Description:
<jats:title>Abstract</jats:title><jats:p>DNA metabarcoding is increasingly used to analyze the diet of arthropods, including spiders. However, high sensitivity to DNA contamination makes it difficult to apply to organisms obtained from mass‐sampling methods such as pitfall traps. An alternative is to hand‐sample spiders, but it is unclear how effectively this prevents external contamination, especially with new knowledge showing the wide spread of eDNA in the environment. Protocols using bleach to remove external DNA have been tested on several invertebrates, though testing with both mass‐sampling methods and spiders is lacking. Here, we used wolf spiders (Lycosidae) to assess the risk of external DNA contamination from pitfall trapping and hand sampling, and the efficacy of bleach decontamination. We first conducted a contamination experiment where we placed spiders in pitfall traps containing trapping medium and a nonprey insect species to simulate external DNA contamination. We also compared sampling methods by collecting spiders using pitfall traps and hand sampling. Spiders from the contamination experiment and sampling method comparison were either bleached or untreated, then metabarcoded using multiple primer pairs. The contamination experiment resulted in the contamination of almost all spiders from pitfall traps, which was successfully eliminated with bleaching. Interestingly, there was no difference in the number of amplicon sequence variants (ASVs) detected per spider between pitfall trapping and hand sampling but bleaching resulted in significantly fewer ASV detections for both methods. Additionally, bleaching, but not sampling method, affected the taxonomic diet composition for both hand‐sampled and pitfall‐trapped spiders, indicating similar levels of external contamination. Our results are the first to confirm that DNA metabarcoding can be used together with bleaching for spiders sampled from pitfall traps, and that hand sampling does not necessarily exclude external DNA contamination. Thus, diet studies using metabarcoding should address the risk of external contamination with field‐sampled arthropods, regardless of sampling method.</jats:p>