**Background**
Many organisms exhibit phenotypic plasticity, producing different phenotypes in response to environmental variation. Predator-induced plasticity is a well-known example where prey develop adaptive, predator-specific traits that reduce predation risk. The Neotropical tadpole *Dendropsophus ebraccatus* is a model for studying such plasticity because it develops opposing phenotypes when exposed to invertebrate predators (e.g., dragonfly larvae) versus vertebrate predators (e.g., fish). Dragonfly cues induce a large, deep red and black tail and reduced activity, while fish cues induce a shallower, achromatic tail and increased activity. These phenotypes are adaptive against the inducing predator but costly against the other. In nature, tadpoles co-occur with multiple predator species, yet little is known about how they respond to graded or mixed predator cues, especially in treefrogs (Hylidae). This study aimed to determine whether *D. ebraccatus* tadpoles can assess variation in cue concentration and how they respond to simultaneous cues from both fish and dragonfly nymphs.
**Methods**
Two experiments were conducted using a laboratory colony of *D. ebraccatus* at Vassar College. Experiment 1 (July–August 2017) used eight families, with 80 tadpoles per family divided into eight groups of 10 in 2 L of water. Treatments included four cue concentrations (none, low, medium, high) crossed with two predator types (mosquitofish *Gambusia affinis* or aeshnid dragonfly larvae). Cue water was generated by housing predators alone in 2 L and feeding them four tadpoles per day, creating a concentration of 2 tadpoles consumed per L per day. Experimental tadpoles received daily water changes: high treatment received 500 mL full-strength cue water (0.5 tadpoles/L/day), medium received 200 mL cue water + 300 mL artificial pond water (0.2 tadpoles/L/day), low received 50 mL cue water + 450 mL artificial pond water (0.05 tadpoles/L/day), and control received 500 mL artificial pond water. Initial tadpole size was 7.3 ± 0.3 mm total length; fish were 43.0 ± 2.6 mm, dragonfly larvae 34.5 ± 2.9 mm. Experiment 2 (September–November 2019) used six families, with 76–88 tadpoles per family divided into four groups of ~20 in 4 L tanks divided into thirds by mesh screens. Treatments were: no-predator control (CC), two fish (FF), two dragonfly nymphs (DD), or one fish and one dragonfly nymph (DF). Each predator was fed one tadpole per day. Initial tadpole size was 6.4 ± 0.7 mm; fish were 28.0 ± 4.0 mm, dragonfly larvae 20.1 ± 1.8 mm. Both experiments lasted 10 days. Tadpoles were then anesthetized with MS-222 and photographed. Morphology was analyzed using geometric morphometrics with 14 landmarks; tail spot area and coloration (hue, saturation, brightness) were measured from photographs. Data were analyzed using mixed effects models with family and tank as random effects. Principal components analysis (PCA) was used for shape and color data.
**Key Results**
In Experiment 1, tadpole morphology changed significantly with increasing cue concentration. PC1 and PC2 explained 42% of shape variation. There was a significant interaction between cue concentration and predator for both PC1 (χ² = 8.1, P = 0.004) and PC2 (χ² = 6.3, P = 0.012). Post-hoc tests showed that predator treatments differed significantly only at the high cue concentration (both P ≤ 0.003). Tail spot coloration showed more graded responses: PC1 (explaining 83% of color variation) showed a significant predator effect (χ² = 15.4, P < 0.001) and interaction (χ² = 19.30, P < 0.001). Post-hoc tests revealed that PC1 differed between predator treatments at medium and high concentrations (both P ≤ 0.03) and marginally at low concentration (P = 0.079). PC2 (hue) varied only by predator (χ² = 6.3, P = 0.01). In Experiment 2, morphology PC1 and PC2 explained 47% of shape variation. There was a significant effect of treatment on PC1 (χ² = 28.1, P < 0.001) and PC2 (χ² = 9.0, P = 0.029). Post-hoc tests showed that the DF and FF treatments did not differ for PC1 (P = 0.63), but all other pairwise comparisons were at least marginally significant (DD vs CC: P = 0.06; others P < 0.05). For PC2, only DD and FF differed (P = 0.038). Tail spot coloration PC1 (87% of variation) showed a significant treatment effect (χ² = 42.3, P < 0.001), with DF and FF not differing (P = 0.33), while all other treatments differed significantly (P ≤ 0.005). PC2 (hue) showed no treatment effect (χ² = 3.1, P = 0.38).
**Clinical Implications**
While this study is not clinical, it provides fundamental insights into how prey organisms assess and respond to multiple predation risks. The finding that tadpoles prioritize the more lethal predator (fish) when exposed to mixed cues suggests that prey can integrate information from multiple sources and produce adaptive phenotypes even in complex environments. This has implications for understanding the evolution of phenotypic plasticity and for predicting how species may respond to changes in predator communities, such as those caused by invasive species or habitat alteration. The graded response to cue concentration also indicates that prey can fine-tune their defenses based on perceived risk, which may help them balance the costs and benefits of plasticity. These results contribute to a broader understanding of predator-prey dynamics and the ecological and evolutionary consequences of phenotypic plasticity.