Comparison of Three Methods for Measuring Dietary Composition of Plains Hog-nosed Snakes

Abstract

Wild animal diets are challenging to quantify, and the various methods for doing so have strengths and weaknesses. Combining multiple methods can allow ecologists to assess their level of confidence in particular results, increase sample size, and investigate diet over varying time scales. The biases of traditional gut content–based methods are mostly well understood. Newer methods may have important biases that can only be worked out through comparison to established ones. We collected data on the diet of wild Plains Hog-nosed Snakes (Heterodon nasicus) using multiple, fundamentally dissimilar methods, combined analytically using a Bayesian framework to describe an ontogenetic dietary shift. Gut contents were the most straightforward, but yielded a small sample size that fell below any reasonable threshold for making generalizations. Stable isotopes indicated an obvious ontogenetic dietary shift, but were labor-intensive, and conclusions are limited by multiple methodological caveats including similarity among prey groups, maternal carryover effects, and uncertainty in trophic enrichment factors. Fecal environmental DNA (eDNA) was intermediate in terms of effort, yielding results congruent with the other two methods, but the interpretation of which would likely have been confounded by contaminants had we not used all three methods in tandem. Several apparent artifacts are discussed. There are some reassuring similarities among methods. There are also several differences. The most complete picture uses data from all methods taken together. Future studies should attempt to compare the biases, expense, and potential drawbacks of these and other methods in greater detail.