How risky is biological control?
- Chinese proverb
An increasingly contentious debate has polarized the pest-control community about potential harmful effects of non-indigenous species introduced for biological control purposes. Howarth (1983, 1985, 1991, Gagne and Howarth 1985) first argued that such introductions had probably already produced extinctions of non-target species, and that existing protocols for assessing likely outcomes of such introductions were woefully inadequate. This view has attracted support from several workers who cite specific cases of insufficient consideration of potential impacts, likely problems from introduced control agents, or both (e.g., Ehler 1991, Simberloff 1992, Lockwood 1993a, b). It has also attracted vigorous rebuttals from biological-control practitioners who argue that the specific problem cases are unproven, that the general problem is greatly overstated, that there is no inherent reason why biological control is risky, and that existing protocols go a long way towards minimizing the already-low probability of unforeseen potential debacles (e.g., Funasaki et al. 1988, Lai 1988, Gonzalez and Gilstrap 1992, Carruthers and Onsager 1993). Many of these biological-control advocates concede that early projects, particularly using generalized vertebrate predators, caused unintended damage, but say that such events are far removed from the world of biological control today. DeLoach (1991) goes so far as to assert that current procedures in the United States completely prevent traditional biological-control projects from affecting non-target species.
Probably a fraction of the heated reaction is visceral, and stems from a sense almost of betrayal. After all, biological control has been advanced for many years as a green alternative to chemical control, and the great majority of its practitioners surely entered the field as idealists, seeking to stem environmental destruction (Center 1995). To then be tarred with the same brush as the pesticide "nozzleheads" must be a cruel blow. Nevertheless, in the light of increasing knowledge of the scope of the environmental damage caused by non-indigenous species (e.g., U.S. Congress OTA 1993), the arguments of both critics and defenders of biological control deserve careful consideration.
PREDATION, PARASITISM, AND HERBIVORY OF NON-TARGET SPECIES
Many generalized predators released for biological control have preyed on non-target species (Simberloff 1992). For example, the small Indian mongoose (Herpestes auropunctatus), introduced to the West Indies, Hawaiian islands, Mauritius, and Fiji to control rats in agricultural fields, has contributed to the decline of native birds in all those areas (Cheke 1987; references in Lever 1985, Simberloff 1992). It has probably eliminated native reptile species in the West Indies (Honegger 1981). The predatory snail Euglandina rosea has been introduced from Florida and Central America to many islands worldwide to control the giant African snail Achatina fulica (Civeyrel and Simberloff 1996). At least in the Hawaiian and Tahitian archipelagoes, it has extinguished several endemic forest snails.
Generalized herbivores introduced for biological control of weeds have similarly had unintended effects on native species. Perhaps the most prominent introductions of this sort are of freshwater fishes (Moyle et al. 1986, Courtenay and Williams 1992, Courtenay 1993). The full effects of these fish introductions are unknown, but many such control agents have so greatly reduced the native vegetation that they have changed the composition of native fish communities. In fact, it appears that every fish introduction for biological control that has been thoroughly studied has had major detrimental effects on non-target organisms, whatever its impact on the target species.
Although it has been asserted that phytophagous insects introduced for terrestrial weed control have never caused the elimination of non-target native species (e.g., Groves 1989, DeLoach 1991), examples are known in which such non-target plants have been so reduced as to cast doubt on the assertion. For example, the semaphore cactus (Opuntia spinosissima) was represented in the United States by only a few plants in the Torch Wood Hammock Preserve of the Nature Conservancy, in the Florida Keys. It is a candidate for listing under the Endangered Species Act; all other records are from the south coast of Jamaica, and there is reasonable doubt that the two populations are conspecific. The individuals in the preserve have been severely impacted by caterpillars of the moth Cactoblastis cactorum (Kass 1990, Robertson 1990). This moth was introduced in 1957 to Nevis in the Lesser Antilles by the Commonwealth Institute of Biological Control to control several Opuntia species (Simmonds and Bennett 1966), then by local authorities from Nevis to nearby islands (Tuduri et al. 1971). It then either island-hopped without human assistance through the West Indies (Simberloff 1992), or was carried inadvertently on ornamental cacti imported to Florida (Pemberton 1995). Either way, it reached the Florida Keys about 1989. Almost all individuals of O. spinosissima now are in botanical gardens; the few in nature are surrounded by wire cages, and, in the Keys, the moth also infests the rare jumping prickly pear, O. triacantha.
It seems clear from these examples that any predator or herbivore maintained at high densities on common alternative hosts can potentially drive a rare non-target species to extinction (cf. Howarth 1991). The density dependence that would vitiate such an effect in a simple two-species interaction is lacking in such cases. For example, the presence of the common alternative host Opuntia stricta in the Florida Keys must compound the threat posed by the cactus moth to the semaphore cactus (Simberloff 1992; D. M. Johnson and P. D. Stiling, unpublished manuscript).
Gagne and Howarth (1985) and Howarth (1985), following earlier suggestions by Zimmerman (1958, 1978), argued that parasitoids introduced to Hawaii for biological control might have been agents of extinction of several lepidopterans. Funasaki et al. (1988) object strenuously to this suggestion. Having analyzed all biological-control introductions into Hawaii between 1890 and 1985, they argue that almost all introduced parasitoids recorded from native moth species arrived in the Hawaiian islands on their own rather than by deliberate introduction. However, many introductions they view as autonomous may actually have been deliberate but not recorded (Howarth 1991). For example, Swezey (1931) noted that the normal practice of many early workers was to record only introductions that were observed to be successful. Further, there were probably a number of unrecorded free-lance biological-control introductions. Perhaps this particular argument would not be terribly germane to risks posed by current biological-control introductions if all parties agreed that current protocols are vastly safer than those of earlier workers, a contention we discuss below (see Would any protocols suffice?).
Funasaki et al. (1988) concede just one exception: the tachinid fly Lespesia archippivora, introduced to control introduced armyworms, attacked the native noctuid Agrotis crinigera, which is among the sixteen species that Gagne and Howarth had suspected of extinction by biological-control agents. The same fly was recorded attacking 10 other native Lepidoptera. Generally, Funasaki et al. (1988) see little reason for concern with current biological-control procedures, given the overall statistics and their assertion that present protocols are vastly better than their predecessors. Of 679 species that they recognize as deliberately introduced for biological control between 1890 and 1985, 243 established populations, and 20 of these are recorded as attacking non-target native species. None of these 20 was introduced after 1967. For insects alone, 533 species were introduced, of which 175 established populations and 15 are known to attack non-target native species.
Howarth (1991) lists several other examples, almost all from islands, of parasitoids introduced for biological control that either substantially reduced populations of non-target native species or eliminated them altogether. These examples seem to be viewed as highly speculative by advocates of biological control (e.g., Center 1995), though the basis of this skepticism has not been published, to our knowledge.
COMPETITION WITH NATIVE SPECIES
Interspecific resource competition, especially for food, is notoriously hard to document in the field even when it occurs; this is one reason why a persistent debate in ecology concerns the importance of this force. Ecologists working with vertebrates have had recourse to various patterns that might indirectly reflect competition (e.g., inverse population trends in time or space). Such data would be available for even fewer invertebrates than for vertebrates. Thus, even if introduced biological-control agents often competed with native species, it would not be surprising that the effects would not be documented. Nevertheless, Ehler (1991), Howarth (1991), and Bennett (1993) provide suggestive examples. One that has attracted much recent attention is the apparent replacement of many native American lady beetles by the European lady beetle Coccinella septempunctata, mass-released for control of Russian wheat aphid (Gordon and Vandenberg 1991, U.S. Congress OTA 1993, Elliott et al. 1996; P. Kareiva, personal communication). As noted below, the full effects of such a dramatic change in a numerous taxon on various community and ecosystem phenomena are unknown, but it would be surprising if no major impact occurred.
We have so far restricted our discussion to straightforward interactions among species, one on one. Introduced species can wreak havoc with native species in much more tortuous ways. Consider the ad hoc attempt to control rabbits in Great Britain by introducing the Myxoma virus (Ratcliffe 1979, Moore 1989). Caterpillars of the large blue butterfly, Maculina arion, must develop in underground nests of the ant Myrmica sabuleti. The ant, in turn, does not nest in overgrown areas. The rabbits were the main means of maintaining open habitats in the wake of changing land-use patterns and reduced grazing by livestock. The virus so reduced rabbit populations that ant numbers declined greatly, leading to the extinction of the butterfly.
COMMUNITY AND ECOSYSTEM EFFECTS
Perhaps one could have predicted that greatly reducing the English rabbit population would have had some major effects, possibly many major effects, even if one could not have predicted precisely what these impacts would be. This is because the rabbit, by maintaining the open habitat, fulfilled the classical definition of a keystone species (Paine 1966, 1969): a species whose grazing or predatory activity markedly changes the composition of a numerically dominant and physically structuring trophic level of a community, thus leading to a dramatic change in the physical structure itself. In fact, Harper (1969) long ago noted the keystone role of rabbits in Great Britain and how myxomatosis consequently led to great community changes, though he focussed on plant species. What is really remarkable about the extinction of the large blue (butterfly, Maculina arion) is not that it happened but that anyone noticed it. We return to this point below (see The problem of insufficient monitoring).
Simberloff (1991) argued that the keystone-species concept might provide an entree into the morass of predicting major environmental impacts of introduced species; a non-native species that destroys or constitutes a classical keystone species or that itself becomes a dominant structural element in the habitat might be expected to have a huge impact, though the particulars of that impact might not be easily predicted. Lately the concept of a keystone species has been attacked (Mills et al. 1993) on the grounds that the definition has been so expanded as to encompass just about any species that someone cares about, and thus has no real utility in indicating which species have some special, disproportionate influence on the entire community. Further, there is often no experimental verification of this influence. However, this argument seems an overreaction (cf. DeMaynadier and Hunter 1994). That some researchers have misused the term does not mean it is devoid of content; the experiments reported by Paine (1966, 1969) and Harper (1969) met the strict criteria and truly elucidated the key roles of particular species in the determination of community composition and physical structure.
The frequent absence of evidence that a putative keystone species plays this role is, of course, a serious problem, but it is simply part of a problem that pervades all community- and ecosystem-level ecology - controlled experiment is difficult and often not attempted. Still, this problem does not render community and ecosystem ecology less "scientific" or mean that events and observations at these levels are either unimportant or uninstructive (Shrader-Frechette and McCoy 1993). Rather, it dictates caution in interpreting and extrapolating results. For the rabbit in Great Britain, both experiments and carefully planned observations (Harper 1969) verified its keystone role. For other species in other systems the designation must be tentative. However, at the very least one would expect that an introduced species that eliminates or greatly modifies an entire large taxon might play this role; the Australian brown tree snake in Guam, the Nile perch in Lake Victoria, and certain ant species in various locations all come to mind (Simberloff 1991).
A controversial program mounted by the United States Department of Agriculture to introduce non-indigenous species to control native rangeland grasshoppers in the West (Carruthers and Onsager 1993, Lockwood 1993a, b) may fall in this category. The plan was to release an Australian scelionid wasp (Scelio parvicornis) and an Australian fungal pathogen (Entomophaga praxibuli). Upon objection that ecological costs had not been adequately assessed, the wasp release was postponed, although the fungus had already been released in two states. Among the 300-odd grasshoppers of western rangelands, there are no introduced species (Lockwood 1993a). Most of these grasshoppers in most years in most places are not harmful to human interests, and some are believed to be ecologically beneficial. But there has been no thorough study of all their exact ecological roles, as with those of so many other insects. They may suppress various weeds. For example, one non-target species, Hesperotettix viridis, feeds on plants poisonous to livestock, which might well increase in their absence. Grasshoppers surely compete with other pest insects and almost certainly contribute to determining the composition and therefore physical structure of plant communities. They vector symbionts such as mycorrhizal fungi, they vector microbial pathogens, and, by sheer numbers, they must contribute greatly to nutrient cycling. What would happen to any of these processes if any non-target or even target species disappeared or greatly declined is simply not predictable given the current state of knowledge of community and ecosystem ecology of grasslands. But it seems very likely that something important would happen, even if we cannot predict what it would be. What is more doubtful is whether we would ever know what happened, and why.
THE "NEW ASSOCIATIONS" AND "NEOCLASSICAL CONTROL" CONTROVERSY
The grasshopper case is particularly controversial because it entails importing non-native species to control native ones, whereas "classical biological control" is typically defined (e.g., Nechols and Kauffman 1992) as controlling non-native pests by importing species, often antagonists from the pests' regions of origin. Lockwood (1993a) suggests the term "neoclassical biological control" when an introduced antagonist of a native species is used. This approach automatically raises warning flags because it is not based, as is much classical biological control, on a close coevolutionary association between pest and enemy. The presence of a wild card seems much more likely than in classical biological control. After all, if the pest and enemy have never seen one another before, is the probability not increased that non-target species will be affected at least as much as the target will?
A guiding principle of neoclassical control is the "new association" hypothesis of Hokkanen and Pimentel (1984, 1989). In essence, the hypothesis is that totally new associations are likely to be more devastating to the pest precisely because the pest has not coevolved with its enemies. One can understand the rationale by analogy to the evolution of benignity in disease-host relationships (e.g., Ewald 1983). Often the first wave of an introduced epizootic disease devastates the host, but subsequent coevolution, including evolution of disease resistance by the host species and of less virulent genotypes of the pathogen, makes successive waves progressively less catastrophic. Because it is not in the pathogen's "interest" to eliminate its host, natural selection inexorably leads to a more benign relationship. The trajectory of the myxoma virus-rabbit interaction in Australia is a good example from biological control (Williamson 1992).
Hokkannen and Pimentel (1984, 1989) marshal literature data to show that many, perhaps most, truly effective biological-control programs involve new associations of host and enemy, and often they entail new-host use by species previously considered monophagous or oligophagous. This claim has been criticized, but the grounds of the criticism are essentially haggling over the exact numbers, not over the fact that new associations are often effective for biological control. For example, Waage and Greathead (1988) reanalyzed a subset of the data analyzed by Hokkanen and Pimentel, using somewhat different criteria for project "success," and concluded that, rather than a typically better result with new associations, there is no difference between new and old associations in likelihood of producing economically important control.
With respect to untoward consequences of a biological-control introduction, the very species that are most likely to produce effective neoclassical biological control are those preadapted to use new hosts, and these are therefore the greatest threats to non-target species (Roderick 1992). Of course, the circumstances of many biological-control releases - small propagules isolated from conspecific populations in new environments - facilitate rapid evolution (Roderick 1992), which might be expected to lessen effectiveness of control for the reasons just stated, but might also lessen effects on non-target species initially attacked.
A common response to concerns about unintended effects of biological-control introductions is that, whatever the ecological costs of a particular project, the costs of not doing the project may well be greater (e.g., Harris 1990, Nechols et al. 1992, Center 1995). For example, the imported fire ant Solenopsis invicta has, in 50 yr, spread over much of the Deep South, with numerous publicized agricultural and medical effects (Tschinkel 1993). Lately it has been recognized as a conservation concern, with the spectre of still-greater damage looming as the polygynous form spreads (Porter et al. 1988, Porter and Savignano 1990). For example, in Florida it has replaced the native fire ant S. geminata in many habitats; in Texas it is believed to have caused a great decline in horned toads (Phrynosoma) through predation and displacement of harvester ants. Much of the native ant fauna seems threatened (Buren 1983), and it appears evident that there have been numerous ecosystem-level effects, though we know of no specific study of these. Chemical control has failed, leading to numerous proposals for biological control (Jouvenaz 1990). Faced with the high probability of ecological damage if biological control is not attempted, what degree of likely ecological damage from a prospective control agent should be tolerated?
In Florida, the Australian tree Melaleuca quinque-nervia is now dominating [approximately equal to]20 new hectares daily in spite of intensive manual-control efforts. They transpire as much as four times as much water as the saw-grass community they are replacing in many areas, and have replaced entire native plant communities (Schmitz 1994). Could potential damage by an introduced biological-control agent be worse than this? One could make a similar argument in Florida alone for Brazilian pepper (Schinus terebinthifolius) and Australian pine (Casuarina spp.). In each instance, extensive tracts of native vegetation have already been replaced and current control methods are not stemming the spread.
Even those most concerned about potential dangers of biological control concede the force of this argument (e.g., Howarth 1991, Simberloff 1992). Assessing likely costs and benefits of the two courses of action, however, is a very difficult matter, and the typical cost-benefit analysis for biological control seems one-sided, with costs of pesticides, costs of collecting trips, costs of host-specificity tests, etc., fairly well detailed, but costs of the loss of isolated populations or even entire species, costs of disruption of community and ecosystem features, etc., barely considered. A part of the problem is that the former set of costs can easily be transformed to dollars, while the latter set of costs cannot. Philosophers, economists, and biologists have all treated the difficulty of valuing natural entities (e.g., Norton 1987, Rolston 1991), but nothing remotely approaching a consensus has emerged. However, the fact that it is difficult to value these things in dollars does not mean that their value is not immense. The fact that Cactoblastis has worked well on some West Indian islands to control pest Opuntia (Simmonds and Bennett 1966) is important, but so would the loss of native Opuntia species in Florida be important. Further, Cactoblastis may have already invaded the Yucatan from Cuba (R. Pemberton, personal communication) and could reach the Opuntia-rich southwestern United States (D. M. Johnson and P. D. Stiling, unpublished manuscript). At the least, these facts mandate a much more rigorous analysis of the real threats posed even by introduced species universally conceded to be horrors.
For example, the imported fire ant has certainly displaced the native fire ant in Florida from some habitats, but these are predominantly disturbed habitats (Tschinkel 1988, McInnes 1994). What ants are truly threatened by the imported fire ants, and in what habitats? No one disputes that Melaleuca replaces native plants, but Melaleuca in Florida has also been described as primarily invading disturbed areas (Ewel et al. 1976, Ewel 1986) and wetlands in which the hydroperiod has been shortened by human activities (Hofstetter 1991). This plant also invades less disturbed areas (Myers 1983, Ewel 1986), but what is really needed, and appears not to have been published, is a full accounting of which species have been eliminated or are threatened over which areas, and what other ecological impacts have occurred or are likely to occur. This is a tall order, but without a thorough attempt at such a synthesis, no cost-benefit analysis is really possible.
The probabilities of an introduced biological-control agent going awry, and the likely costs if it does, are other aspects of cost-benefit analyses that need at least to be more explicit, even if they would have enormous confidence limits.
THE PROBLEM OF INSUFFICIENT MONITORING
Numerous authors (e.g., Funasaki et al. 1988, Lai 1988, Center 1995) point to the small number of known disastrous consequences of biological-control introductions relative to the many projects as evidence that the enterprise is generally safe. This claim is not cogent for two reasons. First, when one ponders the remarkable sets of circumstances allowing the detection of some of these disasters, it becomes clear that the detected cases must be a minuscule fraction of those that have occurred (Simberloff 1992).
Consider, for example, the attack of the cactus moth on the semaphore cactus in the Florida Keys, discussed above (see Predation, parasitism, and herbivory). Because of their unusual biota, the Florida Keys are well-studied botanically; further, they are crawling with tropical-plant enthusiasts. Were this not so, the presence of the cactus on Little Torch Key might never even have been known. Even if it had been known, it is not likely that the attack of the moth would have been recognized by anyone were it not in a heavily managed site like a Nature Conservancy refuge. Further, it would almost certainly have disappeared before the attack was noticed, even in a heavily managed preserve, had not a botanist (C. Lippincott of the Fairchild Tropical Garden) been alerted by a biologist (W.T. Starmer) working on cactus-eating insects at Guantanamo Bay, who recognized the cactus moth as something not recorded from Cuba and wondered if it was also in Florida.
Very few taxa attract sufficient human interest to be observed more than cursorily by the lay public or even scientists. The extinction of the large blue (butterfly) in Great Britain would almost surely have been missed completely if it were a bug or grasshopper rather than a butterfly. Prickly pears and butterflies are highly visible, often beautiful, and collectable. Most species have none of these traits. Seen in this light, the complacency of Funasaki et al. (1988) about the dearth of known harmful effects of biological-control introductions in Hawaii seems unwarranted (Simberloff 1992). They do not describe the effort to find non-target hosts and to determine their mortality factors, or the study of community and ecosystem properties to see if they have been modified. Particularly in remote upland habitats, it is difficult to believe that the survey was at all adequate.
Many extinctions are quite mysterious, particularly the reasons why the last few individuals disappear (Simberloff 1994). The general problem is that insufficient data are gathered to implicate and to eliminate various potential causes. That a few extinctions can confidently be assigned to a biological-control agent, far from being grounds for comfort, should be cause for alarm.
There are numerous plausible hypotheses about the detrimental effect of non-indigenous biological-control agents on native species, but some of them can never be tested because no baseline data were gathered before the introductions, and all of them would require substantially more monitoring than presently occurs. For example, the European parasitic tachinid fly, Compsilura concinnata, was liberated in the United States from 1906 through 1927 in an attempt to control the gypsy moth, Lymantria dispar (Reardon 1981). It rapidly established and spread westward well beyond the range of the gypsy moth and was quickly reared from over 200 other species of Lepidoptera. Massive DDT spraying for gypsy moths in New Jersey in the late 1950s led to the local decline and even disappearance of many moths and butterflies (Brody 1995; D. Schweitzer, personal communication). Some of these, particularly saturniids, sphingids, and Datana species, have been surprisingly slow to recover after the cessation of DDT use. Parasitism by Compsilura is suspected as a major contributing factor (D. Schweitzer, personal communication), but the hypothesis cannot be tested. There was virtually no monitoring of these species before and during the DDT blitz.
A second reason why one might not be too comforted by the fact that relatively few untoward consequences of biological control are known is that, as unlikely as extinctions are to be observed, disruption of various community or ecosystem processes is even less likely to be observed. Except at Long-Term Ecological Research sites or other intensive research sites, who monitors nutrient cycles and flows, decomposition, vectoring of symbionts, etc.? Even intensive research projects rarely monitor more than a small fraction of ecosystem traits and processes.
WOULD ANY PROTOCOLS SUFFICE?
Given the recondite and sometimes tortuous pathways by which various ecological effects occur, are there any protocols for biological-control introductions that would prevent all disasters? Probably not; ecologists simply cannot predict the effects of introduced species well enough ever to be certain (Simberloff 1991). However, we believe that protocols for such introductions could be vastly improved. Most fundamentally, there has to be a perceptual shift from the view that most such introductions are not likely to cause trouble to the view that extensive research is required to justify a judgment that an introduction will probably be innocuous. In other words, guilty until proven innocent. This shift in modus operandi is needed generally in policies related to the environment (Shrader-Frechette and McCoy 1993), not only in biological control.
At the very least, there must be an end to ad hoc biological-control projects, particularly aimed at insect pests, by academics and business interests. Such unofficial actions are still common, in spite of disclaimers by irate biocontrol advocates that official modern biological-control safeguards are rigorous. No one person or informal group, no matter how knowledgeable he/she is or how patently unobjectionable the release is, should be able to mount such a project. Mail-order biological-control operations must be much more tightly regulated. Even if a species has been determined to be safe in one state, no one should be allowed to transport it to another state on an ad hoc basis.
There must also be more concern in government-sanctioned programs for host-specificity of entomophages and potential effects on non-target species. Just this year a Japanese coccinellid, Pseudoscymnus sp., was released in Connecticut for control of Adelges tsugae, the Japanese hemlock woolly adelgid (Grant 1995). The adelgid, which arrived in the U.S. [approximately equal to]40 years ago, is ravaging northeastern forests by killing many individuals of eastern hemlock (Tsuga canadensis), a highly dominant tree (McClure 1991), and chemical control has proven insufficient (M. McClure, personal communication). The speed with which this insect is eliminating hemlock induced the Biological Assessment and Taxonomic Support group of the U.S. Department of Agriculture to approve experimental release of the coccinellid with the assessment that it would have "no significant impact" on non-target species (Grant 1995). Given the observed effects of Coccinella septempunctata, described above (see Competition with native species), and how long it took to detect them (Schaefer et al. 1987), this assurance is a bit surprising. The Pseudoscymnus is believed to specialize on adelgids, although testing of this assertion was minimal, and little consideration was given to affects on native coccinellids, because of the perceived speed of the adelgid invasion (M. McClure, personal communication).
Also, the argument that further safeguards will be so costly as to hinder biological control greatly or to shut it down entirely has to be examined critically. It is true that safeguards may be expensive, just as they are for pharmaceutical testing in the United States. But that does not mean that they are unnecessary, and the precise costs should be made explicit. It is also true that some potentially very useful technologies carry such inherent and irreversible risks that they should not permitted.
Several fundamental difficulties with assessing risks from biological-control agents stem from the fact that they are alive. The first is that all living organisms have dispersal means, so that, if they survive at all, they will not stay in the habitat or even region where they are introduced. The island-hopping by the cactus moth from the Lesser Antilles at least partway to Florida exemplifies the problem. In this case, previously observed island-hopping by the same species in the Hawaiian islands should have suggested the possibility, especially since this project was initiated by the Commonwealth Institute of Biological Control. In addition to spreading to different regions, biological-control agents routinely spread from agricultural or human-influencial habitats to other ones, including pristine ones such as nature preserves (Simberloff 1992).
Dispersal ability complicates greatly the matter of safety testing for host-specificity. For example, if one aims to introduce a phytophage in one region, it does not suffice to test only potential host plants from that region. Three non-native mole cricket species are established in the southern United States; two are important pests of vegetable seedlings and pasture and turf grasses. Three biological-control agents (a digger wasp, a nematode, and a tachinid fly) were introduced in Florida in the 1980s (Frank 1994). Within Florida, the fly and wasp appear to be host-specific to the introduced mole crickets, although larvae can develop on at least two other cricket genera. The nematode attacks several other native species, but not native mole crickets. Not only were the costs to other species attacked in Florida not carefully determined, but no apparent consideration was given to the consequences of potential spread of the introduced control agents. A native North American mole cricket, Gryllotalpa major, is a candidate endangered species. What is the likelihood that these control agents would spread and, if they did, what is the probable effect on this native species?
Another pervasive problem engendered by the fact that biological-control agents are alive is that they evolve. Species do acquire new hosts (e.g., Bush 1975, Prokopy et al. 1988). Even a single gene mutation can change host-specificity (Williamson 1992), and such host range expansions have been observed in nature, as with the Australian gall wasp Trichilogaster acaciaelongifoliae (Dennill et al. 1993). Similarly, species can evolve an expanded tolerance of physical factors, such as soil contaminants (e.g., Walley et al. 1974), that might greatly increase the probability of important ecological effects. Likewise, changes in virulence of a pathogen can occur suddenly with great potential impact. For example, the fungus Entomophaga maimaiga was introduced from Japan to Massachusetts for gypsy moth control in 1910. It was not seen again until 1989, when it caused massive epizootics among gypsy moth populations in the Northeast. Evolution of a strain with increased aggressiveness is one of two viable hypotheses (Hajek et al. 1995). As was discussed above in another context (see The "new associations" and "neoclassical control" controversy), it is very possible that such evolutionary changes are common but generally unobserved. If a phenomenon similar to what happened to the gypsy moth occurred in an obscure non-target lepidopteran of no perceived economic significance, how could one have thought to associate the decline with an earlier introduction? In fact, would the decline or even disappearance have been noted? In any event, the potential evolution of an imported biological-control agent means that the temporal scope of potential ecological problems is much greater than is traditionally assumed, just as dispersal behavior means the spatial scope is larger.
Some of the published safety-testing protocols (e.g., that for the purple loosestrife project [Malecki et al. 1993]) seem quite stringent, but none appear to be adequate in the context of the potential spread and evolution of the proposed control agent. It is difficult to imagine what protocols would be. And safety-testing of entomophages seems even more problematic than that of phytophages. In many countries such testing is restricted to insects of known commercial value, i.e., to a minuscule minority. Given the poor knowledge of so many insects of non-human-influenced habitats, it is difficult to imagine a protocol that would ensure an adequate scan of regional potential hosts, even aside from the possibility of ecosystem impacts, dispersal out of the region, or evolution.
Of course, other courses of action in the face of an invasive non-indigenous species, such as chemical control and benign neglect, carry their own risks. These must be carefully considered, and potential costs weighed, even though forecasting the impacts of non-indigenous species is a very imprecise science. However, as we note above, biological control should be subjected to the same rigorous cost-benefit analysis and not automatically accorded the mantle of the environmentally friendly alternative.
We thank Peter Kareiva, Mark McClure, and Dale Schweitzer for unpublished information.
Bennett, F. D. 1993. Do introduced parasitoids displace native ones? Florida Entomologist 76:54-63.
Brody, J. E. 1995. "Aggressive invader from Asia increases the threat from gypsy moths." New York Times, National edition, 30 May 1995, Section B: pp. 5 (col. 1-5), 9 (col. 1-5).
Buren, W. F. 1983. Artificial faunal replacement for imported fire ant control. Florida Entomologist 66:93-100.
Bush, G. L. 1975. Sympatric speciation in phytophagous parasitic insects. Pages 187-206 in P. W. Price, editor. Evolutionary strategies of parasitic insects and mites. Plenum, New York, New York, USA.
Carruthers, R. I., and J. A. Onsager. 1993. Perspective on the use of exotic natural enemies for biological control of pest grasshoppers (Orthoptera: Acrididae). Environmental Entomology 22:885-903.
Center, T. D. 1995. Review of Selection criteria and ecological consequences of importing natural enemies, edited by W. C. Kauffman and J. R. Nechols. Biodiversity and Conservation 4:524-526.
Cheke, A. S. 1987. An ecological history of the Mascarene Islands, with particular reference to extinctions and introductions of land vertebrates. Pages 5-89 in A. W. Diamond, editor. Studies of Mascarene Island birds. Cambridge University Press, Cambridge, England.
Civeyrel, L., and D. Simberloff. 1996. A tale of two snails: is the cure worse than the disease? Biodiversity and Conservation 5, in press.
Courtenay, W. R., Jr. 1993. Biological pollution through fish introductions. Pages 35-61 in B. N. McKnight, editor. Biological pollution. The control and impact of invasive exotic species. Indiana Academy of Science, Indianapolis, Indiana, USA.
Courtenay, W. R., Jr., and J. D. Williams. 1992. Dispersal of exotic species from aquaculture sources, with emphasis on freshwater fishes. Pages 49-81 in A. Rosenfield and R. Mann, editors. Dispersal of living organisms into aquatic systems. Maryland Sea Grant College, College Park, Maryland, USA.
DeLoach, C. J. 1991. Past successes and current prospects in biological control of weeds in the United States and Canada. Natural Areas Journal 11:129-142.
DeMaynadier, P., and M. L. Hunter. 1994. "Keystone support." (Letter to the Editor). BioScience 44:2.
Dennill, G. G., D. Donnelly, and S. L. Chown. 1993. Expansion of host range of a biocontrol agent Trichilogaster acaciaelongifoliae (Pteromalidae) released against the weed Acacia longifolia in South Africa. Agriculture, Ecosystems & Environment 43:1-10.
Ehler, L. E. 1991. Planned introductions in biological control. Pages 21-39 in L. R. Ginzburg, editor. Assessing ecological risks of biotechnology. Butterworth-Heinemann, Boston, Massachusetts, USA.
Elliott, N. C., R. W. Kieckhefer, and W. C. Kauffman. 1996. Effects of an invading coccinellid on native coccinellids in an agricultural landscape. Oecologia 105:537-544.
Ewald, P. W. 1983. Host-parasite relations, vectors, and the evolution of disease severity. Annual Review of Ecology and Systematics 14:465-485.
Ewel, J. J. 1986. Invasibility: lessons from south Florida. Pages 214-230 in H. A. Mooney and J. A. Drake, editors. Ecology of biological invasions of North America and Hawaii. Springer-Verlag, New York, New York, USA.
Ewel, J. J., R. Meador, R. Myers, L. Conde, and B. Sedlik. 1976. Studies of vegetation changes in South Florida. Report to U.S. Forest Service on Research Agreement 18-492. University of Florida, Gainesville, Florida, USA.
Frank, J. H. 1994. Inoculative biological control of mole crickets. Pages 467-475 in A. R. Leslie, editor. Handbook of integrated pest management for turf and ornamentals. Lewis, Boca Raton, Florida, USA.
Funasaki, G. Y., P. Y. Lai, L. M. Nakahara, J. W. Beardsley, and A. K. Ota. 1988. A review of biological control introductions in Hawaii: 1890 to 1985. Proceedings of the Hawaiian Entomological Society 28:105-160.
Gagne, W. C., and F. G. Howarth. 1985. Conservation status of endemic Hawaiian Lepidoptera. Pages 74-84 in Proceedings of the 3rd Congress on European Lepidopterology, Cambridge 1982. Societas Europaea Lepidopterologica, Karlsruhe, Germany.
Gonzalez, D., and F. E. Gilstrap. 1992. Foreign exploration: assessing and prioritizing natural enemies and consequences of preintroduction studies. Pages 51-70 in W. C. Kauffman and J. E. Nechols, editors. Selection criteria and ecological consequences of importing natural enemies. Entomological Society of America, Lanham, Maryland, USA.
Gordon, R. D., and N. Vandenberg. 1991. Field guide to recently introduced species of Coccinellidae (Coleoptera) in North America, with a revised key to North American genera of Coccinellini. Proceedings of the Entomological Society of Washington 93:845-924.
Grant, S. 1995. "Alien beetle lands to fight hemlock scourge." Hartford Courant, Statewide edition, 13 May 1995: pp. 1 (col. 3-6), 6 (col. 1-4).
Groves, R. H. 1989. Ecological control of invasive terrestrial plants. Pages 437-461 in J. A. Drake, H. A. Mooney, F. diCastri, R. H. Groves, F. J. Kruger, M. Rejmanek, and M. Williamson, editors. Biological invasions: a global perspective. John Wiley & Sons, Chichester, England.
Hajek, A. E., R. A. Humber, and J. S. Elkinton. 1995. Mysterious origins of Entomophaga maimaiga in North America. American Entomologist 41:31-42.
Harper, J. L. 1969. The role of predation in vegetational diversity. Pages 48-62 in G. M. Woodwell and H. H. Smith, editors. Diversity and stability in ecological systems. Brookhaven Symposia in Biology, Number 22. Brookhaven National Laboratory, Upton, New York, USA.
Harris, P. 1990. Environmental impact of introduced biological control agents. Pages 289-300 in M. Mackauer, L. E. Ehler, and J. Roland, editors. Critical issues in biological control. Intercept, Andover, Hampshire, England.
Hofstetter, R. H. 1991. The current status of Melaleuca quinquenervia in southern Florida. Pages 159-176 in T. D. Center, R. F. Doren, R. L. Hofstetter, R. L. Myers, and L. D. Whiteaker, editors. Proceedings of the exotic pest plant symposium (2-4 November 1988). NPS/NREVER/NRTR-91/06 Technical Report. National Park Service, Denver, Colorado, USA.
Hokkanen, H., and D. Pimentel. 1984. New approach for selecting biological control agents. Canadian Entomologist 116:1109-1121.
Hokkanen, H., and D. Pimentel. 1989. New associations in biological control: theory and practice. Canadian Entomologist 121:829-840.
Honegger, R. E. 1981. List of amphibians and reptiles either known or thought to have become extinct since 1600. Biological Conservation 19:141-158.
Howarth, F. G. 1983. Classical biological control: panacea or Pandora's Box? Proceedings of the Hawaiian Entomological Society 24:239-244.
-----. 1985. Impacts of alien land arthropods and mollusks on native plants and animals in Hawaii. Pages 149-178 in C. P. Stone and J. M. Scott, editors. Hawaii's terrestrial ecosystems: preservation and management. University of Hawaii Cooperative National Park Resources Study Unit, Honolulu, Hawaii, USA.
-----. 1991. Environmental impacts of classical biological control. Annual Review of Entomology 36:485-509.
Jouvenaz, D. P. 1990. Approaches to biological control of fire ants in the United States. Pages 620-627 in R. K. Vander Meer, K. Jaffe, and A. Cedeno, editors. Applied myrmecology: a world perspective. Westview, Boulder, Colorado, USA.
Kass, H. 1990. Once a savior, moth is now a scourge. Plant Conservation 5:3.
Lai, P.-Y. 1988. Biological control: a positive point of view. Proceedings of the Hawaiian Entomological Society 28:179-191.
Lever, C. 1985. Naturalized mammals of the world. Longman, London, England.
Lockwood, J. A. 1993a. Environmental issues involved in biological control of rangeland grasshoppers (Orthoptera: Acrididae) with exotic agents. Environmental Entomology 22:503-518.
-----. 1993b. Benefits and costs of controlling rangeland grasshoppers (Orthoptera: Acrididae) with exotic organisms: search for a null hypothesis and regulatory compromise. Environmental Entomology 22:904-914.
Malecki, R. A., B. Blossey, S. D. Hight, D. Schroeder, L. T. Kok, and J. R. Coulson. 1993. Biological control of purple loosestrife. BioScience 43:680-686.
McClure, M. S. 1991. Density-dependent feedback and population cycles in Adelges tsugae (Homoptera: Adelgidae) on Tsuga canadensis. Environmental Entomology 20:258-264.
McInnes, D. 1994. Comparative ecology and factors affecting the distribution of north Florida fire ants. Dissertation. Florida State University, Tallahassee, Florida, USA.
Mills, L. S., M. E. Soule, and D. F. Doak. 1993. The keystone-species concept in ecology and conservation. Bio-Science 43:219-224.
Moore, N. W. 1989. The bird of time. Cambridge University Press, Cambridge, England.
Moyle, P. B., H. W. Li, and B. A. Barton. 1986. The Frankenstein effect: impact of introduced fishes on native fishes in North America. Pages 415-426 in R. H. Stroud, editor. Fish culture in fisheries management. American Fisheries Society, Bethesda, Maryland, USA.
Myers, R. L. 1983. Site susceptibility to invasion by the exotic tree Melaleuca quinquenervia in southern Florida. Journal of Applied Ecology 20:645-658.
Nechols, J. R., and W. C. Kauffman. 1992. Introduction and overview. Pages 1-5 in W. C. Kauffman and J. E. Nechols, editors. Selection criteria and ecological consequences of importing natural enemies. Entomological Society of America, Lanham, Maryland, USA.
Nechols, J. E., W. C. Kauffman, and P. W. Schaefer. 1992. Significance of host specificity in classical biological control. Pages 41-52 in W. C. Kauffman and J. E. Nechols, editors. Selection criteria and ecological consequences of importing natural enemies. Entomological Society of America, Lanham, Maryland, USA.
Norton, B. G. 1987. Why preserve natural variety? Princeton University Press, Princeton, New Jersey, USA.
Paine, R. T. 1966. Food web complexity and species diversity. American Naturalist 100:65-75.
-----. 1969. A note on trophic complexity and community stability. American Naturalist 103:91-93.
Pemberton, R. W. 1995. Cactoblastis cactorum in the United States: an immigrant biological control agent or an introduction of the nursery industry? American Entomologist 41:230-232.
Porter, S. D., and D. A. Savignano. 1990. Invasion of polygyne fire ants decimates native ants and disrupts the arthropod community. Ecology 71:2095-2106.
Porter, S. D., B. Van Eimeren, and L. E. Gilbert. 1988. Invasion of red imported fire ants (Hymenoptera: Formicidae): microgeography of competitive replacement. Annals of the Entomological Society of America 81:913-918.
Prokopy, R. J., S. R. Diehl, and S. S. Cooley. 1988. Behavioral evidence for host races in Rhagoletis pomonella flies. Oecologia 76:138-147.
Ratcliffe, D. 1979. The end of the large blue butterfly. New Scientist 8:457-458.
Reardon, R. C. 1981. Rearing, evaluating, and attempting to establish exotic species and redistributing established species. Attempts before 1960. Pages 348-351 in C. C. Doane and M. L. McManus, editors. The gypsy moth: research toward integrated pest management. United States Department of Agriculture, Washington, D.C., USA.
Robertson, M. 1990. "Moth monitoring." Nature Conservancy Weekly Stewardship News Reports, 20 June: p. 1.
Roderick, G. K. 1992. Postcolonization evolution of natural enemies. Pages 71-86 in W. C. Kauffman and J. E. Nechols, editors. Selection criteria and ecological consequences of importing natural enemies. Entomological Society of America, Lanham, Maryland, USA.
Rolston, H., III. 1991. Environmental ethics: values in and duties to the natural world. Pages 73-96 in F. H. Bormann and S. R. Kellert, editors. Ecology, economics, ethics: the broken circle. Yale University Press, New Haven, Connecticut, USA.
Schaefer, P. W., R. J. Dysart, and H. B. Specht. 1987. North American distribution of Coccinella septempunctata (Coleoptera: Coccinellidae) and its mass appearance in coastal Delaware. Environmental Entomology 16:368-373.
Schmitz, D. C. 1994. The ecological impact of non-indigenous plants in Florida. Pages 10-28 in D. C. Schmitz and T. C. Brown, editors. An assessment of invasive non-indigenous species in Florida's public lands. Florida Department of Environmental Protection, Tallahassee, Florida, USA.
Shrader-Frechette, K. S., and E. D. McCoy. 1993. Method in ecology: strategies for conservation. Cambridge University Press, Cambridge, England.
Simberloff, D. 1991. Keystone species and community effects of biological introductions. Pages 1-19 in L. Ginzburg, editor. Assessing ecological risks of biotechnology. Butterworth-Heinemann, Boston, Massachusetts, USA.
-----. 1992. Conservation of pristine habitats and unintended effects of biological control. Pages 103-117 in W. C. Kauffman and J. E. Nechols, editors. Selection criteria and ecological consequences of importing natural enemies. Entomological Society of America, Lanham, Maryland, USA.
-----. 1994. The ecology of extinction. Acta Palaeontologica Polonica 38:159-174.
Simmonds, F. J., and F. D. Bennett. 1966. Biological control of Opuntia spp. by Cactoblastis cactorum in the Leeward Islands (West Indies). Entomophaga 11:183-189.
Swezey, O. H. 1931. Records of introductions of beneficial insects into the Hawaiian islands. Pages 368-389 in F. X. Williams, compiler. Handbook of insects and other invertebrates of Hawaiian sugar cane fields. Advertiser, Honolulu, Hawaii, USA.
Tschinkel, W. R. 1988. Distribution of fire ants Solenopsis invicta and S. geminata in north Florida in relation to habitat and disturbance. Annals of the Entomological Society of America 81:76-81.
-----. 1993. The fire ant (Solenopsis invicta): still unvanquished. Pages 121-136 in B. N. McKnight, editor. Biological pollution: the control and impact of invasive exotic species. Indiana Academy of Science, Indianapolis, Indiana, USA.
Tuduri, J. C. G., L. F. Martorell, and S. M. Gaud. 1971. Geographical distribution and host plants of the cactus moth, Cactoblastis cactorum (Berg), in Puerto Rico and the United States Virgin Islands. Journal of Agriculture of the University of Puerto Rico 55:130-134.
U.S. [United States] Congress OTA [Office of Technology Assessment]. 1993. Harmful non-indigenous species in the United States. OTA-F-565. U.S. Congress, Office of Technology Assessment, Washington, D.C., USA.
Waage, J. K., and D. J. Greathead. 1988. Biological control: challenge and opportunities. Philosophical Transactions of the Royal Society of London B 318:111-128.
Walley, K. A., M. S. I. Kahn, and A. D. Bradshaw. 1974. The potential for evolution of heavy metal tolerance in plants. 1. Copper and zinc tolerance in Agrostis tenuis. Heredity 32:309-319.
Williamson, M. 1992. Environmental risks from the release of genetically modified organisms (GMOs) - the need for molecular ecology. Molecular Ecology 1:3-8.
Zimmerman, E. C. 1958. Insects of Hawaii. Volume 7. Macrolepidoptera. University of Hawaii Press, Honolulu, Hawaii, USA.
-----. 1978. Insects of Hawaii, Volume 9. Microlepidoptera, Part I. University of Hawaii Press, Honolulu, Hawaii, USA.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Special Features|
|Author:||Simberlof, Daniel; Stiling, Peter|
|Date:||Oct 1, 1996|
|Previous Article:||The Ecology of Fire.|
|Next Article:||Managing tropical rice pests through conservation of generalists natural enemies and alternative prey.|