Selbyana 20(1): 40-48. 1999.

BROMELIAD-EATING WEEVILS

1.R. FRANK Entomology and Nematology Department, University of Florida, Gainesville, FL 32611-0630 USA. E-mail: [email protected]

ABSTRACT. All known bromeliad-eating weevils-24 species in four genera (Cactophagus, Cholus, Dias­ tethus, and Metamasius)-are native to the tropics and subtropics of the Americas. Several attack more than one bromeliad genus. Six are known to attack Ananas comosus. Four (Metamasius dimidiatipennis, M. callizona, M. quadrilineatus, and M. sellatus) have traveled far from their native ranges by hitchhiking in exported/imported bromeliads. Metamasius mosieri (a small weevil) attacks bromeliads of small stature, whereas M. callizona (a larger weevil) attacks bromeliads of larger stature. utriculata popUlations in Florida (USA) have suffered devastating attack by the immigrant weevil M. callizona (from Mexico), perhaps because these plants have evolved in isolation and have no method of resistance. A biological control campaign against M. callizona is in progress.

RESUMEN. Todos los picudos comedores de bromelias conocidas, los cuales pertenecen a 24 especies en cuatro generos (Metamasius, Cactophagus, Diastethus y Cholus), son nativos de los tr6picos y subtr6picos de las Americas. Varios atacan mas de un genero de bromelias. Seis atacan Ananas comosus. Cuatro (Metamasius dimidiatipennis, M. callizona, M. quadrilineatus y M. sellatus) han viajado lejos de sus am­ bientes nativos como "polizones" en bromelias exportadas e importadas. Metamasius mosieri (un picudo pequeno) ataca bromelias de tamano pequeno, mientras M. callizona (un picudo mas grande) ataca bro­ melias de mayor tamano. Poblaciones de en la Florida (Estados Unidos) han sufrido ataques devastadores por el picudo inmigrante de Mexico M. callizona quizas porque dichas poblaciones han evolucionado aisladamente y no tienen ninglin mecanismo de resistencia. Una campana de control bio16gico en contra de M. callizona esta en progreso.

Key words: , , Metamasius, Cholus, Cactophagus, bromeliaphagy

INTRODUCTION The purposes of this paper-to review knowl­ edge of the hostplant relationships of weevils In a natural population of bromeliads, a large that feed on bromeliads, to consider the ecology proportion of dead and dying plants is an un­ of these weevils under natural and altered con­ common sight, at least when are causing ditions, and to document how some of them the damage. Although collectors of bromeliads have become pests-may stimulate additional may encounter numerous kinds of invertebrates observations and research by canopy biologists. living in leafaxils, few if any of these are detrimental to the plants. Even when lubber WEEVILS THAT EAT BROMELIADS grasshoppers, Romalea guttata (Houttuyn), swarm in the Florida Everglades, the damage Weevils (Curculionidae) are almost exclusive­ they do to the larger bromeliads seems to be ly phytophagous. The more than 40,000 de­ largely cosmetic; leaves are severely chewed, scribed species make them one of the largest but the plants are not killed. families of . The diversity of weevil spe­ Such observations, coupled thus far with the cies often is attributed to their special adapta­ study of only a few species and those not tions to diverse hostplants. known to damage bromeliads grown as orna­ Adults and larvae of 24 species of weevils mental plants, may have given rise to the pop­ belonging to four genera (Cactophagus, Cholus, ular misconception among growers that brome­ Diastethus, and Metamasius) currently are liads lack consequential insect pests. Evidence known or suspected to be specialized feeders on in this paper suggests, to the contrary, that bro­ bromeliads (TABLE 1, with distribution records meliads in natural populations in the neotropics compiled from O'Brien & Wibmer 1982 and are attacked by phytophagous insects potentially Wibmer & O'Brien 1986). Each of these 24 spe­ lethal to the plants but that these insects gener­ cies is native to some restricted area of the trop­ ally have sparse populations. Among these in­ ics and subtropics of the Americas. Although sects are weevils (Curculionidae), which upon two of the genera (Cactophagus and Metama­ arrival in a previously unoccupied locality may sius) are closely related, the relationship of the cause severe losses to bromeliad populations, other two genera is remote, indicating that "bro­ whether the plants are growing under natural or meliaphagy" evolved independently in several cultivated conditions. ancestral lineages of weevils. 40 FRANK: BROMELIAD-EATING WEEVILS 41

TABLE 1. Weevils known to eat bromeliads, or that are suspected of doing so, with their distribution and hostplants in their native areas.

Weevil Distribution Hostplants Cactophagus lojanus (Heller) Bolivia, Brazil, Ecuador, Peru ? Cactophagus miniatopunctatus Mexico, Belize, Guatemala, Hon­ Ananas comosus Chevrolat duras, Nicaragua, Costa Rica Cactophagus sanguinolentus Mexico, Belize, Guatemala, Hon­ Aechmea bracteata (Olivier) duras, Nicaragua, Costa Rica, Panama, Colombia, Ecuador Cactophagus transatlanticus Costa Rica, Colombia, Ecuador, ? (Kirsch) Venezuela Cactophagus validirostris Mexico, Guatemala, Honduras, Aechmea bracteata, Vriesea sp. (Gyllenhal) Panama, Colombia, Ecuador Cholus spinipes (Fabricius) Grenada Ananas comosus Cholus vaurieae O'Brien Venezuela Ananas comosus Diastethus bromeliarum Costa Rica ? Champion Metamasius alveolus Vaurie Costa Rica ? Metamasius bromeliadicola Costa Rica ? Champion Metamasius callizona (Chevrolat) Mexico, Guatemala, probably not Tillandsia deppeana, T. ehlersi­ Panama despite old record ana, T. heterophylla, T. limba­ ta, T. roland-gosselinii, T. streptophylla, T. utriculata, Aechmea mexicana, Catopsis sp. (see also text) Metamasius ciliatus (Champion) Mexico Aechmea bracteata Metamasius cincinnatus Nicaragua, Costa Rica, Panama, Werauhia werckleana, Guzmania Champion Ecuador circinnata Metamasius dimidiatipennis Mexico, Guatemala, Nicaragua, Ananas comosus (Jekel) Costa Rica, Panama, Colombia, Venezuela, French Guiana, Ecuador, Peru, Brazil Metamasius fasciatus (Olivier) Costa Rica, Panama, Venezuela, ? Ecuador Metamasius jlavopictus Mexico, Guatemala Tillandsia velickiana (Champion) Metamasius hebetatus (Gyllenhal) Nicaragua, Costa Rica, Panama, ? Colombia, Venezuela, Guyana, Ecuador, Peru, Bolivia Metamasius mosieri Barber Cuba, Dominican Republic, USA Tillandsia balbisiana (see also (Florida) text) Metamasius nudiventris Champion Mexico, Nicaragua, Costa Rica, ? Panama Metamasius quadrilineatus Mexico, Guatemala, EI Salvador, Tillandsia guatemalensis, T. oro­ Champion Honduras genes, T. ponderosa, T. stand­ leyi, T. yunckeri, Racinaea spi­ culosa, Werauhia nephrolepis, W. pectinata, Catopsis hahnii, C. morreniana, Billbergia sp. Metamasius quadrisignatus Montserrat, Guadeloupe, Domini­ ? (Gyllenhal) ca, Martinique, Panama Metamasius ritchiei Marshall Cuba, Jamaica Ananas comosus Metamasius rugipectus Mexico, Costa Rica, Panama ? (Champion) Metamasius sellatus Champion Mexico, Belize, Guatemala, Nica­ Tillandsia guatemalensis ragua, Panama 42 SELBYANA Volume 20(1) 1999

TABLE 2. Weevils found attacking bromeliads in other-than-natural environmental conditions.

Weevil Country Plants attacked Reference Cactophagus miniatopunctatus Mexico' Ananas comosus Vaurie 1967 Cholus spinipes Grenada' Ananas comosus Marshall 1922 Cholus vauriae Venezuela' Ananas comosus Salas & O'Brien 1997 Metamasius callizona Mexico2•3 Tillandsia spp. Frank & Thomas 1994b Mexico' Ananas comosus This paper USA (FL),,4 13 genera Frank & Thomas 1994b USA (FL)5 Native Tillandsia Frank & Thomas 1994b USA (FL)' Ananas comosus Frank & Thomas 1994b Metamasius cincinnatus Panama2 Werauhia werckleana This paper Metamasius dimidiatipennis Venezuela1.6 Ananas comosus Salas et al. 1996 Indonesia (Java)4 Imported bromeliads Vaurie 1966 Metamasius mosieri USA (FL)3 Imported Tillandsia This paper Metamasius ritchiei Jamaica' Ananas comosus Marshall 1916 , Cultivated pineapples in beds or fields. 2 Native bromeliads in habitats disturbed by humans (e.g., coffee plantations). 3 Cultivated bromeliads in shade-houses or greenhouses or gardens. 4 Imported bromeliads attacked by an adventive weevil species. 5 Native bromeliads in native habitats (attacked by an adventive weevil species). 6 Also in Costa Rica, Panama, and Ecuador (Vaurie 1966, Frank & Thomas 1994b.

Weevil larvae mine inside the stem, flowers­ here, organized under the names of the weevils. talk, or (in plants such as Ananas) the fruit. Al­ Information on natural habitats and hosts pre­ though the presence of these larvae demon­ cedes information on weevils in altered habitats, strates bromeliaphagy by a weevil species, iden­ and the latter is summarized in TABLE 2. tification keys for weevils depend on characters provided by adults. Entomologists surveying the Metamasius callizona insect fauna of a region seldom bother to collect larvae, much less take time to rear them to the Vaurie (1966) examined one museum speci­ adult stage for identification. Adults feed exter­ men of M. callizona labeled "Panama, Potreri­ nally on leaves, flowerstalks, and flowers. After llos, Dec., 1935." For the greater part of three developing through immature stages inside a weeks in June 1994, part of the time accompa­ plant, resultant adults disperse to detect other nied by A. Muzzell, the author searched unsuc­ plants suitable for attack. During their dispersal, cessfully for that species in western Panama; nor they may be found in other habitats and on were any specimens found in the insect collec­ plants that are not their hostplants, providing op­ tions housed in the Smithsonian Tropical Re­ portunity for misassociation of the hostplant. search Institute, Panama City. For lack of evi­ Entomologists who collect weevils are not nec­ dence and for lack of records from any other essarily adept at identifying plants, especially to country south of Guatemala, the existence of M. the species level in tropical countries; nor will callizona in Panama is doubtful (the specimen they necessarily collect plant samples for iden­ seen by Vaurie may have been mislabeled). tification. The hostplants of many weevil species During more than two weeks of concentrated thus are unknown. Increased knowledge will searching for Metamasius callizona in southern come gradually aided by botanists who take the Mexico (mainly the state of Veracruz) in July time to collect weevil larvae from identified bro­ 1992, M. Thomas and the author found a mere meliads, rear the larvae, and present the resultant dozen specimens in fallen Tillandsia strepto­ adult weevils to the few entomologists who spe­ phylla Scheidw., T. utriculata L., and/or T. lim­ cialize in weevil identification. bata SchItdl. The weevils were found in coffee plantations near Coatepec (19°26'N, 97°00'W), Veracruz (Frank & Thomas 1994a, 1994b). A WEEVILS IN NATURAL AND ALTERED search during several weekends from May to HABITATS Oc~ober 1995 in that same general area resulted Metamasius in no greater success in numbers of specimens collected. Other Mexican records of M. callizo­ The available information on host-relation­ na are from Tillandsia roland-gosselinii Mez in ships of bromeliad-eating weevils is presented Oaxaca (Frank & Thomas 1994b), from Aech- FRANK: BROMELIAD-EATING WEEVILS 43 mea mexicana Baker in Veracruz and Vriesia sp. Metamasius cincinnatus in Puebla (Zaragoza 1974), and from Catopsis At Fortuna (8°43'N, 82°16'W) and Boquete sp. near Coatepec. (8°47'N, 82°26'W) in the western province of In 1992 and 1995, more than 60 Metamasius Chiriqui, Panama, M. cincinnatus was encoun­ callizona weevils were found attacking brome­ tered by the author in June 1994. During one liads in a shadehouse of a grower in Veracruz, week at Fortuna, a few specimens were found in Mexico. Field-collected bromeliads had been as­ mines in stems of Guzmania circinnata Rauh sembled from various localities and concentrat­ and of an unidentified Vriesia (not in flower). ed in this shadehouse. The plants under attack During three days at Boquete, nearly 100 spec­ included Tillandsia deppeana Steud., T. ehler­ imens were found in mines in Werauhia werck­ siana Rauh, T. heterophylla E. Morren, and T. leana (Mez) 1.R. Grant still attached to fallen streptophylla. Adult weevils have been inter­ limbs of trees. The rarity of M. cincinnatus is cepted several times, at least beginning in the attested not only by my failure to obtain more early 1970s, in shipments of ornamental Tilland­ than 100 specimens during examination of hun­ sia from Mexico at U.S. airports by U.S. De­ dreds of bromeliads at many localities in west­ partment of Agriculture, and Plant ern Panama during a 3-week search, but also Health Inspection Service (APHIS) inspectors because it was unrepresented in the collections (Frank & Thomas 1994b). Discovery of M. cal­ of the Smithsonian Tropical Research Institute. lizona in a nursery in Broward County, Florida, in 1989 (O'Brien et al. 1990) indicated that a Metamasius dimidiatipennis contaminated shipment of ornamental bromeli­ There are no published records of this species ads had escaped detection by APHIS inspectors. from bromeliads in natural habitats. In Costa For the most part, the subsequent spread of M. Rica, Panama, Ecuador, and Venezuela, it has callizona to 12 counties in southern Florida has been reported to attack pineapple (Vaurie 1966, been documented (Frank & Thomas 1991a, Frank & Thomas 1994b, Salas et al. 1996). 1991b, 1994b, Frank 1997). The Florida native Specimens examined by Vaurie (1966) included bromeliads on which M. callizona has been seen two collected in 1923 from Indonesia. Meta­ to feed include T. fasciculata Sw., T. paucifolia masius dimidiatipennis may well have arrived in Baker, and T. utriculata. All three species have Indonesia as a contaminant of planting stock of been placed (or are in process of being placed) ornamental bromeliads or pineapples. on the Florida list of endangered species (Florida Administrative Code) because of attack by this Metamasius mosieri weevil. The most severe losses thus far have A few specimens of M. mosieri infesting Til­ been to T. utriculata whose populations in sev­ landsia balbisiana Schult. f. were encountered eral areas have been devastated (Frank 1997). by the author in southwestern Florida. The lar­ In southern Florida, Metamasius callizona is vae seem to have the habit of mining the flo­ now feral and has attacked cultivated bromeliads werstalks of mature plants. Only four adult spec­ growing outdoors and in greenhouses. The imens are present in the Florida State Collection plants attacked include representatives of Aech­ of , and these lack hostplant data mea, Ananas, Canistrum, Cryptanthus, Dyckia, (Frank & Thomas 1994b). Guzmania, Hohenbergia, Neoregelia, Nidula­ Specimens of Metamasius mosieri seldom rium, Orthophytum, Quesnelia, Tillandsia, and have been collected in nature, and these only in Tillandsia balbisiana; but a bromeliad grower in Vriesia (Frank & Thomas 1994b). The weevil Lee County, Florida-with a collection of im­ therefore may be capable of attacking the rarest ported bromeliads mainly in shadehouses but of Florida's native bromeliads, once it reaches with a few attached to trees-has had some their highly restricted ranges in the Everglades plants attacked by M. mosieri. Imported bro­ region of southwestern Florida. These bromeli­ meliads attacked in that collection include T. ads are Catopsis berteroniana (Schult. f.) Mez, bergeri Mez, T. bulbosa Hook., T. concolor L.B. C. floribunda L.B. Sm., C. nutans (Sw.) Griseb., Sm., T. geminiflora Brongn., T. gardneri Lindl., G. monostachia (L.) Rusby ex Mez, T. flexuosa T. hondurensis Rauh, T. ionantha Planch., T. ju­ Sw., and T. pruinosa Sw. Because the cultivated cunda Castellanos, T. rhomboidae Andre, T. plants attacked include pineapple, it may be streptophylla, T. stricta Sol. ex Ker Gawl., and asked whether this weevil attacks pineapple in T. vemicosa Baker. All of these species, like T. Mexico. An unconfirmed report suggests that it balbisiana, are of sman stature. None of these does so, but the attacks are restricted to pine­ bromeliads could hold water in axils except on apples growing near field edges in recently de­ rainy days or shortly thereafter, because the axils forested areas. have but trivial volumetric capacity, a few 44 SELBYANA Volume 20(1) 1999 drops; and none of them is a true tank brome­ Florida when pineapple cultivation was a major liad, even when fully grown. agricultural activity; M. callizona arrived in­ Metamasius mosieri originally was described stead. Only much later, when the growing of or­ from Florida and Cuba (Barber 1920) and sub­ namental bromeliads had become an important sequently was found in the Dominican Republic. horticultural activity (and pineapple production Its arrival date in Florida from Cuba or the Do­ had almost disappeared) did M. ritchiei arrive; minican Republic is unclear. If hundreds or even and now M. callizona attacks pineapple in Flor­ thousands of years ago, as is possible, then M. ida. mosieri may be considered a species native to Florida. If, however, the weevil arrived within Metamasius sellatus the last 400 years, then it would be considered This species of weevil has been collected adventive. Its small populations, despite having from Tillandsia guatemalensis in Chiapas, Mex­ been present since before 1920, and its restric­ ico (Lucas 1975). Metamasius sellatus is one of tion to southwestern Florida suggest that its nat­ the species that, like M. callizona and M. qua­ ural diet may be limited to Tillandsia balbisiana, drilineatus, has been intercepted in shipments of which likewise has small populations and a re­ ornamental Tillandsia from Mexico or Central stricted distribution. The weevil mayor may not America at U.S. airports by APHIS inspectors have specialist natural enemies in Florida. With (Frank & Thomas 1994b). In 1995, the author so few specimens collected, the limits of detec­ collected several M. sellatus specimens from tion of natural enemies may not have been flowers of Heliconia bourgaeana Petersen in reached. Veracruz, Mexico. The weevils were not ob­ served feeding on the flowers, although they Metamasius quadrilineatus could have been taking pollen or nectar. Heli­ Between January 1995 and July 1996, R. conia thus should not be thought of as a host­ Cave harvested 424 larvae and 60 pupae of M. plant for this weevil. quadrilineatus from 6920 fallen bromeliads in Honduran cloud forests. The sampled plants in­ Other Metamasius species cluded Tillandsia guatemalensis L.B. Sm., T. Metamasius ciliatus has been reported from orogenes StandI. & Williams, T. ponderosa L.B. Aechmea bracteata (Sw.) Griseb. in Veracruz, Sm., T. standleyi L.B. Sm., T. yunckeri L.B. Mexico (Zaragoza 1974); M. jlavopictus has Sm., Racinaea spiculosa (Griseb.) M.A. Spencer been collected from Tillandsia velickiana L.B. & L.B. Sm., Werauhia nephrolepis (L.B. Sm. & Sm. in Guatemala (specimen and hostplant in­ Pittendr.) J.R. Grant, Catopsis hahnii Baker, and formation supplied by D.J. Cathcart). Hostplants C. morreniana Mez (Cave 1997). Additional for the remaining bromeliad-eating species ap­ hostplant records for this weevil are from Chia­ pear to be unreported under natural conditions. pas, Mexico: T. guatemalensis (Lucas 1975), Many Metamasius species feed on plants other Billbergia sp., Tillandsia sp., and W. pectinata than bromeliads (Vaurie 1966). (L.B. Sm.) J.R. Grant (=Vriesia chiapensis Ma­ tuda) (Zaragoza 1971). Adults have been inter­ Other Weevils cepted in shipments of ornamental Tillandsia from Mexico or Central America at U.S. airports by APHIS inspectors (Frank & Thomas 1994b). Cactophagus Host records are C. validirostris from Aech­ Metamasius ritchiei mea bracteata in Veracruz, Mexico, and from Neither adults nor larvae of this species have Vriesia sp. in Puebla, Mexico, and C. sangui­ been reported from bromeliads under natural nolentus from A. bracteata in Veracruz (Zara­ conditions. Its life cycle was investigated in Ja­ goza 1974). Other species of the genus have maica where the weevil was found to attack been reported from bromeliads, but the brome­ pineapple plants growing under shaded condi­ liads were not identified specifically; other Cac­ tions, as when overgrown by weeds (Marshall tophagus species feed on plants other than bro­ 1916, Gowdey 1923, Reid 1960). Discovery of meliads. The name Cactophagus was coined be­ this damage to pineapples aroused concern in cause a few of the species attack cactus. Cac­ Florida, where pineapple production was an im­ tophagus miniatopunctatus has been reared from portant industry, and the perceived threat of the pineapple in Veracruz, Mexico (Vaurie 1967). arrival of M. ritchiei in Florida stimulated cre­ ation of the Florida State Plant Board, later Diastethus called the Division of Plant Industry (Newell The name D. bromeliarum suggests the col­ 1917). Ironically, this species did not arrive in lection of specimens from a bromeliad and per- FRANK: BROMELIAD-EATING WEEVILS 45 haps bromeliaphagy, but no information Was which it is able to feed in its new range (the found about this species nor any of the five other place to which it has immigrated). Evidence Diastethus species known from Mexico and (Frank & McCoy 1992) indicates that pest in­ Central America. sects immigrating to Florida seldom arrive with the specialist natural enemies of their home­ Cholus lands. The founding members of an overseas There is a dearth of information about natural plant population most likely have survived an hostplants of bromeliad-eating Cholus species. arduous journey; they are few in number and Cholus spinipes was reported (under the syno­ healthy (free of parasitoids, parasites, and path­ nym C. wattsi Marshall) as a pest of pineapple ogens). growing in poorly tended, shaded conditions in Grenada, West Indies (Marshall 1922). Eggs are Natural Populations laid in the flowerstaIk, which the larvae mine; they sometimes mine the fruit. Adults feed on Natural populations of epiphytic bromeliads fruits, stalks, and leaves. A second species, C. and weevils have been found in Honduras, Mex­ vaurieae, has similar habits in Venezuela ico, and Panama. Those in Honduras were in (O'Brien 1994, Salas & O'Brien 1997). Host remnants of cloud forests scattered about land plants other than pineapple have not been re­ that had been cleared for agriculture. Those in ported for these two species; their presence was Mexico were in extensive coffee plantations, noted only when they attacked pineapple grow­ where many of the larger trees were left to pro­ ing in monoculture in an altered environment. vide shade, but all undergrowth was cleared and Other species of Cholus attack palms, grasses, planted in coffee. Fallen bromeliads were and orchids; but none has been reported to attack cleared routinely from the ground, together with bromeliads (Vaurie 1976). weeds. Those at Boquete, Panama, (a weevil­ producing site) were on sparse roadside trees in DISCUSSION an area cleared for agriculture. Only those at Fortuna, Panama, (a site with few weevils) were Since native plants and insects of any natural in an extensive forest preserve with no recent area are the result of an evolutionary process disturbance. These few observations do not give that may have taken thousands of years to reach an adequate picture of the density of weevil pop­ its present state, any plant that could not coexist ulations relative to numbers of bromeliads sam­ with insect fauna no longer exists in that area. pled. The current population densities of the plants We can conclude that weevil population den­ and their insect herbivores are interdependent. If sity varied from place to place, but nowhere was a specialist insect herbivore limits the population it as dense as that observed during invasion of of a plant species, then the plant population Metamasius callizona in previously unoccupied would be expected to expand if the insect were areas of Florida. We can perhaps conclude that removed. It is not surprising then that bromeli­ the densest populations were in some way iso­ ads growing in the wild in the neotropics are not lated from continuous forest as at Boquete, Pan­ seen to be heavily attacked and killed by some ama; in cloud forest remnants in Honduras; and native insect. Typically each herbivorous insect in the shadehouse of a grower in Mexico. We species has assorted natural enemies (predators, can perhaps conclude that the traveler in those parasitoids, parasites, and pathogens) that attack countries is unlikely to find dense weevil pop­ it and limit its population. ulations in areas with extensive forest (either When natural ecosystems are disrupted by hu­ natural or with undergrowth cleared for coffee man activity, populations of some species may trees), despite dense bromeliad populations. expand, while others may decline. Humans may introduce non-native species, causing further Weevil Attacks and International change. Other (immigrant) species arriving by Hitchhiking natural means may be able to establish popula­ tions, sometimes as a consequence of changes Six species (Cactophagus miniatopunctatus, caused by humans. Yet other immigrant species Metamasius callizona, M. dimidiatipennis, M. are able to hitchhike in cargoes imported by hu­ ritchiei, Cholus spinipes, and C. vaurieae) have mans or in the vehicles used to transport the car­ been reported to include pineapple in their host­ goes (Simberloff et al. 1997, including Frank et plant range. With some of the weevils, the threat al. 1997). Establishment of an herbivorous insect to pineapple is greatest in shaded areas around species, especially when it comes without the field edges and where pineapples have become natural enemies that hold its population in check shaded by weeds. 1\\'0 instances (the arrival of in its homeland, may threaten the plants on M. dimidiatipennis in Indonesia and of M. calli- 46 SELBYANA Volume 20(1) 1999

zona in Florida) threaten pineapples grown far plants of T. utriculata-simply do not provide from the native range of pineapple. Other in­ adequate tissue for development of the larvae of stances raise questions of how long pineapple M. callizona, whose adults are moderately large. has been cultivated in Costa Rica, Panama, Co­ These plants seem not to be oviposited in by the lombia, Venezuela, Ecuador, and more so in female weevils; the most heavily attacked in na­ Grenada, Jamaica, and Mexico. Another ques­ ture in Florida are large plants of T. utriculata. tion concerns whether any of these weevils ac­ Further, far fewer weevils of this species have complished their distribution with the help of been found in T. fasciculata than in T. utriculata human intervention. in the field in Florida, no matter that T. fascicu­ Risk faces pineapple industries elsewhere. For lata is in some places more abundant and grows . example, Hawaii, among all the states of the almost as large, suggesting a preference for T. USA, has the worst record of invasion by ad­ utriculata. ventive species of animals and plants, propor­ Conversely, female Metamasius mosieri seem tionately exceeding that of Florida (U.S. Con­ to oviposit in Tillandsia balbisiana in nature in gress 1993). Hawaii has an active trade in or­ Florida, avoiding the much larger food source namental bromeliads that could provide a means provided by large T. fasciculata and T. utricu­ of hitchhiking by weevils. For example, a bro­ lata but accepting imported Tillandsia of small meliad-inhabiting mosquito native to Florida to moderate stature as alternative oviposition (Wyeomyia mitchellii [Theobald]) became estab­ sites. Adults of M. mosieri are considerably lished in Hawaii before 1980 (Shroyer 1981). smaller than those of M. callizona, so evidently Native bromeliads assembled in shadehouses their larvae need much less plant material in may be attacked by native weevils (Metamasius which to mine and develop than do larvae of M. callizona in Veracruz, Mexico). Imported, or­ callizona. Larvae of Cholus vaurieae are much namental bromeliads in such locations may be larger than those of M. callizona and presumably attacked by native (M. mosieri in southwestern would need a still greater bulk of plant material Florida) or adventive, feral weevils (M. callizo­ in which to mine and develop; thus far they have na in southern Florida and perhaps M. dimidia­ only been found in pineapple stems and fruits. tipennis in Java, Indonesia). Many other instanc­ Much remains to be learned about food needs es likely have gone unreported. The circum­ and preferences. stances include the packing of many bromeliads into a small area and, perhaps, the exclusion of Do Weevils Kill Plants? natural enemies that might in nature limit pop­ ulations of the weevils. In Florida, mining of Tillandsia utriculata Metamasius dimidiatipennis and M. callizona stems by Metamasius callizona larvae causes the have established populations in areas far from stems to break during wind and rain and the their homelands, in Java (Indonesia) and Florida plants to fall to the ground, dying. Similarly, (USA) respectively. Interceptions by APillS in­ bromeliad growers in Florida report that at least spectors of ornamental bromeliads from Mexico some ornamental bromeliads are killed outright and Central America have shown shipments to by larvae of M. callizona. In contrast, M. mo­ be infested by M. callizona, M. quadrilineatus, sieri seems to mine the ftowerstalks of T. bal­ and M. sellatus. bisiana, leaving the plants alive and able to pro­ duce a new ftowerstalk. Although M. callizona Hostplant Range larvae can destroy parent plants of T. fasciculata and some ornamental bromeliads in Florida, the At present, the weevil species listed in TABLE plants may be stimulated to produce viable off­ 1 are not believed to feed on any plant other than sets. In Florida, T. utriculata seems to rely upon bromeliads. Females of some, such as Cactop­ seed for reproduction and rarely produces off­ hagus validirostris, Metamasius callizona, M. sets. Thus the devastation caused by M. calli­ cincinnatus, and M. quadrilineatus, evidently zona in Florida may be the result of the weevil's will oviposit on (and their larvae develop in) attack on susceptible bromeliad populations that bromeliads of several genera in their natural have not evolved to resist attack by such an in­ habitats. Some, such as M. callizona, have been sect. One method of defense by T. utriculata shown to attack an even wider range of brome­ would be to produce offsets, and it is noteworthy liad genera under unnatural conditions. It may that it does not do so in Florida but does so in not be assumed, however, that all of these wee­ Guatemala where bromeliad-eating weevils exist vils will feed on bromeliads of all genera, or naturally (Isley 1987). even all bromeliads of one genus. Bromeliads of Metamasius callizona has had a devastating small stature-all Tillandsia recurvata (L.) L., effect on natural populations of Tillandsia utric­ T. setacea Sw., T. usneoides (L.) L., or young ulata in Florida by destroying the larger plants, FRANK: BROMELIAD-EATING WEEVILS 47 i.e., the breeding population (Frank & Thomas The localities mentioned in Honduras, Mexi­ 1994b). This effect can only worsen as the wee­ co, and Panama were rich in bromeliads. The vil population spreads. Further, as the weevil collection of thousands would have been easy at population encroaches on the Everglades, it will many of them, but only at a few sites were wee­ encounter Florida's rarest bromeliad species, vils detected, and only at sites in Honduras were some or all of which likely will be susceptible specialist parasitoids detected. There is no ob­ to destruction by it. Without the introduction and vious way to locate sites with abundant parasit­ establishment of a successful biological control oids. They may be present at many forested sites agent to limit the weevil population, Florida with low-density populations of weevils, but if populations of some of these bromeliad species the weevils are uncommon then the parasitoids may be driven to extinction. In contrast, popu­ will be still less common because they are high­ lations of bromeliads of small stature such as T. er in the food pyramid. At some of the few, iso­ recurvata, T. setacea, and T. usneoides likely lated sites with relatively abundant weevils, will suffer only from trivial cosmetic damage these insects may have been abundant precisely caused by occasional nibbling by adult weevils. because of the absence of parasitoids. Detection Feeding by the adult insects alone can destroy of weevil-eating Admontia at Honduran sites, such plants when weevils are confined with therefore, was a fortunate event which otherwise them alone and no other food, but weevil pop­ could have required greatly increased effort and ulations in nature likely will not be dense cost in searching. Funds permitting, research is enough to do this. planned on rearing methods for Admontia and, Bromeliad societies at all levels have contrib­ if it will attack M. callizona but no non-target uted funds and mounted Internet and print cam­ organisms, on introducing it into Florida and paigns to alert growers to the damage caused by measuring its effects (Frank 1997). Metamasius callizona and other bromeliad-eat­ ing weevil pests in Florida. Such publicity is Collaboration by Canopy Biologists spreading the word on the risks of imported bro­ meliads contaminated by weevils and other pest Bromeliads are attacked in nature by weevils insects. The best safeguard may be to dip all and by other pest insects. Among them are lar­ bromeliads to be exported in a suitable pesticidal vae of Lepidoptera (butterflies and moths) which solution and to repeat the dipping at the place could possibly cause highly consequential dam­ of arrival. The best test of the publicity's effec­ age. No review of such insects has yet been writ­ tiveness may be inspection of APHIS records to ten. The same recommendations--dipping of ex­ discover whether weevils intercepted in ship­ ported/imported plants in chemical pesticides­ ments of bromeliads have declined. If no marked should reduce or eliminate risk from them. decline in interceptions occurs, the next step This paper exposes tremendous gaps in would be to deny exporUimport permits to peo­ knowledge of weevils and other insects that eat ple who will not comply with the recommended bromeliads. Those who are in the best position dipping. to fill many of the gaps are canopy biologists who work with epiphytic plants and their asso­ Food Pyramids and Biological Control ciated insect faunas. Entomologists who take lit­ tle interest in plants and botanists who take little If knowledge of weevils that eat bromeliads interest in insects are in an inferior position, es­ is poorer than knowledge of bromeliads, then pecially if they do not have ready access to epi­ knowledge of organisms that kill the weevils is phytic bromeliads growing in tree canopies. even poorer. The fungal entomopathogen Beau­ veria bassiana (Balsamo), which attacks many ACKNOWLEDGMENTS insects, was found on some Jl,fetamasius calli­ zona from Mexico. Ants have been seen eating I thank A. MuzzeU for help collecting weevils accidentally exposed weevil larvae, and perhaps and identifying bromeliads in Panama, J.L. Pe­ some of the predatory organisms found in bro­ tersen for hospitality, and H. Stockwell for al­ meliad leafaxiIs in the neotropics (such as frogs, lowing me to examine weevil specimens in col­ lizards, scorpions, and earwigs) eat weevil lections housed in the Smithsonian Tropical Re­ adults or larvae (Frank & Thomas 1994b). The search Institute. The Florida Council of Bro­ only specialist organism known to attack wee­ meliad Societies provided partial financial vils that eat bromeliads is a yet-undescribed par­ support for these studies. CONACYT of Mexico asitoid fly of the genus Admontia, discovered in provided support during my stay in Mexico in Honduras (Cave 1997). Its only known host at 1995. Identification of weevil specimens was present is the larva of M. quadrilineatus. Per­ made by C.W. O'Brien. Hospitality and infor­ haps this species also will attack M. callizona. mation about bromeliads attacked by Metama- 48 SELBYANA Volume 20(1) 1999

sius mosieri and M. jiavopictus were provided pest likely to be introduced into Florida. Florida by H.L. and S. Smith and D.J. and L. Cathcart. State PI. Bd. Quart. Bull. 1: 47-50, pI. 23. J. Alvarez translated the abstract to Spanish. O'Brien, e.W. 1994. Two new species in the Cholus spinipes group. Trans. Amer. Entomol. Soc. 120: Earlier versions of the manuscript were re­ 412-421. viewed by J.E. Lloyd and R.E. Woodruff. This O'Brien, e.W., M.C. Thomas and J.H. Frank. 1990. A is Florida Agricultural Experiment Station jour­ new weevil pest of Tillandsia in south Florida. J. nal series R-06343. Bromeliad Soc. 40: 203-205, 222. O'Brien, e.w. and G.J. Wibmer. 1982. Annotated checklist of the weevils (Curculionidae sensu lato) LITERATURE CITED of North America, Central America, and the West Indies (Coleoptera: Curculionoidea). Mems. Amer. Entomol. lnst. 34: i-ix, 1-382. Barber, H.S. 1920. A new tropical weevil from Florida Reid, A.L. 1960. Distribution and ecology of the pine­ and Cuba. Proc. Entomol. Soc. Washington 22: apple weevil (Metamasius ritchiei). Min. Agric. 150-152, pI. 8. Lands, Jamaica, w.I., Invest. 1956-1957: 14-18. Cave, R.D. 1997. Admontia sp., a potential biological Salas, J. and C.W. O'Brien. 1997. Cholus vauriae control agent of Metamasius callizona in Florida. O'Brien (Coleoptera: Curculionidae), nueva plaga J. Bromeliad Soc. 47: 244-249. de la piiia en el Estado Lara, Venezuela. Bol. En­ Frank, J.H. 1997. Protection of Florida's native bro­ tomol. Venez. N.S. 12: 157-158. meliads by control of Metamasius callizona. J. Salas, J., C.W. O'Brien and A. Parra. 1996. Metama­ Bromeliad Soc. 47: 60-64. sius dimidiatipennis (Jekel) (Coleoptera: Curcu­ Frank, J.H. and E.D. McCoy. 1992. The immigration lionidae), plaga potencial de la piiia en Lara. Bol. of insects to Florida, with a tabulation of records Entomo!' Venez. N.S. 11: 63. published since 1970. Florida Entomol. 75: 1-28. Shroyer, D.A. 1981. Establishment of Wyeomyia mit­ Frank, J.H., E.D. McCoy, H.G. Hall, G.P. O'Meara and chellii on the island of Oahu, Hawaii. Mosquito W.R. Tschinkel. 1997. Immigration and introduc­ News 41: 805-806. tion of insects. Pp. 75-99 in D. Simberloff et ai., Simberloff, D., D.e. Schmitz and T.e. Brown, eds. eds. Strangers in Paradise. Island Press, Washing­ 1997. Strangers in Paradise. Impact and Manage­ ton. ment of Nonindigenous Species in Florida. Island Frank, J.H. and M.e. Thomas. 1991a. Metamasius ca­ Press, Washington, De. llizona kills bromeliads in southeastern Florida. J. U.S. Congress, Office of Technology Assessment. Bromeliad Soc. 41: 107-108. 1993. Harmful Non-Indigenous Species in the ---. 1991b. Metamasius callizona in four counties United States. OTA-F-565. U.S. Govt. Ptg. Office, in south Florida. J. Bromeliad Soc. 41: 253-255. Washington, De. ---. I 994a. The homeland of Metamasius calli­ Vaurie, P. 1966. A revision of the neotropical genus zona. J. Bromeliad Soc. 44: 173-176. Metamasius (Coleoptera, Curculionidae, Rhyn­ ---. 1994b. Metamasius callizona (Chevrolat) chophorinae). Species groups I and II. Bull. Amer. (Coleoptera: Curculionidae), an immigrant pest, Mus. Nat. Hist. 131: 211-338. destroys bromeliads in Florida. Canadian Ento­ ---. 1967. A revision of the neotropical genus Me­ mol. 126: 673-682. tamasius (Coleoptera, Curculionidae, Rhyncho­ Gowdey, e.C. 1923. The principal agricultural pests of phorinae). Species group III. Bull. Amer. Mus. Jamaica. Dept. Agric., Jamaica, Entomol. Bull. 2: Nat. Hist. 136: 177-268. ---. 1976. Revision of the neotropical Cholinae. 1-80, i-vi. The subgenus Cholus (Cholus). Bull. Amer. Mus. Isley, P.T. 1987. Tillandsia. The World's Most Unusual Nat. Hist. 158: 1-80. Airplants. Botanical Press, Gardena. Wibmer, G.J. and e.W. O'Brien. 1986. Annotated Lucas, K.E. "Tank Bromeliads and Their Macrofauna checklist of the weevils (Curculionidae sensu lato) from Cloud Forests of Chiapas, Mexico." (M.A. of South America (Coleoptera: Curculionidae). thesis, San Francisco State Univ., San Francisco, Mems. Amer. Entomol. lnst. 39: i-xvi, 1-563. CA, USA, 1975). Zaragoza, C.S. 1971. Variabilidad de Metamasius qua­ Marshall, G.A.K. 1916. A new weevil attacking pine­ drilineatus Champ. (Coleoptera: Curculionidae, apples in Jamaica. Bull. Entomol. Res. 7: 197- Rhynchophorinae). An. Inst. BioI., Univ. Nac. Au­ 198. t6n. Mexico (Ser. Zoo!.) 42: 73-75. ---. 1922. Some injurious neotropical weevils ---. 1974. Cole6pteros de algunas epffitas y doce (Curculionidae). Bull. Entomol. Res. 13: 59-71, nuevos registros de especies para la fauna mexi­ pI. 1-2. cana. An. lnst. BioI., Univ. Nac. Aut6n. Mexico Newell, W. 1917. The pineapple black weevil, another (Ser. Zool.) 45: 111-118.