270477.qxd

Oecologia (2000) 125:489–494DOI 10.1007/s004420000477 Robert W. Pemberton
Predictable risk to native plants in weed biological control Received: 1 March 2000 / Accepted: 21 June 2000 / Published online: 17 August 2000 Springer-Verlag 2000 Abstract Data on field host use of 112 insects, 3 fungi,
analysis of the patterns of host use by insects imported to 1 mite, and 1 nematode established for biological control control weeds. One of the primary concerns regarding of weeds in Hawaii, the continental United States, and the safety of biological control is the stability of the host the Caribbean indicate that the risk to native flora can be ranges of the organisms introduced. Biological control judged reliably before introduction. Virtually all risk is can be viewed as a grand experiment in which the stabil- borne by native plant species that are closely related to ity of the host ranges of insects and other organisms em- target weeds. Fifteen species of insects introduced for bi- ployed can be examined. The present analysis is based ological control use 41 native plant species; 36 of which upon 117 natural enemies (112 insects, 3 fungi, 1 mite are congeneric with target weeds, while 4 others belong and 1 nematode) introduced and established for biologi- to two closely allied genera. Only 1 of 117 established cal control of 55 weed species in Hawaii, the continental biological organisms uses a native plant unrelated to the United States and the Caribbean since 1902 (Julien and target weed. Thus the elements of protection for the na- Griffiths 1998). This is the first comprehensive assess- tive flora are the selection of weed targets that have few ment of the risk to non-target, native plants posed by in- or no native congeners and the introduction of biological sects introduced for biological control.
control organisms with suitably narrow diets.
Keywords Biological control of weeds · Non-target use ·
The principal source of information is specialized entomological lit-erature, which is supplemented by unpublished reports and personal communications from researchers familiar with the projects. “Use”is defined as completed life cycle of the introduced agent upon thenon-target plant species. Use does imply harm to either individuals Environmental safety is an important issue for biological or populations of the non-target plants; harm is largely unstudied for control (Andres 1985; Turner 1985; Pemberton 1985a, non-target species in biological control. Except for cacti (Opuntia 1985b, 1995, 1996; Funasaki et al. 1988; Howarth 1991; spp.) used by the moth Cactoblastis cactorum (Bergroth) in Florida, Miller and Applet 1993; Lockwood 1993a, 1993b; Car- the data on non-target use are for within the country where theagents were released. Cactoblastis cactorum was included in the ruthers and Onsager 1993; Center 1995; McEvoy 1996; analysis because its presence in Florida is a result of biological con- Simberloff and Stiling 1996; Morohasy 1996; Onstad trol in the Caribbean biogeographic region (Simmons and Bennett and McManus 1996; Hawkins and Marino 1997; Louda 1966), a region to which southern Florida belongs. Introductions af- et al. 1997; Strong 1997; Thomas and Willis 1998; ter 1994 were excluded because I judged that insufficient time hadpassed for agent population growth and dispersal to potential non- Strong and Pemberton 2000). There is now evidence of target plant species. Data from Hawaii were analyzed separately harm to a few non-target, native species caused by in- from the continental United States and Caribbean data because both sects and other organisms imported to suppress pests, but the weeds and native floras are taxonomically distinct.
a general assessment of the kinds and degrees of risk to This analysis concentrates upon the distinction between weeds with closely-related native relatives and those lacking close rela- native organisms owing to biological control is lacking.
tives in the area of introduction. Close relatives are defined as As a step towards risk management, I offer herein an congeneric species of plants and species in closely related generathat previously have been classified as in the same genus (i.e.
Cirsium and Carduus thistles). Agents established on weeds with USDA-Agricultural Research Service, Aquatic Plant Research, close relatives and on weeds without close relatives have been re- 3205 College Avenue, Ft. Lauderdale, FL 33314, USA leased for similar lengths of time; 21.4 versus 23.8 years in the continental United States and the Caribbean and 47.2 versus Tel.: +1-954-4750541 ext. 106, Fax: +1-954-4769169 50.4 years in Hawaii (calculated from Julien and Griffiths 1998).
Table 1 Known non-target native host plants of introduced biological control agents of weeds in the continental United States, the-Ca-
ribbean and Hawaii
Blataparon (= Philoxerus) vermiculare Cirsium calcareum (=C. pulchellus) Cirsium ciliolatum (Ashland thistle) Cirsium cymosum (peregrine thistle) Cirsium douglasii (Douglas’ thistle) Cirsium eatonii (=C.tweedyi) Cirsium flodmanii (Flodman’s thistle) Cirsium fontinale (fountain thistle) Cirsium hydrophilum (Suisun thistle) Cirsium occidentale (cobwebby thistle) Cirsium ownbeyi (Ownbey’s thistle) Cirsium remotifolium (=C. centaureae) Cirsium scariosum (meadow thistle) Cirsium tioganum (stemless thistle) Cirsium undulatum (wavyleaf thistle) Table 1 (continued)
in the Caribbean Antigua, Opuntia spinosissima Opuntia stricta (erect pricklypear) eae). Blutaparon (= Philoxerus) vermiculare (L.) Mears,adopted by the moth Acrola malloi (Pastrana) introduced Taxonomically isolated weeds provide much safer tar- against alligatorweed [Alternanthera philoxeroides gets for biological control than do weeds with close rela- (Martius) Grisebach], is in the same tribe (Gomphren- tives in the native flora (Table 2). Only 1 of 117 estab- eae) as Alternanthera in the Amaranthaceae. Most of the lished agents has come to use a native, non-target plant native plants (37/41) that have become hosts of biocon- unrelated to the target weeds. Virtually all of the non-tar- trol insects belong to genera of plants used by these in- get, native plant species (40/41) that have been attacked sects in their areas of origin. [Three exceptions are spe- by biological control insects are closely related to the cies of Kallstroemia, a genus limited to the Americas.
target weed species (Table 1); 36 plants belong to the Pre-release host specificity testing on weevils indicated same genus as the target weeds, while the other four spe- that Kallostroemia spp. were acceptable hosts of the cies are in two closely allied genera. (Kallstroemia, with Microlarinus biocontrol weevils (Andres and Angalet three species adopted by the two Microlarinus weevils 1963). The other exception is the alligatorweed moth introduced against puncturevines (Tribulus spp.), was Acrola malloi now using Blutaparon (= Philoxerus) previously included in the genus Tribulus (Zygophyllac- vermiculare as a host, used Philoxerus species in its na- Table 2 Comparison of non-
target use of native plants by
control projects on target weedswith close relatives with pro- previously to belong to thesame genus as the weed Total (combined)% projects with non-target use 117 (5 were released in both regions but counted only once) tive South America (Vogt et al. 1963).] Overall, 12.8% duced for control of Lantana camara L. (Verbenaceae), (15/117) of the established agents attack native plants was reported to use naio [Myoporum sandwicense (DC) (Table 2). Almost a quarter (23%, 14/61) of the agents Gray] an endemic shrub in the Myoporaceae.
established on weeds with close relatives use non-target,native plant species, compared to 1 of 61 insects thathave established on weeds without close relatives. Half (51.6%, 16/31) of the projects on target weeds with closerelatives have resulted in the non-target use of native Teleonemia scrupulosa, the single insect to be recorded plants by biological control agents, compared to 4.2% to use a native plant that is not related to its host, was (1/24) of projects on weeds without close relatives. collected in Mexico and released in Hawaii in 1902 In the continental United States and the Caribbean, at (Funasaki et al. 1988), without host specificity testing. It least 37 native plants have become hosts to ten species has been thought to be a lantana specialist (Winder and of insects introduced for biological control (Tables 1, 2).
Harley 1983). The Myoporaceae and Verbenaceae are in More than half (22/37) of these are native Cirsium this- the same order – the Lamiales (Angiosperm Phylogeny tles used by Rhinocyllis conicus (Frolich), the European Group 1998), but lantana and naio are not closely relat- weevil introduced to the continental United States ed. The true host range of T. scrupulosa is unclear. When against exotic thistles. In mainland United States and the introduced to Uganda for lantana control, it fed on and Caribbean, about a fifth (17.9%, 10/56) of the insects es- damaged sesame (Sesamum indicum L. – Pedaliaceae, tablished on target weeds with close relatives have also in the order Lamiales), and reproduced on the plant adopted non-target hosts, compared to none of the 12 to a limited extent (Davies and Greathead 1967). This agents established on weeds without close relatives.
record and unverified records on Lippia alba (Verbenac- In Hawaii, both projects conducted against weeds eae) in the Antilles, ebony (Diospyros sp., Ebenaceae) in with close relatives resulted in non-target use of native the United States (Drake and Ruhoff 1965), and a Xanth- species; four of the five insect species established now ium species (Asteraceae) in Hawaii (Funasaki et al.
use native plant species as hosts (Tables 1, 2). The pro- 1988) suggest that the insect may not be the specialist ject to control an introduced blackberry (Rubus argutus that it has been presumed to be. Recent searches in an ar- Link) led to the establishment of three insect species in ea of the island of Hawaii, where both naio and lantana the 1960s; all three use the two native Hawaiian black- grow closely together, found much T. scrupulosa damage berry species. The other project in this category, to con- to lantana but none to naio (S. Hight and P. Conant, per- trol purple nutsedge (Cyperus rotundus L.), established two insect species and one of these, a weevil (Athesape- Risks to closely related plants of target weeds are am- uta cyperi Marshall) introduced in 1925, uses a native ply illustrated by the cases of Cirsium thistles and Opun- sedge (Cyperus polystachyos Rottb.). By comparison, tia cacti in North America, each of these speciose genera only 1 of the 18 (5.6%) projects against Hawaiian weeds are threatened by a biological control insect. The Euro- that lack close relatives has produced native plant use pean weevil Rhinocyllus conicus, introduced in 1969, (Tables 1, 2). In these, only 1 of 49 (1.6%) established was first detected using a NorthAmerican native thistle biological control agents now uses a native Hawaiian 20 years ago (Reese 1977), and substantial harm to a na- host. The lacebug Teleonemia scrupulosa Stal, intro- tive thistle was reported in 1997 (Louda et al. 1997).
Table 3 Proportion of congeneric native plant species of target weeds known to be hosts for introduced biological control agents
Eurphorbia sensu lato (Euphoribiaceae) Linaria = Nuttallannthus (Schrophulariaceae) Currently 22 of 90 Cirsium thistles are known hosts of known to be used by insects introduced for biological the weevil and more use of and damage to native and control. For instance, only 1 of 46 Hypericum, 3 of rare Cirsium thistles will likely occur as the weevil 63 Senecio, and none of the 43 Salvia native species, that spreads. Some thistles, such as C. canescens Nutt. in Ne- are broadly sympatric with the target weeds in these gen- braska, may be significantly harmed (Louda et al. 1997), era, are used by the insects (Table 3).
but while others, such as C. hydrophilum (Green) Jepson These patterns of non-target, native plant use by intro- in California, will experience use but not significant duced biological control insects indicate that the risk to harm (Herr 1999). While all 90 native Cirsium spp. may native flora can be judged reliably before introduction.
be in the weevil’s physiological host range, many will The first element of protection for the native flora is the escape use and/or significant damage because they are choice of weed targets that have few native congeners.
not in the weevil’s ecological host range. For instance, Native plants in the same genus as target weeds have a thistles that flower after the adult female weevils no lon- predictable chance of being attacked, while more distant- ger lay eggs will escape use because the eggs are laid on- ly-related plants have little risk. The second element of safety is employing insects and other natural enemies Cactoblastis cactorum, an Argentine moth, was intro- with diets narrow enough to avoid damaging native duced to the Caribbean in 1957 against native cactus plants in the area of introduction. Careful determination weeds (Cock 1985). It was detected in south Florida in of the candidate insect’s field host range in its native ar- 1987 and has now spread northward at least to Georgia ea, coupled with rigorous host plant testing, will predict along the Atlantic coast. C. cactorum may have entered the potential host range in the intended area of introduc- Florida as a contaminate of commercial nursery stocks of tion. The diet needs not just to be narrow, but suitably Opuntia imported from the Caribbean (in which it was narrow. Rhinocyllus conicus, introduced to North Ameri- repeatedly detected) (Pemberton 1996), and/or as a mi- ca and Argentina (Enrique et al. 1983) to control weedy grant from Cuba or other Caribbean islands (Johnson and thistles, illustrates the point. In North America, the wee- Stiling 1996). If it spreads westward, up to 60 native vil threatens native Cirsium thistles because its host species, including ca. 12 rare species, of Opuntia in the range is too broad, but in Argentina, where there are no U.S. may be used and damaged. [Opuntia numbers esti- native thistles, the weevil’s host range is suitably narrow, mated from the U.S. flora PLANTS database (USDA- enabling it to be used without risk to native plants.
NRCS 1999); rare Opuntia from the US Fish and Wild- These data also dispel some concern that the physio- life (Federal Register 1993).] Opuntia species thought to logical, genetically-determined host ranges of herbivo- be at risk grow in the warmer areas of the United States, rous insects employed for biological control of weeds are where C. cactorum can live. Mexico has large numbers unstable. The most obvious indication of the evolution of of Opuntia species that the moth could use and possibly host ranges would be increased, or at least changed, tax- impact (Zimmermann 2000). The interception of C.
onomic breadth after introduction, which is not indicated cactorum in Laredo, Texas in 1995 in Opuntia plants from Mexico indicates that the moth is probably alreadyin Mexico (USDA-APHIS 1999). Both the thistle weevil Acknowledgements I thank P. Boldt and T. Robbins, P. Conant
and the cactus moth effectively controlled their target and S. Hight, E. Coombs, C. Dawson and T. Grant, J. Herr, J. Littlefield, G. Markin, and C. Turner for unpublished data; L.
weeds (Julien and Griffiths 1998; Simmons and Bennett Andres, J. Coulson, F. Howarth, K. Teramoto for information; and G. Buckingham and T. Center for helpful reviews.
Overall, relatively few of the native plants congeneric with the adopted native plants and the target weeds are Louda SM, Kendall D, Connor J, Simberloff D (1997) Ecological effects of an insect introduced for the biological control ofweeds. Science 277:1088–1090 Andres LA (1985) Interaction of Chrysolina quadrigemina and Maehler MR, Ford MR (1955) Teleonemia scrupulosa. Proc Hypericum spp. in California. In: Delfosse ES (ed) Proceed- ings of the V1 international symposium on biological control McAvoy T, Kok L-T, Mays WT (1987) Dispersal of Trichosiroc- of weeds, Vancouver, Canada,1985. Agriculture Canada, pp alus horridus (Panzer) (Coleoptera: Curculionidae) in south- west Virginia. J Entomol Sci 22:324–329 Andres LA, Angalet GA (1963) Notes on the ecology and host McEvoy P (1996) Host specificity and biological pest control.
specificity of Microlarinus lareynii and M. lypriformis (Cole- optera: Curculionidae) and biological control of puncture vine, Miller M, Applet G (1993) Biological control: a little knowledge Tribulus terrestris. J Econ Entomol 56:333–340 is a dangerous thing. Rutgers Law Rev 45:285–334 Angiosperm Phylogeny Group (1998) An ordinal classification for Morohasy J (1996) Host shifts in biological weed control: real the families of flowering plants. Ann Mo Bot Garden 85:531– problems, semantic difficulties or poor science. Int J Pest Bianchi F (1961) Teleonemia scrupulosa. Proc Hawaii Entomol Onstad DW, McManus ML (1996) Risk of host expansion by par- asites of insects. BioScience 46:430–435 Carruthers RI, Onsager JA (1993) Perspective on the use of exotic Pemberton RW (1985a) Native plant considerations in the biologi- natural enemies for biological control of pest grasshoppers cal control of leafy spurge. In: Delfosse ES (ed) Proceedings of (Orthroptera: Acrididae). Environ Entomol 22:885–903 the V1 international symposium on biological control of weeds, Center T (1995) Selection criteria and ecological consequences of Vancouver, Canada, 1984. Agriculture Canada, pp 365–390 importing natural enemies (book review). Biodivers Conserv Pemberton RW (1985b) Native weeds as candidates for biological control research. In: Delfosse ES (ed) Proceedings of the V1 Cock M (1985) A review of biological control of pests in the international symposium on biological control of weeds, Van- Commonwealth Caribbean and Bermuda up to 1982. Tech couver, Canada, 1984. Agriculture Canada, pp 869–877 Pemberton RW (1995) Cactoblastis cactorum (Lepidoptera: Pyral- Davies JC, Greathead DJ (1967) Occurrence of Teleonemia idae) in the United States: an immigrant biological control scrupulosa on Sesamum indicum Linn. in Uganda. Nature 230: agent or an introduction of the nursery industry. Am Entomol Diehl J, McEvoy P (1990) Impact of the cinnabar moth (Tyria ja- Pemberton RW (1996) The potential of biological control for the cobaeae) on Senecio triangularis, a non-target native plant in suppression of invasive weeds in southern environments. Oregon. In: Delfosse ES (ed) Proceedings of the VII interna- tional symposium on biological control of weeds. Rome, Italy, Poinar GO Jr (1964) Observations on nutgrass insects in Hawaii 1988. Ministry of Agriculture, Forestry, Rome, and CSIRO, with notes on the host range of Bactra truculenta Meyrick and Athesapeuta cyperi Marshall. Proc Hawaii Entomol Soc Drake CJ, Ruhoff FA (1965) Lacebugs of the world. Bull U S Nat Reese NE (1977) Impact of Rhinocyllus conicus on thistles in Enrique AE, Cordo HA, de Crouzel IS, Gimenez Tanzi MR (1983) southwestern Montana. Environ Entomol 6:839–842 Importacion de Rhinocyllus conicus Froelich y Trichosirocal- Simberloff D, Stiling P (1996) How risky is biological control? lus horridus Panzer para el control biologico fr lod “Cardos” en la Argentina. Maleza (ASAM, Argentina) 11:232–241 Simmons F, Bennett F, Biological control of Opuntia spp. by Cac- toblastis cactorum in the Leeward Islands (West Indies) En- Funasaki GY, Lai P-Y, Nakahara LM, Beardsley J, Ota AK (1988) A review of biological control introductions in Hawaii: Strong DR (1997) Fear no weevil. Science 277:1058–59 1890–1985. Proc Hawaii Entomol Soc 28:105–160 Strong DR, Pemberton RW (2000) Biological control of invading Goeden D, Ricker DW (1986) Phytophagous insect faunas of the species: risk and reform. Science 288:1969–1970 two most common native Cirsium thistles, C. californicum and Thomas MB, Willis AJ (1998) Biological control - risky but nec- C. proteanum, in southern California. Ann Entomol Soc Am Turner C E (1985) Conflicting interests and biological control of Hawkins BA, Marino PC (1997) The colonization of native phy- weeds. In: Delfosse ES (ed) Proceedings of the V1 interna- tophagous insects in North America by exotic parasitoids.
tional symposium on biological control of weeds, Vancouver, Canada, 1984. Agriculture Canada, pp 203–225 Herr J (1999) Non-target impact of Rhinocyllus conicus (Coleop- Turner CE, Pemberton RW, Rosenthal SS (1987) Host utilization tera: Curculionidae) on rare native California Cirsium spp.
of native Cirsium thistles (Asteraceae) by the introduced wee- thistles. Abstracts of the X International symposium on bio- vil Rhinocyllus conicus (Coleoptera: Curculionidae) in Cali- logical control of weeds, Bozeman, Montana, USA, p 43 Howarth FG (1991) Environmental impacts of classical biological USDA-NRCS (1999) The PLANTS database (http://plants.usda.gov/ Johnson DM, Stiling PD (1996) Host specificity of Cactoblastis cactorum (Lepidoptera: Pyralidae), an exotic Opuntia – feed- Vogt GB, Quimby PC Jr, Kay SH (1992) Effects of weather on the ing moth, in Florida. Environ Entomol 25:743–748 biological control of alligatorweed in the lower Mississippi Julien MH, Griffiths MW (eds) (1998) Biological control of River region, 1973–83. USDA, ARS Tech Bull 1766 weeds: A world catalogue of agents and their target weeds, Winder JA, Harley KS (1983) The phytophagous insects on lant- edn 4. CAB International, Wallingford, UK ana in Brazil and their potential for biological control in Aus- Lockwood JA (1993a) Environmental issues involved in biologi- cal control of rangeland grasshoppers (Orthroptera: Acrididae) Zimmermann HG (2000) A new pest on Opuntiae in wait for with exotic agents. Environ Entomol 22:503–518 Mexico. In: Aguirre Rivera JR, Reyes Aguero JA (eds) Pro- Lockwood JA (1993b) Benefits and costs of controlling rangeland ceedings of the VIII congress and the VI international con- grasshoppers (Orthroptera: Acrididae) with exotic organisms: gress on the understanding and utilization of cactus pear, Uni- search for the null hypothesis. Environ Entomol 22:904–914 versity of San Potosi, Mexico, 1999. pp 333–341 (in press)

Source: http://www.ent.orst.edu/insect_ecology/PDF/Pemberton%202000.pdf