1. Clay model studies are a popular tool to identify predator-prey interactions that are challenging to observe directly in the field. But despite its wide use, the method’s applicability is limited by its low taxonomic resolution. Attack marks on clay models are usually identified visually, which only allows classification into higher taxonomic levels of predators. Thus, the method is often biased, lacks proof and, above all, standardization.
2. Here, we tested whether precise identification of attackers can be provided by amplification and sequencing of mitochondrial DNA left in bite marks on clay models. We validated our approach in a controlled laboratory study as well as in a field experiment using clay models of a common European amphibian, the European fire salamander (Salamandra salamandra). DNA based taxonomic assignments were additionally compared to visual assessments of bite marks.
3. We show that trace DNA of attackers can be routinely isolated and sequenced from bite marks, providing accurate species level classification. In contrast, visual identification alone yielded a high number of unassigned predator taxa. We also highlight the sensitivity of the method and show likely sources of contamination as well as probable cases of secondary and indirect predation.
4. Our standardized approach for species level attacker identification opens up new possibilities far beyond the standard use of clay models to date, including food web studies at unprecedented detail, invasive species monitoring as well as biodiversity inventories.

Amplicon sequencing data, collected from clay models of salamanders