Comparing Morphological and Molecular Diet Analyses and Fecal DNA Sampling Protocols for a Terrestrial Carnivore

1. What they did

Scats are one of the main ways researchers learn what predators eat. The traditional method washes a scat and identifies the hair, bones and other hard parts left behind. A newer method looks for the prey’s DNA in the scat. The authors wanted to know which method finds prey more often in a land carnivore, and where on a scat a DNA sample should be taken.

• Collecting scats (13 January to 19 March 2014). Two researchers walked standardized transects set up for canid (dog family) research in the Utah desert. At each carnivore scat they took a small sample from the side, measured the scat, scored its freshness from 1 (very fresh) to 5 (very old) by smell and colour, and froze the rest.
• Confirming the species. DNA from the side sample confirmed which carnivore left each scat. Only coyote scats were used.
• Choosing the scats. They used 50 fresh or very fresh coyote scats of similar size: 25 in which the hard parts showed leporids (rabbits and hares) and 25 in which they did not.
• Sampling each scat in five ways. Besides the side sample, they took a sample from each end, one from the centre, and a mixed sample made of small amounts from all four places.
• Testing. Each sample was tested for the DNA of six rabbit and hare species. The hard parts of every scat were washed out and identified with reference collections and skull and hair keys.

2. What they found

Hard parts

• The scats averaged 15.45 cm long and 1.98 cm wide, with up to 3 segments; 11 were very fresh and 39 fresh.
• Hard parts showed leporids in 25 scats (50%), but could not tell which rabbit or hare species.

Prey DNA

• Using all five samples together, leporid DNA was found in 38 scats (76%). That included 15 of the 25 scats (60%) where hard parts showed no leporids, and 23 of the 25 (92%) where they did.
• Mountain or desert cottontail was found in 66% of scats, black-tailed jackrabbit in 14% and eastern cottontail in 4%. Pygmy rabbits and white-tailed jackrabbits were not detected.
• 8 scats held DNA from two leporid species. In 13 of the 38 positive scats (34%), the two ends gave different species, or only one end had prey DNA.

Where to sample

• Single samples found prey in 35% (ends, averaged), 38% (centre), 44% (side) and 46% (mixed sample) of scats. These differences were not significant.
• Every single sample did worse than all five together. The odds of detecting prey with all five were 3.67 times those with the mixed sample alone.
• Two samples together came close to all five: centre plus side 64%, mixed plus side 70%. Mixed plus side was significantly better than the mixed sample alone.

DNA versus hard parts

• With all five samples, DNA found prey in 26 percentage points more scats than hard parts (76% versus 50%), a significant difference. With mixed plus side it was 20 points more (70% versus 50%).
• With only one sample per scat, hard parts actually found prey slightly more often than DNA, though the differences were not significant.
• DNA identified the rabbit and hare species; hard parts could not.

3. What they concluded

• DNA analysis can find prey more often than hard-parts analysis, but prey DNA is not spread evenly through a scat, so sampling from more than one place on a scat can make a big difference.
• Taking a side sample plus a mixed sample is a cost-effective choice: it found only 6% fewer positives than all five samples. If only one sample is affordable, the mixed sample is best.
• Fully blending whole scats may detect even more prey, but destroys the hard parts needed for traditional analysis and may be impractical in the field.
• With only one DNA sample per scat, the two methods may give similar results at the population level. This differs from most earlier studies, which were mainly on marine predators. A pilot study comparing methods before a full study, and possibly a correction factor, would help.
• The choice of method also depends on the research question, budget and tolerance for missed prey. DNA work generally costs more, and it is hard to estimate diet proportions from it.

4. How tracking was or was not used

Scats were found by two researchers walking standardized transects set up for canid monitoring. The authors do not say how carnivore scats were recognized in the field; the species of each scat was confirmed by DNA. Freshness was scored from the scat’s smell and colour at collection. Prey were identified from hair and bones washed out of the scats, using reference collections and keys.

5. Tracker or observer certification

Not reported. Two researchers surveyed each transect; M. Smith and M. Melham are thanked for help with fieldwork. No observer skill or experience is described. This matters little to the findings, because every scat’s species was confirmed by DNA.

6. Other technologies, skills or innovations

• Fecal DNA preserved in DETS buffer (a dimethyl sulfoxide (DMSO) and salt solution), with clean-lab methods and negative and positive controls to guard against contamination.
• A mitochondrial DNA (mtDNA) test that tells apart six rabbit and hare species with one forward and five species-specific reverse primers, run by polymerase chain reaction (PCR).
• Scats washed in nylon bags in a standard washing machine to leave only the hard parts, then dried in an oven.

7. Implications for conservation (from the paper)

• Monitoring programs should aim to use analysis methods that detect the most prey.
• DNA methods offer an alternative to hard-parts analysis and may reduce missed prey when no hard parts are left in a scat.
• More comparisons of diet methods across a wide range of predators and places will give better information for wildlife management and conservation decisions.