These results infer that, in a random sample of scats, consumption of test prey is more likely to be detected using DNA than hard parts because excretion of prey DNA not only occurs in a higher proportion of scats than hard parts, but is also more consistent over time.