Choose your meals carefully if you need to coexist with a toxic invader

Invasive species are a major threat to biodiversity and can drive native taxa to extinction or long-term declines. Some native species persist and eventually recover in the presence of invaders, potentially through behavioural avoidance of novel threats, physiological tolerance, or morphological changes that reduce vulnerability. Cane toads (Rhinella marina) have spread across Australia for ~85 years, severely impacting large frog-eating predators, including monitor lizards (Varanus spp.) that are fatally poisoned by ingesting toads. Prior work shows native predators such as blacksnakes can exhibit behavioural, physiological, and morphological shifts after toad arrival, but evidence is mixed for monitors. Two studies in eastern Australia reported that monitors in toad-infested areas refuse to consume toad flesh, suggesting taste aversion may enable coexistence more than morphological or physiological adaptation. However, broadscale evidence across the invasion chronosequence has been lacking. This study tested behavioural responses of two monitor species (V. panoptes and V. varius) to toad legs versus non-toxic foods across transects spanning toad-invaded and uninvaded areas, predicting that lizards from toad-colonised regions would avoid toxic prey.

The literature indicates invasive species are a leading cause of animal extinctions and long-term ecological change. In Australia, cane toads have caused rapid declines in susceptible predators. Native predators can adapt via multiple pathways: behavioural avoidance (e.g., conditioned taste aversion), physiological tolerance to toxins, and morphological changes that reduce fatal ingestion risk. For snakes, cane toad invasions have been associated with rapid evolution of reduced head size and altered toxin tolerance, reducing vulnerability. For varanid lizards, ingestion of toads often proves fatal and leads to population crashes; training to induce aversion can improve survival. Prior localized studies suggested monitors in toad-present areas refuse to eat toad flesh, pointing to learned or innate taste aversion rather than evolved toxin resistance or morphological constraints. Nonetheless, occasional toad consumption in long-invaded areas has been observed, and aversions may wane or be context-dependent. Broadscale, standardized tests across invasion timelines were needed to generalize these findings.

Study system and sites: Cane toads were released on Australia’s east coast in 1935 and now occupy over 2.1 million km². The study targeted two large monitor lizards: lace monitors (Varanus varius) along the east coast and yellow-spotted monitors (Varanus panoptes) across northern Australia. Forty-one sites were selected to span the invasion chronosequence and include yet-to-be-invaded regions (further south for V. varius and west for V. panoptes). Lace monitors: During the 2018–2019 Austral summer, trials were conducted at 15 sites (reported as 5 toad-absent and 12 toad-present). Thirty-six individual lizards were selected for trials (14 from toad-absent, 22 from toad-present sites). When approached, lizards typically climbed trees; at the base, a black metal tray (38 × 26.5 cm) was placed with two food items: a chicken egg (mean ± SE 27.5 ± 2.7 g) and a road-killed toad leg (~15.8 g), with items randomly oriented on either side. A remote-sensing camera (Scoutguard SG560K) was set 1.5 m away; the observer left and returned after 1 hour. Yellow-spotted monitors: In 2019 across the wet/dry tropics, 24 sites were used (6 uninvaded, 18 invaded for 6–84 years). At each site, 16 bait units were deployed (total 384), spaced 100 m apart and left for 48–72 hours over four sessions. A non-consumable lure (80 g sardines in oil) in a PVC canister was attached to a star picket. A consumable chicken egg was placed at the base, and a sardine and a toad leg were placed 30 cm to either side under mesh plastic lids that lizards could flip (excluding smaller scavengers). Remote cameras monitored stations. Individual lizards were identified by morphology to ensure only the first interaction per lizard with a toad leg was analyzed. Equipment was cleaned with ethanol between trials. Data and measures: For each encounter, investigators recorded investigation (tongue-flicking) and consumption (consumed vs not) of toad legs and non-toxic baits (chicken neck/egg or sardine depending on species/protocol). Figure descriptions note that lace monitors were offered a chicken neck or toad leg, and yellow-spotted monitors were offered a chicken egg, sardine, or toad leg. Statistical analysis: Generalized Linear Models with binomial errors and logit link were used to test changes in feeding responses over time. To address quasi-complete separation, Firth-adjusted maximum likelihood estimation was applied. Analyses were run separately by species and prey type (toad leg vs non-toxic bait). Predictors included toad status (presence/absence) and invasion time categories (uninvaded, recently invaded, mid-term, long invaded; based on generation times). An additional analysis in effort-ideal or toad-present areas tested preference between toad legs and non-toxic baits. JMP Pro 14.2 was used. Ethics and permits: University of Sydney ethics approval 2017/1202; permits NSW SLI10977, QLD WITK186321571, NT 63850, WA 008237-2, Commonwealth RK924.

• Across both species, of monitors that consumed at least one prey type, 96% consumed non-toxic control baits, whereas toad legs were consumed by 60% of lizards at toad-free sites and 0% at toad-invaded sites.
• Lace monitors: Of 36 selected, 8 did not interact; among the remaining 28 trials, all investigated and all consumed the chicken neck/egg. At toad-free sites, 60% consumed toad legs; at toad-present sites, 0% consumed toad legs. One toad-naïve individual bit the leg repeatedly over ~40 minutes without consuming it.
• Yellow-spotted monitors: In 43 trials, 40 consumed the non-toxic bait. Only lizards from toad-free areas consumed toad legs; three individuals ate the toxic bait while refusing the non-toxic item.
• Statistical results: Consumption of non-toxic prey did not change with duration of sympatry with toads (lace monitor χ² = 0.03, P = 0.88). Toad presence strongly reduced the probability of consuming a toad leg (lace monitor χ² = 14.69, P < 0.0007). Even at toad-free sites, non-toxic foods were preferred over toad legs (lace monitor χ² = 5.65, P = 0.017; yellow-spotted monitor χ² = 99.7, P = 0.0016). Duration of toad occupancy showed no relationship with responses to toad legs (χ² = 0.00, P = 1.0 for both species).
• Cameras never recorded toad-leg consumption in toad-present areas, though one V. panoptes was observed carrying a road-killed toad; consumption was not confirmed.

Findings support that conditioned taste aversion or learned avoidance underpins the persistence of vulnerable monitor predators after cane toad invasion. Initial willingness to consume toads appears to be eliminated soon after exposure to toads and persists for decades (at least 80 years), enabling coexistence despite the high lethality of toad toxins. Notably, a substantial fraction (~40%) of toad-naïve monitors already refused toad legs while readily consuming non-toxic foods, suggesting innate caution or general neophobia toward unfamiliar prey may contribute. Prior reports of occasional toad consumption in long-invaded areas align with observations that handling and cautious assessment can lead to rejection and that aversions may sometimes be inconsistent. The consistent avoidance across the invasion timeline implies that monitors have not evolved physiological toxin resistance; strong behavioural avoidance would relax selection for such traits. Differences in population impacts between lace monitors and yellow-spotted monitors may reflect differences in acquisition or retention of aversion across encounters or life stages. More broadly, behavioural avoidance can mitigate invader impacts across taxa, but population outcomes may still depend on age-specific responses and exposure histories.

This continent-scale, standardized assessment shows that Australian monitor lizards commonly consume non-toxic foods but avoid cane toad flesh in areas invaded by toads, with avoidance evident across decades since invasion. The ability to recognise and reject toads as prey likely allows these predators to persist alongside a toxic invader without evolving physiological resistance. The results highlight behavioural plasticity as a key mechanism enabling coexistence with invasive species. Future work should experimentally disentangle learning versus heritable components of aversion, assess age-class and individual variation in acquisition and retention of aversion, test management interventions such as conditioned taste aversion training at landscape scales, and link behavioural responses to demographic outcomes over time.

The study is observational rather than experimental, limiting inference about proximate mechanisms (learned versus heritable aversion). Bait-station methods and camera observations may not capture all foraging contexts or rare consumption events. Some methodological inconsistencies (e.g., control bait types described) and site counts in text may reflect reporting errors. Occasional anecdotal observations of monitors handling or carrying toads indicate that aversion may not be absolute. Results from selected sites and time windows may not generalize to all populations or seasons.