We conducted a quantitative and qualitative chemical analysis of cane toad bufadienolides--the cardioactive steroids that are believed to be the principal cane toad toxins. We found complex shifts in toxin composition through toad ontogeny: (1) eggs contain at least 28 dominant bufadienolides, 17 of which are not detected in any other ontogenetic stage; (2) tadpoles present a simpler chemical profile with two to eight dominant bufadienolides; and (3) toxin diversity decreases during tadpole life but increases again after metamorphosis (larger metamorph/juvenile toads display five major bufadienolides). Total bufadienolide concentrations are highest in eggs (2.64 +/- 0.56 micromol/mg), decreasing during tadpole life stages (0.084 +/- 0.060 micromol/mg) before rising again after metamorphosis (2.35 +/- 0.45 micromol/mg). These variations in total bufadienolide levels correlate with toxicity to Australian frog species. For example, consumption of cane toad eggs killed tadpoles of two Australian frog species (Limnodynastes convexiusculus and Litoria rothii), whereas no tadpoles died after consuming late-stage cane toad tadpoles or small metamorphs. The high toxicity of toad eggs reflects components in the egg itself, not the surrounding jelly coat. Our results suggest a dramatic ontogenetic shift in the danger that toads pose to native predators, reflecting rapid changes in the types and amounts of toxins during toad development.

The early life history stages of anurans in the Family Bufonidae often possess chemicals that are noxious or toxic to predators. Predators with no evolutionary history of exposure to bufonids may be particularly susceptible to these toxins. We conducted a series of laboratory experiments to investigate the toxic effects of eggs, hatchlings and tadpoles ofthe introduced toad, Bufo marinus (Linnaeus), on native Australian aquatic predators. There was considerable interspecific and intraspecific variation in these effects. Bufo marinus were highly toxic to some predator species, but were readily consumed by other species without apparent ill effect. Interspecific variation in toxic effects was not related to predator feeding mode or the number of B. marinus ingested by predators, and there was no clear pattern of distribution of vulnerability among species within higher taxa. Intraspecific variation in responses to toxins may result from individual variation in the resistance of predators to B. marinus toxins, or from individual variation in toxicity among B. marinus. Some native species adversely affected by B. marinus appeared unable to detect and avoid B. marinus toxins. This may result from a general inability to assess the toxicity of food items or from a lack of evolutionary exposure to B. marinus toxins.

Adaptive developmental plasticity allows individuals to match their phenotype with their environment, increasing fitness where threats are inconsistently present. However, despite clear advantages of plasticity, adaptive traits are not ubiquitously nor infinitely plastic. Trade‐offs between benefits and costs or limits are therefore theoretically necessary to constrain the evolution of plastic responses. Systems in which extreme risk can be reliably detected are ideal for investigating trade‐offs, as even costly responses may be adaptive where risk is severe. Cane toads (Rhinella marina) are abundant in Australia and produce large clutches (frequently >10,000 eggs), but asynchronous breeding and rapid development result in variable larval densities within breeding pools. In the field, we found that cannibalism by older cohorts often reduces the survival of conspecific eggs and newly hatched pre‐feeding larvae (“hatchlings”) by >99%, as feeding larvae (“tadpoles”) use chemical cues from the relatively immobile hatchlings to locate and consume them. After hatchlings become free‐swimming, however, they cannot be cannibalized. Hatchlings can reduce this period of vulnerability by accelerating development when they detect cannibal cues. However, this developmental acceleration decreases initial tadpole mass, reduces subsequent survival, growth, and development, affects behavior, and compromises feeding structures. Reaction norms differ among clutches, and greater developmental acceleration is followed by greater impairment of larval function in plastic clutches, whereas nonresponsive clutches are unaffected by cue exposure. More plastic clutches ultimately exhibit both poorer performance and greater variation among siblings in exposed and (to a lesser degree) control treatments. Variation among clutches in tadpole viability is driven by differences in plasticity rather than phenotype; fitness reductions are linked to developmental acceleration, not rapid development per se. Clutches with intrinsically slow pre‐feeding developmental rates exhibit stronger acceleration (i.e., steeper reaction norms), but clutches with intrinsically rapid development reach invulnerable stages more quickly than those that accelerate development. As a result, high cannibalism risk may favor canalized rapid development rather than facultative developmental acceleration. Cannibalism plays an important role in the recruitment of this invasive species, and hatchling defenses against this threat demonstrate how the limits and costs associated with an inducible defense can favor canalized defenses over phenotypic plasticity.

Body size at metamorphosis is a key trait in species (such as many anurans) with biphasic life-histories. Experimental studies have shown that metamorph size is highly plastic, depending upon larval density and environmental conditions (e.g. temperature, food supply, water quality, chemical cues from conspecifics, predators and competitors). To test the hypothesis that this developmental plasticity is adaptive, or to determine if inducing plasticity can be used to control an invasive species, we need to know whether or not a metamorphosing anuran’s body size influences its subsequent viability. For logistical reasons, there are few data on this topic under field conditions. We studied cane toads (Rhinella marina) within their invasive Australian range. Metamorph body size is highly plastic in this species, and our laboratory studies showed that larger metamorphs had better locomotor performance (both on land and in the water), and were more adept at catching and consuming prey. In mark-recapture trials in outdoor enclosures, larger body size enhanced metamorph survival and growth rate under some seasonal conditions. Larger metamorphs maintained their size advantage over smaller siblings for at least a month. Our data support the critical but rarely-tested assumption that all else being equal, larger body size at metamorphosis is likely to enhance an individual’s long term viability. Thus, manipulations to reduce body size at metamorphosis in cane toads may help to reduce the ecological impact of this invasive species.

No abstract