THE QUARTERLY REVIEW of Biology

Home , Cane toad, Cane toads in Australia, Delicate mouse, Opisthodon, Ornate burrowing frog

Volume 85, No. 3 September 2010 THE QUARTERLY REVIEW of Biology

THE ECOLOGICAL IMPACT OF INVASIVE CANE TOADS (BUFO MARINUS) IN AUSTRALIA

Richard Shine School of Biological Sciences A08, University of Sydney, NSW 2006, Australia e-mail: [email protected]

keywords alien species, food-webs, introduced species, invasive species, predator-prey, Rhinella marina, toxicity

abstract Although invasive species are viewed as major threats to ecosystems worldwide, few such species have been studied in enough detail to identify the pathways, magnitudes, and timescales of their impact on native fauna. One of the most intensively studied invasive taxa in this respect is the cane toad (Bufo marinus), which was introduced to Australia in 1935. A review of these studies suggests that a single pathway—lethal toxic ingestion of toads by frog-eating predators—is the major mechanism of impact, but that the magnitude of impact varies dramatically among predator taxa, as well as through space and time. Populations of large predators (e.g., varanid and scincid lizards, elapid snakes, freshwater crocodiles, and dasyurid marsupials) may be imperilled by toad invasion, but impacts vary spatially even within the same predator species. Some of the taxa severely impacted by toad invasion recover within a few decades, via aversion learning and longer-term adaptive changes. No native species have gone extinct as a result of toad invasion, and many native taxa widely imagined to be at risk are not affected, largely as a result of their physiological ability to tolerate toad toxins (e.g., as found in many birds and rodents), as well as the reluctance of many native anuran-eating predators to consume toads, either innately or as a learned response. Indirect effects of cane toads as mediated through trophic webs are likely as important as direct effects, but they are more difﬁcult to study. Overall, some Australian native species (mostly large predators) have declined due to cane toads; others, especially species formerly consumed by those predators, have beneﬁted. For yet others, effects have been minor or have been mediated indirectly rather than through direct interactions with the invasive toads. Factors that increase a predator’s vulnerability to toad invasion include habitat overlap with toads, anurophagy, large body size, inability to develop rapid behavioral aversion to toads as prey items, and physiological vulnerability to bufotoxins as a result of a lack of coevolutionary history of exposure to other bufonid taxa.

253 254 THE QUARTERLY REVIEW OF BIOLOGY Volume 85

Introduction NVASIVE SPECIES are widely viewed as Imajor threats to biodiversity, and there are many examples of catastrophic ecological impacts resulting from such taxa (Pimen- tel et al. 2000). However, there are equally abundant examples of introduced species that have had relatively minor ecological impacts on native fauna, despite spreading widely, and still others that have even had positive effects (Wonham et al. 2005; King et al. 2006). Understanding the nature of interactions between native and introduced spe- Figure 1. Variation in Body Size of BUFO MARINUS cies is critical for establishing conservation Cane toads span an enormous range of body sizes, priorities; given limited resources, conserva- even in post-metamorphic stages (i.e., ignoring the eggs tion efforts need to focus on species and and larvae). This photograph shows a newly metamor- areas where the invader’s impact is most se- phosed toadlet on the head of a large adult female toad. vere. Also, knowledge about pathways of im- (Photograph by T. Child, and used with permission.) pact may suggest novel approaches to impact reduction; for instance, it might be possible and adults), although toxin types and con- to manipulate an invasive predator’s re- tent change markedly during toad ontog- sponses to native prey through conditioned eny (Hayes et al. 2009). Toxin levels and, taste aversion learning (Nicolaus et al. consequentially, danger to vertebrate pred- 1983a,b; O’Donnell 2009; O’Donnell et al. ators are high in eggs, decrease through 2010). tadpole life, are lowest at around the time This paper discusses the ecological im- of metamorphosis, and increase rapidly pact of a large anuran amphibian—the thereafter (Hayes et al. 2009). cane toad, Bufo marinus (or Rhinella ma- Brought to Australia from Hawaii in rina) (Frost et al. 2006; Pramuk 2006) (see 1935 to control sugar cane pests in north- Figure 1)—in Australia. The cane toad is eastern Queensland, cane toads have now one of the “true toads” (family Bufo- spread throughout much of Queensland, nidae)—a group with a natural worldwide northern NSW, and the wet-dry tropics distribution in the Americas, Africa, and of the Northern Territory. In 2009, they Eurasia (Tyler 1975; Zug et al. 1975; Cog- crossed the border into Western Australia ger 2000). The cane toad is among the (Figure 2). Because cane toads eat a wide world’s largest anurans (measuring up to variety of prey, have greater fecundity than 23 cm, 1.25 kg) (Tyler 1975), with a natu- native anurans, and develop rapidly in ral range in Central and tropical South tropical regions, colonizing cane toads at- America extending from 27° N latitude in tain very high densities (Freeland 1986; southern Texas and western Mexico to 10° Lever 2001). They tolerate a broad range S latitude in central Brazil (Zug and Zug of environmental and climatic conditions 1979). As in other bufonids, the parotoid and occupy a wide range of habitats (Ur- (shoulder) glands of cane toads contain ban et al. 2007), albeit with a clear prefer- potent bufogenins, bufotoxins, and other ence for anthropogenically degraded sites substances (Zug and Zug 1979), but the (Zug and Zug 1979). Ultimately, the cane active constituents of cane toad toxins toad’s Australian range may encompass an (steroid-derived) differ from those of Aus- area of approximately 2 million km2 (Suth- tralian native frogs (typically peptides) erst et al. 1995). Exact predictions of the (Daly and Witkop 1971; Tyler 1987). The toad’s eventual range within Australia de- cane toad is toxic throughout all of its life pend on the modelling approaches used, stages (i.e., eggs, tadpoles, metamorphs, but all models agree that the Kimberley September 2010ECOLOGICAL IMPACT OF INVASIVE TOADS 255

formation available on the ecological impacts of most invasive species, including the cane toad itself in the many other countries in which it has been introduced (Lever 2001). As a result, the toads’ Australian invasion offers a robust opportunity to investigate exactly what happens when a highly toxic invasive species spreads through a complex tropical ecosystem far outside its native range. The ecological impact of cane toads has aroused great concern not only among conservation biologists and wildlife managers, but also among members of the general public (see Beckmann and Shine 2009; Somaweera et al. 2010). In this paper, I summarize the current understanding on the ecological impacts of invasive cane toads on native fauna in Australia. The topic has been reviewed several times before, notably by Freeland Figure 2. Geographic Range of BUFO MARINUS in (1984, 2004), van Dam et al. (2002), and Australia Webb and Glanznig (2004). However, Spread of the invasive cane toad (Bufo marinus) these reviews were written prior to a re- through Australia (upper panel), and the predicted cent major expansion of research on final distribution of the species based on sites that, cane toad impacts (see below). As the according to climatic data, are likely to sustain toad breeding for three or more months per year (lower toad invasion front moved closer to the panel). The upper map is based on an updated data- city of Darwin, logistical obstacles to re- set assembled by Urban et al. (2008), and the lower search decreased at the same time as map is taken from Kearney et al.’s (2008) study (used public concern—and thus, funding op- with permission). Black crosses show records of toad portunities—increased. This review fo- occurrence; grey shading shows predicted final extent cuses on threats to native fauna rather of toad colonization. than to domestic pets or humans; for toad impact on those latter species, see Knowles (1964), Roberts et al. (2000), region of northwestern Australia will be Sakate and Oliveira (2000), and Reeves suitable for toads (Urban et al. 2007, 2008; (2004). Kearney et al. 2008). The situation in Intuition suggests that evaluating the im- southern Australia is more difficult to pre- pact of cane toads on native fauna should dict, especially given the uncertainty about be straightforward: we merely need to as- future climatic conditions (see Figure 2). sess abundance of native taxa before versus Reflecting its relative recency and the after toad invasion at a site, or compare ease with which cane toads can be distin- faunal abundances in sites containing cane guished from native frogs, the invasion of toads versus those lacking cane toads. Un- cane toads through Australia has been fortunately, such population-level counts documented in enough detail to allow are often subject to wide error and are for precise analysis of speeds and trajec- difficult to interpret. For example, exten- tories of progress of the invasion front sive surveys in the wet-dry tropics by Woin- (Urban et al. 2007, 2008). Research on arski et al. (2004) showed that more than the ecological impact of cane toads in 1000 sample sites would be needed in or- Australia has also been unusually de- der to attain a 90% certainty of detecting a tailed when compared to the sketchy in- 20% change in abundance. This indicates 256 THE QUARTERLY REVIEW OF BIOLOGY Volume 85 that even limited experimental studies may lack a history of evolutionary exposure about mechanisms of impact may reveal to anuran predation during the dry season. more than extensive survey work (Thom- Reductions in the abundance of seed- son 2006); therefore, I will first review harvesting ants could potentially alter information about pathways of impact. Al- plant community dynamics, leading to though the invasion of cane toads may long-term vegetation changes (Andrew affect native species in several ways (e.g., 1986; Andersen 1991; Andersen and Braith- predation on native animals, competition waite 1996; van Dam et al. 2002). Many with native species, and poisoning of native researchers have reported emaciated toads predators), the available evidence has around dry season waterbodies (e.g., Tyler identified one pathway of direct impact— 1994; R. A. Alford, personal communica- namely, lethal toxic ingestion—as the most tion), thus suggesting that toads have sig- important, and I will discuss this pathway nificantly reduced the abundance of inver- in detail. Finally, I will review information tebrate prey in such areas. Experimental on population-level and longer-term im- work to evaluate this effect would be of pacts of cane toads on biodiversity, and I great interest to future research. will attempt to identify generalities emerg- The only direct experimental analysis ing from the accumulated data (Table 1 of the impact of toad predation on inver- summarizes the nature and magnitude of tebrates was performed in outdoor enclo- likely impacts of cane toads on the Austra- sures on the Adelaide River floodplain in lian fauna). the tropical Northern Territory (Green- lees et al. 2006). The work was conducted Pathways of Cane Toad Impact during the wet season, and reported significant declines in the abundance of in- predation on invertebrates by toads vertebrates in enclosures occupied by The diet of cane toads in Australia toads as compared to anuran-free enclo- consists mainly of arthropods, with ants, sures; however, native frogs had almost as termites, and beetles being the most im- much impact as cane toads in this respect. portant numerically (van Beurden 1980a; Greenlees et al. (2006) also reported that Freeland 1984, 1990; Cameron and Cog- cane toad predation caused shifts in the ger 1992; Werren and Trenerry 1993). species composition of invertebrates, with Although cane toads have been reported a decline in the prey types preferred by to consume 200 food items per night— toads, notably small-bodied invertebrates far more than most native frogs (Tyler such as ants. and Cappo 1983; Freeland et al. 1986a,b)— Although cane toads forage preferen- typical feeding rates may be much lower tially on taxa such as ants and beetles (Bailey and more similar to those of sympatric 1976; Freeland et al. 1986b; Lever 2001), frogs (e.g., Greenlees et al. 2007). they are opportunistic feeders and therefore Because they attain high population might significantly impact some taxa that densities and eat frequently, cane toads would normally comprise only a small part of have the potential to cause declines in the their diet. For example, although anurans abundance and diversity of some inverte- rarely prey upon molluscs (Drewes and Roth brate taxa, particularly terrestrial beetles, 1981), cane toads from invasive populations termites, and ants. Such effects might be in other parts of the world feed readily on especially pronounced during the tropical land snails, with terrestrial gastropods consti- dry season, when the risk of desiccation tuting more than 40% of the contents of drives toads to concentrate around the toad stomachs in some samples (Hinckley shrinking margins of waterbodies (Free- 1963; Bailey 1976; Grant 1996). Thus, toad land and Kerin 1988; Alford et al. 1995; predation might potentially imperil threat- Child et al. 2008a,b, 2009). Many native ened species of land snails (Camaenidae) anurans are inactive during this period; endemic to small limestone outcrops of the therefore, the native invertebrate fauna Kimberley region. Laboratory trials indi- September 2010ECOLOGICAL IMPACT OF INVASIVE TOADS 257 cate that under wet conditions, which 2010a). Increases in the abundance of ant encourage activity both by snails and toads, species, such as the dominant dolichodorine cane toads readily consume a wide range of species within the genus Iridomyrmex, that camaenid species (Pearson et al. 2009). feed upon metamorph toads might have Whether or not this vulnerability will trans- flow-on effects to predation rates on other late into population-level impacts depends animals or to interactions among competing upon encounter rates between snails and ant species. toads. Field surveys reveal little habitat Other native invertebrates that can feed overlap between toads and camaenids, sug- upon cane toads without ill effect include gesting that the snail populations may with- crustaceans such as freshwater crayfish stand toad invasion (Pearson et al. 2009). (Cherax spp., and Euastacus hystricosus, E. suttoni, and E. valentulus) (Hutchings 1979; competition between toads and Crossland 1998a; Crossland and Alford invertebrates 1998) and crabs (Holthuisana sp.: Cross- land and Alford 1998; Holthuisana trans- Cane toads also might influence inverte- versa: M. Crossland, unpublished data), as brates through processes other than preda- well as adult dytiscid diving beetles (Cybister tion. For example, the presence of toad godeffroyi, Hydaticus vittatus, Sandracottus larvae in artificial waterbodies has been bakewelli) (Crossland 1998a; Crossland and observed to suppress the development of Alford 1998), dragonfly larvae (Trapezos- mosquito larvae, and to discourage oviposi- tigma sp., Hemianax papuensis: Crossland tion by female mosquitoes (Hagman and and Alford 1998), and mosquitoes (van Shine 2007). In both these respects, the ef- Beurden 1980b). However, not all inverte- fects of cane toads resemble those of native brates are unaffected by toad toxins. frogs, and presumably involve competitive Leeches (Goddardobdella elegans: Crossland interactions between anuran and mosquito and Alford 1998; unidentified NT species: larvae that have favored the ability of female G. P. Brown, personal communication) fre- mosquitoes to detect anuran tadpole pres- quently die after feeding on larval or adult ence and to use that information when se- toads, and aquatic snails (Austropeplea les- lecting oviposition sites (Mokany and Shine soni) die after consuming toad eggs (Cross- 2003a,b). land and Alford 1998). Other invertebrates vary in their susceptibility to toads. Some predation on toads by invertebrates dytiscid diving beetle larvae (Cybister sp., Toad invasion may provide an additional Hydaticus sp.) and notonectid backswim- food source for predatory native inverte- mers (Anisops sp.), for instance, consume brates, many but not all of which tolerate toad hatchlings or tadpoles without ill ef- toad toxins better than do vertebrates. For fect while other individuals die (Crossland example, meat ants of the genus Iridomyrmex and Alford 1998), and belastomatid giant prey upon many metamorph toads (Clerke water bugs (Lethocerus insulanus) are highly and Williamson 1992; Ward-Fear et al. 2009, sensitive to some developmental stages of 2010a,b), and laboratory and field studies toad tadpoles but not to others (Crossland show that cane toad metamorphs are more 1998a; Crossland and Alford 1998). vulnerable to ant attack than are the metamorphs of native frog species (Ward-Fear predation on vertebrates by toads et al. 2009). That vulnerability reflects the Despite frequent speculation to the con- facts that toad metamorphs are smaller and trary, predation by cane toads on other slower than frog metamorphs, are primarily vertebrates is unlikely to be important in diurnal in their activity (an adaptation to terms of broad ecological impact. Cane reduce vulnerability to intraspecific canni- toads do consume vertebrates, but they do balism) (Pizzatto and Shine 2008; Pizzatto et so rarely (e.g., Stammer 1981), and this is al. 2008), and lack effective antipredator re- true in the aquatic as well as the terrestrial sponses to ants (Ward-Fear et al. 2009, phase of the toad life history. Toad tad- 258 TABLE 1 A summary of the ecological impact of cane toads on native animals in Australia Direction Population Reason for level of Lineage Species of effect Pathway effect? impact Reference Invertebrates Ants, beetles, etc. Negative Predation by toads Variable in space Toads feed heavily on these Catling et al. 1999; and time groups; may depress local Greenlees et al. 2006; abundances Freeland & Kerin 1988 Camaenid land Camaenidae Negative Predation by toads Minor? Little overlap between toads Pearson et al. 2009 snails and snails

Mosquitoes Negative Larval competition Minor? Little overlap in breeding Hagman & Shine 2006, BIOLOGY OF REVIEW QUARTERLY THE sites 2007 Meat ants Iridomyrmex spp. Positive Consumption of Minor? No credible reports of impact Clerke & Williamson toads 1992; Ward-Fear et al. 2009 Crayﬁsh, crabs Positive Consumption of Minor? No credible reports of impact Crossland & Alford 1998 toads Leeches Hirudinea Negative Lethal toxic ingestion Minor? No credible reports of impact Crossland & Alford 1998 Giant water bugs Belostomatidae Negative Lethal toxic ingestion Minor? No credible reports of impact Crossland & Alford 1998 Water beetles Dytiscidae Minor? No credible reports of impact Crossland & Alford 1998 Backswimmers Notonectidae Negative Lethal toxic ingestion Minor? No credible reports of impact Crossland & Alford 1998 Snails Austropeplea Negative Lethal toxic ingestion Minor? No credible reports of impact Crossland & Alford 1998 lessoni Tapeworm Proteocephalidae Negative Predation by toads High? Toads destabilize host- Freeland 1993, 2004 parasite equilibrium (host is python Antaresia maculosa) Fishes Native freshwater Negative Lethal toxic ingestion Minor Toad eggs are deadly but Crossland 1997; Grace & species tadpoles are not; some ﬁsh Sawyer 2008; Grace rapidly learn to avoid toads 2008; Nelson 2008 Anurans Tadpoles Hylids, Negative Larval competition Variable in space Limited overlap in breeding Williamson 1999; Limnodynastids and time sites; reduced densities can Freeland 2004; enhance viability Crossland 2000;

Crossland et al. 2009 Volume Negative Lethal toxic ingestion Minor Limited overlap in breeding Crossland 1997; Crossland sites et al. 2008; Crossland & Shine 2010 85 continued TABLE 1 September Continued Direction Population Reason for level of Lineage Species of effect Pathway effect? impact Reference 00259 2010 Native frogs Hylids, Negative Competition for Minor? Non-equilibrial populations, Freeland & Kerin 1988; Limnodynastids terrestrial food great variation in food Greenlees et al. 2007; supply; diets differ Watson & Woinarski 2003a Negative Competition for Minor? Shelter sites may not be Pizzatto & Shine 2009 CLGCLIPC FIVSV TOADS INVASIVE OF IMPACT ECOLOGICAL shelter sites limiting Negative Lethal toxic ingestion Negligible Rapid aversion learning Nelson 2008; Greenlees et al. 2010a,b Negative Parasite transfer Minor Cross-infection rare in the Dubey & Shine 2008 wild Positive Reduced predation ? Fewer varanids to eat frogs Grigg et al. 2006; Watson & Woinarski 2003a Turtles Pignose turtle Carettochelys Positive Reduced egg ? Goannas killed; therefore, Doody et al. 2006 insculpta mortality egg survival enhanced Long-necked Chelodina rugosa Negative Lethal toxic ingestion Minor Limited overlap with toad Greenlees & Shine 2010 turtle eggs Lizards Goannas Varanus panoptes Negative Lethal toxic ingestion High Steep decline post-invasion, Freeland 2004; Grifﬁths & (Varanidae) Varanus gouldii then recovery McKay 2007; Doody et Varanus mertensi al. 2009; Ujvari & Varanus mitchelli Madsen 2009 Other Varanus Negative Lethal toxic ingestion ? Highly vulnerable to toad Freeland 2004; Doody et spp. toxins al. 2009; Ujvari & Madsen 2009 Scincidae Carlia gracilis Negative ? Moderate Watson & Woinarski (skinks) 2003a.b Bluetongue Tiliqua scincoides Negative Lethal toxic ingestion High Highly vulnerable to toad Price-Rees et al. 2010 lizard intermedia toxins Dragons Chlamydosaurus Negative? Lethal toxic ingestion Reports Reports inconsistent van Dam et al. 2002; (Agamidae) kingii inconsistent Ujvari et al. 2010 Diporiphora Positive ? Moderate No known pathway Watson & Woinarski bilineata 2003a,b Amphibolurus Positive ? Moderate No known pathway Doody et al. 2009 gilberti continued 260

TABLE 1 Continued Direction Population Reason for level of Lineage Species of effect Pathway effect? impact Reference Snakes Keelback Tropidonophis Positive Predation on toads Minor Rarely consume toads in the Llewelyn et al. 2009 mairii wild Slatey-grey snake Stegonotus Positive Predation on toads Minor Rarely consume toads in the Phillips et al. 2003 cucullatus wild? Northern death Acanthophis Negative Lethal toxic ingestion High Highly vulnerable to toad Hagman et al. 2009a; BIOLOGY OF REVIEW QUARTERLY THE adder praelongus toxins Phillips et al. 2010b King Brown Pseudechis Negative Lethal toxic ingestion High? Highly vulnerable to toad G. P. Brown et al., snake australis toxins unpublished data Crocodiles Freshwater Crocodylus Negative Lethal toxic ingestion High in some Highly vulnerable to toad Letnic et al. 2008; Doody crocodile johnstoni areas, not toxins et al. 2009 others Birds Rainbow bee- Merops ornatus Negative Usurpation of ? Boland 2004 eater burrows Corvids, raptors Positive Consumption of Minor Rarely consume toads in the Beckmann & Shine 2009 toads wild? Overall Positive ? Moderate No credible reports of impact Watson & Woinarski 2003a,b Eutherian Water rat Hydromys Positive Consumption of Minor Rarely consume toads in the Cassels 1966 mammals chrysogaster toads wild? Pale field-rat Rattus tunneyi Negative ? Moderate No known pathway; may be Watson & Woinarski unaffected 2003a,b Marsupials Northern quoll Dasyurus Negative Lethal toxic ingestion High? Highly vulnerable to toad Oakwood 2003a,b; hallucatus toxins O’Donnell 2009 Smaller dasyurids Planigale, Negative Lethal toxic ingestion Low Rapid aversion learning Webb et al. 2008; Sminthopsis Llewelyn et al. 2010b Fawn antechinus Antechinus bellus Positive ? Moderate No known pathway; may be Watson & Woinarski unaffected 2003a,b Agile wallaby Macropus agilis Positive ? Moderate No known pathway; may be Watson & Woinarski Volume unaffected 2003a,b This table is restricted to species for which there is evidence of either a pathway of impact or a shift in population densities following toad invasion. 85 September 2010ECOLOGICAL IMPACT OF INVASIVE TOADS 261 poles rarely consume the eggs of native icant predators of O. ornatus larvae (Cross- frogs (Crossland 1998b; M. R. Crossland land 1998b), these effects were probably and R. Shine, unpublished data). One ex- due to competition; however, such com- ception may involve predation by toads on petitive priority effects were reversed when the eggs and nestlings of ground-nesting O. ornatus tadpoles encountered toad eggs. birds such as bee-eaters (Merops ornatus: In this situation, survival of toad eggs was Boland 2004), but, in studies following up reduced via predation, while survival of O. on Boland’s work, Beckmann and Shine ornatus tadpoles was also reduced, via poi- (unpublished data) found rather different soning from ingestion of toxic toad eggs. results. In experimental trials in large out- Thus, lethal ingestion and competition play door enclosures, cane toads avoided rather complex and overlapping roles in this in- than selected burrows containing scent teraction; under some circumstances, the cues from birds or their eggs. However, additional mortality inflicted by cane toad field nest-burrows of bee-eaters often contained toads or frogs, hinting that toads presence may reduce overall tadpole densi- may usurp such burrows and reduce bee- ties enough to increase body sizes at meta- eater reproductive success (C. Beckmann morphosis, therefore potentially increasing and R. Shine, unpublished data). rates of Opisthodon recruitment from a waterbody (Crossland et al. 2009). More evidence that cane toad tadpoles competition between toads and can affect the growth of native anuran tad- aquatic vertebrates poles comes from the Darling Downs area of Cane toads are more fecund than Austra- southern Queensland (Williamson 1999). lian frogs (Pengilley 1981; Tyler 1994), Tadpoles of three out of four native frog and toad tadpoles often form large ag- species (Limnodynastes tasmaniensis, L. terrar- gregations (Wassersug 1971; Freeland eginae, Notaden bennetti) grew more slowly in 2000). Such factors may increase the com- the presence of cane toads, and survival was petitive effects of toad tadpoles on the lar- sometimes reduced in the two Limnodynastes vae of native frogs (Freeland 1984). In an (Williamson 1999). A fourth species, Opisth- experiment designed to assess the effects odon ornatus, showed higher rather than of native tadpoles on toad tadpoles, Alford lower rates of growth if kept with Bufo tad- (1999) found that growth rates of toad tadpoles, reflecting lower survival rates and, poles were unaffected by tadpoles of Pe- thus, densities (Williamson 1999). A range ters’ frog (Litoria inermis), but were greatly of aquatic vertebrates and invertebrates reduced (and survival was, consequently, feed on algae, detritus, and suspended or- reduced to nil) in the presence of tadpoles ganic matter (van Dam et al. 2002), and of the ornate burrowing frog (Limnodyn- Freeland (2004) reported that cane toad astes [Opisthodon] ornatus). Crossland et al. larvae can substantially reduce such algal (2009) found that cane toad tadpoles sig- mats, thereby presumably affecting both nificantly affected tadpoles of O. ornatus via food supply and habitat for other anuran size-based competitive effects, but that these effects were highly sensitive to the larvae as well as invertebrates (note the timing of breeding of the two species. studies by Mokany and Shine 2003a,b, and Opisthodon ornatus tadpoles performed bet- Hagman and Shine 2007 for experimental ter and had increased survival and in- evidence of such interactions between an- creased mass at metamorphosis when they uran larvae and mosquito larvae). Hagman preceded cane toad tadpoles into ponds, and Shine (2008, 2009) reported that the but they performed worse (zero survival) tadpoles of native frog species showed little when cane toad tadpoles preceded them overt reaction to alarm pheromones from into ponds, compared with when both spe- cane toads, thus suggesting that the toad’s cies were added to ponds simultaneously. pheromones are unlikely to suppress activ- Because cane toad tadpoles are not signif- ity of native tadpoles. 262 THE QUARTERLY REVIEW OF BIOLOGY Volume 85

competition between toads and an adverse impact on larger terrestrial frogs terrestrial vertebrates (e.g., Cyclorana australis, Limnodynastes con- Cane toads attain high population densi- vexiusculus); however, aestivation by C. aust- ties (Cameron and Cogger 1992), con- ralis during the dry season removes this sume many invertebrates (Freeland 1984, species from competition during the period 2000), and use the same kinds of retreat sites of the year when resources are most likely to as do many native species (Cohen and Alford be limiting (Janzen 1973). 1996; Schwarzkopf and Alford 1996); there- Direct behavioral interference rather than fore, they may compete with native species competition for food may play a role in the for food and shelter sites. Experimental stud- interactions between toads and native frogs. Pizzatto and Shine (2009) showed that 9 out ies to evaluate competition between toads of 10 native frog species tested avoided and frogs have yielded complex results. In retreat sites scented by cane toad chemical the tropical wet season, weather-dependent cues (the exception was Litoria rubella), shifts in invertebrate abundance and avail- and Greenlees et al. (2007) reported that ability from one night to the next reduce the the presence of cane toads reduced noc- potential for effective competition between turnal activity levels of the native frog Cy- cane toads and other predators utilizing in- clorana australis. Occasional observations of vertebrate prey (Greenlees et al. 2007). As a male toads clasping female frogs, and vice result, total food intake rates of native frogs versa, suggest the potential for sexual ha- (Cyclorana australis and Litoria dahlii) were rassment, as is known to occur between not significantly reduced by the presence of native and introduced taxa in other phylo- cane toads in their outdoor enclosures genetic lineages (e.g., Valero et al. 2008). (Greenlees et al. 2007). Competitive effects Female anurans may experience severe fit- may be more important during the dry sea- ness costs (including drowning, in some son, when high densities of toads persist cases) from prolonged amplexus—a situa- long-term around waterbody margins. Most tion that is particularly likely if males and native anurans are inactive during this time females are of different species and, of year, but other invertebrate-eating taxa, hence, do not react to the intraspecific such as lizards, birds, and small marsupials, cues that normally stimulate termination remain active and may experience reduced of amplexus (Bowcock et al. 2008, 2009). prey availability. Experimental studies to Although most studies of competition evaluate this possibility would be of great have involved native anurans, other verte- interest. brate species also might compete with Focusing on dry season interactions, Free- toads for limited resources. Boland (2004) land and Kerin (1988) examined competi- suggested that cane toads may usurp the tive interactions between cane toads and nesting burrows of bee-eaters, Merops orna- three species of native tree frogs (Litoria pal- tus. Boland reported that 33% of all nest- lida, L. rothii, and L. rubella) around water- ing attempts were “lost” due to cane toads, holes in the Gulf of Carpentaria. Cane toads via either nest predation or nest usurpa- had no detectable impact on habitat or food tion, and that average chick production use by the native frogs, or on the species per nest fell from about 1.2 to 0.8 after composition and population sizes of native toad invasion. frog communities active during the dry season. The absence of competitive interactions transfer of parasites from cane may reflect the cane toad’s heavy reliance on toads to native vertebrates ants and termites, whereas native tree frogs Invasive species can transfer pathogens to eat other kinds of prey (Freeland and Kerin native fauna. Zupanovic et al. (1998) re- 1988). Also, the erratic climatic conditions of ported antibodies against ranaviruses in the region resulted in non-equilibrium frog Australian and native-range cane toads, communities (Freeland 1990). Intuition sug- and speculated that the toad invasion may gests that cane toads are more likely to have bring harmful viral pathogens to native September 2010ECOLOGICAL IMPACT OF INVASIVE TOADS 263 fauna. Delvinquier and Freeland (1988) 1995), also consume toads. Raptors (Mil- concluded that 3 of the 14 protozoan spe- vus migrans, Haliastur sphenurus) take many cies carried by cane toads in Australia were road-killed toads from highways on the Ad- of neotropical origin. Cane toads in Aus- elaide River floodplain, but, like snakes, tralia also contain a nematode lungworm selectively take frogs rather than toads if (Rhabdias pseudosphaerocephala) endemic to given a choice (C. Beckmann and R. Shine, the Americas, and presumably brought to unpublished data). Water rats (Hydromys Australia in 1935 with the original founding chrysogaster) and introduced black rats (Rat- toads (Dubey and Shine 2008). Although tus rattus), however, frequently consume this nematode might cause significant im- cane toads (Cassels 1966; Fitzgerald 1990). pact if it were to infect native frogs, Dubey and Shine (2008) found no cases of R. killing native wildlife by lethal pseudosphaerocephala infection in native frogs toxic ingestion in northeastern Queensland. Preliminary trials by Pizzatto and Shine (unpublished data) Because Australian predators have no evo- indicate that the toads’ nematode can in fact lutionary history of exposure to bufonid infect native frogs, at least under laboratory toxins, many of these animals, as well as conditions, but that the infection generally domestic cats and dogs, die after mouthing does not persist in these frogs or cause major or ingesting cane toads (Lever 2001). De- viability decrements. Concern also has been spite occasional reports of toad toxins expressed about the cane toad invasion car- “polluting waterbodies with their poisons” rying chytrid fungus to western parts of Aus- (e.g., Freeland 1984), the only evidence of tralia, but PCR tests of 100 adult cane toads impact involves poisoning of predators via (50 from Cairns, 50 from Normanton) pro- direct ingestion of toads. The strong allom- vided no evidence of any chytrid infection etry of toxin content in cane toads means (B. L. Phillips et al., unpublished data). Any that toad body size strongly influences the impact of parasite transfer, if it does occur, is risk posed to a predator. As a result, the likely to be several years post-invasion, as vertebrates most at risk from cane toads invasion-front populations of toads typically are animals that are large enough to con- lack parasites (Phillips et al. 2010a). sume an adult toad. In practice, this allometry results in vulnerability being greatest predation by vertebrates on toads for large-bodied species of frog-eating snakes (especially elapids), lizards (espe- Cane toads provide a novel additional cially varanids, but also the bluetongue food source for some vertebrates. Snakes, skink), freshwater crocodiles, and northern such as the keelback (Tropidonophis mairii), quolls (Burnett 1997; van Dam et al. 2002; can consume small cane toads without dy- Phillips et al. 2003). Because these are the ing, but nonetheless show clear evidence most dramatic direct impacts of cane toad of ill effects (see Ingram and Covacevich invasion, I will review the data on vulnerabil- 1990; Llewelyn et al. 2009, 2010a). Slatey- ity of species within each vertebrate lineage grey snakes (Stegonotus cucullatus) some- in more detail below. times consume cane toads in the field (Llewelyn et al. 2009, 2010a) and are also relatively resistant to their toxins (Phillips Lethal Toxic Ingestion et al. 2003). Although snakes may take sub- Direct poisoning of predators is the most stantial numbers of small toads under cer- significant pathway by which cane toads tain conditions, this seems to be uncom- impact the Australian native fauna. A mon; given a choice, keelback snakes broad range of native terrestrial vertebrate prefer native frogs to toads (Llewelyn et al. species are known to have died after 2010a). Some birds, such as corvids (Bek- mouthing or ingesting cane toads. For ex- ker 1985; Caswell 1987; Barraclough 1988; ample, Lever (2001) notes 27 native spe- Donato and Potts 2004), egrets (McKilli- cies of terrestrial vertebrates (as well as gan 1984), and raptors (Mitchell et al. domestic cats and dogs, and humans) in 264 THE QUARTERLY REVIEW OF BIOLOGY Volume 85 which individual predators are known to (1992) recorded meat ants (Iridomyrmex pur- have died. Unfortunately, detailed quanti- pureus) consuming small toads in Queens- fication is lacking in most cases. land, while Ward-Fear et al. (2010a,b) reported high rates of predation by bearded aquatic invertebrates meat ants (Iridomyrmex reburrus) on metamorph cane toads in the Darwin area. The toxic effects of cane toad eggs and tadpoles on Australian aquatic invertebrate freshwater fish predators are poorly studied. Because cane toads breed both in temporary and perma- Laboratory studies suggest that most species nent waterbodies (Crossland and Alford of Australian freshwater fish either ignore the 1998; Freeland 2000), many aquatic inverte- early life-history stages of cane toads, or taste brate predators will encounter the early life and reject them without ill effects (Pearse stages of cane toads. Crossland and Alford 1980; Hearnden 1991; Crossland and Alford (1998) found that cane toad eggs, hatch- 1998). Although Alford et al. (1995) con- lings, and/or tadpoles were toxic to native cluded that “no native fish species were ad- snails (Austropeplea lessoni), water beetle lar- versely affected by aquatic stages of B. marinus,” vae (Dytiscidae), backswimmers (Notonecti- more detailed work has shown that consump- dae), and leeches (Hirudinea); individuals tion of the eggs or tadpoles of cane toads is of these taxa died after consuming cane toxic to the fly-specked hardyhead (Cratero- toads. Mortality of the snails (100%) and cephalus stercusmuscarum) (Crossland and Al- leeches (60%) was high, but the other taxa ford 1998), the banded grunter (Amniataba per- showed intraspecific variation in their sus- coides), the spangled grunter (Leiopotherapon ceptibility to cane toads. Other aquatic inver- unicolor) (Hearnden 1991), the purple-spotted tebrate predators, including water scorpions gudgeon (Mogurnda adspersa) (Pearse 1980), (Nepidae), giant water bugs (Belastomati- glassfish (Family Ambassidae), western rain- dae), dragonfly larvae (Odonata), freshwater bowfish (Melanotaenia australis), and black cat- prawns, crabs, and crayfish (Decapoda), ei- fish (Neosilurus ater) (Wilson 2005). Most fish ther consumed cane toad eggs or tadpoles appear to detect the noxious nature of cane without apparent ill effect, or killed cane toad early life stages, and avoid them as a result toads but consumed none or little of the (Licht 1968; Lawler and Hero 1997; Crossland carcass. Similarly, fishing spiders (Dolomedes 2001; Nelson et al. 2010a,b). Barramundi sp.) also kill and partially consume toad tad- (Lates calcifer) and sooty grunters (Haphaestus poles without incurring any apparent detri- fuliginosus) that attacked tadpoles spat them ment (M. R. Crossland, unpublished data). out almost immediately, but displayed signs of Some water beetles and crustaceans (fresh- distress by shaking their heads from side to side water prawns, shrimps, crabs, and crayfish) for up to a minute, while opening and closing can live on cane toad eggs and tadpoles for their mouths (Crossland 2001). These barra- at least four weeks without apparent ill effect mundi rapidly learned to avoid toad tadpoles, and may be potential predators of cane toads and some individuals continued to do so one in nature (M. R. Crossland, unpublished day after their previous encounter (Crossland data). 2001). At least one other fish species—the northern trout gudgeon, Mogurnda mo- terrestrial invertebrates gurnda)—also rapidly learns to avoid toad tad- Some terrestrial invertebrates are success- poles (Nelson 2008; Nelson et al. 2010a,b). ful cane toad predators. Martin and Free- Attributing sources of mortality is difficult un- land (unpublished data) observed a centi- der field conditions, but Grace and Sawyer pede preying on a 60 mm sub-adult cane (2008) reported a case of fish mortality associ- toad without apparent ill effects, while van ated with toad breeding. This involved approx- Beurden (1980a) reported predation by imately 70 spangled perch (Leiopotherapon large wolf spiders and ants on cane toad unicolor) presumably killed by ingesting metamorphlings. Clerke and Williamson toad eggs or tadpoles; black-banded rain- September 2010ECOLOGICAL IMPACT OF INVASIVE TOADS 265 bowfish (Melanotaenia nigrans), however, ability than other native tadpoles to detect were unaffected. the noxiousness of cane toads, or that L. Experimental studies indicate that for most rubella tadpoles do not feed on the early life fishes, toad eggs are more likely to be lethal stages of other anuran species (Crossland than toad tadpoles (Greenlees and Shine 1998b; Crossland and Alford 1998; Cross- 2010). The taxa tested included ambassids land and Azevedo-Ramos 1999). Behavioral (Ambassus macleayi), eleotrids (Mogurnda mo- responses suggest that native frog tadpoles gurnda), melanotaenids (Melanotaenia australis, are unable to detect the toxicity of cane toad M. splendida), plotossids (Neosilurus hyrtlii), and eggs; tadpoles did not avoid cane toad eggs, pseudomugils (Pseudomugil gertrudae, P. incon- but grazed on egg strings until they had pen- spicuus, P. tenellus). Fortunately, the brief dura- etrated the gelatinous string and consumed tion of the egg stage and the restriction of eggs the fertilized eggs inside, after which they to shallow pond edges reduce the rates at always died (Crossland and Alford 1998). which eggs are encountered by native preda- However, the major predator of cane toad tors. Predators encounter toad tadpoles more eggs and tadpoles is the cane toad tadpole often, but the tadpoles are less toxic than eggs, itself (Hearnden 1991). especially late in development (Hayes et al. More recently, anuran species from the 2009), and their toxicity is more easily detect- Northern Territory have been evaluated in able by predators (Greenlees and Shine 2010). this respect. Field and laboratory studies have clarified interactions between native amphibians frog tadpoles and cane toad eggs on the Crossland and Alford (1998) investigated Adelaide River floodplain. Of 14 species the effects of cane toad eggs, hatchlings, and tested in the laboratory (Opisthodon orna- tadpoles on tadpoles of native frog species tus, Limnodynastes convexiusculus, Cyclorana from the Townsville and Cape York Penin- australis, C. longipes, Litoria bicolor, L. caerulea, sula areas. Consumption of cane toad eggs L. dahlii, L. microbelos, L. nasuta, L. pallida, L. was always fatal to native tadpoles, but the rothii, L. rubella, L. tornieri, and Uperoleia litho- number of tadpoles that fed on eggs varied moda), most attempted to eat toad eggs, and among species: Litoria bicolor - 100%, L. all that did died after doing so (the excep- rubella - 30%, L. infrafrenata - 100%, L. nigro- tion to this being tadpoles that were too frenata - 60%, L. alboguttata - 100%, and small to physically penetrate the toad’s egg Opisthodon ornatus - 90%. Similarly, consump- capsule) (Crossland and Shine 2010). The tion of cane toad hatchlings was always fatal only tadpoles that survived were those that to native tadpoles and predation rates varied: did not consume toad eggs, and almost all of L. infrafrenata - 80%, L. alboguttata - 20%, and these animals consumed eggs—and conse- O. ornatus - 10%. Litoria rubella and L. nigro- quently died—when re-tested (Crossland frenata did not consume cane toad hatch- and Shine 2010). In the field, toad spawning lings. No native frog tadpoles ate live cane was followed by mass mortality of tadpoles of toad tadpoles (Crossland and Alford 1998), several native frog species (Crossland et al. but many frog tadpoles will feed on the car- 2008). Counts of species abundance before casses of dead tadpoles; exposure to dead versus after those kill events suggested cane toad tadpoles killed 60% of L. albogut- non-random mortality rates; for instance, tata, 50% of Cyclorana brevipes, and 90% of O. Litoria rothii tadpoles appeared to be par- ornatus (Crossland and Azevedo-Ramos ticularly vulnerable. 1999). In contrast, L. rubella tadpoles (0% The toxicity of cane toads shifts rapidly mortality) and L. gracilenta tadpoles (20% during the course of the animal’s early life mortality) were less likely to feed on dead history, as demonstrated by analyses of bufa- cane toad tadpoles, although they readily ate dienolide content as well as the mortality dead native tadpoles (Crossland and Aze- rates of native tadpoles that consume cane vedo-Ramos 1999). The fact that L. rubella toads during those life-history stages (Hayes rarely ate cane toad eggs, hatchlings, or tad- et al. 2009). The toxin is contained in the poles suggests that they either have a greater cane toad ovum rather than in the jelly coat 266 THE QUARTERLY REVIEW OF BIOLOGY Volume 85

(Crossland and Shine 2010); therefore, virtually 100% of tadpoles (Limnodynastes convexiusculus and Litoria rothii) that consumed toad eggs rather than the less toxic later- stage tadpoles died (Hayes et al. 2009). Anurans are also at risk after metamorphosis, if they attempt to consume small terrestrial-phase cane toads. Although van Dam et al. (2002) speculated that only the largest frog species would be at risk in this way, recent studies show vulnerability even in smaller taxa such as Limnodynastes convexiusculus (Greenlees et al. 2010a). In laboratory trials, fatal toxic ingestion of toad metamorphs by Cyclorana australis, C. alboguttata, C. novaehollandiae, Litoria caerulea, and L. dahlii has also been recorded. De- spite anecdotal reports to the contrary, Litoria dahlii was no more resistant to the toxins of cane toads (in eggs, tadpoles, or metamorphs) than were other local frog species (Shine et al. 2009). Supporting these conclusions, Grace and Sawyer (2008) re- Figure 3. CROCODYLUS JOHNSTONI Attacking a Cane ported mortality of 11 post-metamorphic L. Toad dahlii and two Limnodynastes convexiusculus at Invasive cane toads are lethally toxic to many native a site where cane toads had bred. In the Fogg predators that attempt to consume them. This pho- Dam area, we found recently killed speci- tograph was taken at Fogg Dam, Northern Territory, mens of L. dahlii (n ϭ 3), Cyclorana australis and shows a freshwater crocodile (Crocodylus john- (n ϭ 2), and Limnodynastes convexiusculus stoni) in the act of attacking a toad. Note the milky (n ϭ 4), apparently as a result of toad inges- defensive secretions from the toad’s parotoid tion (Greenlees et al. 2010a and unpub- gland. (Photograph by G. P. Brown, and used with permission.) lished data), as some of these dead frogs contained single metamorph toads in the gut. The impact of toads on terrestrial-phase ported that C. johnstoni actively hunts and frogs can be reduced by rapid aversion learn- ingests cane toads, and Letnic and Ward ing, as documented in Limnodynastes convexi- (2005) provided photographs of this behav- usculus, Cyclorana australis, and Litoria dahlii ior (see also Figure 3). Smith and Phillips (Nelson 2008; Nelson et al. 2010a,b; Green- (2006) showed that physiological resistance lees et al. 2010a,b). Nonetheless, the inva- to toad toxins is lower in freshwater croco- sion of cane toads may exert selection on diles than in saltwater crocodiles, and Doody life-history traits such as breeding seasonality et al. (2009) reported the deaths of some and size at metamorphosis in native anurans freshwater crocodiles in the Daly River due (Greenlees et al. 2010a). to toad ingestion, but to no overall effect on crocodile populations. In a study primarily crocodiles focused on population-level impact, Letnic Covacevich and Archer (1975) reported et al. (2008) reported 34 dead freshwater that saltwater crocodiles (Crocodylus poro- crocodiles in the Victoria River and attrib- sus) could ingest cane toads with impunity. uted these deaths to cane toad ingestion. By contrast, there are reports of freshwater The evidence included a wave of crocodile crocodiles (Crocodylus johnstoni) dying after deaths moving upriver and coinciding with mouthing or ingesting cane toads (Begg et the cane toad invasion front, as well as the al. 2000) (Figure 3). Freeland (1990) re- discovery of toads in the stomachs of some of September 2010ECOLOGICAL IMPACT OF INVASIVE TOADS 267 the dead crocodiles (Letnic et al. 2008). jelly coat of the egg, which lacks toxicity Consistent with Doody et al.’s (2009) study, (Crossland and Shine 2010), may mask the Letnic et al. (2008) reported much lower poisons sequestered within the ovum, such rates of toad-induced mortality for freshwa- that the predator swallows the entire egg ter crocodiles in the Daly River. mass before detecting its toxic nature.

turtles lizards Turtles sometimes ingest cane toads with- The area of Australia invaded by cane out ill effects, whereas other predation toads contains high densities and species attempts—even by the same turtle spe- diversities of lizards belonging to several cies—can be fatal. Covacevich and Archer phylogenetic lineages. The impact of cane (1975) reported that a long-necked turtle toads across those lineages is highly non- (Chelodina sp.) ate a dead cane toad with- random; fatal poisoning is an important out becoming ill, and Crossland and Alford pathway of impact only for species within (1998) reported that saw-shelled turtles (El- three lineages: the Varanidae (goannas), seya latisternum) and Krefft’s river turtles the Scincidae (skinks), and the Agamidae (Emydura krefftii) consumed cane toad tad- (dragons). Varanids often grow to large poles but were unaffected. Hamley and body sizes and readily attack large prey Georges (1985) found toads in the stom- items (Losos and Greene 1988); hence, achs of four field-collected saw-shelled tur- they may attack a large cane toad more tles from the Brisbane area. The same readily than would a lizard from most authors fed toads to this turtle species in other lineages. Thus even relatively small captivity without noting any ill effects, and varanids may be at risk from cane toad even maintained two specimens for four invasion, whereas the impact appears to months or more on a diet composed ex- fall heavily only on a few large-bodied spe- clusively of cane toads. By contrast, there cies within the Scincidae and Agamidae. are unsubstantiated reports from Aborigi- Smith and Phillips (2006) directly measured nal communities on Groote Eylandt and toxin resistance of a range of lizard species the Borroloola region that cane toads have and found that most were highly sensitive to killed “goanna, bluetongue, long-necked cane toad toxins. turtle, geese and everything” (Evans 1999: Varanid lizards (goannas) often die after 16). Similarly, community members from ingesting or mouthing cane toads (Lever Beswick and Burunga, south of Katherine, 2001; van Dam et al. 2002; Phillips et al. included barramundi and long- and short- 2003). There have been many reports of necked turtles in their list of species goannas found dead in the field soon after affected by toads (Begg et al. 2000). Scien- the arrival of cane toads, and the evidence tific studies of toad impacts on taxa such as for causation—a freshly-dead goanna with barramundi and geese suggest that these a toad in its mouth or stomach—is unam- statements may be in error, whereas the biguous. In the wet-dry tropics of the North- statements about goannas and bluetongue ern Territory, most of these cases have lizards are consistent with other evidence. involved the large yellow-spotted monitor Hence, it is difficult to evaluate the reliabil- Varanus panoptes (e.g., Doody et al. 2009; ity of statements about turtle mortality. Ujvari and Madsen 2009; G. P. Brown, per- E. Kruger (personal communication) has sonal communication). Medium-sized va- reported dead long-neck turtles (Chelodina ranid species may also be at risk, based on rugosa) in waterbodies used by cane toads for population-level declines reported for breeding. In our experimental studies, long- Varanus gouldii (Freeland 2004; T. Madsen, necked turtles (Chelodina rugosa) seized but personal communication), V. mitchelli, and rejected toad tadpoles, and survived; how- V. mertensi (Doody et al. 2009), although V. ever, consumption of toad eggs was fatal scalaris remains common in areas of high (Greenlees and Shine 2010). As is the case toad abundance near Darwin (T. Madsen, for other aquatic predators such as fish, the personal communication). Laboratory tri- 268 THE QUARTERLY REVIEW OF BIOLOGY Volume 85 als have confirmed that small varanids, risk from cane toads are frog-eating species such as V. acanthurus, V. glauerti, and V. that cannot tolerate toad toxins and that glebopalma, readily consume small toads frequently encounter toads, can swallow and are likely to die if they do so (R. Shine toads large enough to be fatal, and whose et al., unpublished data), but the restric- geographic distribution overlaps greatly tion of small toads to riparian habitats for with that of the cane toad (Sutherst et al. much of the year may reduce their encoun- 1995; Cogger 2000). A risk assessment ter rates with non-riparian varanid species. based on these criteria identified 49 snake Most species of scincid lizards exhibit species potentially at risk from the toad relatively small adult body sizes. Given the invasion (Phillips et al. 2003). Of these, 26 strong allometry in toxin content of cane were likely to have their range totally en- toads (Phillips and Shine 2006a,b), such compassed by that of the toad (under 2030 small predators are unlikely to be at risk climate change), and three have already from consuming a cane toad large enough had their range totally encompassed. Thus, to kill them. Doubtlessly, occasional indi- the toad invasion constitutes a potential viduals of smaller species eat toads that are threat to 70% of the Australian colubrid large enough to prove fatal; we have re- snakes (7 of 10 species), 40% of the py- corded this outcome in one laboratory en- thons (6 of 15), and 41% of the elapids (36 counter between a cane toad and a striped of 87). Nine of these “at risk” species are skink (Ctenotus robustus: M. Greenlees et al., currently recognized as threatened on a unpublished data). One scincid species at federal or state level (Cogger et al. 1993). considerable risk is the bluetongue skink Despite this, recent work on the feeding (Tiliqua scincoides intermedia), a large omni- responses of snakes in captivity when of- vore that is widely distributed across tropical fered cane toads of edible size is encourag- Australia. Captive individuals of this species ing. Many species that had been previously from northwestern Australia readily con- identified as potentially at risk by Phillips sume small toads and die as a result (Price- et al.’s (2003) study were recently observed Rees et al. 2010); the same outcome has to be reluctant to take cane toads as prey, been observed in free-ranging animals as or released the prey immediately after well (Sanson 2010; S. Price-Rees, personal striking it and, thus, survived the encoun- communication). ter (R. Shine et al. unpublished data). Most dragon lizards (Agamidae) are un- Also, colubrids and pythons appear to be likely to be directly affected by cane toads less at risk than elapid (venomous) snake (Griffiths and Christian 1996; Cogger 2000). species (M. Greenlees et al., unpublished Several years after cane toads reached Lawn data), as, in the laboratory, we have re- Hill National Park, rangers reported remov- corded high mortality rates in death ing cane toads from the mouths of frillneck adders (Acanthophis praelongus), black whip lizards that frequented the camping area snakes (Demansia papuensis), and king brown (van Dam et al. 2002). These lizards were snakes (Pseudechis australis) (M. Greenlees distressed but not dead; however, over the et al., unpublished data). Most of the vari- next few years, they apparently disappeared ation in survival rates reflects snake behav- from the area. In trials with field-collected ior (i.e., tendency to attack versus to ignore animals in captivity, frillneck lizards did not a toad); however, two colubrid species are eat small toads, but readily took insects in- known to have high physiological toler- stead (Ujvari et al. 2010). ance to toad toxins (Stegonotus cucullatus and Tropidonophis mairii) (Phillips et al. snakes 2003). In laboratory trials, individuals of many Laboratory experiments have provided species of frog-eating snakes died after in- detailed data on a particularly high risk gesting tiny amounts of cane toad toxin species—the death adder Acanthophis (Fearn 2003; Phillips et al. 2003; Phillips praelongus. Individuals of all sizes and both and Fitzgerald 2004). The snakes most at sexes, and from woodland as well as flood- September 2010ECOLOGICAL IMPACT OF INVASIVE TOADS 269 plain habitats, consumed cane toads in the extensive review of available data suggests a laboratory and died after mouthing or in- less gloomy scenario for toad impacts on gesting them (Webb et al. 2005; Phillips et Australian birds via lethal toxic ingestion. al. 2010b). Death adders are especially vul- Beckmann and Shine (2009) reviewed all nerable because they are ambush foragers available literature on this topic, and con- that use caudal luring to attract prey within cluded that cane toads appear to have mini- striking range (Cogger 2000). The lure of mal impact on Australian birds. This ability adult northern death adders is black, and it to survive toad invasion appears to result resembles a small wriggling invertebrate from a widespread physiological tolerance of (Webb et al. 2005). Because cane toads are bufotoxins, perhaps reflecting close genetic attracted to black coloured wriggling ob- ties between Australian birds and conspecif- jects (Ingle and McKinley 1978), encounter ics or congeneric taxa in Asia, where many rates between death adders and cane toads bufonid species possess toxins similar to have likely been high since the toad invasion. those of cane toads (Meyer and Linde 1971). Experimental trials confirm that cane toads are readily attracted to the lure of a death mammals adder—indeed, more readily than are native Webb and Glanznig (2004) listed nine frogs (Hagman et al. 2009a). Dead death species of native mammals and two species adders with cane toads in their mouths have of introduced mammals as potentially at been reported from the Mary River region of risk from ingesting cane toads. There are Kakadu National Park (T. Flores, personal many reports of domestic dogs dying or communication) and from the Adelaide becoming ill after mouthing or ingesting River floodplain (Phillips et al. 2010b). toads (Rabor 1952; Knowles 1964; Covacev- ich and Archer 1975; van Beurden 1980a), birds thus hinting that dingos could be at risk. Covacevich and Archer (1975) reported Feral cats and pigs also may be negatively that some individual crows (Corvus sp.) and affected by cane toads (Begg et al. 2000); kookaburras (Dacelo novaeguineae) died af- feral pigs, in particular, frequently con- ter mouthing cane toads, whereas other sume anurans (White 2008). Covacevich individuals of the same species consumed and Archer (1975) collated anecdotal re- young toads or road-killed toads and in- ports of dogs, cats, and quolls dying as a curred no ill effects. Van Beurden (1980a) consequence of attempting to ingest toads. recorded deaths of kookaburras, the little The ghost bat (Macroderma gigas) is thought bittern (Ixobrychus minutus), and the black to eat native frogs (ANPWS/DEST 1991) bittern (Ixobrychus flavicollis) after ingesting and, therefore, may also eat cane toads. Al- juvenile cane toads. At least seven native bird though other bat species may eat native frogs species can eat cane toads successfully, either and potentially cane toads as well, there is because they eat only the non-toxic parts of little information on this topic. the toad or because they are immune to the In captivity, some native rodents (Melomys toxins (Covacevich and Archer 1975; Free- burtoni, Rattus colletti, R. tunneyi) readily at- land 1987; Seabrook 1991; Mitchell et al. tacked and killed small toads, but did not ap- 1995). pear to suffer any ill effects from ingesting Using data on bird foraging habits and these anurans (J. Webb et al., unpublished diets, Dorfman (1997) predicted that 76 data). Other rodents (Mus domesticus, Pseu- bird species from Kakadu National Park domys nanus, Zyzomys argurus) showed no in- were potentially under threat from cane terest in cane toads as prey. Small dasyurid toads. Catling et al. (1999) listed several marsupial species, such as planigales additional species that were potential con- (Planigale ingrami, P. maculata) and dun- sumers of cane toads, and a risk assessment narts (Sminthopsis virginiae), attacked the by van Dam et al. (2002) suggested that 66 first toad they encountered, became ill as a species of birds might be at risk from cane result, and were reluctant to attack toads toads in Kakadu National Park. A more thereafter (Webb et al. 2008 and unpub- 270 THE QUARTERLY REVIEW OF BIOLOGY Volume 85 lished data). Relatively few of these narios are easy to envisage. For example, dasyurids died as a result of attempting to Waterhouse (1974) speculated that preda- ingest toads (Webb et al. 2008). Planigales tion on dung beetles by cane toads could (P. maculata) from northeastern Queens- reduce rates of dung breakdown, with land, where toads have been abundant for flow-on consequences to many aspects of several decades, reacted similarly to cane ecosystem function. Freeland (2004) sug- toads as did naı¨ve animals from the North- gested that toad-induced shifts in host- ern Territory, but they appear to have parasite ecology could destabilize anuran evolved a higher physiological resistance to community structure. More important, how- toad toxins (Llewelyn et al. 2010b). ever, may be the mortality of top-order pred- The species of mammal most often im- ators that are killed when they attempt to plicated as a victim of toad invasion is the ingest these highly toxic anurans. In an ex- northern quoll, Dasyurus hallucatus. Data perimental study, Crossland (2000) demon- on the fate of quolls radio-tracked in the strated that toxic effects of cane toad eggs on field have been interpreted to suggest that populations of predatory native frog tad- lethal ingestion of cane toads is responsi- poles resulted in reduced predation on the ble for the local extinction of populations eggs of other native frog species, and, as a of northern quolls from the Mary River result, larval survival of these species was in- region of Kakadu National Park (Oakwood creased. As I will discuss below, toad invasion 2003a,b). However, only four of 14 quolls can massively reduce numbers of predators whose fates were recorded during Oak- such as varanid lizards, elapid snakes, fresh- wood’s studies were killed by ingesting water crocodiles, and northern quolls. The toads, suggesting that other pressures may consequent reduction of predator pressure also threaten quoll populations. A more on the usual prey of these species may thus recent monitoring study of radio-tracked change the abundance of particular species quolls confirms that many quolls die from and the composition of ecological assem- ingesting toads (O’Donnell et al. 2010). In blages, as may have occurred for small eight cases for which cause of mortality agamid lizards following toad-induced mor- could be confirmed, seven resulted from tality of larger varanids (Doody et al. 2009). ingestion of cane toads (the other quoll A sudden release of predation pressure on was killed by a feral cat). Quolls encoun- specific life-history stages—such as turtle tered and were killed by toads even in areas eggs (Doody et al. 2006)—also might gener- and at times when few toads were seen by ate shifts in age structure within a species. observers (O’Donnell et al. 2010). This re- Indirect effects occur in two distinct ways sult is a discouraging one, hinting that even (Wootton 1994). “Interaction chains” oc- low densities of cane toads may threaten cur when one species changes the abun- quoll survival. dance of another and indirectly affects a third species that directly interacts with Indirect Impacts them both. For example, an introduced The information above all relates to more leafhopper caused a population expansion or less direct impacts of cane toads—that in a parasitoid wasp, thus increasing rates is, outcomes of encounters between toads of predation on a native leafhopper (Settle and native fauna. Toad invasion also may and Wilson 1990). “Interaction modifica- affect native species indirectly, via changes tions,” on the other hand, occur when one to sympatric species (Pace et al. 1999; Mack species affects the interaction between two et al. 2000; Roemer et al. 2002). For exam- other species; for example, a shelter plant ple, if cane toads directly reduce the abun- can modify a predator-prey interaction by dance of insects, they may thereby reduce providing the prey with cover and decreas- the feeding rates of native insectivores. ing its vulnerability to predation. An inter- This relatively simple case is included un- action modification also can occur when der the category of “competition” in the one species affects another via a change in discussion above, but more complex sce- the behavior of a third species. Behavioral September 2010ECOLOGICAL IMPACT OF INVASIVE TOADS 271 shifts like this might have major implica- predators such as varanid lizards (Doody et al. tions for conservation (Curio 1996; Buch- 2006). The magnitude of such impacts will de- holz 2007). For example, in response to pend upon the pre-toad densities of these introduced brown trout (Salmo trutta)in predators, which will be massively dependent New Zealand streams, mayfly nymphs (Ne- upon human activities (e.g., provision of extra sameletus ornatus) reduce the time spent food at campgrounds, subsistence hunting, at- grazing, and this behavioral change subse- tacks by domestic dogs). As a result, toad im- quently increases algal biomass (McIntosh pacts may depend sensitively upon the details and Townsend 1996; Townsend 1996). of prior disturbance regimes imposed by hu- Predicting the indirect effects of invasive man activities. species on ecosystems is extremely difficult. For example, the introduction of pigs Population-Level Impacts to the Californian Channel Islands caused There is a substantial difference between an increase in the abundance of golden observing mortality of individual predators eagles, which then preyed on a major pred- and inferring that such cases are common ator—island foxes—and caused their pop- enough to affect population numbers or vi- ulation to decline dramatically (Roemer et ability (Doody et al. 2009). Many processes al. 2002). Feral house-mice on offshore is- that kill native animals do not threaten pop- lands have complex effects on dasyurids and ulation persistence; abundance and distribu- skinks, and suburban cats simultaneously in- tion are regulated by a host of factors (e.g., crease mortality rates of adult birds while weather conditions, food supply, disease), es- decreasing rates of predation by rodents on pecially in stochastic environments, and in- bird nests (Dickman 2007). Even more sub- creases in mortality due to toad invasion may tle effects can have significant consequences; have little effect on overall population num- for example, if the invasive species modifies bers. Indeed, both ecological theory and ma- the behavior of a native taxon, that change nipulative experiments suggest that, in some can reverberate throughout the system. cases, additional mortality may increase Webb et al. (2008) showed that marsupial rather than decrease population numbers carnivores (Planigale maculata) exposed to (Crossland et al. 2009). This paradoxical re- metamorph toads rapidly began to avoid not sult reflects the role of density-dependent only toad metamorphs, but also frog meta- processes. For example, if tadpoles of native morphs. That aversion to native frogs de- frogs occur at such high densities that com- creased through time (i.e., the predators petition for food resources greatly reduces learned to distinguish noxious prey from ed- growth rates and survival, then any extrinsic ible prey), but the toad-averse planigales in- mortality source—such as cane toads—may vestigated such potential prey more carefully actually benefit the surviving tadpoles by lib- prior to seizing it, plausibly increasing the erating them from competitive effects (Al- chances of escape for frog metamorphs ford et al. 1995). In keeping with this in- (Webb et al. 2008). Similarly, Nelson et al. ference, the addition of cane toads to (2010a) reported that exposure to toad tad- experimental ponds can increase the body poles induced native fishes (Mogurnda mo- sizes of metamorphosing frogs (Crossland et gurnda) to shift their foraging efforts towards al. 2009). However, the magnitude and di- non-tadpole prey, thereby reducing preda- rection of such effects depend upon the tion on native tadpoles and increasing pre- numbers and times of breeding of toads and dation on alternative prey types. frogs, so generalities about the impact of Cane toads actively select human-modified toads in this situation may prove difficult to (i.e., degraded) habitats (e.g., Zug and Zug identify (Crossland et al. 2009). 1979; Lever 2001), and thus their impacts are Another important complication is that tightly intertwined with other anthropogenic the direct impact of cane toads on any modifications. For example, one of the major given native species is likely to be accom- impacts of cane toads likely involves flow-on panied by a suite of indirect impacts, me- effects brought about by the death of native diated via the toads’ impact on other taxa 272 THE QUARTERLY REVIEW OF BIOLOGY Volume 85 with which the target species interacts. For 2006; Brown and Shine 2007), causing pop- example, the invasion of cane toads kills ulation densities of some species to fluctuate (i.e., fatally poisons) many tadpoles of na- enormously from one year to the next, espe- tive frogs, and also a significant number of cially in the case of taxa such as frogs with frog metamorphs (Crossland et al. 2008). high reproductive rates. For example, an- As a result, we might expect frog popula- nual numbers of some native frog species at tions to decline following toad invasion. an intensively studied area of the Adelaide However, cane toads also kill (fatally poi- River floodplain frequently shifted by more son) many frog-eating predators—indeed, than 50% from one year to the next, even from a predator’s perspective, a cane toad before cane toads arrived at the site (Cross- is simply a booby-trapped frog. Therefore, land et al. 2008). Some of the species most frog populations will be increased by the likely to be affected by cane toad ingestion death of these top predators, at the same are large predatory reptiles and mammals, time as frogs are being killed by ingesting with much lower reproductive rates—thus, toads. The change in the population size of we might expect greater temporal consis- frogs will, thus, represent the balance be- tency in population densities. Unfortunately, tween these two effects (plus others, such such taxa also tend to be rare and to be as competition-reduction). In such a situa- highly mobile, so that robust population es- tion, frog numbers might well increase af- timates are almost impossible to obtain. ter cane toad invasion, despite high levels Even if we surmounted these obstacles of direct mortality inflicted upon the frogs and documented a strong correlation be- by interactions with toads. tween cane toad arrival and a reduction in Given that we cannot infer population- the abundance of some native species, that level effects from data on mortality of correlation provides only weak evidence of individual predators, we are left with the a causal connection. Changes in land-use challenge of actually measuring population- patterns, fire regimes, and the presence of level effects directly. This is a difficult chal- invasive plants, animals, and pathogens all lenge; even with a relatively stable “closed” affect native fauna, and such changes are population, detecting a 20% increase or de- occurring rapidly throughout much of the crease (for example) from one year to the cane toad’s Australian range. Thus, a de- next requires very large sample sizes or high cline in some species coincident with toad capture rates per individual (Caughley 1977; arrival might well be due to some third, Woinarski et al. 2004). If individuals can dis- unappreciated factor. Northern quolls pro- perse from one population to the next, the vide a clear example of this ambiguity. challenge becomes much harder because There is no doubt that many of these mar- dispersal—immigration and emigration— supials die when they attempt to ingest may contribute more to changing densities cane toads; radio-tracking studies docu- than any other process acting within the pop- ment this link clearly (O’Donnell 2009; ulation under study. Furthermore, if we add O’Donnell et al. 2010). However, northern in high reproductive rates, spatial heteroge- quoll populations have been in decline for neity, and environmental stochasticity, the many years, and their virtual disappear- logistical difficulties of assigning causation to ance from many areas long predates the any year-to-year change in densities become arrival of cane toads (Braithwaite and Grif- incredibly challenging. Unfortunately, much fiths 1994). Separating out the impact of of the area colonized by cane toads, and toads from the impact of other threatening many of the species affected, have ecological processes is a Herculean task in these com- attributes that preclude exact enumeration. plex and dynamic systems. Year-to-year variation in wet season rainfall One clear implication of these sources of patterns has immense influence on repro- ambiguity is that any scientific studies that ductive rates, growth rates, and survival rates wish to document impacts of cane toads on of animals in the wet-dry tropics (e.g., Mad- native fauna must include replication on a sen and Shine 1999, 2000; Madsen et al. scale that allows cane toad impact to be September 2010ECOLOGICAL IMPACT OF INVASIVE TOADS 273 differentiated from other sources of varia- the value of such opinions. Strong public tion. For example, simply surveying faunal revulsion against cane toads may encourage abundance in toad-colonized versus non- reports of ill-effect even if the actual obser- colonized sites is fraught with problems vations are, at best, equivocal. Even in cases because (a) habitat variables may explain where the data show a consistent decline fol- why some sites have been colonized by lowing toad invasion, causation is rarely toads and others have not, and those same clear-cut. Paradoxically, some of the most variables may well affect the composition readily interpretable data involve cases of the native fauna; and (b) species rich- where densities of native species are not af- ness and abundance of native taxa (and fected by cane toad invasion. In these cases their ease of detection by the observer) (such as for native frogs, as I will discuss vary in complex ways across the landscape, below), parsimony suggests that we accept independent of any effects of toads, so that the null hypothesis that cane toads have no the sites to be compared must be closely significant impact on these native taxa, at matched for all variables that might influ- least in terms of the timescale for which data ence those native species. In practice, this are available, which is typically quite short- is a major challenge, especially when com- term. bined with temporal variation in faunal composition and/or detectability. invertebrates The end result of these factors is that we High densities and feeding rates of cane know relatively little about the population- toads have suggested to many observers level impacts of cane toads on native fauna. that toad invasion may significantly deplete There are many anecdotal reports of de- invertebrate populations (e.g., Pippet 1975); clines in particular species that coincide however, data to support this prediction with the time of cane toad invasion, but are weak or non-existent. Catling et al. typically few data have been provided. For (1999) reported that the abundance of example, Schultze-Westrum (1970) reported beetles was lower in sites colonized by cane that in Papua New Guinea, increases in cane toads as compared to toad-free sites, but toad populations in savannas coincided with variances were high—indeed, the highest declines in native reptile populations, includ- density of beetles was recorded at a toad- ing several species of gecko, skink, and other colonized site. Overall, Catling et al. (1999) terrestrial species that shelter under logs and concluded that there was little evidence that rocks. Following the toad invasion, cane cane toads caused a significant adverse effect toads became the dominant animal found on the diversity and abundance of many of under logs and rocks, thus suggesting that the native fauna examined, and that the competition for shelter sites was involved in apparent effects recorded may well have the decline of these native species (Schultze- been statistical artifacts. Freeland (2004) Westrum 1970). Many similar examples can highlighted further statistical problems asso- be cited from Australia. For instance, ciated with inferring effects based on this Breeden (1963) reported declines in snakes, kind of abundance data, and pointed out monitor lizards, frillnecked lizards (Chlamy- numerous inconsistencies in the results of dosaurus kingii), and quolls following the ap- surveys by Catling et al. (1999) as compared pearance of toads in north Queensland, and to those of Watson and Woinarski (2003a,b) Pockley (1965) and Rayward (1974) claimed in terms of which species were affected by that populations of snakes, goannas, and toads. Freeland (2004) also noted the con- birds had declined following the arrival of tinued presence of some of the purported toads in southeastern Queensland and “negatively impacted” species in areas long northern New South Wales. Unfortunately, after toad invasion. these reports provided no quantitative evi- Freeland (1993, 2004) has suggested that dence. Many of the stories that one hears cane toads may have contributed to a de- about cane toad impact are inconsistent with cline in the proteocephalid tapeworm of the available data, which raises doubts about the python Antaresia maculosa. The inci- 274 THE QUARTERLY REVIEW OF BIOLOGY Volume 85 dence of this parasite in native frogs—an this species was a frequent victim of lethal intermediate host—decreased after toads toxic ingestion of toad eggs in the same invaded, possibly because higher feeding floodplain site: Crossland et al. 2008), and rates allowed cane toads to acquire most of for total anuran biomass and species rich- the tapeworm larvae, thus preventing their ness in that latter area also (Greenlees et subsequent passage to the pythons, as al. 2007). In an innovative and geographi- these pythons rarely consume toads. The cally extensive study, Grigg et al. (2006) parasite’s incidence was higher in invasion- used fixed listening posts to record anuran front populations of cane toads than in calls over long periods both before and populations from long-colonized areas, after toad colonization; however, they re- consistent with the inference of a decline ported no significant effects of cane toads (Freeland et al. 1986b). on the numbers of frog species calling. Watson and Woinarski (2003a,b) com- fishes pared sites in Kakadu National Park the year before versus the year after toad inva- Grace (2008) presented data on numbers sion. No native frog species showed any of barramundi and saratoga in a billabong substantial decline in toad-invaded areas; of the Mary River over an 18-year period instead, there was a tendency for a relative encompassing the arrival of cane toads at increase in species richness and abun- this site. There was no hint in any of these dance following toad invasion. This rela- data of shifts in fish abundance coincident tive increase was most marked for Litoria with cane toad invasion. wotjulemensis, L. rothii, and Uperoleia sp. A broader comparative survey comparing de- amphibians clining versus non-declining anuran spe- The close phylogenetic relatedness be- cies across Australia found a weak overall tween cane toads and native frogs results in association between anuran decline and many broad similarities in ecology, physi- geographic overlap with cane toads, but ology, and behavior; hence, intuition sug- this association disappeared in more de- gests that cane toad invasion might imperil tailed analyses (Murray and Hose 2005). frogs via competition, predation, and disease transfer (see above). Any such effects crocodiles would likely be opposed by positive effects Freshwater crocodiles remain common (or resulting from deaths of frog-eating pred- their numbers have recovered) in areas of ators. Possibly reflecting the balance be- Queensland that have contained high densi- tween these two impacts, and also the ties of cane toads for several decades (Letnic difficulty of accurately enumerating an- et al. 2008). Catling et al. (1999) found no uran densities, research has so far failed to evidence that cane toads affected freshwater detect any effect of cane toad invasion on crocodile populations; however, there are frog populations. A range of methods have numerous anecdotal reports of freshwater been used to look for changes in frog pop- crocodiles dying following cane toad con- ulation densities before versus after toad sumption (van Dam et al. 2002). Letnic et al. invasion, and importantly, all have incor- (2008) and Doody et al. (2009) reported no porated some level of replication in space decline in freshwater crocodiles on the and/or time. Direct counts of anurans Daly River, despite some cases of mortality around waterholes in the Gulf of Carpen- due to toad ingestion. Similarly, Freeland taria during the tropical dry season showed (2004) reported that spotlight surveys in that numbers of Litoria pallida, L. rothii, the Elsey National Park and Nitmiluk Na- and L. rubella did not change significantly tional Park prior to and following the ar- when cane toads invaded the study sites rival of cane toads revealed deaths of some (Freeland and Kerin 1988). The same was crocodiles, presumably due to toad inges- true for numbers of Litoria rothii on the tion, but no overall population-level impact. Adelaide River floodplain (even though In the course of our intensive research at September 2010ECOLOGICAL IMPACT OF INVASIVE TOADS 275

Fogg Dam on the Adelaide River floodplain, spotted goannas (Varanus panoptes) (see my research group has not noted any decline Figure 4), a reduction of 71–97% among in numbers of Crocodylus johnstoni. All of Mitchell’s water monitors (V. mitchelli), and a these cases suggest that mortality rates from decline of 87–93% in Merten’s water moni- toad ingestion are too low to have pop- tors (V. mertensi). Griffiths and McKay (2007) ulation-level effects. similarly reported a major decline in V. The situation is very different, however, mertensi in the Darwin area following toad on the Victoria River (Letnic et al. 2008). arrival: estimates of site occupancy by lizards At this more arid site, the cane toad inva- fell from 95% to 14% over an 18-month pe- sion has caused mass mortality of croco- riod. Radio-tracking of V. panoptes in Kakadu diles, reducing population densities by up National Park suggested a 50–70% mortality to 70%. The divergence in levels of impact due to toad invasion (Holland 2004). Also across sites may reflect differing rates of based upon radio-tracking data, Ujvari and encounter between crocodiles and toads. Madsen (2009) concluded that at least 90% In well-watered areas, toads have many po- of adult male V. panoptes on the Adelaide tential rehydration sites away from the River floodplain were killed by toad inges- river, but, in more arid sites, the toads are tion. In conjunction with abundant evidence forced to enter the river to rehydrate; thus, of goannas being fatally poisoned by toad encounter rates between toads and croco- ingestion (see above), the consistency of diles are higher in more arid conditions these reports provides overwhelming evi- (Letnic et al. 2008). The area currently dence that the invasion of cane toads has being invaded by toads is relatively arid, had serious immediate impacts on goanna and the habitats around Lake Argyle may populations. intensify toad impacts on the very large Encouragingly, some smaller varanid spe- population of C. johnstoni found in that cies (e.g., V. scalaris) still remain common lake. after toad invasion (T. Madsen, personal communication), and at least some of the turtles larger species (e.g., Varanus panoptes) that declined abruptly with toad invasion are To my knowledge, the only quantitative once again abundant several decades later evidence of cane toad impacts on turtles in places such as Townsville and Bor- involves positive effects. On the Daly River, roloola (e.g., J. Llewelyn, B. L. Phillips, and cane toad invasion killed many yellow- L. Schwarzkopf, personal communication; R. spotted monitors and, presumably as a re- Shine, personal observation). Individuals sult, rates of predation by varanid lizards from these populations do not attack cane on the nests of pignose turtles (Carettochelys toads (Freeland 1990); these results suggest insculpta) fell from an average of around that either large varanids have learned to 17–23% to zero after toad invasion (Doody et al. 2006). At Fogg Dam, nightly counts avoid eating cane toads, or that there has of long-necked turtles (Chelodina rugosa) been strong selection against varanids that on the dam wall have tended to increase eat cane toads (Freeland 1990). Freeland since cane toad invasion (G. P. Brown, per- (2004) provided data on the amount of time sonal communication). it took for an expert hunter (with trained dogs) to locate Varanus panoptes before versus after the arrival of cane toads in a study lizards area near Borroloola in the Northern Terri- Burnett (1997) reported a significant pop- tory. The mean time to find a goanna aver- ulation decline of varanid lizards in Cape aged 35 minutes in control (non-toad in- York and Lawn Hill after the arrival of cane vaded) areas, and 32 minutes at the main site toads. Based on data collected during a six in the year prior to toad invasion. The time it to seven year period before and after toad took to locate a lizard increased to an aver- invasion on the Daly River, Doody et al. age of 103 minutes in the year after toads (2009) reported a 83–96% decline of yellow- arrived, but then decreased again to means 276 THE QUARTERLY REVIEW OF BIOLOGY Volume 85

Figure 4. Native Predators of the Cane Toad A phylogenetically diverse array of large-bodied vertebrate predators have been severely impacted by the spread of cane toads throughout Australia. Depicted here, clockwise from upper left, are the northern quoll Dasyurus hallucatus, the yellow-spotted goanna Varanus panoptes, the bluetongue lizard Tiliqua scincoides intermedia, and the king brown snake Pseudechis australis. (Photographs by J. K. Webb [quoll, king brown], R. Somaweera [goanna], and T. Child [bluetongue lizard], and used with permission.) of 56 and 38 minutes over the two following the period 1999 to 2005 (one year after years (Freeland 2004). toad arrival), falling to less than 0.01 liz- Although there have been many reports ards per night since that time (2006–2009; of a decline in the abundance of frillneck Price-Rees et al. 2010). lizards Chlamydosaurus kingii (see above), Watson and Woinarski (2003a,b) com- studies by Ujvari et al. (2010) have re- pared sites in Kakadu National Park the ported that, out of three study populations year before versus the year after toad inva- of frillnecks, one declined when toads ar- sion and reported declines in populations rived, another remained stable, and the of varanid lizards, presumably due to fatal third population increased. Those authors toxic ingestion of toads, as well as declines concluded that toad invasion was unlikely in the terrestrial gecko Gehyra nana and the to be responsible for any of these changes skink Carlia gracilis. The causes for decline in frillneck abundance. in these latter species are unclear. The Our laboratory studies have shown high same surveys revealed an increase in abun- mortality rates (through fatal poisoning) dance of the dragon Diporiphora bilineata, of bluetongue skinks Tiliqua scincoides inter- perhaps reﬂecting decreased predation by media (see Figure 4) from northwestern varanids. A similar phenomenon may ex- Australia when tested in captivity. Encoun- plain population increase following cane ter rates with bluetongue lizards on the toad invasion in another small agamid, Am- Adelaide River ﬂoodplain averaged ap- phibolurus gilberti, in the Daly River region proximately 0.025 lizards per night over (Doody et al. 2009). September 2010ECOLOGICAL IMPACT OF INVASIVE TOADS 277

snakes death adder Acanthophis praelongus—were en- It is notoriously difficult to obtain robust countered more frequently since the arrival population density estimates for snakes, as of toads than had been the case beforehand. these animals are typically rare, secretive, Counts of two other species, the brown tree highly mobile, and infrequently active (Sei- snake Boiga irregularis and the slatey-grey gel 1993). Hence, despite much anecdotal snake Stegonotus cucullatus, showed no evi- evidence that cane toads have been re- dence of significant impact, albeit with a sponsible for declines in population num- trend to increasing encounter rates follow- bers of elapid snakes (Covacevich 1974; ing toad invasion. Encounter rates with Chil- Covacevich and Archer 1975; van Beurden dren’s pythons (Antaresia childreni) and car- 1980; Mirtschin and Davis 1982; Easteal et pet pythons (Morelia spilota) both increased al. 1985; Seabrook 1993), actual data are concurrent with the arrival of toads, and lacking. Shine and Covacevich (1983) doc- then declined. Encounter rates with water pythons (Liasis fuscus) had been increasing umented a decrease through time in the prior to the arrival of toads, but began to number of specimens of toad-vulnerable decrease at about the time of toad arrival. snake species relative to non-vulnerable The remaining species—the king brown snake species registered in the Queensland snake Pseudechis australis (see Figure 4)—had Museum, but noted that the data were con- been declining in encounter rates prior to sistent with habitat changes as well as with toad arrival, and continued this pattern the impact of cane toads. Freeland (2004) post-invasion. No king brown snakes were pointed to a lack of extinctions of snake pop- recorded in the final 23 months of this ulations, even at a local level, by examining study, suggesting possible local extinction data from small islands where any negative of this taxon. effect of cane toads on such snakes would Highlighting the difficulties of inferring likely result in local extirpation. His tabula- causation from correlational data, some of tion of data on the occurrence of snake spe- the species that declined were the ones cies on islands off Queensland does not least likely to experience any direct impact reveal any trend for frog-eating species to be from cane toads. The usual prey of water absent from islands that contain cane toads. pythons—native rats—was in atypically low Quantitative population-abundance in- abundance because of a flood in 2006; formation is available for death adders (Ac- therefore, python numbers were reduced anthophis praelongus) on the Adelaide River by starvation. A lack of rats also forced floodplain. Laboratory studies and field- water pythons to begin consuming other based radio-tracking of death adders at this snakes, particularly keelbacks—a species site showed a toad-induced mortality of whose decline cannot be directly attrib- 48%, and counts of road-killed snakes in uted to toads, as these snakes are physio- the same area indicated an 89% mean re- logically capable of consuming them (Phil- duction in snake numbers concurrent with lips et al. 2003; Llewelyn et al. 2009). The cane toad invasion (Phillips et al. 2010b). increase in numbers of death adders—a In the same general area, we have exten- species at high risk from toads, and whose sive unpublished data on other snake spe- population was recorded to decline precip- cies. Nocturnal counts of active snakes were itously at a nearby site by 89% (see above)— made on the wall of Fogg Dam (Adelaide remains puzzling. The only species that River floodplain) by G. P. Brown over a 12- seems to have experienced a dramatic de- year period (1998 to 2009; more than 3600 cline in numbers on the Adelaide River standardized survey nights, comprising 83% floodplain as a direct result of lethal toxic of all nights over that 12-year period), brack- ingestion of toads is the king brown snake. eting the cane toad’s arrival in 2006. Of the What is clear is that (a) major changes nine most common snake species, three— have occurred in the composition of the the keelback Tropidonophis mairii, the small- snake community, concurrent with cane eyed snake Cryptophis pallidiceps, and the toad invasion, and (b) few of those 278 THE QUARTERLY REVIEW OF BIOLOGY Volume 85 changes are likely to have been causally rows of birds such as bee-eaters (Merops linked to cane toads, but instead reflect ornatus) (Boland 2004); data on the abun- other year-to-year variation in this dynamic dance of this species as a function of toad system. Our data thus paint a cautionary tale presence would be of interest. Similarly, for any attempt to interpret population-level given that toad invasion kills many varanid impacts of cane toads in the absence of lizards, predation on the eggs and nest- detailed information on the ecology, be- lings of ground-nesting birds (such as plo- havior, and physiology of the species in- vers) might be reduced. Information on volved. Even with thousands of nights of abundances of ground-nesting birds could survey data at a single site, standardized for clarify this possibility. search effort and techniques, it is still difficult to detect any unambiguous impacts mammals of cane toad invasion on the overall snake Based on track counts, Catling et al. (1999) fauna at Fogg Dam (G. P. Brown, personal suggested that toad invasion may have re- communication). duced dingo populations in the Borroloola region; however, more robust data from Kak- birds adu sites provided no evidence of any such Catling et al. (1999) reported that nec- impact (Watson and Woinarski 2003a,b). tarivorous/granivorous birds were less Following toad invasion, Watson and Woin- common following cane toad invasion, but arski (2003a,b) recorded fewer northern interpreted this result as a statistical arti- quolls (Dasyurus hallucatus), western chest- fact. In comparisons of sites in Kakadu the nut mice (Pseudomys nanus), and pale field- year before versus the year after toad inva- rats (Rattus tunneyi), but more agile wallabies sion, Watson and Woinarski (2003a,b) noted (Macropus agilis) and feral pigs (Sus scrofa). a general trend for relative increases in the Numbers of fawn antechinus (Antechinus bel- abundance and species richness of birds. lus) also increased, but the increase fell short Species that were more common in the toad- of statistical significance. The other mammal invaded areas included the mistletoebird Di- species that were surveyed—the common caeum hirundinaceum, red-backed fairy-wren rock rat Zyzomys argurus, the northern brown Malurus melanocephalus, white-winged triller bandicoot Isoodon macrourus, the feral buffalo Lalage sueurii, golden-headed cisticola Cisti- Bubalus bubalus, the Kakadu pebble-mound cola exilis, and magpie lark Grallina cyano- mouse Pseudomys calabyi, the grassland melo- leuca. In contrast, numbers declined for the mys Melomys burtoni, the delicate mouse banded honeyeater Cissomela pectoralis, Pseudomys delicatulus, the black-footed tree rat helmeted friarbird Philemon buceroides, olive- Mesembriomys gouldi, the dingo Canis familiaris backed oriole Oriolus sagittatus, and lemon- dingo, the feral horse Equus caballus, and the bellied flycatcher Microeca flavigaster. Watson red-cheeked dunnart Sminthopsis virginiae— and Woinarski (2003a,b) did not survey showed no overall increase or decrease in waterbirds, the group most likely to encoun- toad-colonized sites. ter toads. Reflecting its iconic status and rapid de- I am unaware of any other quantitative cline, the northern quoll (Dasyurus halluca- studies on the impact of cane toads on bird tus) (see Figure 4) has attracted the most populations; given strong public interest in research attention in this respect. Burnett birds, this lack of reports hints that toad (1997) reported serious population de- impacts may not generally be severe. Infor- clines in northern quolls in Cape York and mation on the outcomes of encounters be- Lawn Hill within a few months after the tween birds and toads suggests that despite arrival of cane toads. Unfortunately, it is occasional mortality from fatal poisoning, difficult to tease apart the effects of cane cane toad invasion is unlikely to severely toads from those of other threatening pro- affect most bird populations (Beckmann cesses. Populations of some small mam- and Shine 2009). An alternative pathway of mals, such as the northern quoll and impact involves usurpation of nesting bur- northern bandicoot, have declined rapidly September 2010ECOLOGICAL IMPACT OF INVASIVE TOADS 279 over recent decades, prior to the invasion Queensland, in areas long colonized by cane of cane toads (Braithwaite and Griffiths toads. They found co-occurring quolls and 1994; Woinarski et al. 2001). The toad, toads at several sites, and inferred that coex- however, may exacerbate these declines; istence between quolls and toads seemed to Oakwood (2003a,b) reported the extirpa- be stable in this system. tion of one study population coincident Given that these same species apparently with toad arrival in Kakadu National Park. declined precipitously with cane toad inva- Watson and Woinarski (2003a,b) found no sion in Queensland (Burnett 1997) and in northern quolls in 110 quadrats of Kakadu the Northern Territory (see above), the that cane toads had invaded, whereas 41 current situation suggests recovery post- individuals had been found in 17 of these invasion. How has this happened? Long- quadrats in the previous year. In conjunc- term, at least some Australian predators tion with the results of radio-tracking stud- have adapted to the presence of cane toads. ies (Oakwood 2003a,b; O’Donnell 2009; The most extensive data come from studies O’Donnell et al. 2010), these data leave no of the red-bellied blacksnake (Pseudechis por- doubt that cane toad invasion severely phyriacus) (Phillips et al. 2003, 2004; Phillips threatens the viability of northern quoll and Shine 2004, 2006a,b,c). Comparisons populations. between snakes from toad-colonized areas versus areas not yet invaded by toads re- Long-Term Impacts of Cane Toads vealed differences in a set of traits predicted There are massive logistical obstacles to to enhance a snake’s probability of surviving quantifying the numerical impact of cane an encounter with a toad. Snakes from toad- toads on the Australian fauna over a time- colonized areas refused to attempt to ingest scale of decades or beyond. First, we lack toads, and were more resistant to the toads’ quantitative data on population sizes of na- toxin than were conspecific animals from tive species prior to the toad’s invasion areas lacking cane toads (Phillips and Shine through Queensland. Second, abundances 2006c). Other data suggest that these of most species fluctuate so much through shifts in behavior and physiology reflect time and space, and are so heavily impacted genetically-based adaptive shifts rather by a suite of other anthropogenically- than learning or physiological acclimation imposed challenges (e.g., land-clearing, pes- (Phillips and Shine 2004, 2006c). Black- ticides, changes to water flows, other invasive snakes from toad-colonized areas exhib- species), that it would be difficult to inter- ited smaller head sizes relative to body size, pret those data in the light of cane toad a pattern also seen in the anuran-eating impact, even if we were to obtain precise Dendrelaphis punctulatus, but not in two numbers on prior abundance. All that can sympatric snake species not predicted to be be stated is that: (a) as far as we know, no at risk from cane toads (Hemiaspis signata, species has gone extinct as a result of cane Tropidonophis mairii) (Phillips and Shine toad invasion, and (b) some of the species 2004). Smaller relative head size renders a that declined dramatically soon after toad snake less vulnerable to lethal toxic inges- invasion are now abundant in areas well be- tion, as small-headed snakes can ingest hind the current invasion front. Although only small toads (Phillips and Shine 2004). these data are anecdotal, even superficial ob- Hence, cane toad invasion appears to have servation confirms that “toad-sensitive” spe- favored rapid evolutionary shifts in behav- cies such as yellow-spotted monitors (Vara- ior, physiology, and morphology of snake nus panoptes) are common in northeastern predators in ways that facilitate snake sur- Queensland, where they thrive in areas con- vival in the presence of toads. taining high densities of cane toads (Fearn Such shifts may occur quickly. Phillips et 2003; J. Llewelyn, D. Nelson, L. Schwarzkopf, al. (2010b) demonstrated that cane toad personal communication; R. Shine, personal invasion imposed strong selection on feed- observation). Woinarski et al. (2008) sur- ing responses (propensity to attack toads) veyed populations of northern quolls in and morphology (relative head size) in a 280 THE QUARTERLY REVIEW OF BIOLOGY Volume 85 population of floodplain death adders (Ac- provides an unusually detailed case study anthophis praelongus) near Darwin. In this of how the arrival of a single toxic species study, feeding responses of adders were can affect a speciose tropical fauna. Ac- monitored in captivity, and then the snakes cordingly, we can ask what general lessons were released and radio-tracked in an area emerge from the accumulated data on that had recently been invaded by cane toad impact. I identify several such emer- toads. Snakes that refused to attack toads in gent themes below. the laboratory were more likely to survive 1. Community perceptions of invasive species after release, and snakes with relatively small impacts are often in error. The general public heads were more likely to survive than were holds diverse views about cane toads, most of larger-headed conspecifics (Phillips et al. which are very negative. A widespread abhor- 2010b). Encouragingly, then, Australian rence of the animal, apparently reflecting its snakes can adapt to cane toad invasion; size and appearance, has led some commu- nonetheless, populations of these preda- nity spokespeople to promulgate catastrophe tors may require long periods of time scenarios, whereby the cane toad invasion is before they return to pre-toad levels, if in- predicted to massively affect virtually all deed they ever do. Freeland (2004) sug- native species (this opinion is especially gested that large varanids might return to common in communities where toads have pre-toad abundances within a few years, yet to arrive) (Clarke et al. 2009). Indeed, but anecdotal reports suggest that some surveys of public opinion often rank cane previously-common predator species are toads among the greatest ecological threats still relatively rare in toad-infested areas of to Australian ecosystems (Fitzgerald et al. northeastern Queensland (e.g., Pseudechis 2007). That widespread concern has stim- porphyriacus) (B. L. Phillips, personal com- ulated massive levels of expenditure by munication). state and federal governments—more than Potentially, very low population levels, $9,500,000 between 1986 and 1996 (Shine per se, due to toad invasion might affect the et al. 2006)—and enormous efforts by vol- probability of extinction as well as rates of unteers to eradicate local toad populations subsequent recovery. Reduced genetic vari- by direct collection. In reality, the direct ability due to “bottleneck” effects can im- impact of cane toads falls most heavily on a peril population recovery (Madsen et al. small number of native taxa, rather than 1999), and low population sizes render across a wide spectrum of native fauna. such systems more vulnerable to stochastic Similarly, members of the general public threats (Caughley 1994). In practice, such envisage a wide range of mechanisms of issues are unlikely to be significant for impact, ranging from toads poisoning wa- most Australian native fauna affected by terholes via toxin release, to encouraging cane toad invasion. Most of the animals drug abuse in people who become ad- involved (e.g., varanid lizards, large elapid dicted to licking the toads or smoking their snakes) are highly vagile, enabling signif- dried skins (Clarke et al. 2009). There are icant gene flow to persist even at low no data to suggest that either of these is- population densities. Relatively high popu- sues is a major problem. More realistically, lation numbers at a landscape scale mean competition between toads and native that even severe reductions are unlikely to frogs is often suggested to be a major pose real threats of extirpation. Detailed mechanism of impact (Shine et al. 2006). population genetics analyses to address Experimental studies and surveys do not these questions would be of value. support that hypothesis (Freeland and Kerin 1988; Catling et al. 1999; Grigg et al. 2006; What Factors Influence the Greenlees et al. 2006, 2007), although Ecological Impact of This Invasive more research is warranted on the effects Species? of dense dry season aggregations of toads The intense research on ecological im- on local prey abundances. Available data pacts of invasive cane toads in Australia suggest that a single proximate mecha- September 2010ECOLOGICAL IMPACT OF INVASIVE TOADS 281 nism—fatal poisoning of predators that at- long to a diverse array of phylogenetic lin- tempt to ingest toads—is by far the most eages, with large absolute body size being important mechanism of direct impact of the single common denominator. cane toads on native fauna. Consistent, albeit anecdotal, reports of 2. Impacts are affected by several attributes eventual post-invasion reductions in toad of toad biology. The biology of invasive spe- abundance and recovery of predator popu- cies necessarily influences the type, extent, lations suggest that the impact-determining and magnitude of ecological interactions properties of the toad population may also with native fauna, even if attention is re- change with time since invasion. In keeping stricted to a single mechanism of impact with that inference, the toxin-containing pa- (i.e., lethal toxic ingestion). Importantly, rotoid gland is largest, relative to body size, cane toads have a multiphasic life history, in toads from the invasion front (Phillips and with eggs and tadpoles in waterbodies, Shine 2006b). It is unclear whether this is an metamorphs restricted to riparian areas, adaptive shift, or a phenotypically plastic re- and larger juveniles and adults spreading sponse to larval environments (Hagman et out through the drier landscape (e.g., Al- al. 2009b). Toad abundance may also de- ford et al. 1995; Freeland and Kerin 1988; crease in long-colonized areas, perhaps re- Child et al. 2008a). The types and amounts flecting reduced food availability (Freeland of toxins, as well as body sizes and habitat et al. 1986a; Tyler 1994). Lastly, toad popu- use, change considerably through toad on- lation structure may alter. Toads at the inva- togeny (Hayes et al. 2009). Specific fea- sion front rarely reproduce, apparently tures of toad biology also affect impact. For because of strong spatial selection for traits example, the presence of a non-toxic jelly that enhance dispersal abilities at the cost of coat around the highly toxic eggs is fatal to investment into other processes (Brown et al. aquatic predators that otherwise would be 2007; Phillips et al. 2010a). Thus, one reason able to detect and reject prey items that for the intense ecological impact of toads at contain bufotoxin (Greenlees and Shine the colonization front may be the scarcity of 2010). The end result of the complex life toads small enough to comprise a non-fatal history and wide range in absolute body meal. If small toads are common, predators sizes and habitats of toads is an increase in may have the opportunity for aversion learn- the range of predator species likely to in- ing (e.g., Webb et al. 2008; Greenlees et al. gest them. The breadth of impact would be 2010a). A shift in toad recruitment rates— lower in the case of an invasive anuran and, therefore, in the availability of toads species with larger metamorphs relative to small enough to induce aversion but not adult size, or without an aquatic phase in cause predator death— might render pred- the life history (e.g., the coqui, Eleuthero- ators more likely to survive in long-colonized dactylus coqui, currently invading the Ha- areas than at the toad invasion front. waiian islands) (Kraus 2009). 3. Impacts are affected by several attributes Superimposed on the wide body-size of predator biology. As noted above, any range of terrestrial-phase toads (from Ͻ characteristic that increases the size of 1gtoϾ 1 kg) (see Figure 1) is a strong toads that a predator will attack is likely to positive allometry in toxin content (Phil- greatly increase predator vulnerability be- lips and Shine 2006a). Thus, a large adult cause of the strong positive allometry in toad contains orders of magnitude more toxin content of cane toads. Predator body toxin than a juvenile conspecific. That al- size is the most obvious of such traits, but lometry appears to explain one of the in some species there may also be signifi- strongest patterns to emerge from research cant divergences in predatory tactics on cane toad impact: the disproportionate between the sexes. For example, male mortality of predators that attain large northern quolls (Dasyurus hallucatus) are body sizes themselves and are thus pre- bolder as well as larger than females, are more pared to attack large toads (see Figure 4). willing to attack large toads, and are thus much The anurophagous predators involved be- more likely to die as a result of such encounters 282 THE QUARTERLY REVIEW OF BIOLOGY Volume 85

(O’Donnell 2009). Divergences in prey types ability, and yet others to be under little and foraging tactics may be widespread in effective selection because they already are species that exhibit marked sex divergences able to cope with the novel challenge in mean adult body size, potentially modify- posed by cane toad invasion (Llewelyn et ing their vulnerability to toad invasion (e.g., al. 2010a). floodplain death adders, Acanthophis praelon- 4. Impacts are affected by attributes of the gus) (Webb et al. 2005). local environment. Rates of encounter be- Interspecific (and occasionally intraspe- tween cane toads and potential predators cific) variation in physiological tolerance to are affected by the local habitat. For exam- bufotoxins also plays a role in determining ple, a relatively dry landscape matrix will predator vulnerability. Red-bellied blacksnakes keep most toads close to water (because (Pseudechis porphyriacus) and small marsupial they need to hydrate every few days) (Al- carnivores (Planigale maculata) exposed to ford et al. 1995), thus increasing rates of several decades of sympatry with cane toads encounter with aquatic predators like fresh- show greater tolerance to toad toxins than water crocodiles (Letnic et al. 2008) but do toad-naı¨ve conspecifics (Phillips and decreasing rates of encounter with preda- Shine 2006c; Llewelyn et al. 2010b). Given tor taxa that forage in drier habitats. At the the rapidity of this evolutionary response, extreme arid zone edges of the toad inva- we would expect that Australian predators sion, this situation may generate great het- with strong genetic ties to Asia (i.e., the erogeneity in toad impact both through descendants of relatively recent invasions space (riparian vs. other sites) and time from the north) will be able to tolerate (dry weather vs. rainy periods that allow bufotoxins better than species with a long toad dispersal across the landscape). history of endemicity in Australia (Beck- 5. Impacts are affected by attributes of local mann and Shine 2009; Llewelyn et al. foodwebs. The invasion of cane toads can 2010a,b,c). Asia contains many toad spe- result in the mortality of a high proportion cies with toxins broadly similar in chemical of large anurophagous reptiles and mam- composition to the bufadienolides pro- mals. Because the species most strongly af- duced by cane toads (Meyer and Linde fected in this respect have broad diets 1971). Within groups such as mammals, (e.g., Shine et al. 2006; Letnic et al. 2008; birds, snakes, and crocodiles, taxa with re- O’Donnell et al. 2010), their deaths reduce cent links to Asia are indeed less vulnera- mortality rates for their usual prey taxa ble to toad invasion than are related taxa (e.g., see Doody et al. 2006 for reduced with a longer history of Australian ende- rates of varanid predation on turtle eggs micity. following toad invasion). Presumably, such Aversion learning protects many native effects may flow through multiple trophic predators from the ill-effects of cane toad levels. However, in addition to killing pred- ingestion (Webb et al. 2008; Shine et al. ators, cane toads can also modify the pred- 2009; Greenlees et al. 2010a); therefore, atory tactics of the survivors; toad-averse variation in learning ability may influence predators switch their behaviors to target magnitudes of impact. There is likely to be prey that do not resemble toads (Nelson a complex interplay among the multiple 2008; Nelson et al. 2010a,b), or more care- selective forces induced by toad arrival. For fully evaluate potential prey before seizing example, a predator species capable of them (Webb et al. 2008). Native species rapid aversion learning will be under little thus may benefit from “accidental mim- selection for increased physiological toler- icry” of the invasive toads. For example, a ance to cane toad toxin, because immedi- frog species that closely resembles a toad ate rejection of the toxic prey removes the may be relatively safe, at least until preda- physiological challenge of dealing with the tors learn to distinguish between the toxic toad’s poisons. We might thus expect some invader and the non-toxic native (Webb et species to evolve physiological tolerance to al. 2008; Nelson et al. 2010b). toxins, others to evolve enhanced learning 6. Impacts of invasive species are difficult to September 2010ECOLOGICAL IMPACT OF INVASIVE TOADS 283 demonstrate. In this review, I have relied on Good land management practices can help two types of information—survey data that to ensure that healthy predator populations show reduction in abundance of a native can recover more rapidly and completely taxon before versus after toad arrival, and from the impact of toads. However, there is laboratory trials that document the out- no room for complacency; myriad other comes of staged encounters between pred- threats act on native biodiversity, and chang- ators and toads. Both have major flaws. ing habitats and climate can potentially de- Survey data face two problems: (1) count stabilize these systems. Toads are currently data are so variable, especially for rare spe- moving into areas of Australia with faunistic cies, that it is difficult to obtain statistically assemblages different from any that they robust evidence of decline even for major have yet encountered. Cane toads were first changes in abundance (Watson and Woin- brought to Australia more than 70 years ago, arski 2003a), and (2) the inference of cau- but it is only within the last five years that a sation is based on a temporal correlation clear picture is beginning to emerge about (i.e., faunal decline occurred at the same the consequences of that unfortunate intro- time that toads arrived), that could equally duction for the Australian fauna. Direct as well be explained by many other factors impacts of cane toads are now fairly well- (e.g., weather conditions). Laboratory tri- understood, but more research is needed to als face the difficulty of extrapolation to clarify the indirect, foodweb-mediated effects the field. For example, if local ecological of this invasive species on native biodiversity. circumstances render encounter rates between predators and toads highly unlikely, acknowledgments then whether or not that predator will attack a toad in the laboratory tells us very Much of the information reviewed above comes little about impact in the field. In practice, from the work of my research group, “Team Bufo.” I thank them for their hard work, their insights, the combination of survey data and labora- and their generosity in sharing ideas and data. In tory trials—plus reliable anecdotal reports of particular, our understanding of the ecological im- field encounters and their outcomes, or dis- pacts of cane toads has been substantially enhanced section of stomach contents of predators by the work of my postdoctoral fellows (Greg found dead in the field (e.g., Fearn 2003; Brown, Michael Crossland, Sylvain Dubey, Mattias Letnic et al. 2008; Doody et al. 2009)—is Hagman, Ben Phillips, Ligia Pizzatto, Jonathan stronger than either survey or laboratory- Webb), research students (Christa Beckmann, Ha- trial data alone. ley Bowcock, Travis Child, Matt Greenlees, Crystal Kelehear, John Llewelyn, David Nelson, Stephanie O’Donnell, Sam Price-Rees, Ruchira Somaweera, Conclusion Georgia Ward-Fear), laboratory staff (Melanie El- phick, Adele Haythornthwaite, Christine Rioux), Cane toads have direct impacts on some and collaborators (Ross Alford, Di Barton, Rob Ca- species of Australian native animals, mostly pon, Sean Doody, Andrew Hayes, David Pearson, through lethal toxic ingestion by large pred- Lin Schwarzkopf, Cathy Shilton, Dion Wedd). Our ators (quolls, crocodiles, snakes, varanid and research has been funded primarily by the Austra- scincid lizards). Direct impacts on other spe- lian Research Council, through the University of cies are unlikely to be severe enough to have Sydney. Additional funding has come from the Fed- any major effects at the population level. eral Department of Environment, Water, Heritage However, high mortality rates of top preda- and the Arts, and in collaboration with the Western tors will have flow-on effects, with popula- Australian Department of Environment and Con- tions of some native species increasing as a servation, the Territory Wildlife Park, and the Aus- tralian Reptile Park. Preparation of this review was result, while others decrease. A capacity for funded by the Commonwealth of Australia, under rapid learning of toad-avoidance by some the Caring for Our Country initiative. I thank Tony predators, and for rapid adaptive shifts in Peacock (CRC for Invasive Species) for his encour- others, means that the ecological impact of agement, and the reviewers and editor for their cane toads is likely to decrease through time. helpful suggestions. 284 THE QUARTERLY REVIEW OF BIOLOGY Volume 85

REFERENCES

Alford R. A. 1999. Ecology: resource use, competition, bondage: the costs of amplexus in cane toads and predation. Pages 240–278 in Tadpoles: The Biol- (Bufo marinus). Copeia 2009(1):29–36. ogy of Anuran Larvae, edited by R. W. McDiarmid and Braithwaite R. W., Griffiths A. 1994. Demographic R. Altig. Chicago (IL): University of Chicago Press. variation and range contraction in the northern Alford R. A., Cohen M. P., Crossland M. R., Hearnden quoll Dasyurus hallucatus (Marsupialia: Dasyuri- M. N., Schwarzkopf L. 1995. Population biology of dae). Wildlife Research 21:203–217. Bufo marinus in northern Australia. Pages 173–181 Breeden K. 1963. Cane toad (Bufo marinus). Wildlife in in Wetland Research in the Wet-Dry Tropics of Austra- Australia 1:31. lia, Supervising Scientist Report 101, edited by Brown G. P., Shilton C. M., Phillips B. L., Shine R. C. M. Finlayson. Canberra (Australia): Supervising 2007. Invasion, stress, and spinal arthritis in cane Scientist. toads. Proceedings of the National Academy of Sciences Andersen A. N. 1991. Seed harvesting by ants in Aus- USA 104(45):17698–17700. tralia. Pages 493–503 in Ant-Plant Interactions, ed- Brown G. P., Shine R. 2007. Rain, prey and predators: ited by C. R. Huxley and D. F. Cutler. Oxford climatically driven shifts in frog abundance mod- (UK) and New York: Oxford University Press. ify reproductive allometry in a tropical snake. Oe- Andersen A. N., Braithwaite R. W. 1996. Plant-animal cologia 154(2):361–368. interactions. Pages 137–154 in Landscape and Veg- Buchholz R. 2007. Behavioural biology: an effective etation Ecology of the Kakadu Region Northern Austra- and relevant conservation tool. Trends in Ecology lia, edited by C. M. Finlayson and I. von Oertzen. and Evolution 22(8):401–407. Dordrecht (The Netherlands): Kluwer Academic Burnett S. 1997. Colonising cane toads cause popula- Publishers. tion declines in native predators: reliable anec- Andrew M. H. 1986. Granivory of the annual grass dotal information and management implications. Sorghum intrans by the harvester ant Meranoplus sp. Pacific Conservation Biology 3:65–72. in tropical Australia. Biotropica 18:344–399. Cameron E. E., Cogger H. G. 1992. The Herpetofauna of [ANPWS] Australian National Parks and Wildlife Ser- the Weipa Region, Cape York Peninsula. Technical vice / [DEST] Department of the Environment, Reports of the Australian Museum, No. 7. Sydney Sport and Territories. 1991. Nomination of Kakadu (Australia): Australian Museum. National Park by the Government of Australia for In- Cassels A. J. 1966. Disembowelled toads near water. scription in the World Heritage List. Canberra (Aus- North Queensland Naturalist 34:6. tralia): ANPWS and DASETT. Caswell D. 1987. Australian raven eats cane toad. Bird Bailey P. 1976. Food of the marine toad, Bufo marinus, Observer 663:47. and six species of skink in a cacao plantation in Catling P. C., Hertog A., Burt R. J., Wombey J. C., New Britain, Papua New Guinea. Australian Wild- Forrester R. I. 1999. The short-term effect of cane life Research 3(2):185–188. toads (Bufo marinus) on native fauna in the Gulf Barraclough N. 1988. Dead cane toads. Bird Observer Country of the Northern Territory. Wildlife Re- 672:21. search 26:161–185. Beckmann C., Shine R. 2009. Impact of invasive cane Caughley G. 1977. Analysis of Vertebrate Populations. toads on Australian birds. Conservation Biology Chichester (UK): Wiley. 23(6):1544–1549. Caughley G. 1994. Directions in conservation biology. Begg G. W., Walden D., Rovis-Hermann J. 2000. Report Journal of Animal Ecology 63:215–244. on the Joint ERISS/PAN Cane Toad Risk Assessment Child T., Phillips B. L., Brown G. P., Shine R. 2008a. Field Trip to the Katherine/Mataranka and Borroloola The spatial ecology of cane toads (Bufo marinus)in Regions. Supervising Scientist Internal Report 389. tropical Australia: why do metamorph toads stay Darwin (Australia): Supervising Scientist. near the water? Austral Ecology 33(5):630–640. Bekker R. 1985. Predation on cane toads (Bufo mari- Child T., Phillips B. L., Shine R. 2008b. Abiotic and nus). Herpetofauna 16:52–53. biotic influences on the dispersal behaviour of Boland C. R. J. 2004. Introduced cane toads are active metamorph cane toads (Bufo marinus) in tropical nest predators and competitors of rainbow bee- Australia. Journal of Experimental Zoology 309A(4): eaters: observational and experimental evidence. 215–224. Biological Conservation 120:53–62. Child T., Phillips B. L., Shine R. 2009. Does desicca- Bowcock H., Brown G. P., Shine R. 2008. Sexual com- tion risk drive the distribution of metamorph cane munication in cane toads (Chaunus marinus): what toads (Bufo marinus) in tropical Australia? Journal cues influence the duration of amplexus? Animal of Tropical Ecology 25:193–200. Behaviour 75(4):1571–1579. Clarke R., Carr A., White S., Raphael B., Baker J. 2009. Bowcock H., Brown G. P., Shine R. 2009. Beastly Cane Toads in Communities: Talking Toads in North- September 2010ECOLOGICAL IMPACT OF INVASIVE TOADS 285

ern Australia. Executive Report to the Australian native and exotic Bufo habitats. Herpetologica 55(2): Government, Bureau of Rural Sciences, Canberra, 192–199. Australia. Crossland M. R., Brown G. P., Anstis M., Shilton C., Clerke R. B., Williamson I. 1992. A note on the pre- Shine R. 2008. Mass mortality of native anuran dation of Bufo marinus juveniles by the ant Iri- tadpoles in tropical Australia due to the invasive domyrmex purpureus. Australian Zoologist 28:64–67. cane toad (Bufo marinus). Biological Conservation Cogger H. G. 2000. Reptiles and Amphibians of Austra- 141:2387–2394. lia. Sixth Edition. Sanibel Island (FL): Ralph Cur- Crossland M. R., Shine R. 2010. Vulnerability of an tis Books. Australian anuran tadpole assemblage to the toxic Cogger H. G., Cameron E. E., Sadlier R. A., Eggler P. eggs of the invasive cane toad (Bufo marinus). Aus- 1993. The Action Plan for Australian Reptiles. Can- tral Ecology 35:197–204. berra (Australia): Australian Nature Conservation Curio E. 1996. Conservation needs ethology. Trends in Agency. Ecology and Evolution 11(6):260–263. Cohen M. P., Alford R. A. 1996. Factors affecting Daly J. W., Witkop B. 1971. Chemistry and pharma- diurnal shelter use by the cane toad, Bufo marinus. cology of frog venoms. Pages 497–519 in Venomous Herpetologica 52:172–181. Animals and their Venoms, Volume II, edited by W. Covacevich J. 1974. An unusual aggregation of snakes fol- Bu¨cherl and E. E. Buckley. New York: Academic lowing major ﬂooding in the Ipswich-Brisbane area, Press. south-eastern Queensland. Herpetofauna 7(1):21–24. Delvinquier B. L. J., Freeland W. J. 1988. Protozoan Covacevich J., Archer M. 1975. The distribution of the parasites of the cane toad, Bufo marinus, in Austra- cane toad, Bufo marinus, in Australia and its effects lia. Australian Journal of Zoology 36:301–316. on indigenous vertebrates. Memoirs of the Queens- Dickman C. R. 2007. The complex pest: interaction land Museum 17:305–310. webs between pests and native species. Pages 208– 215 in Pest or Guest: The Zoology of Overabundance, Crossland M. R. 1997. Impact of eggs, hatchlings and edited by D. Lunney et al. Mosman (Australia): tadpoles of the introduced cane toad Bufo marinus Royal Zoological Society of NSW. (Anura: Bufonidae) on native aquatic fauna in Donato D. B., Potts R. T. 2004. Culturally transmitted northern Queensland, Australia. Ph.D. thesis. predation and consumption techniques by Torre- Townsville (Australia): James Cook University of sian crows Corvus orru on cane toads Bufo marinus. North Queensland. Australian Field Ornithology 21:125–126. Crossland M. R. 1998a. Ontogenetic variation in Doody J. S., Green B., Rhind D., Castellano C. M., toxicity of tadpoles of the introduced toad Bufo Sims R., Robinson T. 2009. Population-level de- marinus to native Australian aquatic invertebrate clines in Australian predators caused by an inva- predators. Herpetologica 54(3):364–369. sive species. Animal Conservation 12(1):46–53. Crossland M. R. 1998b. A comparison of cane toad Doody J. S., Green B., Sims R., Rhind D., West P., and native tadpoles as predators of native anuran Steer D. 2006. Indirect impacts of invasive cane eggs, hatchlings and larvae. Wildlife Research 25: toads (Bufo marinus) on nest predation in pig- 373–381. nosed turtles (Carettochelys insculpta). Wildlife Re- Crossland M. R. 2000. Direct and indirect effects of search 33:349–354. the introduced toad Bufo marinus (Anura: Bufo- Dorfman E. 1997. Possible Threats to Birds from Cane nidae) on populations of native anuran larvae in Toads in Kakadu National Park. Unpublished List Australia. Ecography 23(3):283–290. supplied to Parks Australia, Canberra (Australia). Crossland M. R. 2001. Ability of predatory native Aus- Drewes R. C., Roth B. 1981. Snail-eating frogs from tralian ﬁshes to learn to avoid toxic larvae of the the Ethiopian highlands: a new anuran specializa- introduced toad Bufo marinus. Journal of Fish Biol- tion. Zoological Journal of the Linnean Society 72(3): ogy 59:319–329. 267–287. Crossland M. R., Alford R. A. 1998. Evaluation of the Dubey S., Shine R. 2008. Origin of the parasites of an toxicity of eggs, hatchlings and tadpoles of the invading species, the Australian cane toad (Bufo introduced toad Bufo marinus (Anura: Bufonidae) marinus): are the lungworms Australian or Amer- to native Australian aquatic predators. Australian ican? Molecular Ecology 17:4418–4424. Journal of Ecology 23:129–137. Easteal S., van Beurden E. K., Floyd R. B., Sabath Crossland M. R., Alford R. A., Shine R. 2009. Impact M. D. 1985. Continuing geographical spread of of the invasive cane toad (Bufo marinus)onan Bufo marinus in Australia: range expansion be- Australian frog (Opisthodon ornatus) depends on tween 1974 and 1980. Journal of Herpetology 19(2): reproductive timing. Oecologia 158:625–632. 185–188. Crossland M. R., Azevedo-Ramos C. 1999. Effects of Evans S. 1999. Fisheries issues in the Borroloola re- Bufo (Anura: Bufonidae) toxins on tadpoles from gion. Page 16 in Wise Use of Wetlands in Northern 286 THE QUARTERLY REVIEW OF BIOLOGY Volume 85

Australia: Indigenous Use, edited by P. Whitehead et Haddad C. F. B., De Sa´ R. O., Channing A., Wil- al. Darwin (Australia): Northern Territory Univer- kinson M., Donnellan S. C., Raxworthy C. J., sity, Centre for Tropical Wetlands Management Campbell J. A., Blotto B. L., Moler P., Drewes and Centre for Indigenous Natural and Cultural R. C., Nussbaum R. A., Lynch J. D., Green D. M., Resource Management. Wheeler W. C. 2006. The amphibian tree of life. Fearn S. 2003. Pseudechis porphyriacus (red-bellied black Bulletin of the American Museum of Natural History snake). Diet. Herpetological Review 34:253–254. 297:1–370. Fitzgerald G., Fitzgerald N., Davidson C. 2007. Public Grace B. 2008. Cane toads and Top End fish. Fishnote Attitudes Towards Invasive Animals and Their Im- 42:1–3. pacts. Report to the Invasive Animals Cooperative Grace B., Sawyer G. 2008. Dead fish and frogs associ- Research Centre, Canberra (Australia). ated with cane toads near Darwin. NT Naturalist Fitzgerald M. 1990. Rattus rattus: the introduced black 20:222–225. rat, a successful predator on the introduced cane Grant G. S. 1996. Prey of the introduced Bufo marinus toad Bufo marinus in northern New South Wales. on American Samoa. Herpetological Review 27(2): Herpetofauna 20(1):9–13. 67–69. Freeland W. J. 1984. Cane toads: a review of their biology Greenlees M. J., Brown G. P., Webb J. K., Phillips and impact on Australia. Technical Report 19. Dar- B. L., Shine R. 2006. Effects of an invasive anuran win (Australia): Parks and Wildlife Unit, Conser- (the cane toad, Bufo marinus) on the invertebrate vation Commission of the Northern Territory. fauna of a tropical Australian floodplain. Animal Freeland W. J. 1986. Populations of cane toad, Bufo Conservation 9(4):431–438. marinus, in relation to time since colonisation. Greenlees M. J., Brown G. P., Webb J. K., Phillips Australian Wildlife Research 13:321–329. B. L., Shine R. 2007. Do invasive cane toads (Chau- Freeland W. J. 1987. Cane toads and the balance of nus marinus) compete with Australian frogs (Cyclo- nature. Wildlife Australia 24:12–15. rana australis)? Austral Ecology 32(8):900–907. Freeland W. J. 1990. Effects of the Cane Toad Bufo Greenlees, M. J., Phillips B. L., Shine R. 2010a. An marinus on Native Australian Wildlife: A Review of invasive species imposes selection on life-history Past and Current Research. Unpublished Report. traits of a native frog. Biological Journal of the Lin- Darwin (Australia): Conservation Commission of nean Society 100:329–336. the Northern Territory. Greenlees M, Phillips B. L., Shine R. 2010b. Adjusting Freeland W. J. 1993. Parasites, pathogens and the to a toxic invader: native Australian frog learns not impact of introduced organisms on the balance of to prey on cane toads. Behavioral Ecology doi: nature in Australia. Pages 171–180 in Conservation 10.1093/beheco/arq095. Biology in Australia and Oceania, edited by C. Moritz Greenlees, M. J., Shine R. 2010. Impacts of eggs and and J. Kikkawa. Chipping Norton (NSW, Austra- tadpoles of the invasive cane toad (Bufo marinus) lia): Surrey Beatty and Sons. on aquatic predators in tropical Australia. Austral Freeland W. J. 2000. Untitled. Unpublished Report to Ecology doi: 10.1111/j.1442-9993.2010.02116.x. Parks and Wildlife Commission of the Northern Griffiths A. D., Christian K. A. 1996. Diet and habitat Territory, Palmerston (Australia). use of frillneck lizards in a seasonal tropical envi- Freeland W. J. 2004. An Assessment of the Introduced Cane ronment. Oecologia 106:39–48. Toad’s (Bufo marinus Linnaeus) Impacts on the Native Griffiths A. D., McKay J. L. 2007. Cane toads reduce Australian Fauna, with Particular Reference to the Eastern the abundance and site occupancy of Merten’s Kimberley Region. Unpublished Report to the WA De- water monitor (Varanus mertensi). Wildlife Research partment of Industry and Resources. Perth (Austra- 34(8):609–615. lia): HLA-Envirosciences. Grigg G., Taylor A., McCallum H., Fletcher L. 2006. Freeland W. J., Delvinquier B. L. J., Bonnin B. 1986a. Monitoring the impact of cane toads (Bufo mari- Decline of cane toad, Bufo marinus, populations: nus) on Northern Territory frogs: a progress re- status of urban toads. Australian Wildlife Research port. Pages 47–54 in Science of Cane Toad Invasion 13(4):597–601. and Control: Proceedings of the Invasive Animals CRC/ Freeland W. J., Delvinquier B. L. J., Bonnin B. 1986b. CSIRO/Qld NRM&W Cane Toad Workshop, Brisbane, Food and parasitism of the cane toad, Bufo mari- June 2006, edited by K. L. Molloy and W. R. Hen- nus, in relation to time since colonisation. Austra- derson. Canberra (Australia): Invasive Animals lian Wildlife Research 13:489–499. Cooperative Research Centre. Freeland W. J., Kerin S. H. 1988. Within-habitat rela- Hagman M., Hayes R. A., Capon R. J., Shine R. 2009b. tionships between invading Bufo marinus and Aus- Alarm cues experienced by cane toad tadpoles tralian species of frog during the tropical dry affect post-metamorphic morphology and chemi- season. Australian Wildlife Research 15:293–305. cal defences. Functional Ecology 23(1):126–132. Frost D. R., Grant T., Faivovich J., Bain R. H., Haas A., Hagman M., Phillips B. L., Shine R. 2009a. Fatal September 2010ECOLOGICAL IMPACT OF INVASIVE TOADS 287

attraction: adaptations to prey on native frogs im- Kearney M. R., Phillips B. L., Tracy C. R., Christian peril snakes after invasion of toxic toads. Proceed- K. A., Betts G., Porter W. P. 2008. Modelling spe- ings of the Royal Society of London, Series B: Biological cies distributions without using species distribu- Sciences 276(1668):2813–2818. tions: the cane toad in Australia under current Hagman M., Shine R. 2006. Spawning-site selection by and future climates. Ecography 31:423–434. feral cane toads (Bufo marinus) at an invasion King R. B., Ray J. M., Stanford K. M. 2006. Gorging on front in tropical Australia. Austral Ecology 31(5): gobies: beneficial effects of alien prey on a threat- 551–558. ened vertebrate. Canadian Journal of Zoology 84: Hagman M., Shine R. 2007. Effects of invasive cane 108–115. toads on Australian mosquitoes: does the dark Knowles R. P. 1964. The poisonous toad and the cloud have a silver lining? Biological Invasions 9(4): canine. Veterinary Medicine 59:38–42. 445–452. Kraus F. 2009. Alien Reptiles and Amphibians: A Scientific Hagman M., Shine R. 2008. Australian tadpoles do Compendium and Analysis. New York: Springer. not avoid chemical cues from invasive cane toads Lawler K. L., Hero J. M. 1997. Palatability of Bufo (Bufo marinus). Wildlife Research 35:59–64. marinus tadpoles to a predatory fish decreases with Hagman M., Shine R. 2009. Species-specific commu- development. Wildlife Research 24:327–334. nication systems in an introduced toad compared Letnic M., Ward S. 2005. Observations of freshwater with native frogs in Australia. Aquatic Conservation crocodiles (Crocodylus johnstoni) preying upon 19:724–728. cane toads (Bufo marinus) in the Northern Terri- Hamley J., Georges A. 1985. The Australian snapping tory. Herpetofauna 35:98–100. tortoise Elseya latisternum: a successful predator on Letnic M., Webb J. K., Shine R. 2008. Invasive cane the introduced cane toad? Australian Zoologist toads (Bufo marinus) cause mass mortality of fresh- 21(7):607–610. water crocodiles (Crocodylus johnstoni) in tropical Hayes R. A., Crossland M. R., Hagman M., Capon Australia. Biological Conservation 141:1773–1782. R. J., Shine R. 2009. Ontogenetic variation in the Lever C. 2001. The Cane Toad: The History and Ecology of chemical defences of cane toads (Bufo marinus): a Successful Colonist. Otley (UK): Westbury Aca- toxin profiles and effects on predators. Journal of demic and Scientific Publishing. Chemical Ecology 35(4):391–399. Licht L. E. 1968. Unpalatability and toxicity of toad Hearnden M. N. 1991. Reproductive and larval ecol- eggs. Herpetologica 24(2):93–98. ogy of Bufo marinus (Anura: Bufonidae). Ph.D. Llewelyn J., Phillips B. L., Brown G. P., Schwarzkopf thesis. Townsville (Australia): James Cook Univer- L., Alford R. A., Shine R. 2010c. Adaptation or sity of North Queensland. preadaptation: why are keelback snakes (Tropido- Hinckley A. D. 1963. Diet of the giant toad, Bufo nophis mairii) less vulnerable to invasive cane toads marinus (L) in Fiji. Herpetologica 18:253–259. (Bufo marinus) than are other Australian snakes? Holland D. C. 2004. Interim Report for Study of Impacts of Evolutionary Ecology doi: 10.1007/s10682-010-9369-2. Cane Toads, Bufo marinus, on Two Goanna Species, Varanus panoptes and Varanus gouldii. Issues As- Llewelyn J., Phillips B. L., Shine R. 2009. Sublethal sociated with the Progressive Entry into the Northern costs associated with the consumption of toxic Territory of Cane Toads. Volume 1. Report No.1, prey by snakes. Austral Ecology 34:179–184. October 2003. Darwin (Australia): Legislative As- Llewelyn J., Schwarzkopf L., Alford R., Shine R. sembly of the Northern Territory, Sessional Com- 2010a. Something different for dinner? Responses mittee on Environment and Sustainable Develop- of a native Australian predator (the keelback ment. snake) to an invasive prey species (the cane toad). Hutchings R. W. 1979. A native predator of the cane Biological Invasions 12:1045–1051. toad (Bufo marinus). North Queensland Naturalist Llewelyn J., Webb J.K.W., Schwarzkopf L., Alford R., 45:4–5. Shine R. 2010b. Behavioural responses of carniv- Ingle D., McKinley D. 1978. Effects of stimulus con- orous marsupials (Planigale maculata) to toxic in- figuration on elicited prey catching by the marine vasive cane toads (Bufo marinus). Austral Ecology toad Bufo marinus. Animal Behaviour 26(3):885– doi: 10.1111/j.1442-9993.2009.02067.x. 891. Losos J. B., Greene H. W. 1988. Ecological and evo- Ingram G. J., Covacevich J. 1990. Tropidonophis mairii lutionary implications of diet in monitor lizards. vs. Bufo marinus. Memoirs of the Queensland Museum Biological Journal of the Linnean Society 35:379–407. 29:396. Mack R. N., Simberloff D., Lonsdale W. M., Evans H., Janzen D. H. 1973. Sweep samples of tropical foliage Clout M., Bazzaz F. 2000. Biotic invasions: causes, insects: effects of seasons, vegetation types, eleva- epidemiology, global consequences and control. tion, time of day, and insularity. Ecology 54(3):687– Issues in Ecology 5:1–20. 708. Madsen T., Shine R., Olsson M., Wittzell H. 1999. Res- 288 THE QUARTERLY REVIEW OF BIOLOGY Volume 85

toration of an inbred adder population. Nature 402: C. R. 1983a. Taste aversion conditioning of crows 34–35. to control predation on eggs. Science 220:212–214. Madsen T., Shine R. 1999. Rainfall and rats: climatically- Nicolaus L. K., Hoffman T. E., Gustavson C. R. 1983b. driven dynamics of a tropical rodent population. Taste aversion conditioning in free ranging rac- Australian Journal of Ecology 24(1):80–89. coons (Procyon lotor). Northwest Science 56(3):165– Madsen T., Shine R. 2000. Rain, fish and snakes: 169. climatically driven population dynamics of Ara- Oakwood M. 2003a. The Effect of Cane Toads on a Marsu- fura filesnakes in tropical Australia. Oecologia 124: pial Carnivore, the Northern Quoll, Dasyurus hallucatus. 208–215. Unpublished Progress Report February 2003. Dar- Madsen T., Ujvari B., Shine R., Olsson M. 2006. Rain, win (Australia): Parks Australia North. rats and pythons: climate-driven population dy- Oakwood M. 2003b. The Effect of Cane Toads on a namics of predators and prey in tropical Australia. Marsupial Carnivore, the Northern Quoll, Dasyurus Austral Ecology 31(1):30–37. hallucatus. Unpublished Progress Report, August McIntosh A. R., Townsend C. R. 1996. Interactions 2003. Darwin (Australia): Parks Australia North. between fish, grazing invertebrates and algae in a O’Donnell S. 2009. Alien versus predator: taste aver- New Zealand stream: a trophic cascade mediated sion learning as a way of enhancing survival in by fish-induced changes to grazer behaviour? Oe- Northern Quolls (Dasyurus hallucatus). Honours cologia 108(1):174–181. thesis. Sydney (Australia): School of Biological Sci- McKilligan N. G. 1984. The food and feeding ecology ences, University of Sydney. of the cattle egret, Ardeola ibis, when nesting in O’Donnell S., Webb J. K., Shine R. 2010. Conditioned southeast Queensland. Australian Wildlife Research taste aversion enhances the survival of an endan- 11:133–144. gered predator imperiled by a toxic invader. Jour- Meyer K., Linde H. 1971. Collection of toad venoms nal of Applied Ecology 47:558–565. and chemistry of toad venom steroids. Pages 521– Pace M. L., Cole J. J., Carpenter S. R., Kitchell J. F. 556 in Venomous Animals and Their Venoms, Volume 1999. Trophic cascades revealed in diverse ecosys- II: Venomous Vertebrates, edited by W. Bucherl and tems. Trends in Ecology and Evolution 14(12):483– E. E. Buckley. New York: Academic Press. 488. Mirtschin P. J., Davis R. 1982. Dangerous Snakes of Pearse B. W. 1980. Effects of feeding on Bufo marinus by Australia. Adelaide (Australia): Rigby. native and exotic fishes. Report to School of AES. Mitchell D., Jones A., Hero J. M. 1995. Predation on Brisbane (Australia): Griffith University. the cane toad (Bufo marinus) by the black kite Pearson D. J., Greenlees M., Ward-Fear G., Shine R. (Milvus nigrans). Memoirs of the Queensland Museum 2009. Predicting the ecological impact of cane 38:512. toads (Bufo marinus) on camaenid land snails in Mokany A., Shine R. 2003a. Biological warfare in the north-western Australia. Wildlife Research 36(6): garden pond: tadpoles suppress the growth of 533–540. mosquito larvae. Ecological Entomology 28(1):102– Pengilley R. 1981. Observations on the Cane Toad (Bufo 108. marinus) in the Southern Gulf Country, Queensland. Mokany A., Shine R. 2003b. Oviposition site selection by Internal Report. Darwin (Australia): Conservation mosquitoes is affected by cues from conspecific lar- Commission of the Northern Territory. vae and anuran tadpoles. Austral Ecology 28:33–37. Phillips B., Fitzgerald M. 2004. Encounters between Murray B. R., Hose G. C. 2005. Life-history and eco- eastern brown snakes (Pseudonaja textilis) and cane logical correlates of decline and extinction in the toads (Bufo marinus) in northern New South endemic Australian frog fauna. Austral Ecology Wales. Herpetofauna 34:23–25. 30(5):564–571. Phillips B. L., Brown G. P., Shine R. 2003. Assessing Nelson D. W. M. 2008. Bufo for breakfast: how cane the potential impact of cane toads on Australian toads affect native predators and their prey. B.Sc. snakes. Conservation Biology 17(6):1738–1747. Honours thesis. Sydney (Australia): School of Bi- Phillips B. L., Brown G. P., Shine R. 2004. Assessing ological Sciences, University of Sydney. the potential for an evolutionary response to rapid Nelson D., Crossland M. R., Shine R. 2010a. Indirect environmental change: invasive toads and an Aus- ecological impacts of an invasive toad on predator- tralian snake. Evolutionary Ecology Research 6:799– prey interactions among native species. Biological 811. Invasions doi: 10.1007/s10530-010-9729-4. Phillips B. L., Greenlees M. J., Brown G. P., Shine R. Nelson D. W. M., Crossland M. R., Shine R. 2010b. 2010b. Predator behaviour and morphology me- Foraging responses of predators to novel toxic prey: diates the impact of an invasive species: cane toads effects of predator learning and relative prey abun- and death adders in Australia. Animal Conservation dance. Oikos doi: 10.1111/j.1600-0706.2010.18736.x. 13(1):53–59. Nicolaus L. K., Cassel J. F., Carlson R. B., Gustavson Phillips B. L., Kelehear C., Pizzatto L., Brown G. P., September 2010ECOLOGICAL IMPACT OF INVASIVE TOADS 289

Barton D., Shine R. 2010a. Parasites and patho- ication in dogs: 94 cases (1997-1998). Journal of the gens lag behind their host during periods of host American Veterinary Medical Association 216(12): range-advance. Ecology 91:872–881. 1941–1944. Phillips B. L., Shine R. 2004. Adapting to an invasive Roemer G. W., Donlan C. J., Courchamp F. 2002. species: toxic cane toads induce morphological Golden eagles, feral pigs, and insular carnivores: change in Australian snakes. Proceedings of the Na- how exotic species turn native predators into prey. tional Academy of Sciences USA 101(49):17150–17155. Proceedings of the National Academy of Sciences USA Phillips B. L., Shine R. 2006a. Allometry and selection 99(2):791–796. in a novel predator-prey system: Australian snakes Sakate M., Oliveira P. C. L. D. 2000. Toad envenom- and the invading cane toad. Oikos 112(1):122–130. ing in dogs: effects and treatment. Journal of Ven- Phillips B. L., Shine R. 2006b. Spatial and temporal omous Animals and Toxins 6(1):52–62. variation in the morphology (and thus, predicted Sanson A. 2010. No winners in death match. Northern impact) of an invasive species in Australia. Ecogra- Territory News Feb 15, 2010. phy 29(2):205–212. Schultze-Westrum T. G. 1970. Conservation in Papua Phillips B. L., Shine R. 2006c. An invasive species and New Guinea. Final Report on the 1970 World induces rapid adaptive change in a native preda- Wildlife Fund Mission. tor: cane toads and black snakes in Australia. Schwarzkopf L., Alford R. A. 1996. Desiccation and Proceedings of the Royal Society of London, Series B: shelter-site use in tropical amphibians: comparing Biological Sciences 273(1593):1545–1550. toads with physical models. Functional Ecology Pimentel D., Lach L., Zuniga R., Morrison D. 2000. 10(2):193–200. Environmental and economic costs of nonindige- Seabrook W. A. 1991. Range expansion of the intro- nous species in the United States. Bioscience 50(1): duced cane toad Bufo marinus in New South Wales. 53–65. Australian Zoologist 27:58–62. Pippet J. R. 1975. The marine toad, Bufo marinus,in Seabrook W. A. 1993. Habitat use of the cane toad Papua New Guinea. Papua New Guinea Agricultural Bufo marinus: implications for assessment of im- Journal 26:23–30. pact and control strategies. Ph.D. thesis. Sydney Pizzatto L., Child T., Shine R. 2008. Why be diurnal? (Australia): University of Sydney. Shifts in activity time enable young cane toads to Seigel R. A. 1993. Summary: future research on evade cannibalistic conspecifics. Behavioral Ecology snakes, or how to combat “lizard envy.” Pages 395– 19(5):990–997. 402 in Snakes: Ecology and Behavior, edited by R. A. Pizzatto L., Shine R. 2008. The behavioral ecology of Seigel and J. T. Collins. New York: McGraw-Hill. cannibalism in cane toads (Bufo marinus). Behav- Settle W. H., Wilson L. T. 1990. Invasion by the var- ioral Ecology and Sociobiology 63:123–133. iegated leafhopper and biotic interactions: para- Pizzatto L., Shine R. 2009. Native Australian frogs sitism, competition, and apparent competition. avoid the scent of invasive cane toads. Austral Ecol- Ecology 71(4):1461–1470. ogy 34(1):77–82. Shine R., Brown G. P., Phillips B. L., Webb J. K., Pockley D. 1965. The free and the caged. Blackwoods Hagman M. 2006. The biology, impact and con- Magazine 1965:439–466. trol of cane toads: an overview of the University of Pramuk J. 2006. Phylogeny of South American Bufo Sydney’s research program. Pages 18–32 in Pro- (Anura: Bufonidae) inferred from combined evi- ceedings of the Cane Toad Workshop. Canberra (Aus- dence. Zoological Journal of the Linnean Society tralia): Invasive Animals Cooperative Research 146(3):407–452. Centre, University of Canberra. Price-Rees S. J., Brown G. P., Shine R. 2010. Are Shine R., Covacevich J. 1983. Ecology of highly ven- bluetongue lizards (Tiliqua scincoides intermedia, omous snakes of the Australian genus Oxyuranus Scincidae) threatened by the invasion of toxic (Elapidae). Journal of Herpetology 17(1):60–69. cane toads (Bufo marinus) through tropical Aus- Shine R., Greenlees M., Crossland M. R., Nelson D. tralia? Wildlife Research 37:166–173. 2009. The myth of the toad-eating frog. Frontiers in Rabor D. S. 1952. Preliminary notes on the giant toad Ecology and the Environment 7(7):359–361. Bufo marinus (Linn.) in the Philippine Islands. Smith J. G., Phillips B. L. 2006. Toxic tucker: the Copeia 1952:281–282. potential impact of cane toads on Australian rep- Rayward A. 1974. Giant toads: a threat to Australian tiles. Pacific Conservation Biology 12:40–49. wildlife. Wildlife 17:506–507. Somaweera R., Somaweera N., Shine R. 2010. Frogs Reeves M. P. 2004. A retrospective report of 90 dogs under friendly fire: how accurately can the gen- with suspected cane toad (Bufo marinus) toxicity. eral public recognize invasive species? Biological Australian Veterinary Journal 82(10):608–611. Conservation 143:1477–1484. Roberts B. K., Aronsohn M. G., Moses B. L., Burk Stammer D. 1981. Some notes on the cane toad (Bufo R. L., Toll J., Weeren R. 2000. Bufo marinus intox- marinus). Australian Journal of Herpetology 1:61. 290 THE QUARTERLY REVIEW OF BIOLOGY Volume 85

Sutherst R. W., Floyd R. B., Maywald G. F. 1996. The control of a toad pest. Australian Zoologist 20:501– potential geographical distribution of the cane 504. toad, Bufo marinus L. in Australia. Conservation Bi- van Dam R., Walden D., Begg G. 2002. A preliminary ology 10(1):294–299. risk assessment of cane toads in Kakadu National Park. Thomson D. M. 2006. Detecting the effects of intro- Supervising Scientist Report 164. Darwin (Austra- duced species: a case study of competition be- lia): Supervising Scientist Division. tween Apis and Bombus. Oikos 114(3):407–418. Ward-Fear G., Brown G. P., Greenlees M., Shine R. Townsend C. R. 1996. Invasion biology and ecological 2009. Maladaptive traits in invasive species: in Aus- impacts of brown trout Salmo trutta in New Zea- tralia, cane toads are more vulnerable to preda- land. Biological Conservation 78(1–2):13–22. tory ants than are native frogs. Functional Ecology Tyler M. J. 1975. The Cane Toad Bufo marinus:An 23(3):559–568. Historical Account and Modern Assessment. Report to Ward-Fear, G., Brown G. P., Shine R. 2010a. Factors the Vermin and Noxious Weeds Destruction affecting the vulnerability of cane toads (Bufo ma- Board of Victoria and the Agriculture Protection rinus) to predation by ants. Biological Journal of the Board of Western Australia, Melbourne (Austra- Linnean Society 99:738–751. lia). Ward-Fear, G., Brown G. P., Shine, R. 2010b. Using a Tyler M. J. 1987. Frog and cane toad skin secretions. native predator (the meat ant, Iridomyrmex rebur- Pages 329–339 in Toxic Plants and Animals: A Guide rus) to reduce the abundance of an invasive for Australia, edited by J. Covacevich et al. Brisbane species (the cane toad, Bufo marinus) in tropical (Australia): Queensland Museum. Australia. Journal of Applied Ecology 47(2):273–280. Tyler M. J. 1994. Australian Frogs: A Natural History. Wassersug R. J. 1971. On the comparative palatability Sydney (Australia): Reed Publishers. of some dry-season tadpoles from Costa Rica. Tyler M. J., Cappo M. 1983. Diet and Feeding Habits of American Midland Naturalist 86:101–109. Frogs of the Magela Creek System. Unpublished Open Waterhouse D. F. 1974. The biological control of File Record 10. Canberra (Australia): Supervising dung. Scientific American 230:101–109. Scientist for the Alligator Rivers Region. Watson M., Woinarski J. 2003a. A Preliminary Assess- Ujvari B., Madsen T. 2009. Invasion of cane toads ment of Impacts of Cane Toads on Terrestrial Vertebrate associates with a significant increase in mortality Fauna in Kakadu National Park. Unpublished Re- in a naı¨ve Australian varanid lizard. Herpetological port, February 2003. Jabiru (Australia): Kakadu Conservation and Biology 4:248–251. Research Advisory Committee. Ujvari B., Shine R., Madsen T. 2010. Detecting the im- Watson M., Woinarski J. 2003b. Vertebrate Monitoring pact of invasive species on native fauna: cane toads and Re-Sampling in Kakadu National Park, 2002. (Bufo marinus), frillneck lizards (Chlamydosaurus kin- Project RS10. Unpublished Report, March 2003. gii), and the importance of spatial replication. Aus- tral Ecology doi: 10.1111/j.1442-9993.2010.02126.x. Jabiru (Australia): Parks Australia. Webb J. K., Brown G. P., Child T., Greenlees M. J., Urban M. C., Phillips B. L., Skelly D. K., Shine R. 2007. The cane toad’s (Chaunus marinus) increas- Phillips B. L., Shine R. 2008. A native dasyurid ing ability to invade Australia is revealed by a predator (common planigale, Planigale maculata) dynamically updated range model. Proceedings of rapidly learns to avoid a toxic invader. Austral the Royal Society of London, Series B: Biological Sciences Ecology 33(7):821–829. 274(1616):1413–1419. Webb J. K., Glanznig A. 2004. Predation, Competition Urban M. C., Phillips B. L., Skelly D. K., Shine R. and Lethal Toxic Ingestion Caused by Cane Toads 2008. A toad more traveled: the heterogeneous (Bufo marinus). Key Threatening Process Nomi- invasion dynamics of cane toads in Australia. Amer- nation to Commonwealth Government from ican Naturalist 171(3):E134–E148. WWF-Australia. Valero A., Garcia C. M., Magurran A. E. 2008. Het- Webb J. K., Shine R., Christian K. A. 2005. Does erospecific harassment of native endangered fishes intraspecific niche partitioning in a native preda- by invasive guppies in Mexico. Biology Letters 4(2): tor influence its response to invasion by a toxic 149–152. prey species? Austral Ecology 30(2):201–209. van Beurden E. K. 1980a. Report on the Results of Stage Werren G. L., Trenerry M. P.1993. Size and diet of Bufo 3 of an Ecological and Physiological Study of the marinus in rainforest of north-eastern Queensland. Queensland Cane Toad, Bufo marinus. Canberra Memoirs of the Queensland Museum 34:240. (Australia): Report of the Australian National Western Australian Department of Environment and Parks and Wildlife Services. Conservation. 2009. Draft Cane Toad Strategy for van Beurden E. K. 1980b. Mosquitoes (Mimomyia el- Western Australia 2009–2019. Perth: Government egans [Taylor]) feeding on the introduced cane of Western Australia. Kensington (Australia): De- toad Bufo marinus (Linnaeus): implications for partment of Environment and Conservation. September 2010ECOLOGICAL IMPACT OF INVASIVE TOADS 291

White A. 2008. Predation of frogs in northern Austra- Natural Resources, Environment and the Arts, lia by feral pigs. Herpetofauna 38(1):35–38. Darwin (Australia). Williamson I. 1999. Competition between the larvae Wonham M. J., O’Connor M., Harley C. D. G. 2005. of the introduced cane toad Bufo marinus (Anura: Positive effects of a dominant invader on intro- Bufonidae) and native anurans from the Darling duced and native mudflat species. Marine Ecology Downs area of southern Queensland. Australian Progress Series 289:109–116. Journal of Ecology 24(6):636–643. Wootton J. T. 1994. The nature and consequences of Wilson D. 2005. Toads and native fish. Australia New indirect effects in ecological communities. Annual Guinea Fishes Association News 25:8–9. Review of Ecology and Systematics 25:443–466. Woinarski J. C. Z., Armstrong A., Price O., McCartney J., Zug G. R., Lindgren E., Pippet J. R. 1975. Distribu- Griffths A. D., Fisher A. 2004. The terrestrial vertebrate tion and ecology of the marine toad, Bufo mari- fauna of Litchfield National Park, Northern Territory: nus, in Papua New Guinea. Pacific Science 29(1): monitoring over a 6-year period and response to fire 31–50. history. Wildlife Research 31(6):587–596. Zug G. R., Zug P. B. 1979. The marine toad, Bufo Woinarski J. C. Z., Milne D. J., Wanganeen G. 2001. marinus: a natural history resume´ of native popu- Changes in mammal populations in relatively intact lations. Smithsonian Contributions to Zoology 284: landscapes of Kakadu National Park, Northern Terri- 1–58. tory, Australia. Austral Ecology 26(4):360–370. Zupanovic Z., Lopez G., Hyatt A. D., Green B., Ba- Woinarski J. C. Z., Oakwood M., Winter J., Burnett S., trtran G., Parkes H., Whitttington R. J., Speare R. Milne D., Foster P., Myles H., Holmes B. 2008. 1998. Giant toads Bufo marinus in Australia and Surviving the Toads: Patterns of Persistence of the North- Venezuela have antibodies against ‘ranaviruses.’ ern Quoll Dasyurus hallucatus in Queensland. Diseases of Aquatic Organisms 32(1):1–8. Report to the Natural Heritage Trust Strategic Reserve Program, March 2008. Department of Handling Editor: John J. Wiens