Our recent paper (Pizzatto & Shine, 2012) suggested a new approach that may help to manage invasive cane toads (Rhinella marina) in Australia: the use of a native frog species as a ‘Typhoid Mary’ to spread a lungworm Rhabdias pseudosphaerocephala to toads. Both commentaries elicited by that paper (Harris, 2012; Hough-Goldstein, 2012) suggest that we are too optimistic about the effectiveness of lungworm dissemination for toad control, and too dismissive of potential collateral impacts of our method on native species. Nonetheless, the commentaries are in broad agreement with us about the nature of costs and benefits involved. Some of the reservations outlined in the commentaries are ones that have already occurred to us, and have stimulated follow-up studies that (so far at least) suggest that the likely collateral impacts are minor. Below, we consider each point in turn, concluding with the most important but most-difficult-to-answer issue: when do we decide that the potential risks of any new control method outweigh the biodiversity impacts that accrue while we delay?
First, would increasing lungworm densities really control toad populations? We cannot answer that question, because the data to do so are incredibly difficult to obtain. So far as we know, there is no amphibian species worldwide for which we have robust evidence on the role of parasites in controlling population size. What we can say is that the lungworm can severely depress survival, endurance (Kelehear, Webb & Shine, 2009; Pizzatto & Shine, 2011a) and growth in metamorph cane toads, even under favorable laboratory conditions (Kelehear et al., 2009), but does not always do so (Pizzatto & Shine, 2011a,c); and that field as well as laboratory evidence shows that lungworm infection reduces growth rates of juvenile and adult cane toads (Kelehear, Brown & Shine, 2011). Thus, the evidence that lungworms reduce toad viability is stronger than for virtually any other host–parasite system involving an amphibian. The degree to which those effects translate into population-level parameters is a critical question, but very difficult to answer. All that we can confidently state is that increasing parasite numbers will likely reduce viability (growth, survival, endurance) of some proportion of cane toads within a population. Obviously, transmission dynamics will depend upon both biotic (host and parasite) traits and abiotic factors (Pizzatto, Kelehear & Shine, unpubl. data). Harris (2012) suggested that high density of hosts and confinement in our laboratory experiments might have led to unrealistically high transmission rates. However, toad densities in our design were not unrealistic: we have often found toads aggregated at much higher densities in natural conditions in tropical Australia (Fig. 1), and levels of parasite prevalence and intensity in the field often are similar to those in our experiments (Barton, 1998; Pizzatto, Kelehear & Shine, unpubl. data).
Second, does our novel approach offer a feasible way to increase the rate of lungworm infection in cane toad populations at the invasion front? Our experiments in the laboratory may fail to mimic natural conditions, such that (for example) differences in habitat use between cane toads and green tree frogs (Litoria caerulea) will mean that the former species will rarely if ever encounter lungworms as they disperse from feces deposited by the latter species. Our paper also identified this as a critical assumption. Our follow-up radiotelemetry studies have confirmed significant habitat overlap between the two species. Frogs tend to be arboreal and philopatric whereas toads are terrestrial and more dispersive (at least in the invasive front); but when active, L. caerulea sit on fallen branches, move and forage on the leaf litter, and often leave their feces on the ground in areas occupied by cane toads (Pizzatto, Both & Shine, unpubl. data). Will frogs maintain sufficiently high infection rates for long enough to serve as useful Typhoid Marys? First, transmission rates will likely be higher in nature than in our experiments where frogs sat in plastic boxes without substrate, from which contaminated feces were removed; Pizzatto & Shine, 2012). Second, our studies showed that adult frogs maintained infections for at least 120 days; in a management context, this provides ample opportunities to transmit the lungworm.
Third, would the release of lungworm-infected green tree frogs transmit these parasites to the magnificent tree frog (Litoria splendida), which (unlike the green tree frog) is badly affected by these parasites? Again, we identified this as an issue and initiated field studies to evaluate habitat overlap between the two taxa. Our radio-tracking and surveys reveal some overlap (especially in anthropogenically modified areas), but magnificent tree frogs prefer rock outcrops, and green tree frogs are rare in sites where magnificent tree frogs are abundant (Pizzatto, Both & Shine, unpubl. data). Thus, parasite transfer between the two species is unlikely to be common. The possibility that the toads' lungworm may also affect other native anurans certainly merits close attention, especially considering the high diversity of anurans in tropical Australia; fortunately, our detailed studies so far have shown that most common native species are unlikely to be affected by the lungworm (Pizzatto, Shilton & Shine, 2010; Pizzatto & Shine, 2011a,b). The lack of host-switching over the toads' 75-year history in Australia, and its host-specificity within its native range, both argue that host-switching is unlikely (Dubey & Shine, 2008; Pizzatto et al., 2012).
Fourth, would it not be simpler and more effective to infect metamorph toads, rather than native frogs, and release them to spread lungworms? We advocated this method for existing populations, but it is impossible at the invasion front (at least at present) because of opposition by ‘toad-busting’ community groups. These volunteer organizations take a straightforward approach to toad control – they simply collect and kill as many toads as possible – and are politically powerful; they have vigorously opposed the idea of releasing additional toads at the invasion front. Most researchers who work on invasive species will attest to the pervasive influence of politics on the acceptability of alternative management tactics.
Although we do not accept some points raised in the commentaries, we are in broad agreement with both authors. There is no doubt that cane toad invasion is devastating for some, but not all, species of native fauna (Shine, 2010), and that we urgently need new approaches to control toads (Saunders et al., 2010). It is also true that any new approach (such as the one we suggest) may not work; or that it may have collateral impacts on native biota. The obvious way forward is more research. However, the reality is that every year's delay sees cane toads spread at least another 50 km, killing thousands of native animals in the process. When do we decide that enough is enough, and desperate problems require risky solutions? We will never be completely certain of the impact of any new method until we try it. At some point, we will need to start field tests of our new ideas rather than conduct further trials to predict the effectiveness and consequences of our methods. We need a broad debate, involving the general public as well as scientists, to clarify exactly when is the right time to test new approaches to invasive-species control.