To be (cryptic) or not to be? Variation in detectability by prey explains the diet of an ambush predator

Differences in detection of cryptic predators by different animal groups are largely unstudied, limiting our understanding of how variation in detectability may affect the diet of ambush predators and their interactions with other animals. Using a combination of radiotelemetry and fixed videography, we monitored encounters between Bothrops snakes and other vertebrates in an Atlantic Forest community of Brazil. We also assessed snake detection by animal types and characterized snake feeding habits using multiple data sources. Our results showed that animal encounters occurred more frequently during daytime than nighttime, because birds, which are largely diurnal, encountered the snakes more often than the typically nocturnal mammals. We found dramatic differences in snake detectability between animal types: birds detected snakes in 89.2% of encounters compared to only 5.3% for mammals. All predatory strikes targeted mammals and snakes never struck at birds. Further, Bothrops fed almost exclusively on mammals (96.3% of diet), despite birds representing 61% of simulated prey encounters. The detectability bias we observed provides a mechanistic explanation for why birds are severely underrepresented in Bothrops diet relative to their encounter frequency. Our findings indicate that predator detectability can be a key factor shaping the realized diet of cryptic ambush predators.