Scat removal: a source of bias in feces-related studies

1. What they did

Scats are cheap and easy to collect, and researchers use them to study diet, to tell which species are present, to estimate relative numbers and even to identify individuals from DNA. All of this assumes that the scats found are a fair sample of what animals left behind. The authors asked how often scats simply disappear, and what takes them.

The study ran through 2000 on Fort Riley, a United States Army training area in the Flint Hills of northeastern Kansas, in both forest and prairie.

• Tracking stations. They cleared circles 1 m across, smoothed the ground and sifted sand and soil over them so that the tracks of visiting animals would show clearly.
• Scats. Each station got one scat, chosen at random, from one of three sources: free-ranging coyotes (collected fresh along dirt roads on the base), captive coyotes fed dog food, and captive bobcats fed a commercial cat diet (both from zoos). For each type they set 20 stations at least 0.5 km apart, half in forest and half in prairie, plus 12 stations with no scat as controls. Scats were frozen and thawed 24 hours before use.
• Checking. They read the stations every 24 hours, recorded and identified the tracks of visitors where possible, noted whether the scat had gone or been disturbed, then smoothed the surface again. They kept checking until the scat was removed or broke down from rain, drying or wind, over more than 30 days in each season.
• Analysis. For each season they tested whether removal depended on habitat, on the type of scat, or on the two together.

2. What they found

How many scats disappeared

• Winter: 4 of 60 scats were removed (2 bobcat, 1 captive coyote, 1 free-ranging coyote), 3 of them in forest. Too few to test.
• Spring: 9 of 58 were removed, 6 of them coyote scats. Removal depended on the combination of scat type and habitat, but not on either alone.
• Summer: 30 of 60 (50%) were removed and appeared to have been eaten. Removal was higher in forest (67% of removals) than prairie (33%). Scat type made no difference.
• Autumn: 19 scats were removed, nearly three times more often in forest than in prairie; 15 of the 19 were coyote scats, although scat type alone was not significant.
• Scats went quickly: on average in about 1.5 days in summer and about 2.5 to 3 days in the other seasons (read from Figure 1).

Who visited

• Baited stations had 140 visits in winter, 125 in spring, 176 in summer and 271 in autumn. Control stations with no scat were hardly visited: one deer in spring and one rabbit each in summer and autumn.
• Opossums were the most frequent visitors where scats were removed: at 2 of the 3 forest removals in winter, 33% of removal stations in spring, half of them in summer and 37% in autumn.
• Coyotes, bobcats, raccoons, deer, rabbits, small rodents and insects were also present at stations where scats disappeared.

Other ways scats disappeared

• A bobcat buried the study scat and defecated on the station in spring; in autumn a bobcat did the same with a captive bobcat’s scat.
• At 2 spring stations, coyote tracks were present, the scat was gone and the station had been dug up and covered with fresh urine and feces.
• A woodrat removed a coyote scat and cached it at a nearby den.
• Heavy spring rains washed scats away more than once.

The species that removed a scat could be named for certain only when a single species’ tracks were on the station. Where several species had visited, the authors listed them all but could not say which took the scat. Leftover pieces often showed tooth marks and small fragments, suggesting the scat had been eaten.

Comparing diet studies

An older study of 1190 coyote stomachs from Kansas (1948 to 1957) found rabbits and hares in 57% and hoofed animals (mostly livestock) in 38%, against rodents in 36%. A later study of 1389 coyote scats (1990 to 1991) found rodents in 86%, rabbits and hares in only 20% and hoofed animals in about 6%. The authors suggest that animals selectively eating scats holding partly digested rabbit and deer could explain part of the difference, alongside changes in prey and differences in method.

3. What they concluded

• Eating feces and other behaviours that remove scats can bias studies based on scats. Removal rates like those at Fort Riley could distort diet studies if scat contents are taken to represent what animals ate.
• Scat counts used to show presence or relative abundance could also be biased by scats being eaten, cached or removed during territorial marking.
• Scats may be an important seasonal food for opossums, perhaps because of the protein-rich insects on them in summer, and may supplement the diet of other species. Scats from large prey may be richer in protein because large prey are less completely digested.
• Eating or carrying feces may spread parasites and disease, as with the raccoon roundworm and the Allegheny woodrat.
• More physiological and behavioural studies of coprophagy are needed.
• Cameras or video would have identified the animals more surely, but theft and vandalism risks on a busy military base ruled them out.

4. How tracking was or was not used

Tracking was the main way the researchers learned what visited each scat. Each station was a patch of smoothed, sifted sand and soil, read every 24 hours, and the tracks of visitors were recorded and identified to species where possible. All rodent tracks were pooled as “small rodent” and small bird tracks as “small bird”. Tracks were also used to recognise territorial behaviour (a bobcat burying a scat and defecating; coyotes digging up stations and leaving fresh urine and feces). Leftover scat fragments were examined for tooth marks.

5. Tracker or observer certification

Not reported. The paper does not say who read the tracking stations or how their track identification skills were checked. This matters because which animals removed scats rests entirely on track identification.

6. Other technologies, skills or innovations

• Tracking stations (smoothed, sifted plots of sand and soil, 1 m across) used with scats as the bait, rather than scent.
• Comparison of scats from captive animals on known diets with scats from wild coyotes.

7. Implications for conservation (from the paper)

• Scat-based diet studies and population surveys, widely used because they are cheap and non-invasive, may be biased when scats are removed selectively, and removal varies with season and habitat.
• Woodrats foraging in latrines and caching feces may face a higher risk of parasites, which has been linked to the decline of the Allegheny woodrat.