DNA Metabarcoding Complements but Does Not Replace Direct Observations of Penguin Predation by Corvids

Establishing methods that allow for more focused management of wildlife under predator pressure may increase the efficiency of managing problematic predators. Non-invasive dietary analysis and identification of conservation-sensitive prey in the diet of 'culprit' predator individuals could help to facilitate this and is worthy of exploration. Recently on Phillip Island, Australia, Little Ravens Corvus mellori have emerged as a prominent predator on the clutches of burrow-nesting Little Penguins Eudyptula minor. We tested the feasibility of using non-invasive PCR approaches targeting the penguin mitochondrial 16S rRNA marker gene to establish whether penguin DNA could be detected in raven faecal samples, potentially enabling the identification of culprit ravens missed by extensive field observation. Using a metabarcoding approach, we examined the feasibility of non-invasively establishing other dietary items via high-throughput amplicon sequencing. We documented components of raven diet using the universal mitochondrial 16S rRNA, insect-specific 'Chiar' 16S rRNA and plant ITS2. The assemblage of dietary items did not differ with raven culprit status (i.e. a raven previously observed preying upon penguin), sex or date. Penguin was detected in the diet of some individuals classified observationally as non-culprits. Although some cases may conceivably have been false detections, other explanations include missed depredation events, consumption via scavenging or through secondary consumption (e.g. eating invertebrates that have consumed penguin). While this study found metabarcoding unreliable for unambiguous assigning of raven culprit status, at least as we implemented it, it may hold promise complementing observations if consumption via scavenging can be distinguished from direct depredation.