Estimates of population density and diets of predators are often based on detection and examination of feces (scats). Decomposition rates and manipulation or consumption of scats by animals could bias studies that use scat to investigate life-history parameters. We investigated the longevity (persistence), nutritional content, and fates of coyote (Canis latrans) and bobcat (Lynx rufus) scats near Tucson, Arizona from April 2000–July 2001. We established 42 stations containing scats along roads in the Sonoran Desert. We placed half of the scats in protective cages to estimate degradation rates apart from disturbance by other animals. Unprotected scats disappeared sooner than expected based on decomposition rate. When pooled over spring, summer, autumn, and winter, median survival of unprotected wild canid scats was 11 days, but decreased to 4 days in June–July 2001. Some scats were moved to white-throated woodrat (Neotoma albigula) dens. A 1-m increase in distance of scats to woodrat dens resulted in a 2.2% (Wald 95% CI: 0.9–3.4%) increase in persistence (days) of unprotected scats of all types over all seasons. In June and July 2001 a 1-m increase in distance of scats to woodrat dens resulted in a 4% (Wald 95% CI: 2.7–5.4%) increase in persistence (days) of unprotected scats of wild canids. Studies based on detection, counts, or collection of predator scats may be biased by scat persistence. Investigators can minimize bias through frequent sampling and by accounting for local scat-degradation rates.