Control, U.K. Baloch, G. M., Sana-Ullah, and Ghani, M. A. 1980. Some promising insects for the biological control of Hydrilla verticillata in Pakistan. Trop. Pest Manage. 26, 194–200. Buckingham, G. R. 1988. Reunion in Florida: Hydrilla, a weevil, and a fly. Aquatics 10, 19–25. Buckingham, G. R., Okrah, E. A., and Thomas, M. C. 1989. Laboratory host range tests with Hydrellia pakistanae (Diptera: Ephydridae), an agent for biological control of Hydrilla verticillata (Hydrocharitaceae). Environ. Entomol. 18, 164–71. Bush, G. L., and Neck, R. W. 1976. Ecological genetics of the screw-worm fly, Cochliomyia hominovorax (Diptera: Calliphoridae) and its bearing on the quality control of mass-reared insects. Environ. Entomol. 5, 821–826. Center, T. D., Cofrancesco, A. F., and Balciunas, J. K. 1990. Biological control of aquatic and wetland weeds in the southeastern United States. In ‘‘Proceedings of the VII International Symposium on Biological Control of Weeds, 6–11 March 1988, Rome, Italy’’ (E. S. Delfosse Ed.), pp. 239–262. Istituto Sperimentale per la Patologia Vegetale (Ministero dell’Agricoltura e delle Foreste), Rome, Italy. Chambers, D. L., Calkins, C. O., Boller, E. F., Ito, Y., and Cunningham, R. T. 1983. Measuring, monitoring, and improving the quality of mass-reared Mediterranean fruit flies, Ceratitus capitata (Wied.). 2. Field tests for confirming and extending laboratory results. Z. Angew. Entomol. 95, 285–303. Coulson, J. R. 1992. The TAG: Development, functions, procedures, and problems. In ‘‘Regulations and Guidelines: Critical Issues in Biological Control. Proceedings of a USDA/CSRS National Work-

73

shop, June 10–12, 1991, Vienna, VA’’ (R. Charudattan and H. W. Browning, Eds.), pp. 53–60. Inst. Food Agric. Sci., Univ. Florida, Gainesville. Dame, D. A., and Birkenmeyer, D. R. 1975. The dispersal and survival of laboratory-bred and native Glossina moristans moristans Westw. (Diptera: Glossinidae) in the field. Bull. Entomol. Res. 65, 453–457. Deonier, R. L. 1978. New species of Hydrellia reared from aquatic macrophytes in Pakistan (Diptera: Ephydridae). Entomol. Scand. 9, 188–197. Grodowitz, M. J., Lloyd, E. P., and Mckibben, G. H. 1992. Comparison of feeding and olfactory behaviors between laboratory-reared and overwintered native boll weevils (Coleoptera: Curculionidae). J. Econ. Entomol. 85, 2201–2210. Harris, P. 1988. Environmental impact of weed-control insects. BioScience 38, 542–548. Pemberton, R. W. 1980. ‘‘Exploration for natural enemies of Hydrilla verticillata in Eastern Africa,’’ U.S. Army Eng., Waterways Exp. Sta., Misc. Paper A-80-1. Sankaran, T., and Rao, V. P. 1972. An annotated list of insects attacking some terrestrial and aquatic weeds in India, with records of some parasites of the phytophagous insects. Commonw. Inst. Biol. Control Tech. Bull. 15, 131–157. Schmitz, D. C., Nelson, B. V., Nall, L. E., and Schardt, J. D. 1991. Exotic aquatic plants in Florida: A historical perspective in review of the present aquatic plant regulation program. In ‘‘Proceedings of a Symposium on Exotic Pest Plants, Miami, Florida, 2–4 Nov. 1988’’ (T. D. Center, R. F. Doren, R. L. Hofstetter, R. L. Myers, and L. D. Whiteaker, Eds.), pp. 303–323. U.S. Dept. Interior, Tech. Rep. NPS/NREVER/NRTR-91/06, Natl. Park Serv., Washington, DC. Sonder, L. W. 1979. Hydrilla infestations in California. Proc. Calif. Weed Conf. Sacramento, CA. 1979, 122–125. Sutton, D. L., and Vandiver, V. V. 1986. ‘‘Grass carp. A Fish for Biological Management of Hydrilla and Other Aquatic Weeds in Florida,’’ Fla. Agr. Exp. Sta. Bull. 867. Univ. Florida, Gainesville. Wheeler, G. S., and Center, T. D. 1996. The influence of hydrilla leaf quality on larval growth and development of the biological control agent Hydrellia pakistanae (Diptera: Ephydridae). Biol. Control 7, 1–9.