By Jason Van Driesche and Roy Van Driesche
Illustration ©Janusz Kapusta/SIS
If only all conservation crises were this simple. It was 1920. The invasion of prickly pear cacti (Opuntia spp.) in Australia had become an emergency of national proportions. Millions of acres of grassland and forest were covered with these invaders, leaving domestic grazers with nothing to eat and native species with severely degraded habitat. So it was with high hopes that the Australian government began looking for natural enemies of prickly pears. Their goal was to harness the tightly evolved relationships between these natural enemies and their host plants to knock the prickly pears back to incidental levels. As luck would have it, they found an ideal natural enemy in a moth called Cactoblastis cactorum, and within a few years, the prickly pear invasion was nothing but a bad memory. And the best part was, because there are no native cacti in Australia, the introduction of a cactus-specific herbivore posed absolutely no threat to Australia’s native species.
But then there is the complex and unfortunate case of the endangered semaphore cactus (Opuntia spinosissima). A prickly pear native to the Florida Keys, the semaphore is in danger of extinction in part because it is under attack by the same biocontrol agent that was responsible for the successful control of prickly pears in Australia. Following on the heels of decades of habitat loss to development, the arrival of the moth in Florida may prove a deadly blow to what was already a quite rare species. The story is that various Caribbean nations imported Cactoblastis cactorum between 1957 and 1970. They released it into the wild in a misguided effort to control native and introduced prickly pear cacti that had become invasive in overgrazed pastures or that were simply undesirable to ranchers. As in Australia, the reason for introducing the moth was largely economic—but the ecological context was entirely different. The Caribbean is home to many endemic prickly pear species, making it perhaps one of the worst places on earth to introduce a powerful natural enemy of prickly pears. The moth soon migrated to Florida via the horticultural trade and has had devastating impacts on the already threatened semaphore cactus.
The key thing to notice here is the questions that were not asked. None of those responsible for importing C. cactorum to the Caribbean ever stopped to ask what the real problem was—and how best to solve it. Biocontrol never should have been used as a “band-aid” for overgrazing, and it never should have been used in an ecosystem with so many native species related to the target pest. Though we know far more about the ecology of introduced species now than we did 80 years (or even 30 years) ago, it wasn’t lack of ecological knowledge that created the semaphore cactus disaster; it was the blinders that come with a narrow-minded focus on a limited, nonecological goal.
No matter how much we know in theory, it is the thoughtful application of knowledge in context that determines on-the-ground outcomes. Then and now, the trick to navigating the uncertainties inherent in biological control is taking the time to ask the right questions.
And that is the angle we’ve taken in this article. When you’re up against an invasive species crisis, biocontrol can look like the obvious answer. But it’s critical to step back and evaluate the appropriateness of biocontrol to a given management problem—as well as the hazards of doing nothing in a world of invasions that are increasingly unmanageable by other means.
Avoiding Collateral Damage
The history of biocontrol is blemished with what today are judged as catastrophic mistakes. A few—most of them from the early years of the discipline—are the product of nothing but ecological ignorance. The introduction of mongooses to Hawai’i to control rats, for instance, ignored the basic fact that rats are nocturnal and mongooses are generalists and hunt by day. Many of the more recent examples, though, are less a consequence of ecological ignorance than of changing social norms. Until a few decades ago, many nontarget impacts were generally considered acceptable if the benefits to agriculture were large enough.
The first rule of safe biocontrol, then, is to ensure that possible nontarget impacts are not only evaluated with care but also are taken seriously. The case of the non-native thistle-feeding weevil Rhinocyllus conicus is a case in point. R. conicus was introduced to North America in 1968 to control musk thistle (Carduus nutans) and other non-native thistles that are considered significant pasture weeds. It has since proven fond not only of its original non-native target but also of several native (and in some cases, rare) North American thistle species.
Biocontrol researchers attempt to prevent problems like this one in part through host-range testing. Candidate agents are presented with a series of native species more and more distantly related to the invader to determine if the agents will feed on, and more importantly, complete their life cycle on any of those native species. These tests are relatively straightforward to perform. The hard part is evaluating the results.
Have the right species been tested? How much feeding on a native species is “significant?” How will a change of context from the lab to the wild—and from the point of introduction to a different ecological context—affect the behavior of the biocontrol agent? The wrong interpretation could result in ecological disaster.
The scientists responsible for the thistle bio-control effort justified the weevil’s release based on host specificity tests whose basic principles were largely the same as those in use today. These tests indicated that, although the weevil could complete its life cycle on native and introduced thistles in three genera, it had a strong preference for and developed more rapidly and to a larger size on the non-native thistles. On the basis of these results, the researchers argued that the weevil would have no significant impacts on natives. The 35 years since its release have proven the researchers wrong.
The weevil attacks and significantly reduces seed production of the native Platte thistle (Cirsium canescens) as well as other native thistles. In hindsight, this weevil should never have been introduced, for it is too much of a generalist to be considered safe.
The question, then, is whether these impacts could have been foreseen—and therefore avoided. In other words, was it inherent unpredicability of species introductions or unsound evaluation of test results that led researchers to release what has turned out to be such a damaging invader?
In a recent article in Biological Conservation, University of Nebraska researchers Amy Arnett and Svata Louda reported on an experiment designed to answer just that question (1). They hypothesized that the unexpected impacts of R. conicus on native thistles were the result of a post-release expansion in host range; that is, that the weevil had adapted over time to reproduce with greater success on North American thistles than was the case before its introduction to North America. If this proved true, the future for biocontrol would be bleak—no predictions about impacts could be made with any certainty.
What they found was exactly the opposite. When the researchers subjected 28th-generation naturalized weevils collected from a stand of Platte thistles in western Nebraska to exactly the same tests that were used in the original 1968 evaluation, they found that host specificity had not changed at all. R. conicus still showed a strong preference for the non-native thistles it was introduced to control. This meant that the weevil was feeding heavily on Platte thistle and other native thistles simply because in western Nebraska the natives were the only hosts available; the introduced musk thistle is not present in that part of the country. In short, the reduction in fitness that the weevils suffered as a result of this host switch was not enough to reduce its population to nondamaging levels.
Arnett and Louda’s research indicates that even a “strong preference” for a target species may not be sufficient to guard against collateral damage. It also shows the importance of evaluating host specificity data with an eye to worst-case scenarios. If a biocontrol agent is able to reproduce on something other than its target, one must assume that somewhere, at some point, it will. Even if the rate of reproduction is reduced, the impacts of that biocontrol agent on native species can be considerable. The key is to evaluate how impacts might vary across the full range of ecological contexts a biocontrol agent is likely to reach—not just the context in which it is to be introduced.
Notice how much has changed in the last 35 years. The researchers who made the decision to release the weevil discounted what they knew to be the worst-case scenario; the nontarget impacts on native thistle species were unimportant by the standards of their time. Today’s norms are much more attuned to broader ecological consequences, and researchers have begun to take a more precautionary approach. The data remain largely the same. All that has changed is the scientific and cultural consensus on how much collateral damage is acceptable and under what circumstances it makes sense to proceed even when nontarget impacts are likely.
The Costs of Doing Nothing
In a perfect world, biocontrol would only be used for those invaders whose natural enemies attacked absolutely nothing but the target invader. But sometimes you have to bend the rules. In cases where chemical, mechanical, or cultural management tools are useless and impacts are significant, conservationists must weigh the impacts of the biocontrol agent on nontarget species against the consequences of doing nothing.
Such is the case of the imported red fire ant (Solenopsis invicta), an invader in the southeastern United States that inflicts painful bites and causes significant crop damage. It is also the single most important factor in the decline of native fire ants (as well as many other ant species), outcompeting them wherever the invader and the native come into contact. To make matters worse, available poisons do not discriminate between native and imported ants and therefore have no value as a tool for native species conservation. So in the face of serious and unmanageable ecological impacts, protecting native ant species will happen via biocontrol or not at all. Under such circumstances, the current consensus in the conservation community appears to favor biocontrol, so long as the benefits to native species outweigh the costs.
This is the standard that researchers held themselves to as they considered whether to introduce a natural enemy of the imported red fire ant called Pseudacteon curvatus, a species of decapitating fly native to South America. Decapitating flies are parasites that reproduce by attacking fire ants and laying their eggs inside the ants’ head cavity. Most are tightly co-evolved with a specific species of ant, making them ideal candidates for use as biocontrol agents.
Although the fly shows a strong preference for the imported fire ant, host specificity tests indicated that if it were given no other option, the fly would also attack several native fire ant species in the same genus as that of the imported ant (2). Tests also showed that parasitized native ants were adequate hosts for larval development and that at least some fly eggs laid in native ants survived to adulthood. By the standard of “no collateral damage,” it would appear that this particular species of decapitating fly was not an acceptable candidate for introduction.
However, several factors suggested that the net impacts on native ant species would not be as uniformly negative as the raw numbers implied. As in the case of the thistles, the flies showed a strong preference for their non-native target, attacking native fire ants only 6 percent to 35 percent as often as they attacked the imported fire ant. Whereas these results by themselves would not be sufficient to justify release (after all, the fly did attack and kill native fire ants), the rest of the evidence tipped the scales in favor of introduction.
First, the frequency with which the fly attacked native fire ant species was too low to create a self-perpetuating fly population. As such, the fly would not be able to survive in locations where all it had available were native hosts. (Recall that ability to survive in the absence of the target non-native species was the key contributing factor to the R. conicus weevil’s impact on native thistles.) Second, native fire ant species already are subject to parasitism by native decapitating flies in the same genus as P. curvatus, and therefore are evolutionarily adapted to withstand the kind of pressure that such parasites bring to bear. Finally, and most importantly, the single greatest threat to native fire ants is competitive pressure from imported fire ants. In invaded areas, one native fire ant species plummets in numbers, and the other disappears entirely. As a consequence, the researchers reasoned that anything that reduces the relative advantage of S. invicta is likely to have net positive benefits, even for those native ant species that would suffer some degree of parasitism by the biocontrol agent.
So in the spring of 2000, P. curvatus was released in Alabama on a population of imported fire ants. The fly has since established and expanded its range to about ten miles from the point of release. A closely related fly is slated for introduction in spring 2003, and studies on nontarget impacts will begin in 2003 as well. Sanford Porter, one of the lead researchers on the project, is confident that even though the flies attack native species, the reduction in competitive pressure from the non-native fire ants will more than compensate for any incidental parasitism that the native species suffer. Even more important, Porter argues, are two factors that differentiate this case from that of the thistles: the flies are unlikely to survive if nothing but native hosts are present, and the native species that the flies attack are all very common and therefore able to absorb some parasitism without showing any negative population-level effects.
The driving force of this biocontrol project was a desire to do more good than harm. In an invaded world, sometimes that is the best you can hope for. Biocontrol is a delicate balancing act, for it demands that we identify the best course of action while avoiding collateral damage. Few decisions will be without cost. But in a rapidly changing world, inaction is a form of action—and sometimes doing something is better than doing nothing.
Life in an Invaded World
Invasions are the reality of our age. Humans and the species that accompany them have become cosmopolitan, and there is no going back. In many places, we are hard-pressed even to define what or where “back” might be. It is entirely futile to long for a return to an ecosystem free of “invaders,” for as far as a species is concerned, it belongs wherever it can survive.
In this context, biological control is not a tool for restoring community composition per se but rather for restoring a sense of what one might call ecological fair play. When an invasive species abruptly and dramatically alters the rules of the game such that one or many other native species are put at a marked disadvantage, biocontrol may provide a vital counterbalance. One can define biological control as an effort to tilt the field such that native diversity and function are retained.
But this is a difficult and dangerous business, requiring much caution and humility. There is no question about the power of biocontrol. The real issue is the wisdom and the insight that we bring to its application in an ever-more-muddled ecological context.
1. Arnett, A. and S.M. Louda. 2002. Re-test of Rhinocyllus conicus. Host specificity and the prediction of ecological risk in biological control. Biological Conservation 106:251-257.
2. Porter, S. 2000. Host specificity and risk assessment of releasing the decapitating fly Pseudacteon curvatus as a classical biocontrol agent for imported fire ants. Biological Control 19:35-47.
Arnett, A. and S.M. Louda. 2002. Re-test of Rhinocyllus conicus. Host specificity and the prediction of ecological risk in biological control. Biological Conservation 106:251-257.
Food and Agriculture Organization of the United Nations. 1997. Code of Conduct for the Import and Release of Exotic Biological Control Agents. Biocontrol News and Information 18(4):119N-24N.
Louda, S.M. and C.W. O’Brien. 2002. Unexpected ecological effects of distributing the exotic weevil, Larinus planus (F.), for the biological control of Canada thistle. Conservation Biology 16(3):717-727.
Louda, S.M. et al. 2003. Nontarget effects—The Achilles’ heel of biological control? Retrospective analyses to reduce risk associated with biocontrol introductions. Annual Review of Entomology 48:365-396.
Porter, S. 2000. Host specificity and risk assessment of releasing the decapitating fly Pseudacteon curvatus as a classical biocontrol agent for imported fire ants. Biological Control 19:35-47.
Van Driesche, R.G. and T.S. Bellows, Jr. 1996. Biological Control. Chapman and Hall, New York.