Detecting Environmental DNA From Crested and Smooth Newts—Not as Straight Forward as Filtering a Drop of Water

ABSTRACT Environmental DNA methodologies are constantly being developed and optimized for specific purposes. This paper summarizes 3 years of water‐sampling and filtering method development, with the aim to reliably detect and quantify eDNA from great crested newts and smooth newts from established monitoring ponds, using species‐specific assays and droplet digital PCR (ddPCR). Both newt species were caught in traps from all ponds during the three sampling years. For the first year, water was collected from three separate positions within each pond and filtered individually through a 0.45‐μm PES filter. Overall, the pond replicates resulted in 26% false negatives, including three ponds where only one‐third of the samples were positive. Positive results within sites also showed high variance in DNA concentration (the highest variance across three filters and nine ddPCRs: 1987 ± 1789). For the second year, subsamples of water from different positions were collected and mixed before being filtered through two 0.45‐μm PES filters. This sampling approach reduced the overall variation in eDNA concentration from replicates filtered from each water batch (the highest variance: 7355 ± 4147). However, a few samples did still not reliably detect newt eDNA (9%–10%). For the third year, to test for potential effects of stratification and seasonal variation on the probability of detection, subsamples were collected from both the surface layer and 30 cm below the surface four to five times through the main mating period. In addition, the filter type was changed to a 2.0‐μm glass fibre filter to increase water volume. Despite filtering on average 1 L more through each filter compared to the second sampling year, this method produced a higher number of false negatives. Optimizing filtering methods to reliably detect newts from ponds in the mating season is needed before applying the method in managemental monitoring.