The blame game: Using eDNA to identify species-specific tree browsing by red deer (Cervus elaphus) and roe deer (Capreolus capreolus) in a temperate forest

•eDNA can be applied successfully to elucidate red- and roe deer browsing patterns.•Neither habitat, tree species nor height influenced the DNA amplification success.•Silver fir and rowan were equally likely to be browsed by roe deer and red deer.•Differences in foraging habitat preferences were found among roe and red deer.•eDNA tool can influence hunting management in multi-ungulate communities. Increasing deer populations in many temperate regions can affect tree regeneration, resulting in severe long-term impacts on forest structure, composition and diversity. Of the most common deer species in Europe — red deer (Cervus elaphus) and roe deer (Capreolus capreolus) — roe deer are generally thought to have the highest impact on palatable tree species owing to their feeding niche. Although browsing and its potential consequences are well researched, less is known about the influence of specific deer species within multi-species ungulate communities on specific tree species. Environmental DNA (eDNA) allows the determination of species-specific browsing habits without the need for direct observations, facilitating effective targeting of management interventions. In this study eDNA was used to elucidate the browsing patterns of these two deer species in the temperate forest of the Bavarian Forest National Park, Germany and analysed the influence of tree species, management type and height of browsing, on the success rate of the method. Samples were collected from twigs used in feeding trials from enclosures containing red deer or roe deer and from naturally browsed twigs in three different management types within the national park. eDNA was successfully amplified from 98% of the feeding trial samples, and the correct deer species was identified for all samples. eDNA was successfully amplified from approximately 50% of the naturally browsed samples. Neither management type, tree species, nor height of browsing had any significant influence on the success of the method. For silver fir and rowan, no significant difference was found in the proportion of browsing events attributable to roe or red deer, when the two deer species occur at similar densities. These results indicate that roe deer might not always be disproportionately responsible for the browsing of palatable tree species as expected from its food niche. Roe deer were significantly more responsible for browsing at lower heights than red deer. Although not statistically significant, roe deer were more r