Pemberton & Cordo: Biological Control of Cactoblastis 513

POTENTIAL AND RISKS OF BIOLOGICAL CONTROL OF (: ) IN NORTH AMERICA

ROBERT W. PEMBERTON1 AND HUGO A. CORDO2 1USDA-ARS Invasive Plant Research Laboratory, Ft. Lauderdale,

2USDA-ARS South American Biological Control Laboratory, Hurlingham, Argentina

ABSTRACT

Cactoblastis cactorum Berg, an invasive and famous biological control of weeds agent, threatens numerous native and economic prickly pear cacti () in the United States and Mexico. Biological control of the moth, using a variety of approaches, is considered in- cluding: introduction of parasitoids and pathogens from the moth’s native home in South America, introduction of parasitoids from related North American (Pyralidae: Phycitnae), inundative releases of parasitoids known to attack the moth in Florida, and in- undative releases of mass reared generalists parasitoids. The primary risk of employing bi- ological control is the reduction of the many North American cactus moths, some of which probably regulate native Opuntia that can be weedy. The various biocontrol approaches are ranked according to their relative risk to the native cactus moths. The introduction of South American parasitoids or pathogens specific to the Cactoblastis (if they exist) may be the least risky approach. The introduction of South American parasitoids that can attack many cactus moths is the most risky approach because it could result in persistent “control” of these non-target native . Biological control probably can reduce the abundance of C. cactorum populations but is unlikely to prevent the spread of the moth. The relative ben- efits and risks of biological control need to be carefully assessed prior to any operational bio- logical control programs. It will be difficult to reach agreement on acceptable levels of risk, if the likely benefits can’t be predicted. Other management options need to be considered.

Key Words: Opuntia, biological control risk, cactus moths, host specificity, parasitoids, pathogens

RESUMEN

Cactoblastis cactorum Berg, una polilla famosa como agente de control biológico de malezas, amenaza numerosas especies nativas y económicas de cactos del género Opuntia en los Es- tados Unidos de América y en México. Se considera en este trabajo el control biológico de la polilla, utilizando diversas alternativas: la introducción de parasitoides y patógenos de Sud América, el área nativa de la polilla; la introducción de parasitoides de polillas de cactos de Norte América (Pyralidae: ); liberaciones inundativas de parasitoides que atacan a C. cactorum en la Florida, y liberaciones inundativas de parasitoides generalistas criados en forma masiva. El principal riesgo de la utilización del control biológico es el empobreci- miento de las muchas especies de polillas de cactos de Norte América, algunas de las cuales probablemente regulan Opuntia nativas que podrían ser malezas. Las distintas alternativas del control biológico son ordenadas de acuerdo al riesgo relativo hacia las polillas de cactos nativas. La introducción de parasitoides y patógenos específicos del género Cactoblastis (si existieran) en Sud América, sería la alternativa menos riesgosa. La introducción de parasi- toides sudamericanos que ataquen muchas polillas de cactos es la alternativa más riesgosa porque podría resultar en el “control” permanente de estos insectos que no son objetos del control biológico. El control biológico probablemente pueda reducir la abundancia de las po- blaciones de C. cactorum pero es poco probable que pueda prevenir la dispersión de la polilla. Los beneficios relativos y riesgos del control biológico necesitan ser cuidadosamente evalua- dos antes del comienzo de los programas de control. Será tal vez difícil lograr acuerdos sobre los niveles de riesgo aceptables si los beneficios esperados no pueden ser estimados. Otras opciones de manejo necesitan ser consideradas.

Cactoblastis cactorum Berg has successfully Caribbean in 1957 to control native Opuntia spe- controlled prickly pear cacti (Opuntia spe- cies that were weeds of pasture (Simmons & Ben- cies) in Australia (Dodd 1940) and in many other nett 1966). In 1989, C. cactorum was found in places in the world (Moran & Zimmermann Florida (Habeck & Bennett 1990). The insect may 1984). The moth was introduced to Nevis in the have spread on its own from other places in the

514 Florida Entomologist 84(4) December 2001

Caribbean (Johnson & Stiling 1996) or may have tant to explore the potential use and implications been accidentally introduced by the nursery in- of various biological control options. dustry (Pemberton 1995). Since that time the moth has spread throughout the Florida penin- MATERIALS AND METHODS sula where it attacks five of the six Opuntia spe- cies native to the state, including the endangered Searches of the literature were made to detect Opuntia spinosissima Miller (see Stiling & Moon the known parasitoids and diseases of Cactoblas- this volume). There is considerable concern that tis and other species of cactus moths the moth will continue to spread and attack addi- (Pyralidae: Phycitinae). Compilations of these or- tional Opuntia species. There are many native ganisms were created in different categories re- Opuntia in the southwestern U.S. and Mexico lated to various biological control options. Other that could be harmed by the moth (Strong & Pem- literature, primarily parasitoid catalogues, were berton 2000; Zimmermann et al. 2000). An esti- searched to detect records of other host insects of mated 79 Opuntia species native to Mexico and these natural enemies, to help judge their host the United States are at risk (Zimmermann et al. specificity. Criteria related to the potential benefit 2000). In addition, as many as 25 Opuntia species and risk of different biological control approaches in Mexico and three species in the United States were developed and then used to rank these ap- are used by people as food, fodder, and as proaches. the host of the cochineal dye producing scale (Zimmermann et al. 2000). RESULTS AND DISCUSSION Currently in Florida the primary host plant of C. cactorum is O. stricta (Haworth) Haworth Biological control using parasitoids of Cactoblastis which is distributed around the Gulf of Mexico cactorum from its native range in South America from Florida to and Mexico. It appears likely that the moth will spread from Florida to The 8-9 parasitoids associated with the cactus Texas using this plant. If C. cactorum reaches moth in South America are shown in Table 1. Texas and Mexico, many other Opuntia species These include one braconid larval parasitoid, one could become hosts and the moth could continue chalcidid pupal parasitoid, 5-6 ichneumonid its spread via these new hosts. This would also and one tachinid fly. Apparently no egg bring the moth into contact with endangered parasitoids are known. The chalcidid , Bra- Opuntia species that could be harmed. The ability chymeria cactoblastidis Blanchard, is suspected of C. cactorum to quickly and completely control of being a hyperparasitoid (Zimmermann et al. many exotic weedy Opuntia in disparate parts of 1979). The braconid wasp, Apanteles alexanderi the world, and also native weedy Opuntia species Brethes, has been recorded to cause more than in the Caribbean, suggest that the moth could be 30% parasitism of the larvae and the ichneu- particularly dangerous in North America. monid, Temelucha sp., was recorded to cause 5- In addition to the ecological damage caused by 30% parasitism of the larvae (Zimmermmann et the moth, public confidence in biological control al. 1979). Parasitism rates of the other parasi- practice is being injured because of the moth’s toids were not recorded, but two of the ichneu- damage and threat to native Opuntia. Unfortu- monid wasps are rare. Apanteles alexanderi nately, this situation is occurring when biological attacks other cactus moths and at least three control, a critical tool in the fight against the many other genera of Lepidoptera, and probably others , is needed more than ever before. (DeSantis 1967; Mann 1969). The Temelucha sp. The possible use of biological control against and the tachinid, Epicoronimyia mundelli (Blan- C. cactorum in North America was first raised by chard), are known to use other genera of cactus Bennett and Habeck (1992). Biological control moths (Mann 1969; Blanchard 1975; Zimmer- has controlled many insects, including Lepi- mann et al. 1979). No information about other po- doptera that feed within plants. For example, the tential hosts of the remaining parasitoids, the European pine shoot borer, Rhyacionia buoliana four ichneumonid wasps, was found, but this is (Denis & Schiffermuller), and the European corn probably due to a lack of knowledge rather than a borer, Ostrinia nubialis (Hübner), have been suc- true absence of other hosts. cessfully reduced using biological control (Kogan Some of these wasps appear to have the poten- et al. 1999; Dahlsten and Mills 1999). In this pa- tial to reduce C. cactorum populations (e.g., A. al- per, we will consider the possibilities of various bi- exanderi and Temelucha sp.). If used in the ological control approaches that might be useful United States, they probably would be able to ori- to reduce existing populations of the moth in Flor- ent to the plants and locate C. cactoblastis inside ida and adjacent Georgia, and perhaps limit its the pads where they feed. It is, however, unlikely spread. The possible benefits and perceived risks that any of these parasitoids are monophagous, so of each approach will be discussed. We do not wish their introduction for C. cactorum control could to advocate the use of biological control for C. cac- result in use of and harm to non-target Lepi- torum in North America, but we think it is impor- doptera in North America, especially native cac-

Pemberton & Cordo: Biological Control of Cactoblastis 515 specificity Presumed Reference DeSantis 1967 Zimmermann et al. 1979 Zimmermann et al. Bennett & Habeck 1992 Bennett & Habeck 1992 Bennett & Habeck Mann 1969 ? Cactoblastis of Not mentioned Mann 1969 Degree of attack hyperparasitoid Pupae?? Thompson 1943 Mann 1969 ? ??? 1992 Bennett & Habeck ? 1992 Bennett & Habeck DeSantis 1967 ? ? . (Dyar) (Meyr.) MERICA Curt. Dyar A Meyr. OUTH S sp. NATIVE

Lepidoptera sp. tapiacola Tucumania Salambona analamprella Salambona analamprella tapiacola Tucumania Plutella maculipennis Eulia loxonephes Eulia Argyrotaenia sphaleropa THEIR

IN

spp. Rare 1979 Zimmermann et al. spp. spp. spp. >30% Pupae prob. 1979 Zimmermann et al. ? Rare 1979 Zimmermann et al. ? Berg Berg ? Larvae? 1953 et al. Parker broad species Other hosts Stage attacked Heinrich Heinrich Pupae Mann 1969 ACTOBLASTIS C

C. cactorum C. cactorum C. doddi C. Cactoblastis cactorum C. doddi C. Cactoblastis cactorum C. cactorum C. Cactoblastis cactorum C. C. cactorum C. cactorum C. cactorum C. Cactoblastis OF

Cactoblastis

)

(Porter) (Cush- PARASITOIDS

Pseudobra- ( sp. sp. cactorum C. NOWN cactoblastidis ) sp. 1. K Chromocryptus doddi ABLE Chalcididae chymeria Parasitoid species Parasitoid Braconidae Apanteles alexanderi Brachymeria Brachymeria Chromocryptus doddi (Cushman) Cryptus Phyticiplex doddi Brethes Blanchard Phyticiplex eremnus cactorum Podogaster (Cushman) Podogaster man) (Probably a synonym of T 516 Florida Entomologist 84(4) December 2001 specificity Presumed Reference Zimmermann et al. 1979Zimmermann et al. ? 1975 Blanchard 1979 Zimmermann et al. Cactoblastis of 5-30% Rare 5-30% Degree of attack . MERICA A OUTH S Dyar Mann 1969 ? NATIVE

THEIR

spp. IN

Salambona analamprella Tucumania ACTOBLASTIS C

OF

spp. ? species Other hosts Stage attacked PARASITOIDS C. doddiC. cactorum C. tapiacola Tucumania Cactoblastis

Cactoblastis ) NOWN

) K Cremastus = ONTINUED sp. 1. (C ABLE Temelucha Parasitoid species Parasitoid ( Temelucha Diptera Tachinidae Epicoronimyia mundelli (Blanchard) T Pemberton & Cordo: Biological Control of Cactoblastis 517 tus moths. Careful host specificity research might identify some of these parasitoids with narrow enough host ranges to minimize potential non- target risks. Such parasitoids could be a viable Narrow Narrow specificity control option. Parasitoids probably could be ob- Presumed tained from climatic areas of South America that are similar to Florida, which would increase the probability of establishment and control.

Biological control using insect pathogens

Two types of diseases, an entomopathogenic fungus and a protozoan, have been recorded to in- fect Cactoblastis species (Table 2). The fungus Beauveria sp. caused high death rates of larvae in 1948 Pettey 1948 Pettey many places in Australia (Dodd 1940), but Beau- veria probably bassiana (Balsamo) Vuillemin caused only low levels of infection in larvae and pupae in South Africa (Pettey 1948). This fungus is a cosmopolitan disease with a broad host range. AustraliaSouth Africa Dodd 1940 Pettey 1948 Broad Broad It has been isolated from more than 200 hosts, in- South Africa 1939 Fantham cluding spiders (Feng et al. 1994). Beauveria bassiana has been used in biological control of grasshoppers, scarab beetles and other insects (Hajek & Butler 2000). It might be possible to use the fungus in an inundative biological control of C. cactorum in limited situations such as where the moth threatens the few remaining plants of Opuntia spinossisma in the Florida Keys. The in many places to 5.2% (winter) control obtained would be dependent on local 0-18.7% summer weather conditions and perhaps broader climatic conditions as suggested by the differing levels of infection of C. cactorum in Australia and South Africa. Control at best would be temporary. Many pathogens such as B. bassiana have wide host ranges, but fungal pathogens with nar- rower host ranges are known. One such fungus, Entomophaga maimaiga Humber, Shimazu & Soper, is successfully controlling the gypsy moth, Lymantria dispar L., in the U.S. (Hajek et al. 1990). No such fungal pathogens are known from Larvae, winter generationLarvae, Larvae and pupae High death rate 3.3 (summer) C. cactorum but none has been sought. Explora- All stages 0-100% winter; tion for pathogens of the moth could be produc-

tive. Demonstrating the safety of such pathogens sp. All stagesAfrica South 1939 Fantham

probably would be difficult because the laboratory . host ranges and field host ranges are different and host acceptance is not always determined by phylogenetic position of the host insect (Hajek & C. cactorum C. cactorum C. Butler 2000). cactorum C. Two Nosema were described from Cactoblastis ACTOBLASTIS C species in South Africa (Fantham 1939). One of OF these, N. cactoblastis Fantham, was described causing up to 100% infection of winter broods of

larvae in some South African localities (Pettey Fantham DISEASES

1948). This microsporidium was thought to be the bassiana cause of the lack of control of pest Opuntia by C. cactorum in these areas. Surveys for these NOWN sp. prob.

Nosema spp. were recently made in South Africa cactoblastis

sp. or spp. cactorum C. larvaein detected stages, All 0 to 5.8% summer Argentina & Cordo 2001 Pemberton ?

and Argentina, where they may have originated 2. K (Pemberton & Cordo 2001, this volume). Consid-

erations for their use against C. cactorum in ABLE (Bals.) Vuill. (Bals.) AgentFungi Beauveria Beauveria Microsporidia Nosema Host Stage attacked Degree of attack Place of attack Reference Nosema cactorum FanthamNosema Cactoblastis T 518 Florida Entomologist 84(4) December 2001

North America are discussed in that paper. 1992). Inundative releases of Brachymeria pedalis Nosema, like other pathogens, usually have their could also depress prodenialis popula- greatest impact at high host population levels. tions and perhaps other cactus moths, but should This characteristic might limit their impact on impact no other species. B. ovata releases could po- C. cactorum if they were introduced against the tentially affect a wide array of Lepidoptera. moth in North America. The relative lack of parasitoids attacking Bacillus thuringensis Berliner and its prod- C. cactorum in Florida, which could be used in this ucts are commonly used in biological control pro- approach, may reflect the limited knowledge of its grams against pest Lepidoptera (Beegle & natural enemies in Florida and also the relatively Yamamoto 1992). Cactoblastis cactorum larvae short time that the moth has been in the state. In- could probably be killed by the disease and its creased research may reveal additional parasi- products, but they would be unlikely to contact it toids that could be employed in this approach. because they feed entirely within the pads of their prickly pear hosts. Biological control using cosmopolitan generalist parasitoids known to attack Cactoblastis cactorum

Biological control using parasitoids known to attack Perhaps more suitable for the inundative bio- Cactoblastis cactorum in Florida logical control approach are two generalist parasi- toids known to attack C. cactorum (Table 4). A An alternative to importing exotic parasitoids braconid wasp, Bracon hebetor Say, parasitized up for C. cactorum control could be to employ the to 25% of the larvae in South Africa (Pettey 1948) parasitoids already known to attack the moth in and Trichogramma minutum Riley parasitized up Florida (Table 3). Only three species, two chalci- to 32% of the moth’s eggs in Australia (Dodd doid pupal parasitoids and one trichogrammatid 1940). These attack rates were naturally occur- egg parasitoid, are known (Bennett & Habeck ring. Higher rates of attack might be obtained 1992). One of the chalcidoids, Brachymeria ovata with inundative releases into C. cactorum popula- Say, attacked 55% of C. cactorum pupae at one tions. Both parasitoids already occur in North site. The other chalcidoid, B. pedalis Cresson, was America, and T. minutum has been used previ- reared from a single pupa. The host range of ously in inundative biological control of many pest B. ovata is very wide; it has been reared from di- insects (Li 1994), including the sugarcane borer verse butterflies and moths (Peck 1963). Bra- Diatraea saccharalis (Fab.) in Florida (Wilson chymeria pedalis may be limited to cactus moths 1941). Trichogramma minutum is easily reared on (Thompson 1943; Krombein et al. 1979; Mann a variety of insect eggs and is also available com- 1969). This parasitoid probably moved to C. cac- mercially. Bracon hebetor could be mass reared on torum from Walker, a native some of its known hosts that are easily raised, in- cactus moth host of the parasitoid (Krombien et cluding Galleria mellonella (L.), Plodia interpunc- al. 1979) which attacks platyopuntias (prickly tella (Hübner) and Sitotroga cerealella (Olivier). pears) in Florida. The egg parasitoid, an uniden- These parasitoids, like those already recorded tified Trichogramma, was recorded to attack two to attack the cactus moth in Florida, could be re- egg sticks. Since this species was not determined, leased into C. cactorum populations that threaten its other hosts are unknown, but most Tricho- rare cacti, and populations at the leading edge of gramma species have broad host ranges (Pinto & the moth’s expansion. These parasitoid species Stouthamer 1994). would attack non-target insects. Bracon hebetor All three of these parasitoids could be collected, parasitizes at least three families of moths (Krom- reared to increase their numbers, and then re- bein et al. 1979) but prefers concealed hosts like leased against C. cactorum in the field. The Bra- most Bracon species (Askew 1971). Trichogramma chymeria spp. might be reared on easily cultured minutum is known to attack insects in many dif- Lepidoptera such as the pyralid flour moths, Plo- ferent orders (Clausen 1940). The release of this dia and spp., if they proved to be accept- parasitoid should not be made in areas where rare able hosts. It also might be feasible to mass rear butterflies or moths occur. Trichogramma species the Trichogramma sp. on other more easily grown are usually specialists and many have moth eggs and then release on particular C. cac- poor dispersal abilities (Orr et al. 2000). Both con- torum populations. These might be employed trol of C. cactorum and impacts on non-target spe- against C. cactorum attacking vulnerable Opuntia cies brought by these wasps would probably be such as O. spinossisima in the Florida Keys, or temporary because neither wasp may persist in O. stricta, at the leading edge of the moth’s popula- the environment. Periodic releases would be nec- tions in northern Florida and southern Georgia. essary for ongoing control of C. cactoblastis. The suitability of these parasitoids for this ap- proach is unknown. Inundative releases of Tri- Biological control of Cactoblastis cactorum with chogramma could have an adverse impact on non- parasitoids of related North American cactus moths target Lepidoptera, especially rare butterflies such The 14 parasitoids listed in Table 5 are re- as those in the Florida Keys (Bennett & Habeck ported to parasitize seven cactus moth species in Pemberton & Cordo: Biological Control of Cactoblastis 519 specificity Presumed Broad Reference Mann 1969 Peck 1963 Peck C. cactorum C. in Degree of attack Stage Pupae One pupa in 1991 1992 Bennett & Habeck Moderately narrow? . attacked LORIDA F IN

(Hulst) Mann 1969 1 Walker 1979 Krombein et al. (Dyar) Mann 1969 CACTORUM

(Grote) Mann 1969 (Dyar) ACTOBLASTIS C

Ozamia fuscomaculella clarefacta Alberada parabates Other specialised phycitid moths cacti attacking Melitara prodenialis dentata M. junctolineella Olycella ATTACK

TO

pedalis ) KNOWN

Say Diverse butterflies and moths Pupae 55% To 1992 Bennett & Habeck sp. unknown Yes, Egg Two egg sticks 1992 Bennett & Habeck Broad ARASITOIDS 3. P Pseudobrachymeria Recorded host not necessarily in Florida. 1

. ( ABLE Trichogrammatidae Trichogramma Cresson Hymenoptera Chalcididae ovata Brachymeria B Parasitoid speciesParasitoid Other hosts T 520 Florida Entomologist 84(4) December 2001

three genera in the Phycitinae (Pyralidae). These moths include three Melitara, three Oly- cella and one species. All feed within the pads of prickly pear cacti. The 14 parasitoids in- Broad clude: one chalcid pupal parasitoid, five ichneu- specificity Presumed monid parasitoids (four attack larvae and one species is an egg-larval parasitoid), five braconid wasp larval parasitoids, and three tachinid flies that attack both larvae and pupae. Brachymeria pedalis (Chalcididae), Temelucha sinuatus Cush- man (Ichneumonidae), and Apanteles etiellae Vierek (Braconidae) appear to be specialists of

. cactus moths, whereas the other parasitoids ap- pear to be generalists. The degree of attack of

Krombein et al. 1979 Krombein et al. these parasitoids on their native hosts is un- MERICA known to us, so it is difficult to sense how much A parasitism they might induce in C. cactorum. The ability of these parasitoids to attack C. cactorum ORTH is also unknown, except for Brachymeria pedalis N

IN that attacks the moth in Florida (Bennett & Ha-

beck 1992). However, C. cactorum probably would be a suitable host for most of these parasitoids. Place of attack Reference OCCUR Releases of specialist parasitoids of Melitara TO

prodenialis into C. cactorum populations could be effective because the moths are closely related, use Opuntia species, and are partly sympatric, so oc- KNOWN cupy areas of climatic similarity. Introduction of any of the ten North American cactus moth parasi- ARE toids, distributed beyond the current distribution

C. cactorum C. of C. cactorum, would be classical introductions to THAT

in

Degree of attack Florida. These parasitoids are mostly from west- ern cactus moths and may be less suitable for C. cactorum control in Florida and Georgia be- cause they may be unable to survive and develop effective populations in this humid region. ACTOBLASTIS C Potential risks of biological control of Cactoblastis Larvae 25% To Africa South 1948 Pettey cactorum in North America ATTACK The potential risks associated with biological

TO

control of C. cactorum in North America are both direct and indirect. The direct risk is the reduc-

Vitula tion of populations of non-target insects by para- , KNOWN sitoids employed against the moth. The insects most likely to be harmed by the introduction of Plodia

, “specialist” parasitoids from the native South (Noctuiidae), America range of C. cactorum are related cactus moths in North America (Table 6). The level of PARASITOIDS risk will depend on the level of specialization. (Gelechiidae) There are at least 16 species of cactus moths (Pyralidae: Phycitinae) in eight genera that occur in the U.S., Mexico and the West Indies (Heinrich Heliothis obsoleta Ephestia, Galleria Many species Egg 32% To Australia Dodd 1940 Broad (Pyralidae: Phycitinae), Sitotroga

GENERALIST 1939). In addition, in the same geographic areas there are at least seven species of specialist cactus feeding moths in seven genera in four other fami- lies: the Pyraustidae (Megastes, Noctuelia, and

Say Mimorista), Gelechiidae (Metapleura and Aeroty- pia), (Dyotopasta), and Gracillariidae OSMOPOLITAN Riley (Maramara) (Mann 1969). It is probable that the

4. C highest risk of specialist parasitoids from C. cactorum in its native range would be to cactus

ABLE moths most similar to C. cactorum and which co- Hymenoptera Braconidae Bracon hebetor Trichogrammatidae Trichogramma minutum Parasitoid speciesParasitoid Other hosts Stage attacked T Pemberton & Cordo: Biological Control of Cactoblastis 521 specificity Presumed Broad Broad Broad Broad Broad Relatively narrow? . CACTORUM

Krombein et al. 1979 Krombein et al. Krombein et al. 1979 Krombein et al. Muesebeck et al. 1951 et al. Muesebeck Krombein et al. 1979 Krombein et al. 1979 Krombein et al. Krombein et al.1979 ACTOBLASTIS C

OF

Degree of attack Reference ? Mann 1969 AGENTS

CONTROL

AS

Larvae ? 1979 Krombein et al. Broad ?Larvae ? ? 1979 Krombein et al. Broad Mann 1969 Egg (emerges from larva) LarvaeLarvae ? ? Mann 1969 Mann 1969 ? Larvae ? Mann 1969 Larvae ? Mann 1969 Pupae ? Mann 1969 ,

Euzo- (Pyral- , POTENTIAL Barbara

Laphygma (Pyralidae: HAVE Elasmopal- Loxostege

(Pyralidae) , , , Ostrina Ephestia , , (Gelechiidae) MAY Ufa

, (Pyralidae), (Hesperiidae) Larvae ? Mann 1969 Hymenia , Hellula Ephestia , , , (Noctuidae); WHICH Cansarsia Pima ) , Homeosoma , Isophrictis , Psorosina , Eteiella pus Megathymus Pyraustinae) (Pyralidae) phera Anagasta Prodenia (Pyralidae) Heliothis (Torticidae) idae), Crambus HYCITINAE

: P ) ) ) = =

(Dyar) = (Hulst) Larvae ? Mann 1969 ? (Hulst) (Dyar) Walker Walker (Barnes YRALIDAE (P M. doddalis M. doddalis

M. doddalis M. Alberada parabates , (as MOTHS

(as (Grote)

(as Rumatha glaucatella fuscomaculella

Cactus moth hosts (Pyralidae: Phycitinae)(Pyralidae: Other hosts Stage attacked CACTUS

Melitara prodenialis ponderosella Olycella Olycella junctolineella Olycella junctolineellaOlycella Mimorista Olycella nephelepasa Olycella Ozamia fuscomaculella odiosella O. O. odiosella Cahela ponderosella Barnes & McD. dentataM. Ozamia M. dentata M. M. dentata M. Cactobrosis strigalis & McD.), (Hulst) denata M. Melitara prodenialis Melitara prodenialis dentata M. junctolineella Olycella Ozamia fuscomaculella O. odiosella Alberada parabates RELATED

OF

Cush- Cremas- Cresson Cresson = Vierek (Cresson) Muesebeck Pseudobra-

Riley ) ( (Cresson) prob. sp. pedalis ARASITOIDS sp. ) (Cresson) Temelucha 5. P ) C. texanus ABLE maceratum Apanteles mimoristae A. Braconidae Apanteles etiellae A. megathymi A. Mesostenus gracilis (= Temelucha Trichomma Chelonus electus T. facilis T. tus Parasitoid species Parasitoid Ichneumonidae sinuatus Temelucha man) ( chymeria Hymenoptera Chalcididae Brachymeria T 522 Florida Entomologist 84(4) December 2001 . CACTORUM specificity Presumed Broad Broad Broad Broad Broad

ACTOBLASTIS C

OF

Arnaud 1978 Arnaud 1978 Arnaud 1978 Arnaud 1978 Muesebeck et al. 1951 et al. Muesebeck 1979 Krombein et al. Mann 1969 Mann 1969 AGENTS

? Mann 1969 ? Mann 1969 ?? Mann 1969 Mann 1969 Degree CONTROL of attack Reference

AS

POTENTIAL

Larvae, pupae Larvae, pupae Larvae, Pupae pupae larvae, Larvae ? Dodd 1940 Larvae ? Dodd HAVE ,

, MAY

Plodia , , ), WHICH ) Phothorimaea Ephestia , , Epilachna HYCITINAE Cadra (Pyralidae: Pyraustinae)(Pyralidae: ? ? 1951 et al. Muesebeck ? , Laetilia (Gelechiidae) : P , (Pyralidae); Hymenoptera: Diprionidae, Hymenoptera: diverse Lepi- Tenthridindae; doptera Diprionidae Hymenoptera: Cossidae, Lepidoptera: Megathymidae, Pyralidae—Ostrina Coccinellidae ( diverse Lepidoptera diverse Lepidoptera Probably Sitotroga Galleria Anagasta Vitula YRALIDAE (P ) MOTHS

= CACTUS

M. doddalis (as sp., Texas (In cages) RELATED

Cactus moth hosts (Pyralidae: Phycitinae)(Pyralidae: Other hosts Stage attacked OF

Melitara prodenialis dentata M. junctolineella Olycella nephelepasa Olycella dentata M. junctolineella Olycella Melitara prodenialis Olycella junctolineella Olycella Noctuelia M. dentata M. Ozamia fuscomaculella O. odiosella Melitara ARASITOIDS ) P Riley Say ONTINUED sp. 5. (C Rohwer ABLE Diptera Tachinidae Phorocera texana Aldrich & Webber Phorocera comstocki Williston Lespesia aletiae Lespesia Parasitoid species Parasitoid Heterospilus melanoceph- alus Bracon hebetor T Pemberton & Cordo: Biological Control of Cactoblastis 523 BIOLOGICAL

Warm or Warm cold area A

OF

TARGETS - range Size of geographic NON

BECOME

Region of COULD occurrence

AND

PUNTIA site O

Feeding IN

FEED

THAT

NDIES I Gregarious or solitary larvae EST W THE

AND , subgenus 1 . EXICO Host plant in Opuntia PlatyopuntiaPlatyopuntia Solitary Gregarious-SolitaryCylindropuntiaCylindropuntia Pads Pads? Solitary Solitary W US-Mex Hispaniola Large Small? Stems Warm-Cold Stems Warm Texas ? Small Warm Warm U.S., M IN CACTORUM

) HYCITINAE (Wright) Platyopuntia Solitary Fruit W US-Mex Large Warm ACTOBLASTIS : P C

YRALIDAE AGAINST r

(P (Hulst) Platyopuntia Gregarious-Solitary Pads W US-Mex Small Warm ) Barnes & McDunnough Solitary Stems W US-Mex Large Warm ) (Dyar) Cylindropuntia prob. Solitary Stems US Western Large Warm Walke (Dyar) Cylindropuntia Solitary Stems W US-Mex Large Warm

EFFORT ) (Dyar) Cylindropuntia Solitary Stems W US Large Warm

(Walker) Platyopuntia Solitary Fruit FL-W Indies Large Warm MOTHS (Grote) Platyopuntia Gregarious Pads W US Large Warm-Cold

Dyar Cylindropuntia Solitary Fruit CA Small? Warm (Dyar (Dyar) Platyopuntia Gregarious-Solitary Pads W US-Mex Large Warm-Cold Walker Platyopuntia Gregarious PadsTX FL to Large Warm-Cold (Hulst (Dyar) Cylindropuntia prob. Solitary Stems Texas Small Warm? (Dyar ACTUS CONTROL 6. C Extracted from Heinrich, 1939. Extracted from Heinrich, 1 ABLE Olycella junctolineella Olycella nephelepasa O. fuscomaculella Ozamia odiosella = O. Ozamia lucidalis thalassophila O. Rumatha bihinda glaucatella R. polingella R. O. subumbrella O. Moth species Cahela ponderosella prodenialis M. Alberada parabates phryganoides Olyca A. holochlora holochlora A. Alberada bidentella T 524 Florida Entomologist 84(4) December 2001 occur with it in North America. Melitara prodeni- ulate their populations. This increase could allow alis, with gregarious larvae in pads of the same some prickly pear cacti to become artificially hosts, and an overlapping geographic range, abundant, enabling them to displace other spe- would be the most vulnerable cactus moth. It oc- cies and to dominate natural communities. These curs from Florida to Texas, and introduced parasi- plants could also become troublesome on range- toids which adopt the moth could move via this land where some Opuntia already tend to be new host from Florida to Texas, where many other weedy. Prickly pear cacti, including the Florida pad feeding cactus moths in the genera Melitara, native O. stricta, are some of the worst weeds in Olycella, Megastes, Mimorista and Mermara oc- the Old World, which has no native Opuntia. This cur. Cactus moths less likely to be attacked by in- is thought to occur, at least in part, because of the troduced South American parasitoids are lack of regulating natural enemies (Moran and probably those most dissimilar to C. cactorum. Zimmermann 1984). An additional indirect effect These would be species that are not gregarious could be to negate the successful biological control (Olycella and Ozamia), which attack flower buds of weedy Opuntia by C. cactorum in the Carib- and fruits instead of pads (Ozamia, Noctuelia), bean if introduced parasitoids spread to that re- which use cylindropuntias (the chollas) instead of gion (Bennett & Habeck 1992). Finally, it is also platyopuntias (Alberada and Cahela) or other possible that the introduction of parasitoids from genera of cacti (Yosemitia). Cactus moths that South America that attack native cactus moths have many host plants and large geographic could result in competition with native parasi- ranges would probably also be at less risk. Cactus toids that lowers the overall level of control they moths that occur in cold areas, where C. cactorum provide. This might actually result in increased and its introduced parasitoids are unlikely to col- damage by native moth species to native cacti onize, should experience the least risk. (e.g., Ferguson & Stiling 1996). The risks associated with the introduction of The criteria used to consider and rank the risk parasitoids of native cactus moths from the west- to non-target insects by different biological con- ern United States would be somewhat greater trol approaches for C. cactorum control are shown than that of specialist parasitoids introduced from in Table 7. These criteria relate in various ways to C. cactorum in South America, because they would estimating the degree of use of non-target insects. come from genera other than Cactoblastis. As Table 8 shows the rankings of the various ap- such, they would likely have broader host ranges proaches from the least to most risky. Both the cri- and be more likely to attack non-target moths. teria and the resulting rankings are not absolute, The risks associated with inundative releases but are attempts to further compare and contrast of parasitoids already associated with C. cac- relative risks. Both the least risky and most risky torum in Florida should be relatively minor be- approaches in our scheme involve the importation cause neither the geographic range nor the host of agents from the moth’s native range. The least range would be likely to increase. Large numbers risky approach is the introduction of coevolved of parasitoids could temporarily suppress host specialist parasitoids to the genus Cactoblastis. populations of insects in the area of release. The The most risky approach is the introduction of risk of inundative releases of cosmopolitan gener- stenophagous agents from the moth and its rela- alist parasitoids, such as Trichogramma minu- tives in South America, as they would likely have tum and Bracon brevicornis, could be the tem- the greatest and most persistent effects on native, porary suppression of many non-target insects in non-target cactus moths. The degree of host spec- the area of release. Because these parasitoids are ificity of prospective biological control agents is already in Florida, no increase in their geographic the key aspect in our ranking. ranges would occur. A similar degree of risk would Risks can be minimized or avoided by careful be associated with the use of generalists patho- host specificity research on parasitoids considered gens such Beauveria bassiana. The risks involved with the use of the Nosema species could be very limited if they prove to be TABLE 7. CRITERIA TO CONSIDER AND RANK THE RISK TO the narrow specialists that they are suspected to NON-TARGET SPECIES OF BIOLOGICAL CONTROL be, particularly if they have transovarial trans- APPROACHES FOR CACTOBASTIS CACTORUM. mission and kill larvae within the pads where they feed. If they are not species or genus level Degree of host specificity of agent specialists the risk will be commensurate with If new hosts will be exposed their host breadth. Relative number of new hosts that could be adopted The indirect risks of biological control relate to the potential effects caused by reducing popula- If the agent’s geographical range will increase tions of non-target insects, particularly effects to Likely persistence of non-target use their host plants. The main effect could be in- If rare species will be exposed creased populations of some Opuntia species due to the reduction of the cactus moths that help reg- Size of the treatment area Pemberton & Cordo: Biological Control of Cactoblastis 525

TABLE 8. BIOLOGICAL CONTROL APPROACHES FOR CACTO- Biological control practices during that era did not BLASTIS CACTORUM RANKED BY RELATIVE RISK usually consider possible conservation conse- TO NON-TARGET SPECIES (RANKED FROM LEAST quences of introductions. TO MOST RISKY). The use of biological control to try to correct, what might be seen in hindsight, a biological con- 1. Classical introductions from South America of trol error might seem appropriate. But it is impor- parasitoids specific to the genus Cactoblastis tant to carefully consider the actual capability of 2. Inundative releases of cactus moth parasitoids from biological control to reduce the C. cactorum threat Florida in Florida in North America. If the primary goal is to stop the 3. Inundative releases of parasitoids that attack C. cac- spread of the moth, biological control is probably torum in Florida not the best tool. Inundative release approaches, 4. Inundative releases of generalist parasitoids known as discussed above, might be both difficult and ex- to attack C. cactorum that also occur in Florida pensive to apply. The populations of the moth occur mostly along coastal Florida and Georgia. Their 5. Classical introductions of western cactus moth para- sitoids that attack gregarious larvae Opuntia host populations are scattered at various densities in complexes of natural vegetation. This 6. Classical introductions of other western cactus moth situation is quite different from the homogeneous parasitoids row crop environments where inundative biologi- 7. Classical introductions of stenophagous Cactoblastis cal control has been successful. The identification parasitoids from South America of a sex pheromone from C. cactorum, that could be combined with an insecticide to trap and kill the moth, might do a better job of stopping its spread. for introduction for biological control of C. cactorum Biological control may, however, be able to reduce in North America. However, it appears that the cur- existing populations of the moth. The possibility of rently known parasitoids of C. cactorum are un- controlling C. cactorum must be carefully weighed likely to be limited to the genus Cactoblastis, the against the direct and indirect risks of the ap- level of specificity needed to avoid use of non-target proach. Although parasitoids may be able to re- insects. This may preclude their use as biological duce populations of the moth, they probably will control agents of C. cactorum in North America. produce some degree of non-target effects. The use The relative risks and benefits of any such intro- of host-specific biological control agents is the best duction would have to be carefully evaluated. solution since they would pose the least risk. Nosema cactoblastis might be such an agent and it may be useful to control the moth. CONCLUSIONS Whether or not the risks of biological control While it appears that the threat is substantial, are acceptable will depend on the level of the the assumption that C. cactorum will devastate na- C. cactorum threat to native and economic Opun- tive North American Opuntia, as it did to exotic tia as it continues to spread and the degree to weedy Opuntia for which it was employed in Aus- which introduced biological control agents might tralia and elsewhere in the world, may be usefully disrupt native Opuntia . In light of our questioned. Native Opuntia are different than the inability to predict the benefits of a given biologi- exotic weedy Opuntia in several important respects. cal control agent, it might prove difficult to reach Native Opuntia usually occur at lower densities and agreement about what level of risk is acceptable. have complexes of specialist herbivores. These her- bivores might effectively compete with C. cactorum ACKNOWLEDGMENTS and the predators and parasitoids of these native We thank Dan Mahr, the primary organizer of the herbivores could limit the moth. The native herbi- Cactoblastis cactorum Workshop, for bringing together vores might also reduce the suitability of the Opun- the diverse participants from both the Mexico and the tia plants as food for C. cactorum. However, the fact United States to examine the problems caused by the that C. cactorum did devastate native weedy O. moth and to consider some potential solutions for its stricta and other native Opuntia on Nevis (Sim- control. We thank Dan Mahr and Ken Bloem for improv- ing this manuscript. monds & Bennett 1966) suggests that the threat is real. Cactoblastis cactorum was introduced to Nevis REFERENCES CITED Island in the Caribbean more than 40 years ago to control weedy native Opuntia in a part of the world ARNAUD, P. 1978. A host-parasite catalog of North where the cactus is an indigenous and di- American Tachinidae (Diptera). United States De- verse part of the flora (Britton & Rose 1937). Prior partment of Science and Education Ad- ministration Misc. Publ. no. 1319, 860 pp. use of C. cactorum was against weedy exotic Opun- ASKEW, R. R. 1971. Parasitic insects. Heinemann Edu- tia in parts of the world without native cacti (Moran cational Books, London, 316 pp. & Zimmermann 1984). The project in Nevis did not BEEGLE, C. C., AND T. YAMAMOTO. 1992. History of Ba- consider the conservation aspects of introducing the cillus thuringiensis Berliner research and develop- moth to the Caribbean (Fred Bennett, pers. comm.). ment. Canad. Entomol. 124: 587-616. 526 Florida Entomologist 84(4) December 2001

BENNETT, F. D., AND D. H. HABECK. 1992. Cactoblastis north of Mexico. Vol. 1 (Symphyta and Apocrita (Par- cactorum: a successful weed control agent in the asitica). Smithsonian Institution Press, Washington, Caribbean, now a pest in Florida?, pp. 21-26. In E.S. D.C., 1198 pp. Delfosse & R.R. Scott [eds.]. Proc. 8th International LI, L-Y. 1994. World-wide use of Trichogramma for bio- Symposium on Biological Control of Weeds, Canter- logical control in different crops: a survey, pp. 37-54. bury, New Zealand, CSIRO, Melbourne, Australia. In E. Wajnberg and S. A. Hassan [eds.]. Biological BLANCHARD, E. E. 1975. Primera lista anotada de Oestro- Control With Egg Parasitoids. CAB International, muscarios entomofagos argentinos (*). Completada y Oxon, U.K. actualizada por Luis DeSantis.(publicada despues de MANN, J. 1969. Cactus feeding insects and mites. U.S. su muerte, el trabajo es de 1947) Revista de Investiga- National Museum Bull. 256. ciones Agropecuarias. INTA, Bs. As., Rep. Argentina. MORAN, V. C., AND H. G. ZIMMERMANN. 1984. The biolog- Serie V, = Patologia Vegetal, vol XII, No 1, pp. 78. ical control of cactus weeds: achievements and pros- BRITTON, N. L., AND J. N. ROSE. 1937. The Cactaceae. pects. Biocontrol News and Information 5: 297-320. Carnegie Institute, Washington Vol. 1, 236 pp. MUESEBECK, C. F., K. V. KROMBEIN, AND H. K. TOWNES. CLAUSEN, C. P. 1940. Entomophagous insects. McGraw- 1951. Hymenoptera of America north of Mexico, Hill, New York, 688 pp. Snyoptic Catalog. United States Department of Ag- DAHLSTEN, D. L., AND N. J. MILLS. 1999. Biological con- riculture Agricultural Monograph No. 2, 1420 pp. trol of forest insects, pp. 761-788. In T. S. Bellows and ORR, D. B., C. GARCIA-SALAZAR, AND D. A. LANDIS. 2000. T. W. Fisher [eds.]. Handbook of Biological Control, Trichogramma nontarget impacts: a method for bio- Academic Press, San Diego. logical control assessment, pp. 111-125. In P. A. Follet DESANTIS, L. 1967. Catalogo de los Himenopteros Ar- and J. J. Duan [eds.]. Non-target Effects of Biological gentinos de la erie parasitica, incluyendo Bethy- Control. Kluwer Academic Publishers, Boston. loidea, 337 pp. PARKER, H. L., P. A. BERRY, AND A. SILVERIRA GUIDO. DODD, A. P. 1940. The biological campaign against 1953. Host-Parasite and Parasite-Host Lists of In- prickly-pear. Commonwealth Prickly Pear Board, sects Reared in the South American Parasite Labo- Brisbane, Australia, 177 pp. ratory During the Period 1940-1946. Montevideo FANTHAM, B. 1939. Nosema cactoblastis, sp. n. and (Republica Oriental del Uruguay), 100 pp. Nosema cactorum, sp. n. Microsporidian parasites of PECK, O. 1963. A catalogue of the Neartic Chalcidoidea species of Cactoblastis destructive to prickly pear. (Insecta: Hymenoptera). Canad. Entomol. Supple- Proc. Zool. Soc. London (B) 108: 688-705. ment 30, 1092 pp. FENG, M. G., T. J. POPRAWSKI, AND G. G. KHACHATOURI- PEMBERTON, R. W. 1995. Cactoblastis cactorum (Lepi- ANS. 1994. Production, formulation and application doptera: Pyralidae) in the United States: an immi- of the entomopathogenic fungus Beauveria bassiana grant biological control agent or an introduction of for insect control: current status. Biocon. Sci. and the nursery industry? Amer. Entomol. 41: 230-232. Technology 4: 3-34. PEMBERTON, R. W., AND H. CORDO. 2001. Nosema spp. FERGUSON, K. I., AND P. STILING. 1996. Non-additive ef- as potential biological control agents of Cactoblastis fects of multiple natural enemies on aphid popula- cactorum (Lepidoptera: Pyralidae): surveys for the tions. Oecologia 108: 375-379. microsporidia in Argentina and South Africa. Flor- HABECK, D. H., AND F. D. BENNETT. 1990. Cactoblastis ida Entomol. 84: 527-530. cactorum Berg (Lepidoptera: Pyralidae), a Phycitine PETTEY, F. W. 1948. The biological control of prickly new to Florida. Florida Department of Agriculture pears in South Africa. Union of South Africa Depart- and Consumer Services Division of Plant Industry ment of Agriculture Series N. 22. Circular no. 333. PINTO, J. D., AND R. STOUTHAMER. 1994. of HAJEK, A. E., AND L. BUTLER. 2000. Predicting the host the Trichogrammatidae with emphasis on Tricho- range of entomopathogenic fungi, pp. 263-276. In P. A. gramma, pp. 1-36. In E. Wajnberg and S. A. Hassan Follet and J. J. Duan [eds.]. Non-target Effects of Bio- [eds.]. Biological Control With Egg Parasitoids. CAB logical Control. Kluwer Academic Publishers, Boston. International, Oxon, U.K. HAJEK, A. E., R. A. HUMBER, J. S. ELKININGTON, B. SIMMONS, F. J., AND F. D. BENNETT. 1966. Biological con- MAY, S. R. WALSH, AND J. C. SILVER. 1990. Allozyme trol of Opuntia sp. by Cactoblastis cactorum in the Lee- and RFLP analyses confirm Entomphophaga maim- ward Islands (West Indies). Entomophaga 11: 183-189. aga responsible for 1989 epizootics in North Ameri- STRONG, D. R., AND R. W. PEMBERTON. 2000. Biological can gypsy moth populations. Proc. Nat. Acad. Sci. control of invading species—risk and reform. Science U.S.A. 87: 6979-6982. 288: 1969-1970. HEINRICH, C. 1939. The cactus-feeding Phycitinae: a THOMPSON, W. R. 1943. Catalogue of the parasites and contribution toward a revision of the American predators of insect pests. Vol. 1 Parasite host cata- pyralidoid moths of the family Phycitidae. Proc. U.S. logue. The Institute of Entomology, The Imperial Ag- National Museum 86:331-413, pls. 23-51. ricultural Bureau, Belleville, Ontario, Canada. JOHNSON, D. M., AND P. D. STILING. 1996. Host specific- WILSON, J. W. 1941. Biological control of Diatraea sac- ity of Cactoblastis cactorum (Lepidoptera: Pyral- charalis in the Florida Everglades during 1940 and idae), an exotic Opuntia-feeding moth, in Florida. 1941. Florida Entomol. 24: 52-57. Environ. Entomol. 25: 743-748. ZIMMERMANN, H. G., R. E. MCFADYEN, AND H. E. ERB. KOGAN, M., D. GELING, AND J. V. MADDOX. 1999. En- 1979. Annotated list of some cactus-feeding insects of hancing biological control in annual agricultural en- South America. Acta Zoologica Lilloana 32: 101-112. vironments, pp. 789-818. In T. S. Bellows AND T. W. ZIMMERMANN, H. G., V. C. MORAN, AND J. H. HOFF- Fisher [eds.]. Handbook of Biological Control. Aca- MANN. 2000. The renowned cactus moth, Cactoblas- demic Press, San Diego. tis cactorum: its natural history and threat to native KROMBEIN, K. V., P. D. HURD, JR., D. R. SMITH, AND B. Opuntia in Mexico and the United States of America. D. BURKS. 1979. Catalog of Hymenoptera in America Diversity and Distributions 6: 259-269.