SEE ALSO THE FOLLOWING ARTICLES Usually, this is achieved by the introduction of additional Agriculture / Biological Control, of Plants / Integrated Pest natural enemies (either new species or more of the same Management / Ladybugs / Wasps species) or by encouraging an increase in the abundance of local natural enemies by habitat modifi cation. FURTHER READING

Bigler, F., D. Babendreier, and U. Kuhlmann, eds. 2006. Environmental BIOLOGICAL CONTROL STRATEGIES Impact of Invertebrates for Biological Control of Arthropods. Wallingford, Weed biological control efforts can be grouped into UK: CABI Publishing. Clausen, C. P. 1978. Introduced Parasitoids and Predators of Arthropod three different strategies: introduction, augmentation, Pests and Weeds: A World Review. Agricultural Handbook No. 480. and conservation. Introduction, or classical biological Washington, DC: USDA. control, is the movement of selected natural enemies of Greathead, D. J., and A. H. Greathead. 1992. Biological control of insect pests by parasitoids and predators: The BIOCAT database. Biocontrol a targeted plant species from its native range into the News and Information 13(4): 61N–68N. new area invaded by the weed. It is common for exotic Gurr, G., and S. Wratten, eds. 2000. Biological Control: Measures of Suc- weeds to lack natural enemies in their new area of inva- cess. Dordrecht: Kluwer. sion. When a plant is imported as an ornamental or crop Gurr, G. M., S. D. Wratten, and M. A. Altieri. 2004. Ecological Engineer- ing for Pest Management: Advances in Habitat Manipulation for Arthro- plant, effort is taken to ensure it is free of insects, mites, pods. Ithaca, NY: Cornell University Press. and disease. Similarly, plants accidentally introduced Heinz, K. M., R. G. Van Driesche, and M. P. Parrella, eds. 2004. Biocontrol are usually transported as seeds or pieces of stem or rhi- in Protected Culture. Batavia, IL: Ball Publishing. Neuenschwander, P., C. Borgemeister, and J. Langewald, eds. 2003. zome, plant parts too small to include or support natural Biological Control in IPM Systems in Africa. Wallingford, UK: CABI enemies that require leaves or larger pieces of the plant Publishing. to survive. If the plant escapes cultivation or is acciden- Van Driesche, R., M. Hoddle, and T. Center, eds. 2008. Control of Pests and Weeds by Natural Enemies: An Introduction to Biological Control. tally introduced and begins to spread, the reason for its Oxford: Blackwell. success as an invader may be due to the difference in the Waage, J. K., and D. J. Greathead. 1988. Biological control: Challenges level of herbivory it receives compared to plants native to and opportunities. Philosophical Transactions of the Royal Society of the area, which are damaged and infected by their own London B 318: 111–128. Wajnberg, E., J. K. Scott, and P. C. Quimby, eds. 2001. Evaluating Indi- group of natural enemies. This difference in the level rect Ecological Effects of Biological Control. Wallingford, UK: CABI of herbivory or disease has been proposed as one of the Publishing. reasons that exotic plants become invasive and is called the “enemy release hypothesis.” Classical biological con- trol involves the discovery of specifi c natural enemies in a plant’s native range, an evaluation of their safety (through host-specifi city testing) and effi cacy, and the BIOLOGICAL CONTROL, study of their release and establishment in the invaded range. The objective is for the exotic natural enemies to OF PLANTS permanently reduce the weed population. It is generally accepted that the weed will not to be eradicated and MICHAEL J. PITCAIRN that both the weed and biological control agents will California Department of Food and Agriculture, Sacramento permanently persist, but at densities below economic or ecological threshold levels where the weed is no lon- Plant populations are limited by many factors, including ger problematic. Classical biological control is the most abiotic conditions, resource limitations, germination safe common biological control method used against plants sites, plant-to-plant competition, predation (herbivory), and should generally be part of an integrated pest man- pollination, and plant disease. The large and various agement program. groups of herbivores and diseases that consume or infect Augmentative biological control is the addition of a particular plant are called its natural enemies, and the natural enemies, either native or exotic, to provide a damage they impart due to their feeding or infection temporary boost to the background level of herbivory. works together with the other limiting forces to main- Natural enemies released in an augmentative program are tain a plant’s population density around some reduced usually not expected to survive past their life spans or the level. Biological control of invasive plants is a pest control growing season and often do not become permanently method where the natural enemies of an organism are established. The released organisms are mass reared in lab- intentionally manipulated to further reduce its abundance. oratory cultures so that thousands are released at a time.

BIOLOGICAL CONTROL, OF PLANTS 63 From Daniel Simberloff and Marcel Rejmánek, editors, Encyclopedia of Biological Invasions, Berkeley and Los Angeles: University of California Press, 2011.

02_Simberloff10_B_p43-91.indd 63 9/10/10 4:09:50 PM Augmentative biological control was originally developed An example is Collego, a commercial product consist- against insect pests in greenhouses and fi eld crops, where ing of spores of the fungus Colletoctricum gloeosporioides it is economically feasible to produce thousands of para- f. spp. aeschynomene, for control of northern jointvetch, a sitic or predatory insects on high value crops, especially native leguminous weed in rice and soybean crops in the if the pests are resistant to insecticides. There are very southeastern United States. Most of these products have few examples of augmentative biological control being not been economically viable because they are effective used on plants, probably because it is not cost-effective against a single weed species and must compete in a mar- compared to herbicides, which are able to control broad ketplace with broad-spectrum herbicides effective against classes of weeds. For invasive plants that infest large areas, many weed species. The use of plant pathogens as an aug- augmentative strategies are not likely to be cost-effective. mentative biological control tool has great potential and Some have called the use of sheep or goats an augmenta- needs to be explored further. tive control activity, but the use of grazing is tradition- Recently, a stem-boring wasp and a scale insect have ally considered a cultural control method because the been proposed for use in augmentative control releases animals must be herded. Some plant diseases have been against Arundo donax, a giant reed that has invaded the developed for use as bioherbicides and can be classifi ed riparian community along the Rio Grande, the river that as augmentative control. Unlike insect agents, impacts serves as a border between Mexico and the United States. from a bioherbicide may occur for several generations Both insects are exotic and were obtained from Spain, of the pathogen, but they usually do not extend beyond where A. donax is native. The proposed objective is to rear a single fi eld season. Eight pathogens worldwide have hundreds of thousands of these species in a mass-rearing been registered for use as bioherbicides against weeds. facility and then release them early in the growing season

STEPS IN A CLASSICAL WEED BIOLOGICAL must be summarized and submitted for consideration to the regu- CONTROL PROGRAM latory authority before a permit will be issued. The review process can take months to years. 1. Target selection. Identify weed species using morphological and molecular techniques and identify area of origin. Resolve 5. Implementation. Upon approval for introduction, initial release confl icts regarding the commercial or environmental value of the and establishment of the biological control agent will occur in target weed. Perform cost–benefi t analysis. fi eld nursery sites, areas with high densities of the target weed located in climatic areas deemed optimal for the control agent. 2. Foreign exploration in weed’s area of origin. Examine litera- Usually, only a few organisms (usually fewer than 1,000) are ture and explore target weed’s native range to discover and col- available for initial releases. Once they are established and their lect potential biological control agents. When extensive, native numbers increase, collections of surplus agents will be used areas with the most similar climate to the invaded range should to redistribute them throughout the invaded range. Regional be identifi ed as priority. Correct identifi cation of all collected redistribution can be facilitated through outreach events, such material is critical for purposes of safety and project success. as “fi eld days” where local land managers and property owners Plants closely related to the target weed should be examined in are invited to visit the nursery site, learn the biology of the tar- the native range to see if they are damaged by candidate control get weed and control agent, and receive a small quantity of the agents. agent for release on their property.

3. Host specifi city studies. All potential biological control agents 6. Post-release monitoring. Following their initial release, nurs- should be subjected to a series of choice and no-choice tests. ery sites should be monitored to determine whether the agent Results will be used to predict fi eld host range and potential establishes, populations begin to increase, spread occurs into risks to nontarget species after release of a control agent in the nearby plant infestations, and the new populations support col- invaded range. lections for further redistribution. Generally, an agent is consid- ered established once it has survived at least two consecutive 4. Approval of agents by government regulatory agencies. Most years after release. Monitoring should be performed to examine major countries have enacted laws to regulate the introduction of impact of the agent on the target plant population and the occur- exotic plants and animals. These laws also regulate the introduc- rence of any nontarget effects. This kind of monitoring is more tion of biological control organisms. Results of host specifi city tests detailed and should occur for several years following release.

64 BIOLOGICAL CONTROL, OF PLANTS

002_Simberloff10_B_p43-91.indd2_Simberloff10_B_p43-91.indd 6464 99/10/10/10/10 4:09:504:09:50 PMPM along the critical areas of the Rio Grande, in an attempt a plant’s reproduction, competitive ability, and growth to suppress growth of the reed. rate (Fig. 1). Conservation biological control is the manipulation of It is common for a complex of natural enemies to the habitat to encourage more activity by the local com- be introduced against a target weed. Approximately 40 munity of the pest’s natural enemies. This could include percent of past weed control projects introduced at least providing resources needed by the natural enemies (e.g., three exotic natural enemy species. A complex of agents overwintering sites) and reducing interference by other is especially needed for weeds that are widely distributed control strategies, such as mowing or applying herbicides throughout different climatic and geographic regions in when they are likely to harm the weed’s natural enemies. the invaded range, are genetically variable, and have sev- This method is better developed for use against insect pests, eral modes of reproduction. such as the attraction of syrphid fl ies (hover fl ies) using certain fl owering plants. Syrphid larvae are predators on HISTORY aphids, and some control of aphids is achieved by the inter- The fi rst recorded use of a natural enemy to control a plant planting of Alyssum as a nectar subsidy among row crops population was the introduction of a cochineal insect to con- such as lettuce. The use of conservation biological control trol the prickly pear cactus, Opuntia vulgaris, in Sri Lanka needs to be more fully explored for control of weeds. in 1863. Carmine red dye was produced using the cochineal insect, Dactylopius coccus, which can be grown on several spe- CLASSICAL BIOLOGICAL CONTROL cies of Opuntia cactus. Both the cactus and insect are native OF PLANTS to the tropical regions of North and South America and The objective in a classical biological control program were introduced in the early 1800s to India and Australia for is to locate host-specifi c natural enemies from a plant’s commercial production of carmine dye. Some cactus spe- area of origin and introduce them into the plant’s invaded cies became troublesome weeds, and it was discovered that range. Those natural enemies identifi ed for use in a plant’s another cochineal insect, Dactylopius ceylonicus, which had invaded range are called biological control agents. All been mistakenly introduced as D. coccus into southern India, potential control agents are tested for safety prior to intro- appeared to suppress populations of Opuntia vulgaris, one of duction, and a permit for their introduction is required the cactus species that had become a weed. In 1863, prickly from the appropriate regulating agency of the recipient pear pads infested with D. ceylonicus were transported from country (see box). The goal is to establish permanent, India to Sri Lanka where it established and substantially self-sustaining populations of the biological control agent reduced the weedy cactus populations. that increase to critical population levels and reduce the The fi rst classical biological control program in which abundance or impact of the target plant. Control of a tar- a weed’s area of origin was explored for natural enemies get weed can result from death of the attacked plant, but was the program against Lantana camara in Hawaii in this is unusual. More commonly, control is achieved by 1902. This woody shrub is a native of Mexico and was the cumulative stress from nonlethal impacts that reduce introduced into Hawaii as an ornamental. It built up

FIGURE 1 (A) Adult Galerucella calmariensis, a leaf beetle released as a biological control agent on purple loosestrife in northern California. (B) Purple loosestrife plants damaged from feeding by the leaf beetle. (Photographs courtesy of Baldo Villegas, California Department of Food and Agriculture.)

BIOLOGICAL CONTROL, OF PLANTS 65

002_Simberloff10_B_p43-91.indd2_Simberloff10_B_p43-91.indd 6565 99/10/10/10/10 4:09:504:09:50 PMPM dense populations in a wide variety of habitats, displac- In the United States, the fi rst major weed biologi- ing native vegetation. In 1902, the entomologist Albert cal control program was against Klamath weed (Hyperi- Koebele, who was previously involved in the introduction cum perforatum, St. John’s wort) in California. Klamath of natural enemies for the control of the cottony cushion weed is a European herbaceous perennial that became scale in California, searched the native range of Lantana an invasive weed in Australia, South Africa, Chile, New in Mexico for natural enemies. His exploration efforts Zealand, Hawaii, and western North America. It is toxic resulted in shipping 23 insect species to Hawaii, of which to livestock, and invaded rangelands and pastures could 14 were released and eight became established. These no longer be used for cattle and dairy production. Previ- introductions resulted in successful control of Lantana in ously, scientists had introduced six insects into Australia, the drier lowland habitats of Hawaii; however, Lantana some of which proved successful in controlling Klamath continues to be a problem in the wetter, upland habitats. weed. Bolstered by the success in Australia, scientists in The next major program was against a complex of California introduced four of the six insects between Opuntia cacti that were rapidly invading rangelands in 1940 and 1950. The two most important insects were Australia. In 1912, two Australian scientists explored some the chrysomelid beetles Chrysolina quadrigemina and of the native range of prickly pear cacti in South America C. hyperici. Larvae of these beetles burrow through the and introduced fi ve insects into Australia between 1913 stems, and the adults feed on the foliage. Following their and 1914. The cochineal insect, D. ceylonicus, successfully introduction, Klamath weed populations were reduced controlled O. vulgaris, but other species of Opuntia con- to less than 5 percent of their former abundance. Most tinued to spread, especially O. stricta and O. inermis. In local landowners at the time were involved in dairy 1920, a renewed effort was initiated with scientists search- production and credit the Chrysolina beetles with hav- ing Mexico, the southwestern United States, and Argen- ing saved their livelihood. In gratitude, they erected a tina for new natural enemies. The efforts resulted in the monument to the beetles that sits in front of the Agri- discovery of 48 different species of insects, 12 of which were cultural Commissioners Offi ce in Humboldt County in released and established in Australia. The most important northern California. of these was the moth, Cactoblastis cactorum, which was introduced from Argentina in 1925. At the time, it was estimated that over 60 million acres were infested with HOST SPECIFICITY TESTING prickly pear cactus. By 1933, less than ten years follow- The risks associated with attack on nontarget plant spe- ing its introduction, all of the large stands of cactus had cies are reduced by selecting organisms with high host been destroyed (Fig. 2). Control of prickly pear cactus by specifi city—that is, those that damage only the target C. cactorum in Australia continues today. plant. Many insects and pathogens have long, coevolved

FIGURE 2 Before (A) and after (B) photographs of Opuntia inermis following release of Cactoblastis moths at a location in Queensland, Australia. The “before” photograph was taken in October 1926; the “after” photograph was taken in October 1929. (From Dodd, A. P. 1940. The Biological Campaign against Prickly-Pear. Brisbane, Australia: Commonwealth Prickly Pear Board.)

66 BIOLOGICAL CONTROL, OF PLANTS

002_Simberloff10_B_p43-91.indd2_Simberloff10_B_p43-91.indd 6666 99/10/10/10/10 4:09:514:09:51 PMPM associations with their hosts and have developed high TABLE 1 Weed Species under Complete Biological levels of host specifi city. They search out and reproduce Control in All or Part of Invaded Range on only a few plant species and have proven to be safe when introduced into new areas. Acacia saligna Alternanthera philoxeroides Prior to introduction, potential biological control Carduus nutans agents are subjected to a series of tests to examine their Centaurea diffusa host specifi city. Upon establishment in a new habitat, Chondrilla juncea Chromoleana odorata a biological control agent will be exposed to hundreds Cordia curassavica of new plant species. It is not possible to test all plant Eichhornia crassipes species, and so methods have been developed to select Euphorbia esula Hypericum perforatum a sample of plant species to best predict the host range Lythrum salicaria of a potential biological control agent once it is released Mimosa invisa in the new region. A plant list is created, consisting of Opuntia spp. related genera, crop plants, and native species. Observa- Pistia stratiotes Salvinia molesta tions of control agents released in past programs have Senecio jacobaea shown that those plants most closely related to the target Sesbania punicea weed are the most vulnerable to damage, and emphasis Sida acuta Tribulus terrestris is placed on testing representatives of these species. This Xanthium occidentale method of constructing a test list is called the phyloge- netic method. Several types of feeding and oviposition tests are per- formed using all of the plant species listed for testing. habitats, clogging water fl ow in irrigation canals, and If possible, it is important to reject a potential agent as modifying ecosystem services (Table 1). early in the process as possible, to reduce costs and lost time. As a result, the most conservative tests (no-choice TYPES OF ORGANISMS USED FOR CLASSICAL tests) are usually performed fi rst. No-choice tests con- BIOLOGICAL CONTROL OF PLANTS sist of enclosing an organism with a test plant, where Worldwide, there have been over 1,100 releases of bio- it must either use the plant or die. Those plant species logical control agents against 365 weed species in 75 coun- that are not fed on or in other ways damaged are con- tries. Most releases have occurred in fi ve countries: the sidered to be unusable by the organism and are removed United States, Australia, South Africa, Canada, and New from further testing. The remaining plants are exposed Zealand. Insects are the most common type of organ- to the potential control agent in a series of choice tests. ism used, accounting for 98 percent of the species used. In a choice test, the control agent can choose on which The other organisms consist of mites, nematodes, and plants to feed, deposit eggs, or infect. Choice tests are plant pathogens. As a group, beetles (Coleoptera), moths performed in cages in a quarantine laboratory or in out- (Lepidoptera), true bugs (Hemiptera), and fl ies (Diptera) door gardens in the native range of the target weed. are the most commonly used organisms, making up Upon completion, the results of all tests are summa- 65 percent of all species released. rized and submitted to the appropriate governmental regulatory agency as a petition requesting permission to fi eld release the organism. BENEFITS OF CLASSICAL BIOLOGICAL The costs of developing a classical biological con- CONTROL OF WEEDS trol program are very high and can exceed $1 million Classical biological control is an attractive control and take fi ve or more years per agent. The highest method because, when successful, control is permanent costs occur during the exploration and safety testing and requires little human input thereafter. Ongoing of potential biological control agents. Because of the expenditures for pesticides, labor, and specialized equip- high development costs, classical biological control is ment are signifi cantly reduced or removed altogether, sav- usually directed at weeds that infest large regions and ing enormous amounts of time and money. Over time, produce signifi cant negative impacts such as reducing these cost savings accrue and can become substantial. For forage plants in rangelands, poisoning livestock, dis- example, Klamath weed in the western United States has placing native plant and animal species by dominating been successfully controlled since the 1940s following

BIOLOGICAL CONTROL, OF PLANTS 67

02_Simberloff10_B_p43-91.indd 67 10/19/10 9:52:34 AM introduction of four exotic insects. The savings from not of ultimate success, uncertain food web effects, and using herbicides or other control efforts was estimated at the inability to remove a control agent once it has been $5 million annually. Economic analysis of the biocontrol released. Classical biological control targets only one of leafy spurge, a noxious weed of rangeland in the north species and is not effi cacious in habitats (e.g., cultivated central United States, by introduction of several species crops) where many weed species need to be controlled. of exotic Aphthona fl ea beetles, estimated the cost–benefi t The introduction of a new exotic organism has the ratio to exceed 150:1, despite the high prerelease costs for potential to cause both direct and indirect nontarget the program’s development. impacts. Direct impacts consist of feeding by a control Environmental benefi ts of classical biological control agent on nontarget native, economic, or other desir- can include the reduction of pesticide use, an increase able plant species. The risk of direct nontarget damage of biodiversity, a reduction of the occurrence of fi re, and is reduced through the host-specifi city tests that are per- an increase in ecosystem services, such as increased water formed prior to their introduction. Predictions of the fl ow in irrigation canals and access to fi shing in lakes and realized host range from these tests have been mostly rivers. For some weeds, the immense size and spread of accurate. Reviews of past weed biological control pro- weed infestations preclude the use of chemical or other grams report that, of nearly 400 species of control agents traditional control methods as both logistically and eco- released worldwide, 12 (3%) have been recorded attack- nomically impossible. In such cases, classical biological ing nontarget plants. Of these, most are transitory, short- control is especially valuable. term episodes of exploratory feeding on nearby plants Evaluation of success rates in classical biological con- that sometimes occurs when an outbreak population of a trol programs worldwide has been largely subjective. For biocontrol agent has destroyed its local host population. those programs where the target weed is reduced to a However, there are two examples of potentially fraction of its former abundance, such as occurred with significant effects on nontarget plant species from Klamath weed in the western United States and Opuntia biological control agents: the thistle seed head wee- cactus in Australia, the level of success is obvious. How- vil, Rhinocyllus conicus, on native Cirsium spp. in the ever, for many programs, reduction of the target weed is United States, and the cactus moth, Cactoblastis cacto- not dramatic, and other control methods are still needed. rum, on native Opuntia cacti in Florida and Mexico. Still, a reduction in control effort, such as using less her- The seed head weevil, R. conicus, was introduced from bicide or less frequent applications, may be considered a Europe to the United States during 1968 and 1969 partial success. When programs identifi ed as having had for control of musk thistle (Carduus nutans, C. tho- complete or partial success are combined, the success rate ermeri). While there are no native North American can be very high (Table 2), exceeding 80 percent in Aus- Carduus species, there are two native genera (Cirsium tralia, South Africa, and New Zealand. and Sassaurea) in the same tribe Cardueae. Field host records in Europe and results from the prerelease host- DISADVANTAGES AND RISKS OF CLASSICAL specificity testing suggested that four genera (Card- BIOLOGICAL CONTROL uus, Cirsium, Sylibum, and Onopordum), all within There are several disadvantages in a classical biological the tribe Cardueae, can be used as hosts by R. conicus. control project: high development costs, risks of dam- Because none of the ornamental or agricultural crop age to nontarget plant species, the lack of a guarantee species was damaged in the host-specificity studies,

TABLE 2 Successful Weed Biological Control Programs

Location Total Weed Targets Complete Success Partial Success Proportion of Total (%)

South Africa 23 6 13 83 Hawaii 21 7 3 48 Australia completed 15 12 0 80 ongoing 21 4 3 33 New Zealand 6 1 4 83

note: Reported number of weed biological control programs that have achieved complete (no other control options needed) or partial (reduced weed densities, some control options needed) success.

68 BIOLOGICAL CONTROL, OF PLANTS

002_Simberloff10_B_p43-91.indd2_Simberloff10_B_p43-91.indd 6868 99/10/10/10/10 4:09:534:09:53 PMPM and because many thistles (both native and exotic) abundance of generalist predators can lead to increased were considered at the time to be weeds, release of this predation on desirable native species unrelated to the insect was approved. The target weed was successfully target weed system. For example, the fruit fl y, Mesoclanis controlled at many locations; however, attack by R. polana was introduced in 1996 as a seed predator of conicus has been reported on over 20 native Cirsium Chrysanthemoides monilifera (bitou bush), a perennial species. The impact of this weevil on one native this- shrub that invaded the bush country of Australia. After tle, Cirsium canescens, has been studied in detail, and introduction, M. polana developed moderate popula- results showed that populations of this native thistle tion levels, but seed destruction was too low to cause are seed-limited and that population densities are a decline in the host plant’s abundance. A few years reduced following attack by this weevil. after initial release, fi eld observations found high rates What has been learned from this program is that the of parasitism of M. polana by native parasitoid wasps, error in the decision to release R. conicus was not due which normally attack several species of native insect to a misjudgment of its predicted host range but to the seed predators. Recently, fi eld observations have shown societal view of the time that native species need not that species richness and abundance of the local native receive the same level of protection as agricultural crops seed predators declined where the introduced biologi- and ornamental species. Social views on biodiversity have cal control agent increased. Ineffective biological con- changed since the 1960s, and native plants are now seen as trol agents that remain abundant in the community are a valued resource worth protecting. Consistent with this most likely to have persistent, indirect negative effects. change, both Canada and the United States have revoked However, eventual achievement of successful control all permits to collect and distribute R. conicus within their of the target weed will reduce populations of ineffec- country’s borders. tive biological control agents, thereby reducing indirect The second example of direct attack of a weed bio- nontarget impacts to acceptable levels. logical control agent on a new host is the cactus moth Biological control practitioners have developed an (C. cactorum) on native Opuntia cacti in southern Florida. International Code of Best Practices, which provides In 1957 through 1960, the cactus moth was introduced a set of guidelines for individuals engaged in the bio- into several islands in the Caribbean to control native logical control of invasive weeds. The code consists of Opuntia cacti that had infested pastures due to overgraz- 12 guidelines (Table 3) and covers all aspects of classical ing. Several decades later, the moth accidentally spread biological control, including prerelease and postrelease into southern Florida, where it has been observed attack- activities. The code attempts to incorporate the lessons ing native Opuntia cacti, including the endangered cactus learned from past projects and to help identify actions Consolea (formerly Opuntia) corallicola. An even greater that reduce risk and enhance effectiveness. By follow- threat is the potential for attack on the much larger ing this code, it is hoped that the practice of biological Opuntia fl ora of Mexico. control will continue to improve and remain a viable The decision to release the cactus moth in the control option for invasive weeds in the future. Caribbean was primarily economical, not ecological. Its release in Australia was ecologically sound, as there were no native cacti that were vulnerable to attack. In TABLE 3 The International Code of Best Practices for Classical Biological the Caribbean, the cactus moth was released into an Control of Invasive Weeds ecological region that has a high diversity of Opuntia 1. Ensure that the target weed’s potential impact justifi es the risk of cacti, and, once it was established, was able to spread. releasing non-endemic agents The lesson learned from this project is that the degree 2. Obtain multi-agency approval for the target weed of host specifi city required of a biological control 3. Select agents with potential to control the target weed 4. Release safe and approved agents agent is dependent on the ecology of the region of 5. Ensure that only the intended agent is released introduction. 6. Use appropriate protocols for release and documentation Indirect nontarget impacts can occur through 7. Monitor impact on the target weed changes in the food web. If a biocontrol agent builds 8. Stop releases of ineffective agents, or when control is achieved 9. Monitor impacts on potential nontarget species up high populations on a target weed but fails to con- 10. Encourage assessment of changes in plant and animal trol its host, then the biocontrol agent becomes an communities abundant new resource that can be exploited by gen- 11. Monitor interactions among biological control agents 12. Communicate results to the public eralist predators in the community. As a result, higher

BIOLOGICAL CONTROL, OF PLANTS 69

002_Simberloff10_B_p43-91.indd2_Simberloff10_B_p43-91.indd 6969 99/10/10/10/10 4:09:534:09:53 PMPM SEE ALSO THE FOLLOWING ARTICLES about 3 percent of Earth’s ice-free land area. Just over half Biological Control, of Animals / Enemy Release Hypothesis / of global bird introductions have been to Pacifi c islands Freshwater Plants and Seaweeds / Herbicides / Lantana camara / (notably to Hawaii) and to Australasia. Bird introductions Pathogens, Plant / Weeds do not always result in an established population. A case in point is the introduction of more than half a million FURTHER READING common quail (Coturnis coturnix) to more than 30 states Carvalheiro, L. G., Y. M. Buckley, R. Ventim, S. V. Fowler, and in North America between 1875 and 1958. Incredibly, this J. Memmott. 2008. Apparent competition can compromise the safety of highly specifi c biocontrol agents. Ecology Letters 11: 690–700. experiment failed. In the case of Australasia, of the 242 Coombs, E. M., J. K. Clark, G. L. Piper, and A. F. Cofrancesco Jr. 2002. bird species introduced in the past two centuries, 32 per- Biological Control of Invasive Plants in the United States. Corvallis: cent established viable populations. Oregon State University Press. Denoth, M., L. Frid, and J. H. Meyers. 2002. Multiple agents in biologi- BIRD INVADERS cal control: Improving the odds? Biological Control 24: 20–30. Hoffman, J. H. 1996. Biological control of weeds: The way forward, a Most bird introductions took place in the eighteenth and South African perspective (77–89). In C. H. Stirton, ed. Weeds in a Changing World International Symposium, November 20, 1995, nineteenth centuries, during a period of major European Brighton, England. British Crop Protection Council Monograph no. colonization. Because of this, a large proportion of inva- 64. Farnham: British Crop Protection Council. sive (alien, exotic, or introduced) bird species originated Julien, M. H. 1997. Biological Control of Weeds: Theory and Practical Appli- from temperate regions. For instance, 60 percent of intro- cation. Canberra: Australian Centre for International Agricultural Research. duced birds in New Zealand originated from the Palae- Julien, M. H., and M. W. Griffi ths. 1998. Biological Control of Weeds: A arctic and Australasian regions. Birds were introduced by World Catalogue of Agents and their Target Weeds, 4th edition. Walling- settlers to various countries, predominantly for aesthetics, ford: CABI Publishing. McFadyen, R. E. C. 1998. Biological control of weeds. Annual Review of hunting, and biocontrol. Two-thirds of the species cho- Entomology 43: 369–393. sen for introduction have been from 6 (of 145) bird fami- Rose, K. E., S. M. Louda, and M. Rees. 2005. Demographic and evolu- lies: Anatidae (ducks), Columbidae (pigeons and doves), tionary impacts of native and invasive herbivores on Cirsium canescens. Ecology 86: 453–465. Fringillidae (fi nches), Passeridae (sparrows), Phasianidae Van Driesche, R., M. Hoddle, and T. Center. 2008. Control of Pests (pheasants), and Psittacidae (parrots). This overrepresen- and Weeds by Natural Enemies: An Introduction to Biological Control. tation of certain families points to the reasons for these Oxford: Blackwell Publishing. introductions. For example, ducks and pheasants were introduced for hunting, while fi nches, sparrows, and par- rots were introduced as pets. Three bird species included BIOSECURITY among the 100 worst invasive species (see Appendix) are SEE ECOTERRORISM AND BIOSECURITY the European starling (Sturnus vulgaris), red-vented bul- bul (Pycnonotus cafer), and common myna (Acridotheres tristis) (Fig. 1). The negative effects of invasive bird species on native BIOTIC RESISTANCE HYPOTHESIS biodiversity, ecosystems, and humans have been widely SEE INVASIBILITY recognized and are briefl y discussed below. Most long- distance bird introductions to new areas are the direct or indirect result of human activities, and social and eco- nomic factors are often as critical as biological factors in BIRDS the introduction of exotic species. Activities such as log- ging and grazing further enhance establishment of exotics NAVJOT S. SODHI by creating optimal habitat for colonization (e.g., through National University of Singapore range expansion). Agriculture can facilitate bird species invasions when pests in agroecosystems are exposed to Humans are responsible not only for creating a condu- agricultural practices for many generations, resulting in cive environment for invasive species but also for intro- selection for characteristics that make them persist. It ducing many of these species themselves. Worldwide, is possible that bird invasions will increase in the future 1,400 attempts to introduce 400 bird species have been owing to increase in human trade and traffi c, and “global recorded. Approximately 70 percent of bird introduc- warming” may further facilitate species movements to tions have been to islands, although islands make up only new locations.

70 BIRDS

002_Simberloff10_B_p43-91.indd2_Simberloff10_B_p43-91.indd 7070 99/10/10/10/10 4:09:534:09:53 PMPM