1. Laboratory experiments have shown that the viability of embryos of the invasive cane toad (Rhinella marina) can be reduced by exposure to chemical cues from older conspecific larvae. These effects (very strong in laboratory trials) may offer an exciting new approach to controlling this problematic invasive species in Australia. However, the degree to which the method works in natural environments has yet to be assessed.

2. Our experiments in the laboratory and in semi-natural outdoor waterbodies show that chemical cues from tadpoles do indeed suppress the growth, development and survival of conspecific larvae that are exposed as embryos and do so in a dose-dependent manner; higher tadpole densities cause greater suppression of embryos.

3. In semi-natural outdoor waterbodies, suppressor-exposed tadpoles were less than half as likely to survive to metamorphosis as were controls, and were much smaller when they did so and hence, less likely to survive the metamorph stage. Additionally, female cane toads were less likely to oviposit in a waterbody containing free-ranging (but not cage-enclosed) tadpoles, suggesting that the presence of tadpoles (rather than the chemical cues they produce) may discourage oviposition.

4. Broadly, our results suggest that the suppression effect documented in laboratory studies does indeed occur in the field also, and hence that we may be able to translate that approach to develop new and more effective ways to reduce rates of recruitment of peri-urban populations of cane toads in their invasive range.

Field trials of cane toad suppression pheromones

2.1  |  Experimental protocols

2.1.1  |  Laboratory trials on suppressor density

Adult cane toads were collected by hand from Middle Point, Northern Territory (-12.579602, 131.313863) and brought back to a nearby laboratory where they were injected with leuprorelin acetate (Lucrin, Abbott Australasia) to induce breeding (see Hayes et al., 2009 for detailed methods). Newly laid clutches of eggs were placed in individual 18 L tubs in unchlorinated water at a constant temperature (26 °C) and aerated until they reached developmental stage 18 (Gosner, 1960).

Four circular pools (2200 mm diameter) were filled with 1800 L of unchlorinated water. An enclosed mesh net (700 x 400 x 300 mm, mesh size 1 x 1 mm) was placed into the centre of each pool and secured so that the top of the enclosure was 5 cm above the water surface. Twelve hours before the experiment began, each net was filled with 300, 30, 3 or zero (control) live cane toad tadpoles (stage 28–34) from an older Middle Point clutch.

When the experimental clutch reached Gosner stage 18, 10 of the hatchlings were placed in each of 32 enclosed circular plastic containers (diameter 76 mm, height 24 mm) with mesh sides to allow water flow-through. Eight of these containers were then placed into each of the four pools, and weighted so that they sat at the bottom of the water column. Four of the containers were positioned at the edge of the pool (75 cm from the tadpole net, = ‘far’) and four were placed beside the net (‘near’). The containers were left in the pools for 48 h (until hatchlings reached stage 25) to allow exposure to cues from the netted tadpoles. They were then removed and five of the now free-swimming tadpoles from each experimental container were randomly selected and placed together into a clean 1 L plastic container filled with 750 mL of fresh unchlorinated water. The tadpoles were fed crushed algal wafers daily, and water was changed every second day. Ten days later, the tadpoles were weighed and their developmental stages recorded. We ran this experiment three times, using three different clutches of hatchlings, and exposing them to three different clutches of tadpoles.

2.1.2  |  Field and laboratory trials on effects of suppression

Two clutches of eggs were obtained from cane toads collected from Kununurra, and raised as described above. When the tadpoles reached stage 18, 10 hatchlings were placed in each of eight enclosed circular plastic containers with mesh sides (diameter 76 mm, height 24 mm), and this was repeated for each clutch. For the field trials, two of these containers (both containing hatchlings from the same clutch) were then placed into each of the eight ponds (four ponds per clutch) and weighted to sit at the bottom of the water column (20 cm deep). Eight replicate ponds (5 x 4 m, 1 m depth at deepest end, gradient to 0 m at opposite end) were dug 2 m apart, in a clay-based depression in bushland 15 km from Kununurra, Western Australia (-15.827949, 128.856982; Figure 1). The ponds were lined with plastic sheeting (100 μm thick) to help retain water, and covered with 20 mm of natural sediment and 28 L of benthic leaf litter sourced from a nearby waterbody. Ponds were each filled with 7500 L of water, sourced from the local Lake Kununurra (from an area where toads do not breed), and given 48 h to settle. An enclosed mesh net (400 x 300 x 300 mm, mesh size 1 x 1 mm) was placed into each pond and secured so that the top of the enclosure was 5 cm above the water surface. Twelve hours before the experiment began, half of the nets were filled with 30 cane toad tadpoles (Gosner stage 28–34) field-caught from two local populations and half of the nets were left empty (controls). This suppression treatment equalled a density of 0.004 tadpoles/L, falling between the “3 tadpole” (0.002 tadpoles/L) and “30 tadpole” (0.02 tadpoles/L) treatments in the previous laboratory experiment.

For the laboratory trials, another 10 hatchlings were placed into each of 10 plastic aquaria (1 L) containing 750 mL of unchlorinated water, and this was repeated for each clutch. These were assigned to either a ‘suppression’ or ‘control’ treatment. We placed a flyscreen mesh enclosure (60 x 40 x 30 mm, mesh holes 1 x 1 mm) into each aquarium, and added two live cane toad tadpoles (one from each of the wild populations used for the pond enclosures) to the container for each of the ‘suppression’ aquaria. In ‘control’ aquaria the mesh container remained empty. The aquaria were kept in the laboratory at 30 °C.

The pond and laboratory treatments were both left for 48 h (until the hatchlings reached stage 25, when they become capable of swimming) to allow exposure to tadpole-derived cues, after which they were removed. From each container five of the now free-swimming tadpoles were randomly selected and placed into a clean 1 L plastic container filled with 750 mL of fresh unchlorinated water. The tadpoles were fed crushed algal wafers daily and water was changed every second day. Ten days later, tadpoles were weighed and their developmental stages recorded. They were then returned to the laboratory and raised until they either died or reached metamorphosis. If metamorphosis was reached, the days taken to reach metamorphosis and the size of the metamorph at emergence were recorded.

This experiment was repeated with a further two clutches, allowing enough time for the previous suppression cues to no longer be present in the ponds (Clarke et al. 2015, 2016), and using different suppressor tadpoles. Although measurements were taken at 10 days growth for all clutches tested, time constraints meant that only the first two clutches were run to metamorphosis.

2.1.3  |  Trials of oviposition behaviour

We erected walls of plastic sheeting (500 mm high) across the middle of each of the ponds, to divide each pond evenly into two. Each half of the original pond held 2500 L, with no water flow between the two sides but a common bank. Twelve hours before the experiment began we placed a mesh enclosure (200 x 200 x 150 mm) into the middle of each half-pond. One side was randomly allocated to the ‘suppression’ treatment and the other side left as a control. In the suppression sides, 30 tadpoles (stage 28–34) from a mixture of two wild clutches were placed into the mesh enclosure. Fences (600 mm tall) around each pond (enclosing both of the half-ponds as a single unit) excluded any wild cane toads.

Adult cane toads were collected by hand from around Kununurra, and were injected with leuprorelin acetate (Lucrin, Abbott Australasia) to induce breeding. One female and two males were then placed inside the fence of each pond and left overnight. The next morning we collected any clutches laid in the ponds and recorded the side (half-pond) in which they had been laid. If no eggs had been laid the replicate was removed from the data set. This experiment was repeated (with fresh adult toads and fresh suppressor tadpoles) until 10 clutches had been laid.

We then repeated the experiment but instead of restraining the suppression tadpoles (‘enclosed suppression’), the mesh enclosures were removed and suppression tadpoles were placed directly into the pond, allowing them to swim freely (‘non-enclosed suppression’). The control side of the pond contained no tadpoles. This experiment was repeated with fresh toads and fresh suppressor tadpoles until seven clutches had been laid (the work was then terminated because of unsuitable weather).

2.2  |  Statistical analysis

 Data were normally distributed, so we used parametric tests.

2.2.1  |  Laboratory trials on suppressor density

The position of embryos in the pond (near or far) had no effect on mass or stage of tadpoles (ANOVA with location as factor, all p > .05), and so these treatments were combined for the analysis. The mass and stage of the five tadpoles raised in each 1 L container were averaged, to create one replicate. To test for an effect of suppressor density on the mass and stage of tadpoles, we ran two individual ANOVAs in SPSS v21 (IBM, Armonk, NY), with density treatment as the factor, and included clutch as a random factor. We then ran Tukey’s post-hoc tests to determine which treatments contributed to the differences observed.

2.2.2  |  Field and laboratory trials on effects of suppression

To test for an effect of suppression treatment on survival, mass and stage we ran two individual two-way ANOVAs in JMP v9 (SAS Institute, Cary, NC), with ‘Suppression treatment’ and ‘Location’ as the factors, and including ‘Clutch’ as a random factor. To test for an effect of suppression on the mass of metamorphs and on the days taken for individuals to reach metamorphosis, we ran separate ANOVAs in JMP with ‘Suppression treatment’ and ‘Location’ as the factors and including ‘Clutch’ as a random factor.

2.2.3  |  Trials of oviposition behaviour

To test for an effect of enclosed suppression treatment on the likelihood of toads laying in a waterbody, we ran a binomial test in SPSS comparing the number of times toads laid in the suppression treated side of the pond and the number of times they laid in the control side. We repeated this analysis to test for an effect of non-enclosed suppression treatment on the likelihood of toads laying in a waterbody.